Oxford Handbook of Engineering and Technology in The Classical World - 2008

THE OXFORD HANDBOOK OF
ENGINEERING
AND TECHNOLOGY
IN THE CLASSICAL
WORLD
This page intentionally left blank
ENGINEERING
AND
TECHNOLOGY
IN THE
CLASSICAL
WORLD
Edited by
JOHN PETER OLESON
OXFORD
UNIVERSITY PRESS
2008
OXFORD
UNIVERSITY PRESS
Oxford University Press, Inc., publishes works that further

Oxford University's objective of excellence
in research, scholarship, and education.
Oxford New York

Auckland Cape Town Dar es Salaam Hong Kong Karachi
Kuala Lumpur Madrid Melbourne Mexico City Nairobi
New Delhi Shanghai Taipei Toronto
With offices in
Argentina Austria Brazil Chile Czech Republic France Greece
Guatemala Hungary Italy Japan Poland Portugal Singapore
South Korea Switzerland Thailand Turkey Ukraine Vietnam
CopYTight © 2008 by Oxford University Press, Inc.
Published by Oxford University Press, Inc.

198 Madison Avenue, New York, New York 10016
www.oup.com
Oxford is a registered trademark of Oxford University Press
All rights reserved. No part of this publication may be reproduced,
stored in a retrieval system, or transmitted, in any form or by any means,
electronic, mechanical, photocopying, recording, or otherwise,
without the prior permission of Oxford University Press.
Library of Congress Cataloging-in-Publication Data

Oxford handbook of engineering and technology in the Classical world / edited by John Peter Oleson.
p.cm.
Includes bibliographical references and index.
ISBN 978-0-19-518731-1
1. Technology-Rome-History-Handbooks, manuals, etc. 2. Technology-Greece-History-
Handbooks, manuals, etc. 3. Engineering-Rome-History-Handbooks, manuals, etc.
4. Engineering-Greece-History-Handbooks, manuals, etc. I. Oleson, John Peter.
T16.O94 2007
609.38-dc22 2007010727
135798642
Printed in the United States of America

on acid-free paper
For Martha, Olaf, and Patience
"For where your treasure is, there will your heart be also" (Luke 12:34)
ACKNOWLEDGMENTS
The editor is deeply grateful to the scholars who agreed to take time from their
busy schedules of research and publication to contribute to this book. I believe that
our collaboration has been fruitful and that the total of our contributions far
exceeds the sum of the individual parts. I am also grateful to Professor Jamie A.
Cassels, vice president academic of the University of Victoria, and to Professor
Andrew Rippin, dean of humanities, for some teaching release and research
funding associated with the award of a Distinguished Professorship, and to my
colleagues in the department of Greek and Roman Studies for their explicit and
implicit support of my research endeavors. I am also very grateful to Derek Sou,
who provided admirably meticulous proofreading services during the preparation
of this volume, and to Milorad Nikolic, who translated chapter 12 from German
and read over some other materials. Ms. Martina Mussi assisted me in the
translation of chapter 22 from Italian, and I myself translated chapter 23 from the
French manuscript. I benefited from consultation concerning various matters with
my colleagues Dr. Gregory Rowe and Dr. Brendan Burke. I am also grateful to the
many members of the editorial staff at Oxford University Press who assisted me in
improving the manuscript and producing a useful and attractive book.
Several authors have asked that their acknowledgments be included in the
book, and I have listed these here, in order to avoid footnotes and addenda in the
individual chapters.
Dr. Kevin Greene acknowledges with gratitude that his chapters incorporate
research conducted during the tenure of a British Academy Research Readership
with support from the School of Historical Studies of Newcastle University. His
definitions of invention and innovation benefited from discussions with Professor
David Edgerton (Imperial College) and Claudia Durrwachter (Newcastle Uni-
versity).
Professor Andrew I. Wilson wishes to thank the Leverhulme Trust for the
award of a Philip Leverhulme Prize, which funded a period of research leave during
which he prepared three chapters for this book.
Dr. Lynne Lancaster expresses her gratitude for the comments and insights of
Fikret Yegiil on an early draft of the manuscript of her chapter, and to Sandra
Lucore for sharing her work on the bath building at Morgantina, which is the
subject of her doctoral dissertation at Bryn Mawr.
CONTENTS
Contributors, xiii
Abbreviations and Spelling Norms, xvii
Introduction, 3
John Peter Oleson
PART I SOURCES
1. Ancient Written Sources for Engineering and Technology, 15

Serafina Cuomo
2. Representations of Technical Processes, 35

Roger Ulrich
3. Historiography and Theoretical Approaches, 62

Kevin Greene
PART II PRIMARY, EXTRACTIVE TECHNOLOGIES
4. Mining and Metallurgy, 93

Paul T. Craddock
5. Quarrying and Stoneworking, 121

J. Clayton Fant
6. Sources of Energy and Exploitation of Power, 136

6rjan Wikander
7. Greek and Roman Agriculture, 158

Evi Margaritis and Martin K. Jones
8. Animal Husbandry, Hunting, Fishing, and Fish Production, 175

Geoffrey Kron
X CONTENTS
PART III ENGINEERING AND COMPLEX MACHINES
9. Greek Engineering and Construction, 225

Fredrick A. Cooper
10. Roman Engineering and Construction, 256

Lynne Lancaster
11. Hydraulic Engineering and Water Supply, 285

Andrew I. Wilson
12. Tunnels and Canals, 319

Klaus Grewe
13. Machines in Greek and Roman Technology, 337

Andrew I. Wilson
PART IV SECONDARY PROCESSES AND MANUFACTURING
14. Food Processing and Preparation, 369

Robert I. Curtis
15. Large-Scale Manufacturing, Standardization, and Trade, 393

Andrew I. Wilson
16. Metalworking and Tools, 418

Carol Mattusch
17. Woodworking, 439

Roger B. Ulrich
18. Textile Production, 465

John P. Wild
19. Tanning and Leather, 483

Carol van Driel-Murray
20. Ceramic Production, 496

Mark Jackson and Kevin Greene
21. Glass Production, 520

E. Marianne Stern
CONTENTS X1
PART V TECHNOLOGIES OF MOVEMENT AND TRANSPORT
22. Land Transport, Part 1: Roads and Bridges, 551

Lorenzo Quilici
23. Land Transport, Part 2: Riding, Harnesses, and Vehicles, 580

Georges Raepsaet
24. Sea Transport, Part 1: Ships and Navigation, 606

Sean McGrail
25. Sea Transport, Part 2: Harbors, 638

David ]. Blackman
PART VI TECHNOLOGIES OF DEATH
26. Greek Warfare and Fortification, 673

Philip de Souza
27. Roman Warfare and Fortification, 691

Gwyn Davies
PART VII TECHNOLOGIES OF THE MIND
28. Information Technologies: Writing, Book Production,

and the Role of Literacy, 715
Willy Clarysse and Katelijn Vandorpe
29. Timekeeping, 740

Robert Hannah
30. Technologies of Calculation, 759

Part 1: Weights and Measures (Charlotte Wikander), 759
Part 2: Coinage (Andrew Meadows), 769
Part 3: Practical Mathematics (Karin Tybjerg), 777
31. Gadgets and Scientific Instruments, 785

6rjan Wikander
32. Inventors, Invention, and Attitudes toward Innovation, 800

Kevin Greene
Xll CONTENTS
PART VIII ANCIENT TECHNOLOGIES IN THE MODERN WORLD
33. Expanding Ethnoarchaeology: Historical Evidence

and Model-Building in the Study of Technological Change, 821
Michael B. Schiffer
Index, 837
CONTRIBUTORS
Prof. David Blackman Dr. Gwyn Davies

Classics Department, Oxford Department of History
University Florida International University
Old Boys' School, George St. University Park (DM-397)
Oxford OX1 2RL, UK Miami, FL 33199, USA
Prof. Willy Clarysse Dr. Philip de Souza

Katholieke Universiteit Leuven Department of Classics
Afd. Geschiedenis vd Oudheid University College Dublin
Blijde-Inkomststraat 21 Belfield, Dublin 4, Ireland
B-3000 Leuven, Belgium
Dr. Carol van Driel-Murray
Prof. Fredrick A. Cooper Amsterdam Archaeological Centre
Department of Art History Nieuwe Prinsengracht 130
Heller Hall, 271 19th St. S 1018 VZ Amsterdam, Netherlands
University of Minnesota Prof. J. Clayton Fant
Minneapolis, MN 55455-0121, Department of Classical Studies
USA 2376 Olin Hall
Dr. Paul T. Craddock University of Akron
Department of Scientific Research Akron, OH 44325-1910, USA
The British Museum Dr. Kevin Greene
Great Russell St. School of Historical Studies
London WC1B 3DG, UK The University
Dr. Serafina Cuomo Newcastle upon Tyne NEl 7RU, UK
Centre for the Study of Science, Dr. Klaus Grewe
Technology, and Medicine Landschaftsverband Rheinland,
Sherfield Building 447 A Rheinisches Amt Fiir
Imperial College Bodendenkmalpflege
London SW7 2AZ, UK Endenicher StraBe 133
D-53115 Bonn, Germany
Prof. Robert I. Curtis
Department of Classics, Prof. Robert Hannah
Park Hall Department of Classics
University of Georgia Box 56, University of Otago
Athens, GA 30602-6203, USA Dunedin, New Zealand
XIV CONTRIBUTORS
Dr. Mark Jackson Prof. John Peter Oleson

School of Historical Studies Department of Greek and Roman
The University Studies
Newcastle upon Tyne NEl 7RU, Box 3045, University of Victoria
UK Victoria, B.C. V8W 3P4, Canada
Prof. Martin Kenneth Jones Prof. Lorenzo Quilici

Department of Archaeology Istituto di archeologia
University of Cambridge Universita di Bologna
Downing St. Piazza S. Giovanni in Monte, 2
Cambridge CB2 3DZ, UK 40124 Bologna, Italy
Dr. Geoffrey Kron Prof. Georges Raepsaet

Department of Greek and Roman Universite Libre de Bruxelles
Studies Histoire de l'art et archeologie
Box 3045, University of Victoria Batiment NA, bureau NA.3.107
Victoria, B.C. V8W 3P4, ULB CP175, Avenue F.D. Roosevelt 50,
Canada 1050 Bruxelles, Belgium
Prof. Michael Brian Schiffer

Dr. Lynne Lancaster
Department of Anthropology
Classics Department
University of Arizona
Ellis Hall
1009 E. South Campus Drive, Building
Ohio University
30A
Athens, OH 45701-2979, USA
Tucson, AZ 85721-0030, USA
Dr. Evi Margaritis
Dr. E. Marianne Stern
Department of Archaeology
Willibrorduslaan 87
University of Cambridge
1216 PA Hilversum, Netherlands
Downing St.
Cambridge CB2 3DZ, UK Dr. Karin Tybjerg
Department of History and Philoso-
Prof. Carol C. Mattusch phy of Science
Department of Art History University of Cambridge
George Mason University Free School Lane
4400 University Drive Cambridge CB2 3RH, UK
Fairfax, VA 22030, USA
Dr. Roger Ulrich
Prof. Sean McGrail Department of Classics
Bridge Cottage, Chilmark Dartmouth College
Salisbury SP3 5AU, UK Hanover, NH 03755-3526, USA
Dr. Andrew Meadows Dr. Katelijn Vandorp

The American Numismatic Katholieke Universiteit Leuven
Society Afd. Geschiedenis vd Oudheid
96 Fulton St. Blijde-Inkomststraat 21
New York, N.Y. 10038, USA B-3000 Leuven, Belgium
CONTRIBUTORS XV
Prof. Charlotte Wikander Prof. John Peter Wild

Lunds Universitet, Klassiska Archaeology Department
Institutionen The University of Manchester
Solvegatan, 2 Oxford Rd.
22362 Lund, Sweden Manchester M13 9PL, UK
Prof. Orjan Wikander Prof. Andrew Wilson
Lunds Universitet, Klassiska Institute of Archaeology
Institutionen 36 Beaumont St.
Solvegatan, 2 Oxford OX1 ZPG, UK
22362 Lund, Sweden
ABBREVIATIONS AND
SPELLING NORMS
AN attempt has been made throughout the book to address the perennial problem
of the transliteration of Greek personal and place names, although no single so-
lution is ideal or even workable. With the invaluable assistance of my Oxford
University Press copyeditor, Mary Bellino, I have tried for uniformity among
chapters, for clarity and the familiarity of the result to a nonspecialist reader. For
the most part we have opted for Latinized forms. The discussion of technical
processes also frequently brings up the tricky question of hyphenation, an aspect of
modern English that continues to evolve. "Glassblowing," "glass-blowing," and
"glass blowing" are all acceptable in English, as are "watermill" and "water-mill,"
and-perhaps-"water mill." Usage is tending toward, for example, "glasswork-
ing" and "metalworking." The policy of the press is to follow Webster's Collegiate
Dictionary, nth ed., for spelling and hyphenation, followed by The Chicago Manual
of Style, but the result is not always completely logical.
In order to make it easier for nonspecialists to follow up on the citations of
Greek and Roman literature, a particularly important source for our under-
standing of ancient technology, I have not abbreviated the names of classical
authors (except in tables, for the sake of saving space); the titles of the individual
works have, however, been abbreviated according to the system used by the Oxford
Classical Dictionary, 3rd ed. According to this system, the titles of some single
surviving works, such as Herodotus' Histories, are not used, while others are cited,
such as Vitruvius, De arch. None of this, of course, has much value as a logical
system, but consistency with standard systems can assist the reader, and the Oxford
Classical Dictionary is easily available for reference. Where no abbreviations are
suggested in the OCD, abbreviations of the works of Greek authors have been taken
from the Liddell and Scott Greek-English Lexicon (9th ed., rev. by H. Stuart Jones;
suppl. by E. A. Barber; Oxford: Oxford University Press, 1968) and of Latin authors
from the Oxford Latin Dictionary. I have not abbreviated periodical titles. I asked
contributors to avoid the use of footnotes, with the result that there are occasional
clusters of references in the text. But for the most part this approach makes for a
smoother presentation. Most of the translations from Greek and Latin texts not
otherwise attributed have been taken from J. W. Humphrey, J.P. Oleson, and A. N.
Sherwood, Greek and Roman Technology: A Sourcebook (New York: Routledge,
1998), with the kind permission of my co-authors.
XVlll ABBREVIATIONS AND SPELLING NORMS
The following abbreviations appear in the text for corpora, encyclopedias, and
other commonly cited reference works, for papyri, and for a small number of
technical terms.
AB L 'Annee Epigraphique.
ATTA Atlante tematico di topografia antica. Rome: L'Erma di Bretsch-
neider.
CIL Corpus Inscriptionum Latinarum (1863-).
DarSag C. Daremberg and E. Saglio (eds.) 1877-1919. Dictionnaire des
antiquites grecques et romaines d' apres les textes et les monuments. 9
vols. Paris: Hachette.
HS sestertius, sestertii
JG Inscriptiones Graecae (1873-).
ILLRP E. Degrassi (ed.) 1963-1965. Inscriptiones Latinae Liberae Rei Pub-
licae. Vol. i2 (1965); vol. 2 (1963). Florence: La Nuova Italia.
OCD S. Hornblower and A. Spawforth (eds.) 1996. Oxford Classical
Dictionary. 3rd ed. Oxford: Oxford University Press.
OLD P. G. W. Glare (ed.) 1982. Oxford Latin Dictionary. Oxford: Oxford
University Press.
P. Cairo Zeno C. C. Edgar 1925-1940. Zenon Papyri: Catalogue general des anti-
quites egyptiennes du Musee du Caire. 5 vols. Cairo: Imprimerie de
l' Institut Fran~ais dArcheologie Orientale.
P. Lille P. Jouguet, P. Collart, and J. Lesquier (eds.) 1928. Papyrus grecs.
Institute papyrologique de l'Universite de Lille. Paris: Leroux.
P. Petrie J.P. Mahaffy (ed.) 1893. The Flinders Petrie Papyri. Vol. 2. Dublin:
Academy House. J.P. Mahaffy and J. G. Smyly 1905. The Flinders
Petrie Papyri. Vol. 3. Dublin: Academy House.
PG J.-P. Migne (ed.) 1857-1887. Patrologiae Cursus Completus: Series
Graeca. Paris: P. Geuthner.
RE A. Pauly, G. Wissowa, and W. Kroll (eds.) 1893-1978. Real-
Encyclopiidie der klassischen Altertumswissenschaft. Stuttgart:
A. Druckenmiiller.
SEG Supplementum Epigraphicum Graecum (1923-).
SIG W. Dittenberger 1915-24. Sylloge inscriptionum graecarum. 3rd ed.
4 vols. Leipzig: Hirzel.
Tab. Vindol M.A. Speidel (ed.) 1996. Die romischen Schreibtafeln von Vindo-
nissa. Veroffentlichungen der Gesellschaft Pro Vindonissa 12.
Brugg: Gesellschaft Pro Vindonissa.
ENGINEERING
AND TECHNOLOGY
IN THE CLASSICAL
WORLD
INTRODUCTION
JOHN PETER OLESON
THE GREEK AND ROMAN VIEW

OF TECHNOLOGY
THis handbook is designed to survey the role of technology in the Greek and Roman
cultures and their respective technological accomplishments, from approximately
the eighth century B.C. through the fifth century A.D. Such an approach, of course,
leaves out the remarkable technological achievements of the Bronze Age and
previous periods, but their inclusion would have required a second volume. In any
case, the Greco-Roman accomplishments in technology reveal a greater unity than
those of the earlier cultures of the Near East or the Mediterranean, reflecting the
shared attitudes and experiences of the Greek and Roman cultures themselves. The
Greeks and Romans appreciated and occasionally marveled at the structures and
artifacts that had survived from the preclassical cultures, and they understood that
human technologies had even older roots. Ancient anthropologists, like their mod-
ern counterparts, realized that human activities intentionally altering the envi-
ronment to ensure survival or simply a more pleasurable lifestyle-what we call
technological activities-were inextricably bound up with the roots of human
biology. In the fifth century B.C., the philosopher Anaxagoras (Fr. 59.A.102, Diels-
Kranz) stated that "the human race is the wisest of all living creatures because it
possesses hands." A century later, Aristotle (De an. 3.8.432a) was more explicit:
"The hand is the tool that makes and uses tools." Even the stolid Roman architect
Vitruvius (De arch. 2.1.2) is almost poetic in his account of the capabilities enjoyed
by humans because of their very form and nature: "They possessed by nature a gift
beyond all other creatures: they walked not stooped downward, but erect, so they
4 INTRODUCTION
viewed the magnificence of the world and stars. Moreover they could easily handle
whatever they wished with their hands and fingers."
Lucretius (5.925-1025) attributes some of this human vigor to the survival of
the fittest, while Pliny (like all the Roman authors, paraphrasing Greek philosophi-
cal commonplaces) makes a virtue of the helplessness of the infant, naked and
without instinctive survival behaviors (HN 7, preface 1-4):
Humans alone of all animals Nature clothes in borrowed resources ... only a
human on the day of birth does she cast down naked onto the bare ground,
immediately to cry and wail. ... All the rest of the animals know their own nat-
ural abilities, some use their agility, others swift flight, others swim; the human
knows nothing unless taught: not to speak, nor to walk, nor to eat, and in short,
he knows nothing by natural instinct other than to cry.
In addition to recognizing the age-old biological-perhaps even evolutionary-

roots of human technologies, both Greek and Roman intellectuals appreciated
that an enormous period of time must have passed as human communities
formed, spread out across the known world, and mastered an increasing array of
materials and technologies. Their theories of social evolution can easily be com-
pared with nineteenth-century theories of "progress" from Paleolithic to Meso-
lithic to Neolithic manners of living, a neat sequence that has only recently come
into question: Hesiod (Op. 107-78), Lucretius (5.925-1025), Vergil (G. 1.121-46),
Vitruvius (De arch. 2.1.1-7), for example (cf. Cole 1967; many of the passages are
collected and translated in Humphrey et al. 1998: 1-20). Chreia (necessity; [Aris-
totle], Mech. 10-25.847a; Diodorus 1.8.7-9) drove humans to create their technai
(arts, technologies) and the mechanai (devices) to assist with them, while observa-
tion of nature provided models (Lucretius 5.1361-69; Seneca, Ep. 90). In meta-
phorical terms, the Greeks counted fire and the various technologies developed by
humans as the "gifts" of the mythological figure Prometheus, whose name soun-
ded something like "Forethought" to a Greek-speaker (Hesiod, Theog. 565-66, Op.
42-53; Aeschylus, PV 442-506). Plato (Prt. 32oc-322d), however, attributes only fire
and the arts in general to Prometheus, and the rest to human inventiveness, while
Sophocles (Ant. 332-72) composed a splendid "hymn" to the independent accom-
plishments of humans as conscious entities (both passages are quoted in chapter 1).
Given the recent flowering of interest in an environmental basis for the great
disparities in technological achievement among human communities in various
parts of the world (Diamond 1997), it is striking to note that Greco-Roman culture
had already hit upon an adumbration of these principles. A tribe of one-eyed giants,
the Cyclopes, lived on a fertile island with a beneficent climate that provided wheat,
barley, and grapes (the basis for the ancient Greek diet) spontaneously, without any
need for plowing or sowing. As a result, Homer concludes (Od. 9.105-31), they had no
need for assemblies or laws, shipwrights, or other technologies, each family living
apart in caves and herding flocks of sheep in a modified hunting-gathering culture.
Greek poets and playwrights such as Hesiod, Homer, Sophocles, and Aeschylus
marveled at the capacity of humans for technological advance and the techniques
INTRODUCTION 5
they had developed for managing the earth and its resources. Greek philosophers
such as Plato, Aristotle, and Posidonius debated the role of technology in society,
generally reflecting an intellectual posture that thought and dialectic were preferable
to applied research and the arts and crafts (e.g., Plato, Resp. 7.53oe-531a; Plutarch,
Mor. 8.2.718e). Plato (e.g., Resp. 2.368e-374e, Leg. 8.846d-847a) proposed-more as
a metaphor than as a practical suggestion-that society should segregate its
thinkers, soldiers, and producers, and that the simple rhythms of an agricultural
community were preferable to the "diseased" complexity of urban living. In his
Politics, Aristotle considered in somewhat more practical fashion the role of slave
labor, craftsmen, architecture, money, and trade in human society, but he reflects
the same intellectual posturing as Plato (Aristotle, Pol. 8.2.1, 1337b):
It is clear, therefore, that the young must be taught those of the useful arts that are
absolutely necessary, but not all of them. It is obvious that the liberal arts should be
kept separate from those that are not liberal, and that they must participate in
such of the useful arts as will not make the participant vulgar [banausos]. A task or
an art or a science must be considered banausic if it makes the body or soul or
mind of free men useless for the practice and application of virtue. For this rea-
son we term "banausic" those crafts that make the condition of the body worse,
and the workshops where wages are earned, for they leave the mind preoccu-
pied and debased.
The adjective banausos, which could also mean "vulgar" or "in bad taste," de-
rived from baunos, the workman's forge or furnace. The term comes up frequently
in theoretical or philosophical discussion, for example Xenophon's Oeconomicus
(On Household Management, 4.2-3):
To be sure, the so-called banausic arts are spoken against and quite rightly held in
contempt in our states, for they ruin the bodies of those practicing them and
those who supervise, forcing them to sit still and pass their time indoors-some
even to spend the day at a fire. As their bodies are softened, so too their minds
become much more sickly. In addition, the so-called banausic arts leave no leisure
time for paying attention to one's friends or state, so that the persons who prac-
tice them have the reputation of treating their friends badly and being poor
defenders of their homeland. In some states, particularly those with a warlike
reputation, it is forbidden for any citizen to practice the banausic arts.
Roman intellectuals occasionally assumed the same attitude; for example, Ci-
cero (Off. 1.42): "All craftsmen spend their time in vulgar occupations, for no
workshop can have anything liberal about it." (Many of the relevant literary pas-
sages are collected in Humphrey et al. 1998: 579-97.) Although sentiments such as
these have found expression to some degree in most human societies-even in
those, such as nineteenth-century England, that are noted for rapid technological
advance-many twentieth-century scholars hit upon this banausic prejudice as an
"explanation" for a perceived blockage of technological innovation in the Greco-
Roman world. The presence of slave labor was felt to be a related, concomitant
factor. The historian Moses Finley was the most prominent exponent of this now
6 INTRODUCTION
discredited interpretation, and there is an extensive literature on both sides of the

argument (see esp. Greene 2000, and chapters 2, 3, and 23 in this volume).
As the material in this book will make clear, the classical world was marked by
remarkable technological advances in many areas, often fostered by the elite. A
technological literature composed of both sophisticated compendia and workshop
manuals did exist, although much of it has been lost, and both inscriptions and
visual representations show that craftsmen and craftswomen were proud of their
work and their products (see chapters 1 and 2). Over half of the male names on
tomb inscriptions at Korykos in Rough Cilicia are accompanied by an occupation,
and the number of occupations totals no. At least 85 occupations are named in
various inscriptions and graffiti at Pompeii, over 200 on tombstones in Rome, and
Artemidorus, in his Onirocritica, refers to 264 (Hopkins 1978: 72). These numbers
imply self-identification and a certain amount of pride among craftsmen and the
service professions. One goal I had in editing this book was to help put an end to
the myth of a "technological blockage" in the classical cultures.
ORGANIZATION OF THE MATERIAL
This book is not intended to be a compendium of all the technological procedures,

devices, and machines in use in the classical world. I have been fortunate to be able
to enlist 31 expert scholars from 9 countries as contributors, and I asked each of
them to look at the accomplishments of the classical cultures in their particular
technology or family of technologies, or at an issue in the history of ancient tech-
nology, from primarily an analytical rather than a descriptive point of view. The
objective was the creation of a critical summation of our present knowledge of the
Greek and Roman accomplishments in technology and engineering, and the evo-
lution of the technical capabilities of these cultures over the chronological period.
Each chapter was designed to review the issues surrounding that topic and the
recent scholarly contributions, and then to define the capacities and accomplish-
ments of the technology in the context of the society that used it, the available
"technological shelf," and the resources consumed. Since these studies are not sim-
ply descriptive in character, but introduce and synthesize the results of excavation
or specialized research, the volume is not as heavily illustrated as might otherwise
have been the case. Readers who want more information can easily follow up by
using the full bibliographies appended to each chapter.
In working up the outline of the book, I had to make difficult decisions about
the selection and organization of the topics, the relative length of the chapters and
sections, and the allocation of a restricted number of illustrations. For a variety of
reasons I gave particular prominence to mining and metallurgy, Roman engineer-
ing and architecture, hydraulic engineering, land and sea transport, and informa-
tion technologies. The organization of the chapters into eight sections represents a
first sorting of an enormously varied and complex array of data and interpretation,
INTRODUCTION 7
and I hope it will be of particular use to those first approaching the subject. The
first section deals with the ancient written sources for Greek and Roman tech-
nology, the visual representations of technological activities, and the historiography
of the subject. These three chapters set the stage for consideration of the individual
technologies in separate chapters that assemble the evidence from archaeological
remains as well as the literary evidence and visual representations. The archaeolog-
ical evidence, of course, consists of raw materials, tools and equipment, workshops,
by-products, and the primary products. There is a natural hierarchy of technol-
ogies, not always directly linked to the hierarchy of human needs, but generally
tracking human needs closely. The primary, extractive technologies, those activities
that wrest raw materials and foodstuffs or energy directly from the environment-
mining, metallurgy, quarrying and stoneworking, agriculture, animal husbandry,
hunting, and fishing-appear first, in the second section. Not surprisingly, many of
these technologies had already reached a significant stage of development by the
Middle or Late Bronze Age.
The third section includes chapters dealing with engineering and complex ma-
chines, representing dramatic human intervention to shape the physical world to
human needs: Greek and Roman engineering and architecture, hydraulic engineer-
ing and water-supply systems, the engineering of tunnels and canals, and heavy
machinery. The fourth section concerns secondary processing of the raw materials
accounted for in the second: food processing, metalworking, woodworking, the
production of textiles, leather, ceramics, and glass, and the application of the chem-
ical substances that assisted many of these secondary technologies.
The remainder of the book deals with technologies that have a more complex
relationship with their socioeconomic context and generally involve a variety of
motivations and technologies beyond the extraction or processing of raw materials.
The fifth section includes chapters dealing with the infrastructure for land and sea
transport, and with riding, harnesses, vehicles, ships, and navigation. These tech-
nologies represent complex responses to a variety of human needs; ships, for ex-
ample, were the largest machines known to the classical world. The movement of
persons and goods made possible by these technologies forms the basis for the
striking success of Greco-Roman urbanism and the Roman imperial system. The
sixth section, which I have titled "Technologies of Death," presents warfare and
fortification, the most innovative and pervasive human technology from at least the
Early Bronze Age through the present day, for reasons that are all too obvious (cf.
Plato, Resp. 2.373d-e). Ironically, until the invention of nuclear weapons this was
the ultimate "labor-saving" technology: groups that excelled at making war reaped
the labor and took the property of those they defeated. Despite the importance of
the subject and the richness of the literary and archaeological evidence, these two
chapters are relatively short, because there is already so much published literature
available concerning Greek and Roman methods of warfare and fortification.
"Technologies of the Mind," the seventh section, incorporates chapters pre-
senting technologies that have been crucial to the sophisticated development of
human societies, but which deal with abstract ideas rather than making a direct
impression on the physical world: recording and reading information in written
form, measuring the passage of time, measuring the weight, size, or number of
8 INTRODUCTION
objects or substances, and accounting for the value and numerical relationships
among these objects or substances. Another chapter presents the gadgets and
scientific instruments of the Greco-Roman world, devices that represent a telling
interface between the world of theoretical thought and mechanical application. A
related chapter evaluates the attitude in the classical cultures toward inventors,
invention, innovation, and technology.
The last chapter in the book stands on its own, but I hope it will help to stim-
ulate a more theoretical approach to the study of technology in the classical world.
For most of the twentieth century, classical archaeologists and the historians of
Greek and Roman culture approached their subject matter very differently than did
"prehistoric archaeologists," "historical archaeologists," "New World archaeolo-
gists," or "processual archaeologists," to use only a few of the terms applied to those
outside the classical field. The reasons for this divide are numerous and complex,
and mostly irrelevant to the available archaeological data; in any case the argument
has now moved on to "post-processual" approaches to archaeology and history
(Trigger 2006). Beginning in the 1990s, many archaeologists of the classical cultures
began to make use of the theoretical tools developed by archaeologists and histo-
rians outside their own area. This shift of focus, which has naturally also affected the
study of engineering and technology in the classical cultures, has lead to many im-
portant new insights. The concomitant widespread abandonment of training in the
Greek and Latin languages is very unfortunate, but may reverse itself in the future.
The vocabulary and procedures of the "expanded ethnoarchaeology" presented
by Michael Schiffer in chapter 33 may seem foreign to some classical archae-
ologists and historians of ancient technology, but in fact he simply makes explicit
the methods of research that many of us have already used to one degree or an-
other: "If ethnoarchaeology is the study of general relationships between activi-
ties and artifacts when strong evidence is available on both, then in an expanded
ethnoarchaeology researchers can make use of evidence from the historical re-
cord as well." The relevance of this approach to the rich array of evidence for the
technologies of the classical cultures is obvious. Schiffer's discussion of differential
adoption, invention cascades, and the use of a performance matrix to evaluate the
life history of complex technological systems may help to stimulate productive
reevaluation of such Greco-Roman technologies as textile manufacture, glass-
working, ship design, and even construction.
THE FUTURE OF PAST TECHNOLOGIES
In addition to progress in the development of explicit theory, what does the future
hold for the study of ancient technology, in particular that of the classical cultures?
Archaeological excavation has become increasingly expensive and difficult, and the
INTRODUCTION 9
archaeological heritage has been badly diminished by development and illicit ex-
cavation. Nevertheless, remarkable discoveries continue to appear, in part because
of better techniques of excavation and analysis. Ancient representations of tech-
nological activities or machinery continue to turn up, sometimes with very surpris-
ing new information, such as the detailed representation of a pair of water-wheel-
driven saws on a sarcophagus at Pammukale-long exposed to view, but only
recognized in 2006 (Ritti et al. 2007). Ancient workshops and production sites are
now more readily recognized for what they are, and the animate and inanimate
materials, products, or by-products can be subjected to increasingly sophisticated
analytical procedures. Ceramic clay, stone, and metals can be traced to their sources;
bones, phytoliths, and microscopic traces of food residue provide detailed infor-
mation about diet; the analysis of ancient DNA, still in its rudimentary stage, has
already providing astonishing information about the interrelationships of human
groups and the domestication of plants and animals (Jones 2002). The application
of three-dimensional X-ray microfocus computed tomography has extracted im-
portant new data from the corroded remains of the Antikythera Mechanism (Freeth
et al. 2006). Archaeological survey, on the ground or even from space, continues to
reveal important remains relevant to the history of ancient technology without
excavation: the sites of previously unrecognized Roman water-mills, for example
(Wikander 1985), or Roman fortifications (Kennedy and Bewley 2004). Survey in
the deep Mediterranean by means of submersibles or remotely operated vehicles
has revealed a rich supply of well-preserved shipwrecks of all periods (McCann and
Oleson 2004).
Many technologies that were ubiquitous and important in the classical world
have not received commensurate analytical attention: basketry, for example, or ap-
plied chemistry, or textile and leather production. In addition, quantitative ap-
proaches such as product life-cycle analysis should be applied to many of the
ancient manufactured products. Delaine (1997) has reached remarkable conclu-
sions through calculation of the amount of labor and materials required for the
construction of a major imperial building, and the hypothetical process and sched-
ule of completion. This approach should be applied to other regions and periods.
Experiments in the replication of ancient artifacts and materials (e.g., Lierke 1999,
glass), structures (Oleson et al. 2006), and vehicles or boats (Raepsaet 2002; Mor-
rison et al. 2000) also provide scope for further inquiry.
Finally, there is more work to be done in the analysis of ancient literary sources.
Even familiar passages in Pliny the Elder's Natural History that concern technical
processes or attitudes toward technology require further interpretation or decod-
ing. Lewis (1997, 2001) has made remarkable discoveries by carefully mining the
surviving Arabic translations of Greek technical writers whose original works are
lost. The papyri from Egypt continue to require transcription and analysis, and
new methods of imaging have made possible the transcription of previously il-
legible papyrus documents. The digitization of nearly all the works of the Greek
and Latin authors now allows rapid and comprehensive searching for key words
and phrases, and the enormous mass of raw data and synthetic studies on the
10 INTRODUCTION
Internet make it possible for the historian of ancient technology to peruse a breadth
of material unthinkable only five years ago.
Together, all of these resources will allow the continued reconstruction of
Greek and Roman technological procedures and accomplishments, as well as
ongoing analysis of the relationship between technology and the ancient economy
and ancient society. Even our understanding of such an important topic as the role
of women in ancient technological processes is still in its infancy (cf. Kampen 1982;
Stern 1997; Glazebrook 2005). Clearly, there is a bright future for the study of
ancient technology.
REFERENCES
Cole, T. 1967. Democritus and the sources of Greek anthropology. Chapel Hill, NC: Press of
Western Reserve University.
Delaine, J. 1997. The Baths of Caracalla: A study in the design, construction, and economics of
large-scale building projects in Imperial Rome. Journal of Roman Archaeology Suppl. 25.
Portsmouth, RI: JRA.
Diamond, J. 1997. Guns, germs, and steel: The fates of human societies. New York: W.W.
Norton.
Freeth, T., Y. Bitsakis, X. Moussas, J. H. Seiradakis, A. Tselikas, E. Magkou, M. Zafeir-
opoulou, R. Hadland, D. Bate, A. Ramsey, M. Allen, A. Crawley, P. Hockley,
T. Malzbender, D. Gelb, W. Ambrisco, and M. G. Edmunds. 2006. "Decoding the
ancient Greek astronomical calculator known as the Antikythera Mechanism," Nature
444 (30 November 2006): 587-91.
Glazebrook, A. 2005. "Reading women: Book rolls on Attic vases," Mouseion 5: 1-46.
Greene, K. 2000. "Technological innovation and economic progress in the ancient world:
M. I. Finley reconsidered," Economic History Review 53: 29-59.
Hopkins, K. 1978. "Economic growth and towns in classical antiquity," in P. Abrams and
E. A. Wrigley (eds.), Towns in society: Essays in economic history and historical sociology.
Cambridge: Cambridge University Press.
Humphrey, J. W., J.P. Oleson, and A. N. Sherwood 1998. Greek and Roman technology:
A sourcebook. London: Routledge.
Jones, M. 2002. The molecule hunt: Archaeology and the search for ancient DNA. New York:
Arcade Publishers.
Kampen, N. B. 1982. "Social status and gender in Roman art: The case of the saleswoman,"
in N. Broude and M. D. Garrard (eds.), Feminism and art history. New York: Harper
and Row, 60-77.
Kennedy, D., and R. Bewley 2004. Ancient Jordan from the air. London: CBRL.
Lewis, M. J. T. 1997. Millstone and hammer: The origins of water power. Hull: University
of Hull Press.
Lewis, M. J. T. 2001. Surveying instruments of Greece and Rome. Cambridge: Cambridge
University Press.
Lierke, R. 1999. Antike Glastopferei: Bin vergessenes Kapitel der Glasgeschichte. Mainz:
Philipp von Zabern.
INTRODUCTION 11
Mccann, A. M., and J. P. Oleson 2004. Deep-water shipwrecks off Skerki Bank: The 1997
survey. Journal of Roman Archaeology Suppl. 58. Portsmouth, RI: JRA.
Morrison, J. S., J.E. Coates, and N. B. Rankov 2000. The Athenian trireme: The history and
reconstruction of an ancient Greek warship. 2nd ed. Cambridge: Cambridge University
Press.
Oleson, J.P., C. Brandon, L. Bottalico, R. Cucitorre, E. Gotti, and R. L. Hohlfelder 2006.
"Reproducing a Roman maritime structure with Vitruvian Pozzolanic concrete,"
Journal of Roman Archaeology 19 (2006): 29-52.
Raepsaet, G. 2002. Attelages et techniques de transport dans le monde greco-romain. Brux-
elles: Timperman.
Ritti, T., K. Grewe, and P. Kessner 2007. "Stridentes trahens per levia Marmora serras: A
relief of a water-powered stone saw mill on a sarcophagus at Hierapolis of Phrygia,"
forthcoming in Journal of Roman Archaeology 20.
Stern, E. M. 1997. "Neikais: A woman glassblower of the first century AD?" in G. Erath et al.
(eds.), Komos: Festschrift T. Lorenz. Vienna: Phoibos, 129-32.
Trigger, B. G. 2006. A history of archaeological thought. 2nd ed. New York: Cambridge
University Press.
Wikander, 6. 1985. "Archaeological evidence for early water-mills-An interim report,"
History of Technology 10: 151-79.
PART I
SOURCES
CHAPTER 1
ANCIENT WRITTEN
SOURCES FOR
ENGINEERING AND
TECHNOLOGY
SERAFINA CUOMO
THERE are four types of written evidence for engineering and technology in Greek
and Roman antiquity: texts dealing solely or mainly with technical knowledge and/
or written by people who identify themselves as technical practitioners; nontech-
nical texts, that is, any other text that provides us with information about technical
knowledge and its practitioners but whose main aim is not the communication of
technical knowledge and whose author's primary identification is not as technical
practitioner; inscriptions recording technical activities or providing information
about the lives of technical practitioners; and papyri with the same content. There
are several published collections of translations of ancient sources relevant to
engineering and technology. The only relatively comprehensive survey is Humphrey
et al. (1998), but some more specialized sourcebooks are available: Cohen and
Drabkin (1948), J.B. Campbell (1994), Sage (1996), Irby-Massie and Keyser (2002).
The index of 138 ancient authors and inscriptional sources quoted in Humphrey et
al. (1998: 601-12) is striking evidence for the pervasive presence of technological
information in Greek and Latin literature. Only about eight Greek and ten Latin
authors, however, wrote works that survive (at least in part) and can be considered
handbooks of a particular technology or group of technologies: Aeneas Tacticus
(military), the anonymous author of De rebus bellicis (military) and Periplus Maris
Erythraei (navigation), Apicius (food preparation), Biton (siege engines), Cato
16 SOURCES
(agriculture), Columella (agriculture), Dioscorides (pharmacology), Frontinus

(water supply), Hero (mechanics, hydraulics, surveying, siege engines), Hesiod
(agriculture), Hyginus (military), Palladius (agriculture, veterinary medicine),
Philo of Byzantium (mechanics, hydraulics, siege engines), Varro (agriculture),
Vegetius (military), Vitruvius (architecture, machinery), and Xenophon (horse-
manship, hunting, agriculture). Many fragments or short works on surveying have
been collected in the Corpus Agrimensorum (Thulin 1913; B. Campbell 2000). The
following authors wrote works that contain significant information relevant to
technology but are not themselves handbooks of technology: Archimedes (math-
ematics and geometry), Aristotle (various sciences), Galen (medicine), Oribasius
(medicine), Pappus (mathematics), Pliny the Elder (human civilization and the
natural world), Ptolemy (geography), Strabo (geography), and Theophrastus (nat-
ural materials). Numerous short works and fragments of works concerning ge-
ography and navigation have survived, collected in Muller (1855-1861).
These varied textual sources present different problems for the historian. Tech-
nical and nontechnical texts have survived in the form of manuscripts, and have
thus been subject to repeated selection and copying, which has often affected not
only the form in which they survive, but also their content. The first phenomenon
has led to the formation of technical corpora, "bodies" or collections of works that
deal with related topics but are not necessarily related in their composition. For
instance, the Corpus Hippocraticum consists of treatises that early on were grouped
under the name of Hippocrates, but were in fact produced by different authors,
writing between the fifth and second centuries B.C., and not only with very diverse
views, but also perhaps different professional identities. While some Hippocratic
texts were manifestly produced in a context of medical practice, some could equally
well be classified as philosophy or rhetoric-above all, On Ancient Medicine and
On Techne (Lloyd 1987; Schiefsky 2005). Another example is the Corpus Agrimen-
sorum Romanorum, comprising a wealth of material on land-surveying, dating
from possibly the first century B.C. to at least the fourth century A.D. (B. Campbell
2000; M. J. T. Lewis 2001). Finally, manuscripts concerning subjects oflittle interest
to the literate elite could be left uncopied, and thus ultimately lost. It is no accident
that most of the surviving handbooks concern agriculture and military technology.
In the second kind of transformation frequently undergone by manuscripts,
the contents of a technical work can be changed in ways that range from scribal
errors (especially in copying numbers) to the modification of diagrams or the ad-
dition of explanations, corollaries, or lemmas that eventually become incorporated
into the main text. These issues are particularly acute in the transmission of math-
ematical manuscripts (Fowler 1999).
Both technical and nontechnical texts tend to offer a perspective on technical
knowledge that, unsurprisingly, depends on the agenda and circumstances of in-
dividual writers. In the case of nontechnical authors, the historian ought to ask
how well informed they were about the technology they present, and what attitude
they had toward technology in general. In the case of technical authors, again the
historian ought to wonder for whom the text was written-whether it was meant
ANCIENT WRITTEN SOURCES 17
to be read by colleagues, pupils, or patrons-and what function it was meant to

have: manual of instruction, celebration of the subject, or codification of knowl-
edge that up to that point had been transmitted orally.
Inscriptions and papyri avoid some of the pitfalls of manuscript transmission,
in that they come to us more or less directly from the past. We must not forget,
however, that inscriptions were public documents designed to be displayed, so they
again offer a particular, somewhat biased perspective. Papyri, on the other hand,
coming as they do in their great majority from Egypt, do not necessarily allow gen-
eralization to other parts of the Mediterranean world.
With these words of caution in mind, in what follows I aim to give readers an
idea of the range of extant ancient textual sources for engineering and technology.
At the same time, I will try to sketch a broad outline of how the production of texts
dealing specifically with technical matters changed in the course of antiquity (see
Mei:Bner 1999). In order to combine the two aims, I will proceed chronologically,
focusing on two or three examples from each period, chosen to represent both dif-
ferent types of textual evidence and the technological practice of the period in
question.
CLASSICAL ATHENS
Our earliest surviving technical texts from Greek and Roman antiquity date to the
fifth century B.C., and form part of the Corpus Hippocraticum (Nutton 2004: 60-
61). We know that treatises on rhetoric, as well as on architecture, sculpture, and
possibly many other forms of technical knowledge, were also written at this time,
but none have survived (Lanza 1979). The emergence of a written tradition for
forms of knowledge that had thus far existed only in an oral context can be linked
to various phenomena that characterized the cultural and social life of Athens in
the fifth and fourth centuries B.C. The increased importance of literacy is one,
along with the connected significance of putting information into a form that was
perceived as more stable and longer lasting than simple oral transmission (Thomas
1992). The presence of the sophists, and their claim to be able to teach pretty much
anyone pretty much anything for a fee, can also be linked to the production of texts
for instruction, particularly manuals of rhetoric. The existence of a widespread ag-
onistic ideal, by which bearers of knowledge were prepared and expected to defend
their ideas in public against opponents, may also have contributed to the articu-
lation of some forms of knowledge into a written format, where arguments could
be developed in a more systematic and comprehensive fashion. Several of the works
in the Corpus Hippocraticum, for instance, have a marked polemical tone, and dis-
cussions about the status and nature of technical knowledge abound in the works
of philosophers like Plato and Aristotle (Cambiano 1991).
18 SOURCES
A very different kind of text, although one equally meant for public fruition,
deserves our attention, however: a marble stele inscribed on four sides with the
accounts for the construction of one of the most iconic buildings of classical
Athens, the Parthenon, completed around 432 B.C. after many years of work (JG I3
436-51). There are comparable accounts for the Erechtheum (JG I3 474-49; Randall
1953), also on the Acropolis, and for the temple of Asclepius at Epidaurus (JG IV.i2
102; Burford 1969). We know from later sources (Plutarch, Per. 13.9, 31.2) that
financing such a massive and luxurious enterprise was fairly controversial, and that
Pericles, effectively the political leader at the time, was accused, if not of embezzling
funds, at least of promoting his own interests and glory, and of favoring his own
friends (the sculptor Phidias, for instance) in allocating the work.
A written document such as the accounts of the Parthenon provides us with
useful information about the individuals involved with the building site. For in-
stance, we know that the works were supervised, in terms of administration, by a
board of five trustworthy citizens, called epistatai, appointed by the assembly. The
workforce consisted of people with various levels of expertise, from the architects
responsible for the design (apart from Phidias, later tradition has transmitted the
names of Callicrates and Icinus) and general technical supervision, to the headmen
of various teams, to the workers who specialized in tasks such as fluting the col-
umns or quarrying the marble, to nonspecialized laborers in charge of transporting
or clearing materials. The workforce was diverse also in its geographical origin:
even a large city like Athens would not have been able to provide enough special-
ized technicians for the jobs required, and when workers are named, as in the ac-
counts of the Erechtheum, we find several metics (resident foreigners) among them
(Randall 1953; Burford 1963; Korres 1995).
Although the inscriptions do not provide us with details about how the Par-
thenon was built, we can reconstruct the various phases of construction by fol-
lowing what is recorded each year (cf. Burford 1963: 29-32). First the foundations
of the temple were laid, and payments were made not only to quarrymen and
builders, but also to people who carted the marble from the Mount Pentelicon to
Athens, and specifically to the highest point in the city, the Acropolis. From then
on, we find that for several years wood was among the materials being bought, for
use in scaffolding and roof beams. One year's account specifies that work was being
conducted on the columns, and later precious materials (ivory, silver, and gold) are
mentioned, for decoration-including perhaps the massive cult statue of the
goddess Athena. As time passed, surplus material and even equipment were sold
off, perhaps to working teams producing monuments at other sites. The inscrip-
tion also provided information about the costs, but great chunks of the text where
the figures would have been are missing.
The stele was found in the Athenian marketplace, the agora, and was evidently
meant to provide public information about exactly how the money had been spent;
this tells us that the production of monuments at the time was viewed as a matter
of collective concern, and that the community was expected to be directly or in-
directly involved in it.
Further evidence on the status of technology in classical Athens is given by

some nontechnical texts, specifically plays-as with the building inscriptions, texts
meant to be presented to the public. A chorus from Sophocles' Antigone (332-67;
ca. 440 B.C.) and a long passage from Aeschylus' Prometheus Bound (463-522; ca.
430 B.C.) are particularly relevant here. Sophocles presents an encomium of hu-
mans as technological beings:
There are many wonders, but nothing more to be wondered at than humankind.
This creature crosses over the grey sea in the face of the wintry south wind and
ploughs his way through the roaring billows. And they vex the indestructible and
inexhaustible Earth, the eldest of the gods, with their ploughs and the race of
horses going up and down, turning over the earth year after year. And after
snaring the race of nimble birds and the host of fierce wild beasts and the mar-
itime creatures of the sea in the net's meshy folds, this skillful being carries them
off. By his devices he masters the rustic wild animals and makes obedient to the bit
the shaggy-maned, mountain-ranging horse and the unwearying mountain bull,
yoking them about the neck. He has learned speech and lofty thought and pub-
lic speaking, and to flee the arrows of the storm and the barbs of inhospitable frost
in the open air. Always inventive, he never meets the future unprepared. Only
from death he has not created an escape, but he has developed cures for unre-
lenting diseases; skillful beyond all hope are the devices of his art [techne]. And
sometimes he glides toward evil, at other times toward good.
The theme of the moral ambiguity of technical knowledge is embodied by the

protagonist of Aeschylus' play, the rebellious demigod Prometheus, who steals fire
from Zeus to give it to humankind and is subsequently punished by being tied to a
rock, his liver perpetually eaten by an eagle. In a famous monologue, the Titan re-
members how his gift turned humans from a beastly existence to a life character-
ized by shelter, travel, metalworking, medicine, writing, numbers-in short, all the
forms of technical knowledge, which were, he said, "from him":
Their every act was without knowledge until I came. I showed them the risings and
settings of the stars, hard to interpret until now. I invented for them also num-
bering, the supreme skill, and how to set words in writing to remember all
things, the inventive mother of the Muses. I was the first to harness beasts under
the yoke with a trace or saddle as a slave, to take the man's place under the
heaviest burdens; I put the horse to the chariot, made him obey the rein, and
be an ornament to wealth and greatness. No one before me discovered the sailor's
wagon, the flax-winged craft that roam the seas. Such tools and skills I dis-
covered for humans .... All arts [technai] possessed by mortals come from
Prometheus.
Alongside a celebration of human achievements and progress in the face of an

often hostile nature, these two plays recognize the potential of technology for sub-
version of the normal order through an empowering of the weak. It is not a
coincidence that both Prometheus and Antigone are defiant rebels. That these plays
were performed in front of a popular audience also tells us that technical knowl-
edge in its various manifestations was something that the average Athenian citizen
20 SOURCES
would have been familiar with, exposed to when not directly practicing it, and
ready to reflect on from a historical and moral point of view.
HELLENISTIC KINGDOMS
The main factor that differentiates technical texts produced in the fifth and fourth
centuries B.C. in classical Athens from those written from around the third century
B.C. onward, and around the Hellenistic metropoleis-Alexandria, Pergamon,
Rhodes-is above all the presence of patrons. The agonistic context of some of the
medical treatises in the Corpus Hippocraticum, where the author could address a
general public, real or ideal, as ifhe were giving a speech against an opponent, does
not disappear, but it somehow loses prominence in favor of a more or less direct
relationship between the technician and a powerful addressee, often a political
leader. Hellenistic kings were prepared to invest heavily in culture by founding
libraries, financing institutions such as the Alexandrian Museum, and supporting
individual poets, philosophers, and various other "intellectuals" who often lived in
relatively close contact with them at what have been called Hellenistic courts. At
the same time, the practical necessities of ruling states that were much larger than
the classical Greek polis called for experts in land-surveying, water-supply, ar-
chitecture (including fortifications), navigation, and military engine-building.
Sometimes a strong administrative structure preexisted the formation of the
Hellenistic kingdoms, and traditions, including technical traditions, merged. A
good example of the intermingling of tasks, knowledge, and people in Ptolemaic
Egypt is provided by a group of papyri about the architekton Cleon, who lived in
Crocodilopolis/Arsinoe in the Fayum in the third century B.C. (P. Petrie ll.4.1-13;
N. Lewis 1986: chap. 2). As is the case for a surprising number of papyri, a good deal
of Cleon's correspondence consists of complaints. For instance, a group of doc-
uments dating from 255 to 253 B.C. informed him of the problems some Egyptian
men, working in a quarry, were having with one of the supervisors, Apollonius.
The men, who were free and presumably on a corvee, complained through their
headmen that they had not been properly and sufficiently equipped with iron tools,
including wedges, and that they lacked wheat for sustenance and slaves to do some
of the clearing work (P. Petrie ll.4.1-5, 13.1; P. Petrie ll.4.6-9 may relate to the same
quarrymen). Throughout the documents, Cleon is presented as responsible for
procuring materials and specialized personnel for works relating to water supply.
One letter mentions a canal or conduit that was dug the year before but had now
silted up; another discusses sluice-gates, bridges, and even prison walls (P. Petrie
ll.4.10-11, 13.2-5, 13.8-16, 13.18a-b; on water supply works see also P. Lille 1 (259-58
B.C.), with a diagram). Kleon acted through a number of delegates, at various
levels of authority and presumably of expertise: we come across Greek names such
as Diotimus and Clearchus, but also Egyptian ones such as Petesnites and Armas
(P. Petrie II.11.1-2; no date is given). We also hear about Cleon's two sons, Poly-
crates and Philonides (Philonides at P. Petrie 11.13.19, Il.42.c, IIl.42.H.7; perhaps
both brothers at 11.16; Cleon's wife Metrodora at IIl.42.H.8). Polycrates wrote to
him in what the papyri editor calls "a beautifully clear and correct hand," asking to
be introduced to the king with a view to employment, and telling him that he
was going to train as geometres (here probably land-surveyor). We thus learn that
Cleon knew the king personally, and that technical knowledge may have run in
families, although father and son were not necessarily in a master-and-apprentice
relationship.
It is interesting that Cleon's main duties as they appear from these papyri seem
to be administration and supervision, rather than design or even personal care of
the works. Indeed, his successor, Theodorus, is called "architect" and oikonomos
(steward/manager), and his tasks appear to have been very much the same (P. Petrie
Il.9.1-5, 15.1-3 [341-39 B.C.]; also Il.42.a, IIl.43.2-3). No specific reference is made
to Cleon's or Theodorus' specialized knowledge, but that does not mean that it did
not play a crucial role in forging their identity, or that it was not determinant in
procuring them jobs. Rather, we need to adjust our view of what ancient engi-
neers were expected to do, and include administrative and management capacities
alongside technical skills.
Another technician at the service of the state common in the Hellenistic period
is the military expert. Kings and urban communities of the period invested heavily
in city walls, weapons, ships, and the hiring of professional mercenaries, and it
is not surprising that the majority of technical treatises extant from this period
are war-related. We have texts dealing with what to do in the case of a siege (e.g.,
Aeneas Tacticus), texts teaching how to be a good general (e.g., Asclepiodotus,
Tactica), and texts explaining how to build fortifications or machines suitable for
offense and defense. A good example of the latter is the Construction of Catapults
(Belopoeica) by Philo of Byzantium (ca. 240 B.C.; Marsden 1971; Garlan 1974).
Philo was a practicing technician and addressed both the Belopoeica and an-
other text on fortifications to an otherwise unknown Ariston, who appears to be,
rather than a fellow technician or a total layman, a member of the educated elite,
whose interest in military-related technology could at the time almost be taken as a
given. Biton (perhaps 241-197 B.C.) addressed his treatise on siege engines to one of
the kings of Pergamon (Marsden 1971: 6, 78 n. 1; M. J. T. Lewis 1999).
Philo's treatise on catapults was part of a larger Mechanice syntaxis (Synthesis of
Mechanics), and it combines engineering and geometry with epistemology and his-
tory of science. Philo begins by explaining how confusion still reigns among tech-
nicians, who often produce very different results even starting from the same
principles and using the same materials. His aim is to provide a reliable method for
building different kinds of catapults, and for modifying the dimension of a catapult
according to the size of the projectile. This latter procedure was based on the facts
that all the components of a catapult were in proportion to each other, that one of
these pieces (specifically, the cylinder holding the torsion spring) was taken as a
22 SOURCES
module or standard, and finally that a simple numerical relation could be estab-
lished between projectile and module. In other words, in order to modify the di-
mension of a catapult an ancient engineer had to know how to modify the di-
mension of the cylinder holding the torsion spring, and he then had to increase or
decrease all the other pieces accordingly. Philo and, later, Vitruvius refer to complete
tables of specifications for catapults of various sizes. Philo also provides suggestions
on the materials to use, and hints at the existence of accompanying diagrams.
He is quite explicit on the criteria a good catapult should satisfy: it should
shoot far and be powerful, but also present an awesome appearance to the enemy,
and be not too costly. Philo prizes novelty of design, and he presents himself as
having improved on some of his predecessors, including Ctesibius. On the value of
discovery, he says (Belopoeica 58.26-59.1; trans. adapted from Marsden 1971: 121):
One must praise those who at the beginning discovered the construction of these
instruments, for they were the originators of the thing and of its shape; they
discovered something superior to all other artillery, I mean like the bow, javelin,
and sling, in length of the shot and weight of the projectiles. To devise something
at the beginning and to realize the device is of a superior nature; to bring to
correction or adaptation something that exists seems rather easy. But, although
very many years have passed since the putting-together [of the machine] hap-
pened to be discovered and there have been, as usual, both many machine- and
artillery-makers, no one has dared to depart from the established method. We
were the first to do so.
Most importantly for the historian, Philo also provides precious information
on the past and the present of his discipline: he reports how former technicians
experimented with their machines and through trial and error came to the reali-
zation that one piece of the catapult could be used as a module, so that all the other
components could be expressed in proportion to it. Moreover, the modification of
the cylinder holding the torsion spring could be carried out according to mathe-
matical principles, because it basically amounted to a geometrical problem known
as duplication of the cube. In the Belopoeica, Philo provides a solution to this prob-
lem, complete with proof, thus giving the stamp of ultimate mathematical stability
to a result that had been arrived at empirically.
The image of technology that emerges from Philo's treatise is that of a disci-
pline in expansion: growing through time from trial and error to a kind of math-
ematical certainty, accumulating knowledge, again through time, while learning
from previous mistakes, and spreading geographically over the Mediterranean.
Philo informs us about the existence of what we can call a network of catapult
builders across Hellenistic kingdoms and cities, and he comments on the largesse
and philotechnia (love of technical knowledge) of the Ptolemaic kings, thanks to
whom particular progress had been made by the engineers in Alexandria.
On the whole, Philo may be considered representative of the Hellenistic breed
of technicians: emboldened by their newly found importance in military opera-
tions, they address a public beyond their immediate circle of peers or fellow citi-
zens. Their audience can be said to include figures of political relevance, as well as
colleagues or apprentices; their perspective has widened in space (together with
their travels) and in time. Philo is able to look back in order to trace the journey of
his form of knowledge from a past of uncertainty to a present of mathematical

reliability, and to point to a future of further innovation. He comes across as aware
of the importance of what he does, and aware that he is part of a group. His work
has to be understood against this background: Philo's Belopoeica does not simply
provide information on how to build catapults. Indeed, the treatise implies much
tacit knowledge, so that its actual usefulness as a manual of instruction, especially
to a complete beginner, is doubtful. Rather, in an environment where, with the rise
of Hellenistic scholarship and the foundation of big libraries, written sources were
being collected and canonized, the text provides respectability, both for the dis-
cipline of mechanics in general and catapult-building in particular, to which it
gives a history and epistemological and mathematical ratification, and also for its
author, to whom it offers a platform for his claims and his designs.
It is then all the more intriguing that most of the evidence for the technological
feats of possibly the most famous Hellenistic engineer of all, Archimedes of Syr-
acuse, should be provided by nontechnical sources. Famously, and despite the
survival of many of his other works, no treatise by Archimedes is extant about
catapults, burning-mirrors, giant ships, or astronomical globes, all of which he is
reported to have built. Plutarch later (notoriously) explained that Archimedes' all-
consuming passion was mathematics, and that he engaged in machine-building
only at the behest of his king, Hiero of Syracuse, and later on in order to help his
city when it was besieged by the Romans. Polybius (8.3-7) and Livy (24.34), as well
as later historians (Plutarch, Marc. 14.9-17.3), give accounts of how versatile and
impressive Archimedes' catapults were, and how they managed to strike terror into
the hearts of the numerically superior Roman army. When Syracuse fell, it was
thanks to treason, not technical superiority (Dijksterhuis 1987).
Perhaps Plutarch's description of Archimedes is historically accurate, and he
did not deem mechanical matters worthy of written treatment-a situation that
would make him even more unusual among Hellenistic technical experts than he
already is, thanks to his outstandingly sophisticated mathematical achievements.
The circumstances of his life and even of his death, however, also mark him as
typical of engineers contemporary with him: working in the service of a patron (he
addressed a treatise on astronomy and arithmetic to king Gelon of Syracuse), pro-
viding for the safety and welfare of his community, and eventually being assimi-
lated by the new world power, in the shape of the Roman troops who killed him
when Syracuse was taken in 212 B.C.
THE ROMAN EMPIRE
The rise of the Roman Empire coincides with the advent of an even bolder breed of
technician, well exemplified by Vitruvius Pollio and his De architectura. In ten eclec-
tic books that deal with styles of architecture, materials, and planning, but range as
well from what we would call ethnography to the construction of catapults, to that
24 SOURCES
of sundials, to methods for finding water, Vitruvius not only provided information
useful to the specialist, but also, and especially, strengthened the claim that tech-
nical experts like himself played a fundamental part in the empire. In the opening
chapter, Vitruvius famously claimed that architects were well-rounded, well-edu-
cated individuals: he said that they ought to be competent in, among other things,
philosophy, law, astronomy, music, and medicine (De arch. 1.1-10). His treatise is
scattered with anecdotes about the ethical virtues of technicians, and even contains
a potted history of humankind from its brutish origins to the advent of civilization
(De arch. 2.1.1-7). This passage, marked by stepping-stones such as the accidental
discovery of fire, is essentially a tribute to technology, which, far from being con-
trary to nature, imitates the cosmic order and is a part of it (Romano 1987).
Vitruvius dedicated De architectura to the newly installed Octavian Augustus
(1, preface), the man who allegedly turned Rome from a city of bricks to one of
marble, and the treatise can indeed be seen as a celebration of his empire. Not
coincidentally, Vitruvius explicitly drew on a wealth of earlier Greek architectural
manuals, none of which survives. The Roman De architectura erased previous ef-
forts in more senses than one. By comparing throughout, directly or indirectly,
Roman achievements with those of the Greeks or the barbarians, Vitruvius estab-
lished the former as dominant, not only in deeds, but in accumulated knowledge.
A slightly different perspective comes from the works of Hero of Alexandria.
On the one hand, he writes in the mold of Philo of Byzantium and inserts himself
explicitly in the Greek tradition of geometry and mechanics, to the point where
some historians would tend to classify him as Hellenistic, although he lived around
the mid-first century A.D. On the other hand, his treatises contain scattered clues
to the fact that he lived in a world now under Roman rule. Hero's devices, with the
possible exception of his catapults, have often been described as gadgets or toys,
but a more correct interpretation of his contribution to technology would empha-
size their capacity to produce wonder, and to engage the observers in the philo-
sophically validated act of curiosity. For instance, he provides several designs for
automata, that is, things that move apparently by themselves, and in particular for
a device that opens the gates of a temple when someone lights a fire on an altar
(Pneum. 1.38-39). The point of the device for Hero is that it enables him better to
reflect on the properties of matter, in particular of hot air, down to the presumed
characteristics of particles and void, and that it also enables him to produce a
wondrous effect, which will cause amazement in the audience and by reflection em-
power the technician, who alone knows what is going on behind the scenes (Tybjerg
2003). This particular configuration of knowledge and power, achieved through a
manipulation of reality that involved building machines, returns in the introduc-
tion to Hero's own Belopoeica (71-73.11; trans. adapted from Marsden 1971: 19):
The largest and most essential part of philosophical study is the one about
tranquility [ataraxia], about which many researches have been made and still
are being made by those who pursue learning; and I think research about tran-
quility will never reach an end through reasoning [logoi]. But mechanics has
surpassed teaching through reasoning on this score and taught all human beings
how to live a tranquil life by means of one of its branches, and the smallest-I
mean, of course, the one concerning the so-called construction of artillery. By
means of it, when in a state of peace, they will never be troubled by reason of re-
surgences of adversaries and enemies, nor, when war is upon them, will they ever
be troubled by reason of the philosophy which it provides through its engines.
One could hardly be more explicit: pitting himself against the most prestigious
form of knowledge of all, philosophy, Hero asserts the right of his own discipline
and by extension of technicians as a whole to claim a crucial role in society.
As with Philo in the third century B.C., technical treatises in the Roman impe-
rial period were not just about conveying information or instructing the novice. In
fact, if we assume that the sole aim of ancient technical treatises was to enable the
reader to build a house, supervise the water supply, or divide up a piece of land, we
would have to come to the conclusion that many of them simply do not work. This
has led some historians to say that some ancient technical writers did not know
what they were doing or produced pipe dreams, detached from reality. But the in-
capacity of some technical treatises to enable the reader to produce a technical arti-
fact just on the basis of the text should not be a surprise to anyone who has ever
grappled with the instructions that accompany unassembled furniture kits. Techni-
cal training in antiquity relied primarily on oral teaching and on direct experi-
ence through apprenticeship. Even the simplest technical treatises often imply tacit
knowledge-about how to fit pieces together, or how to choose materials, or how
to turn a two-dimensional geometrical description into a working structure-that
needs to be supplied independently of the text. Some scholars (e.g., Oleson 2004,
2005) propose the existence of a class of subliterary technical manuals, how-to-do-it
booklets concerning metalworking, water-lifting, or agriculture that were lost as Ro-
man imperial culture withered away. Unlike the surviving technical literature, such
booklets most likely would have lacked both literary and philosophical pretensions.
We thus have to come to the conclusion that technical texts have to be seen not
only as providing information, but also as constructing a certain way of knowl-
edge, and a certain identity for their authors. In the Roman period even more than
at earlier times, technical knowledge was presented as useful to the commonwealth,
epistemologically solid thanks to its links with mathematics and/or philosophy,
and essential to the establishment or maintenance of the political order. The tech-
nician was correspondingly represented not only as competent and able to carry
out specialized tasks, but also as honest and virtuous, or even as an upright Roman
citizen contributing to the welfare of the empire.
Vitruvius is but one example of this posture; another is Sextus Julius Fronti-
nus, author of De aquaeductu urbis Romae (On the Aqueducts of the City of Rome;
see now Rodgers 2004). Whereas Vitruvius seems to have been, generally speaking,
an upwardly mobile member of the lower middle classes who had come up through
the army, Frontinus, who also had a strong military background, was a senator and
a sometime consul. He was charged in A.D. 97 with the supervision of the water
supply of the empire's capital city, and De aquaeductu stems from that experience.
It can be seen as an administrative pamphlet, written by Frontinus for his successor
26 SOURCES
and possibly as a report to the Senate. It provides invaluable information on the

aqueducts of Rome, their foundation and maintenance; it also provides figures on
their capacity. Above all, Frontinus creates a certain image of the technically-
minded high-level civil servant, who equates expertise with efficiency, good rule,
and the welfare of the people of the urbs (Delaine 1996; Rodgers 2004: 12-14).
A similar tone is found in many of the writers in the collection now known as
the Corpus Agrimensorum Romanorum, which includes texts attributed to Fron-
tinus (B. Campbell 2000; M. J. T. Lewis 2001). One author, Balbus, writes for his
addressee, Celsus, a treatise titled Description and Analysis of All Figures, an expla-
nation of systems of surveying and measuring. Not only does Balbus mention his
apprenticeship (tirocinium) and his and Celsus' shared professio, but also the fact
that he has been on a campaign to Dacia (ca. A.D. 101-6) with the emperor, here
probably Trajan (B. Campbell 2000: 204). Again, we find references to training, to
self-awareness as a group, and to vicinity to the top echelons of political power. The
Description is basically a short treatise on elementary geometry, presenting roughly
the same material as the beginning of the first book of Euclid's Elements, but in a
narrative form rather than an axiomatico-deductive one. Given that Celsus, im-
plicitly presented as more senior than Balbus, would have already known the
contents of the treatise, why did Balbus write it at all? The answer is clear from the
introduction: he was keen to have his name circulate among people that Celsus
could indirectly introduce him to, and he wanted to show that he had some the-
oretical knowledge to support his technical skills (trans. B. Campbell 2000: 204-7):
It seemed disgraceful to me that if asked how many kinds of angle there were, I
should reply "many." Therefore in respect of those points relevant to our pro-
fession, as far as I could in my work, I have set out the types, characteristics,
conditions, measurements, and numbers. If you, a man of considerable influ-
ence, think that this work will benefit those learning [the profession], that will be
sufficient recognition of my modest talent.
It is only appropriate that a conquered province, Dacia, should be the implicit

background to the exchange between Balbus and Celsus. The presence of occupied
territories, non-Roman units of measurements, and pre-Roman land management
traditions is very strong in the Corpus Agrimensorum. Surveyors inevitably had to
negotiate between local realities and the demands of imperial administration. In
their texts, imperial control is often equated with geometrical order, to be imposed,
if only in the form of a map, on an often recalcitrant nature, full of mountains,
ridges, and other obstacles to overcome.
That the Roman Empire was identified with its technical achievements is con-
firmed by some nontechnical sources. For instance, Agricola, the biography Tacitus
wrote of his father-in-law around A.D. 98, describes the Roman colonization of
Britain thus (21; trans. from Lewis and Reinhold 1990: 279):
In order that people who were scattered, uncivilized, and hence prone to war
might be accustomed to peace and quiet through comforts, Agricola gave per-
sonal encouragement and public assistance to the building of temples, forums and
houses .... He likewise provided a liberal education for the sons of the chiefs,
and ... he so encouraged them that although they had lately disdained the Roman
language they now eagerly aspired to rhetoric. Hence too, our style of dress
came to be esteemed, and the toga became fashionable. Step by step they turned
aside to alluring vices, porticoes, baths, elegant banquets. This in their experi-
ence they called "culture," whereas it was but an aspect of their enslavement.
As well as the imposition of tribute and the institution of assize circuits and
law courts, the presence of Roman power was made tangible by baths and forums,
which were, ambiguously, both a means to an easier and more peaceful life and an
instrument of servitude, a way through which the Britons were won over and
ultimately Romanized. Similar sentiments are echoed by a passage from the Ba-
bylonian Talmud (Shabbath 33b; Lewis and Reinhold 1990: 333-34):
Rabbi Judah began and said: "How excellent are the deeds of this nation [Rome].
They have instituted market places, they have instituted bridges, they have in-
stituted baths." ... Rabbi Simeon ben Yohai answered and said: "All that they
have instituted they have instituted only for their own needs. They have instituted
market places to place harlots in them; baths, for their own pleasure; bridges, to
collect tolls."
The Roman reaction to the dissenting rabbis is appropriate to their degree of

technological appreciation: Judah is praised by the authorities, whereas Simeon is
condemned to death. Evidently, criticism of the Roman infrastructures is perceived
as criticism of the empire. The identification of Rome with its forums, baths, and
bridges appears complete, and puts a dent in interpretations that see technologi-
cal achievements, ancient or modern, as "objective" or "neutral." Roman tech-
nological achievements were arguably meant, and arguably perceived, as politically
charged.
Inscriptions
Apart from the architectural monuments that represent the most visible legacy of
ancient Roman technology, the empire left its mark in the form of inscriptions. We
have long texts, often celebrating conquests or reminding the public of a rich in-
dividual's benefactions, and we have shorter texts, often on milestones, boundary
stones, or tombstones. As was the case for classical Athens, epigraphy provides
some of the most interesting written sources for the history of Roman technology,
not only for the content of some inscriptions, but also for their location (when that
can still be reconstructed).
For instance, a set of inscriptions recording the settlement of boundary dis-
putes between Delphi and neighboring communities, in which a surveyor was
involved, was displayed on the walls of the temple of Apollo in the sanctuary. The
documents record how a Roman public officer had resolved disputes between
Greek communities, and are thus appropriately bilingual, in Greek and Latin.
28 SOURCES
Their location juxtaposes them, and their indirect commemoration of superior

Roman expertise, to earlier documents from the Greek past (Rousset 2002). Or,
again, an inscription now found at Bejafa (ancient Saldae) in Algeria records how
the military engineer Nonius Datus, who was attacked by brigands on his way
there, was sent to help the local authorities dig a water supply tunnel through a hill
(CIL 8.2728; see chapter 12). The workforce had initially miscalculated the exca-
vation, to the point where the two teams digging at opposite ends of the hill and
hoping to meet in the middle had in fact been proceeding in divergent directions.
Nonius Datus recalls how he provided not only technical expertise, but also su-
pervising skills, in that he encouraged competition between different teams, pro-
viding them with an incentive to complete the work accurately and quickly. The
inscription related Nonius' celebration at the hands of the local governor, and was
accompanied by the personifications of Patience, Excellence (Virtus), and Hope,
presumably the virtues presiding over Nonius' achievement. The inscription is also
itself a pointed celebration of the triumph of Roman technical ingenuity over the
North African wilderness, embodied not only by the brigands and by the hostile,
possibly water-poor territory, but also by the disorganized workforce in need of
guidance.
Most inscriptions from the Roman Empire, however, come from a funerary
context. The particular interest of epitaphs is that they provide valuable informa-
tion about how technicians chose to be remembered after their death. Many tombs
are engraved with scenes of craftsmen at work, or with images of technical tools; in
others, the epitaph expresses pride in technical activities. I will quote two examples.
The first is the epitaph from a sarcophagus found in Arles and probably dating to
the third century A.D. (CIL 12.722). The text is flanked by D[is] andM[anibus] (To
the gods of the afterlife) and above, in small-scale relief, by an axe and carpenter's
square.
Tomb of Quintus Candidus Benignus, master builder of the Arles association.

He had the full extent of the building art, dedication, knowledge, and discretion;
great technicians on any occasion declared him head of the association; nobody
was more knowledgeable than that; nobody could defeat him; he knew how to
make instruments and direct the flow of waters, he was a cherished guest here; he
knew how to nurture friends with ingenuity and dedication, mild and good-
natured [benignus], [erected by] Candidia Quintina for her sweetest father and
Valeria Maximina for her dearest husband.
The second epitaph was written in Greek on a tablet, originally from Hermopolis
Magna in Egypt and probably dating to the early third century A.D. (Donderer
1996: A8; trans. adapted from Taylor 2003: 1; cf. Bernand 1969: no. 23, with photo).
I am the tomb of Harpalos. Which Harpalos? Why, Harpalos most skillful in

the Daedalian craft. This I know, 0 Fates: his all-inventive art has perished with
him. What other man alive was his peer? He who laid out beetling temple walls,
who raised columns for high-ceilinged porticoes-he would often move the very
mountaintops, servants to his puny ropes, as easily as boys gather twigs. So Am-
phion, so Orpheus once charmed the rocks with song and led them effort-
lessly away.
For all their differences-language, location-these two memorials of technicians

have much in common: they both have literary pretensions (the first contains a
couple of puns, the second takes the form of an epigram), and they both express
not only pride in the profession, but also a certain competitive spirit (nobody
could defeat Benignus, just as no man alive was Harpalos' peer). They provide an
insight into the lives and values of everyday practitioners, and are the small-scale
equivalent of Vitruvius' Architecture in celebrating the achievements made possible
by technical knowledge.
LATE ANTIQUITY
The close relationship between political power and technology changed, but was
retained in late antiquity. Some of the most interesting written documents from
this period come from the legal codices. In A.D. 334, 337, and 344, Constantine
issued decrees according fiscal privileges to specified technical professions (Cod.
Theod. 13-4-1; Cod. Just. 10.66.1 = Cod. Theod. 13-4-2; Cod. Just. 10.66.2 = Cod. Theod.
13.4.3). The last law issued prescriptions:
With these words we compel mechanicians, geometers, and architects, who pre-
serve the divisions and partitions of all pieces <ofland>, and who pull together a
piece of work by designing structures, and those who discover the course of water
and demonstrate its size through skillful leveling, to devote themselves equally
to teaching and learning. Let them enjoy their fiscal exemptions and train teach-
ers, so that there are enough of them.
The usefulness of technicians for the state is recognized by the top political au-
thority, to the point of turning them into privileged individuals from the point of
view of fiscal obligations. As for the provisions made for teaching, there was
evidently a concern about securing the future of some professions by making sure
that the next generation was adequately trained.
A similar, urgent, sense of the significance of technical knowledge for the state
is conveyed by an anonymous treatise, written probably around A.D. 360-68, ad-
dressed to the then emperors, and commonly known as De rebus bellicis (Thomp-
son and Flower 1952; Giardina 1989 is fully illustrated). The author presents himself
as an advisor, reprising the administrative, supervisory role so many technicians
appear to have played in the course of antiquity. He deplores the current state of
things, with chaos reigning supreme within the empire and barbarians pressing
without, and proposes technical solutions: a currency reform, codification of the
laws, the fortification of the borders, and a number of military devices. These have
30 SOURCES
caught the eye of historians also because three of the four main manuscripts of De
rebus bellicis were illustrated, sometimes in lavish color; we can thus readily vi-
sualize, among others, an arrow-shooter made mobile by being installed on a cart,
three types of horse-driven scythed chariots, an especially light and wearable suit of
armor, and even a ship operated by oxen-driven wheels.
With De rebus bellicis, as with earlier and later treatises up to and including the
Renaissance, historians have often asked what was the point of describing machines
that were probably known to be impossible to realize, difficult to implement, or
simply obsolete. The answer is different each time. Philo of Byzantium wanted to
impress his audience and differentiate himself from other engineers in what can
almost be described as a competitive market, with patrons likely to be impressed by
new designs, and technicians traveling from place to place offering their services.
The author of De rebus bellicis was probably not primarily an engineer, but rather a
civil servant, which in the late Roman imperial sense of the term may have implied,
more commonly than modern historians recognize, some specialized knowledge or
expertise. He again wanted to attract the attention of a figure of authority, possibly
more to his financial or administrative reforms than to the weapons he describes
(Liebeschuetz 1994). We do not know how successful his bid for attention was at
the time. As for the modern reader, and perhaps already the manuscript copyist, it
is the weapons, the fancy feats of engineering, that catch the eye in more senses
than one, and convey an attractive image of power against the odds, mechanical
ingenuity triumphing over brutish enemies.
In fact, in late antiquity technical skills appear to have become widespread
among the barbarians-more than one late ancient historian commented on the
technical proficiency, especially with military equipment, of the tribes attacking the
empire (e.g., De rebus bellicis 4; Procopius, Goth. 8.11.27-28). Another group often
characterized as responsible for the decline and fall of the Roman Empire, the
Christians, appear to have been equally proficient. One of the most interesting
phenomena of late antiquity is the Christianization of technology, duly charted in
our written sources but usually neglected by historians of ancient technology.
Beginning around the fourth century A.D., the epitaphs of technicians, especially
architects, not only often invoke God and the Virgin Mary, but also evaluate the
activities of the deceased in relation to what they did in the service of the Church
(Cuomo 2007: chap. 5). St. Augustine himself wrote a panegyric to contemporary
technological accomplishments in De Civitate Dei (22.24), nicely bracketing the list
of accomplishments Aeschylus attributed to Prometheus (PV 463-522, quoted
earlier):
For beside the arts of living well and gaining eternal happiness ... , has not the
human intellect discovered and put to use so many and such great technologies,
in part to serve real needs, in part for the sake of pleasure ... ? What marvel-
ous, stupendous accomplishments human effort has achieved in the fields of
construction and textile production! How far it has progressed in agriculture
and navigation! What accomplishments of imagination and application in the
production of all kinds of vessels, various types of statues and paintings .... What
great inventions for capturing, killing, taming irrational animals! And against
humans themselves, how many types of poisons, weapons, devices! What medi-
cines and remedies humans have discovered for protecting and restoring their
health! How many sauces and appetizers they have found to make eating a
pleasure! What a great number and variety of signs for conveying thought and for
persuasion, the most important of which are words and letters! ... What skill in
measuring and counting! With what sharp minds humans grasp the paths and
arrangement of the heavenly bodies! How great the knowledge of this world
humans have filled themselves with! Who could describe it?
The recognition and institutionalization of Christianity coincided with a mas-

sive building program to provide new places of worship, recorded by, among
others, Procopius. In his treatise Buildings, he narrates how God himself intervened
in the construction of the cathedral of Hagia Sophia in Constantinople by in-
spiring the emperor Justinian with the solution to several technical problems, when
even his handpicked, world-famous engineers were at a loss (Aed. 1.1.66-78).
Procopius' Buildings has been described as a panegyric to the emperor Justinian,
or, in this case, Justinian's technical achievements, with only occasional mention
of the technical experts-architects and engineers-who must have been behind
them (Cuomo 2007).
It would be a mistake to consider all written sources for ancient technology as a

unified group. Even the category of so-called technical treatises is less an actual,
distinct genre recognized as such by the ancients than a convenient means of
classification for modern historians. In the absence of official qualification stan-
dards and in the presence of an often competitive marketplace situation (including
when the marketplace was extended to the whole Mediterranean region), each
technical profession was stratified, with different degrees of expertise, social status,
wealth, literacy, and numeracy. Thus, technical knowledge comprised many more
practices than our present classifications suggest, with varying degrees of closeness
to forms of knowledge that we tend to value, such as mathematics, or forms of
codification that we tend to appreciate, such as the written form itself. For instance,
the techne of the marine pilot-the art of navigation-which played a crucial role
in the economy and politics of the ancient world from classical Greece onward (it
is one of the arts most often mentioned by Plato), goes almost completely unre-
corded, not least because of its lack of written accounts (but cf. Casson 1989). In
other words, ancient technology was too motley and complicated an entity for tech-
nical treatises to be other than partial, problematic, and distorted reflections.
Texts that for their subject matter or the self-identification of their author
seem to us to belong together often were written for completely different purposes
and addressed to very different audiences. Each case has to be considered sepa-
rately, and more finely tuned understanding is required of individual authors. In
particular, the notion ought to be dispelled that technical treatises are instruction
manuals in a straightforward sense, and that their tone and style are "realistic."
This realization would enable us to appreciate them as literary products and,
32 SOURCES
consequently, to explore the issue of their aim and audience in a more nuanced
way than has so far been attempted for many of them.
Having warned against the dangers of generalization, I do think that some
trends are visible in the written sources relative to ancient technology, and thus
that it is possible to offer a brief overview of their shared features. Generally speak-
ing, technicians spoke in a climate of competition, in a marketplace. Even many of
the epitaphs relative to technicians reproduce a sense of comparison with others,
and some of the technical treatises explicitly compare technical disciplines with
other forms of knowledge, such as philosophy. The existence of competition and
debate, and sometimes of open hostility, both among technicians themselves and
between technicians and others, helps to make better sense of some of the evidence.
Several authors throughout antiquity express, in different ways, the opinion that
technical knowledge and its practitioners are or should be despised and margin-
alized, because their form of knowledge is questionable from an epistemological
point of view, and because, as people, they exhibit unethical behavior: technicians
are represented as greedy, arrogant, and base. Those authors include prestigious
names such as Plato, Aristotle, Cicero, and Seneca (cf. Humphrey et al. 1998: 579-
88), so that their views have often been taken to be the dominant opinion of the
society in which they operated. Yet, if we consider that most of the written sources
produced by technicians, including those I have focused on in this section, present
a different, much more positive picture, and factor in the basically constant pres-
ence of competition and debate, we will be able to reread the statements by Plato,
Aristotle, Cicero, and Seneca about technology. They will appear to be just con-
testants in an ongoing argument about knowledge and power that characterized
Greek and Roman antiquity, and is one of its most enduring legacies.
Again generally speaking, in their written texts technicians often addressed not
just other technicians, but the educated general public, and in particular people of
authority-a great number of technical treatises were directed to, or implied direct
familiarity with, kings and emperors. The unique selling point of technology to
these people was its usefulness: the capacity, or alleged capacity, to produce results
rather than mere words. Many of our written sources pair this claim with others
that enable technology to stand side by side with other forms of knowledge: the-
oretical stability, sometimes achieved with the help of mathematics; a respectable
past, including playing a role at the very dawn of civilization; religious affiliations,
from the patronage of Athena or Minerva to the direct intervention of the
Christian God; and the impeccable ethical credentials of its practitioners. All this
was often put at the service of the established authority, with the implication that
technology could help them not only achieve practical ends, but also impose and
retain order at a more symbolic level.
The third, and final, overall feature of technical texts from antiquity is com-
pletely unsurprising. They generally present a very positive, even idealized, view of
the discipline and of its achievements. I would go as far as to say that many tech-
nical treatises were written specifically in order to construct a positive, even ide-
alized, view of the discipline and, by reflection, of the author himself. They cele-
brate as much as inform; they are monuments as much as manuals. Technical

knowledge did not need to be written to be transmitted-oral tradition and ap-
prenticeship took care of that in the majority of the cases. But writing about one's
knowledge made it more visible and respectable, and created a memorial for the
author. Thus, it makes perfect sense for the written sources we have discussed in
this chapter to include both fully-fledged treatises and inscriptions: both Vitruvius'
Architecture or Frontinus' Aqueducts and the epitaphs of Benignus and Harpalos,
while celebrating technology and its achievements, preserve the unique, individual
voices of their authors.
REFERENCES
Bernand, E. 1969. Inscriptions metriques de l'Egypte greco-romaine. Recherches sur la poesie

epigrammatique des Grecs en Egypte. Paris: Les Belles Lettres.
Burford, A. 1963. "The builders of the Parthenon," in G. T. W. Hooker (ed.), Parthenos and
Parthenon. Greece and Rome Suppl. 10. Oxford: Clarendon Press, 23-35.
Burford, A. 1969. The Greek temple builders at Epidauros: A social and economic study of
building in the Asklepian Sanctuary, during the fourth and early third centuries B.C.
Liverpool: Liverpool University Press.
Cambiano, G. 1991. Platone e le tecniche. 2nd ed. Roma/Bari: Laterza.
Campbell, B. (ed.) 2000. The writings of the Roman land surveyors: Introduction, text,
translation and commentary. Journal of Roman Studies Monograph 9. London: Society
for the Promotion of Roman Studies.
Campbell, J. B. 1994. The Roman army, 31 BC- AD 337: A sourcebook. London: Routledge.
Casson, L. 1989. The Periplus Maris Erythraei. Princeton: Princeton University Press.
Celentano, M. (ed.) 2003. Ars/Techne: Il manuale tecnico nelle civilta greca e romana.
Alessandria: Edizioni dell'Orso.
Cohen, M. R., and I. E. Drabkin 1948. A source book in Greek science. Cambridge, MA:
Harvard University Press.
Cuomo, S. 2007. Technology and culture in Greek and Roman antiquity. Cambridge:
Cambridge University Press.
Delaine, J. 1996. "'De aquis suis'?: The 'Commentarius' ofFrontinus," in C. Nicolet (ed.),
Les litteratures techniques dans l'antiquite romaine: Statut, public et destination, tra-
dition. Geneve: Vandreuvres, 117-45.
Dijksterhuis, E. J. 1987. Archimedes. Princeton: Princeton University Press.
Donderer, M. 1996. Die Architekten der spaten romischen Republik und der Kaiserzeit.
Epigraphische Zeugnisse. Erlangen: Universitatsbund Erlangen-Niirnberg e.V., Uni-
versitatsbibliothek Erlangen,
Fowler, D. 1999. The mathematics of Plato's Academy: A new reconstruction. 2nd ed. Oxford:
Clarendon Press.
Garlan, Y. 1974. Recherches de poliorcetique grecque. Paris: Boccard.
Giardina, A. 1989. Anonimo: Le cose della guerra. Milano: Mondadori.
Humphrey, J. W., J. P. Oleson, and A. N. Sherwood 1998. Greek and Roman technology: A
sourcebook. London: Routledge.
34 SOURCES
Irby-Massie, G. L., and P. T. Keyser 2002. Greek science of the Hellenistic era: A sourcebook.
London: Routledge.
Korres, M. 1995. From Pentelicon to the Parthenon: The ancient quarries and the story of a
half-worked column capital of the first marble Parthenon. Athens: Melissa.
Landels, J. G. 2000. Engineering in the ancient world. 2nd ed. Berkeley: University of
California Press.
Lanza, D. 1979. Lingua e discorso nell'Atene delle professioni. Napoli: Liguori.
Lewis, M. J. T. 1999. "When was Biton?" Mnemosyne 52: 159-68.
University Press.
Lewis, N. 1986. Greeks in Ptolemaic Egypt: Case studies in the social history of the Hellenistic
world. Oxford: Clarendon Press.
Lewis, N., and M. Reinhold (eds.) 1990. Roman civilization: Selected readings. Vol. 2, The
Empire. 3rd ed. New York: Columbia University Press.
Liebeschuetz, J. H. W. G. 1994. "Realism and phantasy: The anonymous De rebus bellicis
and its afterlife," in E. Dabrowa (ed.), The Roman and Byzantine army in the East.
Krakow: Jagiellonian University, 119-39.
Lloyd, G. 1987. The revolutions of wisdom: Studies in the claims and practice of ancient Greek
science. Berkeley: University of California Press.
Marsden, E. 1971. Greek and Roman artillery: Technical treatises. Oxford: Clarendon Press
MeiBner, B. 1999. Die technologische Fachliteratur der Antike: Struktur, Uberlieferung und
Wirkung technischen Wissens in der Antike (ca. 400 v. Chr.-ca. 500 n. Chr.). Berlin:
Akademie Verlag.
Millier, C. 1855-61. Geographi graeci minores. 3 vols. Paris: Didot.
Nutton, V. 2004. Ancient medicine. London: Routledge.
Oleson, J.P. 2004. "Well pumps for dummies: Was there a Roman tradition of popular,
sub-literary engineering manuals?" in F. Minonzio (ed.), Problemi di macchinismo in
ambito romano. Como: Comune di Como, Musei Civici, 65-86.
Oleson, J.P. 2005. "Design, materials, and the process of innovation for Roman force
pumps," in J. Pollini (ed.), Terra Marique: Studies in art history and marine archae-
ology in honor of Anna Marguerite Mccann. Oxford: Oxbow, 211-31.
Randall, R. H. 1953. "The Erechtheum workmen," American Journal ofArchaeology 57: 199-210.
Reekmans, T. 1970. "Le salaire de Cleon," Archiv fur Papyrusforschung 20: 17-24.
Rodgers, R. H. 2004. Frontinus de Aquaeductu urbis Romae. Cambridge: Cambridge
University Press.
Romano, E. 1987. La capanna e il tempio: Vitruvio o dell'architettura. Palermo: Palumbo.
Rousset, D. 2002. Le territoire de Delphes et la terre d'Apollon. Paris: Boccard.
Sage, M. M. 1996. Warfare in ancient Greece: A sourcebook. London: Routledge.
Schiefsky, M. 2005. Hippocrates: On Ancient Medicine, translated with introduction and
commentary. Leiden: Brill.
Taylor, R. 2003. Roman Builders: A study in architectural process. Cambridge MA: Harvard
University Press.
Thomas, R. 1992. Literacy and orality in ancient Greece. Cambridge: Cambridge University
Press.
Thompson, E. A., and B. Flower 1952. A Roman reformer and inventor. Oxford: Oxford
University Press.
Thulin, C. 1913. Corpus Agrimensorum Romanorum. Berlin: Teubner.
Tybjerg, K. 2003. "Wonder-making and philosophical wonder in Hero of Alexandria,"
Studies in History and Philosophy of Science 34: 443-66.
CHAPTER 2
REPRESENTATIONS
OF TECHNICAL
PROCESSES
ROGER ULRICH
ANCIENT depictions that include images of technological processes offer vital ev-
idence for the reconstruction of methods, equipment, scale of operation, labor
force, and sequencing of tasks, along with a direct indication of social interest in
the relevant technology. Where literary descriptions are lacking, representations
offer the single most important resource for modern investigation of process. The
images that survive prove unequivocally that the subject of ancient technology
enjoyed a long life in a range of artistic media. Only the most durable examples are
extant-bas-reliefs from Egyptian tombs, Greek images on painted pottery, and
Roman-period relief sculpture constitute the lion's share of this evidence-but we
must imagine that there existed a highly developed tradition of pictorial repre-
sentation on less durable materials that evolved in tandem with the development of
writing and literacy. In addition to the illustrations that appear within this chapter,
I will naturally refer to illustrations that appear in other chapters in the book.
The importance of ancient depictions of technological processes for the gen-
eral study of the subject is attested by their prominence in virtually every modern
publication that addresses issues of tools, processes, engineering, and technology
from the ancient Mediterranean. Both Hugo Bliimner and Albert Neuburger relied
on published images of relevant artifacts in their publications of 1875-1887 and 1919,
respectively, although they depended on some drawn reproductions that contained
errors, and on descriptions of artifacts that were lost or inaccessible at the time they
compiled their evidence. Nevertheless, these early studies remain useful, written at
36 SOURCES
the nascence of an industrial age that then employed many of the same techniques
and tools utilized by the ancient workforce. Over the last century, new discoveries,
high-quality photographic reproductions, public stewardship of artifacts, and even
ease of travel have provided new opportunities to reexamine evidence long rec-
ognized as seminal for the topic and to identify new objects to add to the corpus of
ancient images of ancient technologies. General studies, such as Singer's History of
Technology (1954-1956), White's Greek and Roman Technology (1984), and Adam's
Roman Building (1994) incorporate relevant illustrations of ancient devices in use
(see also Forbes 1955-1964, Klemm 1964). The increasingly specialized nature of the
field has resulted in focused studies that include welcome compilations of images,
although few are primarily concerned with the examination of the manner of an-
cient representation per se. Included in this group are museum exhibitions that
have been accompanied by important catalogs such as Homo Faber, a project of the
Museo Nazionale of Naples (Ciarallo and De Carolis 1999) and the Fire of He-
phaistos exhibit at Harvard University (Mattusch 1996), as well as monographs
on subjects like Attic vase painting (Noble 1988; Schreiber 1999) or hand tools
(Gaitzsch 1980; Zimmer 1985). To attempt even a brief summary of this broad range
of evidence is beyond the scope of this chapter. Instead I hope to offer some general
comment as to the nature of these representations, their intent, and how they may
be categorized over the period of their use.
ARTISTIC INTENT
Any modern attempt to understand an ancient artist's intent or motivation with

regard to a given depicted subject is decidedly problematic. Furthermore, the con-
text, relative chronology, and media of representations of early processes suggest
that the artistic motivations or intents for reproducing such images vary signifi-
cantly over the twelve centuries under consideration. Even with these caveats re-
garding interpretation, a review of the evidence currently available suggests strongly
that there exists no single representation that includes a technical process in which
the artist's primary purpose was to represent in any instructive or analytical
fashion the technology portrayed. There are indeed a few representations that
come close: the well-known Berlin Foundry Cup is a good example (figures 16.1-3),
but here, as in virtually every example cited in this chapter, the technological
process portrayed is ancillary to something else: the image serves to emphasize an
activity for which divine protection is sought from a patron deity (thus we find
customized votive offerings or carved altars) (figure 2.1), the image is one part of a
funerary commemoration of a deceased craftsman or his patron (figures 2.2, 17.1),
the image identifies the presence of a specialized guild (figure 17.3), or is simply a
scene of industrial activity rendered as a charming if not humorous decorative
REPRESENTATIONS OF TECHNICAL PROCESSES 37
Figure 2.1. Corinthian pinax from Penteskouphia; mining scene, end of seventh century
B.C. Height 10.4 cm. Berlin no. 871; Bildarchiv Preussischer Kulturbesitz/Art Resource.
(ART180877.)
motif-this last a frequent subject of both Greek vase painting and Roman mural
decoration (figure 2.3).
The representation of a building crane included in a stone relief from the Tomb
of the Haterii in Rome (late first century) is illustrative in this regard (figure 2.4) .
The crane is carved into the left side of a marble panel that offers as its central focus
the image of a monumental tomb whose construction, it appears, necessitated the
use of a heavy lifting device. While other depictions of cranes exist (see Adam 1994:
44-47) , the machine from the Haterii tomb provides the most detailed rendition:
two workmen perched on the jib of the crane affix an inverted wicker basket,
perhaps to protect the tip of the jib, or possibly a reference to a Dionysiac mystery
cult (Jensen 1978; Landels 2000; Ulrich 2007). The level of detail is striking: braided
ropes, heavy pulley systems, a great planked wheel turned by the shifting body
weights of the laborers (five are shown), even ropes around the edge of the wheel
intended for braking or applying extra force. Close examination reveals some of
the wooden treenails employed to hold the rig together. Yet, despite the plethora of
details, the image of the crane is incidental to the overall composition, itself con-
ceived to place emphasis on the deceased, the adjacent monumental temple-tomb,
and perhaps the family's role and wealth from its construction business. The crane
serves primarily as a dramatic prop for greater pictorial themes whose exact pur-
poses are now elusive; but certainly technical didacticism is not one or these.
No original copies of Greek or Latin technical treatises have survived; we
can only speculate about the kind of graphic aids they may have included. That
Figure 2.2. Baking scenes from the Tomb of Eurysaces, Rome. Top: Baking and knead-
ing operations; bottom: packing and weighing. Late first century B.C. (Author, composite
images after Rossetto 1973.)
Figure 2 .3. Frieze of erotes and psychai engaged in metalworking, House of the Vettii,
Pompeii, first century. (Photograph by Michael Larvey, by permission of Michael Larvey
and the Soprintendenza Archeologica di Pompeii.)
Figure 2.4. Haterii relief of a lifting crane, late first century, Museo Gregoriano Profano,
Vatican Museums. Scala/Art Resource, New York. (ART85226.)
illustrations once accompanied such literary texts or how-to-do-it handbooks

seems a reasonable assumption (Oleson 2004, 2005: 222-23). At least some of the
intricate prose descriptions of military equipment, water-screws, or the layout of a
theater we find in Vitruvius were originally accompanied by drawings (e.g., De arch.
3.4.5; 10.6.4). We know that Vitruvius considered Hellenistic commentaries now
lost; he describes devices such as the pneumatic pump and flap-valve invented by
Ctesibius in the third century B.C. (De arch. 10.7). Similarly, the Syracusan Ar-
chimedes' (d. 212 B.C.) lost works on elementary mechanics and the properties of
mirrors must have been accompanied by diagrams, just as the mathematical for-
mulae that computed the correct parabola for a burning-mirror used to direct the
rays of the sun on enemy ships, published by Anthemius of Tralles much later in the
sixth century A.D., would have required some form of illustration (Huxley 1959).
Heron of Alexandria, who lived in the first century, describes numerous gadgets
and mechanical devices, including devices making use of steam power. He published
a monograph, now lost (titled Baroulkos) that described a lifting machine that
operated with a combination of gears capable of hoisting great weights. Although
40 SOURCES
Heron's texts are thoroughly keyed to illustrations (now preserved only in early
Arabic translations), no ancient depictions of gears or geared machinery survive in
the durable media of stone or ceramic. Among the technical treatises that have
survived, the topic of agriculture is paramount, but there is no reason to doubt that
other handbooks on a number of practical technologies were composed and widely
distributed (White 1984: 12; Oleson 2004, 2005).
The extant Greek literary corpus does not include references to drawn archi-
tectural plans or sketches; verbal specifications (syngraphai) might include dimen-
sions (for example, an inscription pertaining to the fourth-century B.C. arsenal at
Piraeus, JG2 2.1668); in addition, physical models (paradeigmata) of buildings
similar to those found at Perachora from the eighth century B.C., or of parts of
buildings, were apparently used to communicate general ideas about layout and
architectural embellishment (Coulton 1977: 54-60). Roman-period architectural
plans in mosaic and inscribed stone reveal that the practice of rendering buildings
both as plans (sing. ichnographia) and elevations (sing. orthographica) was familiar
to builders in Italy (Vitruvius, De arch. 1.2.2; Taylor 2003, 26-29). Full-scale ele-
vations, inscribed on existing stone paving, were used on site (or nearly on site) for
the preassembly of the entablatures or pediments of the Pantheon (Haselberger
1994). The scaled-down comprehensive plan of Rome known as the Forma Urbis
that has survived in a fragmentary state, carved on sheets of marble in the early
third century A.D. and mounted on a public wall near the Roman Forum, indicates
that such renderings were in public circulation and legible to at least some of the
population (Carettoni et al. 1960). Hadrian (emperor 117-38) not only exhibited a
keen interest in architecture but also apparently designed buildings himself, at least
on paper, and solicited the opinions of professional architects such as Apollodorus
of Damascus (Dio Cassius 69,4.4). In the later first century B.C., Vitruvius urged
that aspiring architects master the art of drafting (De arch. 1.1.3).
Representations of Greco-Roman technologies play an important role in fill-
ings gaps left even when related tools or special-purpose sites have been excavated.
Particularly informative compositions include operators; the representation of the
human agent can aid immensely in interpretation. Consider images of drilling in
wood or stone. For boring small-diameter holes in wood, both Greek and Roman
depictions indicate that a bow-and-drill could be operated by a single deft crafts-
man (figures 2.5-2.6). Larger diameter holes or grooves in wood or stone required
the help of an assistant who could power the bit while the primary operator di-
rected the cut. In this case, evidence includes not only a Homeric literary reference
(see chapter 17), but also an unambiguous image from Rome of a sarcophagus-
maker using such a tool (Gaitzsch 1980: cat. 318; Strong and Brown 1976: 200; figure
2.7). The rendering of a heavy strap (and no bow) instead of a thin thong held in
tension by the bow reveals a significant variant of tool type even though the re-
ciprocating principle of operation is identical. The recovery of isolated iron drill
bits, without handles, bows, or straps, or the examination of tool marks on the sur-
face of stone cannot by themselves lead to recognizing a significant variant of the
drill that is so readily apparent via the relief in Urbino.
Figure 2.5. Vatican gold-glass vessel: detail of an operator of a bow and drill; early fourth
century. Museo Biblioteca Apostolica Vaticana, inv. 60788. (Drawing by author.)
Representations of tools in use provide additional information about technical

processes beyond that which artifacts or workshops can yield, such as the posi-
tioning of tools and the stances assumed by the operator, or the layout of the
workshop. Painted depictions at Pompeii of fullers finishing cloth and cleaning
clothing, for example, are of great interpretive value when otherwise unremarkable
masonry tanks are excavated (Reg. Vl.8, 20; Ciarello and De Carolis 1999: cat. 120 );
a stone funerary relief from Sens (Yonne, Roman Agedincum Senonum) provides
corroborating evidence (figure 18-4). Ancient depictions can provide the appear-
ance of components missing from excavated examples, such as wooden handles
and stands, the number of operators required for a given function, or information
about the age, gender, status (i.e., freeborn or slave), and ethnicity of the operators.
The role of women in the workforces of Greece and Rome, for example, has been
more fully understood through ancient depictions (see below).
The most apparent value of ancient depictions of technical processes is the
recovery of images of tools, machines, or applications for which no physical traces
or literary descriptions survive (we may know the name, and thus the existence, of
a device from briefliterary mention, but description is another matter). There is no
extant written source, for example, that describes the process of throwing or
decorating a ceramic vessel, yet the subject was well known to the Greek vase
painter. We know, for example, from literary references and from the remains of
turned bowls (of stone and wood) or stone columns that Greeks, Etruscans, and
Romans made and operated both large and small lathes, but the appearance and
operation of the machine itself is still uncertain, now only partially solved through
42 SOURCES
Figure 2.6. Attic red-figure painting by the Gallatin Painter depicting a carpenter making
a chest to hold Perseus and his mother Danae, ca. 490 B.C. Height 41.7 cm. MFA Boston
inv. 13.200. (Photograph © 2007 Museum of Fine Arts, Boston.)
the examination of fragmentary depictions (see chapter 17). The appearance of

large lifting cranes powered by a "great wheel" turned by the shifting weight of
human operators treading within is known exclusively through carved stone de-
pictions (figure 2-4). A schematic rendition of a water-lifting wheel is known from
a fragmentary second-century A.O. mosaic from Apamea in Syria, in a region
where such devices once fed irrigation and aqueduct systems (Oleson 1984: fig. 41).
The use and appearance of the cylindrical water-screw (cochlea) invented by Ar-
chimedes, the design of which is known through Vitruvius (10.6.1-4), has been
captured, somewhat improbably, in exotic depictions of Nilotic scenes on two
frescoed walls from Pompeii and a terracotta relief from Roman Egypt now in the
British Museum (Ciarello and De Carolis 1999: cat. 410; Oleson 1984: fig. 86, 101;
BM inv. EA-37563). The Egyptian connection is fitting, since it was along the Nile
that such machines were in common use, according to Diodorus, who wrote about
both the Egyptian water-screws and others used in the Roman mines of Spain in
the first century B.C. (1.34.2; 5.37.304).
Yet even depictions of activities that include a high degree of detail can open
debate on the nature of the specific technology being depicted. The pot-decorating
scene painted by the Leningrad Painter on the Caputi Hydria has been interpreted
since its publication in 1876 as a scene of vase painting being undertaken by both
young males and one female (figure 2.8; Richter 1923). Green (1961: 74), however,
has attempted to demonstrate that the scene is one of workers decorating metal
vessels (see also Venit 1988); the traditional interpretation is still defended (Kru-
Figure 2.7. Strap drill depicted on the sarcophagus of Eutropos, originally from the Via
Labicana, Rome, mid fourth century. Museo Archeologico Lapidario, Urbino. (Author.)
kowski 1990). A metalworking scene from the House of the Vettii in Pompeii, dis-
cussed in more detail below, provides another example of interpretive controversy,
where scholars have argued for over a century on the specific nature of the activity
so carefully rendered (figure 2.3). Two terracotta reliefs from a second-century
tomb (no. 29) at Isola Sacra near the Roman port town of Ostia appear to show
artisans shaping or grinding tools, but the workstations vary significantly in ap-
pearance, and the exact nature of the activity portrayed is difficult to evaluate
(D'Ambra 1988: 91; Zimmer 1985: pl. 4).
In the assessment of such representations of technical processes, there is no
guarantee that the artisan rendering the image was himself knowledgeable about the
technique he was representing or even possessed the artistic skill to reproduce the
process accurately. If an anomaly is suspected in a particular representation, we must
assess whether it is an artistic fault or a previously unobserved variant practice. The
operator of the bow-and-drill depicted on a gilt glass vessel in the Vatican clearly
holds the middle of the stock of the drill in his left hand, where the thong of the bow
belongs; we expect to see him grasp the freely-rotating nave above the stock so that
the drill could turn under the action of the bow (figure 2.5). Either the artist has
made a mistake or an unknown variant has been depicted (the shaft of the drill bit
runs entirely through the stock and is spun from above). In this case the efficacy of
the drill would be greatly reduced by eliminating the downward pressure on the
nave; it seems most likely that the representation is in error (Goodman 1964: 161).
Certain depictions of complicated processes or their outcomes were necessarily
simplified or compressed due to the size of the rendition, the medium being used
by the artist, or the limitations of his ability. A common distortion of relative scale
must also be taken into account: a given tool, machine, or object being made was
routinely rendered disproportionately larger or smaller in relation to the artisan
depicted or the surrounding environment. For example, one may examine the
44 SOURCES
Figure 2.8. Attic red-figure hydria (Kalpis); the Caputi Hydria by the Leningrad Painter.
The male artisans are about to receive crowns by Athena and two winged victories. A
female artisan is at work on the far right of the scene. Ca. 475 B.C. Height 32 cm. Col-
lezione Banca Intesa, Milan, inv. F.G. 00002A-E/BI. (Line drawing adapted from Cloche
1931: pl. 21.)
carved rendition on Trajan's Column (A.O. 113) of the great bridge that spanned
the Danube to understand the principles used to construct broad, segmental arches
composed from beams of wood (figure 17.6). When compared to literary accounts,
such as that in Cassius Dio (13-1), or the actual site of the bridging project at Turnu-
Severin, there can be no doubt that the sculptor simplified his representation of the
actual structure by carving only five piers between the two riverbanks; Dio men-
tions that twenty were built. We may reasonably guess that the number of arced
segments rendered between each span was also reduced, but by how much? Is the
triangular bracing crowning each of the stone piers rendered accurately, or do
the diagonal beams we see, as Richmond has pointed out (1982: 35), seem to lean in
the wrong direction? How did a professional sculptor working in Rome know what
the bridge in Dacia looked like? Had sketches-or drawings of a more technical
nature-been brought back from the frontier? Frequently these problematic issues
cannot be definitively resolved, although considering them carefully will often help
to frame the correct questions when evaluating a representation of a given tech-
nical process.
Even when these limitations have been taken into account, however, one is
often struck by the level of detail and accuracy achieved in the smallest of rendi-
tions. A frescoed frieze course only about 20 cm high, for example, discovered in
1895 in the House of the Vettii in Pompeii, depicts, somewhat improbably, teams
of winged cupids and psychai engaged in a series of industrial activities: pressing
grapes and making wine, fulling, wreath-making, producing scented oils or per-
fumes, baking, and metalworking (figure 2.3). The images were located in one of
the most important rooms of the house, a formal reception and dining room
(oecus), situated to delight the guests and presumably to celebrate the lucrative
enterprises of the house's successful owners (Clarke 1991: 215; 2003: 101). In the
metalworking scene we see a cupid before a furnace as he holds not only the kind
of tongs seen in similar depictions but also a blowpipe to direct a stream of fresh air
on the hot charcoal; in this way the heat of the furnace can be rapidly elevated for
brief periods of time-useful for, among other things, soldering work. Despite the
fanciful nature of the operator, the wall-painter had probably observed a craftsman
working with metal at a hot furnace. To the left of the furnace scene another cupid
sits on a cushioned stool while working intently with a small hammer and anvil.
Before him is a box-shaped workbench that holds a set of scales and a small cabinet
with partially opened drawers. This is one of the most detailed depictions of such a
workstation from the Greco-Roman world, but its very uniqueness has made its
precise interpretation the subject of long debate: Do we see an office where coins
were minted, as some scholars have argued? Or is the scene that of a goldworker?
The answer to the question, never fully resolved, depends on modem interpretation
of the four principal activities and seven figures depicted. Marvin Tameanko (1990)
has argued that the apparatus resting on the top of the workbench is that of a Roman
caelator, or engraver of die-struck jewelry, and includes both scales and a unique
depiction of a magnifying glass suspended over the engraver's block. The validity
of this interpretation will not be explored here; the point is that an ancient wall-
painter creating a decorative frieze has included carefully rendered details of a spe-
cific technique-this is no generic depiction. It is important to remember that the
Pompeian wall-painter enjoyed no higher status as an "artist" than did the met-
alworker; presumably there were opportunities for observing each other's activities
and frequent social contact (for a discussion of status see Burford 1972).
The accuracy of the depiction of a given technology can be further tested by
comparing artifacts and site features with the image in question, as well as by con-
sidering variations over a range of images that show a given technique. The op-
eration of a kiln during the archaic period in the vicinity of Greek Corinth, for
example, is illustrated on at least eight painted plaques from the sanctuary of
Poseidon at Penteskouphia (Noble 1988: 151; Richter 1923: 76). Each depicts a single
nude male, in profile, tending to the firing process. These plaques and others like
them were evidently left as votive offerings for prosperity by potters and other
craftsmen. Although the images were created in the same area and are of roughly
the same date, variations in the otherwise similar depictions indicate that the artists
were not simply recreating an accepted "type" through mass copying. Where the
details have survived (in five of the examples), the single essential tool held by the
craftsman is a rod terminating in a hook, used variously to stoke the firebox or to
open or close the vent hole, which itself is always shown at the very top of the
beehive-shaped kiln. The door for inserting or removing the pottery is consistently
depicted in the side of the kiln, and is correctly shown about halfway up the wall of
the structure. In all but one example, the kiln is shown to be of modest height,
perhaps slightly taller than its operator, who needs to stand on the stoking cham-
ber to reach the vent hole. But in one example (Berlin inv. 802B), a ladder is needed
to reach to the top vent, itself rendered as a broken amphora; the impression is that
some establishments produced enough pots to warrant a kiln capable of firing
oversized pots or larger-scale operation (figure 2.9). In the most remarkable image,
unfortunately only partially preserved, the artist has provided us with a kind of
46 SOURCES
Figure 2.9. Corinthian pinax from Penteskouphia; a large kiln is depicted. The opera-
tor climbs a ladder to adjust the vent hole, which has been fashioned from the top of an
amphora. Sixth century B.C. Height 10.5 cm. Berlin inv. 8028. Bildarchiv Preussischer
Kulturbesitz/Art Resource. (ART317389.)
X-ray view of the kiln, which includes a depiction of the pots tightly stacked inside
the chamber, the perforated floor above the stoking shaft, the vent hole, and per-
haps even the test-pieces-fragments of unfired pottery-placed for easy access
during the firing so that the oxidation and reducing stages of the firing could be
carefully supervised (figure 2 .10; Berlin inv. 893; Cook 1961; Noble 1988).
RANGE OF DEPICTIONS
Not all technologies are represented in the extant corpus of representational art.
We lack, for example, depictions of gears and timekeeping devices, even though
artifacts and literary sources attest to their existence. Technologies that are por-
trayed are not necessarily depicted continuously over time, and for some industries
only one or two visual representations may survive (e.g., mosaic manufacture and
the water-powered water-lifting wheel). Since many ancient renditions are pre-
sumably lost, there may be no significance to the absence from the visual record of
a certain category of technological process. A surviving scene of glassblowing on a
mold-made Roman lamp, for example, suggests that multiple images might be
made of a relatively esoteric practice (figure 21.7). It is clear that the most signif-
Figure 2.10. Fragment of a Corinthian pinax from Penteskouphia depicting a section of a

kiln. Test pieces can be seen near the vent hole (top); sixth century B.C. Height 10.1 cm.
Berlin inv. F893. Bildarchiv Preussischer Kulturbesitz/Art Resource. (ART180898.)
icant initial development of representation in the Greco-Roman world first took

place in the sixth century B.C. with the maturation of Attic black-figure vase
painting and, especially after the second half of that century, the vase painter's
interest in including representations of everyday working life (table 2.1) . This
period of florescence lapsed by the end of the fifth century B.C., as artists moved to
media (such as wall painting) that have not survived well for study. There are few
depictions available from the Hellenistic world. Their absence might seem enig-
matic, since this period (late fourth through second centuries B.C.) was one that
fostered many technical innovations, including the experiments of Alexandrian
engineers and mathematicians. These Greek scientists were apparently prolific in
their inventions and their written commentaries, even though, as we have seen,
only a few of their works have survived through later copies. During the Hellenistic
period, however, ceramic artists turned from the representation of scenes of daily
life to floral motifs or to the imitation of embossed silverware by means of molds,
eliminating one of our best sources of representations of technology.
A given technology may not have been equally represented by both Greek and
Roman artists, or it may even be exclusive to one alone. Images of pottery pro-
duction were a well established theme for the Greek vase painter, but they are rare
in Roman contexts, and unlike the Greek examples are not shown on Roman pot-
tery itself but preserved in wall painting (e.g., Pompeii l.8.10, perhaps a shop sign;
Ciarallo and De Carolis 1999: cat. 178) and relief sculpture (Dobbins 1985; Kleiner
and Matheson 1996: cat. 55) . Similarly, while scenes including architecture are
shown on Greek artifacts, those depicting the process of construction are limited
to the Roman sphere, although even here there are enormous gaps; there exists no
depiction of the wooden centering framework for vault construction, for exam-
ple. Such disparities must reflect cultural attitudes; the incomplete nature of the
48 SOURCES
evidence alone does not suffice to explain the uneven pattern. The Roman's docu-
mentary approach to scenes of warfare and camp construction (Column of Trajan),
or his propensity to decorate his tomb with an image of his profession, has resulted
in a great number of accompanying portrayals of techniques, tools, and applications.
CATEGORIES OF DEPICTIONS
Table 2.1 outlines the cultural range and media of surviving depictions of technical
processes from the cultures of Greece, Etruria, and the greater Roman world and
also offers their rough chronological distribution. Images that offer information
about products of technology but nothing of process are not included (e.g., a
wooden ship versus an image of a ship under construction). Discussion below cat-
egorizes this body of material in a more thematic and formal way. The organizing
principle selected is their order of complexity, from the simplest of graphic depic-
tions that show a single element to those that incorporate multiple graphic com-
ponents within a single composition.
Tools Depicted in Isolation

Individual tools, including adzes, calipers, tongs, miniature plows, saws, planes,
and mining picks, are represented as three-dimensional miniature models or two-
dimensional representations on relief sculptures, mosaics, and coins. Instruments
shown in isolation or in groups (the latter commonly related by function) are gen-
erally found in specific contexts; most common is the realm of Roman funerary art
from the western provinces. The Greek vase-painter generally preferred to include
human agents along with depictions of tools, thus tools in isolation (unless part of
a larger scene) are missing from this otherwise important medium.
Small-scale models of tools tend to be found as votive offerings; examples exist
from Etruscan and Roman contexts. There may well have been similar objects
made as toys for which we have no record. Minute renditions of woodworking
tools are included as mint-marks on Roman republican-period silver coinage (Fava
1969). The miner's pick is depicted on the reverse of coins from the Greek city of
Damastion in Epirus (fourth century B.C.; Healy 1978). One, two, or more tools
carved on funerary reliefs of the Roman period served as pictograms that identified
the profession of the deceased. Instead of or in addition to a tool, the sculptor may
represent the product of a given technology, such as a spoked wheel, a leather shoe,
a boxwood comb, the hull of a ship-although, it must be stressed, many indus-
tries are never represented (e.g., Aquileia inv. 1231: wheel; Uffizi, Florence inv. 1914:
comb). In some cases these images are shown in isolation, in others the tools may
be accompanied by a portrait of the deceased (see below). Isolated tools are gen-
Table 2.1. Depiction of technological processes (G = Greek, E = Etruscan,
R=Roman).
Vase Minor Date Range
Technology Sculpture Painting Fresco Arts (by century)
Agriculture
Harvesting/pruning R G R 6th B.C.-4th A.D.
Plowing, Digging E,R G R 6th B.C.-4th A.D.
Presses R G R 6th B.C.-4th A.D.
Wine-Making R G R R 6th B.C.-4th A.D.
Boatbuilding R,G R 2nd B.C.-4th A.D.
Capstan R 1st A.D.
Ceramics
Pottery wheel G R 6th B.C.-1st A.D.
Pottery painting R G 6th B.C.-2nd A.D.
Pottery firing G 6th-5th B.C.
Pottery burnishing R G 6th B.C.-2nd A.D.
Construction Process
Laying Brick, Stone R R 2nd-4th A.D.
Turf walls R 2ndA.D.
Framing R 2ndA.D.
Scaffolding R R 2nd-4th A.D.
Cranes, Hoists R R 1st A.D.
Fishing, hunting, trapping G,R G E,R R 6th B.C.-4th A.D.
Food Preparation
Butchering R G E 6th B.C.-2nd A.D.
Baking & Cooking G,E,R E G 5th B.C.-3rd A.D.
Furnaces/Ovens/Kilns R G R G 6th-4th B.C.
Glassblowing R 1st-2nd A.D.
Levers R G R 6th B.C.-1st A.D.
Leatherworking, shoes R G 6th B.C.-2nd A.D.
Metalworking
Forge R G R 6th B.C.-2nd A.D.
Bellows G,R G 6th B.C.-1st A.D.
Chasing G R 6th/5th B.C.
Assembly (statues) G 6th/5th B.C.
Coin/Minting R R? 1st B.C.-1st A.D.
Embossing R G 6th/5th B.C.-1st A.D.
Gold Beating R 1st A.D.
Jewelry Production R 1st A.D.
Military Machines R 2ndA.D.
Milling
Animal power R 1st B.C.-2nd A.D.
Pestles or Querns G G 6th B.C.
Water Power R R 3rd-5th A.D.
Mining R G G 6th B.C.-2nd A.D.
Mosaic Manufacture R 4thA.D.
(continued)
Table 2.1. (continued)
Vase Minor Date Range
Technology Sculpture Painting Fresco Arts (by century)
Painting and Plastering R G R R 5th B.C.-2nd A.D.

(portraits, walls)
Pulleys, Block and Tackle R 1st-2nd A.D.
Rope Making R 1st-2nd A.D.
Stone Carving R G R 5th B.C.-4th A.D.
Textiles 6th B.C.-1st A.D.
Bleaching, Pressing R 1st A.D.
Fulling, Dyeing R R 1st-3rd A.D.
Rolling, Spinning R G E 7th B.C.-2nd A.D.
Looms/Weaving R G R E,R 7th B.C.-4th A.D.
Shearing/cropping R 1st-2nd A.D.
Water Supply
Aqueducts R 2ndA.D.
Fountains R G E,R 6th B.C.-2nd A.D.
Water-Lifting
Compartmented wheel R 1st-2nd A.D.
Archimedean Screw R R 1st B.C.-3rd A.D.
Weights/Scales R G R R 6th B.C.-3rd A.D.
Wheelwright R 2ndA.D.
Woodworking
Tools only R R 5th B.C.-4th A.D.
Craftsman/men & tools R G R R 6th B.C.-4th A.D.
erally not found on public monuments of Greece and Rome, with the main ex-
ception being those carved on the faces of Roman-period altars. Of this group
many depictions (e.g., a knife) were connected directly with the ritual of sacrifice
performed at the altar, while others on altars made reference to the guild re-
sponsible for the donation of the altar (e.g., Capitoline Museums inv. 1909).
Single Operators with a Tool or Machine

This group includes small models of figurines engaged in specific activities such as
plowing, sawing wood, or shoemaking. Such figurines had a wide distribution, in-
cluding Bronze Age sites from Egypt and the Aegean, classical Greece, and Italy. The
context of such finds is generally funerary or religious in nature; the figurine may
represent activity in the afterlife (here the Egyptian material seems unambiguous), or
the votive is left by the petitioner to protect a given activity. Attic red-figure vase
painters from the late archaic and early classical periods portrayed a single craftsman
building furniture or carving stone columns (Gallatin Painter, figure 2.6; Antiphon
Painter, ca. 475 B.C., Boston MFA 62.613). We do not find, however, Greek depictions
of operators engaged in industrial activities on stone funerary markers from the
archaic and classical periods, surely because such representation was reserved for
wealthy aristocrats, who were not depicted in association with such manual skills.
The situation is entirely different in Roman Italy, particularly from the period of the
early empire onward and the concomitant appearance of an affluent artisan class of
former slaves and their descendants (George 2006). On Roman-period funerary
reliefs, depictions of single operators and their tools focus on the deceased practicing
his trade. An outstanding example is that of Longidienus from Ravenna, which not
only portrays a Roman boat-builder hard at work with his adze, but also provides a
descriptive caption: "P. Longidienus, son of Publius, pushes on with his work"
(figure 17.1; Ravenna, inv. 7; CIL 11.139; Ulrich 2007; Zimmer 1985: pl. 7). In other
examples the operator may only pose with one of his tools in hand, identifying
himself as a skilled craftsman; the presence of the handheld object identifies the
primary occupation of the deceased. Such images were distributed widely in the
Roman West and are indictors of affluence; the vast majority of Roman manual
laborers could afford only a simple burial in a communal sepulcher, perhaps with a
notebook-sized marble plaque recording his name and profession. Monuments of
artisans with depictions of their trades must have been relatively rare, or clustered in
cemeteries where there was an established pictorial tradition. Cemeteries connected
with Roman legionary posts tend to yield such concentrations of funerary images; the
site of Roman Aquileia in northern Italy provides an outstanding example.
Depictions of Complex Operations

The portrayal of a group of individuals involved in a more or less common en-
terprise was a favorite topic of the Greek vase-painter, and of the Roman sculptor
and fresco-painter working for private patrons. The layout tends to fall into the
category of the continuous frieze, and occasionally superimposed registers, from
the small scale of the figural band around a Greek pot to the larger sculpted and
painted examples seen in both Roman private and, rarely, public art. Even small
objects measuring only a few centimeters on a side can contain imagery that pro-
jects several stages of a single operation. The mining scene from a Corinthian pinax
of the mid-sixth century B.C. provides an excellent example (figure 2.1; Berlin inv.
871). On a field measuring only 10.4 x 13.2 cm, the artist has depicted several stages
of a mining operation for the extraction of ore or clay, showing four figures who
swing a pickaxe, gather the raw material in baskets, and remove them from the pit,
while including such details as the steps cut into the sides of the mining shaft and
an amphora strung up alongside the miners, presumably to provide drinking water
to relieve the hot and dirty work. More typical from the Greek world are the black-
and red-figure vase paintings that depict activities such as weaving, bronze-casting,
or pottery-making (figure 2.11: Amasis Painter, with eleven figures; figure 4.4:
Foundry scene, with two figures; figure 16.1-3: Foundry cup, with eight figures; not
illustrated: Leagros Group, with eight figures, Munich inv. 1717).
The most important Roman scenes to have survived include wall paintings and
relief sculpture. From the late first century B.C., we have a continuous frieze of a
large-scale bread-making operation carved on the tomb of the baker Eurysaces from
Rome, which offers a unique portrayal of a large-scale industrial operation that
ack-figure scene of women weaving, attributed to the Amasis Painter. Metropolitan Museum of Art, New '
politan Museum of Art, P32367.)
includes a kind of assembly-line mentality for the milling of wheat, kneading, baking,
weighing, and distribution of a common and essential commodity (figure 2.2; Ros-
setto 1973). With this relief Eurysaces proclaimed publicly that his accumulation of
wealth was due not to a highly skilled craft but to the perfection of an orchestrated
process of mass production that created wealth from high volume sales of a simple
product with a relatively low profit margin. Such scenes of large-scale production
emphasize the success of a particular individual in organizing and managing the
laborers who work below him (George 2006: 24). Eurysaces' tomb was in fact built to
hold the ashes of the baker's wife, Atistia, but his full-length portrait was probably
included in the decoration of the tomb, and a figure who supervises the kneading
process, dressed in a toga to emphasis his elevated status, may be that of Eurysaces
himself. Another first-century marble relief found in Rome, of a furniture workshop,
shows six laborers, their tools-which include a large mechanized sawing machine,
frame saw, calipers, and square-and furniture being assembled. The proud patron
stands off to one side, in the presence of his divine patroness, Minerva (figure 17.3). A
fragmentary frescoed scene of pottery production from Pompeii, perhaps intended as
a shop sign, depicts at least four workers shaping vessels in one frame while in another
a seated female appears to offer one of the pots for sale (see above).
Trajan's Column in Rome provides an important glimpse into large-scale
building activities undertaken by a disciplined and organized military labor force.
It seems unusual that more such scenes were not set up in public spaces by
Hellenistic kings or later Roman dictators and emperors, particularly since both
Assyrian kings of the eighth century B.C. and Egyptian pharaohs had clearly
promoted art depicting monumental engineering feats involving large numbers of
orchestrated workers (Barnett et al. 1998). Other canonical iconography of eastern
kingship had been adopted by Greek and Roman potentates. The actions of Tra-
jan's soldiers must of course be seen in the larger context of the Column's grand
purpose, to celebrate the emperor's military victory over the Dacians, but the
ultimate victory has been portrayed not only in terms of prowess on the battlefield
but also of preparation and logistics: clearing forests, building camps and abut-
ments for artillery, and crossing rivers.
TECHNOLOGIES REPRESENTED
IN MYTHOLOGICAL OR
FANTASTIC SETTINGS
Certain Greek myths and their Roman adaptations, as well as Rome's own native
tradition of stories, include colorful accounts of technological and engineering
feats. These are reproduced primarily on extant vase painting, frescoes, and mosaics.
54 SOURCES
Some of the associated technological wonders portrayed, such as the wings worn by
Icarus and his master-craftsman father Daedalus, or the wooden shell of the Trojan
horse, are clearly renditions of the artist's imagination, his familiarity with a long
pictorial tradition, or extrapolation from parallel and contemporary technologies
(thus the planked hull of a wooden ship could form the basis for the rendition of
the torso of a wooden horse). Other compositions include images of tools and
devices clearly expropriated from a given artist's contemporary world.
Images related to the feats of the hero Heracles offer illustrative examples. The
hero and his companion Iolaos attack the monstrous Hydria with a long sword and
a toothed sickle, respectively, on an early Corinthian skyphos now in Paris (Louvre
CA 3004; Kron 1988: 190). Another Attic black-figure pot from around 520 B.C.
shows Heracles carrying the Cercopes suspended from either end of a straight
wooden yoke balanced on his shoulders (Brommer 1984, 29). This must have been
a common method of transporting heavy loads: on another pot from around 500
B.C. we see what appears to be a contemporary Greek carrying two large tuna sus-
pended exactly as the Cercopes were carried, and in yet another scene a two-man
rig bears a heavy storage vessel suspended from a pole (Hopper 1979: figs. 23, 24).
The iconography associated with Hephaestus and his forge provides images of
tools and procedures associated with smithing, particularly on Greek painting and
sculpture. The Berlin Foundry Cup shows the god putting finishing touches on the
armor of Achilles (figure 16.1); a black-figure oinochoe from the sixth century B.C.,
now in London, appears to show the god (or a mortal practitioner) working at his
forge while associated tools seemingly "float" in the background (figure 4.4;
Mattusch 1980; BM B507). In London and on the frieze of the Siphnian treasury
at Delphi (ca. 525 B.C.), we can see depictions of a bellows made from an inflatable
animal skin (Brommer 1978: pl. 50). The myth of Minerva and Arachne is the set-
ting for scenes of women spinning and weaving on a marble frieze from Domitian's
Forum in Rome (late first century; D'Ambra 1993).
A related category of images does not represent specific myths but uses fanciful
or exotic figures in otherwise quotidian settings, including activities of a techno-
logical nature. In Greek art, examples of this genre are numerous but rather limited
in scope: the Attic vase painter was fond of satyrs and winemaking scenes for
decorating vessels intended for use in symposia. The Roman affection for "charm-
ing" scenes involving beings such as cupids or pygmies has roots in the Hellenistic
artist's attraction to subjects of infants, of allegory, and of abnormal physique and
ethnic physiognomy (Pollitt 1986: 127). The Nilotic scenes favored by Roman-
period painters and mosaicists, often populated by caricatures of the pygmy race,
likely represent a category first popularized in hellenized Alexandria in the late
third or early second century B.C. (Pollitt 1986: 208). It has already been noted that
this theme was imported to include renditions of the Archimedean water-screw,
apparently a device that was a picturesque feature of the Nile delta. Similar is the
portrayal of cupids (or erotes) and occasionally psychai (their female winged coun-
terparts) in activities of an agricultural and even "industrial" nature; the richly-
detailed frieze in the House of the Vettii (see above) is a prime example (figure 2.3).
The use of erotes is so widespread in a range of contexts, public and private, sacred
and funerary, that their presence as agents in scenes of shoemaking, pressing grapes,
or blacksmithing would not have caused much comment. Nonetheless, their em-
ployment in such scenes has the effect of elevating rather gritty tasks to a level suit-
able for the decoration of a wealthy artisan's reception room-or that of one who
has derived his wealth through commerce. There is a playfulness and charm to such
scenes that is hard to imagine in the poorly lit and often dangerous conditions we
suspect characterized the ancient workshop.
WORKFORCE
With few exceptions, images of industrial activities were made by private indi-
viduals for private commemoration. In a sense these are self-portraits of the
working class made by those who shared their social station. The images place in
center stage the human element that is so integral to these early machines and
processes. Once again we must exercise caution in drawing conclusions from the
manner in which the ancient craftsman or builder is portrayed; contemporary
artistic convention accounts for much. With such caveats in mind, however, the
visual record can provide information otherwise not available about the men,
women, and also children engaged in ancient technologies.
One is not surprised by the dominance of women in Greek scenes of spinning
and weaving, given the well-documented domestic duties of women in ancient
Greek society (Barber 1992; Keuls 1983). Adolescent and adult males are predom-
inant in industries outside the home; this is true of both the Greek and Roman
workforce. Nevertheless, women and children are not excluded from activities one
might predict as exclusively male domains. The Corinthian pinax from Berlin cited
above (figure 2.1) clearly shows a woman-perhaps a slave-taking part in re-
moving material from a mine, the Caputi Hydria shows a woman at work in a
potter's studio (even though the scene, replete with winged Nikai, is highly ide-
alized; cf. Venit 1988). In the Roman context, we find women involved with the
distribution of food and drink on the retail level (Amedick 1991: cat. 97; Kampen
1981) and women portrait-painters, and a Roman relief in Naples shows a child
alongside other laborers in a shop that manufactured metal containers (Ciarallo
and De Carolis. 1999: cat. 201, 335; Veyne 1987: 54; Naples inv. 6575 and 9018).
Within both Greek and Roman scenes where multiple figures are at work, a
simple gradation of status that might simply be described as manager versus la-
borer is often apparent. An owner or overseer, if present, watches over, but does
not actively engage in, the activity shown; two examples from Rome have been
discussed above (figure 17.3). The patron is inconsequential to the process be-
ing portrayed, but his presence may well account for the very existence of the
56 SOURCES
depiction. The ancient viewer would have identified the patron immediately from
his dress; the garment worn, usually of full length, is not suitable for a manual
laborer or his class, whose "uniform," if a garment is worn at all, is utilitarian. Even
so, the full drapery of the patron or overseer may well be worn by a figure who is
himself of servile origin.
CLOTHING AND PERSONAL

ACCESSORIES
The clothing and special equipment worn by the ancient operator is not only an
indicator of social status. Working conditions, as well as issues of utility and even
of safety, can be surmised from the pictorial evidence. The Attic potter showed
many of his subjects at work firing pottery, mining, woodworking, blacksmithing,
and plowing wearing nothing at all. In some such scenes (woodworking, black-
smithing, harvesting) other workers within the same composition are wearing a
short strip of drapery around the waist, but are otherwise unclad. It has been
recognized that Greek athletes commonly exercised while nude, but one must
wonder about the comfort and safety of working without any protective clothing.
Nudes and the semi-draped in industrial scenes may indicate slave and free labor,
respectively, or the nudity may simply be part of the artistic convention that
frequently portrayed the male as unclad. In every scene that shows a Greek work-
man in front of a furnace or kiln, he is nude. Greek women at work, it should be
noted, are always fully garbed, this, too, a convention for other images of women
made in the archaic and classical periods. The peplos or chiton worn is the same as
those worn by women in other painted scenes. Those worn in scenes of weaving or
spinning might be rendered with decorative patterns, as if to emphasize the skill of
the female weavers (figure 2.11). Since women of status in the Greek world spent
working hours at their home-based looms, the fine garments are hardly out of
place.
The Roman worker, however humble, is always garbed. The simplest covering
is similar to that worn by his Greek counterpart: the short, kiltlike garment that
leaves him bare-chested (milling, baking, woodworking, construction scenes, etc.).
The fact that many artisans working in Italy were in fact of Greek origin, slaves and
freedmen, may be reflected in this common and uncomplicated costume. The bare
upper torso is common to scenes where the working environment was hot, such as
around the baker's oven and the smithy's forge; I have noted above the Greek
artist's treatment of analogous scenes. The most common form of artisan's cos-
tume of the Roman period is the short tunic, a one-piece garment, about knee
length, sleeveless or with short sleeves; it seems to have changed little over centuries
of use. Clothing that can be described as protective in function can be cited from a
Figure 2.12. Sandstone relief depicting miners from Palazuelos, near Linares, Spain. Ro-
man period, first or second century. The garments worn by the figures appear to include
panniers. (Sandars 1905: pl. 69.)
small and diverse group of monuments from Roman sites: an Ostian tomb shows a
man wearing a heavy apron while honing a tool (Zimmer 1985, pl. 4), as does a
miller depicted on a funerary monument from Senon (Belgica). A helmet of sorts is
worn by one of the men rigging the crane on the Tomb of the Haterii, while a stone
relief of miners found in Spain shows the workmen wearing a heavy belt, appar-
ently with deep pockets suitable for tools or even chunks of ore (figure 2.12;
Rickard 1928; Sandars 1905). A bronze statuette of a plowman from Trier shows the
field laborer protected by a hooded, heavy cape, probably made of leather (figure
2.13; Wild 1985).
A serendipitous combination of factors created the conditions under which a rich

corpus of graphic material has survived from the Greco-Roman world for modern
scholars of ancient technologies to study. The cultures of both Greece and Rome
fostered an enduring tradition in the figural arts. No less important is the fact that
representations of human activity included so-called genre scenes that documented
activities associated with daily life. These everyday scenes were for the most part
painted, carved, or modeled on cheap but durable media: ceramics or stone. It is
ironic that other representations by famous artists on the finest of materials-
precious metals, fine linens, or polished and fragile frescoes-have, in the majority
of cases, not survived. By the Roman period, the rise of an affluent artisan class that
tapped into this long-established figural tradition and imitated the elite's penchant
58 SOURCES
Figure 2.13. Bronze statuette of a plowman from Trier, Germany, wearing what appears to
be a leather cape. Height n.8 cm. Second-third century. Rheinisches Landesmuseum,
Trier. (Photograph: Rheinisches Landesmuseum Trier.)
for commemoration produced many of the seminal pieces now so important for
the reconstruction of ancient technologies.
REFERENCES
Adam, J.-P. 1994. Roman building: Materials and techniques. Bloomington: Indiana Uni-
versity Press.
Amedick, R. 1991. Die Sarkophage mit Darstellungen aus dem Menschenleben. Vol. 4, Vita
Privata. Berlin: Gehr. Mann.
Barber, E. J. W. 1992. "The Peplos of Athena," in Goddess and polis: The Panathenaic
Festival in ancient Athens. Princeton: Princeton University Press, 103-17.
Barnett, R. D., E. Bleibtreu, and G. Turner 1998. Sculptures from the Southwest Palace of
Sennacherib at Nineveh. 2 vols. London: British Museum Press 1998.
Bli.imner, H. 1875-1887. Technologie und Terminologie der Gewerbe und Kunste bei Griechen
und Romern. 4 vols. Leipzig: B. G. Teubner.
Brommer, F. 1978. Hephaistos: Der Schmiedegott in der antiken Kunst. Mainz: Philipp von
Zabern.
Brommer, F. 1984. Herakles II: Die unkanonischen Taten des Helden. Darmstadt: Wis-
senschaftliche Buchgesellschaft.
Burford, A. 1972. Craftsmen in Greek and Roman society. Ithaca, NY: Cornell University
Press.
Carettoni, G, A. M. Colini, L. Cozza, and G. Gatti 1960. La Pianta Marmorea di Roma
Antica. Rome: Comune di Roma.
Ciarallo, A., and E. De Carolis. 1999. Homo Faber: Natura, scienza e tecnica nell'antica
Pompei. Milan: Electa.
Clarke, J. 1991. The houses of Roman Italy, 100 B.C.-A.D. 250. Berkeley: University of
California Press.
Clarke, J. 2003. Art in the lives of ordinary Romans. Visual Representation and non-elite
viewers in Italy, 100 B.C.-A.D. 315. Berkeley: University of California Press.
Cloche, P. 1931. Les classes, les metiers, le trafic. Paris: Les Belles Lettres.
Cook, R. M. 1961. "The 'double-stoking Tunnel' of Greek kilns," Annual of the British
School at Athens 56: 64-67.
Coulton, J. J. 1977. Ancient Greek architects at work. Ithaca, NY: Cornell University Press.
D'Ambra, E. 1988. "A myth for a smith: A Meleager sarcophagus from a tomb in Ostia,"
American Journal of Archaeology 92: 85-100.
D'Ambra, E. 1993. Private lives, imperial virtues: The frieze of the Forum Transitorium in
Rome. Princeton: Princeton University Press.
D'Ambra, E., and G. P.R. Metraux 2006. The art of citizens, soldiers and freedmen in the
Roman world. British Archaeological Reports, Intl. Series S1526. Oxford: Archaeopress.
Dobbins, J. J. 1985."A Roman funerary relief of a potter and his wife," Arts in Virginia 25:
24-33.
Fava, A. S. 1969. I simboli nelle monete argentee republicane e la vita dei romani. Torino:
Soprintendenza alle Antichita del Piemonte e Museo Civico di Torino.
Forbes, J. 1955-1964. Studies in ancient technology. 9 vols. Leiden: Brill.
Gaitzsch, W. 1980. Eiserne romische Werkzeuge: Studien zur romischen Werkzeugkunde in
Italien und den nordlichen Provinzen des Imperium Romanum. 2 vols. British Ar-
chaeological Reports, Intl. Series S78. Oxford: BAR.
George, M. 2006. "Social identity and the dignity of work in freedman's reliefs," in
D'Ambra and Metraux 2006: 19-29.
Goodman, W. L. 1964. The history of woodworking tools. London: G. Bell and Sons.
Green, R. 1961. "The Caputi Hydria," Journal of Hellenic Studies 81: 73-75.
Haselberger, L. 1994. "Ein Giebelriss der Vorhalle des Pantheon: Die W erkrisse vor dem
Augustusmausoleum," Romische Mitteilungen 101: 279-308.
Healy, J. 1978. Mining and metallurgy in the Greek and Roman world. London: Thames and
Hudson.
Hopper, R. J. 1979. Trade and industry in classical Greece. London: Thames and Hudson.
Humphrey, J., J.P. Oleson, and A. Sherwood 1998. Greek and Roman technology: A
sourcebook. New York: Routledge.
Huxley, G. 1959. Anthemius of Tralles: A study of later Greek geometry. Greek, Roman and
Byzantine Monographs 1. Cambridge, MA: Harvard University Press.
Jensen, W. 1978. "The sculpture from the Tomb of the Haterii," 2 vols. Diss., University of
Michigan.
Kampen, N. 1981. Image and status: Roman working women in Ostia. Berlin: Gebr. Mann.
Keuls, E. 1983. "Attic vase painting and the home textile industry," in Moon 1983: 209-30.
60 SOURCES
Kleiner, D. E., and S. B. Matheson (eds.) 1996. I, Claudia: Women in ancient Rome. New
Haven: Yale University Art Gallery.
Klemm, F. 1964. A history of western technology. Cambridge, MA: MIT Press.
Kron, U. 1988. "Sickles in Greek sanctuaries: Votives and cultic instruments," in R. Hagg,
Ancient Greek cult practice from the archaeological evidence. Acta Instituti Atheniensis
Regni Sueciae 8.15: 187-216.
Krukowski, S. 1990. Pots on pots: Images of pottery-making processes on ancient Greek vases.
http://www.cm.aces.utexas.edu/faculty/skrukowski/writings/pots.html.
California Press.
Mattusch, C. 1980. "The Berlin Foundry Cup: The casting of Greek bronze statuary in the
early fifth century B.C.," American Journal of Archaeology 84: 435-44.
Mattusch, C. 1996. The fire of Hephaistos: Large classical bronzes from North American
collections. Cambridge, MA: Harvard University Art Museums.
Moon, W. (ed.) 1983. Ancient Greek art and iconography. Madison: University of Wisconsin
Press.
Neuburger, A. 1930. The technical arts and sciences of the ancients. Trans. H. L. Brose. New
York: Macmillan. Originally published as Die Technik des Altertums (Leipzig: 1919).
Noble, J. V. 1988. The techniques of painted Attic pottery. London: Thames and Hudson.
Oleson, J.P. 1984. Bronze Age, Greek and Roman technology: A select, annotated bibliog-
raphy. New York: Garland.
Oleson, J.P. 2004. "Well pumps for dummies: Was there a Roman tradition of popular,
ambito romano. Como: Comune di Como, Musei Civici, 65-86.
Oleson, J.P. 2005. "Design, materials, and the process of innovation for Roman force
pumps," in J. Pollini (ed.), Terra marique: Studies in art history and marine archaeology
in honor of Anna Marguerite Mccann. Oxford: Oxbow, 211-31.
Pollitt, J. J. 1986. Art in the Hellenistic age. Cambridge: Cambridge University Press.
Richter, G. M.A. 1923. The craft of Athenian pottery. New Haven: Yale University Press.
Rickard, T. A. 1928. "The mining of the Romans in Spain," Journal of Roman Studies 18:
129-43.
Richmond, I. 1982. Trajan's army on Trajan's Column. London: British School at Rome.
Rossetto, P. C. 1973. Il Sepolcro del Fornaio Marco Virgilio Eurisace a Porta Maggiore. Rome:
Istituto di Studi Romani Editore.
Sandars, H. 1905. "The Linares bas-relief and Roman mining operations at Baetica," Ar-
chaeologia 59: 311-32.
Schreiber, T. 1999. Athenian vase construction: A potter's analysis. Malibu, CA: J. Paul Getty
Museum.
Singer, C., E. J. Holmyard, A. R. Hall, and T. Williams (eds.) 1954-1956. A history of
technology. Vol. 1, From early times to fall of ancient empires; Vol. 2, The Mediterranean
civilizations and the Middle Ages. Oxford: Oxford University Press.
Strong, D., and D. Brown (eds.) 1976. Roman crafts. New York: New York University Press.
Tameanko, M. 1990. "Goldsmith's, mint, or jewelry factory? A new interpretation of the
wall painting from the House of the Vettii, Pompeii," Minerva 1: 42-46.
Taylor, R. 2003. Roman builders: A study of architectural process. Cambridge: Cambridge
University Press.
Ulrich, R. 2007. Roman woodworking. New Haven: Yale University Press.
Venit, M. S., 1988. "The Caputi Hydria and working women," Classical World 81: 265-72.
Veyne, P. (ed.) 1987. A history of private life. Vol 1, From pagan Rome to Byzantium.
Cambridge, MA: Harvard University Press.
White, K. D. 1984. Greek and Roman technology. Ithaca, NY: Cornell University Press.
Wild, J.P. 1985. "The clothing of Britannia, Gallia belgica and Germania inferior," in
Aufstieg und Niedergang der romischen Welt ll.12.3. Berlin: Walter de Gruyter, 362-422.
Zimmer, G. 1982. Romische Berufsdarstellungen. Archaologische Forschungen 12. Berlin:
Gebr. Mann.
Zimmer, G. 1985. "Romische Handwerker," in Aufstieg und Niedergang der romischen Welt
ll.12.3. Berlin: Walter de Gruyter, 205-28.
CHAPTER 3
HISTORIOGRAPHY
AND THEORETICAL
APPROACHES
KEVIN GREENE
THE BEGINNINGS OF A "HISTORY OF

ANCIENT TECHNOLOGY"
Historiography
In Ancient Society, Lewis Morgan (1877) outlined the evolution of society through
stages of savagery, barbarism, and civilization on anthropological rather than ar-
chaeological grounds, and he subsequently played an important part in shaping
the ideas of Engels and Marx. His opening chapter concluded by looking back from
the 1870s, by which time North America, along with Europe, was caught up in the
Industrial Revolution. He presented an interesting list of contributions made by
Greece and Rome-despite being "found deficient in great inventions and dis-
coveries." The emphasis on social and political as well as material items conforms
to modern expectations and is also reminiscent of Douglass North's institutional
approach to economic history (1990).
Passing over the mediaeval period, which gave Gothic architecture, feudal aris-
tocracy with hereditary titles of rank, and a hierarchy under the headship of a
pope, we enter the Roman and Grecian civilizations. They will be found deficient
in great inventions and discoveries, but distinguished in art, in philosophy, and in
HISTORIOGRAPHY AND THEORETICAL APPROACHES 63
organic institutions. The principal contributions of these civilizations were im-

perial and kingly government; the civil law; Christianity; mixed aristocratical and
democratical government, with a senate and consuls; democratical government
with a council and popular assembly; the organization of armies into cavalry and
infantry, with military discipline; the establishment of navies, with the practice of
naval warfare; the formation of great cities, with municipal law; commerce on the
seas; the coinage of money; and the state, founded upon territory and upon
property; and among inventions, fire-baked brick, the crane, the water-wheel for
driving mills, the bridge, acqueduct and sewer; lead pipe used as a conduit with
the faucet; the arch, the balance scale; the arts and sciences of the classical period,
with their results, including the orders of architecture; the Arabic numerals, and
alphabetic writing. (Morgan 1877: 30--31)
Morgan reminds us that-unlike Neolithic farming communities, urban civi-

lizations on the Euphrates, or early modern European states engaged in global
trade-classical civilization did not rely on new forms of settlement, technological
developments, or new sources of raw materials. Indeed, a modern consensus con-
siders Greek and Roman culture to have been deeply conservative, disdainful of
labor, and lacking in the spirit of enterprise associated with the growth of capitalism
(Finley 1973). Much of the interest of Greek and Roman technology arises from a
contradiction between this apparent ideological devotion to the pursuit of honorable
leisure and the practical achievements of Greek city-states, Hellenistic kingdoms,
and the Roman Empire. All three states equipped their armies with plentiful weapons
and body armor, built fortifications and civic buildings designed by skilled archi-
tects, installed systems for the supply and disposal of water, and maintained de-
pendable communications to deliver agricultural and industrial goods by land, river,
or sea. Analysis of this contradiction is fundamental to understanding Greek and
Roman technology in its wider socioeconomic context, and it is not surprising
to find variations in interpretation at different stages in the history of its study. This
chapter will begin by examining work carried out on Greek and Roman technology
in recent decades, before considering how technology has come to be perceived as
an integral component of the society and economy of the classical world.
Where Are We Now?

In 1990, I wrote an article about progress in recent publications concerning an-
cient technology and its interpretation. The opening sentence-"1984 was the year
in which the study of Roman technology emerged from a period of relative
obscurity"-referred to publications by K. D. White, Wikander, and Oleson that
differed in their attitude to Moses Finley's influential assessment of ancient tech-
nological stagnation (1965). Oleson (1984), having reviewed Greek and Roman
water-lifting devices, concurred with Finley that ancient technology was, in general,
a failure for social reasons (although he later changed his opinion; Oleson 2000:
287-88). Wikander made use of growing numbers of archaeological discoveries of
Roman water-mills to challenge Finley's judgment that ancient inventions were
64 SOURCES
not innovated to become technology-in-use. White's Greek and Roman Technology

brought together a lifetime's work on archaeological evidence and classical texts-
much of it summarized in more than 60 pages of appendixes and tables-whose
sheer quantity invited a paradigm shift away from Finley. "The reader of these
pages, having taken measure of the innovations and development of inventions
reviewed in the chapters and tables, may be inclined to take a less unfavourable
view of the standards reached by classical architects, builders and engineers, as well
as by the farmers, the food processors and the men who built the ships that carried
the products of their labours" (K. D. White 1984: 173).
These three publications were singled out because they combined technical and
documentary evidence and concurred (implicitly or explicitly) with White's "less
unfavourable view." More have appeared since 1984: Hill (A History of Engineering
in Classical and Medieval Times, 1984; Islamic Science and Engineering, 1993) and
M. J. T. Lewis (Millstone and Hammer, 1997; Surveying Instruments of Greece and
Rome, 2001) have made particular use of Arabic sources, which are in some cases the
only record of Greek technical treatises. Oleson, Lewis, and Wikander were all
contributors to a substantial volume titled Handbook of Ancient Water Technology
(Wikander 2000 ), which consolidated an enormous amount of archaeological and
documentary work conducted in the later twentieth century. Meissner published a
comprehensive survey of the nature and context of ancient technical literature
in 1996, and specific Roman texts have been subjected to close scrutiny. Fleury's
La mecanique de Vitruve (1993), which analyzed mechanical devices described in
book 10 of De architectura, was followed by his study of Vitruvius' technical terms
(with Callebat, 1995); a new edition of the entire work, enlivened by diagrams, has
been edited by Rowland and Howe (2001). Frontinus' De aquis urbis Romae has also
been the subject of discussion in a collection of papers edited by Blackman and
Hodge (2001), a convincing analysis of its context and purpose by Peachin (2004),
and comprehensive new editions by Rodgers (2004) and Del Chicca (2004). An-
drew Wilson's paper "Machines, Power and the Ancient Economy" (2002) showed
that scholars with expertise in ancient technology could place specialist findings in a
clear social and economic setting-and communicate them to a wider audience in
the influential Journal of Roman Studies.
Many studies of specific technologies appeared during the 1990s and the early
years of this century. A good example is harnessing and traction, where analyses of
artifacts and iconographic evidence have been followed up by experimental studies
with animals. Raepsaet and Rommelaere (1995) Raepsaet and Lambeau (2000), and
Raepsaet (2002) have helped counter the widespread condemnation of premedieval
harnessing by historians of technology (cf. chapter 23). New understanding of
traction methods used in transport and in farming-like Wikander's review of
evidence for water-mills-has contributed to a more positive perception of Greek
and Roman capabilities. Architecture can also be seen in a new light, thanks to the
deduction of working practices from surviving structures in Coulton's Ancient
Greek Architects at Work (1998), Hellmann's L'architecture grecque (2002), Adam's
La construction romaine (1984; English translation 1994), and Lancaster's Concrete
Vaulted Construction in Imperial Rome (2005). The logistics of a major Roman

imperial building project have been elucidated in DeLaine's The Baths of Caracalla
(1997), while Shirley's Building a Roman Legionary Fortress (2001) takes a similar
approach to the very different problem of marshalling building materials (mainly
timber) in newly conquered territory in Scotland in the late first century A.D.
The availability of detailed technical works has fortunately been accompanied
by a number of introductions to Greek and/ or Roman technology. French, German,
and Italian readers enjoy a range of accessible and authoritative books: Sciences et
techniques a Rome (Chevallier 1993), Einfiihrung in die antike Technikgeschichte
(Schneider 1992, with expert guidance on further reading), La tecnica in Grecia ea
Roma (Traina 1994, with a strong cultural and historiographical perspective), and
Tecnica e tecnologia nelle societa antiche (Gara 2002). In addition, a revised edition
of Landels' Engineering in the Ancient World appeared in 2000. Access to original
sources referred to in such books has been facilitated by a well-organized annotated
sourcebook, Greek and Roman Technology (Humphrey et al. 1998). Working rep-
licas of Greek and Roman technical devices have become a popular feature of
television programs and resulted in a book (Wilkinson 2000) linked to the BBC's
series What the Romans Did for Us, which subsequently broadened its scope to
include Greece and other civilizations (Hart-Davis 2004). Conversely The Day the
World Took Off-a sophisticated analysis of the origins of the Industrial Revolution
broadcast by Channel 4 in Britain in 2000-skipped directly from the medieval
period to Neolithic farming. Hellenistic automata were not even mentioned when
comparable eighteenth-century devices were credited with preparing minds for
mechanization (Dugan and Dugan 2000: 40-42). At the end of 2005, automata and
other products of Greek science and technology were displayed in a major exhi-
bition (unsurprisingly called Eureka!) at the National Archaeological Museum in
Naples, accompanied by a lavishly illustrated catalogue containing scholarly essays
(Lo Sardo 2005). Thus, the academic study of sources and evidence for Greek and
Roman technology-and awareness among nonspecialists-has made excellent
progress in recent years.
How Did We Get Here?

The history of archaeology can only be understood in the context of wider intel-
lectual history (Trigger 2006), and the same is true of the study of technology. A
sense of awe in the face of classical art, science, technology, and literature charac-
terized European culture from the early medieval period. Schneider (1992: 17-30)
traced the history of the subject back to the Renaissance scholar Polydor Vergil (De
rerum inventoribus, 1499), who included technology in a much wider study of
inventions. From Cyriac of Ancona in the fifteenth century to Piranesi or Stuart and
Revett in the eighteenth, Greek and Roman structures and works of art were
recorded by artists, architects, and antiquarians, whether for their picturesque
qualities or for imitation in new buildings (Greene 2002: 8-n). Although Roman
66 SOURCES
authors such as Pliny the Elder and Vitruvius provided extensive information
relevant to antiquarians, technical studies from Renaissance treatises to the ency-
clopaedias of eighteenth-century France maintained a contemporary focus even
when they made use of ancient literary sources.
In 1715 a London publisher thought it worthwhile to issue The History of Many
Memorable Things Lost, Which Were in Use among the Ancients, consisting of a series
of lists of "both natural and artificial" items, originally compiled in Italian for the
Duke of Savoy by Guido Pancirolli (1523-99), and translated into Latin around
1600. The lost items include manners, clothing, and forms of buildings, but few that
would now be considered technological, and the compiler points out that ancient
military hardware has been superseded by guns. "Lost" also meant "inaccessible,"
such as sources of marble under the control of the Turks. The author knew about
very large quadriremes and quinqueremes unparalleled in his day, as well as the ox-
powered liburnian warship described in the anonymous De rebus bellicis (17.1-3),
and he retold Roman stories about "unbreakable" glass without skeptical comment
(cf. Pliny, HN 36.195; Petronius, Sat. 50-51; Dio Cassius 57.21.6). Although he knew
about the odometer from Vitruvius (De arch. 10.9.1-4), evidence for Roman water-
mills came only indirectly from Procopius and Pliny rather than Vitruvius' full
description (De arch. 10.5.1-2). The book's contemporary focus was emphasized by
a companion volume, Many Excellent Things Found, Now in Use among the Mod-
erns. The novelties "now in use" begin with the New World and porcelain, but
bezoar stones, rhubarb, sugar, and manna come next, long before printing, spec-
tacles, or guns. The curious arrangement of Pancirolli's lists-and the inclusion
of clothes, medicaments, food, and leisure activities-underlines the lack of a
perception of technology as a distinct entity in preindustrial Europe.
Ancient technical information was taken sufficiently seriously to be included
alongside current mining and metallurgical processes by Agricola in his De re
metallica, published in Germany in 1556 (Hoover and Hoover 1912; Domergue
1989). Archimedes' water-screw appears to have been reintroduced into early
modern Italy as a result of ancient documents (Drake 1976). Pioneering agricultural
reformers in eighteenth-century Britain made use of the Roman agronomists, and
the vallus described by Pliny inspired a labor-saving harvesting machine in Australia
as late as the nineteenth century (Jones 1980). Roman technical achievements and
organizational abilities were highly regarded, alongside the intellectual and artistic
accomplishments of ancient Greece. Daniel Defoe observed the remains of a Roman
road at a time when Britain's road system was very poor:
It is true that the Romans being Lords of the World, had the Command of the
People, their Persons and their Work, their Cattle and their Carriages; even their
Armies were employ'd in these noble Undertakings .... But now the Case is al-
ter'd, Labour is dear, Wages high, no Man works for Bread and Water now; our
Labourers do not work in the Road, drink in the Brook; so that as rich as we are, it
would exhaust the whole Nation to build the Edifices, the Causways, the Aque-
ducts, Lines, Castles, Fortifications, and other publick Works, which the Romans
built with very little Expence. (Defoe 1724-1726: 520-21)
It was not until the eighteenth century that Europeans became certain that the
changes being brought about by the incipient Industrial Revolution were creating a
material existence that not only differed significantly from that of the ancient world
but exceeded its achievements. Growing awareness of modern progress threw up a
paradox: how could Greece and Rome have been so productive in literature, phi-
losophy, art, architecture, and infrastructural engineering but so deficient in eco-
nomic and technological awareness (Traina 1994: 10-13)? A comprehensive study of
inventions (ancient and modern) was compiled-before the full effects of the
Industrial Revolution became discernible-by Johann Beckmann, whose three-
volume Beytriigen zur Geschichte der Erfindungen was published between 1782 and
1804; further editions and translations soon followed. His command of classical
sources was thorough, and accounts of topics such as milling already include most
of the references to Antipater, Vitruvius, and others familiar to us today. Other
scholars based in Gottingen followed Beckmann's lead in recognizing the impor-
tance of socioeconomic factors in inventions and their diffusion, and in assigning a
significant role to technology in history (Schneider 1992: 17-18).
Nineteenth-century German philologists took an interest in everyday life and
laid the foundations for the history of technology. The most significant publication
was Bliimner's Technologie und Terminologie der Gewerbe und Kunst bei Griechen
und Romern (1875-1887 and subsequent editions), an exemplary combination of
documentary, iconographic, and archaeological evidence. This work coincided with
the publication of comprehensive classical dictionaries in France and Germany-
Daremberg and Saglio's Dictionnaire des antiquites grecques et romaines d'apres les
textes et les monuments (1877-1919) and the first volumes of Wissowa's revision of
Pauly's Real-Encyclopiidie der klassischen Altertumswissenschaft (1893-1978)-that
made occasional use of material evidence. Bliimner was particularly interested in
materials and production processes, from baking and textiles to stoneworking and
metallurgy (Schneider 1992: 19), and the comparative neglect of transport and
energy was compensated by the publication in 1899 of Die Ingenieurtechnik im
Alterthum by his contemporary Merckel, who paid most attention to construction
associated with water, roads, and bridges. Although Merckel considered the Greek
and Roman periods to be a high point in engineering history, he included earlier
civilizations and undertook a concise overview of developments up to his own day
(1899: 12-23). In the twentieth century, further German writers (notably Neuberger
and Feldhaus) extended their coverage to other ancient civilizations and to eth-
nographic studies of prehistoric technology, while another specific study of the
classical world was provided by Diels' Antike Technik (1914)-highly praised by
Schneider (1992: 19). Neuberger's wide-ranging Die Technik des Altertums (1919)
appeared as The Technological Arts and Sciences of the Ancients in 1930. Among other
early twentieth-century general writers, the American A. P. Usher was unusual in
including a discussion of classical technology in his History ofMechanical Inventions
(1929), preceded by a profound psychological consideration of the process of in-
vention (expanded in 1954). The second of the five volumes of A History of Tech-
nology (C. Singer et al. 1954-1958) was devoted to the Mediterranean civilizations
68 SOURCES
and the Middle Ages, but by this date expectations had risen: Finley condemned
some of its contributors "who, on both external and internal evidence, have never
studied, in a systematic way, the ancient world ... and so we have a work by
museum-keepers and technicians" (1959: 123-24).
Many twentieth-century works concentrated on specific areas of technology or
spheres of production, from Drachmann's Ancient Oil-Mills and Presses (1932) to
Davies's Roman Mines in Europe (1935), and combined close scrutiny of ancient
sources with growing quantities of archaeological evidence. Specialization contin-
ued throughout the twentieth century, and included the remarkable series Studies
in Ancient Technology produced by Robert Forbes from 1955 to 1964; he consid-
ered standard topics such as agriculture and hydraulic technology, alongside
themes such heating and lighting, as well as energy and various raw materials. While
Schneider was right to describe these volumes as "unzusammenhangende Detai-
luntersuchungen" (1992: 21), Forbes also wrote general works that drew conclu-
sions about the nature of classical technology (for example, Man the Maker, 1950).
Drachmann's The Mechanical Technology of Greek and Roman Antiquity (1963)
improved access to key literary sources, building on his Ktesibios, Philon and Heron
of 1948. Popular general works included Klemm's Technik: Eine Geschichte ihrer
Probleme (1954, available in English from 1959) and de Camp's The Ancient En-
gineers in 1960. Both books made extensive use of ancient sources, and discussed
Greece and Rome within a wider chronological survey of engineering. By the 1970s,
monographs were becoming available on an extraordinarily wide range of specialist
topics: examples from 1974 alone include Blast-Power and Ballistics (Lindsay), Gears
from the Greeks (Price), The Warpweighted Loom (Hoffman), Papyrus in Classical
Antiquity (N. Lewis) and Sur la taille de la pierre antique, medievale et moderne
(Varene). This trend continued into the 1980s and included the works discussed at
the beginning of this chapter. In 1986 Oleson published an extensive, annotated
bibliography of Bronze Age, Greek, and Roman technology.
THEMES IN THE EMERGENCE OF A

"HISTORY OF TECHNOLOGY"
The Ancient Economy

Because economics is a relatively recent discipline, economic histories of periods
earlier than the Industrial Revolution were infrequent before the twentieth century,
and technology was rarely incorporated into them. Economists, however, have
persistently associated innovations (whether technological or organizational) with
raising productivity, usually by reducing the costs of human labor, motive power,
or raw materials. Marx did much of his writing in London, capital of the world's
first industrial nation, while Engels experienced the effects of the Industrial Re-
volution at first hand while working in Manchester in a textile company-one of
the industries most dramatically affected by labor-saving technology between 1750
and 1850. In the early twentieth century, Schumpeter conceptualized economics in
terms of business cycles characterized by periods during which "clusters of inno-
vations" prompted recovery after depressions (1939; Ruttan 2001: 64-65). His
contemporary Kondratieff defined economic waves of prosperity, recession, de-
pression, and recovery (1935). After 1950, studies of innovation were situated in
equilibrium or (to a much smaller extent) evolutionary economics, and the subject
was given a boost after 1970 by the desire to stimulate recovery from a worldwide
recession.
"Economic history" should be distinguished from sociological studies of "the
ancient economy," involving polarized primitive and modern interpretations,
which preoccupied German scholarship in the nineteenth and early twentieth
centuries (Finley 1979; Derks 2002). A lack of technological inventiveness could be
explained quite easily by categorizing Greek and Roman society in terms of
primitive households whose economies lacked any capacity for growth. Campaigns
against slavery in the nineteenth century highlighted the fact that Greek "democ-
racy" and its esteemed arts and crafts were products of a slave-owning society.
Thus, writers who did not share the primitive view of Greek and Roman economic
activity tended to attribute a lack of technical development to slavery (a view
repeated frequently in the twentieth century-Schneider 1992: 22-29).
Growing demand for economic history in the 1920s was illustrated by the
launch of a History of Civilisation-comprising more than 50 monographs-which
included translations of three French works: Ancient Greece at Work (Glotz 1926),
Ancient Rome at Work (Louis 1927), and The Economic Life of the Ancient World
(Toutain 1930). All three paid little attention to technology as a component of
economic history and sought explanations for failure rather than achievements:
Ancient industry was never governed by machines, in Alexandria, or earlier in

Athens, or later in Rome. There was no inducement to adopt them, for human
labour was not scarce or expensive .... The absence of machinery prevents the
workshop from becoming a big factory, labour from concentrating in great
numbers, and wholesale production from killing the work of the family. For
machinery is irreconcilable with slavery.... The institution of slavery, inherent in
the very conception of the city-that is what creates an essential difference be-
tween ancient economy and modern. (Glotz 1926: 352, 381)
A recurrent stereotype of Romans as mere imitators accompanies the diagnosis of

slavery as a limit to progress:
The warlike expeditions which led [the Roman] into contact one by one with the
most diverse peoples and hewed him paths into the whole world, as it was then
known, brought with them technical knowledge of all kinds besides territorial
conquests, booty and slaves .... To be sure, the small field of knowledge and the
70 SOURCES
primitive simplicity of manufacturing processes hardly tended to engender that

distribution of labour which is the basis of modern industry, but such a divi-
sion was in any case incompatible with the dominant institution [slavery] ....
Thus, long before America in the 19th century reached the same conclusion, the
Roman world discovered that servile labour in spite of its apparent economic
advantages was both onerous and unproductive. (Louis 1927: 5, n)
The ideas of Louis and Glotz are constrained by a view of industrial development-
shared by the contemporary economist Schumpeter-that identified innovation as
the key to economic progress.
Finley's The Ancient Economy (1973) attributed technological stagnation not to
slavery but to a mindset that could not conceptualize economic development in
modern terms. This psychological disability included disdain for anything associ-
ated with work, including practical applications of technology. Finley's perception
was supported not only by ancient texts (cf. Humphrey et al. 1998: 579-97) but also
by substantivist economic anthropologists such as Karl Polanyi, a former colleague
in the 1930s. His "primitive" stance contrasts with the "modern"/formalist Rosto-
vtzeff, whose Social and Economic History of the Roman Empire (1926) made ex-
tensive use of archaeological evidence and detected signs of capitalism in the ancient
world. The polarized positions of Finley and Rostovtzeff remain highly influential
(Storey 2004: 106-9). In 1986 I published The Archaeology of the Roman Economy to
assess the contribution of archaeological evidence-which had grown rapidly in
recent decades-to debates between ancient historians:
Quite simply, I believe that the level of economic activity revealed by archaeo-
logical research makes the "minimalist" approach of historians such as Finley
untenable. The economy does not show signs of advance or evolution, simply an
intensification of everything that had already existed in Greek and Roman re-
publican times .... What is absolutely clear is that no economic history of the
Roman empire can ever be written again which does not give the same detailed
attention to the results and technical problems involved in archaeology as it does
to the textual criticism of Roman documentary sources. (Greene 1986: 170-71)
Since I had not yet scrutinized technological evidence, convergence with the in-
dependent publications of K. D. White and others discussed at the beginning of
this chapter was encouraging.
Revolutions and Determinism

As Finley and Schneider's critical comments about the Oxford History of Tech-
nology and Forbes' Studies in Ancient Technology demonstrate, twentieth-century
publications were increasingly expected to engage with the historical and cultural
context of technology. The 1930s saw the appearance of influential works with a
monocausal determinist outlook, notably two by Lefebvre des Noettes: L'attelage,
a
le cheval de selle travers les ages (1931) and De la marine antique la marine a
moderne: La revolution du gouvernail (1935). Ancient harnessing and steering oars

were implicitly condemned as inefficient by associating the medieval horse collar
and the sternpost rudder with the word "revolution"-an opinion echoed in 1935
by Bloch's claims for the medieval "advent and triumph" of the water-mill,
published in the new socioeconomic periodical Annales. It was sustained over the
next fifty years by Lynn White, Jr. (e.g., Medieval Technology and Social Change,
1962) and Jean Gimpel (La revolution industrielle du Mayen Age, 1975). Powerful
narratives of this kind require the accumulation of overwhelming contrary evi-
dence before a paradigm shift takes place. Wikander and Wilson's documentary
and archaeological investigations of water-mills have not yet changed accounts of
medieval technology (e.g., Gies and Gies 1994; Landers 2003).
It is hardly surprising that technological determinism became a feature of the
period between the two world wars of the twentieth century. Nineteenth-century
Scandinavian archaeologists had organized the past on the basis of raw materials
into Stone, Bronze, and Iron Ages. Anthropologists had reinforced theories of
social evolution with ethnographic descriptions of social structure and material
culture. Engels and Marx combined production and social structure to give tech-
nical change a dynamic revolutionary role. World War I provided a frightening
mirror image of the Industrial Revolution and mass production as enormous
numbers of soldiers faced new technical methods for killing them. After the Russian
Revolution, competition with the capitalist powers increased Soviet demands for
the transformation of industrial and agricultural production by technology. The
Marxist archaeologist V. Gordon Childe, who visited the Soviet Union in 1935, had a
particularly enduring impact on the interpretation of the past (Greene 1999;
Gathercole 1994). Childe's Neolithic and Urban revolutions defined in Man Makes
Himself (1936) and What Happened In History (1942) gave Marxist social evolution a
longer pedigree by identifying prehistoric revolutions-unknown to Marx-on the
basis of archaeological evidence.
Unsurprisingly, the technology of Greece and Rome appeared insignificant to
general commentators who considered that the most important revolutions in
technology and economics had taken place in the prehistoric Near East and/or
eighteenth-century England. Destree's Histoire des techniques (1980: 41) identified
"deux revolutions essentielles: l'une a partir de 7000 ACN, la revolution agricole;
l'autre entre 1750 et 1850, la revolution industrielle," a view shared by two Italian
surveys (Forti 1963, Mondini 1973). Destree and other writers emphasize energy and
materials; like the Scandinavian Three Age System (anticipated by Greek philoso-
phers), such categorizations bypass the classical world. Lewis Mumford's Technics
and Civilization described pseudo-archaeological "eotechnic," "paleotechnic," and
"neotechnic" phases of technology based on water and wood, coal and iron, and
electricity and alloys, respectively (1934: 109-51). For Morgan, the production ofiron
was "the event of events in human experience, without a parallel, and without an
equal, beside which all other inventions and discoveries were inconsiderable" (1877:
43), while Buchanan saw direct production of cast iron around A.D. 1500 as the
critical division between the Old Iron Age and the New Iron Age (1992: 16). Merckel,
72 SOURCES
who knew more about Greek and Roman technology than any of these writers,
identified an even more precise turning point-6 October 1829-when the success
of Stephenson's Rocket at the Rainhill trials ushered in the railway age (1899: 4).
Relevance and Continuity

By the late 1940s, it seemed timely to reconnect technology with humanity, and at the
same time to emphasize its continuity and benign aspects, since World War II had
risen out of the unstable politics of radical regimes of the Right and Left. The war had
been brought to an end by atomic weapons of unparalleled power, developed by a
well-organized program of scientific and technical research. By the 1950s, a new
conflict was growing between capitalism and communism, which would be fought
as much through technology and economics as through military might. Charles
Singer, a prominent British historian of science and medicine, was commissioned to
prepare A History of Technology for Oxford University Press (in five volumes, with
the help of an endowment from Imperial Chemical Industries) in the late 1940s:
The main object of this work is to provide students of technology and applied
science with some humane and historical background for their studies. They may
thus be helped to realize that the subjects of their special training are parts of a
very ancient process and are rooted in many civilizations.... [The editors] are
convinced of the human value, in our technological civilization, of an under-
standing of the methods and skills by which man has attained a gradual easing of
his earthly lot through mastery of his natural environment. (C. Singer et al. 1954-
1958:1: v)
Singer's emphasis on "a gradual easing" of human conditions by means of tech-

nology differs from the inevitable revolutions associated with Marxism. Never-
theless, Childe still employed revolutionary vocabulary in his contributions to
Singer's first volume-at a time when an avowed Marxist would have lost his job in
the United States.
In response to reviews, Derry and Williams produced a single-volume sequel to
Singer's five volumes in 1960. It aimed to be "as much a technological history as a
history of technology," following an "important modern trend," because techno-
logical factors were still "far too little recognized" {v-vi). In common with most
general overviews it had a western (primarily European) focus and a tapered chro-
nological range with a large modern section (434 pages on the period from 1750 to
1900) and a small prehistoric "tail" (27 pages); early civilizations occupied 40 pages,
Greece and Rome 61, and the Middle Ages 71. Other books followed Beckmann
by adopting a thematic approach; the Routledge Encyclopaedia of the History
of Technology (McNeil 1990) was divided into sections, each beginning with a brief
summary of its prehistoric or early historical origins. McNeil (1990: 1) echoed
the feeling of neglect voiced 30 years earlier by Derry and Williams: "It is strange
that, in the study and teaching of history, so little attention is paid to the history of
technology.... Technology is all around us: we live in a world in which everything

that exists can be classified as either a work of nature or a work of man. There is
nothing else." Technology in Western Civilisation (Kranzberg and Pursell 1967),
which included such distinguished contributors as Forbes, Usher, and Lynn White,
Jr., was designed for teaching in the U.S. Armed Forces Institute because "tech-
nology has been one of the major determinants in the development of Western
civilization; yet only recently has there been a recognition of this fact .... No single
available text stressed sufficiently the cultural, economic, and social implications of
technology and history" (v). In France, Daumas' five-volume Histoire generale des
techniques was more metaphysical about continuity: "Si les civilisations sont
mortelles, chacune d' elles, avant de succomber ason destin, a prepare un heritage
qui n'a jamais ete ignore de celles qui lui ont succede" (1962: v, repeated on the
cover of a low-priced paperback edition in 1996).
Ethnography and Material Culture

Advances in Greek and Roman archaeology m the nineteenth century were
founded on sociological or ethnographic descriptions of everyday life based on
ancient written sources but enlarged by material evidence (e.g., Bliimner 1911;
Marquardt 1879). Ethnography, or more specifically ethnoarchaeology, is now a
well-established approach to ancient technologies (e.g., Peacock 1982; cf. chapter 33
here). Scientific analyses of materials and experimental simulations of processes are
integrated into mainstream archaeological theory and practice, although confi-
dence in the ability of ethnoarchaeology and experiment to generate direct anal-
ogies and explanations has waned since processual approaches have been modified
or replaced by symbolic/contextual interpretations (Shennan 1996). The dangers of
collecting data in an ethnographic manner and embedding it in explanations de-
rived from ancient documentary sources were highlighted by Bradley (2002) in a
study of evidence for Roman fulling (although his use of empirical data was
criticized by Wilson 2003). Bradley challenged the "fantasy reconstruction" of a
single standardized narrative from scattered and fragmentary material evidence:
A classical dictionary's "Fulling" presents a bold advertisement of Roman tech-

nical ingenuity and aggressive cleanliness. There is no single account in Roman
literature of the process of fulling. What we have is a handful of isolated texts and
images collected and fused together, as if each bit of evidence were a piece of a
jigsaw, just one fraction of a long and rigorous washing-machine programme ....
What, archaeologically speaking, counts as a fullonica is usually represented by the
presence of features which might conceivably have reproduced some or all of
the functions described in the dictionary synopsis. (Bradley 2002: 24-25)
Ethnography has enhanced the significance of the material world and artifacts
in both anthropology and archaeology: How Things Tell the Stories of People's Lives
was the subtitle of Janet Hoskins' influential book Biographical Objects, which
74 SOURCES
concluded that "an object can thus become more than simply a 'metaphor for the
self.' It becomes a pivot for reflexivity and introspection, a tool of autobiographical
self-discovery, a way of knowing oneself through things" (1998: 198). Anthro-
pologists adopt a far broader definition of technology than most historians,
probably because of the influence of philosophers such as Heidegger, whose defi-
nition has been summarized as "the various ways in which I-as-body interact with
my environment by means of technologies" (Ihde 1990: 72). Broad definitions are
frequently indicated by the plural form "technologies," exemplified in an archae-
ological context by Crummy and Eckardt's description of sets of Roman cosmetic
tools (for nail-deaning, etc.) as "technologies of the self" (2003).
For several centuries archaeologists have approached material culture through
the classification of collections. Classification was accompanied by typology from
the early nineteenth century, and typological sequences or stylistic attributions were
reinforced by whatever chronological evidence could be obtained from documents
or from archaeological excavations (Greene 2002: 140-51). Devices-such as the
wood-block force-pump-that were not described in ancient sources still neces-
sitate this kind of approach (Oleson 2004; Stein 2004). Since most of the evidence,
however, comes from accidental survivals of pumps in the bottom of wells, em-
phasis is placed primarily on their construction and performance rather than their
invention and innovation.
Evolution and Progress

Alongside ethnographic and contextual studies of technology, reflection about the
nature of change-whether or not it is considered to have been "progress"-has
stimulated the exploration of evolutionary ideas. Nineteenth-century ideas about
"survival of the fittest" have been refined in the twentieth century into investiga-
tions of the selective mechanisms that determine whether businesses and econo-
mies succeed or fail, for example in the evolutionary or institutional approaches of
modem economists Nelson and Winter (1982; on technology, Nelson 1987) and
North (1990 ). Much thought has also been given to establishing the limits of using
biological evolutionary theory to explain technological innovation and change,
from pioneers such as George Basalla and Joel Mokyr in the late 1980s and early
1990s to a volume edited by Ziman in 2000. Evolutionary studies are at particular
risk from "futurism," whereby the choice of technologies and institutions for study
is determined by their contribution to the present, rather than their relevance to the
past (Edgerton 1999: 123).
A capacity for adaptation is highly regarded by those who have a teleological
perception of human evolution. Some general overviews present technological change
in linear, social-evolutionary terms and regard developments before around A.D. 1500
as necessary steps toward the modem world, rather than significant components of
their own time. Indeed, Buchanan's The Power of the Machine (1992) and Williams'
The Triumph of Invention: A History of Man's Technological Genius (1987) embody
this progressive narrative in their titles. Studies of twenty-first-century global con-

sumerism place considerable emphasis on the origins of the modem world economy,
which are generally seen to lie in the early modem period between the Voyages of
Discovery and the Industrial Revolution (Berg 2004). Inventions and innovations are
plentiful in this period, from infrastructural concerns (such as ship design or banking
systems) to patterns of consumption (such as tea drinking and polite table behavior
involving fine ceramics). Technological innovations are even more visible with the
growth of modem industrial production, when the names of engineers and inventors
become well known-Arkwright, Stephenson, Parsons and others.
Rostovtzeff's Social and Economic History of the Roman Empire (1926), written
by a Russian emigre who had witnessed the 1917 revolution, made many references
to social revolution, but had little to say about technology. Negative assessments of
Greek and Roman stagnation, rather than progress, in technology were consolidated
by Finley in the widely-read periodical Economic History Review in 1959 and 1965.
The frequency with which his views have been repeated by general historians of
technology since the 1960s led me to publish in the same periodical a review of
Finley's position in the light of new archaeological evidence (Greene 2000).
Alternative Technology
In 1992, I published a short paper (Greene 1992) about technology transfer in the
Roman Empire that explored the vocabulary of modem development economics
and drew attention to useful concepts such as the "technology shelf" (the range of
known solutions to technical problems that might be drawn on in differing cir-
cumstances: H. Singer 1977). This concept was helpful in explaining why technology
did not advance on a single front, and why less "efficient" devices such as hand-
mills could coexist with water-powered mechanisms of the kind described by Vi-
truvius in the first century B.C. A key element in this concept, however, was
knowledge: the metaphorical technology shelf only existed in the minds of people
aware of the range of products laid out on it. There is a growing consensus that
Greek and Roman technical treatises, and perhaps a range of simpler instruction
and maintenance manuals, played a major role in promoting awareness of technical
solutions to practical problems (e.g., Oleson 2004). In the context of Hellenistic
kingdoms and especially in the expanding Roman Empire, the combination of lit-
eracy and reference works produced a result observable from archaeological re-
mains: similar devices for water-lifting really were deployed from Egypt to Britain.
Development economics in the second half of the twentieth century was in-
creasingly concerned with mechanisms of technology transfer, and failures led to
increased awareness of the importance of contextual factors (as well as tacit
knowledge) in people's willingness to entertain change. This awareness paralleled
the growth of institutional economics, which related economic progress (measured
in essentially free-market capitalist terms) to the existence of appropriate legal and
administrative frameworks for businesses to risk investment (North 1990). An
76 SOURCES
opposite effect was the emergence of the appropriate technology movement, in-
spired by Schumacher's Small Is Beautiful (1973), which anticipated the devastating
effects of globalization on local social systems, and favored the introduction of
technology that would allow a slower form of endogenous growth that benefited
internal participants rather than external investors.
In the early twenty-first century, concern has deepened about the conflict
between people's addiction to the material benefits of First World industrialization
and its unsustainability in the face of global warming and competition for resources.
The difficulties involved in bringing about change underline the role of psycho-
logical factors, and illustrate how the conceptualization of knowledge is based on
belief (Winder 2005: 154). Widely shared beliefs become institutionalized and dif-
ficult to change without a phase of major disruption that clears the ground for the
acceptance of new beliefs. Winder has described these events as "epiphanies" and
the process as the "Phoenix Cycle" (2005: 10-14). Political fear that a Dark Age is
a necessary precursor to a Renaissance has led the European Union to invest in
research into the nature of innovation in the hope of finding a less disruptive
manner of transition (TiGrESS: Time-Geographical approaches to Emergence and
Sustainable Societies: Winder 2005).
THEORY: THE Loss OF INNOCENCE
By the late twentieth century, the history of technology had become a subject of
central importance that could no longer be neglected. Meanwhile, specialized
studies of both modern technology and prehistoric material culture had witnessed a
rapid growth in sociological and anthropological approaches, frequently combined
with postmodern deconstruction (Greene 2004: 157-60). Publications of such work
were unlikely to be read by classical archaeologists or ancient historians, and met
with resistance from some technological historians (Buchanan 1991). In advocating
the practical study of industrial archaeology, Gordon and Malone claimed that the
material record is independent and lacks inherent bias (1997: 13-14). Rejection of
this stance is central to postprocessual archaeology (e.g., Hodder 1999), for reasons
explored in David Clarke's elegant essay "Archaeology: The Loss of Innocence":
"Theory exists, in however unsatisfactory a form, in everything that an archae-
ologist does regardless of region, material, period and culture.... It is this perva-
sive, central and international aspect of archaeological theory, multiplied by its
current weakness, which makes the whole issue of major importance in the further
development of the discipline" (Clarke 1973: 17-18).
The very act of compiling a handbook of Greek and Roman technology involves
theoretical choices both in its subject matter and its period of study. Division of the
past relies on a consensus that a specific period has a coherence that allows it to be
considered as an entity-Finley has been criticized for treating the Greek and
Roman period as an amorphous whole that allowed him to make generalizations
based on evidence and examples plucked from different parts of it. It is difficult to
avoid social-evolutionary judgments: Are Greece and Rome direct antecedents of
Western civilization, and therefore of special significance in understanding the
globalization of European culture after A.D. 1500? And why focus on engineering
and technology? Has the rhetoric of Singer or Williams persuaded us that because
science and technology are the most significant components of modern industrial
civilization, they will provide the key to understanding the classical world? Since
explicit terminology is a basic requirement in theoretical discussion, key terms
such as discovery, invention, and innovation will be discussed before further ap-
proaches to technology are examined.
Discovery and Invention

Discovery suggests the revealing of something that already existed but had not been
recognized or conceptualized. In the simplest terms, a technical discovery is most
likely to take place in the context of science or as a by-product of a practical process
such as smelting metal ores. Invention implies originality, and a conscious act of
implementing an idea in a new device or process that may well rely on a prior
discovery. Invention is an action (although it may require sustained effort) while
innovation is the process whereby an invention is brought into use: "Nylon ... was
first invented in 1928, but not innovated until 1939" (Shorter Oxford English Dic-
tionary, 5th ed.). The meanings of "invention" and "inventor"-from the Latin
invenire ( to come upon, to discover )-have changed little over recent centuries and
have never been restricted to technical rather than general intellectual endeavors.
The Greek verb heurisko is not as specific as the English term "invent," and can
mean "found" or "observed." "Innovation"-from innovare (to renew, to alter)-
has long-established political and economic associations, but its force has been
diminished by hyperbolic use of the adjective "innovative" for "new" since the later
twentieth century.
Usher included Greek and Roman technology in his pioneering History of
Mechanical Inventions (1929). Vernon Ruttan (2001: 66) follows Usher (1955) in
dismissing both the transcendentalist concept (flashes of brilliance from remarkable
individuals) and the mechanistic view (inventions arising in response to perceived
needs), and finds Usher's cumulative synthesis approach most attractive. The
definition of four steps involved in invention-perception of the problem; setting
the stage; the act of insight; and critical revision-provides "a unified theory of the
social processes by which 'new things' come into existence, a theory broad enough
to encompass the whole range of activities characterized by the terms science,
invention, and innovation" (Ruttan 2001: 67-68).
78 SOURCES
Technology
"Invention" and "technology" tend to be considered as givens in modern studies;
attention is focused on the circumstances that will prompt them and favor their
implementation. The focal points are commercial institutions, which are fully
conscious of innovation as a key to competitiveness, operating in market econo-
mies. Institutions may include research and development (R&D) facilities designed
to generate inventions and innovations that might reduce production costs or
introduce profitable new products; such activities are encouraged by systems of
copyrights and patent laws unknown in ancient Greece or Rome (May 2002).
Archaeological and anthropological approaches to technology understandably fo-
cus on material objects (Greene 2004; Sigaut 1994), but in the case of innovation
theory "the macro-oriented part ... will rarely relate to the specific form of tech-
nology" (Sundbo 1998: 23). Rogers employs modern computing terminology to
broaden his definition: "A technology is a design for instrumental action that
reduces the uncertainty in the cause-effect relationships involved in achieving a
desired outcome. Most technologies have two components: (1) hardware, con-
sisting of the tool that embodies the technology as a material or physical object, and
(2) software, consisting of the knowledge base for the tool" (Rogers 1995: 35).
Innovation
While studies of discovery and invention are concerned with the history of science
and/or technology, modern studies of innovation only begin once a new element
has been devised. It is important to remember that-in economic thinking-
innovation does not require inventions of a technical kind; any new ways of doing
things in organizations are significant if their implementation leads to economic
growth.
"Innovation" will here mean the work of developing an invented element for
practical and commercial use and of ensuring that the introduction of the element
is accepted. In other words, it does not involve the actual invention, or the sub-
sequent introduction of the element by other companies. (Sundbo 1998: 19--20)
The innovations can also have different characters: 1. Technological (objects). 2.
Intellectual, e.g. consultancy. 3. Physical movements (which are not technology),
e.g. a new transport (but without a change in technology). 4. Behavioural, e.g. a
new strategy for the company's market behaviour or a new organizational struc-
ture. (Sundbo 1998: 21)
Concepts of this kind were not articulated in the ancient world (Finley 1973).
Furthermore, most artifacts and monuments studied by archaeologists are the
results of a complete trajectory of invention, innovation, and diffusion. Classical
Greek intellectual innovations in philosophy and science certainly assisted the
development of Hellenistic astronomical and mechanical technology, and Roman
imperialism involved major changes in behavior, organizational structure, physical

movement, and transport. Modern studies of innovation reinforce the view that the
context of Greek and Roman technology was as important as its content (Greene
1994). As an economic activity, innovation is clearly a social phenomenon requiring
a "renewal of social behaviour" on the macro scale as well as individual activities at
the micro scale (Sundbo 1998: 1; Winder 2005).
Invention has greater significance if innovation is path-dependent, in that
the original form of an invention (technological or organizational) exerts a strong
controlling influence over all subsequent developments or patterns of innova-
tion. In practice, few ancient technologies were sufficiently complex to exclude
alternative inventions. An evolutionary approach to the success of innovations in-
corporates the Darwinian concept of selection, but questions arise over pre-
cisely what undergoes selective processes-machines? organizations? working
routines? (Ziman 2002). Nevertheless, analyses of invention and innovation that
are aware of path dependence and selection are more likely to acknowledge social
and behavioral factors and to avoid judgments based on cost factors alone (Ruttan
2001: 117).
Diffusion and Technology Transfer

Diffusion is the process by which an innovation is communicated through cer-
tain channels over time among the members of a social system. Diffusion is a
special type of communication concerned with the spread of messages that are
perceived as new ideas. Communication is a process in which participants create
and share information with one another in order to reach a mutual under-
standing. (Rogers 1995: 35)
Rogers emphasizes the two-way nature of communication and describes the cre-
ation and sharing of information as convergence (1995: xvi-xvii). The implications
lead directly into a social rather than technological perception of the diffusion of
innovations, which are "gradually worked out through a process of social con-
struction" (Rogers 1995: xvii). "Technology transfer" has become associated with
the process of diffusing material equipment and intellectual knowledge from de-
veloped economies to the Third World in the later twentieth century (e.g., H. Singer
1977; Greene 1992).
Ruttan has drawn attention to the fact that "the tacit character of much tech-
nical knowledge can represent a major obstacle to the rapid interregional or inter-
national transfer of technical capability" (2001: 168). The preclassical civilizations of
Mesopotamia and Egypt were characterized by a considerable amount of tacit
knowledge encoded in written form, accessible to a literate minority, which be-
came available to early Greece only through interactions with its eastern neighbors.
The Hellenistic eastern Mediterranean region (and successor Roman provinces and
the Byzantine Empire in that area) inherited much technical and scientific literature
80 SOURCES
written in Greek that was not translated into Latin, and reached western Europe only
during the Renaissance. Thus, innovation is inextricably connected to technology
transfer, given that an invention or discovery cannot be innovated without the
communication of knowledge between receptive practitioners (Winder 2005). That
communication must include essential tacit knowledge, of course, for example the
industrial processes involved in making molded terra sigillata understood by skilled
slaves who were established in branch workshops in Lyon by their Italian owners
(Desbat et al. 1997).
Technological Theory: Social Construction

By its nature contextual history is a vulnerable process in which the historian is
deeply affected by the humanity of the subject matter. To reject as ahistorical the
ideology of autonomous progress is to recognize that technological designs are
intimately woven into the human tapestry and that all of the actors in the drama,
including the storyteller, are affected by tensions between design and ambience.
(Staudenmaier 1985, 201)
Like Foucault's general history of discourses (1970, 1972), Bradley's specific study of
fulling underlined the way that interpretations have been constructed according to
a changing consensus about the constitution of knowledge. This concept is a
commonplace in anthropological approaches to modem technology (e.g., Latour's
Science in Action: How to Follow Scientists and Engineers through Society, 1987; Grint
and W oolgar' s The Machine at Work, 1997). These writers are particularly interested
in the sociocultural or infrastructural side of technology, notably where it impacts
on technology transfer or innovation (e.g., Tann 1995, or Hughes 1992). Such works
rarely look beyond modem industrial societies, although Schiirmann's specialized
study of Hellenistic mechanical devices (1991) did organize them according to their
social or ritual functions.
If much theoretical writing about the history of technology has little time depth,
how can it assist in the study of Greece and Rome? The approach most compatible
with the thinking of archaeologists and ancient historians is the Social Construction
of Technology (conveniently abbreviated to SCOT) or Social Shaping of Technology
(SST) (Fox 1996; Greene 2004). SCOT is associated with a collection of "classic"
papers edited by Bijker, Hughes, and Pinch in 1987, and its agenda is encapsulated in
the title of one chapter: "The Social Construction of Facts and Artifacts: Or How the
Sociology of Science and the Sociology of Technology Might Benefit Each Other"
(Pinch and Bijker 1987). It requires acceptance of the view that technology is not an
independent force that shapes society (i.e., technological determinism: Smith and
Marx 1994); the focus of interest is the social shaping of technology.
A shift from determinism to social production reflects wider intellectual
changes. Critical Theory, hermeneutics, and phenomenology undermined posi-
tivism throughout the social sciences and humanities in the second half of the
twentieth century. In archaeology, a division opened up between New Archaeology
(processualism) and a diffuse range of approaches described as interpretive or
postprocessual between 1970 and 1980 (Greene 2002: 243-66). The focus on hu-
mans shifted from groups, caught up in ecological systems that determined the
form of their social and economic organization, to "knowledgeable actors" en-
gaged in a dynamic interaction whereby they shaped the world in addition to being
shaped by it (Johnson 2004; Bintliff 2004). A collection of studies edited by
Lemonnier (1993) conveyed this active role in its title, Technological Choices:
Because technical actions under construction as well as changes in technology are

in part determined or encompassed by social representations or phenomena that
go far beyond mere action on matter, societies seize, adopt or develop certain
technological features (principles of action, artefacts, gestures) and dismiss others.
It is as though societies chose from a whole range of possible technological avenues
that their environment, their own traditions and contacts with foreigners lay open
to their means of action on the material world. (Lemonnier 1993: 6)
While the concept of individuals as actors might seem far removed from
Greco-Roman technology, M. J. T. Lewis's observations about the products of the
Museum at Alexandria indicate considerable complexities in the relationships
between royal patrons and scientists/engineers (2000). Engineering simultaneously
involved prestige, gained from working on military equipment and automata for
court displays, and the negative associations of utility and work. Although the
primary purpose of the devices associated with Hellenistic Alexandria was to en-
tertain diners and to impress worshippers in temples (Schiirmann 1991), Lewis
emphasized their importance in demonstrating important scientific principles.
Thus, Greek "fine engineering" was not only concerned with "respectable" pursuits
such as the design of instruments for astronomy; interactions took place between
practical work and theoretical science, and between peasant labor and royal leisure.
Oleson (2004, 2005) explored the relationship between Ctesibian force-pumps
made from metal tubes (which are well documented in ancient sources and known
from a small number of archaeological finds) and force-pumps created by drilling
chambers in a solid block of oak (known from a larger number of archaeological
finds but undocumented in antiquity). The latter display considerably greater
uniformity in design and dimensions than bronze pumps, leading Oleson to posit
the existence of "subliterary" technical manuals that have not survived alongside
the treatises of Heron and others. This genre presumably circulated among the
large numbers of literate military and civilian architects, engineers, and craft
workers who moved around the Roman Empire, leaving ample signs of literacy at
relatively humble social levels (Harris 1989, 1993). Thus, the variability and com-
plexity of the social embeddedness of science, technology, and craft activities in the
Greek and Roman periods make it difficult to study technology from any per-
spective other than "social construction."
82 SOURCES
"Technology-in-Use,,
The histories of innovation and of technology-in-use are remarkably different, in
terms of geography, chronology, and sociology. (Edgerton 1999: 115)
Although it is clearly essential to have achieved a proper understanding of the

significance of the terms invention and innovation, even more is to be gained from
exploring use. David Edgerton, whose work has an entirely twentieth-century focus,
first published "From Innovation to Use: Ten Eclectic Theses on the Historiography
of Technology" in French in Annales in 1998; it subsequently appeared in English
(1999) and several other languages. He believes that "most (Anglo-Saxon) histori-
ography of technique is concerned with innovation rather than technology, which,
because of a failure to differentiate the two, leads to very unfortunate results" (111).
Edgerton's Thesis I provides considerable encouragement to the archaeological
study of technology:
The study of the relations of technology and society must necessarily deal with
technology which is in widespread use. However, most writing on the history of
technology and on the relations of technology and society is concerned with
innovation, with the emergence of new technologies. It fails to distinguish this
from the study of technology in widespread use, which is necessarily old, and is
often seen as out-of-date, obsolete, and merely persisting. (Edgerton 1999: 112)
Archaeological discoveries of technology-in-use have given us indications of the

geographical diffusion of devices described in treatises originating in (or inspired
by) the Museum at Alexandria. Thus, we know that large compartmented wheels
were employed to drain deep mines in Spain and Wales, bucket-chains lifted water
at Cosa in Italy and London in Britain, and geared water-mills ground wheat at
Roman villas in Provence and near forts on Hadrian's Wall. The study of tech-
nology-in-use rather than innovation "also involves a massive shift in social class,
social status, gender and race of people involved with technology" (Edgerton 1999:
116). The apparent paradox that most "Greek" technology is found in the Roman
period is elucidated by Edgerton's observation that "rapid innovation need not
correspond to periods of rapid productivity growth; innovations will have the
greatest impact on productivity growth at the time of the fastest diffusion, which
typically takes place long after innovation" (1999: 115).
On a broader cultural scale, Thesis III-"The conflation of innovation and
technology is especially apparent in national histories" (Edgerton 1999: 117)-helps
to release historians from narratives contrasting Greek inventiveness with Roman
practicality. The contribution of Mesopotamian and Egyptian water technology to
the work of individuals such as Archimedes or Philon is undisputed, but Helle-
nistic Greece is never criticized for benefiting from technology transfer. The fact
that Hellenistic technology was innovated and diffused much later in a Roman
imperial context should be seen in a positive light. The use of automata invented in
Alexandria for courtly entertainments and as "temple toys" merely illustrates the
time lag between the development of new technology and its impact on economic
growth. The diffusion of Hellenistic inventions and innovations took place in the
context of Roman political institutions, and was facilitated by the long-distance
communication of know-how by the movement of people, literacy, and technical
manuals in the context of an expanding empire. The disjointed trajectory of
technology transfer, invention, innovation, and diffusion tells us nothing about the
relative quality of thought in Greece and Rome, especially when it is measured by
modern historians of technology who put priority on novelty and sophistication
over established functionality.
Archaeologists are constantly involved in the study of "processes of mainte-
nance, repair, remodeling, re-use, and re-cycling [that] have been fundamental to
material culture, but are obliterated by the emphasis on initial creation" (Edgerton
1999: 120). Technical changes brought about as a result of everyday use and expe-
rience rather than abstract thought may explain the development oflewis holes for
lifting heavy blocks of stone, force-pumps made from wooden blocks rather than
bronze tubes, parchment codices rather than papyrus rolls, or hollow glass vessels
blown from molten blobs rather than molded or formed around a core. Thus,
archaeological studies oflong-term change accord well with Edgerton's Thesis VIII:
"Invention and innovation rarely lead to use, but use often leads to invention and
innovation" (1999: 123). Studies of use rather than invention also avoid the pitfall of
"technological futurism" because they avoid being "systematically biased by the
future in that we study innovations which succeed later." Futurist studies place
unreasonable expectations on the ancient world by scrutinizing Greek and Roman
inventions for signs of progress or stagnation (Greene 1993, 2000).
As noted above, judgments about inventiveness also lead to ideas about tech-
nological determinism, which Edgerton's Thesis VI usefully dismisses as "primarily
a theory of society, not a theory of technology." He notes, however, that "refuting
innovation-determinism is much simpler. First, only a small minority of innova-
tions are widely used; second, the extent of use surely determines the extent of the
effect, not the act of innovation" (Edgerton 1999: 121). Andrew Wilson's clear
exposition of the importance of machinery in supporting the political structure of
the Roman Empire by guaranteeing supplies of bullion for coinage is good example
of a study of effect-and therefore use-rather than innovation (Wilson 2003: 29):
The corollary is that, by contrast, the economic performance of the first and
second century, and to a certain degree the high level of imperial or state
spending-on the army, construction works, the annona, etc.-was partly de-
pendent on the use of advanced technology and industrial-scale operation in the
mines. The importance of technology to the ancient economy, and to wider
historical processes, can be measured by what happened when the larger-scale
hydraulic mining operations were no longer active.
The lesson to be learned from these brief intimations of theoretical complexity is

that it is important for researchers, writers, and readers to have an awareness of their
own ways of understanding history-starting with an appreciation of the subtleties
of terms such as "innovation." Does natural selection govern cultural as well as
84 SOURCES
biological processes? Is progress inherent in the activities of humans? Do political

systems move from one equilibrium state to another, or does chaotic collapse
normally occur? Does economic change take place in a smooth linear manner,
proceed by steps, undergo revolutionary jumps, or pass through repeating cycles of
growth, flowering and decline? Abstract questions of this kind should be considered
as part of any investigation of evidence for Greek and Roman technology.
REFERENCES
Adam, J.-P. 1984. La construction romaine. Paris: Picard.

Adam, J.-P. 1994. Roman building: Materials and techniques. Trans. A. Mathews. London:
Routledge.
Basalla, G. 1988. The evolution of technology. Cambridge: Cambridge University Press.
Beckmann, J. 1782-1804. Beytriige zur Geschichte der Erfindungen. 3 vols. Leipzig: Kummer.
Berg, M. 2004. "In pursuit of luxury: Global history and British consumer goods in tlie
eighteentli century," Past and Present 182: 85-142.
Bijker, W., T. P. Hughes, and T. Pinch (eds.) 1987. The social construction of technologi-
cal systems: New directions in the sociology and history of technology. Cambridge, MA:
MIT Press.
Bintliff, J. 2004. "Time, structure, and agency: The Annales, emergent complexity, and
archaeology," in J. Bintliff (ed.), A companion to archaeology. Oxford: Blackwell,
174-94.
Blackman, D.R., and A. T. Hodge (eds.) 2001. Frontinus' legacy: Essays on Frontinus'
'De aquis urbis Romae.' Ann Arbor: University of Michigan Press.
a
Bloch, M. 1935. "Avenement et conquetes du moulin eau," Annales 7: 538-63.
Bliimner, H. 1875-1887. Technologie und Terminologie der Gewerbe und Kunst bei Griechen
und Romern. 4 vols. Leipzig: Teubner.
Bliimner, H. 1911. Die romischen Privataltertumer. Munich: Beck.
Bradley, M. 2002. "'It all comes out in the wash': Looking harder at the Roman fullonica,"
Journal of Roman Archaeology 15: 20-44.
Buchanan, R. A. 1991. "Theory and narrative in the history of technology," Technology
and Culture 32: 365-76.
Buchanan, R. A. 1992. The power of the machine: The impact of technology from 1700 to the
present. London: Viking.
Callebat, L., and P. Fleury 1995. Dictionaire des termes techniques du 'De architectura' de
Vitruve. Hildesheim: Olms-Weidmann.
Chevallier, R. 1993. Sciences et techniques a Rome. Paris: Presses Universitaires de France.
Childe, V. G. 1936. Man makes himself. London: Watts Library of Science and Culture.
Childe, V. G. 1942. What happened in history. Harmondswortli: Penguin.
Clarke, D. L. 1973. "Archaeology: The loss of innocence," Antiquity 47: 6-18.
Coulton, J. J. 1988. Ancient Greek architects at work. Oxford: Oxbow.
Crummy, N., and H. Eckardt 2003. "Regional identities and technologies of tlie self: Nail-
cleaners in Roman Britain," Archaeological Journal 160: 44-69.
Daumas, M. (ed.) 1962. Histoire generale des techniques. Vol. 1, Des origines au XVe siecle.
Paris: Presses Universitaires de France.
Davies, 0. 1935. Roman mines in Europe. Oxford: Clarendon Press.
de Camp, L. Sprague. 1960. The ancient engineers. New York: Doubleday.
Defoe, D. 1724-1726. A tour through the whole island of Great Britain. London. Edited by
G. D. H. Cole. London: Davies, 1927. 2 vols.
Delaine, J. 1997. The Baths of Caracalla: A study in the design, construction, and economics
of large-scale building projects in imperial Rome. Journal of Roman Archaeology Suppl.
25. Ann Arbor, MI: JRA.
Del Chicca, F. 2004. Frontino, De aquae ductu Urbis Romae: Introduzione, testo critico,
traduzione e commento. Rome: Herder.
Derks, H. 2002. "'The ancient economy': The problem and the fraud," The European
Legacy 7.5: 597-620.
Derry, T. K., and T. I. Williams 1960. A short history of technology from the earliest times to
AD 1900. Oxford: Oxford University Press.
Desbat, A., M. Genin, and J. Lasfargues (eds.) 1997. Les productions des ateliers de potiers
antiques de Lyon, 1ere partie: Les ateliers precoces. Gallia 53. Paris: CNRS.
Destree, A. (ed.) 1980. Histoire des techniques. Brussels: Medders.
Diels, H. 1914. Antike Technik. Leipzig/Berlin: Teubner.
Domergue, C. 1989. "Les techniques minieres antiques et la De re metallica d'Agricola," in
C. Domergue (ed.), Mineria y metalurgia en las antiguas civilizaciones mediterraneas
y europeas. Madrid: Ministerio de Cultura, 2:76-95.
Drachmann, A.G. 1932. Ancient oil-mills and presses. Copenhagen: E. Munksgaard.
Drachmann, A.G. 1948. Ktesibios, Philon and Heron: A study in ancient pneumatics. Co-
penhagen: E. Munksgaard.
Drachmann, A. G. 1963. The mechanical technology of Greek and Roman antiquity: A study of
the literary sources. Copenhagen: E. Munksgaard.
Drake, S. 1976. "An agricultural economist of the late Renaissance," in L. T. White,
B. S. Hall, and D. C. West (eds.), On pre-modern technology and science: A volume
of studies in honor of Lynn White Jr. Malibu, CA: Undena.
Dugan, S., and D. Dugan 2000. The day the world took off: The roots of the Industrial
Revolution. London: Channel 4 Books.
Edgerton, D. 1999. "From innovation to use: Ten eclectic theses on the historiography of
technology," History and Technology 16.2: m-36.
Finley, M. I. 1959. "Technology in the ancient world," Economic History Review 12: 120-25.
Finley, M. I. 1965. "Technical innovation and economic progress in the ancient world,"
Economic History Review 18: 29-45.
Finley, M. I. 1973. The ancient economy. London: Chatto and Windus.
Finley, M. I. 1979. The Bucher-Meyer controversy. New York: Arno Press.
Fleury, P. 1993. La mecanique de Vitruve. Caen: Centre d'Etudes et de Recherche sur
l'Antiquite, Universite de Caen.
Forbes, R. J. 1950. Man the maker: A history of technology and engineering. London: Con-
stable.
Forbes, R. J. 1955-1964. Studies in ancient technology. 9 vols. Leiden: Brill.
Forti, U. 1963. Tecnica e progresso umano. Vol. 1, La tecnica nelle preistoria e nell'antichita.
Milan: Fratelli Fabbri Editori.
Foucault, M. 1970. The order of things: An archaeology of the human sciences. London:
Tavistock Publications.
86 SOURCES
Foucault, M. 1972. The archaeology of knowledge. London: Tavistock Publications.

Fox, R. 1996. "Introduction: Methods and themes in the history of technology," in R. Fox
(ed.), Technological change: Methods and themes in the history of technology. Am-
sterdam: Harwood Academic, 1-15.
Gara, A. 2002. Tecnica e tecnologia nelle societa antiche. Milan: CUEM.
Gathercole, P. 1994. "Childe in history" (Sixth Gordon Childe Memorial Lecture), Bulletin
of the Institute of Archaeology, University of London 31: 25-52.
Gies, R., and J. Gies 1994. Cathedral, forge, and waterwheel: Technology and invention in the
Middle Ages. New York: Harper Collins.
Gimpel, J. 1975. La revolution industrielle du Moyen Age. Paris: Editions de Seuil.
Glotz, G. 1926. Ancient Greece at work: An economic history of Greece from the Homeric
period to the Roman conquest. London: Kegan Paul, Trench, Trubner and Co.
Gordon, R. B., and P. M. Malone 1997. The texture of industry: An archaeological view of the
industrialization of North America. New York: Oxford University Press.
Greene, K. 1986. The archaeology of the Roman economy. London: Batsford.
Greene, K. 1990. "Perspectives on Roman technology," Oxford Journal of Archaeology 9:
209-19.
Greene, K. 1992. "How was technology transferred in the Roman empire?" in M. Wood and
F. Queiroga (eds.), Current research on the Romanization of the western provinces.
British Archaeological Reports, Intl. Series S575. Oxford: BAR, 101-5.
Greene, K. 1993. "The study of Roman technology: Some theoretical constraints," in
E. Scott (ed.), Theoretical Roman archaeology: First conference proceedings. Aldershot:
Avebury, 39-47.
Greene, K. 1994. "Technology and innovation in context: The Roman background to
mediaeval and later developments," Journal of Roman Archaeology 7: 22-33.
Greene, K. 1999. "V. Gordon Childe and the vocabulary of revolutionary change," Anti-
quity 73: 97-109.
Greene, K. 2002. Archaeology: An introduction. 4th ed. London: Routledge.
Greene, K. 2004. "Archaeology and technology," in J. Bintliff(ed.), Blackwell companion to
archaeology. Oxford: Blackwell, 155-73.
Grint, K., and S. Woolgar 1997. The machine at work: Technology, work and organisation.
Cambridge: Polity Press.
Harris, W. V. 1989. Ancient literacy. Cambridge, MA: Harvard University Press.
Harris, W. V. (ed.) 1993. The inscribed economy: Production and distribution in the Roman
empire in the light of instrumentum domesticum. Journal of Roman Archaeology Suppl.
Hart-Davis, A. 2004. What the ancients did for us: A brief history of ancient inventions.
London: BBC Books.
Hellmann, M.-C. 2002. L'Architecture grecque. Vol. 1, Les principes de la construction. Paris:
Picard.
Hill, D. R. 1984. A history of engineering in classical and medieval times. London: Routledge.
Hill, D.R. 1993. Islamic science and engineering. Edinburgh: Edinburgh University Press.
Hodder, I. 1999. The archaeological process: An introduction. Oxford: Blackwell.
Hoffman, M. 1974. The warpweighted loom: Studies in the history and technology of an
ancient implement. Oslo: Robin and Russ Handweavers.
Hoover, H. C., and L. H. Hoover 1912. Agricola's De re metallica, translated from the first
Latin edition of 1556. London: The Mining Magazine.
Hoskins, J. 1998. Biographical objects: How things tell the story of people's lives. New York:
Routledge.
Hughes, T. P. 1992. "The dynamics of technological change: Salients, critical problems and
industrial revolutions," in G. Dosi, R. Giannetti, and P. Toninelli (eds.), Technology
and enterprise in a historical perspective. Oxford: Clarendon, 97-118.
Humphrey, J. W., J.P. Oleson, and A. N. Sherwood 1998. Greek and Roman technology:
Ihde, D. 1990. Technology and the life world: From garden to earth. Bloomington: Indiana
University Press.
Johnson, M. 2004. "Archaeology and social theory," in J. Bintliff (ed.), Blackwell com-
panion to archaeology. Oxford: Blackwell, 92-109.
Jones, L. J. 1980. "John Ridley and the south Australian 'stripper,'" History of Technology 5:
55-101.
Klemm, F. 1954. Technik: Eine Geschichte ihrer Probleme. Freiburg: K. Alber.
Klemm, F. 1959. A history of western technology. London: George Allen and Unwin.
Kondratieff, N. D. 1935. "The long waves in economic life," Revue of Economic Statistics
17.6: 105-15.
Kranzberg, M., and C. W. Pursell (eds.) 1967. Technology in western civilization. Vol. 1, The
emergence of modern industrial society, earliest times to AD 1900. New York: Oxford
University Press.
Lancaster, L. 2005. Concrete Vaulted Construction in Imperial Rome: Innovations in Context.
Landels, J. G. 2000. Engineering in the Ancient World. 2nd ed. Berkeley: University of
California Press.
Landers, J. 2003. The field and the forge: Population, production, and power in the pre-
industrial west. Oxford: Oxford University Press.
Latour, B. 1987. Science in action: How to follow scientists and engineers through society.
Cambridge, MA: Harvard University Press.
Lefebvre des Noettes, Ct. 1931. L'attelage, le cheval de sellea travers les ages. 2 vols. Paris: Picard.
Lefebvre des Noettes, Ct. 1935. De la marine antique a la marine moderne: La revolution du
gouvernail. Paris: Masson.
Lemonnier, P. (ed.) 1993. Technological choices: Transformations in material cultures since
the Neolithic. London: Routledge.
Lewis, M. J. T. 1997. Millstone and hammer: The origins of water power. Hull: Hull Uni-
versity Press.
Lewis, M. J. T. 2000. "The Hellenistic period,'' in Wikander 2000: 631-48.
University Press.
Lewis, N. 1974. Papyrus in classical antiquity. Oxford: Clarendon
Lindsay, J. 1974. Blast-power and ballistics: Concepts offorce and energy in the ancient world.
London: Frederick Muller.
Lo Sardo, E. (ed.) 2005. Eureka! Il genio degli antichi. Naples: Electa.
Louis, P. 1927. Ancient Rome at work: An economic history of Rome from the origins to the
Empire. London: Kegan Paul, Trench, Trubner and Co.
Marquardt, J. 1879. Das Privatleben der Romer. Leipzig: S. Hirzel.
May, C. 2002. "Antecedents to intellectual property: The European pre-history of the
'ownership' of knowledge," History of Technology 24: 1-20.
McNeil, I. (ed.) 1990. An encyclopaedia of the history of technology. London: Routledge.
Meissner, B. 1996. Die technologische Fachliteratur der Antike. Berlin: Akademie Verlag.
88 SOURCES
Merckel, C. 1899. Die Ingenieurtechnik im Alterthum. Berlin: Julius Springer.

Mokyr, J. 1990. The lever of riches: Technological creativity and economic progress. Oxford:
Oxford University Press.
Mondini, A. 1973. Storia della tecnica. Vol. 1, Dalla preistoria all'anno mille. Turin: Editrice
Torinese.
Morgan, L. H. 1877. Ancient society; or, Researches in the lines of human progress from
savagery through barbarism to civilization. New York: Henry Holt.
Mumford, L. 1934. Technics and civilization. London: Routledge and Kegan Paul.
Nelson, R.R. 1987. Understanding technical change as an evolutionary process. Amsterdam:
New Holland.
Nelson, R. R., and S. Winter 1982. An evolutionary theory of economic change. Cambridge,
MA: Harvard University Press.
Neuberger, A. 1919. Die Technik des Altertums. Leipzig: Voigtlander.
Neuberger, A. 1930. The technological arts and sciences of the ancients. Trans. H. L. Brose.
London: Metlmen.
Nortli, D. C. 1990. Institutions, institutional change and economic performance. Cambridge:
Oleson, J.P. 1984. Greek and Roman mechanical water-lifting devices: The history of a
technology. Toronto: University of Toronto Press.
Oleson, J.P. 1986. Bronze Age, Greek and Roman technology: A select, annotated bibliog-
raphy. New York: Garland.
Oleson, J.P. 2000. "Water lifting," in Wikander 2000, 217-302.
Oleson, J.P. 2004. "Well-pumps for dummies: Was tliere a Roman tradition of popular
ambito romano. Archaeologia dell'Italia Settentrionale 8. Como: Comune di Como,
65-86.
Oleson, J.P. 2005. "Design, materials, and tlie process of innovation in Roman force
pumps," in J. Pollini (ed.), Terra marique: Studies in art history and marine archaeology
in honor of Anna Marguerite Mccann. Oxford: Oxbow Press, 211-31.
Pancirolli, G., and H. Salmutli 1715. The history of many memorable things lost, which were in
use among the ancients, and an account of many excellent things found, now in use
among the moderns, both natural and artificial. 2 vols. London: John Morphew.
Peachin, M. 2004. Frontinus and the curae of the curator aquarum. Stuttgart: Steiner.
Peacock, D. P. S. 1982. Pottery in the Roman world: An ethnoarchaeological approach.
London: Longman.
Pinch, T. J., and W. E. Bijker 1987. "The social construction of facts and artifacts: Or how
tlie sociology of science and tlie sociology of technology might benefit each oilier," in
W. Bijker, T. P. Hughes, and T. Pinch (eds.), The social construction of technological
systems. Cambridge, MA: MIT Press, 17-31.
Price, D. J. de Solla. 1974. Gears from the Greeks: the Antikythera Mechanism-A calendar
computer from c Bo BC. Transactions of tlie American Philosophical Society, n.s. 64.7.
Philadelphia: American Philosophical Society.
Raepsaet, G. 2002. Attelages et techniques de transport dans le monde greco-romaine.
Brussels: Le Livre Timperman.
Raepsaet, G., and F. Lambeau (eds.) 2000. La moissonneuse gallo-romaine. Bruxelles:
Universite Libre de Bruxelles.
Raepsaet, G., and C. Rommelaere (eds.) 1995. Brancards et transport attele entre Seine et
Rhin de l'Antiquite au Mayen Age: Aspects archeologiques, economiques et techniques.
Treignes: Ecomusee de Treignes.
Rodgers, R. H. 2004. Frontinus De aquaeductu urbis Romae. Cambridge Classical Texts and
Commentaries 42. Cambridge: Cambridge University Press.
Rogers, E. M. 1995. Diffusion of innovations. 4th ed. New York: Free Press.
Rostovtzeff, M. 1926. Social and economic history of the Roman empire. Oxford: Oxford
University Press.
Rowland, I. D., and T. Noble Howe (eds.) 2001. Vitruvius: Ten Books on Architecture.
Ruttan, V. W. 2001. Technology, growth, and development: an induced innovation perspec-
tive. New York: Oxford University Press.
Schneider, H. 1992. Einftlhrung in die antike Technikgeschichte. Darmstadt: Wissenschaf-
tliche Buchgesellschaft.
Schumacher, E. F. 1973. Small is beautiful: A study of economics as if people mattered.
London: Blond and Briggs.
Schumpeter, J. A. 1939. Business cycles: A theoretical, historical and statistical analysis of the
capitalist process. 2 vols. New York: McGraw-Hill.
Schiirmann, A. 1991. Griechische Mechanik und antike Gesellschaft: Studien zur staatlichen
Forderung einer technischen Wissenschaft. Boethius 27. Stuttgart: Franz Steiner.
Shennan, S. (ed.) 1996. Symbolic aspects of early technologies. Special issue, World Archae-
ology 27.1. London: Routledge.
Shirley, E. 2001. Building a Roman legionary fortress. Stroud: Tempus.
Sigaut, F. 1994. "Technology," in T. Ingold (ed.), Companion encyclopedia of anthropology.
London: Routledge, 420-59.
Singer, C., E. J. Holmyard, A. R. Hall, and T. I. Williams (eds.) 1954-1958. A history of
technology. 5 vols. Oxford: Clarendon.
Singer, H. 1977. Technologies for basic needs. Geneva: UN International Labour Office.
Smith, M. Roe, and L. Marx (eds.) 1994. Does technology drive history? The dilemma of
technological determinism. Cambridge, MA: MIT Press.
Staudenmaier, J.M. 1985. Technology's storytellers: Reweaving the human fabric. Cambridge,
MA: Society for the History of Technology and MIT Press.
Stein, R. 2004. "Roman wooden force pumps: A case-study in innovation," Journal of
Roman Archaeology 17: 221-50.
Storey, G. R. 2004. "Roman economies: A paradigm of their own," in G. M. Feinman and
L. M. Nicholas (eds.), Archaeological perspectives on political economies. Salt Lake City:
Utah University Press, 105-28.
Sundbo, J. 1998. The theory of innovation: Entrepreneurs, technology and strategy. Chel-
tenham: Edward Elgar.
Tann, J. 1995. "Space, time and innovation characteristics: The contribution of diffusion
process theory to the history of technology," History of Technology 17: 143-63.
Toutain, J. 1930. The economic life of the ancient world. London: Kegan Paul.
Traina, G. 1994. La tecnica in Grecia ea Roma. Roma: Laterza.
Trigger, B. G.2006. A history of archaeological thought. 2nd ed. Cambridge: Cambridge
University Press.
Usher, A. P. 1929. A history of mechanical inventions. Cambridge, MA: Harvard University
Press.
Usher, A. P. 1954. A history of mechanical inventions. 2nd ed. Cambridge, MA: Harvard
University Press.
Usher, A. P. 1955. "Technical change and capital formation," in Universities National
Bureau Committee for Economic Growth, Capital formation and economic growth.
Princeton: Princeton University Press, 423-550.
90 SOURCES
Varene, P. 1974. Sur la taille de la pierre antique, medievale et moderne. Dijon: Universite de
Dijon.
White, K. D. 1984. Greek and Roman technology. London: Thames and Hudson.
White, L. T. 1962. Medieval technology and social change. Oxford: Clarendon Press.
Wikander, b. 1984. Exploitation of water-power or technological stagnation? A reappraisal of
the productive forces in the Roman empire. Scripta Minora m. Lund: C. W. K. Gleerup.
Wikander, b. (ed.) 2000. Handbook of ancient water technology. Leiden: Brill.
Wilkinson, P. 2000. What the Romans did for us. London: Boxtree.
Williams, T. I. 1987. The triumph of invention: A history of man's technological genius.
London: Macdonald.
Wilson, A. I. 2002. "Machines, power and the ancient economy," Journal of Roman Studies
92: 1-32.
Wilson, A. I. 2003. "The archaeology of the Roman fullonica," Journal of Roman Archae-
ology 16: 442-46.
Winder, N. 2005. Breaking the Phoenix cycle: An integrative approach to innovation and
cultural ecodynamics. University ofNewcastle upon Tyne: School of Historical Studies.
Ziman, J. (ed.) 2000. Technological innovation as an evolutionary process. Cambridge:
Ziman, J. 2002. "Introduction: Selectionist reasoning as a tool of thought," in M. Wheeler,
J. Ziman, and M. Boden (eds.), The evolution of cultural entities. London: British
Academy, 1-8.
PART II
PRIMARY,
EXTRACTIVE
TECHNOLOGIES
CHAPTER 4
MINING AND
METALLURGY
PAUL T. CRADDOCK
TRADITION AND INNOVATION
The first millennium B.C. saw great developments in all aspects of metal produc-
tion, resulting from changing patterns of demand brought about by the spread of
iron usage and the introduction of coinage. The tremendous increase in the pro-
duction of metals across the Old World, particularly in the Roman, Mauryean-
Gupta, and Han Empires, is recorded in the heavy metal content of the cores taken
through the Greenland ice sheet (Hong et al. 1996). The established economic
foundations of the Bronze Age cultures of the Mediterranean, based on long-
distance trade in copper and even more on tin, were disrupted forever by the
emergence of iron. By the beginning of the period covered by this book, iron had
supplanted bronze, extending metal usage well beyond what had previously been
feasible. Iron above all else was available; most regions had some ore deposits.
The introduction of coinage in the mid-first millennium B.C. created an
enormous demand for precious metals: gold and, above all, silver. The great mines
of the Bronze Age had been worked for copper, but in the succeeding millennium
they were worked for silver. Rio Tinto, Laurion, and others to the east through to
the mines of northwest India flourished in the first millennium B.C. and were
primarily producing silver destined for coinage (Craddock et al. 1989). This new
demand for deep-mined ore encouraged developments in mining technology aided
by new developments in knowledge.
The first millennium saw the rise of the study of secular knowledge, or phi-
losophy, and its application to the world, with theories to cover the nature and
94 PRIMARY, EXTRACTIVE TECHNOLOGIES
origins of matter and the material world. This was an obvious area where obser-
vations made during mining could help the formulation of theory, and conversely
where a real understanding of geological matters could have helped with mine
strategy. Although theories on the origin of rocks, and in particular on the origins of
metals, were developed, there is very little mention of phenomena observed during
mining. This lack of oversight is reflected in the mines, where there is very little
evidence of an understanding of the nature and extent of the ore bodies (see below).
It is in the developing knowledge of mathematics and all aspects of mechanics
to move large quantities of air, water, and other materials in difficult conditions
that the application of the new knowledge is most evident. There was probably a
synergy between the problems posed by winning ore from ever deeper and more
difficult deposits, and the possible solutions offered by the new knowledge. The
familiar adage of the Industrial Revolution, "Science owes more to the steam engine
than the steam engine owes to science," could well be expressed in the classical
world as "Science owes more to mining than mining owes to science."
SOURCES OF METAL ORES
Northern and Western Europe

On the peninsula ofltaly, the main source of metals seems to have been in Tuscany,
with many mines in the Colline Metallifere (the ancient Massa Metallorum) inland
from Piombino (Badii 1931; Davies 1935: 63-75; Cambi 1959; Minto 1955). On Sar-
dinia there are many copper and silver/lead deposits where, although activity is now
mainly associated with the Bronze Age, there was certainly Punic and probably
Roman production (Tylecote et al. 1983).
Rather surprisingly, with the exception of gold, there does not seem to have
been much Roman activity in the mineral deposits of the Alps compared to that in
the prehistoric and medieval periods. The Romans exploited copper in southern
France (Davies 1935: 76-92). In the Central Massif region of central France there
were important gold-workings in the first millennium B.C. (Cauuet 1995, 1999), and
there were also tin workings (Penhallurick 1986: 85-94). Further north, it seems
likely that the lead/zinc deposits at Stolberg near Aachen in Germany were worked
from the beginning of the Roman occupation for zinc ore to make brass (see
below). Together with the deposits just over the border in Belgium at La Vieux
Montainge, these provided the zinc minerals for the major early medieval and
medieval dinanderie, or brass industry (Dejonghe et al. 1993). The distant province
of Britannia certainly produced some iron and copper, and large quantities oflead
(noted by Pliny, HN 34.164), although the silver content was probably disap-
pointing. The gold mines at Dolaucothi in central Wales are well known, and recent
MINING AND METALLURGY 95
work has supported their probable Iron Age origins and emphasized their scale and
importance in Roman times (Lewis and Jones 1969; Burnham and Burnham 2004).
Iberia was probably the greatest source of mineral wealth through the first
millennium B.C., certainly in the western Mediterranean (J.C. Allan 1970; Davies
1935: 94-139; Domerque 1990). Silver was the metal that generated the first outside
interest, attested by the Phoenicians in the south from the end of the second
millennium B.C., and it continued to be the most important metal produced
through the first millennium B.C., but gold, lead, copper, and to a lesser extent tin
and mercury were also produced. The silver occurred in the form of both jarosite
and argentiferous lead ores, especially in the south. Gold was mined principally in
the northwest (J.C. Allan 1970: 29-33; Lewis and Jones 1970; Domerque and Herail
1999), and tin (J.C. Allan 1970: 23-29; Penhallurick 1986: 95-104) and mercury were
worked in northern and central Iberia. Copper and lead were mined all over the
peninsula (Davies 1935; Domerque 1990).
The silver and copper mines at Rio Tinto were first operated on a substantial
scale by the Phoenicians and their successors, the Carthaginians (Blanco and Luzon
1969; Kassianidou et al. 1995). The famous mine ofBaebelo described by Pliny (HN
33.97) is almost certainly to be identified with Rio Tinto. This was one of the most
important mines of the early Roman Empire (J. C. Allan 1970: 3-9; Rothenberg et
al. 1989). The underground workings explored so far probably date from the end of
the first millennium B.C., but were of considerably earlier origin (Willies 1997).
The Balkans and Greece

Although base metals were produced, the region was more famed overall for pre-
cious metals. In the mid-first millennium B.C., silver from the mines at Laurion in
Attica and gold and silver from Siphnos and Thasos (Wagner and Weisgerber 1988)
in the Aegean were of great importance, but thereafter declined. Latterly the Bal-
kans supplied much of the gold used in the Roman and Byzantine Empires. The
gold mines oflllyria, at Urbas in Dalmatia, operated until about the second century
A.D., and thereafter Dacia, supplying gold from the Aurariae Dacicae, must have
been one of the most advanced and productive mining regions of late antiquity
(Wollmann 1976).
Cyprus and Anatolia

Cyprus, whose very name is derived from "copper," was one of the great sources of
that metal through the Bronze Age and classical antiquity (Muhly et al. 1982).
Comments by Galen (Walsh 1929) suggest it may also have been an early source of
zinc minerals for the brass industry. Anatolia was another important source of
metals (de Jesus 1980; Pernicka et al. 1984). In the first part of the first millennium
B.C., there was clearly a major gold strike in the alluvial deposits of the Pactolus
River in the immediate vicinity of Sardis. Geophysical surveys carried out in the
early 2000s (as yet unpublished) have revealed major buried trenches in the gravels
of the Pactolus just to the north of the city. These have been interpreted as docks,
but it seems more likely that they are the old gold workings. By the first century
B.C., Strabo described these gold mines as exhausted (13.1.23). While they lasted,
they were the source of the fabled Lydian wealth and the gold for the new inno-
vation of coinage (Ramage and Craddock 2000; Craddock et al. 2005).
Copper is relatively abundant at many places, with evidence of major early
mining along the Black Sea hinterland from Kiire to Rize, and also at Ergani in
the southeast (Wagner and 6ztunali 2000). Silver/lead/zinc ores are also found
in the Black Sea region as well as further west, as is exemplified by the Bayla Madan
mine (see below).
Syria and Palestine

The principal mines on both sides of the Wadi Arabah exploited copper. These are
the mines in the Wadi Feinan, now in Jordan, which include the well-preserved
Roman mining system at Umm el Amad (Hauptmann and Weisgerber 1992), and
further south on the other side, the mines in Beer Ora and the Wadi Amran
(Rothenberg 1972: 208-23; Willies 1991). The deposits are somewhat unusual in that
they are either of malachite (copper carbonate hydroxide) or of chrysocolla (hy-
drated copper silicate), rather than the more common sulphidic deposits.
Egypt
The mines of the eastern deserts of Egypt were a major source of gold from
Pharonic times through classical antiquity, as is attested by the writings of Aga-
tharchides (Burstein 1989) and their physical remains (Klemm and Klemm 1989).
MINING TECHNOLOGY
The first millennium B.C. saw developments in mining technology from Spain to
China that were radical in both scale and concept. Mining had previously been
constrained by problems of ventilation and drainage, which meant that a mine could
not penetrate far below the water table and a gallery could not extend far from the
shaft into the deposit. Thus a major Bronze Age mine such as Timna, in the Negev
Desert of southern Israel, was no more than a large collection of very small mines,
with thousands of shafts often only a few meters apart (Conrad and Rothenberg
1980). In the first millennium B.C., much of this changed. At just the time when
mathematics and mechanics were beginning to receive serious study through the
Greek world, major mines were being systematically laid out for the first time
(Craddock 1995: 69-92; Healey 1978). Deep shafts were now sunk that penetrated far
below the water table, with galleries stretching out for many hundreds of meters. The
sciences of hydrostatics, pneumatics, and mechanics were being applied, if only
empirically. (For a collection of relevant Greek and Latin literary sources, see
Humphrey et al. 1998: 173-92.)
Where the deposits lay in mountainous country, there was always the possi-
bility of driving a passage (known in English as an adit) into the hillside up from the
valley bottom to link with the workings inside. This would drain the workings
above the adit. Rio Tinto provides good examples of a variety of drainage systems
(Palmer 1926-1927; Salkield 1987: 10, 40; Willies 1997; Guerrero Quintero 2006). An
adit was driven for over 3 km through barren rock just to meet up with and drain
the mine workings. When the mines were reopened in the nineteenth century, this
adit was cleared and for the next century continued to drain the modern workings.
Where the workings were below the level of the adit, the water had to be raised
up to it. In antiquity this could be achieved in several different ways (Oleson 1984). A
series of small reservoirs could be built up an incline and water bailed from a lower
to the next higher reservoir and so on until the drainage adit was reached. Such
systems were still in use in Japan in the nineteenth century (Gowland 1899). Pliny
(HN 33.97) seems to describe something similar: "The Baebelo mine ... now dug
1,500 paces (2,200 m) into the mountain. Along this distance watermen are posi-
tioned day and night, pumping out the water in shifts measured by lamps, and
making a stream." There is probably some conflation here. It is more likely that the
watermen bailed the water up to a long adit, which, if it was anything like the existing
adit at Rio Tinto, sloped down, allowing the water to flow away out of the mine.
A variety of mechanical devices for the raising of water seem to have been
developed in the Hellenistic world from about the third century B.C. (Oleson 1984,
2000; Wilson 2002; see chapter 11). Water-wheels were another method of raising
water used in the mines of the south of Spain and elsewhere (Palmer 1926-1927;
Weisgerber 1979). The series of wheels at Rio Tinto raised water through about 30 m.
They were operated by slaves treading the rim, and it is estimated that one person
could raise about 80 liters of water per minute through about 4 m. Another method
of raising water used at this period employed the Archimedean screw, examples of
which have been found in ancient mines in France and Spain. A force pump with
moveable nozzle was found in the Roman mine at Sotiel Coronoda (figure 13.5).
The movement of air through the mine workings was probably achieved mainly
by the use of carefully placed fires at the bottom of shafts in conjunction with
shuttering and doors within the mines. Air could be drawn up one shaft by the fire at
its base, pulling air through the workings drawn in from another shaft. This result
could be achieved with a single shaft with a partition down the middle, sucking the air
in one side, through the workings, and up the other side (Healy 1978: 83). Pliny (HN
31.49) also describes the use oflinen sheets, shaken in the manner of an Indian punka.
The usual method of breaking up hard rock since the inception of mining had
been fire-setting, and this continued long after the introduction of gunpowder for
blasting in seventeenth-century Europe (Craddock 1992, 1995: 33-37; Willies 1994;

Weisgerber and Willies 2000). The traditional perception has been that it was nec-
essary to extinguish the fire suddenly with water, but more recent experimental
work suggested that this was unnecessary and often impracticable. However, some
classical sources, including Pliny (HN 23.57, 33.71), not only describe dousing but
also suggest it should be done with vinegar. This seems inherently unlikely, but
experiments on limestone showed it worked remarkably well (Shepherd 1992).
Pliny (HN 33.72-73) also describes a method of exposing mineral deposits by un-
dermining them with tunnels allowed intentionally to collapse (arrugiae).
The main innovation in mining technology in the first millennium B.C. was the
introduction of iron and steel tools, completely replacing the stone mining ham-
mers and array of bone, antler, and bronze tools that had been used previously. By
the end of the first millennium, the complete premodern kit of mining tools was in
use: picks, hammers, wedges, chisels, hoes, and rakes (figure 4.1). Metallographic
examination of a selection of Roman tools from Rio Tinto has shown that many are
of heat-treated steel, with properties comparable to their modern counterparts
(Maddin et al. 1996). Pliny (HN 33.72) described the use of machines with ham-
merheads of iron weighing 150 Roman pounds (49 kg) for breaking the silices
(quartz, not flint, in these instances). He could be referring to the iron-shod stamps
for crushing the ore (see below), but the passage implies that they were used
underground against the rock face, so possibly large battering rams are meant, such
as have been found in the Indian mines of the same period. The mined material was
removed in small baskets or trays. The broad straight galleries in some of the larger
mines would be suitable for wheeled transport, and at Tres Minas parallel grooves
in the floor seem to indicate some form of controlled trackway (Wahl 1998: fig. 3).
The wheelbarrow was unknown to the Roman world.
In addition to the deep mines, open cast pits were also worked, especially for
gold. Good examples survive in Spain (Lewis and Jones 1970; Domerque and Herail
1999; Bird 2004). The overburden was washed from these workings by the sudden
release of millions of liters of water stored in reservoirs above the workings, known
in English as hushing. Pliny (HN 33.74-76) provides a dramatic description of the
Spanish operations. This method of mining is apparently a Roman innovation,
relying on their practiced ability to collect, store, and control the release of huge
volumes of water.
In contrast to the application of the mathematical and engineering sciences,
there seems to have been but little appreciation of the geology of the deposits. From
the post-medieval period on, miners attempted to define the extent of a deposit,
then to sink a shaft or to drive an adit beneath it. They then excavated upward
(known as overhand stoping), bringing down the material to the base of the shaft or
adit from where it could be conveniently raised to surface. In antiquity an ore body,
once located, was dug out directly (known as underhand stoping), the ore and
waste being brought to surface through the worked-out deposit. Very often de-
posits are faulted, which means that the deposit has slipped and continues some
distance away. Sometimes there clearly was an understanding that the vein was
likely to continue somewhere in the vicinity and efforts were made to try and
Figure 4.1. Selection of Roman iron and steel tools from Rio Tinto mine, including wedges
or gads (bottom left), picks (bottom right) and hoes (top) . Huelva Museum. (Photograph
by P. Craddock.)
relocate it. At the Roman mines in the Wadi Amran (Israel), a series of exploratory
holes each several meters in length had been cut in several directions at the end of
gallery where the deposit terminated at a fault (Willies 1991).
MINE ORGANIZATION
The organization and administration of mining in classical antiquity have been

much discussed but are still imperfectly understood (Davies 1935: 1-16; Healy
1978: 103-32; Hopper 1979: 164-89). The evidence is derived from a large number
of disparate sources, chance references, and inferences that are now difficult to
interpret and from which it is difficult to obtain an overall coherent picture, if such
a concept is even realistic. Contemporary accounts that do survive often seem to be
in serious error, or seriously prejudiced. Hagiographers, for example, recorded the
apparently appalling working conditions suffered within the mines by the Christian
martyrs. Few direct records of real legislation and administration exist, the famous
bronze tablets from a mine of the second century A.D. at Aljustrel in Portugal being
an important exception (Elkington 2001), from which a very different picture of a
more organized and workable management regime emerges.
In most societies, the ruling authority lays claim to all valuable mineral deposits
in the ground, and in classical antiquity the state took an active interest in getting
the maximum yield, especially from the deposits of precious metals. It seems that
the state sometimes operated the mines directly, at least in their developmental
stages in newly won territories, as exemplified by the apparent military involve-
ment in both the lead and gold mines in Britain (cf. Tacitus, An. 11.20; Aristotle,
Ath. pol. 47.2).
From some surviving descriptions and contracts, particularly from Greece and
Spain, it appears that the mines were often leased to private concerns, although they
were still answerable to the state mining department. At Laurion, for example,
several different concessions operated within one mine (Crosby 1941, 1950). This
system could explain why the mining operations seem to have been confined to
digging in the ore. There would be little incentive to develop the infrastructure of
the mine if a concession of only a few years was in operation. It is clear, however,
from the large-scale drainage works in the mines at Rio Tinto and elsewhere that
long-term development work away from the immediate ore body was carried out,
and it does seem that state involvement went far beyond just collecting the dues,
ensuring that the deposits were worked to the best overall advantage. The second-
century A.D. Aljustrel tablets from Spain are important because they list all sorts of
detailed regulations regarding the maintenance of drainage systems, support tim-
bers, and ore dumps (Mommsen et al. 1909: 293-95, no. 113). This was not done out
of any altruistic concern for the welfare of the miners, but pragmatically to ensure
maximum efficiency. In the major silver and gold mines, the state monitored the
operations quite carefully and probably sometimes ran the mines directly. Pliny,
who had been a fiscal procurator in Hispania Tarraconensis, wrote what are clearly
detailed firsthand descriptions of the gold mining operations, and his compre-
hensive and technically accurate descriptions of the silver smelting processes show
that state officials could be fully conversant with the processes involved.
Much of the labor force seems to have been made up of slaves, and criminals
could be sentenced to the mines as a form of essentially capital punishment (Xe-
nophon, Vect. 4.14-18; Plutarch, Nie. 1.1-2; Diodorus 3.12-3.14.4). As a result, there
are very few representations of miners (figure 2.12; cf. figure 2.1). In remote areas,
the local population could be conscripted to work in the mines on a corvee system.
As well as this enforced labor, there were clearly also specialist technical staff who
would be necessary to operate the mine and who, even if not free, were living a more
tolerable life than is popularly perceived (Mrozek 1989).
BENEFICIATION OF ORES
This is an important stage in the production of metal, as Merkel (1985, 1990) and
Marechal (1985) have stressed. The ores of most metals needed to be reduced to
between the size of a pea and a walnut for smelting and would usually have been
broken up manually with hammers. Trip hammers were possibly in use by the
Roman period, as the distinctive stone anvils (figure 4.2) survive at a number of
mines in Iberia and elsewhere (Wahl 1993, 1998; Sanchez-Palencia 1989), and they
may have been water-powered (M. J. T. Lewis 1997, and below). Gold ores needed
to be finely ground to release the tiny particles of metal, and for this grinding mills
were used, based on the mills used for grinding food grains but with harder
grinding stones (Diodorus 3.13.2-14-4; Wilson 2002) . For other metals, where the
ore had to be smelted, fine powders would probably have been mixed with animal
dung to make discrete lumps to charge into the furnace. This was the practice in
the Middle East and India until the nineteenth century (Craddock 1995: 152, 161).
The crushed ores could be concentrated by washing and allowing the lighter
dross to wash away. This process could be effected by shoveling the ore against a
controlled flow of water, but more elaborate arrangements are also known, most
famously the washeries from the mines at Laurion in Attica (Conophagos 1980;
Jones 1984; Photos-Jones and Ellis Jones 1994). These are very sophisticated and
well-preserved systems, but their true function is still not fully understood, and they
are unique to Laurion. At Rio Tinto, more simple washing floors for concentrating
the jarosite silver ore prior to smelting have been excavated (figure 4.3); in these,
water was introduced from the left, washing dross down the large drain on the right.
Figure 4.2. Anvil of an ore-crushing stamp battery from the Roman gold mines at Tres
Minas. Deutsches Bergbau Museum, Bochum. (Photograph by P. Craddock.)
Figure 4.3. Roman washing floor at Rio Tinto. (Photograph by P. Craddock.)
SMELTING
In contrast to mining, the smelting methods that had evolved in the eastern
Mediterranean and Middle East during the third and second millennia B.C. re-
mained largely unchanged in the classical cultures (for a collection of Greek and
Latin literary sources, see Humphrey et al. 1998: 205-33) . The typical smelting
furnace through antiquity was a cylindrical shaft of clay and stone (Craddock 1995:
169-74). The dimensions were quite small, typically about 30 to 60 cm internal
diameter, and 1-2 m tall. Taller and presumably broader furnaces are recorded by
Strabo (3.2.8) for the smelting of silver ores. Based on Strabo and the evidence from
Laurion, Conophagos believed some of the furnaces could have been between three
and four meters in height (see below) .
Early furnaces were very impermanent structures; archaeologists usually have
to reconstruct their form from little more than their bases. Furnaces are re-
presented on some Greek vases depicting scenes of metalworking, but there are
problems with interpretation. The furnaces usually appear to have a pot set on the
top, sometimes complete with lid (Oddy and Swaddling 1985). In this type of
furnace, however, the objective is to ensure an unimpeded flow of air through the
chimneylike system to remove spent gases up the shaft, and a pot set on top would
prevent this (figures 4-4, 16.2). Possibly these representations all illustrate furnaces
used specifically for annealing purposes, but a simple hearth would seem better
suited for that purpose. The illustrations remind us that there is still much to learn
about early metallurgical processes.
One of the few real improvements to take place in smelting technology was the
introduction of more efficient bellows. Early bellows, in use from about the third
Figure 4.4. Shaft furnace on a Greek oenochoe, 525 B.C. BM inv. B507. (From Oddy and
Swaddling 1985; by permission of Swaddling.)
millennium B.C., were either bag-bellows or pot-bellows that could deliver no

more than about 200 liters a minute of air (Tylecote 1981; Merkel 1990). The well
known carving of the workshop of Hephaestus on the north frieze of the Siphnian
Treasury at Delphi depicts bag-bellows (Lullies and Hirmer 1960: fig. 48). In the
first millennium B.C., the familiar hinged concertina bellows were introduced into
the Mediterranean area (Weisgerber and Roden 1985, 1986), but they seem not to
have spread beyond Europe. Although still manually operated, they were capable of
delivering about 500 liters of air/minute (Merkel 1990). The inability to deliver air
at sufficient pressure using manually operated bellows was the major constraint on
furnace size. Only in the medieval period did powerful water-driven bellows al-
lowed much larger furnaces to evolve in Europe, initially in large part for the
smelting of iron.
There are no detailed contemporary descriptions of the smelting processes, but
replication experiments, such as those carried out by Merkel (1990), coupled with
scientific examination of the surviving smelting debris, have gone a long way
toward reconstructing the operations. First, the furnace was brought up to a tem-
perature of about 1,000°C by burning charcoal alone for about an hour, after which
roasted ore and more charcoal were added. In the smelting process, the metal
minerals in the ore were reduced to droplets of molten metal, and the waste
minerals formed a molten slag. Much of the success of the operation depended on
the slag (Bachmann 1982). Even the richest ore after beneficiation would still be
made up of about 60 to 80 percent of waste material, typically silica and/or iron
oxides. By themselves, these minerals have very high melting temperatures and
would rapidly choke the furnace. Silica and iron oxides, however, react to form iron
silicates, which have a much lower melting temperature, enabling them to be run
out of the furnace as a liquid, known in English as tap slag. Occasionally, the ore had
just the right proportions of silica and iron minerals to remove all of the waste
material without intervention, but usually there was an imbalance. If there was too
much silica, then a little iron oxide would be added; conversely if there was too
much iron then a little silica would be added (known as the flux) to create a free-
flowing slag. The slag was periodically drained from furnace, and the process
continued with the addition of more ore and fuel. The metal would drain through
the slag to form an irregular mass at the base of the furnace. Typically a smelt lasted
for about six to ten hours, allowing some tens of kilograms of metal to accumulate
in the base of the furnace. This activity would usually take place at night, both to
take advantage of the cooler conditions and to enable those in charge of the smelt to
be better able to judge the color of the flames at the top of the furnace, which was
the best indicator of the conditions within.
The smelting of silver was somewhat more complicated, because by the first
millennium B.C. even the richest remaining silver ores in the Mediterranean region
contained no more than traces of silver, which would certainly have been lost in the
conventional process outlined above. Much silver was obtained from the traces
contained in lead ores (Bachmann 1991; Conophagos 1980, 1982). The lead was
smelted in the usual manner, and the silver went into the lead. The argentiferous
lead was then heated to about 1,000°C and exposed to a blast of air, causing it to
oxidize to lead oxide, in the process known as cupellation (Conophagos 1980, 1989;
Craddock 1995: 221-31). The molten silver, however, did not oxidize, but floated on
the molten lead oxide "like oil on water," according to Pliny (HN 34.159). After
cooling, the litharge was broken and a lump of silver released.
Ancient silver regularly contains from several hundred ppm to a few percent of
gold, leading Meyers (2004) to suggest that the usual sources exploited in antiquity
were oxidized ores such as cerussite (hydrated lead carbonate) or jarosite (a mixed
sulphatic ore), both ores that have enhanced gold contents. It has long been debated
whether the main ore smelted at Laurion was cerussite or galena. Photos-Jones and
Jones (1994) stated the ore found in the washeries was cerussite, but Rehren et al.
(2002) claim that it had been galena originally. With ores such as the jarosites
from Rio Tinto, which contained little or no lead, it was necessary to add lead to
the smelting charge (Craddock et al. 1985; Craddock 1995: 216-21; Dutrizac et al.
1983).
Gold usually occurs as metal, so it could be separated by finely crushing the
rock in which the tiny particles were embedded and then washing the result to
separate the heavier metal. There is some evidence that the Romans were smelting
auriferous iron pyrites, in which the gold is very finely distributed almost at the
molecular level. Silica was added to the roasted iron pyrite to form a slag allowing
the gold to separate and coalesce. Pliny mentions gold smelting (HN 33.69, possibly
33.79), and slag heaps from the process have been found at the Roman mines at
Tres Minas in northern Portugal (Bachmann 1993).
METALLURGY AND THE ENVIRONMENT
A new area of archaeometallurgical research concerns the effects of mining and

smelting on the environment and on the health of those involved in the work. The
fuel used in antiquity was invariably charcoal, and the major smelting establishments
required enormous quantities, necessitating the establishment oflarge areas of man-
aged woodland to supply the timber (Craddock 1995: 193-95; J.C. Allan 1970: 9-11;
Fulford and Allen 1992). In desert areas, such as Palestine, it is possible that smelting
operations had to be curtailed periodically because the charcoal sources in the vicinity
had become exhausted. Strabo (14.6.5) mentions deforestation on Cyprus resulting
from copper refining. The conclusion of the studies of Mighall and Chambers (1993)
on smaller, prehistoric iron and copper mining and smelting operations in temperate
Europe was that they did not bring about long-term degradation.
Another potential source of environmental pollution was noxious fumes. Most
of the nonferrous ores smelted in classical antiquity were sulphidic, and the roasting
and smelting operations produced sulphur dioxide in quantity. In the absence of tall
chimneys, these fumes were not dissipated but remained in the vicinity of the mines.
Xenophon (Mem. 3.6.12) and Strabo (3.2.8) both allude to the problem of fumes and
consequent environmental devastation. This aspect of ancient metal refining is
difficult to study now, but sulphur dioxide pollution certainly was a major problem
in many nineteenth-century mining areas, such as Rio Tinto (Salkield 1987: 43-44).
The highly acidic groundwaters that resulted effectively controlled the mosquito
population, keeping the area free of malaria. Yet another environmental concern is
the poisoning of the surrounding land with heavy metals, particularly with lead and
cadmium from silver production by means of the cupellation of argentiferous lead,
and to a lesser extent from copper production. This very persistent pollution has
been studied by Mighall (2003) for prehistoric copper mining sites and by Pyatt et
al. (2000) for the much larger Roman operations at Feinan. At the latter site there
are enhanced levels of heavy metals in the soil that must inevitably have entered the
local plants, including those grown for food. The ores are near surface, however, and
in places are cut by the streams; thus the local alluvial fans would have enhanced
levels of heavy metals anyway. Similarly, attempts to study how this pollution
affected the local population by analyzing bones from the cemeteries are com-
promised by the propensity of bone to absorb heavy metals from the soil.
REFINING
Copper and the other base metals were refined by remelting in an open crucible and
allowing impurities such as iron, sulphur, arsenic or antimony to oxidize. Some
elements, such as the arsenic and sulphur, evaporated from the molten metal;
others, notably iron oxide, could be removed by skimming the surface, possibly
aided by adding a little silica to form a slag. Fire-refining remained the usual
method of refining until the twentieth century.
One of the major technical developments in classical antiquity was the intro-
duction of gold refining in response to the introduction of gold coinage by the
Lydians. The remains of a gold refinery of the sixth century B.C. have been exca-
vated at their capital, Sardis (Ramage and Craddock 2000; Craddock et al. 2005).
The study of the refinery at Sardis, coupled with Agatharchides' description of gold
refining as practiced in Ptolemaic Egypt in the second century B.C. (Diodorus
3.13.2-14.4; Burstein 1989), has illuminated the ancient gold refining process
(known in English as parting). Native gold typically contains between 5 and 30
percent silver. The gold granules and dust were mixed with salt and possibly alum in
an earthenware vessel. This was placed in a furnace and heated at a carefully
controlled moderate heat to between 600° and 8oo°C for many hours or even days.
The astringent vapors generated by the hot salt penetrated the gold and removed
the silver as a vapor of silver chloride, leaving behind pure gold. The silver could be
recovered from the clay parting vessels and furnace walls by cupellation.
The necessity of producing metal of uniform weight and purity required a
change in the perception of the very nature of metals and introduced the concept of
elements. Our idea of gold as an element with precisely defined properties is based
on relatively modern scientific concepts, in particular on the Law of Constant
Composition, which states that each element has precise and invariant properties.
To us this seems no more than stating the obvious, but the ancients did not have
such a concept of materials as elements. The widely held belief, as set down by
Empedocles, was that metals "grew" in the ground. Thus it would seem logical to
expect the properties of the metal to depend on their environment; metals such as
gold, coming from various sources, could have widely differing properties but still
be gold. For example, a light-colored gold from a given locality would be regarded
as the intrinsic gold of that place, rather than as a naturally occurring alloy of gold
with a rather high silver content.
Coinage introduced the concept of an invariant standard of purity (usually
expressed by the metal refiners in antiquity in practical terms when repeated re-
fining failed to reduce the weight any further). It was also noted that natural gold
could be recreated by adding silver to the refined gold, and conversely that silver
could be recovered from the refining waste. Thus, although natural gold was still
"growing" and still contained silver that would in time turn into gold, there was an
ultimate gold, and through refining the gold could attain that state.
This new perception, brought about by the need to refine precious metals
for coinage, was already extending to other metals by Roman times. For example,
Pliny (HN 34.94) stated that the Cypriot mines produced ductile "bar" copper that
could be hammered without breaking, unlike other mines that produced "fused"
copper, which was too brittle to be hammered. But he continued, "in other
the other mines, this difference of bar copper from fused copper is produced by
treatment; for all copper after impurities have been rather carefully removed by fire
and melted out of it become bar copper." That is, the difference was not just an
intrinsic property of Cypriote copper but was also associated with the degree of
refining.
IRON AND STEEL
In the first centuries of the first millennium B.C., the use of iron spread very rapidly
throughout the Old World, supplanting bronze as the usual material for tools and
weapons, as is exemplified by the tools used in mining (see above). Through most
of its history iron has been used in a variety of states: wrought iron, phosphoric
iron, cast iron, steel, and crucible steel (Craddock 2003).
In Europe, the Mediterranean world, and the Middle East, iron was always
smelted by the bloomery process (Tylecote 1987; Craddock 1995). The furnaces and
smelting conditions were not dissimilar to those used to smelt other metals, al-
though some of the furnaces were of considerable size by the Roman period (Crew
1998). The iron was produced as a solid pasty mass, called the bloom, invariably
containing some slag. This was removed while still white-hot and vigorously ham-
mered to squeeze out as much of the slag as possible, and also to consolidate it
(Diodorus 5.13.1-2). The wrought iron so produced could also contain small
quantities of carbon, sometimes sufficient to be classified as steel.
If the iron ore contained phosphorus, as most bog ores do, then this could
enter the iron. Small quantities of phosphorus have much the same effect as carbon
in iron, but can lead to serious brittleness. For this reason phosphorus is now
regarded as a harmful impurity, but in Iron Age Europe phosphoric iron was
deliberately selected for blades such as sickles and knives, since they kept their edge,
even if they occasionally broke (Craddocki995: 238; Ehrenrich 1985). The smith can
have had no notion of the presence of phosphorus but knew that certain ores, when
smelted in a particular way, resulted in an iron suitable for blades.
If the temperature of the furnace was increased and the conditions made more
reducing, then carbon could dissolve in the iron as it formed, reducing the melting
temperature down to about 1,200°C. Thus, instead of a solid bloom, liquid iron
could be run from the furnace containing about 3.5 to 4.5 percent of carbon (called
"cast iron"). Cast iron is not a very useful material, and it seems not to have been
used in the classical world; the use of cast iron was a great achievement of Chinese
metallurgy. Iron with a small amount of carbon (typically between about 0.2 and
1 percent) is known as steel, and when correctly heat-treated it has properties much
superior to wrought iron (see below); getting a regular amount of carbon into the
solid iron, however, was very difficult.
The ideal material would be iron with about 1 percent of uniformly distributed
carbon and containing no slag. This result could be produced by reacting wrought
iron with charcoal and wood in small sealed crucibles at very high temperatures
(1,500°C) to produce crucible steel. This technology was in widespread use through
central and southern Asia over 2,000 years ago, and it seems that the classical world
had some inkling of the process and possibly imported sword blades made from it
(see below). Crucible steel was always forged to shape, never cast.
Wrought iron could be shaped by hammering at red heat (hot forging) using
methods and tools not dissimilar to those used by traditional blacksmiths to this day
(Tylecote 1987; Sim and Ridge 2002). The iron could not be cast, but good met-
allurgical joins could be made by hammering pieces together at red heat (hammer
welding) or by hot riveting. The archaeological record shows that iron was a very
familiar low-cost material. Estimates have been made for iron production of 2,250
tons per annum in Roman Britain, and 82,500 tones per annum through the rest of
the Empire (Sim and Ridge 2002: 23; cf. Cleere and Crossley 1985: 57-86).
Iron was used on such a scale that technologists such as David Sim have
speculated whether some more advanced production techniques might have al-
ready been in use. For example, it is usually held that the production of wire
through a draw plate is a post-Roman technique, and that iron wire would have had
to be laboriously made by hammering a thin rod. Even a simple tunic of chain mail
(the lorica hamata), however, contained hundreds of meters of wire that it would
not have been feasible to make except by drawing. Draw plates have now been
recognized, appropriately enough in the floor of armorers' workshops inside Roman
frontier forts (Sim 1997, 1998).
Wrought iron could be turned into steel by heating it for many hours in close
contact with a variety of organic materials and charcoal dust (carburizing). There
were two techniques: small pieces of iron could be totally carburized and then
welded to the wrought iron of the remainder of the tool (the tip of a chisel, for
example), or the entire artifact could be forged and then the surface carburized all
over (case hardening).
The famous Noricum iron praised by Pliny (HN 34.145) and others was a
natural steel (Davies 1935: 173; Tylecote 1987: 168). The iron ores in the vicinity of
Magdelensburg in Austria are rich in manganese, and thus the slag had a great deal
of manganese oxide. Ordinary iron slag contains iron oxide, which reacts with any
carbon dissolved in the iron in the furnace. Manganese oxide does not react in this
way, and thus there is a better chance of any carbon in the iron surviving; in the
Noricum iron sufficient carbon remained for it to be classified as a steel. Steel can be
greatly improved by heat treatments, as the Greeks knew as early as the eighth
century B.C. The steel was heated to a cherry-red heat and held at that temperature
for some time before being plunged into water. The resulting steel was very hard,
suitable for files but too brittle for other uses. To produce a steel more suitable for
chisels or saw blades, for example, the quenched steel was gently heated (tempered).
The smith judged the degree of tempering by the color of the oxidized layer on
the steel.
With so many imponderables in the production of good steel, it is not sur-

prising that things often went wrong, and the proportion of tools and weapons that
had been correctly heat-treated is low. Studies of prestige blades, however, have
shown that they are regularly of good-quality, correctly tempered steel (Lang 1995).
Swords could be made by hot-forging and folding the iron. The folding could be
repeated several times, resulting in a blade made up of many thin layers of iron
(Lang 1988), a process known as piling; it was probably done to homogenize the
iron. From the end of the first millennium B.C. in Europe the practice began of
twisting the iron sheets or rods into decorative patterns during the forging, some-
times alternating iron and steel. Pattern welding, as this is called, is usually associated
with the post-Roman barbarians, but the technique was certainly practiced in the
late Roman Empire.
Studies of Roman armor have shown that steel and iron were intelligently and
deliberately combined to produce a superior product (Williams 1977). For example,
the lorica segmenta plates of Roman armor were fabricated from wrought-iron sheet
to which steel sheet had been welded (Sim and Ridge 2002: 96). The outer hard steel
layer would stop penetration, and the inner tough iron layer would prevent the
lorica from breaking and absorb the energy of the blow.
As early as the first millennium B.C., the Chinese were producing iron castings
on a prodigious scale, and crucible steel was being produced in central and southern
Asia (Craddock 2003). What knowledge the Romans had of these materials, or what
access they had to them, is uncertain. Tylecote (1987: 325-27), for example, listed
finds of iron castings from Roman sites and suggested that they could have been
imports from China. Ongoing work in southern Germany has shown that the blast
furnace process developed from about the eighth century A.D., much earlier than
previously believed, although the liquid iron so produced was not used for castings
but instead was treated (fined) to produce solid wrought iron (Craddock 2003). The
ancestry of the process in Europe is not known, but excavations of Iron Age and
Roman iron-smelting sites have shown a hitherto unappreciated variety of processes
(Crew and Crew 1997). At the major iron-smelting site at Oulche (France), there is
evidence that granules of high carbon iron may have been separately collected from
the ordinary iron and forged to produce high carbon steel (Dieudonne-Glad 1997).
It has been claimed that various passages in classical literature refer to liquid
iron or steel, but these can be discounted (Craddock 2003), with the exception of
the writings of the Alexandrian alchemist Zosimos, who lived in the third century
A.D. (Berthelot 1893: 332). He gave a detailed and accurate description of how the
Indians produced crucible steel, and he indicated that this material was used by the
Persians to make high-quality swords that were then traded to the Romans. Blades
from central Asia dating from the third century A.D. (Feurerbach 2001), and from
Sassanian Persia (Lang et al. 1998), have been identified as crucible steel, but so far
no crucible steel blades have been reported from the Roman or Byzantine Empires.
Since literary evidence suggests that crucible steel was already known in the late
Roman and Byzantine world, it is very likely that continuing research will uncover
more evidence of sophisticated processes and products.
BRASS: THE NEW METAL
Bronze-the alloy of copper and tin-appeared in the fourth millennium B.C. and
was in universal use by the first millennium B.C. Throughout the classical period,
lead was added in ever greater quantities to bronze intended for castings (Craddock
1976, 1977, 1986). The major change in copper alloys in the ancient world, however,
was the introduction of brass-an alloy of copper and zinc-sometime within the
first millennium B.C. By the late Roman and Byzantine periods, once the tech-
nology for its preparation had been developed, brass had largely replaced bronze as
the usual copper alloy (Caley 1964; Bayley 1998; Craddock 1978; 1995: 292-302; 1998;
Craddock and Eckstein 2003). Zinc ores are abundant compared to those of tin. The
reason for the relatively late introduction of brass as an alloy is the volatility of zinc.
At the temperatures necessary to smelt it, zinc is a very reactive gas. As a result, the
pure metal was not produced commercially until about A.D. 1000, in India
(Craddock et al. 1998a).
Almost from the inception of metallurgy, some copper artifacts from all over
the world have been found to contain varying quantities of zinc (Welter 2003). They
always seem to be isolated examples among items that otherwise consist of copper
and bronze, and they are probably the unintentional result of smelting zinc-rich
copper ores. From the end of the second millennium B.C. in Anatolia and the
Middle East, copper artifacts containing between about 5 and 15 percent of zinc
become more common, contemporary with references to a special copper known as
"copper of the mountain" in Assyrian and other documents. From the mid-first
millennium B.C. there were also references in Greek literature to oreichalkos,
"copper of the mountain," as a distinct metal, and by the first century B.C. this
term, and its Latin form aurichalcum, certainly referred to brass (Halleux 1973).
Early production seems to have been centered in western Anatolia, although pro-
duction in Persia and east as far as India is a distinct possibility. The first series of
artifacts regularly made of brass that have been identified so far are the base metal
coins of Mithradates VI, who ruled western Anatolia during the early first century
B.C. (Craddock et al. 1980). These issues include both copper and brass coins and
are directly ancestral to the Roman reformed coinage of 27 and 23 B.C., which saw
the introduction of brass dupondii and sestertii (Burnett et al. 1982) (figures 4.5,
30.1). It seems likely that through the first century B.C. brass rapidly gained in
popularity and was extensively used by the military. Istenic and Smit (forthcoming)
have shown that the Alesia type of brooches, associated with the military, are of
brass from about 60 B.C., and by the first century A.D. military brooches, trappings,
and fittings were regularly made of brass (Jackson and Craddock 1995; Ponting and
Segal 1998). These brasses and the contemporary brass coins (Calliari et al. 1998) are
usually an alloy of copper with about 20 to 25 percent of zinc and minor amounts of
lead derived from the zinc ore. The contemporary civilian metalwork is often of an
alloy containing less zinc but more lead along with some tin, suggesting the in-
discriminate mixing of scrap bronze and brass.
•• --
••
.
.
'
•
Figure 4.5. Early brass coins. Top: Early first century B.C., Pergamum (BMC 130, 144).
Middle: Roman SC coin of the reformed eastern coinage of 27 B.C. Bottom: dupondius
and sestertius of the general reform of 23 B.C. (After Craddock 1995.)
The ancient written sources do not state how brass was made. The only physical
evidence is a number of distinctive small, lidded crucibles, heavily vitrified and
massively impregnated with zinc (figure 4.6) (Bayley 1998). It has always been
assumed that the alloy was made by cementation, the process that was used in post-
medieval Europe up to the nineteenth century (Rehren 1999). In this process the
calcined zinc ore together with charcoal was added to finely divided solid copper in
a closed crucible and heated to temperatures of about 900°C, whereupon the zinc
oxide was reduced to zinc gas that dissolved into the copper. The first mention of
cementation, however, is by Biringuccio in the 1540s (Smith and Gnudi 1942: 70-
76), and the only medieval accounts-those of Theophilus, dating from the twelfth
century (Hawthorne and Smith 1963), and those of al-Hamdani and al-Kashani,
dating from the tenth and twelfth centuries, respectively (J. W. Allan 1979; Crad-
dock et al. 1998b )-all make it clear that the zinc ore and charcoal were added to
molten copper in the crucible, rather than to solid shavings. This process was much
simpler and quicker, but less zinc entered the copper (Craddock and Eckstein
2003) . That this was the process used in classical antiquity is supported by the
composition of the ancient brasses. Roman and medieval brasses contain up to a
maximum of 28 percent of zinc, but after the sixteenth century this proportion rises
to a maximum of 33 percent.
There was a further complication to the production of brass. The ore added to
the copper must contain no sulphur. The zinc deposits of western Europe that were
very probably worked by the Romans (at Stolberg in Germany, Vieux Montaigne in
Belgium and the Mendip Hills in Britain) are carbonates and could have been used
directly, but the zinc ores of Anatolia and the Middle East are sulphides. The
sulphidic ores first had to be roasted to produce zinc oxide. As zinc oxide had
'
. •.·..
. \
r.i:.i
• Ii.
. •
'.
•
I
f5~./
-·--2'~
~ ... I
I ' \
'
.~ :-~~
~·
.
%-.. ;: .. '
'i•i-~,. t ... ,'/;
I
/1
J
I
\
., 't-·..-,.>«,.,.. ·---· !' /

I I
I I
a J
~//
b
o~---~s
ems
d
Figure 4.6. Small Roman lidded crucibles probably used to make brass; a-b from Culver,
St. Colchester; c from Palace St. Canterbury; d from extramural settlement at Ribchester.
(After Bayley 1998; copyright English Heritage, by permission.)
medicinal uses-the familiar zinc ointment or calamine lotion (Lehmann 2000 ) -

there are detailed descriptions of its preparation in contemporary pharmaceutical
texts, especially that of Dioscorides (Gunther 1934; Riddle 1985). The sulphidic ore
was roughly roasted to convert most of it to the oxide, and then smelted in an or-
dinary shaft furnace. The zinc vapor ascended the furnace flue where it met oxygen
and promptly reoxidized. Complex arrangements of iron bars suspended in the flue,
or even of whole chambers above the flues, were set up to catch and condense the
fumes of zinc oxide. Debris of this process has been excavated at Zawar in India,
dating from the last centuries B.C. (Craddock and Eckstein 2003). The process
continued in use in the Middle East at least until the thirteenth century, when it was
observed by Marco Polo at Kerman in Persia.
The zinc content of the dupondii and of the sestertii declined during the first
century, leading Caley (1964) to believe that brassmaking had already become a lost
art. It is true that coins of the central mint cease to be made of brass, but analyses of
other classes of metalwork show that it continued in production (Craddock 1978;
Dungworth 1996). More interestingly, the major analytical study of thousands of

brooches from Roman Britain (Bayley and Butcher 2004) has shown that different
types of brooch used different and very specific alloys, some of brass and others of
leaded bronze. Similarly, local issues of copper-base coins in the eastern empire
were often struck in brass, whereas the corresponding issues from the central mints
were of bronze (Cowell et al. 2000). After the collapse of the western empire, with
the attendant loss of the tin producing provinces of Iberia and Britannia and easy
access to the tin fields of the Erzgebirge in central Europe, brass became the usual
alloy (Hook and Craddock 1996; Schweizer 1994). The earliest Islamic metalwork
reverted to bronze (Schweizer and Bujard 1994; Ponting 1999), presumably until the
capture of the brass-making sources in Anatolia, after which brass replaced bronze
almost completely (Craddock et al. 1998b).
SUR VIV AL AND DEVELOPMENT
The production of metals survived the collapse of the Roman Empire in the west,
with every indication, for example, that tin production in Britain and brass pro-
duction in Flanders continued. The Byzantine successors of the Roman Empire in
the East seem to have maintained production, and this in turn was continued by the
Islamic regimes. These later mines and smelters have been poorly investigated, and
so comparison with earlier classical practice is difficult. Not until the sixteenth
century do detailed descriptions appear for comparison, especially those of Bir-
inguccio (Smith and Gnudi 1942) and Agricola (Hoover and Hoover 1912). It is clear
that Agricola (a pseudonym for Georg Bauer) was familiar with Strabo's Geography
and Pliny's Natural History, and in some ways his De re metallica contains striking
parallels (Domerque 1989b). There are also striking and significant differences, not
least between the authors and their intended audiences. Strabo and Pliny wrote as
interested and educated observers, and their work was intended for gentlemanly
diversion, not for practical instruction. Agricola was trained as a doctor with strong
interests in geology and metallurgy, and he wrote his works with official encour-
agement as real working manuals.
Clearly, both the classical and Renaissance mining ventures could be on a
considerable scale. Investigations at Rio Tinto, Tres Minas, and other sites have
shown that sophisticated drainage methods were employed. The major difference
seems to have been the prevalence of water-powered, and to a lesser extent animal-
powered, machinery (M. J. T. Lewis 1997). Water power certainly was harnessed in
the classical world, but how widespread was its use? This has been a major issue in
the history of technology for many years (see chapters 6 and 13); some scholars, such
as Wilson (2002), believe that water-powered devices were quite common even in
mining operations. A thousand years after Rome, the pages of De re metallica are
full of descriptions and illustrations of drainage pumps, crushing machinery, and

bellows all powered by water wheels. These do seem to be generally absent in
Roman mines, although it is dangerous to assume their absence. In the few years
since the stamps at Tres Minas were announced, several more have been recognized,
including one at Dolaucothi in Wales adjacent to a Roman water mill and sur-
rounded by heaps of crushed debris (Burnham and Burnham 2004). Thus water
power might have been more prevalent than believed at present, although it does
seem that earlier mining relied far more on large labor forces.
There are more clearcut differences between classical and medieval smelting
practices because of the application of water power. As noted above, the early fur-
naces were necessarily small because of the limitations of manually operated bel-
lows to produce a blast of sufficient volume and power to penetrate a larger furnace
charge. With the introduction of water power the bellows grew to enormous
dimensions, and the furnaces expanded to over a meter in internal diameter and
several meters in height. Especially in iron smelting, a single operation could last
for weeks, producing many tons of metal with just a few highly skilled operators,
totally different from the small-scale, labor-intensive operations of the classical
world.
REFERENCES
Allan, J. W. 1979. Persian metal technology, 700-1300 AD. London: Ithaca Press.
Allan, J. C. 1970. Considerations on the antiquity of mining in the Iberian peninsula. London:
Royal Anthropological Institute.
Bachmann, H-G. 1982. The identification of slags from archaeological sites. London: Institute
of Archaeology.
Bachmann, H-G. 1991. "Archaometallurgie des Silbers." Die Geowissenschaften 9.1: 12-16.
Bachmann, H-G. 1993. "Zur metallurgie der romischen Goldgewinnung in Trees Minas
und Campo de Jales in Nordportugal," in H. Steuer and U. Zimmermann (eds.),
Montanarchiiologie in Europa. Sigmaringen: Jan Torbecke, 53-60.
Badii, G. 1931. "Le antiche miniere del Massetano," Studi Etruschi 5: 455-73.
Bayley, J. 1998. "The production of brass in antiquity with particular reference to Roman
Britain," in Craddock 1998: 7-25.
Bayley, J., and S. Butcher 2004. Roman brooches in Britain. London: Society of Antiquaries
of London.
Berthelot, M. P. E. 1893. La Chemie au Moyen Age. Paris: Georges Stenheil.
Bird, D. 2004. "Water power in Roman gold mining," Mining History: Bulletin of the Peak
District Mines Historical Society 15.4/5: 58-63.
Blanco, A., and J.M. Luzon 1969. "Pre-Roman silver mining at Rio Tinto," Antiquity 43:
14-31.
Burnett, A. M., P. T. Craddock, and K. Preston 1982. "New light on the origins of or-
ichalcum," in T. Hackens and R. Weiller (eds.), Proceedings of the 9th International
Congress of Numismatists. Louvain-la-Neuve: Association internationale des numis-
mates professionnels, 263-68.
Burnham, B., and H. Burnham 2004. Dolaucothi-Pumsaint: Survey and excavations at a

Roman gold-mining complex 1987-1999. Oxford: Oxbow.
Burstein, S. M. (trans. and ed.) 1989. Agatharchides of Cnidus: On the Erythraean Sea.
London: The Hakulyt Society.
Caley, E. R. 1964. Orichalcum and related ancient alloys. Numismatic Notes and Mono-
graphs 151. New York: American Numismatic Society.
Calliari, I., M. Magrini, and R. Martini 1998. "Characterization of Republican and Imperial
Roman coins," Science and Technology for Cultural Heritage 7.2: 81-89.
Cambi, L. 1959. "Problemi della metallurgia," Studi Etruschi 27: 415-32.
Cauuet, B. 1995. Les mines d'or gauloises du Limosin. Limoges: Association Culture et
Patriomoine en Limoussin.
a
Cauuet, B. 1999. "L'exploitation de l'or en gaulle l'Age du Fer," in B. Canuuet (ed.), L'or
dans l'antiquite. Toulouse: Aquitania, 31-86.
Cleere, H., and D. Crossley 1985. The iron industry of the weald. Leicester: The University
Press.
Conophagos, C. E. 1980. Le laurion antique et la technique grecque de la production
del'argent. Athens: Ekdotike Hellados.
Conophagos, C. E. 1982. "Smelting practice at Laurion," in T. A. Wertime and S. F. Wertime
(eds.), Early pyrotechnology. Washington, DC: Smithsonian Institution Press, 181-91.
Conophagos, C. E. 1989. "La technique de la coupellation des Grecs anciens au Laurium,"
in Y. Maniatis (ed.), Archaeometry. Amsterdam: Elsevier, 271-89.
Conrad, H-G., and B. Rothenberg 1980. Antikes Kupfer im Timna Tal. Der Anschnitt
Monograph 1. Bochum: Bergbau Museum.
Cowell, M. R., P. T. Craddock, A. W. G. Pike, and A. M. Burnett 2000. "An analytical
survey of Roman provincial copper-alloy coins and the continuity of brass production
in Asia Minor," in B. Kluge and B. Weisser (eds.), XII. Internationaler Numismatischer
Kongress Berlin 1997. Berlin: Association Internationale des Numismates Profes-
sionnels, 70-77.
Craddock, P. T. 1976. "The composition of the copper alloys used by the Greek, Etruscan
and Roman civilisations, 1: The Greeks before the archaic period," Journal of Archae-
ological Science 3.2: 93-113
and Roman civilisations, 2: The archaic, classical and Hellenistic Greeks," Journal of
Archaeological Science 4.2: 103-24.
and Roman civilisations, 3: The origins and early use of Brass," Journal of Archae-
ological Science 5-1: 1-16.
Craddock, P. T. 1986. "The metallurgy and composition of Etruscan bronze," Studi
Etruschi 52: 211-71.
Craddock, P. T. 1992. "A short history of firesetting," Endeavour 16.3: 145-50.
Craddock, P. T. 1995. Early metal mining and production. Edinburgh: Edinburgh University
Press.
Craddock, P. T. (ed.) 1998. 2000 years of zinc and brass. 2nd ed. British Museum Occasional
Paper 50. London: British Museum.
Craddock, P. T. 2003. "Cast iron, fined iron, crucible steel: Liquid iron in the ancient
world," in Craddock and Lang 2003: 231-57.
Craddock, P. T., A. M. Burnett, and K. Preston 1980. "Hellenistic copper-base coinage and
the origins of brass," in W. A. Oddy (ed.), Scientific Studies in Numismatics. British
Museum Occasional Paper 18. London: British Museum, 53-64.
Craddock, P. T., and M. J. Hughes (eds.) 1985. Furnaces and smelting technology in an-
tiquity. British Museum Occasional Paper 48. London: British Museum.
Craddock, P. T., I. C. Freestone, N. H. Gale, N. D. Meeks, B. Rothenberg, and M. S. Tite
1985. "The investigation of a small heap of silver smelting debris from Rio Tinto,
Huelva, Spain," in Craddock and Hughes 1985, 199-218.
Craddock, P. T., I. C. Freestone, L. K. Gurjar, A. Middleton, and L. Willies 1989. "The
production of lead, silver and zinc in ancient India," in Hauptmann, Pernicka, and
Wagner 1989: 51-70.
Craddock, P. T., I. C. Freestone, L. K. Gurjar, A. P. Middleton, and L. Willies, L. 1998a.
"Zinc in India," in Craddock 1998: 27-72.
Craddock, P. T., S. C. La Niece, and D.R. Hook 1998b. "Brass in the medieval Islamic
world," in Craddock 1998: 73-114.
Craddock, P. T., and K. Eckstein 2003. "Production of brass in antiquity by direct re-
duction," in Craddock and Lang 2003: 216-30.
Craddock, P. T., and J. Lang (eds.) 2003. Mining and metal production through the ages.
London: British Museum Press.
Craddock, P. T., M. R. Cowell, and M.-F. Guerra 2005. "Controlling the composition of
gold and the invention of gold refining in Lydian Anatolia," in -0. Yal<rin (ed.),
Anatolian metal, III.Der Anschnitt Monograph 11. Bochum: Bergbau Museum, 67-77.
Crew, P. 1998. "Laxton revisited," Journal of the Historical Metallurgy Society 32.2: 49-53.
Crew, P., and S. Crew (eds.) 1997. Abstracts: Early ironworking in Europe. Maentwrog:
Snowdonia National Park.
Crosby, M. 1941. "Greek inscriptions: A poletai record of the year 367/6 B.C.," Hesperia
10: 14-30.
Crosby, M. 1950. "The leases of the Laureion mines," Hesperia 19: 189-312.
Davies, 0. 1935. Roman mines in Europe. Oxford: Oxford University Press.
Dejonghe, L., E. Ladeuze, and D. Jans 1993. Atlas des gisements plombo-zinciferes syncli-
norium de Veviers (Est de la Belgique). Brussels: Service Geologique de Belgique.
Dieudonne-Glad, N. 1997. "Oulches, a late Roman iron smelting and smithing workshop,"
in Crew and Crew 1997: 17-20.
Domerque, C. (ed.) 1989a. Mineria y metalurgia en las antiquas civilizaciones mediterraneas
y Europas I & II. Madrid: Institutio de Conservaciones y Restauracion de Benes
Culturales.
Domerque, C. 1989b. "Les techniques minieres antiques et le De re metallica d'Agricola," in
Domerque 1989a: II, 76--95.
Domerque, C. 1990. Les mines de la Peninsule Iberique dans l'Antiquite romaine. Bib-
liotheque de l'Ecole fran<raise de Rome 127. Rome: Ecole fran<raise de Rome
Domerque, C., and G. Herail 1999. "Conditions de gisement et exploitation antique Las a
Medulas (Leon, Espagne)," in B. Cauuet (ed.), L'or dans l'antiquite. Toulouse:
Aquitania, 93-116.
Dungworth, D. 1996. "Caley's 'zinc decline' reconsidered," The Numismatic Chronicle 156:
228-34.
Dutrizac, J.E., J. L. Jambor, and J.B. O'Reilly 1983. "Man's first use of jarosite: The pre-
Roman mining and metallurgical operations at Rio Tinto, Spain," Bulletin of the
Canadian Institute of Mining and Metallurgy 76.859:78-82.
Ehrenreich, R. M., 1985. Trade, technology and the ironworking community in the iron age in
southern Britain. British Archaeological Reports 144. Oxford: BAR.
Elkington, D. 2001. "Roman mining law," Mining History: Bulletin of the Peak District
Mines Historical Society 14.6: 61-65.
Feurerbach, A. M. 2001. Crucible steel in central Asia. PhD Thesis, Institute of Archaeology,
University College, London.
Fulford, M. G., and J. R. L. Allen 1992. "Iron-making at the Chesters Villa, Woolaston,
Gloucestershire: Survey and excavations 1987-91," Britannia 23: 188-91.
Gowland, W. 1899. "The early metallurgy of copper, tin and iron in Europe, as illustrated
by ancient remains, and the primitive processes surviving in Japan," Archaeologia 56.
2: 267-322.
Guerrero Quintero, C. (ed.) 2006. La rueda elevadora de aqua romana de Rio Tinto en el
Museo de Huelva. Seville: Junta de Andalucia, Consejeria de Cultura.
Gunther, R. T. (ed.) 1934. The Greek Herbal of Dioscorides: John Goodyer's translation of
1655. Oxford: Oxford University Press.
Halleux, R. 1973. "L'orichalque et le laiton," Antiquite Classique 42: 64-81.
Hauptmann, A., E. Pernicka, and G. A. Wagner (eds.) 1989. Old world archaeometallurgy.
Der Anschnitt Monograph 8. Bochum: Bergbau Museum.
Hauptmann, A., and G. Weisgerber 1992. "Periods of ore exploitation and metal pro-
duction in the area ofFeinan, Wadi Araba, Jordan," in M. Zaghoul et al. (eds.), Studies
in the history and archaeology of Jordan, vol. 4. Amman: Department of Antiquities,
61-66.
Hawthorne, J. G., and C. S. Smith (eds.) 1963. On divers arts: The treatise of Theophilus.
Chicago: Basic Books.
Healy, J. F. 1978. Mining and metallurgy in the Greek and Roman world. London: Thames
and Hudson.
Hong, S., J.-P. Candelone, C. C. Patterson, and C. F. Boutron 1996. "History of ancient
copper smelting pollution during Roman and medieval times recorded in Greenland
ice," Science 272: 199-206.
Hook, D. R., and P. T. Craddock 1996. "The scientific analysis of the copper-alloy lamps:
Aspects of classical alloying practices," in D. M. Bailey, A catalogue of the lamps in
the British Museum: IV. Lamps of metal and stone and lampstands. London: British
Museum Press, 144-63.
Hoover, H. C., and L. H. Hoover (trans. and eds.) 1912. Agricola: De re metallica. London:
The Mining Magazine.
Hopper, R. J. 1979. Trade and industry in classical Greece. London: Thames and Hudson.
Humphrey, J. W., J. P. Oleson, and A. N. Sherwood 1998. Greek and Roman technology:
Istenic, J., and Z. Smit forthcoming. "The beginning of the use of brass in Europe with
particular reference to the south-eastern Alpine region," in S. La Niece, D.R. Hook,
and P. T. Craddock (eds.), Metals and mines: studies in archaeometallurgy. London:
Archetype.
Jackson, R. P., and P. T. Craddock 1995. "The Ribchester Hoard: A descriptive and tech-
nical study," in Raftery et al. 1995: 75-102.
de Jesus, P. S. 1980. The development of prehistoric mining and metallurgy in Anatolia.
British Archaeological Reports, Intl. Series S74. Oxford: BAR.
Jones, J. E. 1984. "Ancient Athenian silver mines, dressing floors and smelting sites,"
Journal of the Historical Metallurgy Society 18.2: 65-81.
Kassianidou, V., B. Rothenberg, and P. Andrews 1995. "Silver production in the Tartessian
period: The evidence from Monte Romero," ARX: World Journal of Prehistoric and
Ancient Studies 1: 17-34.
Klemm, D. D., and R. Klemm 1989. "Antike Goldgewinnung in der Ostwiiste Agyptens," in
Hauptmann, Pernicka, and Wagner 1989: 227-34.
Lang, J. 1988. "Study of the metallography of some Roman swords," Britannia 19: 199-216.
Lang, J. 1995. "A metallographic examination of eight Roman daggers from Britain," in
Raftery et al. 1995: 119-32.
Lang, J., P. T. Craddock, and St. J. Simpson 1998. "New evidence for early crucible steel,"
Journal of the Historical Metallurgy Society 32.1: 7-14.
Lehmann, J. 2000. "Seit wann wird mit zincweio gemalt? Chemische Untersuchungen,
mittelalterliche Quellenschriften," Restauro 106.5: 356-60.
versity Press.
Lewis, P.R., and G.D. B. Jones 1969. "The Dolaucothi gold mines I," Antiquaries Journal
49.2: 244-72.
Lewis, P.R., and G.D. B. Jones 1970. "Roman gold-mining in North West Spain," Journal
of Roman Studies 60: 169-85.
Lullies, R., and M. Hirmer 1960. Greek sculpture. New York: Abrams.
Maddin, R., A. Hauptmann, and G. Weisgerber 1996. "Metallographische Untersuchungen
an ri:imischen Gezahe Rio Tinto, Spanien," Metalla 3-1: 27-44.
Marechal, J. 1985. "Methods of ore roasting and the furnaces used," in Craddock and
Hughes 1985: 29-42.
Merkel, J. 1985. "Ore beneficiation during the Late Bronze Age/Early Iron Age at Timna,
Israel," MASCA Journal 3.5: 164-69.
Merkel, J. F. 1990. "Experimental reconstruction of Bronze Age copper smelting based on
archaeological evidence from Timna," in B. Rotherberg (ed.), The ancient metallurgy
of copper. London: Institute for Archaeo-metallurgical Studies, 78-122.
Meyers, P. 2004. "Production of silver in antiquity: Ore types identified based upon
elemental compositions of ancient silver artifacts," in L. van Zelst (ed.), Patterns and
process. Suitland: Smithsonian Center for Materials Research and Education, 271-88.
Mighall, T. 2003. "Geochemical monitoring of heavy metal pollution and prehistoric
mining," in Craddock and Lang 2003: 43-51.
Mighall, T., and F. M. Chambers 1993. "Early mining and metalworking: Its impact on the
environment," Journal of the Historical Metallurgy Society 27.2: 71-83.
Minto, A. 1955. "L'antica industria mineraria in Etruria ed il porto di Populonia," Studi
Etruschi 23: 317-23.
Mommsen, T., 0. Gradenwitz, and C. G. Bruns 1909. Fontes iuris Romani antiqui. 7th ed.
Tubingen: Mohr.
Mrozek, S., 1989. "Le travail des hommes libres dans les mines romaines," in Domerque
1989: 163-71.
Muhly, J. D., R. Maddin, and V. Karageorghis (eds.) 1982. Early metallurgy in Cyprus, 4000-
500 B.C. Nicosia: Peirides Foundation.
Oddy, W. A., and J. Swaddling 1985. "Illustrations of metalworking furnaces on Greek
vases," in Craddock and Hughes 1985: 43-58.
Oleson, J.P. 1984 Greek and Roman water-lifting devices: The history of a technology.
Toronto: University of Toronto Press.
Oleson, J.P. 2000. "Water Lifting," in 6. Wikander (ed.), Handbook of ancient water
technology. Leiden: Brill, 217-302.
Palmer, R. E. 1926-1927. "Notes on some ancient mine equipments and systems," Trans-
actions of the Institute of Mining and Metallurgy 36: 299-336.
Penhallurick, R. D. 1986. Tin in antiquity. London: The Institute of Metals.
Pernicka, E., T. C. Seeliger, G. A. Wagner, F. Begemann, S. Schmitt-Strecker, C. Eibner,
0. Oztunali, and I. Baranyi 1984. "Archai:imetallurgische Untersuchungen in Nord-
westanatolien," Jahrbuch der Romisch-Germanischen Zentralmuseum Mainz 31:

553-600.
Photos-Jones, E., and J. Ellis Jones 1994. "The building and industrial remains at Agrileza,
Laurion, and their contribution to the working of the site," Annual of the British School
at Athens 89: 307-58.
Ponting, M. J. 1999. "East meets West in post-classical Bet Se'an: The archaeometallurgy of
culture change," Journal of Archaeological Science 26.11: 1311-21.
Ponting, M. J., and I. Segal 1998. "Inductively coupled plasma-atomic emission spec-
troscopy analyses of Roman military copper-alloy artefacts from the excavations at
Masada, Israel," Archaeometry 40.1: 109-22.
Pyatt, F. B., G. Gilmore, J.P. Grattan, C. 0. Hunt, and S. McLaren 2000. "An Imperial
legacy? An exploration of the environmental impact of ancient mining and smelting in
southern Jordan," Journal of Archaeological Science 27: 771-78.
Raftery, M. B., J. V. S. Megaw, and V. Rigby (eds.) 1995. Sites and sights of the iron age:
Essays on fieldwork and museum research presented to Ian Mathieson Stead. Oxford:
Oxbow.
Ramage, A., and P. T. Craddock 2000. King Croesus' gold. London: British Museum Press.
Rehren, T. 1999. "Small size, large scale: Roman brass production in Germania Inferior,"
Journal of Archaeological Science 26.8: 1083-87.
Rehren, T., D. Vanhove, and H. Mussche 2002. "Ores from the washeries in the Lavriotiki,"
Metalla 9.1: 27-46.
Riddle, J. M. 1985. Dioscorides on pharmacy and medicine. Austin: University of Texas.
Rothenberg, B. 1972. Timna. London: Thames and Hudson.
Rothenberg, B., F. G. Palomero, and H.-G. Bachmann 1989. "The Rio Tinto Enigma," in
Domerque 1989a: 57-70.
Rinuy, A., and F. Schwiezer (eds.) 1994. L'Oeuvre d'art sous le regard des sciences. Geneva:
Musee d'art histoire and Editions Slatkine.
Salkield, L. U. 1987. A technical history of the Rio Tinto Mines: Some notes on exploitation
from pre-Phoenician times to the 1950s. London: The Institution of Mining and Me-
tallurgy.
Sanchez-Palencia, J.-F. 1989. "La explotacion del oro en la Hispania Romana: Sus inicios y
precedents," In Domerque 1989a: II, 35-49.
Schweizer, F. 1994. "Etudes techniques de laitons byzantins," in Rinuy and Schwiezer 1994:
175-86.
Schweizer, F., and J. Bujard 1994. "Aspect Metallurgie de quelques objets byzantins
et omeyyades decouverts recemment en Jordanie," in Rinuy and Schwiezer 1994:
191-203.
Shepherd, R. 1992. "Hannibal the rock breaker," Minerals Industry International 1992: 39-47.
Sim, D. 1997. "Roman chain mail: Experiments to reproduce the techniques of manu-
facture," Britannia 28: 359-71.
Sim, D., 1998. Beyond the bloom. British Archaeological Reports, Intl. Series S725. Oxford:
BAR.
Sim, D., and I. Ridge 2002. Iron for the eagles. Stroud: Tempus.
Smith, C. S., and M. T. Gnudi (trans. and eds.) 1942. The Pyrotechnia of Vannoccio Bir-
inguccio. Chicago: Chicago University Press.
Tylecote, R. F. 1981. "From pot bellows to tuyeres," Levant 13: 107-18.
Tylecote, R. F. 1987 The early history of metallurgy in Europe. London: Longman.
Tylecote, R. F., M. S. Balmuth, and R. Massoli-Novelli 1983. "Copper and bronze metal-
lurgy in Sardinia," Journal of the Historical Metallurgy Society 17.1: 63-78.
Wagner, G. A., and G. Weisgerber (eds.) 1988. Antike Edel- und Buntmetallgewinnung auf
Thasos. Der Anschnitt Monograph 6. Bochum: Bergbau Museum.
Wagner, G., and 6. Oztunali 2000. "Prehistoric copper sources in Turkey," in -0. Yal<rin
(ed.), Anatolian MetaVol. 1. Der Anschnitt Monograph 11. Bochum: Bergbau Museum,
31-66.
Wahl, J. 1993. "Tres Mifias-Vorbericht iiber die archaologischen Ausgragungen in Bereich
des Romischen Goldwerks 1986-7," in H. Steuer and U. Zimmermann (eds.), Mon-
tanarchiiologie in Europa. Sigmaringen: Jan Torbecke, 123-52.
Wahl, J. 1998. "Aspectos technol6gicos da industria minera e metalurgica Romana de Tres
Minas e Campo de Jales," Actas do Seminario Arquelogia E Museologia Minerias.
Lisboa: Museu Geol6gico, 57-68.
Walsh, R. 1929. "Galen visits the Dead Sea and the copper mines of Cyprus (166 AD),"
Bulletin of the Geographical Society of Philadelphia 25: 92-110.
Weisgerber, G. 1979. "Das romisch Wasserheberad aus Rio Tinto in Spanien im British
Museum London," Der Anschnitt 2-3: 52-79.
Weisgerber, G., and C. Roden 1985. "Romische Schmiedeszenen und ihre Geblase," Der
Anschnitt 37.1: 2-21.
Weisgerber, G., and C. Roden, C. 1986. "Griechische Metallhandwerker und ihre Geblase,"
Der Anschnitt 38.1: 2-26.
Weisgerber, G., and L. Willies 2000. "The use of fire in prehistoric and ancient mining:
Firesetting," Paleorient 26.2: 131-49.
Welter, J-M. 2003. "The zinc content of brass," Techne 18: 27-36.
Williams, A., 1977. "Roman arms and armour: A technical note," Journal of Archaeological
Science 1.4: 77-87.
Willies, L. 1991. "Ancient copper mining at Wadi Amran, Israel," Mining History, special
issue of Bulletin of the Peak District Mines Historical Society 11.3: 109-38.
Willies, L. 1994. "Firesetting technology," in T. D. Ford and L. Willies (eds.), Mining before
powder, special issue of Bulletin of the Peak District Mines Historical Society 12.3: 1-8.
Willies, L. 1997. "Roman mining at Rio Tinto, Huelva, Spain, and appendix: The 1981
archaeological survey," Mining history: Bulletin of the Peak District Mines Historical
Society 13.3: 1-30.
92: 1-32.
Wollmann, V. 1976. "Romischer Goldbergbau in Alburnus Major (Ro§ia Montana) in
Drakien, Der Anschnitt 6: 182-91.
CHAPTER 5
QUARRYING AND
STONEWORKING
J. CLAYTON FANT
THE impulse toward monumental religious architecture in Greece during the ar-
chaic period and the consequent shift from building in wood to stone construction
made quarrying and stoneworking activities central to communal identity.
Nevertheless, down to the Hellenistic period quarrying remained relatively small in
scale, static in technology, and local in reach. The notable exceptions, such as Mt.
Pentelicon near Athens, and the quarries of fine statuary marble on the island of
Paras, also flag the growing preference in this period for marble over the abundant
supplies of local limestone (conventionally called poros). These centers of activity
foreshadow the larger quarrying ventures of operations of the Hellenistic period
and the enormous quarries and empire-wide distribution of the Roman epoch.
GREEK QUARRYING AND

STONEWORKING
The origins of Greek quarrying are not entirely clear, but by 500 B.C. the techniques
were already well established. There seems to have been no carryover from the
Bronze Age across two centuries of deep cultural discontinuity. The common
assumption has been that the source was Egypt (Ward-Perkins 1971a: 143, 1971b: 528;
Boardman 1976: 19, 77), where quarrying goes back into the third millennium B.C.
Although the basic technique of freeing a block by trenching around it does seem to
have been pioneered in Egypt, the absence of iron wedges and quarry picks before
about 600 B.C. calls this assumption into question (Waelkens 1992; Klemm and
Klemm 1981: 36). Important new work points to connections with the Middle East
(Waelkens et al. 1988, 1990), particularly the Hittite and neo-Hittite tradition of the
Levant. This suggestion gains support from a modem sculptor's observation that
early Greek carving techniques owe more to Hittite work than to their presumed
Egyptian models (Rockwell 1993: 198-99 ). In fact, quarrying practices were so refined
by the end of the sixth century B.C. that there was little subsequent change, at least
down to the early Roman Empire. The result is that quarry faces can rarely be dated
without related evidence, such as ceramics, graffiti, or obvious association with a
datable building. Roman innovations were few and subtle, so ancient quarrying
retained its character virtually from beginning to end, and evolved with few obvious
changes into medieval quarrying.
Trenching Technique
The basic method of separating a block oflimestone or marble from the parent rock
was common to all quarrying areas and all periods of ancient extractive technology.
Variations in the means of freeing a block once it was isolated above and on four
sides were considerable, creating the appearance of greater divergence in funda-
mental practice than was actually the case. Moreover, there was no basic difference
in technique between marble and limestone (some varieties of which are as hard as
marble), or for that matter other relatively soft stones. Distinct techniques, however,
were necessitated by the harder igneous stones, of which there were few in the Greek
world.
To free a block, a column drum (basically a block), or a monolithic column (in
the Roman period), the quarryman used a quarry pick, shaped like a modem
sledgehammer but with a longer handle and with a lighter iron head. One or both
ends of the iron implement were brought to a dull point. The motion of the
quarryman was to raise the pick above his head and bring it down to a point of
impact between his legs, so the worker faces forward and the movement of his body
is an easy bend at the waist. Twisting motions such as are seen today when using an
axe, scythe, or leaf rake, were avoided, reducing stress on the back and making it
more feasible to work long hours without injury. The pick used for most of classical
antiquity had a long handle and a sharp point (or two points) on the iron head. The
impact crushed the crystals in a small area, allowing precise work with a minimum of
waste. Because the point of the pick did not penetrate very deeply at each stroke, the
quarryman had to edge back constantly. The effect, in skilled hands, was to create a
very shallow trench in a straight line, leaving a tool mark on the quarry wall resem-
bling that created by a point chisel driven across a flat surface. The quarryman then
moved laterally to widen the trench, until finally a separation trench about 40--80 cm
wide was created to the desired depth (figure 5-1). Since the iron pick head was
narrow, it was possible to create a vertical face with little deviation. Quarry surfaces
QUARRYING AND STONEWORKING 123
Figure 5,1. Quarry trenches at Docimium. (Photograph by J.C. Fant.)
show traces of this light pick as early as the late sixth century, as graffiti at Belevi near
Ephesus reveal, in the extensive quarries opened for construction of the Artemisium
(figure 5.2).
Once a block was isolated, it was separated by splitting it free from the bedrock
base. From the beginning of Greek quarrying, the standard tool was the iron wedge,
but there is greater variation in this specific technique than in trenching around
blocks. Wooden wedges, which are recognizable by the large dimensions of the holes
cut for them ( up to 60 x 60 cm) and their wide spacing, are very rare and perhaps
restricted to deposits of stone that could be easily split (e.g., Doliana: Waelkens et al.
1990: 65). Holes chiseled for iron wedges were much smaller and somewhat more
closely spaced. Sometimes they were connected by a more irregular cavity that
permitted hammering directly on the wedges in place. This system is evident already
at the quarry for the colossus at Apollonas on Naxos, the archaic-period quarry at
Phanari on Thasos, and the sixth-century quarry for the tumulus at Belevi near
Ephesus. Splitting the block off was one of the riskiest moments in quarrying, and
several means were devised to provide better control of the direction and angle of the
break. Runs of heavy chisel lines connecting the wedge holes in a pointilli technique,
as in the Roman quarries of Aliki on Thasos, can be seen as early as the sixth century
at Apollonas on Naxos (Waelkens et al. 1990: 62-65). Another method was to place
the wedge holes within a continuous chisel-cut groove, which would weaken the
stone along the intended break line prior to hammering on the wedges themselves.
Highly fissile stone such as the marble of Carystus in southern Euboea was some-
times split free simply by the pressure of a pry bar, underlining the fact that the key
factor was the nature of the stone deposits, rather than a chronological development
of technique documented by, for instance, size or shape of wedge holes (Roeder 1965:
518-34; Nylander 1968: 6-10).
Figure 5.2. Quarry face with marks of the light pick, sixth century B.C. Belevi, near
Ephesus. (Photograph by J. C. Fant.)
Most Greek quarrying, being tied to specific construction projects, was episodic
and small. Quarrymen sought deposits easily accessible without extensive prelim-
inary work, an investment that would be repaid only over a long periods of pro-
duction. Many Greek quarries thus give the appearance of small-scale and shallow
workings, where only one or two layers of trenches were cut. Larger projects pro-
ducing greater quantities of debris would quickly choke such flat quarry beds, and
so quarry schemes cutting into slopes, which facilitated removal both of debris and
finished blocks, were often preferred. The great vein of fine marble on the western
side of Mt. Pentelicon, which provided material for the Periclean Acropolis and a
variety of structures down to the stadium of Herodes Atticus in the mid-second
century A.O., is a spectacular example (Korres 1995). Few early quarries were as
large as Belevi, with its 35 m vertical faces, since few architectural projects were on
the scale of the Artemisium. Nor was all quarrying purely for local uses. As early as
the late sixth century B.C., quarries on Paros were in a position to export substantial
quantities of marble to help Cleisthenes win over the Delphians (Herodotus 5.62.3).
Transport
Once a block was freed, the next challenge was transport; this obvious and difficult
requirement explains why most quarries in the premodern era served very local
needs. Since quarries were often sited on ridgelines-because prospecting was easy
and removal of the overburden greatly simplified-gravity, used with caution,
provided an advantage. Runways, often bordered with heavy bollards to control the
paying out of lines attached to sledges or directly to the blocks, were probably
developed as early as quarrying itself. Mt. Pentelicon near Athens provides early
examples. Probably the most spectacular of these lizza paths (the common term
comes from Carrara), however, date to the Roman period, as at Mons Porphyrites
in the eastern desert of Egypt. Intra-quarry transport was by ox or mule cart;
extended overland carriage was obviously avoided, but evidence from Egypt shows
that carts with very wide axles (2.22 m and 2.66 m), possibly drawn by camels,
carried the products of the numerous quarries of the Eastern Desert down to the
Nile (Peacock and Maxfield 1997: 261) (cf.figure 23.10 ). More desirable was a seaside
location, such as the quarries ofVathy, Phanari, and Aliki on the island of Thasos,
or proximity to a harbor like that of Saraylar, which served the ancient quarries of
Proconnesus. It was no accident that these two sources of white marble were among
the earliest to export; Proconnesian marble was used for the Mausoleum at Hali-
camassus (Pliny, HN 36.47). Cranes using a fixed mast and mobile boom studded
harbors and waterfronts (Vitruvius, De arch. 10.2.10; Sodini et al. 1980). Seaborne
transport of stone, blocks, and monolithic columns was a significant proportion of
trade in the Mediterranean during the Roman Empire, judging by the represen-
tation of such cargoes among recorded shipwrecks (Parker 1992). In the late Hel-
lenistic period, finished sculptural and architectural objects predominate (for ex-
ample, the cargo of the Mahdia wreck: Hellenkemper Salies 1994); during the early
empire, however, cargoes consist of blocks and monolithic columns, and in the later
empire prefabricated church kits appear (Kapitan 1961).
Shaping of quarry products, whether for sculpture, architecture, or the fabri-
cation of objects such as basins or sarcophagi, followed a common procedure in
which tools of tempered iron were essential. Preliminary shaping could still be done
with the quarry pick or with a chisel with a heavy point. The point chisel was the
chief tool for all rough shaping work in architecture as well as sculpture. With the
chisel held at about 70 degrees, large volumes of stone can be removed quickly. At 45
degrees, strokes producing splits with less propagation are used to approach the final
surface (Rockwell 1993: 39). Fine dressing, however, was the task of the flat chisel; a
flat chisel with teeth, essentially a row of very small flat chisels, was apparently a
Roman innovation. Very fine shaping was done with abrasive tools, rasps, files, and
at the end, emery and pumice if a reflective surface was wanted (rare until Antonine
times). The last category of tool was the drill, of no great significance until horizontal
motion was introduced; then the deep continuous channels of the running drill
become characteristic of the chiaroscuro effects of Roman sculpture from the second
century A.D. on (figure 2.7; Strong and Claridge 1976; Rockwell 1993: 37-38).
ROMAN QUARRYING
During the Roman period the principal development was the growth of a central
organization for fine stone supplies. This led in tum to some technical innovations,
but more important was the exponential rise in scale of stone quarrying and stone
use. Under the emperors, new lithic resources from Egypt were explored and
exploited more extensively than they had been even during the Middle Kingdom. A
new taste for colored stone spurred the opening of new quarries, and the volume of
building projects in Rome from Augustus (31 B.C.-A.D. 14) onward, mirrored over
time by prosperity in the provinces, made marble the dominant material for
building and decoration.
Whereas research on Greek quarrying has generally been piecemeal, with
studies often linked to investigation of entire sites, research on the Roman marble
trade has become a subject in its own right since its launch in the essays of John
Ward-Perkins and the quarry surveys ofJosefRoeder in the 1960s and 1970s. Roeder
directed our attention to the neglected sources of well-known Roman lithic ma-
terials, while Ward-Perkins sketched a picture of an empire-wide system with
particular units finding specialized niches, a process he repeatedly characterized in
the language of capitalism.
Using criteria like workshop identification in architectural carving and sar-
cophagus production, Ward-Perkins formulated a list of factors that he highlighted
as fundamental to the marble trade under the empire. Rationalization, standardi-
zation of practices-including dimensions of quarry products, use of prefabrica-
tion, especially in sarcophagus production, stockpiling rather than production for
individual projects, and centralization under imperial ownership and administra-
tion. Although the mercantile mentality this model assumes must be modified,
Ward-Perkins' work stimulated a generation of new research and, most valuably, it
challenged art historians to look beyond the hand of individual artists to the nature
of the raw material and the production processes.
CENTRALIZED MARBLE TRADE

IN THE ROMAN EMPIRE
By the second century A.D., the Roman emperors owned most of the major sources
of fine stone. But the idea of a centralized trade embraces more than the actual
imperial organization, about whose nature there are still areas of doubt. The
prestige of imperial buildings as tastemakers, and the scale of imperial production
of marble caused the rest of the marble trade-other than the purely local-to orbit
around it.
Both the idea of using marble and the material itself were imports to Rome in
the second century B.C. Proconsuls returning from victorious and lucrative cam-
paigns in the eastern Mediterranean found that a dramatic way to celebrate their
success was to build a pious monument with imported marble. In 190 B.C., Scipio
Africanus and his brother built an arch (Livy 37.3.7) that may have been in marble.
The phenomenon expanded in the second half of the century with manubial
temples (built from spoils of war), most of them designed and built by Hermodorus
of Salamis (Gros 1976). Where we know it, the source of the marble, and certainly of
the masons schooled in working it, was Mt. Pentelicon in Attica. The reference to
Athens suited the programmatic aspect of these dedications well. One of their
purposes was to advertise the culture as well as the wealth of the donors, and ever
since the treaty with the Aetolians of 212 B.C. the Romans had deferred to Athens as
the cultural center of Greece. Marble was also imported as columns for the tem-
porary wooden theaters that aediles (junior magistrates eager to gain public ap-
proval) constructed for public festivals; these had the additional advantage that it
could be resold when the sets were broken down, or even moved into the private
house of one of the sponsors (Pliny, HN 17.6, 36.7). Polychrome marble began to
appear in Rome in the next century, associated in anecdote with leading political
figures-orange Numidian used as a house threshold by M. Aemilius Lepidus,
consul in 78 (Pliny, HN 36.49), red and black Lucullan imported by its namesake,
M. Licinius Lucullus-while others became familiar at first in Italy as carved legs for
atrium tables or wall cladding. During their propaganda war in the 50s B.C., Caesar
responded to Pompey's construction of the spectacular Theatrum Pompeii by
preparing for his own building campaign, opening Italy's first white marble quar-
ries, near Luna in Liguria, near modern Cararra (Strabo 5.2.5; Fant 1989). Although
Caesar did not live to see the result of his investment in infrastructure, his political
heir took full advantage of it and demonstrated that he understood the value of
marble displays as a weapon in political campaigning.
One component of the regime that Augustus set about establishing after a
century of intermittent civil wars was a massive building campaign. Infrastruc-
ture projects and smaller public buildings were delegated to Agrippa or other
members of the Augustan circle, but Augustus himself took responsibility for the
most conspicuous and emblematic projects. The Forum of Augustus used both
colored marbles already well known and those imported only recently. Later ev-
idence labels all of them imperial property, and the clear inference is that the
emperor's agents both gathered up the famous stones of the late Republic and
searched out a number of new ones. Surveys in the 1990s have produced evidence
of an extensive campaign of prospecting and development at Mons Porphyrites
in the Eastern Desert of Egypt, where one Cominius Leugas, position not named
but beyond doubt an imperial agent, congratulated himself in A.D. 18 on having
discovered porphyry, black porphyry (knekites), and other colored stones. Given
the logistical difficulty of working in the Eastern Desert and the number of finds
he had made, the inference is inescapable that the effort had begun well be-
fore the demise of Augustus in A.D. 14 (Peacock and Maxfield 1997: 23; Fant 1999:
279-80).
Thus was created the imperial institution that, in the second century A.D., was
probably called the ratio marmorum, or marble bureau (CIL 6.8631, 301, 8482, 33790;
11.3199; Fant 1993: note 3). A provisional list of stone sources owned and exploited
by the Bureau includes the following (the traditional Italian name follows the Latin)
(Fant 1993: 163-67):
Colored marbles (and other stones)
marmor Numidicum, giallo antico: Numidian, from Chemtou in western

Tunisia.
marmor Luculleum, africano: Lucullan, from Teos, near Smyrna/Izmir.
marmor Synnadicum or Phrygium, pavonazzetto: Phrygian, from Docimium.
marmor Lacedaemonium (or lapis Lacedaemonius), serpentino or porfido verde:
green porphyry, from Laconia.
marmor Carystium, cipollino: Carystian, from southern Euboea.
marmor Chium, portasanta: Chian.
marmor Troadense, granito violetto: Troad granite, from Kozak Dag in the
Troad.
All exported Egyptian stones, most notably lapis Thebaicus or pyrropoikilos
(Aswan granite), marmor Claudianum (granito del Poro, Mons Claudianus
granite), marmor Porphyrites (purple and black porphyry), lapis basanites
(bekhen stone from Wadi Hammamat), alabasters, and many igneous
stones quarried in limited quantities.
White marbles
marmor Lunense, marmo di Carrara: Luna, Carrara in northwestern Italy.

marmor Parium, lychnites and lesser qualities: island of Paros.
marmor Proconnesium, Proconnesian: Sea of Marmara.
Likely imperial, but poorly documented
marmor Thasium: white marble from Thasos.

marmor Scyrium, breccia di Settebassi: colored marble from Scyros.
cipollino rosso, rosso brecciato or africanone: polychrome marble of Iasos.
marmor Heracleioticum: Heraclea sub Latmus.
(no known Latin name): a coarse white marble from St. Beat in the Pyrenees.
Such well-known quarries of white marbles as Mt. Pentelicon and Hymettus,

Aphrodisias, and Belevi, and other quarries near Ephesus seem to have remained in
the hands of their cities. Stones with clear imperial associations were the most
desirable, and within that group the most prestigious were Numidian and Phry-
gian, the dominant colors in the interior of the Pantheon, then Lucullan, Chian,
Carystian, and the breccia of Scyros, in about that order. The Egyptian granites
were also highly prestigious but in such short supply that few individuals outside
the court could expect to obtain them. A partial exception was Aswan granite, the
quarries of which were accessible by water (the first cataract of the Nile is a dike of
this granite), and thin slabs of the purple and green porphyries for floor compo-
sitions (opus sectile) were also available to the well connected (Jongste et al. 1992).
Since these stones were not widely obtainable, other similar examples entered the
market as substitutes (marmi sostituivi): granite from Sardinia in place of Aswan
granite; granite from Kozak Dag near Pergamon for Mons Claudianus granite;
breccia corallina from Bithynia for breccia di Settebassi, and so forth. In this way, the
imperial marble trade indirectly controlled private as well as public markets.
Quarries in Egypt deserve special mention. New research in the Eastern Desert
has provided the most dramatic advance in quarry studies in the last two decades.
We now have a picture of an immense effort, starting with the prospecting cam-
paign mentioned above, continuing with the development of an extensive system of
roads and forts (Maxfield 1996) apparently centered at first on Mons Porphyrites,
where a large infrastructure had to be built in very difficult terrain, to an eventual
network of dozens of small quarries and several large ones like Mons Claudianus,
which was staffed (as we know from more than 7,000 ostraka recovered; Bingen et
al. 1992 and 1997) by a force of about one thousand workers and guards (Peacock
and Maxfield 1997).
Like many institutions of the Roman Empire, the imperial marble bureau
reached its height in the second century A.D., both in scope and bureaucratic
organization. But in terms of technology, both in quarrying and transport, inno-
vation was modest and elicited directly by the needs of scale and of adaptation to
the uses for which Roman quarry products were intended. And it never became a
modern entity, in the terms that Ward-Perkins used to describe it.
Innovation in the Roman Period

One technological change, first noticed by W aelkens, was the introduction of a
heavier quarry pick at the end of the first or beginning of the second century A.D.
The thicker iron point of this pick could penetrate farther at each blow; conse-
quently the quarryman could execute many blows before having to move his feet,
and this speeded trenching. The cost was greater waste and, to some extent, less
precise work. The characteristic tool marks left by this pick are tracks forming
segments of circles, rather than the nearly horizontal lines of the light pick. To
counteract the tendency of the heavy pick to stray to the outside as it cuts down-
ward, the workmen often reversed direction, leaving alternating bands of marks
resembling festoons (hence the Italian description a festoni). The heavy pick also
contributed to the typical visual impression of a Roman quarry with its huge hills of
debris (figure 5.3). The major Roman quarries were active on a large scale for two
centuries or more, so it was crucial to plan exploitation so that crews would not
block each other, and that debris would not choke the working faces. In fact, at
Docimium dated blocks in the debris show that the piles were moved around from
time to time in order to build up a working platform to attack a face at a higher level
or to expose a new working area (Fant 1989: 42-47). Central organization at the
Figure 5.3. Quarry face with debris. Bacakale, Docimium. (Photograph by J.C. Fant.)
quarries is visible also in the practice of inscribing blocks at the moment of

extraction; it was standard at most polychrome marble quarries, but products at
white marble quarries were usually anepigraphic (Fant 1993: 157-62). These in-
scriptions record the date, crews responsible for the work, and the site within the
quarry. In the first third of the second century A.D., blocks on hand were marked
and perhaps graded at irregular intervals, probably in response to requests for
details on inventory. Blocks were shipped from the quarries and to imperial
stockpiles, first at the Marmorata in Rome and later (when the Marmorata became
overcrowded) along the south bank of the canal connecting Trajan's harbor at
Portus with the Tiber (Maischberger 1997; Fant 2001).
Another Roman innovation was the taste for monolithic columns, which
developed during the first century A.D. This technique may simply have been seen
as a better way to display the colors and patterning of the new imperial polychrome
marbles, although both gray and white marble columns were rendered as mono-
liths just as early. One may even suspect that the popularity of the monolithic
column lay at least in part in the utterly unnecessary additional effort involved in
production and transport. This effect was redoubled in the case of the spectacular
40-Roman-foot monoliths of Mons Claudianus and Aswan granite supporting the
porch of the Pantheon, since any citizen of Rome could estimate the distance and
terrain over which they had been safely transported-especially since Agrippa's
map of the world could be consulted in the Porticus Vipsania nearby (Pliny, HN
3.16-17). The imperial stockyard and workshop at Portus has produced some
columns with elaborate patching, which demonstrates the astonishing effort ex-
pended to salvage monolithic columns of very expensive stone (figure 5.4).
Column drums were simply blocks that could be quarried like any other, and
then rounded. But monolithic shafts had to be isolated as such in the quarry beds
Figure 5,4. Column of Lucullan marble at Portus with numerous patches to correct or
enhance surface coloration. (Photograph by J. C. Fant.)
before being broken free, a much more chancy process. Columns in the making
and still attached to bedrock show that the quarrymen did as much rounding as
possible even before separation, and a set of unfinished columns at Proconnesus
has permitted the entire process to be reconstructed (Asgari 1992; for finishing on
site see Pensabene 1992; Wilson Jones 2000: 130-32). Other architectural elements,
(capitals and bases), as well as sarcophagus blanks, were also roughed out in the
quarries. Ward-Perkins saw in this an industrial process of prefabrication in which
continuous production to standardized modules was stockpiled as an inventory
against which architects all over the Mediterranean could confidently draw without
having to place special orders and wait for them. He pointed to the number of
columns at quarries and in finished buildings in standard lengths, especially 40
Roman feet, as evidence for this, but it has always seemed unlikely that quarry
practices would constrain an architect's choices. Subsequently Wilson Jones pro-
vided a better explanation based on the developing canon of the Roman Corin-
thian order during the Augustan age, in which a proportion of shaft length to total
column length of 5:6 tended to favor easily computed lengths in whole numbers of
Roman feet (Wilson Jones 2000: 155).
The Roman emphasis on veneer as opposed to structural blocks had both
technical and artistic results. Since Roman marble use was never structural-apart
from columns and architraves-marble was needed only as sheathing. Architects
continued to use traditional materials such as native tuff and travertine even
alongside concrete. As colored stones became available in the first century A.D., cut
stone floors (opus sectile) became popular as inserts within floors of limestone or
slate tesserae, or, when possible, as the entire paving material. Both these uses
required accurately sliced panels. The stone saw, a long iron blade with flat edges,
was mounted in a heavy frame allowing it to travel back and forth in a level plane
while a worker trickled in a slurry of sand and water to provide the cutting agent.
Pliny (HN 36.47) attributes the invention of sawn veneer to Mausolus of Caria in
the fourth century B.C., but he focuses his discussion not on the tool but on the
sand: coarse sand produces rough cuts and leaves more work for the polishers in the
next stage. If the stone saw did originate with Mausolus, it was likely a freehand saw;
the addition of a frame to direct the motion should be a Roman refinement. Roman
saw-cuts in fact produce surfaces that seem polished, and this had to be an im-
portant advantage. Saws were set up at construction sites so that veneer could be cut
from blocks and fitted directly to the building in progress. A dramatic example is at
the Temple of Venus at Pompeii, where blocks of Proconnesian marble were being
sawn when work was interrupted (Bruno et al. 2002). Stocks of veneer prepared for
a renovation project were found in the Villa of the Sulpicii near Pompeii (de
Simone and Ciro Nappo 2000: 190). Saws were not a common quarrying tool, but
there is one case of a sawn quarry face at Docimium (Roeder 1971). Ausonius
(Mosella 359-64) mentions water-powered marble saws near quarries in the fourth-
century A.D. Rhineland. The discovery of water-wheel-driven saws for sawing stone
veneer at the site ofJerash in Jordan, at the other side of the empire, now vindicates
this text (figure 6.5; Seigne 2002; Wilson 2002), along with a very clear represen-
tation of a water-wheel-driven stone saw identified on a Roman sarcophagus at
Pammukale (Ritti et al. 2007). Another innovation of uncertain date is the lathe for
turning stone columns, tabletops, and vessels. Large columns or column drums
may have been given a final finish on a slow lathe, smaller elements on a fast lathe.
Pliny (HN 36.90) mentions the production of large column drums on a lathe
(probably for the Temple of Hera on Samos): "The drums for which [150 columns]
were so well balanced in the workshop that they were turned with a boy supplying
the force" (quarum in officina turbines ita librati pependerunt ut puero circumagente
tornarentur). Both Theophrastus (De lap. 5, 41) and Pliny (HN 44.159) also allude to
the finishing of stone vessels with a lathe on the island of Siphnos. The marks of
lathe turning have been noticed on a variety of objects in Italy (Blanc and Monthel
2006; Metz 1989: 171-81).
DIOCLETIAN'S PRICE EDICT: THE

TWILIGHT OF ROMAN QUARRYING
A chapter of the emperor Diocletian's Edict on Maximum Prices lists some famous
stones, but this does not necessarily mean that they were still being quarried in the
same manner or at the same rate as earlier in the high empire. Since most of the
quarries and stones were under state control, their original price could be set
without recourse to decree. It is, however, likely that actual new supplies were
earmarked for imperial projects. The inference should be that prices were being set
for reused quantities of these stones, and a persuasive argument that the unit of this
chapter of the Edict was the linear square foot, not a cubic measure ( Corcoran and
Delaine 1994), points strongly to veneer, an easily portable and reusable format of
stone. Veneer was not a product rough-cut at the quarry (thin slabs were simply too
breakable) but one of secondary workshops. Thus the marble trade of the fourth
century was already foreshadowing that of the medieval period, when most quarries
had become inaccessible and crumbling Roman structures became the chief source
of supply.
REFERENCES
Asgari, N. 1992. "Observations on two types of quarry-items from Proconnesus: Column-

shafts and column-bases," in Waelkens et al. 1992: 73-80.
Bingen, J., A. Billow-Jacobsen, W. E. H. Cockle, H. Cuvigny, L. Rubinstein, and W. Van
Rengen 1992. Mons Claudianus, Ostraca Graeca et Latina I (O.Claud.1-190). Cairo:
Institut francais d'archeologie orientale.
Bingen, J., A. Billow-Jacobsen, W. E. H. Cockle, H. Cuvigny, L. Rubinstein, F. Kayser, and
W. Van Rengen 1997. Mons Claudianus, Ostraca Graeca et Latina II (O.Claud.
191-414). Cairo: Institut francais d'archeologie orientale.
Blanc, N., and G. Monthel 2006. "Le tournage des elements porteurs en Pierre dans
l' Antiquite." Conference abstract. La Pierre dans tous ses etats, Collogue International,
Association for the Study of Marble and Other Stones in Antiquity, Aix-en-Provence,
June 12-18, 2006.
Boardman, J. 1976. Greek sculpture: The archaic period. New York: Oxford University Press.
Bruno, M. et al. 2002. "Pompeii after the A.D. 62 earthquake: Historical, isotopic, and
petrographic studies of quarry blocks in the Temple of Venus," in J. J. Herrmann, N.
Herz, and R. Newman (eds.), Association for the Study of Marble and Other Stones in
Antiquity, V:. Interdisciplinary studies on ancient stone. London: Archetype Publica-
tions, 282-88.
Corcoran, S., and J. DeLaine 1994. "The unit measurement of marble in Diocletian's Prices
Edict," Journal of Roman Archaeology 7: 263-73.
de Simone, A., and S. Ciro Nappo (eds.) 2000. Mitis Sarni Opes. Naples: Denaro.
Fant, J. C. (ed.) 1989. Cavum antrum Phrygiae: The organization and operations of the
Roman imperial marble quarries in Phrygia. British Archaeological Reports, Intl. Series
S482. Oxford: BAR.
Fant, J. C. 1993. "Ideology, gift and trade: A distribution model for the Roman imperial
marbles," in W. V. Harris (ed.), The inscribed economy: Production and distribution in
the Roman Empire in the light of instrumentum domesticum,Journal of Roman Ar-
chaeology Suppl. 6. Ann Arbor. MI: JRA, 145-70.
Fant, J. C. 1999. "Augustus and the city of marble," in M. Schvoerer (ed.), Archeomateriaux,
marbres et autres roches. Association for the Study of Marble and Other Stones in
Antiquity, W: Actes de la iv"me conference internationale, Bordeaux France, 9-13 octobre

Talence: CRPAA, 277-80.
1995.
Fant, J.C. 2001. "Rome's marble yards," Journal of Roman Archaeology 14: 167-98.
Gros, P. 1976. Aurea Templa: Recherches sur l'architecture religieuse de Rome a l'epoque
d'Auguste. Rome: Ecole fran<raise de Rome.
Hellenkemper Salies, G. (ed.) 1994. Die Antike Schiffsfund von Mahdia. Koln: Rheinland
Verlag.
Jongste, P. F. B., J.B. Jansen, L. Moens, P. De Paepe, and M. Waelkens 1992. "The use of
marble in Latium between 70 and 150 A.D. IPCAES for determination of the prove-
nance of white marbles," in Waelkens et al. 1992: 263-67.
Kapitan, G. 1961. "Schiffsfrachten antiker Baugesteine und Architekturteile vor den Kiisten
Ostsiziliens," Klio 39: 276-318.
Klemm, R., and D. Klemm 1981. Die Steine der Pharaonen. Munich: Staatliche Museen
Aegyptischer Kunst.
Korres, M. 1995. From Pentelicon to the Parthenon. Athens: Melissa.
Maischberger, M. 1997. Marmor in Rom: Anlieferung, Lager- und Werkpliitze in der Kai-
serzeit. Wiesbaden: Reichert.
Maxfield, V. A. 1996. "The Eastern Desert forts and the army in Egypt during the prin-
cipate," in D. M. Bailey (ed.), Archaeological research in Roman Egypt. Journal of
Roman Archaeology Suppl. 19. Portsmouth, RI: JRA.
Metz, A. 1989. "Ein gedrehter Sandsteintisch aus Augst," Jahresberichte aus Augst und
Kaiseraugst 6: 171-81.
Nylander, C. 1968. "Bemerkungen zur Steinbriichgeschichte von Assuan," Archiiologischer
Anzeiger 1968: 6-10.
Parker, A. J. 1992. Ancient shipwrecks of the Mediterranean and the Roman provinces. Ox-
ford: Tempus Reparatum.
Peacock, D. P. S., and V. A. Maxfield 1995. The Roman imperial porphyry quarries: Gebel
Dokhan, Egypt. Interim Report, 1995. Southampton: University of Southampton.
Peacock, D. P. S., and V. A. Maxfield 1997. Survey and excacavation Mons Claudianus. Vol.
I, Topography and quarries. FIFAO 37. Cairo: Institut francais d'archeologie orientale.
Pensabene, P. 1992. "On the method used for dressing the columns of the Colosseum
portico," in Waelkens et al. 1992: 81-89.
Ritti, T., K. Grewe, and P. Kessner 2007. "Stridentes trahens per leviam marmora serras. A
Rockwell, P. 1993. The art of stoneworking. Cambridge: Cambridge University Press. 1993.
Roeder, J. 1965. "Zur Steinbriichgeschichte des Rosengranits von Assuan," Archiiologischer
Anzeiger 1965: 467-552.
Roeder, J. 1971. "Marmor phrygium: Die antiken Marmorbriiche von Iscehisar in Westa-
natolien," Jahrbuch des deutschen archiiologischen Insituts 86: 253-312.
Seigne, J. 2002. "A sixth-century water-powered sawmill at Jarash," Annual of the De-
partment of Antiquities of Jordan 46: 205-13.
Sodini, J.-P., A. Lambraki, and T. Kozelj 1980. Aliki: Les carrieres de marbre a l'epoque
paleocretienne. Etudes thasiennes 9. Paris: Ecole fran<raise d'Athens.
Strong, D, and A. Claridge 1976. "Marble Sculpture," in D. Strong and D. Brown (eds.),
Roman Crafts. New York: New York University Press, 195-207.
Waelkens, M. 1992. "Bronze Age quarries and quarrying techniques in the Eastern Med-
iterranean," in Waelkens et al. 1992: 5-20.
Waelkens, M., P. de Paepe, and L. Moens 1988. "Patterns of extraction and production in
the white marble quarries of the Mediterranean: History, present problems and
prospects," in Fant 1989: 81-116.
Waelkens, M., P. de Paepe, and L. Moens 1990. "The quarrying techniques of the Greek
world," in M. True and J. Podany (eds.), Marble: Art historical and scientific per-
spectives on ancient sculpture. Malibu, CA: J. Paul Getty Museum, 47-72.
Waelkens, M., N. Herz, and L. Moens (eds.) 1992. Ancient stones: Quarrying, trade and
provenance. Acta Archaeologica Lovaniensia monographia 4. Leuven: Leuven Uni-
versity Press.
Ward-Perkins, J.B. 1971a. "Quarrying in antiquity: Technology, tradition and social
change," Proceedings of the British Academy 57: 137-58.
Ward-Perkins, J. B. 1971b. "Quarries and stone-working in the early Middle Ages: The
heritage of the ancient world." Settimane di Studio del Centro Italino di Studi sull' Alto
Medievo XVIII Spoleto 2-8 Aprile 1970. Spoleto: Centro Italiano di Studi sull'Alto
Medievo, 525-44.
92: 1-32
Wilson Jones, M. 2000. Principles of Roman architecture. New Haven, CT: Yale University
Press.
CHAPTER 6
SOURCES OF ENERGY
AND EXPLOITATION
OF POWER
ORJAN WIKANDER
HUMAN history is directly related to the sources of available energy, since human
muscular strength is extremely restricted. The "degree of efficiency" of a working
man or woman amounts to less than 25 percent; that is, more than three-quarters of
the power input (food) is needed to keep the individual alive. In consequence,
during most of human history, humans have been compelled-like most animal
species-to use a considerable part of their time just to stay alive. No essential
change in this situation occurred until the Neolithic period, when the domestica-
tion of several animal species made available for exploitation the muscular power of
animals. From that time onward, the growth in available energy sources made it
possible for humans to divert a greater portion of their energy to activities other
than the gathering or production of food. There have been two periods of partic-
ularly major change: the centuries around the birth of Christ, when water power
was harnessed, and the last three centuries, when steam power, internal combustion
engines, and nuclear power were exploited and new scientific advances made it
possible to use earlier sources of energy more efficiently.
Prior to the twentieth century, almost all useful energy sources were ultimately
dependent on solar energy. The Greco-Roman cultures were well aware of the
electromagnetic energy manifested by magnetism (Radl 1988) and electricity
(Pernot 1983; Keyser 1993), but neither played any real part in the energy supply.
The use of geothermal energy was realized in limited areas, for example in pro-
SOURCES OF ENERGY, POWER 137
viding hot water to Roman baths in volcanic regions, particularly at the Bay of
Naples (Strabo 5.4.5; Dio Cassius 48.51.1-2). In practice, however, derivatives from
solar energy dominated completely. Solar radiation is converted into the com-
bustible hydrocarbons stored in wood (and, in the longer run, in coal, bitumen, and
petroleum). In addition, solar radiation nurtures human and animal musculature
indirectly through their consumption of vegetation. The evaporation of water by
the sun creates the potential energy necessary for the exploitation of water power,
and, by changes in the atmosphere, the air flow necessary for wind power.
Human muscle power is virtually negligible compared to most other forms of
available energy. According to usual calculations (e.g., Forbes 1955b: 83), a mule can
provide about five times as much energy as a working man, an ox seven, and a horse
ten (see also chapter 23). Compared to a water-mill, however, even the power of a
horse remains insignificant. These figures are, of course, approximate, but they do
show the impact the harnessing of draft animals and water power had on human
society. The common assertion that the classical cultures did not need water power
because of their abundant supply of slaves has now been decisively repudiated. Even
the most ruthless exploitation of enslaved humans would not have increased the
amount of available energy more than marginally. Slaves represent a considerable
economic investment and must be nourished even when sick and old, and the true
gain of energy remains limited. In fact, slavery is a zero-sum game that transfers
surplus human energy from one group (the slaves) to another (their owners) but in
no way alters the total amount of energy available. Furthermore, an average modern
household consumes more energy than did a Roman upper-class home employing
hundreds of slaves. In consequence, this survey will concentrate on five groups of
nonhuman energy sources: direct solar radiation, chemical energy, animal power,
water power, and wind power.
DIRECT SOLAR RADIATION
The importance of sunlight for achieving a pleasant indoor climate must have been
obvious as early as the Palaeolithic period. The Roman architect Vitruvius devotes an
entire chapter to the varying influence of climate on men and architecture depending
on their position on the earth, and therefore the angle of the sun (De arch. 6.1; cf.
Xenophon, Mem. 3.8.8-10; Pliny, Ep. 2.17.23). In the classical world, several authors
recommend the building of private houses with deep, south-facing porches, so that
they would admit the heat of the sun in winter but be bathed in shade in the summer
(Xenophon, Mem. 3.8.8-10). By observing these simple rules, two aims were attained:
cooler temperatures in summer and more effective heating in winter (Butti and
Perlin 1980). Other buildings were also designed to make good use of the sun,
particularly the Roman baths; parts of which had to be heated to high temperatures,
requiring great amounts of fuel. Vitruvius comments on the need for solar heating in
baths (De arch. 5-10.1; Ring 1996): "First of all, one must choose the warmest possible
place, that is, turned away from the north and northeast. The hot and tepid rooms
should get their light from the southwest, but if the nature of the place prevents this,
at least from the south, as the time for bathing is mostly set from noon to evening."
Sunlight has been important from time immemorial for drying various mate-
rials, for instance, washed clothes. The intentional drying of foodstuffs such as meat,
fish, or fruit was probably customary from pre-Neolithic times onward, making it
possible to save food for longer periods by preventing putrefaction. In littoral areas,
salt was produced by allowing evaporation of seawater in artificial pools.
The concentration of solar radiation in order to create very high temperatures
was known by at least the classical period. Curved glass vessels or lenses and mirrors
were used to focus sunlight to kindle fire (Aristophanes, Nub. 771-73; Theophrastus,
De igne 73; Pliny, HN 2.239, 36.199, 37.28). Nevertheless, the report found in some
late sources that Archimedes set enemy ships on fire at a distance with the help of
burning-mirrors seems impossible, given the relatively poor quality of the available
reflective surfaces (Simms 1977).
CHEMICAL ENERGY
The exploitation of chemical energy began with the taming of fire by humans about
half a million years ago (Rehder 2000). Through combustion (an oxidation pro-
cess), hydrocarbons are converted into heat and light, while emitting carbon di-
oxide and water. Cooking makes both vegetable and animal food more digestible,
wholesome, and nourishing and allowed humans a greater selection of foodstuffs.
Fire also made it possible for humans to live in temperate and subarctic climate
zones. For heating dwellings, not only did wood prove useful but also various
vegetable fibers, reeds, straw, and dried cow dung. Moreover, fire provided artificial
lighting, first by the use of burning sticks and torches, and later of lamps with
animal or, more commonly, vegetable oils. With artificial light, the dark hours of
night could be used for other productive activities (Forbes 1958: 119-93).
In the Greek and Roman world, the growing need for higher temperatures for
the production of metals, pottery, terracotta, and glass necessitated the use of more
efficient fuels. The most important of these was charcoal, created from wood by dry
distillation (pyrolysis), which liberates the wood from water through heating in a
reducing atmosphere. Charcoal is documented in Egypt at least from the early third
millennium B.C. In the classical world, it is first mentioned by Homer (Il. 9.213).
The technical procedure for transforming wood into charcoal is described in detail
by Theophrastus (Hist. pl. 4.8.5, 5.9.1-6; De igne 37.75), who also recommends
charcoal from various tree species for particular uses (cf. Pliny, HN 16.23). The
benefits of using charcoal rather than wood are not restricted to the more efficient
combustion with higher temperatures but also involve the fact that its weight is
considerably less, which makes it much easier to transport. Considering the wealth
of written and archaeological information at our disposal, it is obvious that enor-
mous amounts of charcoal were produced in the ancient world.
Heating, the firing of terracotta and pottery, and the metal industry required
increasing amounts of fuel in antiquity, a fact that inspired contemporary com-
ment. For instance, Strabo (14.6.5) attributes deforestation in Cyprus to the pro-
duction of copper and silver and notes that, during the Empire, iron ore from Elba
had to be transported over the sea to Populonia, where there were still large enough
forests to supply wood for the smelting (5.2.6). Pliny (HN 34.20.96) comments on
the similar situations causing the shortage of fuel in Campania and Gaul. More-
over, since the 1970s, growing interest in the technology of Roman iron-making has
pointed out the far-reaching destruction of vegetation that resulted from the
production of charcoal (e.g., Cleere 1976; Horne 1982).
Even though this established view of the environmental effect of wood har-
vesting has been questioned, there is little doubt that both Greece and Italy were
subjected to serious deforestation during the last half-millennium B.C. Import
from well wooded areas in the north may have lessened the shortage, but some
argue that the eventual moving of pottery and glass production first to Gaul and
later to Britain and the Germanic provinces in the first centuries A.D. was at least
partly the result of the shortage of fuel in the Mediterranean area.
There were other fuels available to the classical cultures, but neither petroleum
and bitumen nor coal were exploited more than marginally. Bitumen was found in
several parts of the Middle East but was used primarily to provide a waterproof
coating for wood and the like (Forbes 1955a: 1-120; Seguin 1938). The rich potential
of petrochemicals for heating remained almost unexploited, although a late Byz-
antine author states that Septimius Severus constructed a bath in Byzantium (later
Constantinople) fueled with bitumen or petroleum (Humphrey et al. 1998: 43).
Coal was known in many parts of Europe, but in Mediterranean areas mostly in the
form of lignite, which is of much inferior quality. Nevertheless, it was used for
metalworking in Hellenistic Greece and Liguria (Theophrastus, Lap. 16; Forbes
1958: 25-26). Mineral coal of a higher quality was extracted and used with some
frequency in Britain, where it is known from a great number of archaeological sites.
The finds are particularly rich at military forts along the Hadrianic and Antonine
Walls, but coal was used extensively also in villas and towns in central and western
England (Webster 1955). The exploitation of British coal started as early as the first
century A.D. and was spread over large parts of Britain in the second. It was mostly
used for domestic heating, particularly in hypocaust furnaces, and in iron forges,
both military and civil (Dearne and Branigan 1994).
ANIMAL POWER
Although human muscular power is very limited in potential compared to most

other available energy sources, human energy nevertheless provided two essential
advantages over the others. First of all, mobility, particularly the flexibility of our
hands, made it possible for humans to carry out actions that required a complex
series of movements; second, humans could be trained by oral or written in-
struction to execute complex tasks independently.
In their natural form, animal, water, and wind power can be used only to
perform horizontal, linear work, while most human activities are much more
complicated. Today, industrial robots can imitate almost any human movement,
but it was not until the third century B.C. that the first important steps were taken
in this direction. The invention, on the one hand, of devices that could convert a
movement from horizontal to vertical and from linear to rotary (or the other way
around), and, on the other hand, of machines that could adapt these movements to
various activities, created entirely new potentials for harnessing other sources of
energy than human muscles (see chapter 13).
It seems natural that this development first took place to provide for the
needs of agricultural production. The intense urbanization process discernible
in areas around the eastern Mediterranean from the mid-fourth century B.C.
onward, with great numbers of newly-founded cities and rapidly expanding old
ones, made heavy demands on the agricultural sector. Above all, the search for
new power sources involved the two most energy-consuming agricultural activi-
ties: the lifting of water to irrigate otherwise uncultivable land and the grinding of
flour.
In the technology of water-lifting, an earlier invention had made it possible to
transfer at least some work to animals: the so-called cerd, in which a draft animal
pulls a rope running over a vertical pulley wheel mounted over a well. By walking
away from the well, the animal lifts a bucket with water. But since the animal has to
return to the starting point in order to repeat its effort, the operation is rather
inefficient and requires human supervision; in consequence it did not offer a
sufficient solution to the problem. The cerd is known in Persia from at least the
early fourth century B.C., but it is nowhere attested in the classical world (Oleson
1984: 39-41, 387-88). The technological boom in the third century B.C. gave rise to
far better solutions: more efficient use of human power with the water-screw and of
animal power with the saqiya, and above all the transmission of effort to water
power with the noria. Even the grinding of grain could now be effected with the
help of animals by means of the hourglass-shaped Pompeian mill, or by water
power.
WATER POWER
At a very early stage, humans must have become aware of the possibilities offered by
the potential energy of water. Throwing a wooden stick into running water clearly
indicates that even bigger and heavier pieces of wood may be moved downstream
with no more than a minimum muscular effort. In fact, the driving of timber is
known from various parts of the earth from very early times-in eighth-century
B.C. Assyria, for example, this activity is attested in a great number of documents
(e.g., Lanfranchi and Parpola 1990: nos. 4, 6, 25, 33,111,127, fig. 2). By replacing the
wooden logs with rafts, barges, or boats that could drift downstream and be rowed,
sailed, or pulled upstream, even stone and other heavy goods could easily be
conveyed on navigable streams. An early example is the transport of granite from
Upper Egypt, documented from at least the time of the Old Kingdom in the third
millennium B.C. Later examples are the conveyance of enormous amounts of brick
and tile to the rapidly expanding city of Rome in the beginning of the empire from
brickworks located far up the Tiber (Lancaster 2005: 17) and the importation of vast
quantities of marble from the eastern Mediterranean to Rome during the high
empire (cf. chapter 5).
Exploitation of water power for more diversified purposes than transport
needed a particular device: the paddle-driven water-lifting wheels, first mentioned
in the late third century B.C. by Philo of Byzantium (Pneum. 61, 65), but probably
invented in Egypt almost a century earlier (see chapter 13). The next step in the
utilization of water power, the vertical-wheeled water-mill, deserves a special place
in the history of technology. As a true prime mover, it contributed much to the late
Roman economy and later constituted the basis of the industrial breakthrough of
the Middle Ages and, thus, of the modern, economic development in general.
It has never been seriously questioned that the water-mill was invented in clas-
sical antiquity. Historians of technology from Beckmann (1788: 1-68) to Bliimner
(1912: 45-49) took it as confirmation of the advanced technological standard of that
period. From 1935 to about 1975, however, research was hampered by a group of
scholars who forcibly pursued the thesis that water-powered mills did not play any
important part until the fifth or even sixth century A.D. (esp. Bloch 1935, Finley
1965). Since the 1980s, however, the pace of water-mill studies has been fast. More
archaeological finds of pre-medieval water-mills have been published than the total
number known before, and our appreciation of the importance of water power in
the Roman empire has steadily deepened. Various summaries of the ancient de-
velopment have appeared during this period, but they have all soon come into
conflict with new consensus (see particularly Wikander 1979; Smith 1983-1984;
Oleson 1984: 370-80; Wikander 1985, 2000; Wilson 1995; Lewis 1997; for a critical
review of the research, see Wikander 2004; Lucas 2006). The presentation that
follows here may be better founded than its forerunners, but it, too, should be
taken for what it is: a working hypothesis.
The invention of the water-powered mill has, for a long time, been dated
somewhere between the invention of the right-angle gear (now generally dated
about 270 B.C.) on the one hand, and the first three unambiguous literary mentions
of the device (all to be dated around the end of the first century B.C.) on the other.
Considering the fact that the three authors describe the water-mill as brand new,
uncommon, and more or less unique, respectively, it is only natural that most
scholars have placed the invention in the last century B.C., or possibly slightly earlier.
It is also worthy of note that all three texts seem to locate the mills in the eastern
Mediterranean world, that is, in Greece or Asia Minor (Antipater of Thessalonica,
Anth. Pal. 9.418; Vitruvius, De arch. 10.5.2; Strabo 12.3.30; Wikander 1992: 10-11).
In 1997, however, a totally new theory was presented, assigning the invention of
the vertical-wheeled mill to Alexandria in the 240s B.C. and that of the horizontal-
wheeled mill to Byzantium ten or twenty years earlier (Lewis 1997: 13-73, esp. 56-
61). The argument is based on close scrutiny of various technical treatises origi-
nating from the third century B.C. Some dubious points notwithstanding, the new
date has been accepted by a number of scholars, particularly because it locates the
invention in the dynamic, scientific environment of the third century B.C., where it
logically belongs.
If we accept the new, early date of invention, we must also recognize that it is
more than two hundred years earlier than the next evidence we find of the existence
of water-mills, in the three Augustan authors referred to above. This time gap,
together with the mentions of the machine as new after it, needs an explanation. It is
easiest to blame the gap on the problems necessarily involved in transforming an
elegant, scientific model to a full-scale operating machine with a wooden right-
angle gear. There are, in fact, various dubious points in the three Augustan texts
that may indicate initial problems in adapting the invention to real life (Wikander
1992: 12-13; Wikander 2004: 109-11).
After the reign of Augustus, reliable literary testimonia for water-mills are
mysteriously absent for three hundred years, but some inscriptions and a growing
number of archaeological finds conveniently fill out the record. It should be noted
that even the earliest two mills attested by archaeological remains, the Augustan
mill at St. Doulchard "Les Avrillages" (Cher) in France and at Avenches in Swit-
zerland (the latter dated by dendrochronology to A.D. 57/58; Castella et al. 1994;
Champagne et al. 1997), had a mill-channel, millstones of ordinary shape, and
presumably a right-angle gear that increased the velocity of the stones. In other
words, they did not differ markedly from vertical-wheeled mills from early modern
times. The alleged early-first-century A.D. mill at S. Giovanni di Ruoti is apparently
not a water-mill at all (Greene 1999; Wilson 2002: 11 n. 56.)
During the second and third centuries, the number of archaeologically known
water-mills increases, and the epigraphical evidence for collegia of water-millers at
such disparate localities as Hierapolis in Phrygia (Pleket 1988: 27-28) and Giinzburg
in Bavaria (CIL 3.5866) proves beyond doubt that the water-mill had already begun
to play so central a part in the Roman economy that the empire would have met
with serious problems without it.
In the Price Edict of Diocletian (A.D. 301), significantly, the water-mill is

mentioned beside the other main mill types: the horse-, donkey- and hand-mill
(15.54; cf. Lauffer 1971: 257). In the fourth and fifth centuries, water-mills are in
evidence over most of the Roman Empire, from Britain in the northwest to Pa-
lestine in the east and Tunisia in the south. But, as far as we know today, the device
did not pass beyond the imperial frontiers until the sixth century A.D. India may
be an exception, if the story in Cedrenus' Historiarum Compendium (516, p. 295
Bekker), of a certain Metrodorus, who constructed baths and water-mills in India
after A.D. 325, is based, as is generally believed, on Ammianus Marcellinus.
The water-mill needs running water, but in practice this is not enough: it needs
a more or less constant water supply to make the water-wheel efficient, and it must
be feasible to tum off the water completely for inspection and repairs. In other
words, it is not enough just to immerse the wheel into a natural stream in the way
actually suggested by Vitruvius (see below). The difficulties of solving this problem
may have contributed to the apparent delay in the breakthrough of the water-
powered mill. The most simple solution is leading a mill-channel from a nearby
stream; if this is accomplished at a fall or rapids, it also offers the possibility of
creating a substantial water-head. In some cases, channels have been dug for several
kilometers in order to transfer the potential energy to the preferred location; for
instance, at Martres-de-Veyre, France (late second century A.D.), the water was
probably derived from a river about 2.5 km away (Romeuf 2001: 25).
After this realization, it was a natural step to power the mills with water from
ordinary aqueducts as, for instance, the Janiculum mills in Rome (third century
A.D.; Bell 1994; Wilson 2000; Wikander 2002), or from aqueducts constructed for
this very purpose as, for instance, at Barbegal in southern France (early second
century A.D.; Benoit 1940; Leveau 1996). Wikander (1991) provides a general in-
troduction to the relationship between water-mills and aqueducts. By means of
sluice-gates, the amount of water admitted to the wheel could be varied, and when
necessary completely cut off.
When the slope of a natural stream was too shallow, the construction of a weir
across it could raise the water level of the millrace enough to power the wheel. Such
a weir was constructed as early as the third century A.D. at the Haltwhistle Bum mill
in northern England (Simpson and Wilson 1976: 33-35, pl. I), and a detailed de-
scription of the construction of a weir around A.D. 450 is preserved in the Vita S.
Romani abbatis (18 Krusch = 57 Martine). The ultimate solution, the building of a
dam creating a true millpond, is not known with certainty until the sixth century,
but the technique was probably in use even earlier. The millpond had a double
function: not only did it create a substantial water-head, but it also made it possible
to save the potential energy of the water for periods when the water supply became
scarce.
Two more peculiar methods of securing the necessary amount of running water
may also belong to the early medieval period rather than classical antiquity: floating
mills, in which the machinery was placed on small boats anchored in rivers that
turned their paddles, and tide-mills using inflowing tidewater along the Atlantic
coasts. The former is first attested in Rome in A.D. 537 (Procopius, Goth. 1.19.19-22;
Wikander 1979: 29-32; Lohrmann 1992), the latter at Strangford Lough and Little
Island in Ireland, around A.D. 620 and 630, respectively (Rynne 1992).
During the Roman Empire, a series of water-mill types evolved suited for
different volumes of water supply and different water-heads. All water-powered
mills may be divided into two main types: the vertical-wheeled ("Vitruvian") mill
and the horizontal-wheeled ("Greek" or "Norse") mill. Both use a water-wheel to
convert the linear power of the running (or falling) water to a circular movement.
In the horizontal-wheeled mill, this movement is transferred directly (by means of
a vertical shaft or spindle) to the rotating, upper millstone (the runner), while, in
the vertical-wheeled mill, the vertical, circular movement must be converted (by a
right-angle gear) to a horizontal one.
There are three functional variants of the vertical-wheeled mill (figure 6.1). The
simplest (and certainly earliest) is the undershot mill, described in detail by Vi-
truvius between 30 and 10 B.C. (De arch. 10.4.1, 5.1-2):
(5-1) Even in rivers, there are wheels of the same construction as described
above. Wings [i.e., paddles] are attached around the periphery which, when they
are hit by the impact of the river, force the wheels to rotate by their forward
movement.
(4.1) The shaft, manufactured on a lathe or with compasses, its ends shod
with sheet-iron ... is placed on wooden blocks that have iron sheets on them
beneath the ends of the shaft.
(5.2) Water-grinders, too, are turned in the same manner. In them everything
is the same, except that a toothed drum is fixed to one end of the shaft. This drum,
set vertically on its edge, is turned together with the water-wheel. Next to this
drum, there is a larger, horizontal one, also toothed, with which it meshes. Thus,
the teeth of the wheel fixed to the shaft, by driving the teeth of the horizontal
drum, necessarily bring about the revolution of the millstones. A hopper hanging
in this machine supplies grain to the millstones.
This is also the type of mill used in the earliest archaeological finds, those at St.
Doulchard "Les Avrillages" and Avenches (both constructed before A.D. 60;
Castella et al. 1994).
The overshot mill needs a water-head higher than the wheel, so that the water
can be led above it on a wooden chute and move the wheel by both weight and
impulse. Mills of this subtype were apparently less common in antiquity than the
undershot ones, but they are in evidence just as early, in an epigram by Antipater of
Thessalonica dated to between 20 B.C. and A.D. 10 (Anth. Pal. 9.418.4-6). The
breastshot mill represents a middle course, being powered by water that hits the
back of the wheel. This variation was never common, and I know of only one
obvious example from antiquity, at Martres-de-Veyre, France, from the second
century A.D. (Romeuf 2001: 25, pls. 5-7).
The horizontal-wheeled mill, too, has functional variants (figure 6.2). The
simplest one, in which the water falls through a steep chute on the oblique vanes of
the wheel, is not known with certainty until the early seventh century A.D., in
Figure 6.1. Reconstruction of the double vertical-wheeled mill in the Baths of Caracalla,
Rome. 1. Wooden chute. 2. Overshot water-wheel. 3. Wheel-pit. 4. Paddles (vanes).
5. Buckets. 6. Shroud. 7. Sole. 8. Spokes. 9. Wheel-shaft. 10. Vertical cog-wheel (pit-wheel).
n. Bridge-tree. 12. Lantern pinion 13. Vertical shaft (iron spindle). 14. Lower stone
(bedstone). 15. Upper stone (runner). (Drawing: 6 . Wikander, after Schi0ler and
Wikander 1983: fig. 15.)
Ireland (Rynne 1992: 58-66, figs. 3, 6). A similar mill, combined with a tower filled
with water (functioning more or less as a millpond), has been used from an early
period in Algeria, Palestine, Jordan, Arabia, Iraq, and Iran. A number of these
"drop-tower" mills (or arubah penstock mills, as they were called by Avitsur [1960:
39-44]) have been dated to the third to sixth centuries A.D. (Wikander 1985: 161-
63; Wilson 1995: 507-9), but only one of them can be shown with reasonable
certainty to be earlier than the Arab conquest (a mill at the Crocodilion River,
Palestine, dated by carbon-14 analysis to A.D. 345/380, see Schi0ler 1989: 133-36,
figs. 1-3).
A third variant of the horizontal-wheeled mill is attested by two triple-helix
turbines at Chemtou and Testour in Tunisia (Rakob 1993; Roder and Roder 1993;
Wilson 1995: 500-506). This design is evidence of a highly advanced technology,
otherwise not known until the sixteenth century, but the installations are dated
without reservation by the excavators to the late third or early fourth century A.D.
They resemble the drop-tower mills in using a cylindrical pit filled with water, but
here the horizontal wheel was placed in the bottom of this pit. The rotating water
column makes this mill a true turbine.
GREEK (NORSE) Mill
''
:----------J:
'
J-7_-::)
/
14
11
1o'Q /
HELIX-TURBINE
' 2
4 -
DROP-TOWER MILL
Figure 6.2. The three variants of horizontal-wheeled mills. 1. Wooden chute. 2. Horizontal
water-wheel. 3. Paddles (vanes). 4. Bearing stone. 5. Wheel-hub/vertical shaft. 6. Iron
spindle. 7. Rynd. 8. Lower stone (bedstone). 9. Upper stone (runner). 10. Hopper. n .
Arubah penstock shaft. 12. Sluice-gate. 13. Rotating water column. 14. Bridge-tree.
(Drawing: 6. Wikander, after Curwen 1944: fig. 5; Avitsur 1960: fig. 2; Rakob 1993: fig. 1.)
In sum, of the ancient water-mills whose type can be ascertained, almost two-
thirds were undershot and one-third overshot, while only a few examples were
breastshot or with horizontal wheels.
In vertical-wheeled mills, the millrace continued into the wheel-pit, which, for
undershot mills, was narrow and conforming as closely as possible to the shape and
size of the wheel, in order to use the water more efficiently. In overshot wheels, on
the other hand, this is not necessary, and the wheels are often found in larger,
rectangular wheel-pits with lengths varying from 4.20 to 6-40 m and widths from
0.80 to 2.20 m.
The shape of the wheels themselves is fairly well known from representations
and archaeological finds-in some cases including preserved parts of the wood.
Overshot wheels differ from the undershot inasmuch as they must have soles and
shrouds (i.e., bases and side walls) that converted the compartment between two
paddles into a true bucket. Otherwise, overshot and undershot wheels apparently
differed little, at least as far as their dimensions were concerned. Obviously, ancient
mill wheels diverged in one fundamental aspect from their modern counterparts:
the wheels were constructed with real spokes (between 12 and 28 in preserved
wheels and representations) joining the periphery with the wooden hub. Most
diameters (10 out of 13) are estimated at between 1.60 and 2.56 m; two exceed 3 m,
and one allegedly attains 6 m. The widths of the water-wheels vary significantly,
from 0.22 to 1.65 m, without any apparent system-typological, chronological, or
other (for details and references, see Wikander 2000: 384-89).
The traces left by the wheel-shafts indicate that these were normally con-
structed more or less according to the description given by Vitruvius (quoted
above). Their lengths were mostly between 2.90 and 3.60 m, their thickness less
than 30 cm, tapering to between 18.5 and 21 cm at the iron-shod ends (Wikander
2000: 289-90).
The most complicated part of the power transmission was the right-angle gear
consisting of a vertical cogwheel (the "toothed drum" of Vitruvius) and a hori-
zontal one, described as "larger" by Vitruvius, but in later mills always replaced by a
small, lantern pinion. The text of Vitruvius has been questioned, but nowadays
most scholars tend to believe that it is correct. Some ancient mills have actually been
reconstructed with a larger horizontal cogwheel, although in the case of the well-
known Agora II mill at Athens (fifth century A.D.), this is probably wrong (Parsons
1936: 83; cf. critical comments by Spain 1987: 341-42, 348-51), but for the most
recently published Crocodilion River mill, the idea is at least worth considering: the
unique hourglass-shaped millstones found there may well have justified a gearing
down of the speed (see below).
The size of the pit-wheel can be estimated at four mills (all diameters being
between 0.95 and 1.11 m), that of the lantern pinion in one only: it was apparently in
use already in the third-century mill in the Baths of Caracalla in Rome, where its
width was less than 35 cm (Schi0ler and Wikander 1983: 54, fig. 11).
The vertical shaft or spindle connecting the horizontal cogwheel with the
runner stone is not mentioned by any written source until Gregory of Tours ( Vitae
patrum 18.2) around A.D. 500, but three finds of late Imperial iron spindles in
England and Germany (between 53 and 92 cm long) show that they were of the same
appearance as their medieval and later counterparts. The spindle rotated on a
bearing stone protecting the bridge-tree, and its upper end passed through the
bedstone to be fixed to the runner by an iron rynd.
The millstones are the best known part of ancient water-mills. Apart from the
many stones discovered in excavations of early mill sites, hundreds of un-
provenanced stones are found in museums all over the former Roman Empire.
Some probably derive from horse or mule-powered mills, which used a machinery
almost identical to that of the water-mills, but the majority were certainly water-
powered. Ancient millstones were much smaller than their modern successors.
Their diameter varies from 48 to 109 cm, the great majority being between 55 and 85
cm. There is a clear tendency for mills dating to the first two centuries A.D. to have
conical beds tones, while those from late antiquity are almost flat (as later stones
invariably are), perhaps because the speed of the early mills was not high enough to
make the grain pass between the stones by centrifugal force alone.
It is worthy of note that Antipater of Thessalonica wrote about the "hol-
low weights of the millstones," a fact that has led some scholars to believe that
he described hourglass-shaped stones, similar to the "Pompeian" donkey-mills
(Wikander 1992: 12-13) . This hypothesis would also help to explain why Vitruvius
wrote that the horizontal cogwheel should be larger than the pit-wheel: hourglass-
shaped stones would run a greater risk of cracking if the speed was high. But the
finds of "ordinary" stones at the early mill at Avenches made the hypothesis less
attractive.
Subsequently, however, the publication of some further water-mills at the
Crocodilion River gave reason to reconsider the question (Ad et al. 2005: 162, 169,
figs. 6-n). There, two mills powered by one common, vertical wheel obviously used
stones lower but strikingly reminiscent of hourglass mills (figure 6.3). The date of
the mills is quite late (between the fourth and the mid-seventh century), but the
excavators point out that similar stones, presumably deriving from water-powered
mills, have been found in another site in the Hula Valley, at Kibbutz Ma'ayan
Barukh. They suggest that these hourglass-stones are, in fact, "a local survival from
0!!!!!!!!!!!!!!!!!!!!iiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiii~
2 Metara
Figure 6.3. Reconstruction of the twin, vertical-wheeled mill at the Crocodilion River
(Nahal Tanninim), Israel. (Ad et al. 2005: fig. 6; courtesy of R. Frankel, University of
Haifa.)
earlier times." If so, it could very well be that the text of Antipater presents a correct
description of the very earliest, eastern water-mill stones.
Today, we may state with confidence that the breakthrough of the water-
powered mill did not take place, as maintained by Bloch (1935: esp. 545) and his
followers, in the early middle ages, but rather half a millennium earlier, in the first
century A.D., or perhaps even slightly earlier. The large numbers of water-mills that
happen to be attested in sixth and seventh-century Merovingian charters and ha-
giography do not indicate a sudden rise in the actual number of mills, but are just
the first documentation of a situation already established in the Roman Empire.
Water-mills were constructed wherever the necessary conditions were at hand: a
suitable stream and a large enough concentration of people, possibly between 200
and 400, to justify the expense (cf. Wikander 1985: 26-32). Where the concentration
of population was even greater, clusters of mills or even multi-unit mill estab-
lishments were built to supply them with ground flour.
Two such establishments have long been known, the Janiculum mills in Rome
and the mill building at Barbegal, near Arles in southern France. At Barbegal
(figure 6.4), an aqueduct supplied water to a large establishment with 16 overshot
--···· '
D
·················· ···· · ·•®
~
~-®,
1-,--· ::::::::::::~ ·., -::i:::r:r:o:::::.:..-.. :::: :; n:::r:cr::r:::r.-.

:;
E-Fil
'LI ;;
Figure 6.4. Barbegal mill complex, plan and section (Drawing: by 6. Wikander, after
Benoit 1940: fig. 3.)
wheels, constructed in the early second century A.D. (Benoit 1940; Leveau 1996:
138-49). On the slope of the Janiculum hill in Rome, large numbers of mills used
aqueduct water to provide flour to the city from the third century up to A.D. 537
(Wikander 1979: 15-29; Bell 1994; Wilson 2001; Wikander 2002).
For long, the Barbegal and Janiculum mills remained isolated exceptions, but
today there is reason to believe that such large-scale milling establishments were
not uncommon. The existence of collegia of water-millers in Hierapolis and
Giinzburg (above) and oflarge numbers of water-mills in Promona (Dalmatia, CIL
3,14969.2), and Orkistos (Phrygia, Chastagnol 1981: 406-9), during the third and
fourth centuries show convincingly that great efforts were made in towns with
enough running water to exploit this asset. Moreover "a multiple mill complex
akin to the great Barbegal mill is suspected at Colossae ... in Phrygia" (Lewis 1997:
71, with a reference to the London Times, 15 August 1991), and a passage in Am-
mianus Marcellinus (18.8.11; A.D. 359) has been shown to refer to another mill
cluster at Amida in eastern Turkey (Wilson 2001).
The economic consequences of this development are difficult to evaluate, but,
considering the enormous amounts of muscular work that could be saved by the
breakthrough of the water-powered mill, it must necessarily have been of utmost
importance for the Roman economy, particularly from the third century onward,
when the supply of labor started to grow short. This is true even though the
"breakthrough" was in no way complete: mills powered by animals or men re-
mained important up to the twentieth century. Wikander (2004: 111) provides a
discussion of the meaning of the word "breakthrough" in this context. The eco-
nomic importance of the ancient water-mill is obvious. The next question to be
answered is whether we may actually talk of a "power revolution" in the classical
world, comparable to the one earlier alleged for the Middle Ages. But here our
evidence still remains ambiguous.
Besides water-lifting and milling, ancient sources mention both marble-saws
and grain-processing pestles powered by water. The evidence of sawmills in a
tributary of the Moselle (Ausonius, Mos. 362-64) and in Cappadocia (Gregory of
Nyssa, In Ecclesiasten 3, 656A Migne), as well as the mention of a serratoria machina
(sawing machine) by Ammianus Marcellinus (23,4.4) are all to be dated around
A.D. 370-390 (Moselle: Ludwig 1981; Simms 1983; Wikander 1985. Cappadocia:
Wikander 1981: 99. Sawing machine: Rosumek 1982: 135). These testimonies in
literature have all been questioned, but the literary evidence for this labor-saving
machine is now supplemented by archaeological finds at Jerash/Gerasa, Jordan,
from the sixth century (Seigne 2002a, 2002b) and Ephesus from the seventh
(Vetters 1984: 225). In addition, a second- or third-century relief found at Pam-
mukale provides a clear representation of a water-wheel powered stone saw (Ritti
et al. 2007).
The written sources are of little assistance when trying to ascertain the tech-
nical design of the saws. Various suggested reconstructions have been rejected, but
a careful study of the remains at Jerash indicates a possible solution (figure 6.5):
Figure 6.5. Reconstruction of the water-powered saw-mill at Jarash (Gerasa), Jordan.

(Courtesy of Prof. Jacques Seigne, CNRS-Universite de Tours.)
Each saw assembly would have had four blades mounted in parallel on a stout
wooden frame at least 2.5 m long and 2.om wide, and weighing about 300 kg.
These saws must have cut vertically down through the stone, with the blades
moving back and forth horizontally, so as to take advantage of their
weight ... . The evenness of the saw-cut channels shows that the saw frame would
have been mounted within a fixed structure. (Seigne 2002a: 15)
Our knowledge of ancient water-powered pestles rests entirely on two ancient

texts: Pliny (HN 18.23.97, ca. A.D. 70) and the Vita S. Romani abbatis (17 Krusch = 52
Martine, ca. A.D. 520). Both texts involve uncertainties, but the most probable
readings favor the use of water-powered pestles-as argued convincingly by Lewis
(1997: 101-2). So far, no archaeological evidence supports the identification, but
there is little reason for doubt, as the existence of cam-operated tilt-hammers in
small-scale gadgets from Hellenistic times onward is generally acknowledged (Le-
wis 1997: 84-115; cf. chapter 31). In the Middle Ages, the tilt-hammer appears in a
whole series of water-powered machines, none of which, however, has so far been
attested with certainty in the classical world. In the field of metallurgy, a number
of excavations in Britain have yielded finds that may be connected with water-
powered trip-hammers (Lewis 1997= 106-11). At Marseille, the excavators suggest
that a water-wheel discovered in close vicinity to a pile of iron slag powered the
bellows of an iron furnace (Guery and Hallier 1987: 274, fig. 4).
Other industrial applications of water power in antiquity remain qualified
guesses-as, for instance, the attempts at detecting fulling-mills in first-century
A.D. Antioch on the Orontes (Lewis 1997: 95-99). Moreover, Vitruvius (De arch.
10.5.2) mentions water-powered dough-kneading machines, but it is very doubtful
whether the idea was ever transferred into practical use (Wikander 1981: 95-96).
Nevertheless, a mule-powered dough-kneading machine illustrated on the tomb of
Eurysaces in Rome might well have been adapted to water power (late first century
B.C.; chapter 14; figure 2.2).
In consequence, for the time being we must be cautious when speaking of
an ancient power revolution. However, the slow, but constant, growth of evidence
for exploitation of water power for varying industrial purposes in the Roman
empire is justification enough for not abandoning the thought altogether. It may
well be that we have still no idea of the true economic importance of water power
in antiquity.
WIND POWER
The most important use of wind power dates back to the fourth millennium B.C.,
when sails were adopted for driving ships. The art of sailing was gradually refined,
particularly during classical antiquity, when advances in sail rigs and ship design
made it possible to exploit wind power even when the wind came from an unfa-
vorable direction (see chapter 24). A probably much earlier use of the wind was a
necessary part of the process of threshing. Well into the twentieth century, the
wheat was sifted from the chaff by throwing the threshed grain in the air to allow
the wind carry the light chaff away.
Apart from these two areas of use, absolutely basic for ancient populations, we
have only one written document that mentions possible further uses of this source
of energy: Hero of Alexandria's description of the wind-powered organ (Pneum.
1.43). The text is a direct continuation of his preceding chapter, where Hero de-
scribes the water-organ, showing how the air pressure necessary for the organ might
instead by created by a wind-wheel which, byway of cams on the wheel-shaft, drives
a piston up and down a cylinder. It goes without saying that an organ that needs
favorable wind to be played would be of little practical use, and Hero's device is
generally dismissed as a toy or as sheer nonsense. This attitude, however, leaves two
important circumstances out of account. First, the wind-organ was admittedly
hardly meant for use, but Hero's intention was certainly not to construct a toy but
to illustrate the potentials of a physical principle. Second, the text includes a
comment that has mostly been overlooked: the wind-wheel should, according to
Hero, be provided with vanes "like the so-called anemouria." The last word is found
only here in Greek literature. It has often been translated as "windmill," but this is
hardly what Hero meant. If windmills had been known in antiquity, we would
perhaps not find much trace of the buildings and the machinery but-as windmills
power the stones from above-we should have found runners with a central hole
and bedstones without, a fact that would clearly distinguish them from stones from
horse or water-powered mills. The word anemourion is apparently a compound of
anemos and ourion, both of which mean "wind." The precise meaning remains
obscure, but a reasonable guess would he "windwhirl" or the like. Nor is it obvious
what this "windwhirl" would be: a toy for children or, perhaps, a useful machine? In
the latter case, one would rather think of a water-raising plant of the simple kind
still common in the Italian countryside. Wind power remains a dark horse in the
field of ancient energy exploitation (Drachmann 1961; Lewis 1993).
The economy of the Roman Empire declined markedly from the end of the second
century A.D. onward. Many factors in combination brought about this decline, and
this is not the place to enter deeply into the process in general-only into the
question to what extent the available energy sources may have affected the devel-
opment. On the one hand, we have good reason to believe that the central parts of
the empire suffered from a serious lack of fuel, but on the other hand, the diffusion
and breakthrough of water power must have played an important, positive role.
As I have already pointed out, deforestation in Mediterranean countries may
have been of consequence for the moving of various industries northward to Gaul,
Britain, and Germany, where the supply of wood (charcoal) and mineral coal was
far better. This migration of industries, however, provided no definite solution to
the problem, perhaps because the exploitation of energy depended not only on the
amounts available but on other factors, too, particularly the potentials for the
storage, transport, and conversion of energy. Storage did not constitute a major
problem as far as chemical and animal energy is concerned. Water power could
initially be hampered by low water levels (and even total desiccation) during
summer and autumn, but as soon as the construction of millponds became com-
mon this problem could be minimized
Transport of energy was more complicated. Draft animals could walk them-
selves to the desired locality, fuel could be transported both on sea and land,
particularly charcoal, which has a notably small weight compared to its energy
content. Here, too, water power constituted a greater problem. Water could be led
to a mill through long channels or aqueducts but, on the whole, the exploitation
remained restricted to the sites offered by nature.
The main problem was the conversion of energy and, particularly, the con-
version of chemical energy in wood and coal to power. True, there was less access
to fuel in the center of the empire, but if a practicable solution to the conversion
problem had appeared, large parts of the central and western European forests and
coal deposits would have been at the empire's disposal. But these resources were
never utilized in an intensive fashion.
Several of Hero's automata include the conversion of chemical energy to power

via the heating of water and pressure of hot air or steam. Most interesting is the
aeolipile, a kind of simple reaction turbine, which in theory should be able to act as
a prime mover (Pneum. 2.11). This device proves that ancient scientists understood
the principles behind this kind of energy conversion, but it also reveals their in-
ability to transfer the theory to practical use. The steamball rotates when water is
heated, but it is unable to drive anything else (Landels 1978: 28-31; Keyser 1994).
Only through the steam engines ofNewcomen and Watt in the eighteenth century
were the possibilities created for the conversion of fuel to power that became the
basis for the Industrial Revolution in Europe.
REFERENCES
Ad, U., 'A. Sa'id, and R. Frankel 2005. "Water-mills with Pompeian-type millstones at
Nahal Tanninim," Israel Exploration Journal 55: 156-71.
Avitsur, S. 1960. "On the history of the exploitation of water power in Eretz-Israel," Israel
Exploration Journal 10: 37-45.
Beckmann, J. 1788. Beitrage zur Geschichte der Erfindungen. 2 vols. Leipzig: Kummer.
Bell, M. 1994, "An Imperial flour mill on the Janiculum," in Le ravitaillement en ble de
Rome et des centres urbains des de'buts de la Republique jusqu'au Haut Empire. Col-
lections Centre Jean Berard n. Naples and Rome: Centre Jean Berard, 73-89.
Benoit, F. 1940. "L'usine de meunerie hydraulique de Barbegal (Arles)," Revue arche-
ologique 15: 19-80.
a
Bloch, M. 1935. "Avenement et conquetes du moulin eau," Annales d'histoire economique
et sociale 7: 538-63.
Bliimner, H. 1912. Technologie und Terminologie der Gewerke und Kunste bei Griechen und
Romern. 2nd ed. Vol. 1. Leipzig and Berlin: Teubner.
Butti, K., and J. Perlin 1980. A golden thread: 2500 years of solar architecture and technology.
New York: Van Nostrand Reinhold.
Castella, D., et al. 1994. Aventicum VI. Le moulin hydraulique gallo-romain d'Avenches "En
Chaplix," Fouilles 1990-1991. Cahiers d'Archeologie Romande 62. Lausanne: Cahiers
d'archeologie romande.
Champagne, F., A. Ferdiere, and Y. Rialland 1997. "Re-decouverte d'un moulin eau a
augusteen sur l'Yerre (Cher)," Revue archeologique du Centre de la France 36: 157-60.
Chastagnol, A. 1981. "L'inscription Constantinienne d'Orcistus," Melanges de l'Ecole
franraise de Rome et de Athenes 93: 381-416.
Cleere, H. F. 1976. "Some operating parameters for Roman iron-works," Bulletin of the
Institute of Archaeology of the University of London 13: 233-46.
Curwen, E. C. 1944. "The problem of early water-mills," Antiquity 18: 130-46.
Dearne, M., and K. Branigan 1994. "The use of coal in Roman Britain," Antiquaries Journal
74: 71-105.
Drachmann, A. G. 1961. "Heron' s windmill," Centaurus 7: 145-51.
Forbes, R. J. 1955a. Studies in ancient technology. Vol. 1. Leiden: Brill.
Forbes, R. J. 1955b. Studies in ancient technology. Vol. 2. Leiden: Brill.
Forbes, R. J. 1958. Studies in ancient technology. Vol. 6. Leiden: Brill.

Greene, K. 1999. "Review of A. M. Small and R. J. Buck, The excavations at San Giovanni
di Ruoti. Vol. 1, The villas and their environment," American Journal of Archaeology 103:
577-79
Guery, R., and G. Hallier 1987. "Reflexions sur les ouvrages hydrauliques de Marseille
antique retrouves sur le chautier de la Bourse," in L'eau et les hommes en Mediterranee,
4. Paris: CNRS, 265-82.
Horne, L. 1982. "Fuel for the metal worker: The role of charcoal and charcoal production in
ancient metallurgy," Expedition 25:1: 6-13.
Keyser, P. T. 1993. "The purpose of the Parthian galvanic cells: A first-century A.D. electric
battery used for analgesia," Journal of Near Eastern Studies 52: 81-98.
Keyser P. T. 1994. "Classics and technology: A re-evaluation of Heron's first century A.D.
steam engine," in S. Wisseman and W. Williams (eds.), Ancient technologies and
archaeological materials. London: Gordon and Breach, 71-86.
Lancaster, L. 2005. Concrete vaulted construction in imperial Rome: Innovations in context.
Landels, G. 1978. Engineering in the ancient world. London: Chatto and Windus.
Lanfranchi, G. B., and S. Parpola 1990. State archives ofAssyria. Vol. 5, The correspondence
of Sargon II, Part 2: Letters from the northern and northeastern provinces. Helsinki:
Helsinki University Press.
Lauffer, S. 1971. Diokletians Preisedikt: Texte und Kommentare. Berlin: de Gruyter.
Leveau, P. 1996. "The Barbegal water mill in its environment: Archaeology and the eco-
nomic and social history of antiquity," Journal of Roman Archaeology 9: 137-53.
Lewis, M. J. T. 1993. "The Greeks and the early windmill," History of Technology 15: 141-89.
Lewis, M. J. T. 1997. Millstone and hammer: The origins of water power. Hull: University of
Hull Press.
Lohrmann, D. 1992. "Schiffmiihlen in spatantiken und mittelalterlichen Stadten," in
Medieval Europe: A conference on medieval archaeology in Europe, 21st-24th September
at the University of York. Pre-printed papers, Vol. 3: Technology and innovation. York:
Society for Medieval Archaeology, 27-29.
Lucas, A. 2006. Wind, water, work: Ancient and medieval milling technology. Leiden: Brill.
Ludwig, K.-H. 1981. "Die technikgeschichtlichen Zweifel an der 'Mosella' des Ausonius
sind unbegriindet," Technikgeschichte 48: 131-34.
Parsons, A. W. 1936. "A Roman water-mill in the Athenian agora," Hesperia 5: 70-90.
Pernot, L. 1983. "Pour une prehistoire de l'electricite: la Grece antique et l'ambre jaune,"
Bulletin d'histoire d'e1ectricite 2: 19-30.
Pleket, H. W. 1988. "Greek epigraphy and comparative ancient history: Two case studies,"
Epigraphica Anatolica 12: 25-37.
Radl, A. 1988. Die Magnetstein in der Antike: Quellen und Zusammenhiinge. Stuttgart:
Steiner Verlag.
Rakob, F. L. 1993. "Der Neufund einer romischen Turbinenmiihle in Tunesien," Antike
Welt 24: 286-87.
Rehder, J. E. 2000. Mastery and uses of fire in antiquity. Montreal: McGill-Queens Press.
Ring, J. W. 1996. "Windows, baths and solar energy in the Roman empire," American
Journal of Archaeology 100: 717-24.
Ritti, T., K. Grewe, and P. Kessner 2007. "Stridentes trahens per levia Marmora serras: A
Roder, J., and G. Roder 1993. "Die antike Turbinenmiihle in Chemtou," in F. Rakob (ed.),
Simitthus. Vol. 1, Die Steinbrucke und die antike Stadt. Mainz: von Zabern, 95-102.
Romeuf, A. M. 2001. Le quartier artisanal gallo-romain des Martres-de-Veyre (Puy-de-
Dome). Cahier du Centre Acheologique de Lezoux 2. Lezoux: Revue Archeologique
Sites.
Rosumek, P. 1982. Technischer Fortschritt und Rationalisierung im antiken Bergbau. Bonn:
Habelt.
Rynne, C. 1992. "Early medieval horizontal-wheeled mill penstocks from Co. Cork,"
Journal of the Cork Historical and Archaeological Society 97: 54-68.
Schi0ler, T. 1989. "The watermills at the Crocodile River: A turbine mill dated to 345-380
A.D," Palestine Exploration Quarterly 121: 133-43.
Schi0ler, T., and 6. Wikander 1983. "A Roman water-mill in the Baths of Caracalla,"
Opuscula Romana 14: 47-64.
Seguin, A. 1938. "Etude sur le petrol dans l'antiquite grecque et latine: Gisements de petrole
connus des Latins et des Grecs," Revue des Questions Historiques 261: 36--70.
Seigne, J. 2002a. "A sixth-century waterpowered sawmill," International Molinology 64:
14-16.
Seigne, J. 2002b. "A sixth-century water-powered sawmill at Jarash," Annual of the De-
partment of Antiquities of Jordan 46: 205-13.
Simms, D. L. 1977. "Archimedes and the burning mirrors of Syracuse," Technology and
Culture 18: 1-24.
Simms, D. L. 1983. "Water-driven saws, Ausonius and the authenticity of the Masella,"
Technology and Culture 24: 635-43.
Simpson, F. G., and Lord Wilson of High Wray 1976. Watermills and military works on
Hadrian's Wall: Excavations in Northumberland, 1907-1913. Edited by G. Simpson.
Kendal: T. Wilson.
Smith, N. A. F. 1983-1984. "The origins of water-power: A problem of evidence and
expectations," Transactions of the Newcomen Society 55: 67-84.
Spain, R. J. 1987. "The Roman watermill in the Athenian agora: A new view of the evi-
dence," Hesperia 56: 335-53.
Vetters, H. 1984. "Ephesos: Vorlaufiger Grabungsbericht 1983," Anzeiger der 6sterreichischen
Akademie der Wissenschaften in Wien, Philologisch-historische Klasse 121: 209-32.
Webster, G. 1955. "A note on the use of coal in Roman Britain," Antiquaries Journal 35: 199-
216.
Wikander, 6. 1979. "Water-mills in ancient Rome," Opuscula Romana 12: 13-36.
Wikander, 6. 1981. "The use of water-power in classical antiquity," Opuscula Romana 13:
91-104.
Wikander, 6. 1985. "Archaeological evidence for early water-mills-An interim report,"
Wikander, 6. 1991. "Water-mills and aqueducts," in A. T. Hodge (ed.), Future currents in
aqueduct studies. Leeds: Francis Cairns, 141-48.
Wikander, 6. 1992. "Water-mills in Europe: Their early frequence and diffusion," in
Medieval Europe: A conference on medieval archaeology in Europe, 21st-24th September
at the University of York. Pre-printed papers, Vol. 3: Technology and innovation. York:
Society for Medieval Archaeology, 9-14.
Wikander, 6. 2000. "The water-mill," in 6. Wikander (ed.), Handbook of ancient water
Wikander, 6. 2002. "'Where of old all the mills of the city have been constructed': The
capacity of the Janiculum mills in Rome," in Ancient history matters: Studies presented
to J. E. Skydsgaard on his seventieth birthday. Analecta Romana Instituti Danici Suppl.

30. Rome: Danish Institute, 127-33.
Wikander, b. 2004. "Invention, technology transfer, breakthrough-The ancient water-
mill as an example," in F. Minonzio (ed.), Problemi di macchinismo in ambito romano.
Archeologia dell'Italia Settentrionale 8. Como: Commune di Como, 107-12.
Wilson, A. I. 1995. "Water-power in North Africa and the development of the horizontal
water-wheel," Journal of Roman Archaeology 8: 499-510.
Wilson, A. I. 2000. "The water-mills on the Janiculum," Memoirs of the American Academy
in Rome 45: 219-46.
Wilson, A. I. 2001. "Water-mills at Amida: Ammianus Marcellinus 18.8.11," Classical
Quarterly 51: 231-36.
92: 1-32.
CHAPTER 7
GREEK AND ROMAN

AGRICULTURE
EVI MARGARITIS
MARTIN K. JONES
IN the period for which we have written records, much of the agricultural tech-
nology of those parts of the world that came under European influence finds its
roots in the classical world. This legacy includes the usage of a range of heavy metal
implements, in particular the moldboard plow, long sickle, and scythe, as well as
the management of drainage schemes, meadows, water-driven mills, and grinding
equipment. These important technological legacies, however, were not typical of
farms in the classical world itself. Several of the more enduring technologies de-
veloped quite late in the classical epoch, and moreover around the boundaries of
the Roman world, where classical and local traditions came together and jointly
contributed to agrarian development.
A far more typical form of classical agriculture has its roots instead in a long-
standing harmonization between Mediterranean farmers and the particular chal-
lenges of their environment. It entailed the implementation of simpler and more
ancient technologies, based in wood and with much sparser use of metal, powered
either by human energy or by a single cow, donkey, or mule. Throughout the
course of the classical epoch, the great majority of farmers were engaged in these
smaller-scale agrarian technologies rather than in the capital-intensive technolo-
gies associated with the later Roman world. In this chapter, we explore the nature
of agriculture in the Mediterranean epicenter, paying special attention to the com-
bination of arboriculture and arable agriculture typical of the region. In the context
of that account, we also follow developments on the periphery of that world that
led to a range of enduring innovations in agricultural technology. Our evidence
GREEK AND ROMAN AGRICULTURE 159
comes from two principal sources, texts and archaeology, and the interpretation of
that evidence draws repeatedly on ethnographic observation of more recent small-
scale Mediterranean agriculture.
A great deal of the debate on the roots of the ancient economy has been drawn
from incidental references in Athenian legal speeches and philosophical dialogues.
Sources that directly describe Greek agriculture and farming practices are extremely
scarce, which is paradoxical considering the multitude of sources referring to the
importance of agriculture on a theoretical, idealistic basis. Thus, the methods of
cultivating and processing various crops, their different uses, their individual eco-
nomic and social value and status, and the character of the rural landscape are barely
mentioned (cf. !sager and Skydsgaard 1992: 4). Sources such as the Athenian le-
gal speeches and philosophical dialogues are in any case likely to reflect cultural ideals
as much as historical realities (Cahill 2001: 223). The literary sources generally con-
cern the upper strata of society, those involved in legal cases, mortgages ofland, and
similar situations (Cahill 2001: 224). In addition, most of the available information
concerns the large urban centers, especially Athens, and therefore the paucity of
written sources from the rural, farming areas of Greece, does not allow for a closer
investigation of agricultural systems and patterns in the agricultural countryside.
The textual evidence is more direct during the Roman period, with a much
richer body of works specifically on agricultural matters, by such authors as Cato,
Varro, and Columella. These, however, must also be read in the context of both the
regions of the Roman world and the strata of Roman society (again, the upper
stratum of estate-owning society) about which they were writing. A complemen-
tary picture, with different regional and social emphases, is coming from a growing
body of archaeological evidence from throughout the classical world and beyond,
and particularly the direct bio-archaeological evidence of the remains of crops,
livestock, and implements. Visual representations of ancient farmers at work con-
stitute another invaluable source (e.g., figure 2.13).
THE MEDITERRANEAN EPICENTER
Mediterranean ecosystems are characterized by abrupt topography and marked

seasonality, features that have the effect of fragmenting the accessibility of natural
resources and partitioning them across space and time (Horden and Purcell 2000).
Much of classical agrarian practice was aimed at containing and controlling this
spatial and temporal fragmentation. It began with the choice of combinations of
plant and animal species that were themselves well adapted to coping with such
fragmentation in space and time. These species were contained and nurtured within
modestly sized plots to which the supply of water was often managed and moderated.
Distinct altitudinal zones of this rugged landscape were used in complementary
manners for crop growth, animal feed, and woodland resources (figure 7.1).
Figure 7.1. Mediterranean polyculture: olives in the forefront, vines in the middle dis-
tance, arable fields and hillside grazing in the distance. (Photograph by Graeme Barker.)
There were some broad differences in the manner in which Greek and Roman
societies adapted to the Mediterranean ecosystem (Foxhall et al. 2007). Agriculture
on Roman estates was geared toward a control and mastery of nature. Through
interaction with neighboring communities, technologies were developed that
could be taken beyond the Mediterranean, to cooler and wetter environments in
the north and hotter, more arid environments in the south and east. By contrast,
Greek agriculture was more a process of harmonization with, rather than a mastery
of, the Mediterranean ecosystem. The Greek world remained more closely tied to
that particular region, as can be seen from the correlation between the distribu-
tion of Greek settlements and that of the olive tree (Foxhall et al. 2007) .
Two models have been proposed for this harmonizing interpretation of Greek
agriculture. Hodkinson described them as "models of symbiosis and divorce" (1988:
38). The first, "traditional" model uses records of Mediterranean agriculture and
animal husbandry from recent centuries as analogues for the more distant classical
past (Semple 1932: 386; Finley 1973: 108; Duncan-Jones 1982: 49) . According to this
model, agriculture was practiced in a biennial cycle in which cereal crops alternated
with bare fallow. Bare fallow involves the plowing of weeds before they set seed, and
therefore this method would have limited weed growth and resulted in the accu-
mulation of two years' moisture in the field for the succeeding cereal crop (Semple
1932: 386). This conservation is important, considering the limiting factor of the
availability of water in the arid environment of Greece (Halstead 1987: 78, 81). Fol-
lowing the ramifications of this model, the maintenance of soil fertility was depen-
dant neither on manuring nor on pulse rotation; in some cases these were even con-
sidered disadvantageous (Semple 1932: 411) . As agricultural experiments have shown,
however, manured annual cropping is far more productive than a bare fallow-cereal
rotation, provided, of course, that sufficient manure is available (Littlejohn 1946).
The cereal-to-bare-fallow rotation would have resulted in the scarcity of lowland

grazing sources, in turn forcing the livestock to move to other areas for sustenance.
This model envisages the seasonal transhumance of sheep and goats, migrating from
the arid environment of the lowlands, to the cooler and more lush highlands (Semple
1932: 297; Michell 1963). The removal of the livestock from the lowlands would also
have deprived the arable fields of manure from the grazing animals. Consequently,
this model envisages a separation between arable farming and pastoralism, a sepa-
ration that could be seen in traditional economies of Greece and other Mediterranean
countries, and a feature that has also been taken to distinguish traditional farming in
the Mediterranean and temperate European regions (Halstead 1987: 79). Reinders
and Prummel (1998: 82) have inferred specialized pastoralism and transhumance in
Hellenistic Thessaly and the site of New Halos on the basis of faunal remains, two
literary sources attesting to the use of summer pastures in the Orthis Mountains
during the Hellenistic period, and contemporary ethnographic observations. The
same variety of faunal remains (sheep, goats, cattle, pigs, equids, and dogs) found at
New Halos (Reinders and Prummel 1998: 82) were also found at Cassope in western
Greece from the fourth to first century B.C. (Boessneck 1986), and they have been
interpreted as compatible with mixed farming (Halstead 1990: 71).
An alternative model to the one described above questions the uncritical use of
contemporary traditional practices as analogies for antiquity. Drawing instead on
direct bio-archaeological evidence for crop harvests and livestock, it posits a closer
connection between animal husbandry and arable agriculture (Halstead 1981, 1987;
Gallant 1982; Garnsey 1988: 93-94). In this model, agriculture and animal husbandry
are seen as integrated components in an intensive farming system involving the
raising of fodder crops, manuring, and crop rotation. The rotation of cereals and
pulses is emphasized, resulting in the availability of fodder in the lowlands during
summer. This availability of fodder would in turn obviate the need for the seasonal
movement of the animals, who would consume weeds and produce manure, en-
hancing at the same time the fertility and productivity of the land (White 1970: 119;
Halstead 1981: 328).
WATER MANAGEMENT
AND IRRIGATION
The alternative model described here presupposes some degree of water manage-
ment, and there remain doubts about the intensity and extent of water manage-
ment on Greek farms (Isager and Skydsgaard 1992). The role of irrigation in ancient
Greek agriculture has been debated extensively, and many researchers have argued
against complex water management in classical Greece (Gallant 1991: 56-57). On
the other hand, it has been posited that classical farmhouses had developed private
rather than communal sources of water, such as wells, which suggests an inten-
sification of agriculture (Hanson 1999: 60). Others have advocated for the use of
intensive watering and irrigation systems only at the gardens located near the city
houses and the independent farmhouses (Krasilnikoff 2002). The extent of irri-
gation and water control in classical Greece remains a matter for debate (Oleson
2000a: 205-8), but many Greek authors allude to the problem (e.g., Plato, Leg.
6.761a-b ), and Aristotle uses garden irrigation as an analogy for the human circu-
latory system (Part. an. 3.5.668a).
The situation in the Roman world is dearer, and it seems that the expansion of
Roman influence to the east and the south facilitated the flow of technological
ideas from western Asia and the Nile corridor to various arid regions around the
Mediterranean. A number of Roman authors refer to the irrigation of some arable
fields, orchards, vineyards, pastures, and gardens (Butzer et al. 1985). Water man-
agement was also applied to large-scale schemes involving drainage reclamation,
flood control, and navigation (White 1970).
A series of devices for lifting and moving water were adopted from the eastern
and southern boundaries of the Roman world (see chapters 11 and 13). In Andalucia
during the first century B.C., the geographer Strabo (3.2.9) recorded the use of a
water-lifting device that Archimedes had invented two centuries earlier in Egypt
(Dalley and Oleson 2003). The Archimedean screw or "water snail" was a hand-
turned helical device, originally used to lift water out of irrigation ditches and sub-
sequently adapted for a variety of purposes (Oleson 2000b: 242-51). A similar tra-
jectory can be followed for other water-lifting devices first recorded in Egypt. The
shaduf, a device involving a bucket on a counterweighted beam, appears in the Near
East by the seventeenth century B.C. and remained common throughout the Med-
iterranean nearly up to the present. Pliny (HN 19.60) mentions its use in Italy in the
first century A.D., and it had become common in Spain by at least the late Roman
period (Butzer et al. 1985). Animal-propelled water-lifting wheels worked through
an angle gear (saqiya), known in Egypt from the third century B.C., probably formed
the basis for the undershot water-mills familiar to Vitruvius (De arch. 10.5,1) and
Pliny (HN 19.4), and subsequently the basis for water-powered milling (Oleson 1984:
371-80; see also chapter 6). Elaborate irrigation systems based on underground
tunnels (the qanat) are first recorded in ancient Persia then subsequently in Roman
Tunisia (Hodge 1992: 20-24). While the shaduf and qanat are products of the arid
Near East, the mechanical water-lifting devices were all invented in Hellenistic
Alexandria (Oleson 1984, 2000b ), and the most intensive use of a number of these
irrigation technologies dearly follows the arid perimeters of the Roman world in
North Africa and Spain rather than the central regions of Greece and Italy.
ARBORICULTURE
Perennial woody crops have continuously constituted a central element of Med-

iterranean agriculture. A wide variety of species may be involved, sometimes in-
terplanted with other crops or pasture, at other times grown alone in orchards, and
utilized for food, fodder, fuel, or construction materials. Their ecological advan-
tage is that by their perennial habit and extensive root systems, their long-term
growth in itself ameliorates the sharp climatic seasonality and its effects on soil and
water. Alongside almonds, figs, carobs, oak, pistacia, pine, and many others, the
two perennial crops that became most closely associated with the classical Medi-
terranean world are the olive and grape (cf. chapter 14).
Models for the technology of classical oil and wine production have often drawn
on analogues from contemporary ethnography (Loukopoulos 1938; Aschenbrenner
1972; Forbes 1982, 1992; Gavrielides 1976; Zivas 2000). Much information can be
extracted from ancient literature and archaeological data, but many factors related
to these commodities are still uncertain, such as the area of land planted, the labor
expended in their production, and the scheduling of agricultural activities. Al-
though modern data requires critical evaluation, ethnographic studies offer some
basis for estimating labor as well as yield and production rates in ancient agricul-
tural communities.
The Grapevine
For the classical and postclassical periods, both Greek and Roman ancient writers
provide information about the winemaking procedures and the cultivation of the
vine. In the texts, numerous instructions are given on how to make good wine and
how to add the right flavorings (Humphrey et al. 1998: 154-59). They specified the
optimal way to store products and classify different wines in relation to their place
of production, variety of grapes, color, taste, and ways of being processed. Drawing
from written sources, Kissopoulos (1947, 1948, 1949) provided details about tools,
pottery, and other equipment used during and after vintage and wine making.
Both pictorial representations and written sources, however, can be biased, and
critical evaluation is fundamental (!sager and Skydsgaard 1992).
Among Greek writers, Theophrastus remains the main source of our knowl-
edge of Greek viticulture through two major works: Historia plantarum and De
causis plantarum (!sager and Skydsgaard 1992: 29). As a botanist, Theophrastus
took a natural interest in the growing requirements and cultivation of plants and
therefore provided general information about the vine. Only in rare cases did he
provide a more thorough account of viticulture and give detailed descriptions of
planting or pruning, or information about suitable soils. Theophrastus took it for
granted that the reader was fully familiar with the meaning of terms, and we there-
fore have no explanation or definition of the words employed (!sager and Sky-
dsgaard 1992: 28). Despite these difficulties, it is evident from his presentation that
viticulture required a great degree of care and tending. Theophrastus gives us the
first consistent description of the continuous process from pruning until the con-
cluding work prior to the ripening of the grapes. Only Homer and Hesiod mention
the actual vintage, although it was frequently shown in art together with the
treading of the grapes (!sager and Skydsgaard 1992: 33). In his Works and Days,
Hesiod deals also with the agrarian calendar and the pruning of vine, while Xe-
nophon in his Oeconomicus (19) describes the planting process in detail. In the
Deipnosophistae (11.10), Athenaeus explains how wine should be mixed and how to
produce vessels suitable for wine; many names for the vessels used in viticulture are
preserved in this work. Roman viticulture is discussed by a number of authors,
principally Cato, Varro, Columella, and Pliny. Their various discussions of the
many ways of training the vine, planting, cultivation, and pruning are summarized
in White (1970: 229-46; see also Humphrey et al. 1998: 114-20).
As with many crops, an important and laborious task in viticulture was dig-
ging. This took place in the winter while the vine was still dormant. Deep digging
was carried out around the trunk of each vine to allow rain, air, and manure to
break through to the roots. In this way, greater nourishment could reach new,
fruit-bearing growth. If digging was not done, or was not done in the correct way,
the new roots would not develop properly and the yield of the vine would be
greatly reduced (Pliny, HN 15.188; Columella, Rust. 4.28.2; Palmer 1994: 14). Re-
peated digging also reduced the weed growth, which could otherwise compete with
trees or vines for moisture and soil nutrients (Foxhall 1990: 107). During summer,
when grapes and other fruits expand and ripen, they depend entirely on the
moisture retained in the subsoil by the layer of broken soil at the surface. In com-
parisons made by Forbes (1976: 9) on two variously managed contemporary vi-
neyards, the one that was carefully and repeatedly dug yielded almost twice the
amount of grapes. It is common practice to manure the vineyard while digging.
Manuring usually involved animal dung, although various components such as
bones, scraps of leather and other domestic waste, charred grape pips, ash, and
charred wood could also be used. The spread of potsherds over ancient agricultural
fields is usually the incidental result of the dumping of household debris as fer-
tilizer. Even the addition of (dry) soil, transported from elsewhere, is mentioned in
the ancient literature (Palaiologos 1838: 19-20). Tilling the soil is another major
task in the vineyard, and Theophrastus (Hist. pl. 3.9.5, 3,10) is conscious of the
importance of aerating the soil and controlling weeds by tilling.
Vines require a greater degree of attendance and control of the environment
than any other ancient crop. They require skilled labor to plant, prune, and tend
throughout their growing period, and perhaps to guard when the grapes are rip-
ening, as appears to have been the case in ancient Athens (Morris and Papado-
poulos 2005: 179). Extra labor is required during harvesting and processing for the
production of either raisins or wine. Archaeobotanical methods applied to the site
of Komboloi in southern Macedonia have charted the harvesting and processing
of grapes at a first-millennium B.C. farmstead (Margaritis and Jones 2006).
The Olive Tree

As olives produce a full crop in alternate years, Columella proposed that farmers
should divide olive groves into two parts that would give an equal return each year
(Rust. 5.99, 2.7). At present, many farmers gather their olives green and immature
in an attempt to make the trees produce the following year as well, as it is believed
that when olives are gathered mature (black) the tree does not have enough re-
sources and strength to produce olives for two successive years.
The maintenance and cultivation of olive groves involves digging, weeding,
pruning, harvesting, fertilizing, and maintaining terraces, if necessary (Pansiot and
Rebour 1961). As with viticulture, the most laborious and time-consuming part of
tree cultivation was digging. This activity is best done during late autumn, although
many farmers do it in springtime. The purpose of the digging is to assist surface
water reaching the roots. Many farmers merely dig a ditch around the trees to the
same diameter as the spread of the branches, so that the rainwater can accumu-
late there (Gavrielides 1976: 83). Repeated digging destroys roots close to the surface,
thus encouraging the development of roots at lower levels where more moisture
is available and they are less likely to suffer from sun-scorch or exposure, hence
allowing trees to utilize available water more efficiently. This procedure is partic-
ularly important for olives, which naturally have wide-spreading roots close to the
surface (Foxhall 1990: 106-7). Repeated digging also reduces the growth of weeds
that would otherwise compete with trees or vines for moisture and soil nutrients.
Pruning is also a very important task for olive cultivation, as it encourages
the growth of new shoots and helps in the concentration of the juices of the tree
(Isager and Skydsgaard 1992: 32). If the yield is not abundant, pruning is done at
the same time as harvesting. When the yield is average or good, pruning begins
after the harvest and continues during the months of December and January.
Weeding begins in early September and lasts all month. During this period, the
trees are loaded with fruit ready to be harvested. Weeding is done in order to clear
the land under the trees from weeds and thorns in order to facilitate harvesting the
portion of the crop that happens to fall on the ground.
A further aspect of olive grove maintenance is the repair of any existing terrace
walls. The amount of time expended depends on the number of terraces in a given
orchard and the time available. It is hard, time-consuming work and provides no
immediate tangible returns (Foxhall 1990: 90 ). The culmination of the agricultural
year is the time of the harvest. Harvesting could be carried out by several methods,
such as hand-stripping, beating the branches, or gathering from the ground (Brun
1986: 36-38).
ARABLE CULTIVATION
The principal cereal of the Greek world was barley, a crop that was well adapted to
the Mediterranean environment. Its relative tolerance of drought and salinity suited
it well to the typical hydrological conditions. The only cereals that can tolerate a
shorter growing season are the millets, which were grown in irrigated fields, as
their water demands during that growing season are higher. Legumes are also well
attested in the archaeobotanical record, as well as a range of oilseeds such as sesame

and flax. Wheat grew steadily in prominence during the classical epoch. A number
of the key wheat species are more demanding than barley and require a more care-
ful management of the Mediterranean ecology. The impetus behind the expanded
use of wheat may well have been more cultural than ecological, and there is much
textual evidence for the status of different types of bread in Greek and Roman
society, for example in Athenaeus' Deipnosophistae (3.108-16; cf. Pliny, HN 18.105-
6). Pliny (HN 18.86-92) discusses the relationship between types of wheat and
varieties of bread.
Before planting, the ground could be prepared with a range of manual sod-
breaking tools, including wooden mallets and mattocks. These were particularly
important in the cultivation of small plots. In larger fields, the majority of Greek
and Roman farmers throughout the classical period cultivated their soils with the
same implement that had been in use since the Neolithic period: the scratch plow,
or ard, with a single, simple share, light enough to be pulled by a single cow,
donkey, or mule (figure 7.2). In Georgics (1.169-75), Vergil provides a detailed de-
scription of what is probably a bow-ard, constructed from elm, linden, and beech
wood, with aures or "ears" attached to the share (figure 7.3; Varro, Rust. 1.29.2
terms them tabellae, "boards"; Aitken 1956; Manning 1964). Like moldboards, these
"ears" probably helped turn the soil. Bronze models of such ards with turning
boards are known from Britain and Germany, and complete wooden ards have
been recovered from the peat deposits of Northern Europe, though generally
lacking the aures, which appear to have been more common in the south.
Pliny (HN 18.48) refers to an invention that was recent at the time he wrote
(the mid-first century A.D.), the addition of two small wheels, implying that the
plow was becoming heavier. From the late third century A.D., we find archaeo-
logical evidence for the development of an implement that went on to become the
principal tool of arable cultivation, first in Europe, and then accompanying the
expansion of European influence around the world. The moldboard plow, with a
heavy cutting blade at its front, followed by an asymmetrical share and sod-turning
moldboard, and assisted by a variety of wheels, makes its appearance during the
later stages of the Roman epoch (figure 7.4; Jones 1981, 1991).
Little, if any, metal was used in the construction of an ard-in line with the
wood-based technology of farmers in general. Metal was scarce and conserved, used
only where a durable cutting edge was absolutely necessary. Even then, obsidian
flakes often provided a more accessible material for a sharp edge in areas with the
appropriate geology, blades of obsidian being inserted both in threshing sleds and
sickle mounts. The heavy investment in metal seen in the later plow was simply not
within the resources of the great majority of classical-period farmers. Indeed, the
need for it only arose beyond the limits of the Mediterranean region, where the
soils were heavier.
North of the Alps, within temperate Europe, the environmental challenges
faced by farmers were significantly different from those faced in the Mediterranean.
While the annual level of sunshine is lower in more northerly latitudes, peaks of
Figure 7.2. Modern cross-cultivation with an ard. (Photograph by M. Jones.)
sunshine and rainfall are no longer segregated to separate seasons. In fact, neither
water nor the unevenness of its supply are factors limiting biological productivity.
The different conditions had significant implications for animal husbandry, par-
ticularly for animals with large water requirements, such as horses and cattle. There
is much archaeological evidence for prehistoric traditions of cattle husbandry on a
variety of scales in Europe north of the Alps, sometimes indicating large herd sizes.
There is, furthermore, growing evidence for long traditions for the production of
secondary bovine products such as milk and butter (chapter 8).
Temperate Europe is characterized by much larger expanses of plains and
gently rolling topographies than the Mediterranean. Rather than gardens and ter-
races (cf. Van der Veen 2005), the emphasis was on extensive fields growing a
diverse range of cereal and legume crops. While small enclosures and paddocks are
widespread, we know of no clear evidence of what might be described as a "garden"
in prehistoric temperate Europe. As we move north and east across Europe, the
limiting factors to growth in these fields were depressed sunlight hours and con-
straints on both ends of the growing season by late spring frosts and late summer
storms. The winter and spring frosts in particular were ameliorated on the Atlantic
face of Europe by the Gulf Stream, and, significantly, it was along the Atlantic face
of Europe, in Spain, France, and Britain, that the Roman Empire proved ecolog-
ically able to extend furthest beyond the Mediterranean (Cunliffe 2001: 365-421) .
There is now clear evidence that mixed farming with cattle and cereal agriculture is
attested in many parts of prehistoric temperate Europe.
However, the imperial limes only stabilized where cereals were not simply
grown, but were already being mobilized as surplus on at least a local scale. Rather
than being attached to a particular set of farming practices and environmental
Figure 7.3. A traditional "eared" plow similar to that described by Vergil. (Photograph by
Martin Millett.)
resources, the ecology of Roman farming was instead attached to a style of living
and of agrarian management, and, critically, an exchange network to which these
were attached. In Germany, the Netherlands, and northern Britain that ecological
pattern fluctuated around the limits of its sustainability, ultimately settling along
the Rhine valley on the east and the line of Hadrian's Wall on the north (Van der
Veen 1992, Zacchariasse 2003).
Archaeobotanical evidence from this period of temperate Europe suggests a
continuity of traditional agrarian practice, of extensive field agriculture using the
same crops that had been grown in the region for centuries, and the same farming
techniques. Roman imperial expansion northward was clearly manifest in styles of
consumption, networks of exchange, military appropriation, and rural villa archi-
tecture, but the agrarian base initially remained the same, and quite distinct from
Mediterranean practice. With time, however, these styles of ecological engagement
were themselves transformed.
The first discernible transformation was in the use of metal. Iron production
was greatly stimulated by the imperial process, and unprecedented quantities of
this hitherto scarce resource were distributed around the imperial network. Some
was used to fashion a new generation of harvesting tools, which have been inter-
preted as hay scythes (Rees 1979). The harvesting equipment attested in prehistory
would have been poorly suited to gathering herbaceous hay, which can only be ef-
fectively harvested with a much longer blade than was available to prehistoric
farmers. There is scattered evidence that the branches of leafy fodder that can be
harvested with more modest blades was the prevailing winter-feed of livestock
(Jones 1991). The long iron blades that appear in the early Roman Empire allowed
the plant community we know as hay meadows to proliferate. This technological
innovation may even have led to their original genesis as distinct plant commu-
Figure 7-4- A contemporary moldboard plow with wheel and coulter before the share.
(Photograph by M. Jones.)
nities (Jones 1991). Although the first appearance of hay meadows seems linked to
Roman expansion, the Roman favorites of pork, chicken, and oysters are little aug-
mented by the maintenance of hay meadows. This novel ecological product of
metal technology, the hay meadow, was of more relevance in the sustenance (and
especially the winter survival) of those essentially northern symbols of prestige and
wealth, horses and cattle.
Otherwise, the clearest indications of change in food have to do with handling
and mobilization on an unprecedented scale. The limes forts along Hadrian's Wall
contain a series of vast granaries storing cereals in unprecedented quantities, and
the traces of mechanical mills that transformed their contents into flour. More
valuable and transportable foods such as wine and olive oil were traveling across
the Empire and beyond (Bakels and Jacomet 2003). The peppercorns that have
been recovered from a number of sites in Roman Britain may well have traveled
from the Malibar coast in southern India, entering the imperial network at the
African port of Berenice (Cappers 1999).
MEADOWS, GARDENS,
AND MOLDBOARD PLOWS
Ornamental gardens had appeared in conjunction with the earliest villas and
"palaces" within the northern limes, notably attested at Fishbourne in Sussex
(Cunliffe 1971). We are not sure to what extent, if at all, they had a strictly economic
dimension, just as we are unclear whether the hay meadows attested by hay scythes
and fragments of the hay itself were of widespread ecological significance or some-
thing specifically for raising prestigious animals. From the late third century A.D.
onward, however, there is recurrent evidence from villa plans, planting pits, "dark
earths," and the diversification of subsidiary food plants that an economic garden
was becoming a recurrent feature of the northwest provinces (Jones 1989). This was
also a period in which the balance of investment between the town and the coun-
tryside was shifting, and the imperial margins, in both the north and the south,
were witnessing a surge in conspicuous rural wealth. In the northwest provinces,
this provided the context for a highly significant ecological shift. This shift was
the consequence of a fusion between the core elements of temperate European
agriculture-adequate water, gentle topography, deep soils, and a long tradition of
cattle management-and the core elements of Roman agrarian ecology, particu-
larly the ease of mobilization and exchange of goods, materials, and investment.
Just as with the earlier hay scythes, a crucial material that was mobilized was iron.
Three iron implements that recur on Romano-British sites of the late third and
fourth centuries A.D. are the long harvesting scythe, the coulter, and the asym-
metrical plowshare (Rees 1979). All three can be linked to the intensification of
cereal production, and the latter two to the moldboard plow, designed to cut deep
into the soil and invert the furrow (figure 7.4; Jones 1981). Contemporary inverted
furrows have also been found with a buried soil at the north German site of
Feddersen Wierde (Korber-Grohne 1967). Like some of the metalwork finds, this
latter site lies beyond the boundary of the Empire itself. In a manner that mirrors
the development of irrigation techniques in the south and east, these find spots
beyond the limes in the north and west suggest that moldboard plowing arose from
a fusion between ecological traditions rather than as a simple introduction. Fur-
thermore, the moldboard plow was developed to deal with environmental con-
ditions specific to European landscapes north of the Alps-the simple ard, which
did not invert the soil and was associated with shallow tillage, continued as the
cultivation implement of preference in much of the Mediterranean until the twen-
tieth century A.D. (figure 7.2). Interestingly, Korber-Grohne (1967) has interpreted
some of the economic plant remains from Feddersen Wierde as deriving from what
were essentially managed "gardens" within the settlement. The crop repertoire also
shifts with the spread of deep moldboard plowing, with a particular emphasis on
bread wheat in the western regions of temperate Europe, and rye in the East, both
crops favored by deep cultivation.
Although the full impact of deep plowing and bread-wheat cultivation came
after the end of the classical period, a key element of the late classical world,
Christianity, took bread wheat and the plow first across Europe, and then across
the world, as hallmarks of the new faith. In the process it raised bread wheat to the
status of the primary food source of the human species (figure 7.5). The early roots
of that process lie in a particular style of Roman ecology that emphasized control,
management, and distribution, rather than overstamping or eradicating any ex-
isting agrarian practice.
Figure 7.5. Raised stone-built military granary at Corbridge fort, Northern England.
(Photograph by Martin Millett.)
The technology of classical agriculture owed much to its roots in adaptation to the
seasonal fluctuations of the Mediterranean ecosystem and its effect on water avail-
ability and soil stability. Conversely, the dynamics of change through time owed
much to the economic structures of the classical world. These in particular related
to the extensive nature of trade and exchange, whereby both biological resources,
such as seed stock, and technological resources, crucially metal, as well as expertise
in oral and written forms, reached a wide range of farms in many regions. These
structures also related to the potential of land ownership and long-term arbori-
culture. Some of the most enduring legacies of classical agriculture, however,
particularly extensive water management and the heavy plow, owe much to the in-
teraction between agricultural technologies in the Mediterranean heartland of the
classical world and long-standing practices in the geographical regions to which
classical influence subsequently spread.
REFERENCES
Aitken R. 1956. "Virgil's plough," Journal of Roman Studies 46: 97-106.

Alcock, S., J. Cherry, and J. Davis 1994. "Intensive survey, agricultural practice and
classical landscape of Greece," in I. Morris (ed.), Classical Greece: Ancient histories
and modern archaeologies. 2nd ed. Cambridge: Cambridge University Press,
137-70.
Aschenbrenner, S. E. 1972. "Contemporary community," in W. A. McDonald and G. R.

Rapp Jr. (eds.), The Minnesota Messenia expedition: Reconstructing a Bronze Age re-
gional environment. Minneapolis: University of Minnesota Press, 47-63.
Bakels, C., and S. Jacomet 2003. "Access to luxury goods in central Europe during the
Roman period: The archaeobotanical evidence," World Archaeology 34: 542-57.
Boessneck, J. 1986. "Zooarchaologische Ergebnisse an den Tierknochen- und Mollusken-
funden," in W. Hoepfer and E.-L. Schwandner (eds.), Haus und Stadt im Klassischen
Griechenland. Munich: Deutscher Kunstverlag, 136-40.
Brun, J.P. 1986. L'Oleiculture antique en province: Les huileries du departement du Var.
Paris: CNRS.
Butzer, K. W., J. F. Mateu, E. Butzer, and P. Kraus 1985. "Irrigation agrosystems in eastern
Spain: Roman or Islamic origins?" Annals of the Association of American Geographers
75: 479-509.
Cahill, N. 2001. Household and organization at Olynthus. New Haven: Yale University
Press.
Cappers, R, 1999. "Trade and subsistence at the Roman port of Berenike, Red Sea coast,
Egypt," in: Van der Veen (ed.), The exploitation of plant resources in ancient Africa.
New York: Kluwer, 185-97.
Cunliffe B. W. 1971. Fishbourne: A Roman palace and its garden. London: Thames and
Hudson.
Cuniffe, B. W. 2001. Facing the ocean: The Atlantic and its peoples. Oxford: Oxford Uni-
versity Press.
Dalley, S., and J.P. Oleson 2003. "Sennacherib, Archimedes, and the water screw: The
context of invention in the ancient world." Technology and Culture 44: 1-26.
Duncan-Jones, R. 1982. The economy of the Roman Empire. Cambridge: Cambridge Uni-
versity Press.
Finley, M. I. 1973. The ancient economy. Berkeley: University of California Press.
Forbes, H. A. 1976. "The 'thrice-ploughed field': Cultivation techniques in ancient and
modern Greece," Expedition 19.1: 5-11.
Forbes, H. A. 1982. "Strategies and soils: Technology, production and environment in the
peninsula of Methana, Greece." PhD diss., University of Pennsylvania.
Forbes, H. A. 1992. "The ethnoarchaeological approach to ancient Greek agriculture," in
B. Wells (ed.), Agriculture in ancient Greece. Proceedings of the Seventh Interna-
tional Symposium at the Swedish Institute at Athens, 16-17 May 1990. Acta Instituti
Atheniensis regni Sueciae 42. Stockholm: Paul Astroms Forlag, 87-101.
Foxhall, L. 1990. "Olive cultivation within Greek and Roman agriculture: The ancient
economy revisited." PhD diss., University of Liverpool.
Foxhall L., M. K. Jones, and H. Forbes 2007. "Human ecology and the classical land-
scape," in S. Alcock and R. Osborne (eds.), Blackwell guide to classical archaeology.
Oxford: Blackwell.
Gallant, T. W. 1982. "Agricultural systems, land tenure, and the reforms of Solon," Annual
of the British School at Athens 77: 111-24.
Gallant, T. W. 1991. Risk and survival in ancient Greece: Reconstructing the rural domestic
economy. Stanford: Stanford University Press.
Garnsey, P. 1988. Famine and food supply in the Graeco-Roman world: Responses to risk
and crisis. Cambridge: Cambridge University Press.
Gavrielides, N. E. 1976. "A study in the cultural ecology of an olive-growing community:
The southern Argolid, Greece." PhD diss., Indiana University.
Halstead, P. 1981. "Counting sheep in Neolithic and Bronze Age Greece," in I. Hodder, G.
Isaac, and N. Hammond (eds.), Pattern of the past: Studies in honour of David Clarke.
Cambridge: Cambridge University Press, 307-39.
Halstead, P. 1987. "Traditional and ancient rural economy in Mediterranean Europe: Plus
ra change?" Journal of Hellenic Studies 107: 77-87.
Halstead, P. 1990. "Present to past in the Pindos: Diversification and specialisation in
mountain economies," Rivista di Studi Liguri 56: 61-80.
Hanson, V. D. 1999. The other Greeks: The family farm and the agrarian roots of Western
civilization. 2nd ed. Berkeley: University of California Press.
Hodge, A. T. 1992. Roman aqueducts and water supply. London: Duckworth.
Hodkinson, S. 1988. "Animal husbandry in the Greek polis," in C.R. Whittaker (ed.),
Pastoral economies in classical antiquity. Cambridge Philological Society Suppl. 14.
Cambridge: Cambridge Philological Society, 35-74.
Horden, P., and N. Purcell 2000. The corrupting sea: A study of Mediterranean history.
Oxford: Blackwell.
Humphrey, J., J. P. Oleson, and A. N. Sherwood 1998. Greek and Roman technology: A
!sager, S., and J. E. Skydsgaard 1992. Ancient Greek agriculture: An introduction. London:
Routledge.
Jones, M. K. 1981. "The development of crop husbandry," in M. K. Jones and G. W.
Dimbleby (eds.), The environment of man: The Iron Age to the Anglo-Saxon period.
British Archaeological Reports 87. Oxford: BAR, 95-127.
Jones, M. K. 1989. Agriculture in Roman Britain: The dynamics of change," in M. Todd
(ed.), Research on Roman Britain, 1960-1989. London: Society for the Promotion of
Roman Studies, 127-34.
Jones, M. K. 1991. "Agricultural productivity in the pre-documentary past," in B. Campbell
and M. Overton (eds.), Land, labour and livestock: Historical studies in European
agricultural productivity. Manchester: Manchester University Press, 78-93.
Kissopoulos, D. 1947. "Greek and Roman wine making," Chemistry Journal 12A: 22-28.
Kissopoulos, D. 1948. "Wine making of the classical, Roman and Byzantine periods,"
Chemistry Journal 13A: 24-30.
Kissopoulos, D. 1949. "Wine making of the classical, Roman and Byzantine periods,"
Chemistry Journal 14A: 13-17.
Korber-Grohne, U. 1967. Geobotanische Untersuchungen auf der Feddersen Wierde. Wies-
baden: Steiner.
Krasilnikoff, J. A. 2002. "Water and farming in classical Greece: Evidence, method and
perspectives," in K. Ascani, V. Gabrielsen, K. Kvist, and A.H. Rasmussen (eds.),
Ancient history matters: Studies presented to Jens Erik Skydsaggard on his seven-
tieth birthday. Analecta Romana Instituti Danici Romae Suppl. 30. Rome: L'Erma
di Bretschneider, 47-61.
Littlejohn, L. 1946. "Some aspects of soil fertility in Cyprus," Journal of Experimental
Agriculture 14: 123-34.
Loukopoulos, D. 1938. Agricultural practices in central Greece. Athens: Folklore Library,
Sideris Editions.
Margaritis, E. E., and M. K. Jones 2006. "Beyond cereals: Crop-processing and Vitis vinifera
L: Ethnography, experiment and charred grape remains from Hellenistic Greece,"
Journal of Archaeological Science 33.6: 784-805.
Manning, W. H. 1964. "The plough in Roman Britain," Journal of Roman Studies 54: 54-65.
Michell, H. 1963. Economics of Ancient Greece. 2nd ed. Cambridge: Cambridge University
Press.
Morris, S., and J. Papadopoulos 2005. "Greek towers and slaves: An archaeology of ex-
ploitation," American Journal of Archaeology 109: 155-225.
Oleson, J.P. 2000a. "Irrigation," in Wikander 2000, 183-215.
Oleson, J.P. 2000b. "Water-lifting," in Wikander 2000, 217-302.
Palaiologos 1838. Vine cultivation and wine making. Athens: Royal Editions.
Palmer, R. 1994. Wine in the Mycenaean palace economy. Liege: Universite de Liege.
Pansiot, F. P., and H. Rebour 1961. Improvement in olive cultivation. Rome: F.A.O.
Rees, S. 1979. Agricultural implements in prehistoric and Roman Britain. British Archae-
ological Reports 69. Oxford: BAR.
Reinders, H. R., and W. Prummel 1998. "Transhumance in Hellenistic Thessaly," En-
vironmental Archaeology 3: 81-95.
Semple, E. C. 1932. The geography of the Mediterranean region: Its relation to ancient history.
London: Constable.
Van der Veen, M. 1992. Crop husbandry regimes: An archaeobotanical study of farming
in northern England, 1000 BC-AD 500. Sheffield Archaeological Monographs 3.
Sheffield: J. R. Collis.
Van der Veen, M. (ed.) 2005. Garden agriculture. Special issue, World Archaeology 37.2.
Wikander, b. (ed.) 2000. Handbook of ancient water technology. Leiden: Brill
White, K. D. 1970. Roman farming. London: Thames and Hudson.
Zachariasse, F. 2003 "Lowland farming: A comparative study of agricultural variation
and change in the Netherlands during the 1st millennium A.D." PhD diss., Cambridge
University.
Zivas, D. A. 2000 Pre-industrial agricultural buildings at the island of Zakynthos. Athens:
ETBA Publications.
CHAPTER 8
ANIMAL HUSBANDRY,
HUNTING, FISHING, AND
FISH PRODUCTION
GEOFFREY KRON
ARCHAEozooLOGICAL research offers impressive, long-neglected evidence for the

technical sophistication and productivity of Greco-Roman animal husbandry. His-
torians once portrayed ancient livestock farming as crippled by summer drought
and a chronic shortage of fodder (White 1970: 272; Frayn 1984), with farmers forced
to raise a limited number of animals predominantly for their wool or for work rather
than for meat (White 1970: 276--77). In fact, Greek and Roman farmers, graziers,
and shepherds supplied large urban populations with meat from an impressive
range of livestock, game, domestic fowl, birds, fish, and shellfish. Animal bones
reveal a significant increase in the size of most Greco-Roman domestic animal spe-
cies over those found on Bronze Age, Iron Age, or medieval sites (Peters 1998; Kron
2002; MacKinnon 2004), often reaching levels that were consistently found again
only in nineteenth-century Europe (Kron 2002: 63; cf. Moriceau 1999: 47). The writ-
ings of the Roman agronomists consistently reflect sound and intelligent princi-
ples of management, differing little from contemporary organic animal husbandry.
Moreover, evaluations of the writings of Greek and Roman veterinarians by their
modern counterparts show knowledge of most of the surgical procedures used on
livestock as late as the mid-twentieth century and of the properties of medicinal
herbs appropriate for many veterinary ailments.
A case can be made that the classical and Hellenistic Greeks should be credited
for many of the critical innovations in animal husbandry (Kron 2002: 65-68), game-
farming, and both fishing and fish-farming (Trotta 1996: 242-43; Colin-Bouffier
1999). The Greek origin of many of the most highly prized livestock breeds is well
known (Magerstedt 1859: 20-22, 94-98, 100-101; Ryder 1983: 147-50), and the Ro-
man agronomists made no secret of their reliance on methods pioneered by the
Greeks and Carthaginians (Varro, Rust. 3.3.6-7, 3.10.1; Columella, Rust. 8.1.3-4,
8.2.4, 8.2.13), citing a long list of now-lost works (Varro, Rust. 1.1.8-10; Columella,
Rust. 1.1.7-12; Georgoudi 1990: 65-72). The scattered references to animal husbandry
in Greek literary sources and inscriptions (Osborne 1987; Hodkinson 1988; Chan-
dezon 2003) likewise suggest that many of the methods of animal husbandry de-
scribed by Columella and Varro date back to the fifth and fourth centuries B.C., if
not earlier. This hypothesis is supported by the archaeozoological evidence of the
increased size of classical and Hellenistic Greek livestock (Benecke 1994a: 304; Kron
2002: 65-68), but definitive proof will remain elusive until more studies can be
carried out on post-Bronze Age Greek sites (Payne 1985; Reese 1994; Kotjabopoulou
et al. 2003). Improved Greco-Roman livestock did not entirely supplant smaller
breeds, some of which were still prized, presumably for their adaptation to the cli-
mate, hardiness, good wool, or high milk production (e.g., cevae from Altinum; cf.
Columella, Rust. 6.24.5). The transition to more intensive methods was likely less
complete in regions that were only gradually hellenized or romanized, and increases
in size varied by species.
The Greeks and Romans also developed sophisticated new techniques to im-
prove the capture, farming, or fattening of a large range of game, wild birds, and
fish. The farming of game and fish for sale in urban markets was particularly
sophisticated and intensive. Efficient commercial game farming of the sort described
by the Greeks and Romans and attested by archaeozoological studies would not be
revived in Europe or North America until the late twentieth century (Fletcher 1989:
325; Binder 1971). Roman aquaculture of the late Republic and early Principate was
even more impressive. Fish farmers invested large amounts of capital into building
elaborate fish-tank complexes for mariculture and experimented with some of the
most important techniques of modem fish farming.
ANIMAL HUSBANDRY
Greek and Roman farmers kept a wide variety of domesticated animals: cattle,
sheep, goats, pigs (figure 8.1), horses, donkeys, and mules, as well as dogs and cats,
chicken, ducks, geese, rock doves, and even camels (table 8.1). The best methods of
Greco-Roman animal husbandry were highly intensive, often matching the rela-
tively modem methods of improved husbandry introduced in Holland and Bra-
bant in the sixteenth and seventeenth centuries (de Vries 1974; Moriceau 1999: 47)
and England in the eighteenth and nineteenth (Trow-Smith 1959; Davis 1997). Less
intensive traditional methods continued, of course, and were attractive for farmers
ANIMAL HUSBANDRY, HUNTING, AND FISHING 177
Figure 8.1. Roman livestock. Bull (not to scale), sheep, and pig from relief showing
Tiberius carrying out the Suovetaurilia sacrifice, ca. A.D. 14, Rome. Louvre Museum.
(Alinari photo: 22685.)
who did not have access to large urban markets and needed to keep costs low.
As historians of technology, however, our primary interest is in state-of-the-art
techniques, and limitations of space make it impossible to describe the full spec-
trum of approaches in detail.
The innovations in Greco-Roman animal husbandry can be broken down into
four main areas: breeding, nutrition, housing, and health and veterinary care. In each
field, the classical agronomists strongly emphasize intensification of production-
e.g., the battery farming of poultry in dark confined cages (Varro, Rust. 3.9.19-20;
Columella, Rust. 8.7.1-3; Martial 13.62), so similar to modern factory farming (Peters
1998: 201; cf. Fox 1984: 14-18; 21-24)-and dose observation of the normal behavior
of the livestock being raised (Columella, Rust. 8.8-4-5, 10.5, 14.4-5; Varro, Rust. 3.7.6,
10.7).
Breeding
Although some have downplayed the use of selective breeding by Roman farmers
(e.g., White 1970: 272-75), the references in the agronomists show that they had
considerable expertise (Vigneron 1968: 30-31; Ryder 1983: 164; Bokonyi 1984: 21-22) .
They were not aware of the science of genetics, but neither were legendary eigh-
teenth-century English cattle breeders such as Bakewell or many of his successors,
who were responsible for setting many of the principal livestock breeds of today
(Pawson 1957). Even modern breeders of thoroughbred racehorses rely largely on
intuition and experience rather than genetics to make their decisions (Wynmalen
Table 8.1. The principal Greco-Roman domestic aninial species.
Common
name Scientific name References
Cattle Bos taurus Keller 1909: 329-71; Toynbee 1973: 149-62; Flach 1990: 290-96;
Benecke 1994a: 260-88; Peters 1998: 25-71; MacKinnon 2004:
76-99; King 2002: 408-10
Pig Sus scrofa Keller 1909: 388-404; Toynbee 1973: 131-36; Flach 1990: 311-15;
domesticus Benecke 1994a: 248-60; Peters 1998: 107-34; MacKinnon 2001:
138-62
Sheep Ovis aries Keller 1909: 309-28; Toynbee 1973: 163-64; Ryder 1983: 117-81;
Flach 1990: 301-9; Benecke 1994a: 228-38; Peters 1998: 71-106;
MacKinnon 2001: 100-37
Goat Capra hircus Keller 1909: 296-308; Toynbee 1973: 164-66; Flach 1990:
309-11; Benecke 1994a: 239-48; Peters 1998: 71-106; MacK-
innon 2001: 100-137; King 2002: 415-16
Horse Equus caballus Keller 1909: 218-58; Toynbee 1973: 167-85; Flach 1990: 297-99;
Benecke 1994a: 288-310; Peters 1998: 135-65; King 2002: 423-26
Ass Equus asinus Keller 1909: 259-70; Toynbee 1973: 192-97; Flach 1990: 296-97;
Benecke 1994a: 310-18; Peters 1998: 135-65; King 2002: 421-22
Mule Equus caballus Keller 1909: 259-70; Toynbee 1973: 185-92; Flach 1990:
X asinus 299-300; Benecke 1994a: 318-23; Peters 1998: 135-65; King
2002: 422-23
Camel Camelus Keller 1909: 275-76; Toynbee 1973: 141-42; Benecke 1994a:
dromedarius; 323-32; Peters 1998: 189-90; Gilbert 2002: 18-21
Camelus bactrianus
Dog Canis familiaris Keller 1909: 91-151; Benecke 1994a: 208-28; Peters 1998: 167-87;
King 2002: 410-14
Cat Pelis catus Keller 1909: 64-80; Benecke 1994a: 344-52; Peters 1998: 187-88;
King 2002: 426-27
Chicken Gallus gallus Keller 1913: 131-45; Thompson 1936: 33-44; Andre 1967:
domesticus 130-32; Zeuner 1963: 449-55; Toynbee 1973: 256-57; Flach 1990:
315-20; Benecke 1994a: 362-73; Watson 2002: 380-81; Benecke
et al. 2003: 78-79
Greylag Anser anser Keller 1913: 220-26; Thompson 1936: 325-30; Andre 1967:
goose domesticus 132-33; Zeuner 1963: 466-70; Toynbee 1973: 261-64; Salza Prina
Ricotti 1987: 95-6; Flach 1990: 322-23; Benecke 1994a: 373-79;
Olson and Sens 2000: 213-14; Watson 2002: 365-66; Benecke et
al. 2003: 79
Mallard Anas platyrhynchos Keller 1913: 226-35; Thompson 1936: 205-6; Andre 1967: 133;
duck domesticus Zeuner 1963: 470-71; Toynbee 1973: 264-73; Salza Prina Ricotti
1987: 95; Benecke 1994: 379-83; Watson 2002: 364-65
Rock dove Columba livia Keller 1913: 122-31; Thompson 1936: 238-46; Andre 1967: 124;
Zeuner 1963: 460-62; Toynbee 1973: 258-59; Pollard 1977: 104;
Salza Prina Ricotti 1987: 96; Flach 1990: 320-22; Benecke 1994:
383-90; Watson 2002: 372-74; Benecke 2003: 79-80
1950: 88, 99). Like their Greek and Roman predecessors (White 1970: 328-29;
Columella, Rust. 6.29.2-3, 36.1-5; 7.2.5-3.2; Varro, Rust. 2.2.3-5, 7.5), they still rely
heavily on points (Wynmalen 1950: 96-107), and their advice often broadly cor-
roborates the recommendations of the ancients (White 1970: 185-86; Peters 1998:
28-29, 77-78, 199). Modern historians of horse breeding often single out Xeno-
phon's explanation of the points of the horse (Eq. 1.9-10), like much of his advice,
as exemplary (Vigneron 1968: 5-9; Goodall 1977: 117).
As Vigneron points out (1968: 35-38), the ancient agronomists had ably resolved
all the technical problems involved in breeding and raising the young of that most
delicate of domestic animals, the horse, and the same care was given to breeding the
other major domestic animals. For example, Columella advised farmers not only
to use the best breeding stock, but also to use only sexually mature animals in their
prime (Rust. 7.3.6, 6.3, 6.8, 9.2); to control the breeding process (Rust. 6.37.1-2, 8-10;
7.3.11-16, 6.6-8); to provide prompt veterinary help in cases of difficult births (Rust.
7.3.16); and to show great care for the newborn young (Rust. 7.3.16-9; cf. Ryder
1983: 685-87). Unfortunately, the agronomists only occasionally allude to the more
sophisticated methods used by breeders to consciously manipulate and improve
their livestock. They allude to crossbreeding with new breeds and even wild species
(Columella, Rust. 6.37.3ff; 7.2.4-5) and show an awareness of recessive traits that had
to be guarded against (Columella, Rust. 7.3.1-2; Vergil, G. 3.355; cf. Ryder 1983: 143,
164), but the skill of Greek and Roman breeders must be inferred primarily from
their results. They were very successful, for example, in breeding sheep with a wide
range of fleece types. These included several distinct breeds with fine wool, com-
parable to some of the finest modern merino wool (Ryder and Hedges 1973; Ryder
1983: 154-55); as well as true medium wool (long wool) and short wool, both of which
had long been credited to medieval breeders (Ryder 1983: 177-80).
An important stimulus to classical animal breeding came from its vigorous
commercial development (Rinkewitz 1984: 21-23). There is circumstantial evidence
of a significant market not only in breeding stock, but also presumably in stud ser-
vices, as pioneered in eighteenth-century England by Bakewell (Pawson 1957: 70),
whose breeding bulls (worth only about £8 as beef) could command stud fees of £152
for a four-month season. Certainly, the extremely high prices obtained by certain
Roman breeding animals imply a hierarchy of breeders, some public knowledge of
pedigrees and performance, and the potential to recoup one's investment through
fees as well as sales of one's own livestock. To cite just one example, a pair of fine-
quality pedigree pigeons, presumably a monogamous pair proven to be compati-
ble and to perform well as breeders (Naether 1964: 155-56), could command 1,000
HS or even 1,600 HS (Varro, Rust. 3.7.10; cf. Columella, Rust. 8.8.9-10). The vigorous
trade and broad contacts of the Greeks and Romans were also instrumental in
spreading new breeds of livestock and innovative methods throughout the Medi-
terranean and, with the Roman conquest, western Europe (Columella, Rust. 6.1.1-2,
7.2.3-4, 8.2.4-8; Benecke 1994a: 304-5; Anderson 1961: 15-39). Fine-wooled sheep
breeds from Attica and Asia Minor spawned new fine-wooled breeds in Tarentum,
Apulia, Parma, Pollentia, and Mutina (Columella, Rust. 7.2.3-4; Magerstedt 1859: 2:
95-98), and large long-horned cattle from Italy were crossed with small short-
horned Celtic breeds in Pannonia, Germany, and France (Bokonyi 1984: 21-24; but
cf. Kokabi 1988: 475). The larger and more diverse flocks and herds of the Greeks
and Romans are reflected in the improved genetic health oflivestock, as is indicated,
for example, by far fewer teeth abnormalities in Roman-era cattle than in the Celts'
herds (Peters 1998: 70).
Nutrition
Although improved breeding most likely played a significant part in increasing the
size of Greco-Roman livestock, any gains would prove transient without continued
improved nutrition, as the marked decline in the size of Roman cattle breeds over
the course of the Middle Ages shows (Riedel 1994; Forest and Rodet-Belarbi 2000).
Improved Greek or Roman cattle typically showed gains of as much as 20 percent
in height at the withers over Bronze and Iron Age as well as medieval cattle (Kron
2002; MacKinnon 2004: 85, table 24; Sternberg 2005b), and significant increases in
bone thickness, indicative of sharply increased weight (Forest and Rodet-Belarbi
2002). Studies show increases in withers heights of 10 or even 20 percent in some
improved sheep breeds (Lepetz 1996: 100; Peters 1998: 94-98; MacKinnon 2004:
104-5), and large sheep appear in Greece from at least the fifth century B.C.
(Leguilloux 1999). Although the remains of horses are much less common (Peters
1998: 148-49, 164-65), and likely exclude the "noble" breed of race-horses (Colu-
mella, Rust. 6.27.1-2), they too show a consistent increase in withers heights, one
that was already well advanced in classical Greece (Peters 1998: 149-52; de Grossi
Mazzorin et al. 1998; Chiliardi 2000). The increases in withers heights of improved
Greek or Roman pigs are often more modest (Peters 1998: 124-25; MacKinnon
2004: 148, table 44), but there is a more marked increase in bone thickness, indic-
ative of greater weight (Peters 1998: 126; cf. 125, tables 24-25). We also see the ap-
pearance of a short-snouted, white, smooth, quick-fattening breed (Peters 1998:
124, 126; MacKinnon 2001, 2004: 150) similar to the Chinese breeds introduced by
eighteenth-century European pig breeders (Zeuner 1963: 268-69) alongside the
more bristly and feral, slow-fattening pig phenotype (Peters 1998: no). The Ro-
mans, presumably adapting the methods of the renowned Greek breeders (Colu-
mella, Rust. 8.2.4), raised chickens in significant numbers (Lauwerier 1983; Benecke
1994b: 114-16; Lauwerier 1993) and to a high standard (Peters 1998: 197-213). Ar-
chaeozoological studies demonstrate good health (Brothwell 1997) and a dramatic
improvement in size, with Roman hens clustering in the middle range of modern
breeds at about 2 kg, significantly heavier than the hens of 1 to 1.5 kg occasionally
found on Celtic sites (Bokonyi 1984: 93-94; Brothwell 1997; Peters 1998: 222-26).
Improved nutrition is not only critical in optimizing the growth and health of
livestock but also in maintaining a high level of reproduction (Lemming 1969 ), and
evidence from both Greek and Roman sources, as well as from animal bones,
suggests that ancient domestic animals were extremely fecund and thus must have
been well fed. Aristotle claims that Greek domestic sheep often bore twins and
sometimes had as many as three or four lambs (Hist. an. 593b 19-20). This is a very
high reproductive rate (Ryder 1983: 145), significantly superior to most medieval
and many early modern sheep, of which only the best fed, improved breeds reg-
ularly bore twins, and many failed to lamb altogether (Ryder 1983: 448,496). Greco-
Roman pigs were equally prolific (Aristotle, Hist. an. 573a 30; Columella, Rust. 7.9.13;
Peters 1998: 114), and Columella makes it clear that Roman intensive pig farmers
induced their sows to farrow twice a year (Columella, Rust. 7.9.4; MacKinnon 2004:
150-51), as has been confirmed archaeozoologically (Ervynck and Dobney 2002).
Roman villas specializing in pig production will therefore have produced a great
deal of pork. Bokonyi estimates the yield on one of the less productive sites at over
330 percent of the breeding stock, even on the conservative assumption of just one
farrowing (Bokonyi 1988: 174). Like modern livestock farmers, the ancient agron-
omists were very conscious of the value of high-protein feed for breeding and
nursing animals (Columella, Rust. 8.5.1-2; cf. Farrington 1913: 56). While they did
not have access to some of the industrial processed feeds or the additives and
antibiotics used today, such artificial feeds can cause serious behavioral and health
complications from lack of roughage (Fox 1984: 161-63; Houpt 2005: 98-100 ), and
antibiotics are generally unnecessary if management methods are sound (Fox 1984:
76-77).
The Greeks and Romans improved animal nutrition partly through new fodder
sources and improved fodder management, but mostly through greater integration
of livestock into arable farming. The most intensive system of animal husbandry,
and the dominant method of mixed farming today, is convertible husbandry-ley
farming-in which part of the land is seeded with artificial grass, generally for
several years, and used as pasture or meadow, while the rest is cropped continu-
ously with a rotation of cereals and leguminous fodder crops. Agrarian historians
have plausibly identified this system as the single most important factor in the ag-
ricultural revolution in modern western Europe (Kerridge 1967). Regarding the
ancient world, several historians, most notably Hodkinson (Hodkinson 1988), have
acknowledged the importance of mixed livestock farming and crop rotations in
Greek and Roman animal husbandry, while denying that the ancients used con-
vertible husbandry (Hodkinson 1988: 50-51; Pleket 1993: 324 n. 8, citing Sallares
1991). A careful reading of the Roman agronomists, however, seems to prove that
they fully understood and applied the principles (Kron 2000; cf. Carandini 1984).
Convertible husbandry offers two key advantages. The incorporation of le-
gumes in the continuous crop rotation ensures a large quantity of highly nutritious
fodder for livestock, supplementing the pasture during winter dormancy or sum-
mer drought, and the artificial leys provide much better grazing or hay produc-
tion than permanent pasture. One can therefore feed more livestock better, while
producing more manure and achieving higher yields for cereals, as well as (often) a
surplus of quality fodder (Kron 2004b: 312). The classical Greeks had already iden-
tified many of the best leguminous fodder crops for arable cultivation (Hodkinson
1988: 44-45). With Varro's and Columella's careful attention to Hellenistic (and
Carthaginian) agronomy, the inventory of Roman fodder crops expanded to in-

clude most of the species native to western Europe that were used during the past
century (White 1970: 213-19; Kron 2004b: 276-77). Planting artificial leys demands
awareness of the best forage plants, and the Roman literary sources (Kron 2004b:
276-81) and studies of the remains of carbonized hay (Kron 2004b: 281-307) suggest
that an impressive array of excellent forage legumes were exploited by Greek and
Roman farmers and graziers. Some of the most notable included alfalfa (Medicago
sativa)-considered the best forage by ancient and contemporary experts alike-
first imported into Greece from Persia in the fifth century B.C. (Georgoudi 1990:
171; Kron 2004b: 278-79), tree medick (Medicago arborea), a highly nutritious shrub
ideally suited for sheep and milk production, publicized by the fourth-century B.C.
Athenian agronomist Amphilochus (Pliny, HN 13,130; Hodkinson 1988: 45; Kron
2004b: 279-80), subterranean clover, one of the best drought-resistant forage le-
gumes for seeding in pasture (Kron 2004b: 297-98), and a significant number of the
best clover and medic species (Kron 2004b: 298-300), as well as such proven fodder
species as vetch, yellow serradella, vetchling, and lupines.
There was clearly a strong preference on the part of the agronomists in favor
oflabor-intensive mixed farming, but mountainous terrain, macchia, and marginal
rangelands covered a great deal of Italy and Greece, and many livestock were cer-
tainly pastured extensively. While there is little scholarly consensus, it is likely that
sedentary and transhumant pastoralism, often integrated with arable agriculture
(Rosada 2000: 107-11), played a significant role in Greco-Roman animal husbandry
(Pasquinucci 2004), albeit a much more modest one than in the medieval Mezzo-
giorno (Brun 1996; Chandezon 2003: 391-97; Kron 2004a). The technical expertise
that the Greeks and Romans developed from mixed farming also helped to increase
the productivity of permanent pasture and rangelands. With effective management,
ranching can be ecologically sound and highly productive (Heady and Childs 1994).
The Roman agronomists give excellent advice for the management of range-
lands, very much in accord with modern research (Kron 2004b: 311-17). They
recommend the most effective technique for entirely renewing and improving worn
out rangeland: to plow it up, carry out a brief arable rotation, and then reseed the
land with leguminous forage plants (Kron 2004b: 315; cf. Pawson 1957: 79-80 ). Such
methods have proven highly effective in modern studies, increasing forage yield
from two to five times, and we have evidence that saltus were indeed plowed for the
sake of improving the pasture or cultivated periodically (Lirb 1993: 272 n. 10; Festus
p. 392, 33 Lindsay). The most effective, albeit expensive, way of boosting forage
production is to irrigate meadows (Heath et al. 1973: 627-93; Columella, Rust. 2.17.1;
Pliny, HN 18.258-63). Although such water meadows typically were created on
arable farms as a source of extra hay, the Romans made very effective use of aq-
ueducts for irrigation (White 1970: 146-72), and some very large tracts were irri-
gated, as has been confirmed archaeologically (Quilici-Gigli 1989). Other valuable
Roman recommendations for improving the productivity of pasture (Kron 2004b:
311-17) included protecting the range from overgrazing and physical damage
through controlled grazing (cf. Heady and Childs 1994: 248-50 ); seeding it with
forage legumes; fertilizing it, particularly with potash or lime; and periodic burning
(Corbier 1999: cf. Heady and Childs 1994: 333-35).
Housing
The Roman agronomists show considerable care in providing appropriate housing
for their livestock, reflecting once again their capital- and labor-intensive approach
to animal husbandry. Modern research suggests that the provision of shelter gen-
erally improves the health of livestock and their reproductive and growth perfor-
mance (Fox 1984: 58-70, 86-87, 91-92, 106, 108), provided they are not confined in
overcrowded conditions, where lack of exercise and stress cause many behavioral
and health problems (Fox 1984: 75-77, 151-61). Few excavators have shown interest
in investigating purely utilitarian farm buildings, but we can glean enough infor-
mation from the agronomists and archaeological evidence to get some idea of the
design of many shelters (Rinkewitz 1984: 27-29; Carandini 1984: 160-62; Flach 1990:
227-45). The discovery and exemplary study (Badan et al. 1996) of approximately
130 Roman sheepfolds on the famous sheep pastures of the Crau (Pliny, HN 21.57;
Conges 1997; Leveau and Segard 2003; Leveau 2004) in southern France show the
large investments made by Roman sheep farmers or perhaps landowners (Corbier
1999) eager to maximize the productivity of even transhumant sheep raising. Ex-
tremely large-one example at Negreiron-Negres covers 288 square meters, and
most are capable of holding from 700 to 900 sheep-they are remarkably similar
in their design and construction to the sheepfolds first built on the Crau in the
twentieth century (figure 8.2; Badan et al. 1996: 290-95, figs. 21-23).
While housing for sheep and cattle is generally relatively simple (Fox 1984: 88;
Ryder 1983: 682-85), the success of intensive pig and chicken farming depends more
on the design of their sties and coops. The recommendations of the Roman au-
thorities for pigsties (Columella, Rust. 7.9.9-10; Varro, Rust. 2-4-13-15), several ex-
amples of which have been excavated (Morris 1979: 52; Carandini 1984: 182-84),
represent a good illustration of the sophistication of their methods. The size of the
stalls; the inclusion of hay or litter; the clever design to keep nursing piglets seg-
regated while permitting the sows to leave; the attention to cleanliness; the provi-
sion of clean food and water and the care taken to prevent sows from crushing their
young-all correspond closely to modern recommendations for making pig farm-
ing more productive and minimizing the stress and high mortality common in
many modern confinement systems (Fox 1984: 58-60, 65-70, 75).
Health and Veterinary Care

We can see further evidence of the intensive nature of Greco-Roman animal
husbandry in the high level of veterinary care available. Studies of animal bones not
only suggest good nutrition, as we have seen, but also show low levels of disease
and generally good health (Peters 1998: 69-71, 86-87, 133-34; MacKinnon 2004:
~ , - -·
Figure 8.2. Roman sheepfold at Negeirion-Negres, La Crau plain, Proven~e, reconstruc-

tion. Dimensions: 46.3 x 9.55 m. (From Badan et al. 1996: 294, fig. 25.)
96-97, 148-49, 159) and even reveal direct evidence of veterinary intervention
(Udrescu and Van Neer 2005). The agronomists paid careful attention to the be-
havioral and hygienic needs oflivestock (Columella, Rust. 6.23.1-3; 7.3.8, 9.14; Varro,
Rust. 2.2.7; cf. Senet 1953: 26, 80-81) . They also gave detailed, generally sound,
prophylactic and veterinary advice (Peters 1998: 36-38, 85-89, 365-68), which was
widely disseminated for the use of farmers and their slaves, shepherds, or tenants
(Varro, Rust. 2.2.20, 5.18; Georgoudi 1990: 80-81; Adams 1995: 72-79). Good nutri-
tion and management as well as simple prophylactic measures helped to prevent
disease. For example, as in the recent past (Ryder 1983: 708, 781), farmers controlled
scab by dipping sheep using medicinal ingredients of proven effectiveness against
insects and parasites (Peters 1998: 87-88; Frizell 2004; cf. Ryder 1983: 708-9). Varro
also gave sound advice to protect sheep from possible sunstroke and to avoid
pasturing them on dewy grass, a possible cause of foot rot (Varro, Rust. 2.2.7; cf.
MacKinnon 2004: 114) and a breeding ground of the larvae of Haemonchus contor-
tus or the barber pole worm, a dangerous parasite (Peters 1998: 79, 85-86).
Farmers in classical and Hellenistic Greece and the late Republic could also call
on skilled professional veterinarians (Adams 1995: 51-65, 72-102), provided, of
course, such treatment was cost-effective. Several extant treatises on veterinary
medicine (see Moule 1891: 9-50; Georgoudi 1990: 60-63; Peters 1998: 135) give a
good idea of their expertise, although they arguably underrepresent the treatment
of the other domestic animals in favor of horses (Senet 1953: 78-79; cf. Columella,
Rust. 7.3.5-6). The study of anatomy was very sophisticated (Adam 1995: 362-63),
based on dissection and even the performance of post-mortem analysis (Adams
1995: 48 n. 155). The Greek veterinary writers formulated a large technical vocab-
ulary that covered both anatomical features and many of the diseases identified by
modern veterinarians (Adams 1995: 239-331), and the symptoms of these diseases
were accurately described (Leclainche 1936: 9-96; Adams 1995: 49-50).
The most impressive achievements of the practitioners of classical veterinary
medicine lay in their surgical expertise. The ancient manuals describe a remarkable
number of the most common surgical procedures performed by twentieth-century
veterinary surgeons, including virtually all of the standard obstetrical operations
and even eye surgery (Moule 1891: 147-66; Senet 1953: 43, 83-86; Bourdy 1995; Peters
1998: 209-10). Veterinarians used many different specialized surgical instruments,
most remarkably similar to modern instruments, and even a special machine de-
signed to restrain animals during surgery (Moule 1891: 147-50; Senet 1953: 83-84,
86). Medicines prescribed for veterinary use were largely herbal or mineral (Moule
1891: 174-77), and frequently demonstrate an accurate understanding of a number
of tannins and anti-inflammatory, antipyretic, diuretic, or analgesic plants (Bourdy
1995: 207-8; Peters 1998: 209-10). Although Greco-Roman veterinarians are often
criticized for their ignorance of effective treatments of bacterial and viral infections
(Senet 1953: 25-26), the scientific knowledge necessary to achieve workable cures was
unknown to modern veterinary science prior to the revolutionary discoveries of
Koch and Pasteur in the late 1870s and 1880s (Pattison 1984: 77-78, 87-89). The
ancient veterinarians recommended the quarantine or slaughter of infected live-
stock (Columella, Rust. 6.5.1), which was the most rational approach in the cir-
cumstances. Modern veterinarians ultimately gave the same advice, despite popular
resistance (Pattison 1984: 60-61).
HUNTING, FOWLING,
AND GAME FARMING
Ancient hunting has generated many studies, including a number of substantial

monographs (Aymard 1951; Hull 1964; Anderson 1985), but we will focus here
primarily on its economic function as a source of meat and secondary products. In
Europe during the Middle Ages and the ancien regime, only the nobility and large
landowners enjoyed hunting rights, and poachers were subject to draconian pun-
ishments including castration, blinding, mutilation, and death (Trench 1967: 26;
114-21, 128-29; E. P. Thompson 1975)-although these did not prevent poaching

or the virtual extinction of the red deer in England (Trench 1967: 130; Grant 1988:
139). Greco-Roman hunting, on the other hand, was not restricted to a social elite.
Roman law granted ownership of game and wild birds to the hunters who captured
them, regardless of their social status or the ownership of the land on which they
hunted (Dig. 41.1.pr-4), although Roman law did acknowledge the ownership of
wild animals raised on game farms (Dig. 41.1.3.2, 1.5-6). Most Greek and Ro-
man hunters therefore practiced the "petit bourgeois" method of hunting on foot
(Aymard 1951: 373), employing the relatively inexpensive and efficient methods
employed by professional hunters who captured game for sale (Arrian, Cyn. 3-1;
Rinkewitz 1984: 103-4; Bortuzzo 1990: 100 ).
This traditional Greco-Roman hunt employed nets (Aymard 1951: 207-18; Hull
1964: 10-18), foot snares (Hull 1964: 18-19), and the formido or scare (Aymard
1951: 218-28; Hull 1964: 19), along with highly trained foxhounds and tracking dogs
(table 8.2) to trap the quarry. Hunters on foot would then dispatch the game, using
spears (Hull 1964: 5-6), javelins (Hull 1964: 6), longbows (Hull 1964: 7-8), cross-
bows (Baatz 1991), swords (Hull 1964: 6; Aymard 1951: 316-17), or the specialized
boar spears (Aymard 1951: 5-6) in the highly hazardous but prestigious boar hunt
(Aymard 1951: 310-16, 323-30; Hull 1964: 103-5). As illustrated in the classic ac-
count in Xenophon's Cynegetica, classical Greek hunting techniques were already
very highly refined, based on the selection and training of appropriate hunting dog
breeds (Aymard 1951: 246-54, 369-71; Hull 1964: 39-58) and close scientific ob-
servation of the behavior of the game (Aymard 1951: 366; Hull 1964: 60-75).
The most important Greco-Roman innovation increasing the supply of game
was surely game farming, which reached a level of sophistication not matched until
the late twentieth century. Carthaginian and Hellenistic agronomists were already
adept at raising wild fowl and game (Varro, Rust. 3.2.13; Peters 1998: 197), coining
the term theriotropheion to describe these farms (Hudson 1989: 22-23), and Co-
lumeau's analysis of the age and sex distribution of bones from Kassope (2000: 155)
reveals that the Greeks farmed red deer in the classical and Hellenistic periods.
Other sites identified as supplying game to urban markets include the Roman villa
at Neftebach in Switzerland (Olive and Deschler-Erb 1999). Here again, many deer
were slaughtered as subadults, which is consistent with game farming (Columella,
Rust. 9.1.7; cf. Fletcher 1989: 331) rather than hunting.
By the first century B.C., game farming had become a very profitable business
in Italy (Rinkewitz 1984, 21-23 and passim; Bortuzzo 1990 ). Such prominent Roman
aristocrats as L. Licinius Lucullus (Rinkewitz 1984: 81-82, no. 17) and Q. Hortensius
(Rinkewitz 1984: 80, no. 14) participated enthusiastically, and the most up-to-date
techniques were widely disseminated in agronomic treatises (Varro, Rust. 3.2.13 and
passim; Columella, Rust. 9.1.1-9). There was a healthy market among Romans eager
to incorporate new game species into their diet (table 8.3). Large game were raised
in a forest habitat, enclosed with walls of rough stone and mortar or mud brick, or
wooden post fences, which represented one of the greatest expenses (Columella,
Rust. 9.1.2-5), as in modern game farming (Hudson 1989: 17; Fletcher 1989: 327). The
Table 8.2. Greco-Roman hunting dog breeds.
Hunting Dog Breed Ancient sources Modern authorities
Iberian Opp., Cyn. 1.371; Poll. 5.37;

Nemes., Cyn. 228
Indian Arist., Hist. an. 8.23; Xen., Cyn. Keller 1909: 109-10; Aymard 1951:
9.1; Hdt. 1.192; Diod. Sic. 17.92; 244-45; Toynbee 1973: 103; Peters
Poll. 5.37-38; Ael., NA. 4.19; 8.1 1998: 171
Molossian Arist., Hist. an. 9.3; Mart. 12.1; Keller 1909: 103-7; 111-13; Aymard
(Epirus) Nemes., Cyn. 107ff.; Ael., NA 3.2, 1951: 251-54; Hull 1964: 29-30;
10.1; Lucan,, Phars. 4.402; Luer. Toynbee 1973: 103; Peters 1998: 171,
5.1076; Verg., G 3-404-5; Ath. Abb. 63
201 B; Opp., Cyn. 1.375; Grat.,
Cyn. 181, 197; Hor., Epod. 6.5,
Sat. 2.6.19; Poll. 5.37, 39
Cretan Xen., Cyn. 10.1; Grat., Cyn. 212; Keller 1909: 117-18; Aymard 1951:
Arr., Cyn. 2.5, 3-4; Claud., Cons. 246-51; Hull 1964: 34; Toynbee
Hon. Cons. Stil. 3.300; Poll. 1973: 103; Peters 1998: 172
5.4of.; Philostr., VA 8.30.2
Vertragus Arr., Cyn. 3.6, 5.7f.; Grat., Keller 1909: 101-3; Aymard 1951:
(Gaul) Cyn. 203 265-66; Hull 1964: 24-26; Toynbee
1973: 104; Peters 1998: 172
Sicilian Keller 1909: 125-26; Peters 1998: 171
Laconian Xen., Cyn. 3.1, 4.1; 10.1; Varro, Keller 1909: 120-24; Aymard 1951:
Rust. 2.5.5; Grat., Cyn. 212; 2354-57; Hull 1964: 31-33; Toynbee
Oppian, Cyn. 1.372; Arist., Hist. 1973: 103-4; Peters 1998: 171
an. 6.20, 8.28, 9.1; Ael. 8.2; Verg.,
G 3.403; Pl., Prm. 128 B; Soph. Aj. 5
Molossian-Epirote Arist., Hist. an. 9.3 Keller 1909: 103
sheepdog
Tuscan Opp., Cyn. 1.396; Nemes., Cyn. Aymard 1951: 261-62; Hull 1964: 24
231-37
Umbrian Verg., Aen. 12.753; Seneca, Thy Keller 1909: 124-25; Aymard 1951:
497ff.; Grat., Cyn. 171-73; 194 263-64; Hull 1964: 24; Toynbee
1973: 104; Peters 1998: 172-73
Segusius Grat., Cyn. 171; Catull. 42.9 Keller 1909: 102-3; Aymard 1951:
267-68; Hull 1964: 24; Peters 1998:
173
Scottish (Britain) Strabo 4.5.2; Nemes., Cyn. 225 Aymard 1951: 268-70; Toynbee
1973: 104-5; Peters 1998: 173
Agassaeans Strabo 4.5.2; Opp., Cyn. Aymard 1951: 268-70; Hull 1964:
(Britain) 1-468-78; Claud., Cons. Hons. 26; Toynbee 1973: 104-5; Peters
Cons. Still. 3.301 1998: 173 and Abb. 57c
Tibetan Grat., Cyn. 159 Keller 1909: 108-9; Hull 1964: 27
Carian Ael., NA 7.38, VH 14.46; Arr., Hull 1964: 28; Peters 1998: 171
Cyn. 3.1ff.; Dio Chrys. 1.371, 373,
396; Poll. 5.37, 47
natural sources of food in the forest could generally cover most of the nutritional
needs of the game, but supplemental feeding, often by hand, was used in winter,
when mast was scarce, as well as when the animals were breeding or rearing young
(Columella, Rust. 9.1.6-8), as in modern practice (Fletcher 1989: 326-27, 329). The
performance of modern game farming shows the dramatic improvements in pro-
ductivity that Greco-Roman game farming could have achieved. For example, 100
wild Scottish red deer hinds will generally produce 40 to 46 calves, only 30 of which
will live to a year, but with good management a similar number of farmed deer can
generally raise up to 90 calves to weaning. Likewise, shelter and adequate feeding
can increase the live weight of wild red deer stags from 120 kg to more than 185 kg
(Fletcher 1989: 328-29).
Like poultry and pork, game was a prestigious food and appears to have been
most popular among the well off and those who adopted Roman cultural food
habits (Lehman and Breuer 1997: 490-92; Olive and Deschler-Erb 1999), but its
consumption was by no means restricted to an elite. Ordinary Greeks and Ro-
mans were free to hunt, of course, but they could also take advantage of the ample
supply of farmed game meat sold in many local markets, where it was not much
more expensive than many ordinary domestic animals. Note, for example, that red
deer and roe deer sold for 12 denarii per Roman pound according to Diocletian's
price edict (4.44-5), only 50 percent more than beef or mutton and equal to pork.
Moreover, while game did not often exceed s to 10 percent of the meat consumed
on Italian sites, game meat was common and rarely fell below about 1 to 3 percent of
total meat consumption on sites where it was found (McKinnon 2004: 228-29, App.
13). Game was also widely consumed in the transalpine Roman provinces (Lepetz
1996: 225-26, 228; Peters 1998: 241, 246-48), particularly on Roman military sites, as
one would expect (Aymard 1951: 469-81; Davies 1971; King 1984; Epplett 2001), and
in luxury villas, vici, and civil settlements. With the exception of western Euro-
pean countries such as France and Italy (Ministry of Agriculture and Forestry, New
Zealand 1994; cf. Binder 1971), game consumption today is generally much lower,
constituting only about 0.5 percent of total meat consumption worldwide in 1978,
falling to barely half as much in 1984 (Hudson 1989: 46).
The art of fowling was very highly developed by the Greeks and Romans, and
the fattening of birds captured in the wild-or raised from eggs gathered from the
nests of wild birds (Lindner 1973: 96)-was a significant feature of Roman pastio
villatica. Our literary sources often allude to the practice and methods of fowling
(Lindner 1973: 15-28), and a number of mosaics, frescoes, sculptures, lamps, and
gems illustrate it in detail (Lindner 1973: 29-77, figs. 1, 15). Nets (Capponi 1959: 729;
Lindner 1973: 92-93; Cox 1686: 104-6) and snares, particularly for birds such as the
plover or woodcock (Cox 1686: 111-13), were effective tools, widely used by pro-
fessional fowlers. But most ancient literary and artistic depictions emphasize the
capture of birds using specially made reeds topped by twigs coated with birdlime, a
viscous substance manufactured from myrtle berries and oil (Pliny, HN 9.94.248;
Lindner 1973: 95). Capponi gives a particularly lucid analysis of the method, based
al Greco-Roman game species with references to ancient and modern consumption
l by their prominence in archaeozoological excavations).
Scientific name Ancient References References Modern Refen
Cervus elaphus Apicius 8.2.1, 3-8; Varro, Rust. Keller 1909: 277-79; Aymard 1951: Montagne 196
3.13.3; Columella, Rust. 9.1.1; Celsus, 16-17, 331-69; Toynbee 1973: 143-45; Klein 1991; Ha
Med. 2.18.2; Galen, Viet. Att. 4.664; Andre 1981: 119-20; Hull 1964: 66-72
Plin., HN 8.115, 119; 10.182; Verg., G. 76-81; King 2002: 417-19
1.307; Mart. 1.49, 13.94; Justin., Dig.
9.2.28.pr; 41.1.5.5; SHA, Heliogab. 8.3
Capreolus capreolus Apicius 8.3.1-3; Columella, Rust. Keller 1909: 277-79; Aymard 1951: Montagne 196
9.1.1; Varro, Rust. 3.3.3, 13.3; Celsus, 17-18; 331-50; Andre 1981: 120; Hull Catta 2002: 72
Med. 2.18.2; Juv. 11.142; Auson. 18; 1964: 82-83; Toynbee 1973: 143-45;
Verg., G. 2.374 King 2002: 416-17
Lepus europaeus Verg., G. 1.308; Mart. 1.49.25; 3.27; Keller 1909: 210-17; Aymard 1951: Montagne 196
13.92; Varro, Rust. 3.3.2; 3.12.1-6; 364-76; Hull 1964: 59-67; Toynbee Catta 2002: 17
Columella, Rust. 9.1.8; Mart. 13.92; 1973: 200-202; Andre 1981: 118-19;
Hor., Sat. 2.4.44, 8.89; Epod. 2.35; Benecke 1994b: 356-62; Olson and
Apicius 8.8.1-3; Juv. 11.139; Petr., Sens 2000, 207-8; King 2002: 430-32
Satyr. 56.9; Ar., Ach. mo; Eq.
1192-93; 1199; Vesp. 709; Pax 1150,
1196; Ee. 843; Telecl., fr. 34; Antiph.,
fr. 131; Alex., fr. 168
Sus scrofa scrofa Apicius 1.8.1-9; Columella, Rust. Keller 1909: 389-93; Aymard 1951: Montagne 196
9.1.1; Hor., Sat. 2.2.89; Pliny, HN 13-17; 299-316; Toynbee 1973: Catta 2002: 6c
8.210; 10.182; Juv. 1.140-41; Mart. 131-36; Andre 1981: 118-19; King
14,221.2; Plin., Ep. 1.6.1; Mart. 1.43, 2002: 443-45
49, 3.77, 7.27.1, 10.45.3-4; Hor., Sat.
2.4.40, 42, 2.8.6; Carm. 1.1.28; Stat.,
Silv. 4.6.1; Petron., Sat. 40.3
1,1ed)
Scientific name Ancient References References Modern Refer«
Vulpes vulpes Keller 1909: 88; Hull 1964: 96-97; Montagne 19t
Toynbee 1973: 102; King 2002: 446
Ursus arctos Plut., Quaest. Nat. 22; Tert. Apol. Keller 1909: 175-81; Aymard 1951: Montagne 19t
9.11; Petron., Sat. 66.5; Justin., Dig. 12-13; Hull 1964: 94-95; Toynbee
9.2.28.pr 1973: 93-100; King 2002: 445-46
Bos primigenius Plin., HN 8.38; Sen., Phaed. 64 Keller 1909: 342-43; Hull 1964: Montagne 19t
85-86, 89; Toynbee 1973: 148-49
Alces alces Caes., B Gal 6.27.1; Plin. 8.39 Keller 1909: 281-83; Zeuner 1963: Montagne 19t
425-29; Toynbee 1973: 145; King 3 no. 6
2002: 408
Oryctolagus cuniculus Catull. 37; 18; Plin., HN 8.217; 226; Keller 1909: 217-18; Zeuner 1963: Montagne 19t
Strabo 3.2.6, 5.2; Ael., NA 13,15; 409-15; Hull 1964: 59-67; Toynbee Catta 2002: 17
Polyb. 12.3; Varro, Rust. 3.12.6-7; 1973: 202-3; King 2002: 436-37
Apicius 2.2.6; Mart. 13.60
Capra ibex Keller 1909: 299-301; Aymard 1951: Montagne 19t
19-20; Toynbee 1973: 147 2002: 92-95
Dama dama Columella, Rust. 7.12.8, 9.1.1; Mart. Keller 1909: 277-79; Aymard 1951: Montagne 19t
1.49, 3.58, 4.74, 13.94; Juv. 11.121; 18-19; Andre 1981: 120; Zeuner 1963: no. 5; Boisaut
Plin., HN 8.214; Verg., G 1.308, 3.410 429-33; Hull 1964: 81-82; Toynbee 202-5
1973: 143-45; King 2002: 417
Myoxus glis Apicius 9.1.1; Mart. 3.58; Amm. Keller 1909: 191-93; Zeuner 1963: Montagne 19t
Marc. 28.4.13; Varro, Rust. 3,15; Plin. 415-16; Hull 1964: 92; Toynbee 1973:
HN 8.209, 211, 224, 16.18; Petron., 204; Andre 1981, 119-20; Colonnelli
Sat. 31.10 et al. 2000; King 2002: 428-29
Bison bonasus Plin., HN 8.38; Sen., Phaed. 64 Keller 1909: 341-42; Toynbee 1973: Montagne 19t
148
Capra aegagrus, Varro, Rust. 2.3.3; 2.1.5; Columella, Keller 1909: 296-99; Aymard 1951: Hansen-Catta
Rupicapra rupicapra Rust. 9.pr.1; Celsus, Med. 2.18.2 19-20; Hull 1964: 84-85; Toynbee
1973: 147
Gazella dama Juv. 11.138; Hdt. 4.192; Plin., HN Keller 1909: 286-88; Andre 1981: 121; Montagne 19t
Gazella dorcas 8.214; Ael., NA 7.19 Zeuner 1963: 434; Hull 1964: 83;
Toynbee 1973: 147; Gilbert 2002,
23-24; King 2002: 437
Oryx leucoryx Columella, Rust. 9.1.1, 1.7; Juv. 11.140 Keller 1909: 292-93; Andre 1981: Montagne 19t
120-21; Hull 1964: 83-84; Toynbee
1973: 146; Gilbert 2002: 21-23; King
2002: 437
Oryx beisa beisa Keller 1909: 191-92; Hull 1964: 83-84
Ovis orientalis Varro, Rust. 2.1.5; Apicius 8.4.1, 3; Youatt 1837: 133-34; Keller 1909: Montagne 19t
Polyb. 12.3; Plin., HN 8.199, 28.151; 317-18; Aymard 1951: 19-20; Andre 2002: 88-91
30.146; Strabo, 5.2.7 1981: 121; Hull 1964: 84-85; Toynbee
1973: 163
Equus africanus Verg. G 3-410; Celsus, Med. 2.18.2; Ridgeway 1905: 46-54; Keller 1909: Perlin 1989: 7
Plin., HN 8.69; Xen., An. 1.5-1-3 271-74; Toynbee 1973: 192-93; An-
dre 1967: 121; King 2002: 420-21
on the ancient sources and his own experience of fowling using birdlime in
twentieth-century Italy (Capponi 1959).
Although fowling was a popular sport, most birds were likely captured by
professional fowlers for sale in markets (Aristophanes, Aves 526-38; Plato, Leg. 824a;
Pollard 1977; Longo 1989: 62-72). The impressive scale of Greco-Roman fowling
and the astonishing range of birds offered for sale in Greek or Roman markets is
illustrated in many mosaics (Tammisto 1997; Watson 2002), wall paintings (De Caro
and Boriello 2001; Watson 2002), and references in the literary sources (Pollard
1977; Watson 2002), and can be confirmed through archaeozoological research. For
Italy and the northern provinces of the Roman Empire we can document the cap-
ture and exploitation of dozens of bird species. More than 70 of these species are
attested by ancient or modern sources as being game birds of some interest for
human consumption (table 8.4; Parker 1988).
The farming and fattening of game birds greatly enhanced the stock available in
Greek and Roman markets and became a very profitable business (Columella, Rust.
8.1.2; Varro, Rust. 3.2.13-4; Rinkewitz 1984: 21-23, 111-30). Varro himself claimed a
revenue of 60,000 HS per year, more than a typical 200-iugera farm, from his aviary
stocked with 5,000 thrushes (Varro, Rust. 3.2.15), and such profits were by no means
atypical (Varro, Rust. 3.6.1, 6.6; Pliny, HN 10.45). Varro (Rust. 3.7.11) claimed that
some dove breeders had invested as much as 100,000 HS in equipment and physical
plant alone, yet still managed to achieve good returns on their investment. Archa-
eologists have studied towers furnished with small niches matching ancient de-
scriptions of dovecotes (Varro, Rust. 3.7.1-4, Plato, Tht. 197c--d) in Etruria (Car-
andini 1984: 125; Chamoux and Hillier 1996: 56 n. 5), at Apollonia in Libya (Chamoux
and Hillier 1996), and at Thysdrus in Tunisia (Leveau et al. 1993: 103), as well as
several locations in Israel and Egypt (Zissu 1995). Many are extremely large, more
than 9 m in diameter and several stories tall, and capable of holding more than a
thousand niches for pairs of breeding pigeons (figure 8.3). Columella's recom-
mendations for habitats for the breeding of wild ducks are no less elaborate. The
pond is paved in part with pebbles or concrete to keep some clear water for the ducks
to swim, furnished with appropriate aquatic vegetation and plastered breeding
enclosures, and surrounded with 20 feet of grassy banks. It is even fitted with special
channels for flushing fresh water and the ducks' favorite foods including crayfish,
pickled river fish, and small aquatic animals (Columella, Rust. 8.15.1-7; cf. Varro,
Rust. 3.11.1-4; Rinkewitz 1984: 35-37). Modern experts offer remarkably similar, ifless
elaborate, recommendations (Coles 1971: 237-42, 263-64, 268-92).
The Roman agronomists hand down detailed instructions for raising a number
of wild species (Rinkewitz 1984: 29-73): doves, geese, mallard ducks, teal, pochards
and coots, turtle-doves, song thrushes, peafowl, pheasants, quail, blackbirds, and
buntings. The Greeks and Romans also fattened other species in captivity for later
sale: partridges (Varro, Rust. 3.11.4), cranes (Varro, Rust. 3.2.14), and ortolans
(Varro, Rust. 3.5.2) certainly, and presumably many others. The technical quality of
the agronomists' advice is impressive. The enclosures and the breeding and feeding
regimen recommended for peafowl (Columella, Rust. 8.11.3-17; Varro, Rust. 3.6.3-5)
:e for the consumption of Greco-Roman gamebird species (ranked according
1ber of archaeological sites with remains).
tific Name References-Ancient Sources References-Modern Scholarship References-Mod,
· anser Col., Rust. 8.13-4; Pall. 1.30; Varro, Keller 1913: 220-26; Thompson 1936: Montagne 1961: 4
Rust. 3.10.6; Gp. 14.22; Apic. 6.8; Juv. 325-30; Andre 1981: 132-33; Zeuner 478-79; Martin 1!
5.114; Plin., HN 10.56; foie gras: Ath. 1963: 466-70; Toynbee 1973: 261-64; 1998: 99-105
384; Gp. 14.13; 22; Plin., HN. 10.52; Salza Prina Ricotti 1987: 95-96; Flach
Hor., Sat. 2.8.88; Juv. 5.114; Mart. 1990: 322-23; Benecke 1994b: 373-79;
13.58; Apic. 6.5; 8.87; Pers. 6.71; Olson & Sens 2000: 213-14; Watson
Archestr. fr. 58; Ar., Pax 1004; 2002: 365-66; Benecke et al. 2003: 79
Anax:andr. fr. 42.64; Antiph. fr. 295;
Cratin. fr. 49
Hdt. 2.77; Col., Rust.8.15; Gp. 14.23; Keller 1913: 226-35; Thompson 1936: Montagne 1961: 3
·hnynchos Apic. 6.fa.1-6; Petr., Satyr. 93; Mart. 205-6; Andre 1981: 133; Zeuner 1963: 480-90; Martin 1
13.52; Varro, Rust. 3.11; Gp. 14.23; 470-71; Toynbee 1973: 264-73; Salza 1998: 108-19
Auson., Epist. 3.12 Prina Ricotti 1987: 95; Benecke 1994b:
379-83; Watson 2002: 364-65
~ax rusticola Mart. 13.76; Anth. Lat. 2.884; Keller 1913: 181; Thompson 1936: Montagne 1961: 11
Nemes., Cyn. 2.1-10; Plin., HN 261-62; Prummel 1987: 194-95; Salza 530-33; Aksakov
10.56, 111 Prina Ricotti 1987: 98
crecca Ar., Av. 880, 1300; Col., Rust.8.15.1; Keller 1913: 233; Thompson 1936: 205-6; Montagne 1961: 3
Varro, Rust. 3.3.3, 3.11-4; Macrob., Andre 1981: 126; Toynbee 1973: 264-73; 1998: 127-29
Sat. 3.13.12 Salza Prina Ricotti 1987: 95
,s monedula Ath. 65e, 393b; Gp. 14.24; Ath. 65c-d Keller 1913: 92, no; Andre 1981: 129; Perlin 1989: HN.
Pollard 1977: 105; Bokonyi 1984: 99
ued)
tific Name References-Ancient Sources References-Modern Scholarship References-Mod,

grus Ath. 131,338; Dionys., Av. 3.11; Plut., Keller 1913: 184-93; Thompson 1936: Montagne 1984: :
De esu earn. 997a; Varro, Rust. 74-75; Zeuner 1963: 474; Toynbee 1973: 157-62
3.2.14; Hor., Sat. 2.8.86-87; Epod. 243-44; Pollard 1977: 105-6; Rinkewitz
2.35; Verg., G. 1.307; Stat., Silv. 4.6.9; 1984: 73; Salza Prina Ricotti 1987:
Gell. 6.16.5; Plin., HN 10.60; Apic. 99-100
6.2.1-6; Cels. 2.18.2
nba livia Varro, Rust. 3.7; Col., Rust. 8.11.1-7; Keller 1913: 122-31; Thompson 1936: Montagne 1961: i
Gp. 14.1-6; Mart. 2.37.6, 13.66; Apic. 238-46; Andre 1981: 124; Zeuner 1963: 153-68; Escoffier
6.4.1-4; Pall. 1.24; Hor., Sat. 2.8.91; 460-62; Toynbee 1973: 258-59; Pollard
Philumen. 2.126 1977: 104; Salza Prina Ricotti 1987: 96;
Flach 1990, 320-22; Benecke 1994b:
383-90; Watson 2002: 372-74; Benecke
et al. 2003: 79-80
nba oenas Montagne 1961: i
153-68; Escoffier
,s corone Zacharias, Epist. 13 Keller 1913: 92; Thompson 1936: 168-72; Montagne 1961: 3
Andre 1981: 128; Watson 2002: 375-76
nba palumbus Apic. 6.4.1-4; Ar., Ach. 1105, 1107; Keller 1913: 122-31; Thompson 1936: Montagne 1961: i
Varro, Rust. 3.9; Cato, Agr. 90.1; 300-302; Andre 1981: 133; Zeuner 1963: 362-65; Escoffier
Cels. 2.30; Mart. 3.58, 13.67; Petr. 460-62; Toynbee 1973: 258-59; Pollard Aksakov 1998: 2s:
Satyr. 70.2; Plaut., Poen. 676 1977: 104; Salza Prina Ricotti 1987: 96;
Flach 1990, 320-22; Benecke 1994b:
383-90; Watson 2002: 374-75
>.: perdix Ath. 390b; Dionys., Av. 3.7; Hor., Keller 1913: 156-60; Thompson 1936: Montagne 1961: 7
Epod. 2.53-54; Mart. 3.58.15, 13.61, 234-38; Toynbee 1973: 255-56; Pollard 177-95; Escoffier
65, 76; Varro, Rust. 3.11.4; Gp. 1977: 104; Rinkewitz 1984: 55; Salza 1998: 181-88
14.19-21; Hist. Aug. 41.7; Apic. Prina Ricotti 1987: 98--99
6.3-1-3; Plin., HN 10.101-2;
Philumen. 1.112; 2.126
penelope Keller 1913: 226-35; Thompson 1936: Montagne 1961: 3

205-6; Toynbee 1973: 264-73; Salza 77-81; Aksakov 11
Prina Ricotti 1987: 95; Watson 2002:
364
1s frugilegus Keller 1913: 91; Thompson 1936: 168-69; Montagne 1984: !
Bokonyi 1984, 99 27-36
us vulgaris Diocletian, Price Edict 4.42; Ath. Keller 1913: 90-91; Thompson 1936: Montagne 1961: 9
65e; Plin., HN 28.110; Mart. 9.5; 334-35; Toynbee 1973: 276; Salza Prina
Philumen. 2.126 Ricotti 1987: 100; Watson 2002: 395
ilis apricaria Ar., Av. 265; Plt., Grg. 494b Thompson 1936: 311-14; Prummel 1987: Montagne 1961: 7
194-95 Aksakov 1998: 18!
~s merula Dionys, Av. 3.13; Arist. Av. 1080; Keller 1913: 75-76; Thompson 1936: Montagne 1961: 1
Hor., Sat. 2.8.91; Ars Poetica 458-59; 174-76; Toynbee 1973: 277; Pollard 1977: 1984: 527
Varro, Rust. 3.5.6-7; Philumen. 104; Prummel 1987: 195; Watson 2002:
1.112, 2.126; Ath. 65 D; Plin., HN 397--98
10.141-42
ued)
rifle Name References-Ancient Sources References-Modern Scholarship References-Mod,

~s philomelus Ath. 64f-65b; Ar., Av. 1080; Ach. 961, Keller 1913: 76-79; Thompson 1936: 149; Montagne 1961: s
1007; Nub. 339; Apic. 5.3.1, 8.7.14; Andre 1981: 125; Toynbee 1973: 277-78; 525-27
Col., Rust. 8.10; Varro, Rust. 3.3.3; Salza Prina Ricotti 1987: 100; Watson
Hor., Ep. 1.15; 41; Sat. 1.5.72, 2.5.10; 2002: 398-99
Ov. Ars. am. 2.269; Plin., Ep. 5.2.1;
Pall. 1.26; Petr. Satyr. 40; Mart.
3.47.10; 13.51, 92; Gp. 14.24, 15.1.19;
schol. Pers. 6.24; Macr. 2.4.22;
Philumen. 1.112
r sp. Diocletian, Price Edict 4.37 Keller 1913: 88-90; Thompson 1936:
268-70; Salza Prina Ricotti 1987: 100
·nix coturnix Dionys., Av. 1.60, 3.9; Ath. 393a; Keller 1913: 161-64; Thompson 1936: Montagne 1961: 7
Varro, Rust. 3.5.2; Pliny, HN. 10.69; 215-19; Andre 1981: 125-26; Zeuner 517-27; Aksakov :
Diocletian, Price Edict 4.41; Gp. 1963: 458; Toynbee 1973: 256-57;
14.24; Ar., Pax 789; Juv. 12.97; Plaut., Pollard 1977: 105; Rinkewitz 1984:
Asin. 666; Luer. 4.640 67-68; Salza Prina Ricotti 1987: 100;
Watson 2002: 376-77
querquedula Keller 1913: 233; Thompson 1936: 205-6; Montagne 1961: 3
Toynbee 1973: 264-73; Salza Prina 94-98; Aksakov 1
Ricotti 1987: 95; Watson 2002: 365
mius arquata Keller 1913: 183; Thompson 1936: 207; Montagne 1961: 3
Prummel 1987: 194-95 170-78
~elis or Diocletian, Price Edict 4.34; Ath. Keller 1913: 86-87; Thompson 1936:
la sp. 65c-d 266; Andre 1981: 128; Pollard 1977: 107
1lus glandar- Ath. 65e; Zacharias, Epist. 13 Keller 1913: 112-14; Andre 1981: 128 Coles 1971: 27-36
cristatus Varro, Rust. 3.6.1-6; Col., Rust. Keller 1913: 148-54; Thompson 1936: Montagne 1961: 7
8.11.1-7; Pall. 1.28; Gp. 14.18; Plin., 277-81; Zeuner 1963: 456-57; Toynbee
HN 10.45; Ael., NA 3.42; Diocletian, 1973: 250-53; Salza Prina Ricotti 1987:
Price Edict 4.39-40; Hor., Serm. 96; Watson 2002: 388-89
1.2.115, 2.2.23-31; Apic. 2.2.6; SHA,
Alex. Sev. 41.6; Ael., NA 5,4;
Heliogabalus 20.7; Cic., Fam. 9.18.3;
Suet., Vit. 13.2; Tib. 60.1; Calig. 22.3;
Juv. 1.143, 7.32; Mart. 3.58, 70; 14.67,
85; Varro, Sat. Men. 13.18; Gell.
6.16.5; Ath. 58b; Curt. Ruf. 9.1.13;
Petr. Satyr. 33.5
)ica Mart. 3.60.8 Keller 1913: 112-14; Andre 1981: 128 Coles 1971: 27-36
topelia turtur Varro, Rust. 3.8.3; Col., Rust.8.9; Keller 1913: 125, 127; Thompson 1936: Montagne 1961: 7
Apic. 6.2.1, 3.3; Pall. 1.25; Gp. 14.24; 290--92; Andre 1981: 124; Zeuner 1963: 15-16, 303-16, 365
Juv. 6.39; Mart. 3.60.7, 82.21; 13.53; 460-62; Pollard 1977: 104; Salza Prina 500-505; Aksakm
Ael., NA 13.25; Plaut., Most. 46; Ricotti 1987: 98; Watson 2002: 394--95
Philumen. 1.112, 2.126
11s cygnus Ath. 393c-d; Plut., De esu earn. 997a; Keller 1913: 213-20; Thompson 1936: Montagne 1961: s
Hor., Sat. 2.2.49; Plin., HN 10.60, 179-86; Andre 1981: 127; Toynbee 1973: note 166; Aksako
30.30, 30.69; Orib. 4.77; Zacharias, 259-61; Pollard 1977: 106-7; Rinkewitz
Epist. 13 1984: 71; Watson 2002: 377
11s olor Ath. 393c-d; Plut., De esu earn. Keller 1913: 213-20; Thompson 1936: Montagne 1961: 9
997A; Hor., Sat. 2.2.49; Plin., HN 179-86; Toynbee 1973: 259-61; Pollard MacGregor 1996;
10.60, 30.30, 30,69; Orib. 4.77 1977: 106-7; Rinkewitz 1984: 71
11s merganser Keller 1913: 244 Martin 1993: 168-
136-39
ued)

1 atra Ath. 65e; 393c; Varro, Rust. 3.3.3, Keller 1913: 235-37; Thompson 1936: Montagne 1961: 3
11.4; Col., Rust.8.15.1; Ar., Av. 565; 298; Andre 1981: 126; Watson 2002: 379 140-42
Macr. 3.13.12; Plin., HN 10.67
· albifrons Keller 1913: 224 Montagne 1961: ~
38-43
'ia ciconia Hor., Serm. 2.2.49-50 and schol. ad Keller 1913: 193-202; Thompson 1936:
loc.; Gell. 6.16.3; Mart. 13.75; Plin., 221-25; Toynbee 1973: 244-45; Watson
HN 10.60-61, 68; Petr. Satyr. 55.6 2002: 371-72
•aferina Keller 1913: 226-35; Thompson 1936: Montagne 1961: 7
205-6; Toynbee 1973: 264-73; Salza 110-14
Prina Ricotti 1987: 95
,ha/a clangula Keller 1913: 226-35; Thompson 1936: Montagne 1961: 3
icrocorax Keller 1913: 239; Thompson 1936: 164;
Andre 1981
anus colchicus Apic. 2.2.1, 2.6; Ar., Nub. 108; Ath. Keller 1913: 145-46; Thompson 1936: Montagne 1961: 7
386e, 387b; Mart. 3.77, 3.58.16; 13.72; 298-300; Zeuner 1963: 458; Toynbee 60-72, 107-76; Es
Plin., HN 10.132; Juv. 11.139; Pall. 1973: 254-55; Rinkewitz 1984: 54-55; 67; Hansen-Catta 20,
1.29; Gp. 14.19; Silv. 1.1.77-78; 96; Salza Prina Ricotti 1987: 98-99;
Ammian. Marc. 16.5.3; Paul, Dig. Watson 2002: 389-90
32.1.66; SHA, Alex. Sev. 41.6; SHA,
Pert. 12.6
iago gallinago Prummel 1987: 194-95; Watson 2002: Montagne 1961: 8
379-80 293-96; Escoffier
1998: 24-30, 39-4
o urogallus Plin., HN 10.56; Suet., Calig. 22.3; Keller 1913: 165-66; Thompson 1936: Montagne 1961: 4
Nemes., Cyn. 1.1-15; Nemes., De 283; Toynbee 1973: 256; Rinkewitz 1984: 222-26; Hansen-<
Aucupio fr. 1 29; Watson 2002: 396
o tetrix Keller 1913: 156-60 Montagne 1961: 4
227-44; Hansen-I
•afuligula Keller 1913: 226-35; Thompson 1936: Martin 1993: 115-

205-6; Toynbee 1973: 264-73; Salza 132-34
streptera Keller 1913: 226-35; Thompson 1936: Montagne 1961: 3
;a limosa Montagne 1961: 4
296-98;Aksakov
,s iliacus Ath. 65a Keller 1913: 76; Thompson 1936: 121 Montagne 1961: 8
arda; Otis Cyn. Anab. 1.5,1-3; Ael., HA. 4.24; Keller 1913: 175-77; Thompson 1936: Montagne 1961: Ii
Ath. 393d; Gal., Viet. Att. 3 = Med. 338-39; Andre 1981: 126-27; Pollard Aksakov 1998: 15~
Gr. 6.703.9; Plin., HN 10.57, 30.61; 1977: 106; Anderson 1985: 61-62
Synesius, Epist. 4.165
,s pilaris Arist., Hist. an. 6.1.6 Keller 1913: 76; Toynbee 1973: 277-78; Montagne 1961: 9
Prummel1987: 195
us tetrix Keller 1913: 156-60 Montagne 1961: 4
HN. 752
ued)

riza sp. Ath. 65e; Ael. 13.25; Varro, Rust. 3.5; Keller 1913: 72; Thompson 1936: 136; Montagne 1961: 1
5.11; Varro, Ling. 5.76 Andre 1981: 127 1998: 291
·a leucopsis Plin., HN 10.54 Keller 1913: 220-26; Thompson 1936: Martin 1993: 57-(
325-30; Toynbee 1973: 261-64; Salza
Prina Ricotti 1987: 95-6
acuta Keller 1913: 226-35; Thompson 1936: Montagne 1961: 7
clypeata Keller 1913: 226-35; Thompson 1936: Montagne 1961: 3
205-6; Toynbee 1973: 264-73; Salza 1993: 99-103; Aks
1ris rufa Keller 1913: 156-60 Montagne 1961: 7
55-59; Hansen-O
la arvensis Dionys., Av. 3.15; Arist. HA. 617 B Keller 1913: 86; Pollard 1977: 104 Montagne 1961: 5
ilis squatarola Ar., Av. 266; Enn. Heduph. 48 Keller 1913: 179-80; Thompson 1936: Montagne 1961: 7
311-14
stes bonasia Mart. 13.61; Apic. 6.3.3; Ath. 398c-f; Thompson 1936: 282-83; Salza Prina Montagne 1961: ~
Ar., Av. 885 Ricotti 1987: 99 534; Aksakov 199:
Hansen-Catta 20,
ex Ath. 393a Keller 1913: 208; Thompson 1936: 214-15 Aksakov 1998: 191
Keller 1913: 220-26; Thompson 1936: Montagne 1961: ~

rhynchos 325-30; Toynbee 1973: 261-64; Salza 34-7
Prina Ricotti 1987: 95-96
,us lagopus Keller 1913: 156 Montagne 1961: 7
299-323; Aksakov
Hansen-Catta 201
,us mutus Hor., Sat. 2.2.22; Plin., HN 10.133 Keller 1913: 156; Thompson 1936: 190; Montagne 1961: 7
Rinkewitz 1984: 55 Hansen-Catta 201
us canorus
1 Plin., HN 10.27; Arist., HA 564a 3-4 Keller 1913: 63-67; Thompson 1936: Montagne 1961: 3
151-53; Andre 1981: 128; Watson 2002:
377
"na tadorna Plin., HN 10.29.56; Ath. 58B Keller 1913: 235; Thompson 1936: 205-6; Montagne 1961: 3
Toynbee 1973: 264-73; Pollard 1977: 65; 65-71
Salza Prina Ricotti 1987: 95; Watson
2002: 396
!ea leucorodia Keller 1913: 203; Thompson 1936: 193; Perlin 1989: 64 H
Toynbee 1973: 264-73; Salza Prina
Ricotti 1987: 95
'da meleagris Apic. 6.4.1-4; Col., Rust.8.2.2-3; Keller 1913: 154-56; Thompson 1936: Montagne 1961: 4
Varro, Rust. 3.9.18; 8.12; Ath. 655; 197-200; Andre 1981: 133-34; Zeuner
Hor., Ep. 2.53; Juv. 11.142-43; Petr. 1963: 457; Toynbee 1973: 253-54;
Satyr. 93; Mart. 3.77.3, 13.45.1, 13.73; Lamblard 1975; Rinkewitz 1984: 53;
Plin., HN 8.223, 10.74; Gp.14.19; Watson 2002: 384-85
Suet., Calig. 22.3
nachus Plin., HN 10.74 Keller 1913: 181; Thompson 1936: Aksakov 1998: 71-
IX 200-201; Prummel 1987: 194-95
rgris gallopavo Montagne 1961: 9

1ia Hor., Sat. 2.3.235; SHA, Heliogab. Keller 1913: 73-75; Thompson 1936:
vnchos 20.5 16-22; Toynbee 1973: 276-77; Watson
2002: 383-84
ued)

us oriolus Mart. 13.68 Keller 1913: 120; Thompson 1936: Montagne 1961: ~
332-23; Andre 1981: 128; Watson 2002:
386-87
:Zia came/us Apic. 6.1.1-2; SHA, Heliogab. 22.1, Keller 1913: 166-75; Thompson 1936: Montagne 1961: f
28.4, 30.2; SHA, Firmus 4.2; Ath. 270-73; Zeuner 1963: 476-77; Toynbee
145d 1973: 237-40; Pollard 1977: 106; Salza
1icopterus Mart. 3.58.14; 13.71; Apic. 6.6.1-2; Keller 1913: 209-13; Thompson 1936: Ferlin 1989: 54
Juv. 11.139; Seneca, Epist. 110.12; 304-6; Andre 1981: 128; Toynbee 1973:
Suet., Vit. 13; Calig. 22; Ar., Av. 273; 246; Salza Prina Ricotti 1987: 99;
Cels. 2.18.3 Watson 2002: 390
ratricapilla/ Ath. 64b-c; Mart. 13.49; Gell. 15.8.2; Keller 1913: 118-20; Thompson 1936:
Juv. 14.9; Petr. Satyr. 33.8; Suet., Tib. 274-75; Rinkewitz 1984: 72; Salza Prina
42.2; Macr. Sat. 3.13.12; Diocletian, Ricotti 1987: 98; Watson 2002: 396
Price Edict 2.5
cus Krameri Apic. 6.6.1; SHA, Heliogab. 20.6; Keller 1913: 45-49; Thompson 1936:
!lensis Claud. in Eutrop. 2.329 336-38; Toynbee 1973: 247-49
·olinus Apic. 6.3.3; Ath. 387f, 388b; Ar., Ach. Keller 1913: 158; Thompson 1936: 61;
1/inus 875; Mart. 2.37, 13.61; Ov. Fast. 6.175; Andre 1981: 126; Rinkewitz 1984: 56
Plin., HN 8.83, 10.68, 10.133; Aul.
Gell. 6.16.5; Hor., Ep. 2.54; Mart.
2.37; 13.61; Gp. 14.19
ryrio Plin., HN 10.129, 135; Mart. 13.78; Andre 1981: 135 Ferlin 1989: 63 H
yrio Ath. 388d
iula chloropus Aksakov 1998: 75·
tlus vanellus Keller 1913: 178 Montagne 1961: 5
81-83
· fabalis Martin 1993: 30-:
Figure 8.3. Roman columbarium at Apollonia in Cyrenaica, reconstruction. (From

Chamoux and Hallier 1996: fig. 9.)
and pheasant (Palladius 1.29-4; Rinkewitz 1984: 54, 67), for example, find many
parallels with methods used today (Delacour 1964: 29-39; Coles 1971: 96-100). The
precautions against aggression (Columella, Rust. 8.11.7; cf. Delacour 1964: 32), and
the use of broody hens of the domestic chicken to incubate the eggs of these and
many other wild species (Columella, Rust. 8.11.10-14, 15.7), thereby encouraging
further egg production, are especially important and very much in accord with
modern practice (Farrington 1913: 119, 122; Coles 1971: 75-80).
FISHING AND FISH FARMING
Research has begun to document in increasing detail the impressive scale and
sophistication of the fishing industry in classical and Hellenistic Greece and late
Republican Rome (Salza Prina Ricotti 1999; Sahrhage 2002; Bekker-Nielsen 2005b).
The remains of fish processing plants (chapter 14; Trakadas 2005) reflect fishing on
an industrial scale for mackerel, anchovies, sardines and other gregarious migra-
tory pelagic fish in the eastern Mediterranean and along the coasts of Spain, North
Africa, and Sicily (Sahrhage 2002: 76-79; H0jte 2005). Greek and Roman tuna
fishing was equally impressive (Sternberg 1998: 98-103; Sahrhage 2002: 58-69;
Shepherd and Dallai 2003). The vivid interest in fishing on the part of the Greco-
Roman public is evident from Aristotle's superb work on ichthyology (Mair 1928:
xxvi-xxviii, citing Cuvier and Valenciennes 1828-1848), the popularity of repre-
sentations of fish and fishing in the decorative arts (Reese 2002a-b; Bekker-Nielsen
2002), and the many accounts of fishing-both lost technical manuals like that of
Leonidas of Tarentum, and popularizing works such as those of Oppian, Aelian,
Archestratus, and Ovid (Mair 1928: xxxvii-xxxviii; Bekker-Nielsen 2005b). Finally,
a significant proportion of the demos of several Greek cities, most notably By-
zantium, Tarentum, and Syracuse, made their living from fishing (Trotta 1996:
234-43), and fishermen often enjoyed the same social status and pride in their craft
as many other skilled artisans (Lubtchansky 1998).
As in the other Greco-Roman skilled trades, one can observe a great deal of
innovation and creativity in the development of new methods and technologies.
Del Rosso, a leading expert on contemporary Italian fisheries and on Roman fish-
ing, argued forcefully at the beginning of the twentieth century that there had still
been few meaningful advances in Italian fishing methods over those already known
in Roman times (Del Rosso 1905: 26-27, 32-36, 61-70). Dietrich Sahrhage's brief
account of fishing techniques (Sahrhage 2002: 41-56), informed by his own expertise
in modern fisheries, shows just how many important modern fishing techniques
can be attested in Greek or Roman use.
Ancient references for these techniques, along with comparative evidence from
the contemporary Mediterranean, are listed in table 8.5. Although it is sometimes
claimed (Gallant 1985: 25) that Greco-Roman net fishing was a predominantly
shore-based activity, literary sources and mosaics make it absolutely clear that
fishing nets were used from a range of small seacraft (Bekker-Nielsen 2002), as well
as from ships of significant size (Bekker-Nielsen 2002: no. 9; Sahrhage 2002: 92 pl.
4, top; Bekker-Nielsen 2005b: 88). Likewise, despite claims that deep sea fish were
not exploited (pace Gallant 1985: 27), archaeozoological studies of fish landed at
Greek and Roman sites in southern France clearly show a marked increase in the
number of benthic and pelagic species (Sternberg 1995, 1998, 2005a: 247). One ef-
fective method for catching large numbers of fish simultaneously at almost any
depth is longline fishing (Gabriel et al. 2005: 114-35) referred to clearly several times
by our sources (Oppian, Hal. 3.78; 3.468-81; Aristotle, Hist. an. 532b25; 621a15;
Aelian, NA 12.53; 15.10 ). Fishermen fix a large number of shorter branch lines called
snoods, each with up to a hundred hooks, to a single main line, thereby permitting
hundreds or thousands of hooks to be set. The labor of baiting so many hooks is
the main limitation on the size of the catch, but the markedly improved design,
manufacture, and standardization of Roman fish hooks, including small stream-
lined ones similar to those now used in longlining (Cleyet-Merle 1990: 166-70),
confirms the importance of this method in Roman fishing.
The most significant limitation on Greco-Roman fishing technology was pre-

sumably the absence of steam or internal combustion power for trawling (Gabriel
et al. 2005: 402-5) or winching (Gabriel et al. 2005: 9). Nevertheless, given that there
were only four steam-powered fishing boats in the whole of the Mediterranean as
late as 1905 (Del Rosso 1905: 34), we should not exaggerate the significance of this
factor. Moreover, while modern industrial trawling produces large catches, these
often cannot be sustained over the long term, but soon lead to a reduction in the
size of fish and, eventually, a severe depletion of fish-stocks. While they fished on a
very large scale, particularly for the garum and salt-fish industry, and did so over
centuries, Greco-Roman fishermen seem generally to have avoided the destructive
overfishing typical of modern trawling. Several studies have shown that fish cap-
tured by Greco-Roman fishermen were often significantly larger than are typically
caught today (Desse-Berset 1993). Moreover, the highly prized sturgeon (Martial
13.91; Pliny, HN 9.60) still represented about half of the total number of fish caught
by Hellenistic and Roman fishermen in the Black Sea in some surveys, compared to
barely 3 to 4 percent of the fish caught in the same waters today (Ivanova 1994: 280).
Fish Farming
The Greeks and Romans did not rely solely on fishing to supply their great demand
for seafood, but also put great efforts into improving existing fish stocks and into
fish farming. Gallant's use of modern fisheries' statistics to place limits on possible
Greco-Roman fish production (Gallant 1985) has been criticized on a number of
grounds (Jacobsen 2005), as have a number of the claims in his book (Purcell 1995;
Sahrhage 2002: 5 n. 1; Bekker-Nielsen 2005b: 84-86). Perhaps the most critical
objection, however, comes from his neglect of the significance of fish farming.
Modern fish farmers, using techniques which were widely applied by the Romans,
produce quantities of a number of popular Mediterranean sea fish far in excess of
those that can be harvested from natural fish stocks (Kron forthcoming). For
example, according to Food and Agricultural Organization statistics (PAO 2002)
Mediterranean fishing of gilthead seabream, gray mullet, and European sea bass
provides barely one-tenth to one-fifth as many fish as aquaculture.
The Greeks and Romans put considerable effort into enhancing and even
managing their fish stocks. They introduced eggs, fingerlings, and adult fish, in-
cluding new fish species, into fresh water lakes and the Tyrrhenian Sea to enhance
fish stocks (Pliny, HN 9.62-63; Lafon 2001: 162). Moreover, in addition to intensive
aquaculture, they practiced a more extensive system of exploiting and enhanc-
ing stocks of fish in brackish lagoons, first alluded to by Greek sources (Aristotle,
Hist. an. 504b33). Similar techniques were widely used in Italy since the late
medieval period under the title of vallicoltura (Del Rosso 1905: 59; De Angelis
1959; Bevilacqua 1987), and equally widely in modern Greece (Apostolides 1883,
Guest-Papamanoli 1986). McCann's study of the massive Tagliata canal at Cosa
and the associated fish processing plant and fish farming tanks offers the best
loman fish-catching methods.
Ancient Sources Modern Scholarship Comparative Evie
Sahrhage 2002: 40-41 Gabriel et al. 200
neral Sahrhage 2002: 43-44 Gabriel et al. 200
:>m Opp., H 3.543-67; Polyb. 34.3-1-8; Str. D' Arrigo 1956; Gallant 1985: 13; Longo Apostolides 1883:
1.24-25; Poll., Onom.10.133; Pl., Soph. 1989: 27-30; Trotta 1996: 242; Olson Sisci 1992: 139-45
220c; Ath. 7.314e-f. and Sens 2000: 171; Bekker-Nielsen Gabriel et al. 200
2002: nos. 2, 8; Sahrhage 2002: 43-44,
fig. 17, 18a-b; Bekker-Nielsen 2005b: 89;
H0jte 2005: 135-36
Opp., H 4.640-84; Pliny, HN 25.98, Mair 1928: 452-53, notes a-c; Gian- Gabriel et al. 200
120; Arist., Hist. an. 602b31 frotta 1999: 10; Sahrhage 2002: 41-2
Cleyet-Merle 1990: 166-70; Rossi 1990: Gabriel et al. 200
114, no. 255; Gianfrotta 1999: 26-27, fig.
17; Sahrhage 2002: 45, fig. 18c-d; H0jte
2005: 135-36
Opp., H 1.72 f.; 3.76-77; Hom., II. Mair 1928: xxxix; 350-51, note f Gabriel et al. 200
16.406
Opp., H 3.74-75 Mair 1928: xxxix; Gallant 1985: 14-15, Gabriel et al. 200
pl. 3, 4, 6; Sahrhage 2002: 17 fig. 3, 46
fig. 19, 47 fig. 21, 92 pl. 4, bottom;
Bekker-Nielsen 2002: notes 9, 12;
Bekker-Nielsen 2005b: 89-90
Opp., H3.77 Mair 1928: xxxix; 351 note g; Cleyet-

Merle 1990: 169; Sahrhage 2002: 45 fig.
18e-g
1,1ed)
Ancient Sources Modern Scholarship Comparative Evit
Opp., H 4.78-125 Gallant 1985: 15; Sahrhage 2002: 49 Gabriel et al. 200
Opp., H 3.78, 468-81; Arist., Hist. an. Mair 1928: xxxix-xl; Gallant 1985: Gabriel et al. 200
532b25, 621a15 15-66; Sternberg 1998: 96-8 tab. 5-6;
Bekker-Nielsen 2005b: 89-90
Opp., H 4.439-49; Opp., H 3.529-41 Rossi 1990: 114-15, no. 254; Purpura Gabriel et al. 200
(swordfish); Ael., NA 14.26 (S. Russia 1992: 92; Gianfrotta 1999: 26-27, fig. 17;
sturgeon) Sahrhage 2002: 49-51
ecies Opp., H 3,176-93, 443-528ff., 4.308f. Sahrhage 2002: 47-8
Ael.,NA15,1 Sahrhage 2002 50-1 Gabriel et al. 200
le of fences Opp., H 3.73-90, 342, 401; Ael., NA Gallant 1985: 13; McCann et al. 1987 Romdhane 1998: :
12.43 220-21
.e nets Opp., H 3.85-7 (generic); Arist., Hist. Mair 1928: xlvi, 352 notes a, b; Gallant Apostolides 1883:
an. 534a11; b10; Opp., H 3.338ff.; Sil., 1985: 13; Sahrhage 2002: 56-57; 67; Gabriel et al.
Pun. 4.47; Pliny, HN 9.132; Pl., Leg. Bekker-Nielsen 2002: nos. 1-5
823e, Ti. 79d; Theocr. 21.11; Poll.,
Onom. 10.132; Anth. Pal. 6.23
mbers Aesch. Pers. 424-26; Opp., H 3.623-27; Dumont 1976-77; Gallant 1985: 21-23; Romdhane 1998:
3.637-46; Ael. NA 13.16f.; 15.5-6; Trotta 1996: 231-33; Sternberg 1998: 2005: 207-13
Philost. Imag. 1.13 98-103; Mastromarco 1998; Sahrhage
2002: 58-69; Shepherd and Dallai 2003;
Bekker-Nielsen 2005b: 93
Opp., H 4.64-71 Gianfrotta 1999: 21-22, figs. 11-3 Rendini 1997; Ro
Gabriel et al. 200
Opp., H 3.98-116 Apostolides 1883:
73-74; Gabriel et
Opp., H 4.251 Gallant 1985: 18-19; Bekker-Nielsen Gabriel et al. 200
2002: no. 9; Sahrhage 2002: 54, pl. 4
bottom
Mair 1928 xliii-liv; Trotta 1990: 36-46; Gabriel et al. 200
1996: 231; Sternberg 1998: 96-98,
tab. 5-6
Opp., H 3.83; Hom. Il. 5.487 Trotta 1990: 36-46; 1996: 231 Gabriel et al. 200
Opp., H 3.79-84, 4.468 ff. (sardines), Mair 1928: xliv-v; Trotta 1996, 231-32; Gabriel et al. 200
4.491-96; Alciphr. 1.13; 20; 21; Plut. Sternberg 1998: 96-98, tab. 5-6;
Mor. 977f Sahrhage 2002: 52-54; Bekker-Nielsen
2005b: 92
Ael., NA 14.26; Str. 7.3.18 Sahrhage 2002: 55-56 Gabreil et al. 200
Opp., H 3.124; 4.490-503; Plut. SA Mair 1928: xliv-v; Gallant 1985: 19; Romdhane 1998: :
977 f. Bekker-Nielsen 2002; Sahrhage 2002: 439-40
53, pl. 5 top
Mair 1928: xliv-v; Bekker-Nielsen 2002: Romdhane 1998:
fig. 4, nos. 20-29; Sahrhage 2002: 53-54, 441-44
pl. 4 bottom, fig. 26
,eine or Opp., H 4.566-76 Bekker-Nielsen 2002: no. 6 Gabriel et al. 200
Ael.,NA 1.39 Gallant 1985: 19-20; Sahrhage 2002: 54, Gabriel et al. 200
abb. 29 1998: 72-73
Opp., H 3.82; Aesch., Ch. 494 Mair 1928: xlv Apostolides 1883:
1,1ed)
Ancient Sources Modern Scholarship Comparative Evit
Opp., H3.80 Mair 1928: xii-xiii; Gallant 1985: 18-19; Apostolides 1883:
Trotta 1990: 36-46; 1996, 231; Sahrhage 72; Bekker-Nielse
2002: 54-55, figs. 3, 20, 29; Bekker-
Nielsen 2002: no. 7; Bekker-Nielsen
2005b: 91
Opp., H 3.577--95 Gallant 1985: 20-21; Sahrhage 2002: 54 Gabriel et al. 200
ixual lures Opp., H 4.127-46; Arist., Hist. an. Mair 1928: 412 note d, 414 note a; Apostolides 1883:
541a19 Sahrhage 2002: 49 1998: 65-66; Gabi
nun or Sahrhage 2002: 47-48 Gabriel et al. 200
¥}t Pl., Soph. 220d; Poll., Onom. 7.138; Mair 1928: xlvii; Gianfrotta 1999: 21; Apostolides 1883:
Opp., H 4.640-46; Quint. Smyrn. Sahrhage 2002: 48 gar-fish (Belone l
7.569-76; Plin., HN 9.33; Ael. NA 2.8 Mair 1928: 450-1
waters" in Scotia
2005: 151-53
mnd Sahrhage 2002: 56 Sahrhage 2002: 51
155-57
tificial Opp., H 4.404-36 Romdhane 1998:
2005: 190-92
st large Aelian, NA 14.25 Cleyet-Merle 1990: 161-62; Sahrhage Sahrhage 2002: 4,
2002: 48-49, pl. 5 in Danube in rec
archaeological evidence and analysis, with extensive attention to the many parallels
with modern practice (McCann et al. 1987: 35-43, 137-59). The creation by the
Roman state of a massive mole to control the flow of seawater, and the ingress and
egress of fish, between the Tyrrhenian Sea and the Lago Lucrino (Servius, ad. Verg.
G. 2.161; Pagano 1983-1984: 127-29) shows the economic importance of fishing in
brackish lakes and lagoons. Finally, the so-called Emissario Romano connecting
the Lago di Paola with the sea between Monte Circeo and the modern Sabaudia
was likely motivated (at least in part) by similar considerations (Lugli 1928: 31-34,
no. 32; figs. 37-38).
The Greco-Roman fish-farmers soon developed more intensive methods to in-
crease production even further, spawning fish in artificial ponds or tanks and feeding
them to maximize stocking density and growth. They raised many different species.
Higginbotham lists the common eel, conger eel, moray eel, several species of gray
mullet, sea bass, gilthead seabream, red mullet, and the rhombus, most likely either
sole or turbot (Higginbotham 1997: 41-53). In addition to increasing supply, the
Roman practice of fish farming helped bring about some significant changes in the
way Romans marketed and ate fish. Higginbotham has identified more than 70 small
pools in the gardens or peristyle courts of houses in Pompeii and Herculaneum
alone that show evidence of being used as fishponds (Higginbotham 1997: 198 n.
257), and several sources (Martial 10.30; cf. Lafon 2001: 310) imply that it was fash-
ionable to catch live fish from one's own fishpond to ensure their freshness.
The impressive remains of fish farms all along the Tyrrhenian coast (figure 8.4),
from Faleria in the north to Briatico in the south (catalogued in depth in several
syntheses: Giacopini et al. 1994; Higginbotham 1997; Lafon 2001) reveal the lavish
capital investment and technological innovation of Roman fish farming. Although
many literary sources allude to the popular fascination with fish-farming (Conta
1972; Kajava 1998-1999 ), the high productivity of Greco-Roman aquaculture can be
illustrated most clearly through an analysis of the fish farms themselves in the light
of modern fish farming practice. Research still makes limited use of comparative
evidence (Higginbotham 1997: 9-40; Giacopini et al. 1994: passim), but a few ten-
tative observations are in order. In order to maximize overall production, artificial
feeding is generally critical (Huet 1986: 334-85; Shepherd and Bromage 1988: 78-83),
and it is clear from the ancient sources that this practice was very common (Varro,
Rust. 3,17.2-3; 3.17.6-7; Columella, Rust. 8.17.12-13; 15; Pliny, HN 37.2). Some fish
tanks were equipped with gangways (Higginbotham 1997: 133), allowing one to feed
the fish by hand and observe their behavior (Shepherd and Bromage 1988: 79). A
critical limiting factor on the stocking rate of fish tanks is oxygen, which requires
constant water exchange (Shepherd and Bromage 1988: 69-71, 317; Huet 1986: 8-9)
or artificial aeration (Boyd and Tucker 1998: 306-53). The Romans recognized the
problem (Pliny, HN 9.56) and had practical means to address it (Varro, Rust. 3.17.8-
9; Columella, Rust. 8.17.1-6), and many excavated fish tanks show careful design to
maximize water flow (e.g., Higginbotham 1997: 152-57, fig. 62-65; Gianfrotta 2002:
77, figs. 11-12). A number of tanks used fountains and inflow canals well above water
level (e.g., Higginbotham 1997: 125-28, fig. 44, 45; Cunliffe 1971: 131) that would have
Figure 8.4. Roman concrete piscina at Torre Astura. B: main seawater inlet; F: freshwater
canal; G: breaches in perimeter mole to admit seawater; H: cistern for fresh water.
(Higginbotham 1997: 145, fig. 57.)
served very effectively to aerate the water and further enhance the oxygen level
(Boyd and Tucker 1998: 372 with table 7.3). Other features that seem to correspond
closely with modern practice include the creation of smaller tanks for spawning and
as nursery ponds for the growth of fingerlings (figure 8-4; Higginbotham 1997: 140-
51, figs. 55-61, 153-54, figs. 162-63; cf. Huet 1986: 6-7), and measures to recreate the
fish's natural habitat (Columella, Rust. 8.16.6-9.).
The artificial cultivation of oysters, mussels, and other shellfish was highly
developed, using techniques very similar to those used today (Higginbotham 1997:
125-28, figs. 44-45, 202-4, figs. 94-95, 210-13; Cunliffe 1971: 131). Aristotle is aware of
the scientific background (Aristotle, Gen. an. 763b), and the cultivation of shellfish
may well have been underway when he wrote (Del Rosso 1905: 147). Pliny credits
the invention of oyster cultivation to the enterprising Roman fish farmer C. Sergius
Orata, who created artificial oyster beds or ostriaria at Baiae and in the Lago
Lucrino (Lafon 2001: 177-79). A series of glass souvenir flasks (Kolendo 1977: fig. ia-c;
Ostrow 1979: 129, figs. 6-8) illustrate the famous Roman ostriaria along the sea
coast of Baiae. They illustrate a technology remarkably similar in form to modern
bouchot culture, or even more so to modern oyster culture at Taranto (Kolenda

1977: 120-22, fig. 8), or to Spanish rope culture (Hurlburt and Hurlburt 1980: 79-
85). Kolenda (1977: 122) and Pagano (1983-1984: 126) make the attractive suggestion
that these flasks show canals channeling fresh water into the ostriaria, as is done by
some contemporary French oyster farmers (Pagano 1983-1984: 126) to enhance the
growth and taste of the shellfish. Orata also cultured oysters on terracotta tegulae
(Valerius Maximus 9.1.1; Cicero, Hortensius fr. 69 Miiller), as is still done today
in Medoc (Pagano 1983-1984: 124 n. 49). The discovery of an oyster still fused to a
small piece of potsherd, identified in this case as part of an amphora, was likely the
result of this practice (Pinto-Guillaume 2001). Innovative methods of preserving
shellfish were also developed, similar to the practice of keeping harvested oysters
alive in shallow freshwater pools (pares a huitres) (Del Rosso 1905: 152-53; Abad
2002: 522-24), which revolutionized the consumption of shellfish in France in the
late eighteenth century (Abad 2002: 519-64, especially 549-61). Excavations near
Aquileia have revealed large numbers of oysters laid out carefully in pits irrigated
with fresh water from a nearby canal (Balista and Sainati 2003).
It is not surprising, therefore, that shellfish figured prominently in the Greco-
Roman diet. Oyster shells have been found on more than a hundred sites, mostly
associated with Roman villas or military encampments, in Germany, Switzerland,
and Austria alone (Thury 1990 ), and a single deposit of more than a million oysters
was found at Silchester (Sahrhage 2002: 101-3). Although no similar survey yet
exists for Italy, dozens of sites provide evidence for the consumption of oysters and
shellfish, occasionally in considerable quantity (Balista and Sainati 2003). One of
the few detailed regional studies, of southern Provence, shows a dramatic increase
in the consumption of shellfish, primarily the most popular cultivated species-
oysters, mussels, and scallops-following the Roman conquest, penetrating even
into very modest villages (Brien-Poitevin 1996).
REFERENCES
Abad, R. 2002. Le grand marche: L'approvisionnement alimentaire de Paris sous l'Ancien

Regime. Paris: Fayard.
Adams, J. N. 1995. Pelagonius and Latin veterinary terminology in the Roman Empire.
Leiden: Brill.
Aksakov, S. T. 1998. Notes of a provincial wildfowler. Evanston, IL: Northwestern University
Press.
Anderson, J. K. 1961. Ancient Greek horsemanship. Berkeley: University of California
Press.
Anderson, J. K. 1985. Hunting in the ancient world. Berkeley: University of California Press.
Andre, J. 1967. Les noms d'oiseaux en latin. Paris: C. Klincksieck.
Andre, J. 1981. L'alimentation et la cuisine a Rome. 2nd ed. Paris: C. Klincksieck.
Apostolides, N. C. 1883. La peche en Grece: Ichthyologie, migrations, engins et manieres de
peche, produits, statistique et legislation. Athens.
Aymard, A. 1951. Essai sur les chasses romaines des origines a la fin du siecle des Antonins.
Paris: Ecole Frarn;:aise de Rome.
Baatz, D. 1991. "Die ri:imische Jagdarmbrust," Archologisches Korrespondenzblatt 21: 283-90.
Badan, 0., J.-P. Brun, and G. Conges 1996. "Les bergeries romaines de la Crau d'Arles,"
Gallia 52: 263-310.
Balista, C., and C. Sainati 2003. "Ostrea non pectines ad Altino: Le evidenze arche-
ologiche," in G. C. Marrone and M. Tirelli (eds.), Produzioni, merci e commerci in
Altino preromana e romana: Atti del convegno, Venezia 12-14 dicembre 2001. Rome:
Qasar, 331-46
Bekker-Nielsen, T. 2002. "Nets, boats and fishing in the Roman world," Classica et Med-
ievalia 53: 215-33.
Bekker-Nielsen, T. (ed.) 2005a. Ancient fishing and fish processing in the Black Sea region.
Aarhus: Aarhus University Press.
Bekker-Nielsen, T. 2005b. "The technology and productivity of ancient sea fishing," in
Bekker-Nielsen 2005a: 83-96.
Benecke, N. 1994a. Der Mensch und seine Haustiere: Die Geschichte einer jahrtausendealten
Beziehung. Stuttgart: Theiss.
Benecke, N. 1994b. Archiiozoologische Studien zur Entwicklung der Haustierhaltung: In
Mitteleuropa und Sudskandinavien von den Anfiingen bis zum ausgehenden Mittelalter.
Berlin: Akademie Verlag.
Benecke, N., P. Donat et al. (eds.) 2003. Fruhgeschichte der Landwirtschaft in Deutschland.
LangenweiBbach: Beier and Baran.
Bevilacqua, E. 1987. "Diffusion des techniques de l'eau du monde antique au monde
contemporain: Les 'valli da pesca' dans la Lagune de Venise," in A. de Reparaz (ed.),
L'eau et des hommes en Miditirranie. Paris: CNRS, 67-76.
Binder, F. 1971. Organisation, production et controle de l'e1evage du gibier en France. Diss.
Ecole Nationale Veterinaire d'Alfort. Paris: Copedith.
Boisaubert, B., and J.-M. Boutin 1988. Le Chevreuil. Paris: Hatier.
Bi:iki:inyi, S. 1984. Animal husbandry and hunting in Ttic-Gorsium-The vertebrate fauna of a
Roman town in Pannonia. Budapest: Akademiai Kiad6.
Bi:iki:inyi, S. 1988. "Animal breeding on the Danube," in C.R. Whitakker (ed.),
Pastoral economies in classical antiquity. Cambridge: Cambridge Philological Society,
171-76.
Bonnet, G., and F. Klein 1991. Le Cerf. Paris: Hatier.
Bortuzzo, N. 1990. "11 leporarium: Un esempio di pastio villatica nella Gallia Cisalpina di
tarda eta repubblicana-prima eta imperiale? Alcune ipotesi sulla Transpadana or-
ientale," Aquileia Nostra 66: 85-112.
Bourdy, F. 1995. "L'ophthalmologie equine dans l'antiquite tardive d'apres Vegece," in
R. Chevallier and F. Bourdy (eds.), in Homme et animal dans l'antiquiti romaine.
Tours: Centre de recherches Andre Piganiol, 205-15.
Boyd, C. E., and C. S. Tucker 1998. Pond aquaculture water quality management. Boston:
Kluwer.
Brien-Poitevin, F. 1996. "Consommation des coquillages marins en Provern;:e dans l'epo-
que romaine," Revue Archiologique de la Narbonnaise 29: 313-20.
Brothwell, D. 1997. "Interpreting the immature chicken bones from the Romano-British
ritual complex on West Hill, Uley," International Journal of Osteoarchaeology 7.4: 330-32.
Brun, J.-P. 1996. "La grande transhumance a l'epoque Romaine apropos des recherches sur
la Crau d'Arles," Anthropozoologica 24: 31-44.
Capponi, F. 1959. "Le role de l'arundo dans l'oisellerie," Latomus 18: 724-41.
Carandini, A. 1984. Settefinestre: Una villa schiavistica nell'Etruria meridionale.Vol. 1, La

villa nel suo insieme. Modena: Panini.
Chamoux, F., and G. Hallier 1996. "Le colombier d'Apollonias," in Scritti di antichita in
onore di Sandro Stucchi. Vol. 1, La Cirenaica: La Grecia e l'Oriente mediterraneo. Rome:
L'Erma di Bretschneider, 51-60.
Chandezon, C. 2003. L'e1evage en Grece, fin Ve - fin Ier siecle a.C. L'apport des sources
epigraphiques. Bordeaux: Ausonius.
Chiliardi, S. 2000. "I cavalli siracusani in eta ellenistica," in Atti del 2 Convegno Nazionale di
Archeozoologia, Asti, 14-16 Novembre 1997. Forli: Abaco, 285-92.
Cleyet-Merle, J.-J. 1990. La prehistoire de la peche. Paris: Errance.
Coles, C. L. (ed.) 1971. The complete book ofgame conservation. London: Barrie and Jenkins.
Collin-Bouffier, S. 1999. "La pisciculture dans le monde grec: Etat de la question," Melanges
de l'Ecole Franraise de Rome m: 37-50.
Colonnelli, G., G. M. Carpaneto, and M. Cristaldi 2000. "Uso alimentare e allevamento
del ghiro (Myoxus glis) presso gli antichi romani: Materiale e documenti," in Atti del
2 Convegno Nazionale di Archeozoologia, Asti, 14-16 Novembre 1997. Forli: Abaco,
315-25.
Columeau, P. 2000. "Sacrifice et viande dans les sanctuaires grecs et chypriotes (VIie
siecle-Ier siecle av. J.C.) et l'apport de !'habitat de Kassope," Pallas 52: 147-66.
Conges, G. 1997. "Bergeries et transhumance dans la Crau antique: Innovation et adap-
tation," in D. Garcia and D. Meeks (eds.), Techniques et economie antiques et midi
evales: Le temps de l'innovation. Actes du colloque d'Aix-en-Provence, mai 1996. Paris:
Errance, 149-52.
Conta, G.D. 1972. "Note sulle pescherie marittime nel mondo romano," in G. Schmiedt
(ed.), Il livello antico del Mar Tirreno: Testimonianze dei resti archeologici. Florence:
Olschki, 215-36.
Corbier, M. 1999. "La transhumance: Aperyus historiographiques et acquis recents," in E.
Hermon (ed.), La question agraire a Rome: Droit romain et societe: Perceptions his-
toriques et historiographiques. Como: Edizioni New Press, 37-57.
Cox, N. 1686. The gentleman's recreation in four parts, viz. hunting, hawking, fowling, fishing.
London: Freeman Collins.
Cunliffe, B. 1971. Excavations at Fishbourne, 1961-69. London: Society of Antiquaries.
Cuvier, G., and M.A. Valenciennes 1828-1948. Histoire naturelle des poissons. 22 vols. Paris:
F. G. Levrault.
D'Arrigo, A. 1956. "La pesca del pescespada in Calabria dal secondo sec. av. Cr. ai nostri
tempi," Archivio Storico di Calabria e la Lucania 25 (1956) 101-21.
Davies, R. W. 1971. "The Roman military diet," Britannia 2: 122-42.
Davis, S. J. M. 1997. "The agricultural revolution in England: Some zoo-archaeological
evidence," Anthropozoologica 25-26: 413-28.
de Vries, J. 1974. The Dutch rural economy in the golden age, 1500-1700. New Haven: Yale
University Press.
De Angelis, R. 1959. Fishing installations in brackish lagoons, General Fisheries Council for the
Mediterranean. Rome: PAO.
De Caro, S., and M. Boriello (eds.) 2001. La natura morta nelle pitture e nei mosaici delle
citta vesuviane. Naples: Electa Napoli.
de Grossi Mazzorin, J., A. Tagliacozzo, and A. Riedel 1998. "Horse remains in Italy from the
Eneolithic to the Roman period," Proceedings of the XIII International Congress of
Prehistoric and Protohistoric Sciences, Forlt' (Italia) 8-14 september 1996. Forli: Abaco,
87-92.
Del Rosso, R. 1905. Pesche e peschiere antiche e moderne nell'Etruria Marittima. 2 vols.
Firenze: Osvaldo Paggi.
Delacour, J. 1964. The pheasants of the world. London: Country Life.
Desse-Berset, N. 1993. "Contenus d'amphores et surpeche; l'exemple du Sud-Perduto," in
J. Desse and F. Audoin-Rouzeau (eds.), Exploitation des animaux sauvages a travers
le temps: XIIIe Rencontres internationales d'archiologie et d'histoire d'Antibes, IVe
colloque international de l'homme et l'animal, 15-17 octobre 1992. Juan-les-Pins: Editions
APDCE, 341-46.
Dumont, J. 1976-1977. "La peche du thon a Byzance a l'epoque hellenistique," Revue des
Etudes Anciennes 78-79: 96-116.
Epplett, C. 2001. "The capture of animals by the Roman military," Greece and Rome 48:
210-22.
Escoffier, G.-A. 1984. Ma cuisine. New York: Random House.
Everynck, A., and K. Dobney 2002. "A pig for all seasons? Approaches to the assessment of
second farrowing in archaeological pig populations," Archaeofauna n: 7-22.
Farrington, E. I. 1913. The home poultry book. Toronto: McClelland and Goodchild.
Perlin, G. R. 1989. Elsevier's dictionary of the world's game and wildlife. Amsterdam: Elsevier.
Flach, D. 1990. Romische Agrargeschichte. Munich: Beck.
Fletcher, T. J. 1989. "Deer farming in Europe," in R. J. Hudson et al. (eds.), Wildlife pro-
duction systems: Economic utilisation of wild ungulates. Cambridge: Cambridge Uni-
versity Press, 323-33.
Food and Agricultural Organization 2002. FISHSTAT Plus: Universal Software for Fishery
Statistics Time Series, Version 2.3. Rome: PAO.
Forest, V., and I. Rodet-Belarbi. 2000. "Osteometrie et morphologie des bovins medievaux
et modernes en France meridionale," in M.-C. Marandet (ed.), L'homme et l'animal
dans les sociitis miditiraniennes: 4e journie d'itudes du Centre de Recherches Histor-
iques sur les Sociitis Miditerraniennes et du Pole Universitaire Europien de Montpellier
et du Languedoc-Roussillon. Perpignan: Presses Universitaires, 27-92.
Forest, V., and I. Rodet-Belarbi 2002. "A. propos de la corpulence des bovins en France
durant les periodes historiques," Gallia 59: 273-99.
Fox, M. W. 1984. Farm animals: Husbandry, behaviour, and veterinary practice (Viewpoints
of a critic). Baltimore: University Park Press.
Frayn, J. M. 1984. Sheep-rearing and the wool trade in Italy during the Roman period.
Liverpool: Cairns.
Frizell, B. S. 2004. "Curing the flock: The use of healing waters in Roman pastoral econ-
omy," in B. S. Frizell (ed.), Pecus: Man and animal in antiquity. Proceedings of the
conference of the Swedish Institute in Rome, September 9-12, 2002. Rome: Swedish
Institute in Rome, 84-97.
Gabriel, 0., K. Lange, E. Dahm, and T. Wendt (eds.) 2005. Von Brandt's fish catching
methods of the world. 4th ed. Oxford: Blackwell.
Gallant, T. W. 1985. A fisherman's tale. Gent: Peeters.
Georgoudi, S. 1990. Des chevaux et des boeufs dans le monde grec: Rialitis et representations
animalieres a partir des livres XVI et XVII des "Gioponiques." Paris and Athens: Daedalus.
Giacopini, L., B. B. Marchesini, and L. Rustico 1994. L'itticoltura nell'antichita. Rome:
Istituto Grafico Editoriale Romano.
Gianfrotta, P.A. 1999. "Archeologia subacquea e testimonianze di pesca," Me1anges de
l'Ecole franraise de Rome m: 9-36.
Gianfrotta, P.A. 2002. "Ponza (puntalizzazioni marittime)," Archeologia Subacqea 3:
67-90.
Gilbert, A. S. 2002. "The native fauna of the ancient Near East," in B. J. Collins, A history of
the animal world in the ancient Near East. Leiden: Brill, 3-78.
Goodall, D. M. 1977. A history of the domestic horse. London: R. Hale.
Grant, A. 1988. "Food status and religion in England in the Middle Ages: An archae-
ozoological perspective," in L. Bodson (ed.), L'animal dans l'alimentation humaine -
les criteres du choix. Paris: Anthropozoologica Numero Special, 139-46.
Guest-Papamanoli, A. 1986. "Ethnographie ou Ethnoarcheologie des ressources marines
de sites cotiers: Le cas de la peche aux muges dans les lagunes de la Grece Occidentale,"
in L'exploitation de la mer de l'antiquite a nos jours, I: La mer, lieu de production:
V Rencontres internationales d' archeologie et d'histoire d'Antibes, 24, 25, 26 octobre 1984.
Juan-les-Pines: Editions APDCA, 381-403.
Hansen-Catta, P.-H. 2002. Larousse de la chasse d'aujourd'hui. Paris: Larousse.
Heady, H. F., and R. D. Childs 1994. Rangeland ecology and management. Boulder, CO:
W estview Press.
Heath, M., D.S. Metcalfe, and R. F. Barnes 1973. Forages: The science of grassland agri-
culture. Ames: Iowa State University Press.
Higginbotham, J. 1997. Piscinae: Artificial fishponds in Roman Italy. Chapel Hill: University
of North Carolina Press.
Hodkinson, S. 1988. "Animal husbandry in the Greek polis," in C.R. Whitakker (ed.),
Pastoral economies in classical antiquity. Cambridge: Cambridge Philological Society,
35-74.
H0jte, J.M. 2005. "The archaeological evidence for fish processing in the Black Sea region,"
in Bekker-Nielsen 2005: 133-60.
Houpt, K. A. 2005. "Maintenance behaviours," in D. S. Mills and S. M. McDonnell
(eds.), The domestic horse: The origins, development and management of its behaviour.
Hudson, R. J. 1989. "History and technology," in R. J. Hudson et al. (eds.), Wildlife
production Systems: Economic utilisation of wild ungulates. Cambridge: Cambridge
University Press, 11-27.
Huet, M. 1986. Textbook of fish culture: Breeding and cultivation of fish. 2nd ed. Farnham:
Fishing News Books.
Hull, D. B. 1964. Hounds and hunting in ancient Greece. Chicago: University of Chicago
Press.
Hurlburt, C. G., and S. W. Hurlburt 1980. "European mussel culture technology and its
adaptability to North American waters," in R. A. Lutz (ed.), Mussel culture and harvest:
A North American perspective. Amsterdam: Elsevier, 69-98.
lvanova, N. V. 1994. "Fish remains from archaeological sites of the northern part of the
Black Sea region (Olvia, Berezan)," Offa 51: 278-83.
Jacobsen, A. L. L. 2005. "The reliability of fishing statistics as a source for fish stocks and
catches in antiquity," in Bekker-Nielsen 2005a: 97-104.
Jashemski, W. F., and F. G. Meyer (eds.) 2002. The natural history of Pompeii. Cambridge:
Kajava, M. 1998-1999. "Murenae, oysters and gilt-heads. Fish for name, table, and show in
ancient Rome," Acta Classica Debrecensis 34-35: 253-68.
Keller, 0. 1909-1913. Die antike Tierwelt. 2 vols. Leipzig: Engelmann.
Kerridge, E. 1967. The agricultural revolution. London: Allen and Unwin.
King, A. C. 1984. "Animal bones and the dietary identity of military and civilian groups
in Roman Britain, Germany and Gaul," in T. F. C. Blagg (ed.), Military and civilian in
Roman Britain. Oxford: BAR, 187-217.
King, A. C. 2002. "Mammals: Evidence from wall paintings, sculpture, mosaics, faunal
remains, and ancient literary sources," in Jashemski and Meyer 2002: 401-50.
Kokabi, M. 1988. "Viehhaltung und Jagd in romischen Rottweil," in Arae Flaviae W.
Stuttgart: Miiller and Graff, 105-234.
Kolendo, J. 1977. "Pares a huitres a Baiae sur un flacon en verre du Musee National de
Varsovie," Puteoli: Studi in storia antica 1: 108-12.
Kotjabopoulou, E., Y. Hamilakis, P. Halstead, et al. (eds.) 2003. Zooarchaeology in Greece:
Recent advances. British School at Athens Studies 9. Oxford: Oxbow.
Kron, G. 2000. "Roman ley-farming," Journal of Roman Archaeology 13: 277-97.
Kron, G. 2002. "Archaeozoology and the productivity of Roman livestock farming,"
Munstersche Beitriige zur antike Handelsgeschichte 21.2: 53-73.
Kron, G. 2004a. "Roman live-stock farming in southern Italy: The case against environ-
mental determinism," in M. Clavel-Leveque and E. Hermon (eds.), Espaces intigris et
gestion des ressources naturelles dans l'Empire romain. Franche-Comte: Presses Uni-
versitaires de Franche-Comte, 119-34.
Kron, G. 2004b. "A Deposit of carbonized hay from Oplontis and Roman fodder quality,"
Mouseion 4: 275-331.
Kron, G. forthcoming. "L'importance du poisson dans l'alimentation romaine et son
effet sur le developpement de la pisciculture a Rome," Revue en Ligne, Points de vue,
Chaire de recherche du Canada en interactions societe-environment dans l'empire
romain (http://www.chaire-rome.hst.ulaval.ca/revue.htm).
Lafon, X. 2001. Villa maritima: Recherches sur les villas littorales de l'Italie romaine, Ille siecle
av. J.C. - Ille siecle ap. J.C. Paris: Bibliotheque des Ecoles franc;aises d'Athenes et
de Rome.
Lamblard, J.-M. 1975. "Les etapes de la domestication de la Pintade Numida meleagris
Linne," in L'homme et l'animal: 1er colloque d'Ethnozoologie. Paris: Institut Interna-
tional d'Ethnosciences, 421-30.
Lauwerier, R. G. C. M. 1983. "Bird remains in Roman graves," Archaeofauna 2: 75-82.
Lauwerier, R. G. C. M. 1993. "Twenty-eight bird briskets in a pot: Roman preserved food
from Nijmegen," Archaeofauna 2: 101-13.
Leclainche, E. 1936. Histoire de la midecine vitirinaire. Paris. Office du Livre.
a
Leguilloux, M. 1999. "Sacrifices et repas publics dans le sanctuaire de Poseidon Tenos: Les
analyses archeozoologiques," Bulletin de Correspondence Hellenique 123: 423-55.
Leguilloux, M. 2000. "L'alimentation carnee au Ier millenaire avant J.C. en Grece con-
tinentale et dans les Cyclades: Premiers resultats archeozoologiques," Pallas 52: 69-95.
Lehman, P., and G. Breuer 1997. "The use-specific and social-topographical differences
found in the composition of animal species found in the Roman city of Augusta
Raurica (Switzerland)," Anthropozoologica 25-26: 487-94.
Lemming, G. E. 1969. "Nutrition and reproduction," in D. P. Cuthbertson (ed.), Nutrition
of animals of agricultural importance. Vol. 1, The science of nutrition of farm live-
stock. Oxford: Pergamon Press, 411-53.
Lepetz, S. 1996. L'animal dans la sociiti gallo-romaine de la France du Nord. Amiens: Revue
Archeologique de Picardie.
Leveau, P. 2004. "L'herbe et la pierre dans les textes anciens sur la Crau: Relire les sources
ecrites," Ecologia Mediterranea 30: 25-33.
Leveau, P., and M. Segard 2003. "Le pastoralisme en Gaule du Sud entre plaine et mon-
tagne: De la Crau aux Alpes du Sud," Pallas 55: 93-113.
Leveau, P., P. Sillieres, and J.-P. Vallat 1993. Campagnes de la Miditerranie Romaine:
Occident. Paris: Hachette.
Lindner K. 1973. Beitriige zu Vogelfang und Falknerei im Altertum. Berlin: de Gruyter.

Lirb, H. J. 1993. "Partners in Agriculture: The pooling of resources in rural societates in
Roman Italy," in H. Sancisi-Weerdenburg (ed.), De Agricultura: In memoriam Pieter
Willem de Neeve (1945-1990). Amsterdam: J. C. Gieben, 263-95.
Longo, 0. 1989. Le forme della predazione: Cacciatori e pescatori nella Grecia antica. Naples:
Liguori Editore.
Lubtchansky, N. 1998. "Le pecheur et la metis: Peche et statut social en Italie centrale a
l'epoque archa'ique," Me1anges de l'Ecole Francaise de Rome no: m-46.
Lugli, G. 1928. Forma Italiae Vol. 1, part 2, Ager Pomptinus, Circeii. Rome: Danesi.
MacGregor, A. 1996. "Swan rolls and beak markings: Husbandry, exploitation and regu-
lation of Cygnus olor in England, c. 1100-1900," Anthropozoologica 25-26: 39-68.
MacKinnon, M. R. 2001. "High on the hog: Linking zooarchaeological, literary and artistic
data for pig breeds in Roman Italy," American Journal of Archaeology 105: 649-73-
MacKinnon, M. R. 2004. Production and consumption of animals in Roman Italy: In-
tegrating the zooarchaeological and textual evidence. Journal of Roman Archaeology
Suppl. 54. Portsmouth, RI: JRA.
Magerstedt, A. F. 1859. Die Viehzuchtder Romer fur Archiiologen und wissenschaftliche ge-
bildete Landwirthe. Walluf: Sandig.
Mair, A. W. 1928. Oppian, Colluthus, Tryphiodorus, with an English translation. Cambridge,
MA: Harvard University Press.
Martin, B. P. 1993. Wildfowl of the British Isles and north-west Europe. Devon: Newton
Abbot.
Mastromarco, Giuseppe 1998. "La pesca del tonno nella Grecia antica: Dalla realta quo-
tidiana alla metafora poetica," Rivista di Cultura Classica e Medioevale 40: 229-36.
Mccann, A. M., J. Bourgeois, E. K. Gazda, J.P. Oleson, and E. L. Will 1987. The Roman port
and fishery of Gosa: A center of ancient trade. Princeton: Princeton University Press.
Ministry of Agriculture and Forestry, New Zealand 1994. "Market dynamics for venison,"
http://www.maf.gov.nz/mafnet/rural-nz/profitability-and-economics/structural-
change/market-dynamics-for-venison/ conf4-03.htm.
Montagne, P. 1961. Larousse Gastronomique. New York: Crown Publishers.
Moriceau, J.-M. 1999. L'e1evage sous l'Ancien Regime: Les fondements agraires de la France
moderne, XVIe-XVIIIe siecles. Paris: SEDES.
Morris, P. 1979. Agricultural buildings in Roman Britain. British Archaeological Reports 70.
Oxford: BAR.
Moule, L. 1891. Histoire de la medecine veterinaire. Premiere periode: Histoire de la medecine
veterinaire dans l'antiquite. Paris: Bulletin de la Societe Centrale de Medecine
Veterinaire.
Naether, C. A. 1964. The Book of the pigeon and of wild foreign dove. 5th ed. New York:
D. McKay.
Olive, C., and S. Deschler-Erb 1999. "Poulets de grain et rotis du cerf: Produits de luxe pour
les villes Romaines," Archiiologie der Schweiz 22: 35-38.
Olson, S. D., and A. Sens 2000. Archestratos of Gela: Greek culture and cuisine in the fourth
century BCE. Oxford: Oxford University Press.
Osborne, R. 1987. Classical landscape with figures: Ancient Greek city-states and their
countryside. London: Sheridan House.
Ostrow, S. 1979. The topography of Puteoli and Baiae on the eight glass flasks," Puteoli:
Studi di Storia Antica 3: 77-140.
Pagano, M. 1983-1984. "11 Iago Lucrino: Ricerche storiche e archeologiche," Puteoli: Studi di
Storia Antica 7-8: 113-226.
Parker, A. J. 1988. "The birds of Roman Britain," Oxford Journal ofArchaeology 7: 197-226.
Pasquinucci, M. 2004. "Montagna e pianura: transumanza e allevamento," in M. Clavel-
Leveque and E. Hermon (eds.), Espaces integres et gestion des ressources naturelles
dans l'Empire romain. Franche-Comte: Presses Universitaires de Franche-Comte,
165-78.
Pattison, I. 1984. The British veterinary profession, 1791-1948. London: Allen.
Pawson, H. C. 1957. Robert Bakewell: Pioneer livestock breeder. London: Lockwood.
Payne, S. 1985. "Zoo-archaeology in Greece: A reader's guide," in N. C. Wilkie and W. D. E.
Coulson (eds.), Contributions to Aegean archaeology: Studies in honor of William A.
McDonald. Dubuque, IA: Kendall/Hunt Publishing, 211-44.
Peters, J. 1998. Romische Tierhaltung und Tierzucht: Eine Synthese aus archaozoologischer
Untersuchung und schriftlich-bildlicher Uberlieferung. Rahden: Leidorf.
Pinto-Guillaume, E. M. 2001. "Observations on a very singular oyster," in G. Missineo
(ed.), Ad Gallinas Albas: Villa di Livia. Rome: L'Erma di Bretschneider, 209-11.
Pleket, H. W. 1993. "Agriculture in the Roman Empire in comparative perspective," in H.
Sancisi-Weerdenburg (ed.), De Agricultura: In memoriam Pieter Willem de Neeve
(1945-1990). Amsterdam: Gieben, 214-37.
Pollard, J. 1977. Birds in Greek life and myth. London: Thames and Hudson.
Prummel, W. 1987. "Poultry and fowling at the Roman era castellum," Palaeohistoria 29:
183-201.
Purcell, N. 1995. "Eating fish: The paradoxes of seafood," in J. Wilkins, D. Harvey, and M.
Dobson (eds.), Food in antiquity. Exeter: Exeter University Press, 132-49.
Purpura, G. 1992. "Pesca e stabilimenti antichi per la lavorazione di pesce nella Sicilia
occidentale: IV un bilancio," in V Rassegna di archeologia subacquea, Atti (Giardini
Naxos 19-21 ottobre 1990). Messina: Edizioni P&M Associati, 87-101.
Quilici-Gigli, S. 1989. "Paesaggi storici dell'agro falisco: I prata di Corchiano," Opuscula
Romana 17: 123-35.
Reese, D.S. 1994. "Recent work in Greek zooarchaeology," in P. N. Kardoulias (ed.),
Beyond the site: Regional studies in the Aegean area. Lanham, MD: University Presses of
America, 191-221.
Reese, D.S. 2002a. "Fish: Evidence from specimens, mosaics, wall paintings, and Roman
authors," in Jashemski and Meyer 2002: 274-91.
Reese, D. S. 2002b. "Marine invertebrates, freshwater shells, and land snails: Evidence from
specimens, mosaics, wall paintings, sculpture, jewelry, and Roman authors," in Ja-
shemski and Meyer 2002: 292-314.
Rendini, P. 1997. "Vasi per la pesca del polpo?" inAtti del convegno nazionale di archeologia
subacquea, Anzio 30-31 maggio e 1° giugno 1996. Bari: Edipuglia, 75-78.
Ridgeway, W. 1905. Origin and influence of the thoroughbred horse. Cambridge: Cambridge
University Press.
Riedel, A. 1994. "Archaeozoological investigations in north-eastern Italy: The exploitation
of animals since the Neolithic," Preistoria Alpina 30: 43-94.
Rinkewitz, W. 1984. Pastio Villatica: Untersuchungen zur intensiven Hoftierhaltung in der
romischen Landwirtschaft. Frankfurt am Main: Lang.
Rocher, C. 1979. Les chasses des palombes et des tourterelles. Bordeaux: Editions de l'Oree.
Romdhane, M. S. 1998. "La peche artisanale en Tunisie: Evolution des techniques ances-
trales," Me1anges de l'Ecole Franraise d'Athenes no: 61-80.
Rosada, G. 2000. "La centuriazione di Padova nord (Cittadella-Bassano) come assetto
territoriale e sfruttamento delle risorse: Una riflessione dallo studio di Plinio Frac-
caro," Aquileia Nostra 71: 85-122.
Rossi, R. 1990. "Gli atrezzi di pesca," in F. Berti (ed.), Fortuna maris: La nave romana di
Cornacchio. Bologna: Nuova Alfa Editoriale, 114-15.
Ryder, M. L. 1983. Sheep and man. London: Duckworth.
Ryder, M. L., and J. W. Hedges 1973. "Ancient Scythian wool from the Crimea," Nature
242: 480.
Sahrhage, D. 2002. Die Schiitze Neptuns: Eine Kulturgeschichte der Fischerei im romischen
Reich. Frankfurt am Main: Lang.
Sallares, R. M. 1991. The ecology of the ancient Greek world. London: Duckworth.
Salza Prina Ricotti, E. 1987. "Alimentazione, cibi, tavola e cucine nell'eta imperiale," in
L'alimentazione nel mondo antico. Rome: Istituto poligrafico e Zecca dello Stato,
71-130.
Salza Prina Ricotti, E. 1999. "L'importanza del pesce nella vita, nel costume e nell'industria
del mondo antico," Rendiconti: Atti della Pontificia Accademia Romana di Archeologia
71: 111-65.
Schmiedt, G. 1972. Il livello antico del Mar Tirreno: Testimonianze dei resti archeologici.
Florence: Olschki.
Senet, A. 1953. Histoire de la medecine veterinaire. Paris: PUP.
Shepherd, C. J., and N. R. Bromage 1988. Intensive fish farming. Oxford: Blackwell.
Shepherd, E. J., and L. Dallai 2003. "Attivita di pesca al promontorio di Piombino (I
sec.a.C.-XI sec.d.C.)," in Atti del II Convegno di Archeologia Subacquea, Castiglioncello
7-9 settembre 2001. Bari: Edipuglia, 189-207.
Sicsi, R. 1992. "La pesca nell'area dello Stretto di Messina nell'antichita: Continuita fra
presente e passato," in V Rassegna di archeologia subacquea. V Premio Franco Papo.
Atti. Giardini Naxos 19-21 ottobre 1990. Messina: P&M Associati, 127-46.
Sternberg, M. 1995. La peche a Lattes dans l'antiquite a travers l' analyse du ichtyofaune.
Lattes: Lattara.
Sternberg, M. 1998. "Les produits de la peche et la modification des structures halieutiques
en Gaule Narbonnaise du Ille siecle av. J.-C. au Ier siecle ap. J.-C. Les donnees de
Lattes (Herault), Marseille (Bouches-du-Rhone) et Olbia-de-Provence (Var)," Me1-
anges de l'Ecole Franraise de Rome et d'Athenes no: 81-109.
Sternberg, M. 2005a. "La peche," in M.-P. Rothe and H. Treziny (eds.), Marseille et ses
alentours. Carte Archeologique de la Gaule 13/3. Paris: Belles Lettres, 245-47.
Sternberg, M. 2005b. "L'elevage," in M.-P. Rothe and H. Treziny (eds.), Marseille et ses
alentours. Carte Archeologique de la Gaule 13/3. Paris: Belles Lettres, 247-49.
Tammisto, A. 1997. Birds in mosaics: A study on the representation of birds in Hellenistic and
Romano-Campanian tessellated mosaics to the early Augustan age. Rome: Gummerus
Kirjapano Oy.
Thompson, E. P. 1975. Whigs and hunters: The origin of the Black Act. London: Lane Allen.
Thompson, D. W. 1936. A glossary of Greek birds, 2nd ed. London: Oxford University
Press.
Thiiry, G. E. 1990. "Romische Austerfunde in der Schweiz, im rechtsrheinischen Siid-
deutschland und in 6sterreich," in Festschrift fur Hans R. Stampfli - Beitriige zur
Archiiozoologie, Archiiologie, Anthropologie, Geologie und Paliiontologie. Basel: Helbing
and Lichtenhahn, 285-301.
Toynbee, J.M. C. 1973. Animals in Roman life and art. Ithaca, NY: Cornell University Press.
Trakadas, A. 2005. "The archaeological evidence for fish processing in the western Med-
iterranean," in Bekker-Nielsen 2005a: 47-82.
Trench, C. C. 1967. The poacher and the squire: A history of poaching and game preservation
in England. London: Longmans.
Trotta, F. 1990. "Le reti di pesca nell'antichita classica," in Quel filo azzuro tra l'uomo e il
lago. Bergamo: Gruppo Editoriale Walk Over, 34-46.
Trotta, F. 1996. "La pesca nel mare di Magna Grecia e Sicilia," in F. Prontera (ed.), La
Magna Grecia e il mare: Studi di storia marittima. Taranto: Istituto per la Storia e
L' Archeologia della Magna Grecia, 227-50.
Trow-Smith, R. 1959. A history ofBritish livestock husbandry, 1700-1900. London: Routledge
and Kegan Paul.
Udrescu, M., and W. Van Neer 2005. "Looking for human tlierapeutic intervention in
tlie healing of fractures of domestic animals," in J. Davies et al. (eds.), Diet and health
in past animal populations: Current research and future directions. Oxford: Oxbow,
24-33.
Vigneron, P. 1968. Le cheval dans l'antiquite greco-romaine: Des guerres mediques aux
grandes invasions. Contribution a l'histoire des techniques. Nancy: Faculte des lettres et
des sciences humaines de l'Universite.
Watson, G. E. 2002. "Birds: Evidence from wall paintings, mosaics, sculpture, skeletal
remains and ancient autliors," in Jashemski and Meyer 2002: 357-400.
White, K. D. 1970. Roman farming. New York: Thames and Hudson.
Wynmalen, H. 1950. Horse breeding and stud management. London: Country Life.
Youatt, William. 1837. Sheep: Their breeds, management, and diseases. London: Baldwin and
Craddock.
Zeuner, F. E. 1963. A history of domesticated animals. London: Hutchinson.
Zissu, B. 1995. "Two Herodian dovecotes: Horvat Abu Haf and Horvat 'Aleq," in The
Roman and Byzantine Near East: Some recent archaeological research. Journal of Roman
Archaeology Suppl. 14. Ann Arbor, MI: JRA, 57-69.
PART III
ENGINEERING
AND COMPLEX
MACHINES
CHAPTER 9
GREEK ENGINEERING
AND CONSTRUCTION
FREDERICK A. COOPER
THis discussion for the most part deals with public structures, in particular temple
architecture. This bias results from the emphasis the Greeks themselves placed on
public structures at the expense of private homes and commercial establishments, a
focus which resulted in structures that pushed the boundaries of the materials and
the procedures for design and construction. Only in the mid-Hellenistic period did
this pattern begin to change. The best temples, stoas, and theaters represent vividly
what the Greek architect and engineer could accomplish, making them paradigms
for a study of Greek engineering technology. Many of the modern handbooks cited
in this chapter provide further information on domestic and utilitarian architec-
ture, and several chapters in this book will provide further information on the con-
siderable Greek expertise in hydraulic engineering (chapter 11), tunnel and canal
construction (chapter 12), construction cranes (chapter 13), and harbor design
(chapter 25). Since ancient Greek architects based their designs on engineering and
structural theory and practice, the organization of this chapter intentionally fol-
lows the section topics typical of a modern engineering and construction hand-
book (e.g., Merritt and Rickets 2000; cf. Merritt 1975, 1976).
As early as Theodorus of Samos (ca. 575 B.C.; Vitruvius, De arch. 7, preface 12),
architects wrote books on architecture, presumably treatises setting forth their
ideas on the facts and theory of construction. The handbook De architectura by the
Roman architect Vitruvius (late first century B.C.) cites earlier texts, but his own
approach emphasizes architectural design, especially the history and use of the
canons, the application of proportions and modular dimensions, designs dictated
by function, architectural refinements, and the advantages of siting buildings for
exploitation of sun, wind, and natural resources. Vitruvius devotes shorter sections
226 ENGINEERING AND COMPLEX MACHINES
to building materials, construction machinery, foundations, and surveying. The

former topics are not customarily included in a builder's engineering handbook,
but the latter certainly are. From the time of the renaissance scholarship on Greek
architecture, the study of Greek architecture has been wedded to Vitruvian precepts
on appearance and proportion and the spatial planning of buildings as a point of
departure to a study of the topic. But while Vitruvius touches on topics that extend
from mathematics to medicine, the fact is that he provides little insight for an un-
derstanding of the architectural engineering that went into ancient Greek building
practice.
I propose that the engineering observed in ancient building practice expresses
a coherent set of rules and procedures founded in experience and preserved and
passed along in an oral tradition, or better, by written handbooks. The works of
Theophrastus (ca. 370-288/5 B.C.), pupil and successor to Aristotle, especially his
Inquiry into Plants (Peri phyton, or Historia plantarum) and On Stones (Peri lithon,
or De lapidibus), provide clear evidence for a body of scientific theory and engi-
neering practice behind the applications of scientific technology to building. The
distinction between engineering and construction technology should be empha-
sized. Certainly the theories embodied in engineering lead to the technology, but
the reverse process does not always follow. Empirical applications of a technology,
even a sophisticated technology, can happen in the absence of engineering theory
and computation.
The invaluable surveys of Martin (1965), Orlandos (1966-1968), White (1984),
Coulton (1977), and Hellmann (2002) provide a wealth of evidence for the tech-
nologies behind the practice of architecture in ancient Greece. Landels (2000: 210)
omits architecture but compliments Vitruvius for his "common sense and guess-
work." Published abstracts by the occasional symposium on ancient Greek tech-
nology recognize the breadth and range of ancient Greek technology in all its
aspects; few, however, address whether or not this technology comes at the front
end of an underlying scientific, engineering practice (e.g., Tassios and Palyvou
2006). This chapter rests on the proposition that construction theory, especially the
mechanics of building materials-principally wood and stone, but others as well-
and aseismic design constituted the starting point for Greek architectural design.
Proportions, siting, propriety of architectural decoration, and use of the orders
certainly played a part in the design process, but the genius lay in the engineering,
not in the tweaking of style. My aim is to show the existence of an ancient scientific
approach through juxtaposition of the evidence for Greek practice with pertinent
excerpts from contemporary engineering construction handbooks.
For example, Theophrastus' Inquiry into Plants, especially book 5, contains all
the earmarks of a modern-day handbook on wood construction. He has collected
testimonia from lumbermen: lumberjacks, sawyers, jobbers, carpenters, wholesal-
ers, and architektones, specialists working with species of wood indigenous to their
home regions. Theophrastus understood that various grades of the same wood, or
the same timber cut in different ways, can have different strengths and character-
GREEK ENGINEERING AND CONSTRUCTION 227
istics when used as a building material. He set out clear guidelines as to appropriate
use of species of oak-kermes, cork, holm, and turkey-noting that some oaks are
hard, some easily worked, while others are only good for firewood (cf. also chapters
10 and 17).
BACKGROUND:THE
ARCHAEOLOGICAL EVIDENCE
My involvement in the inventory and recovery of demolished ancient Greek build-

ings has led in all cases to reconstructions on paper and to actual reconstruc-
tion or anastylosis for some (F. A. Cooper 1992-1996). This process has afforded
many opportunities to observe the ancient architect's understanding of founda-
tions and aseismic design, strength of materials, paving, ceiling and roof systems-
a set of topics far closer to what is found in a modern building engineering
and construction handbook than to what Vitruvius presents (F. A. Cooper 1983,
1999). Another modern-day experience is relevant. With the AutoCAD software
application it is possible to program scripts in Auto LISP. In using this system, my
purpose was to generate plan and elevation drawings of not only the standard
corpus of Greek temples, but also the lesser known and the newly discovered.
The AutoLISP program prompts for basic dimensions of the kind found in the
tables in Dinsmoor (1950: fold-out chart) and Robertson (1964: 322-46): length
and width of stylobate, facade, and flank; bottom column diameters; height of
colonnade and entablature. Missing measurements can be supplied by secondary
information, such as proportions indexed by approximate date. The program
computes a Drawing Unit and generates a plan and an elevation. A plan of the cella
and other details are added to the AutoCAD drawing. This computer generation of
a plan and elevation simply automates what the ancient Greek architect must
have done routinely: manipulation of proportions and modules, alteration and
improvement of details such as the glyph design or an extra triglyph over the in-
tercolumniation. This was the humdrum part of architectural practice; the real
expertise lay in the engineering that underlay the design of structurally successful
Greek architecture.
The scholarship on architectural construction in ancient Greece frequently
reflects the widely held but erroneous assumption that architects adhered to a pan-
Hellenic standard of measurement or an absolute length (or lengths), such as a
Doric or Ionic foot or an Ionian ell. These measurements, however, varied from
town to town and from project to project, but were usually a value between 0.27 m
and 0.36 m (Pakkanen 2004, 2006: 227; cf. chapter 30, part 1).
DESIGN STAGE
Vitruvius (De arch. 3.4.5) writes of stylobate curves in temples, achieved by the use
of scamilli impares, and other kinds of subtle deflections from the true: the tilting
and curving of verticals and horizontal lines and planes, called refinements (Ha-
selberger 1999). These deflections began to appear in the sixth century B.C. as
exaggerations visible to the naked eye; over time the refinements became increas-
ingly subtle, and not perceptible except by careful measurement with modern in-
struments. Such refinements required the utmost control in surveying. On the one
hand, there is no apparent structural or visual necessity; on the other hand, the
slightest amount of settlement in any of a building's components renders useless
the optimal effort and care to achieve one or more of the refinements. In other
words, a stable foundation construction permitted successful, in the sense of long-
lasting, architectural refinements. There is a standing assumption that earthquakes
felled many buildings across the Greek mainland. In fact, in most cases, human
demolition, not natural forces, leveled Greek architecture.
Structural Design
Coulton (1977: 140) provides a sweeping judgment of Greek architectural design:
"Structural conservatism is inherent in the Greek conception of architecture as
concerned primarily with external form rather than internal space.... Changes in
structure could not make it cheaper." Coulton qualifies this statement by brief
references to quality control of stone and to the evolution of clamp design. At the
same time, he and others overlook the impressive accomplishments of the Greek
architect as engineer: the loading of a colonnade on a continuous base on fill under
a foundation requires an understanding of the engineering problems of differential
loading-stress. This is just one aspect of engineering in which the Greek architect
continuously made improvements over time. It should also be stressed that, with a
few exceptions, successful engineering applications did not come and go. The sim-
ple truss system, for example, seems to appear in structures only during the Roman
Empire (cf. chapters 10 and 17)
Business, Art, and the Profession of Greek Architecture

The financial practice of an ancient Greek architect differed little from that of a
successful architectural practice today: a professional who manages his personal
finances and the budget of his project. Some ancient contracts contained provi-
sions to award the architect for a project coming under budget, while specifying a
penalty and personal assessment for projects more than a quarter over estimate.
The reasons for a client's choice of an architect in ancient Greece are as unfath-
omable as a client's motivations now. Reputation counted a great deal, but there
may have been a preference for a local professional-such as Libon of Elis, "a local
man" according to Pausanias (5-10.3), for the design of the temple of Zeus at
Olympia (ca. 457 B.C.). More broadly, reconstruction of the venerable temple of
Apollo at Delphi, destroyed by a landslide in 373 B.C., was undertaken by the
Amphictyonic League, an alliance of Greek states, which raised funds from pub-
lic and private sources. The league appointed Spintharus of Corinth as architect
(Pausanias 10.5.13). Construction lingered for 40 years, and successor architects
included Xenodorus and Agathon.
Among the several hundred ancient Greek architectural projects known by
name, many were collaborations; for example, Antistates, Callaeschrus, Antima-
chides, and Porinus for the design of the giant temple of Zeus Olympius at Athens
(ca. 525 B.C.; Vitruvius, De arch. 7, preface 15). These architects were members of
the intellectual circle of the Pisistratids, which included philosophers. A bias to-
ward teams of intellectuals certainly persisted to the age of Pericles, who engaged
the likes of Phidias, Ictinus, Callicrates, and other artists and architects in the
building programs in Athens in the third quarter of the fifth century B.C. (Plu-
tarch, Per. 12.1-13.8). There are instances in which slaves were appointed architects
for a major temple. Examples include Demetrias of Ephesus (ca. 350 B.C.; Vi-
truvius, De arch. 7, preface 16) and, quite possibly, Ictinus, architect of the Par-
thenon, a phase of the Telesterion at Eleusis, and the temple of Apollo at Bassae (ca.
450-410 B.C.; F. A. Cooper 1992-1996: I, 377). In antiquity, as today, the term ar-
chitektones (architect) could indicate any number of roles: master designer, princi-
pal contractor, overseer, master craftsman, architect-builder, or builder-architect.
"Architect" is here taken to mean the professional who contributed to all aspects of
the building project from concept and design through execution (Coulton 1977: 15;
Mccredie 1979).
The ancient architect-designer often produced a model or drawing for public
discussion and approval (Coulton 1977: 51-73; Petronotis 1972; Haselberger 1997). A
few inscribed decrees survive in which the architect is directed to draw up speci-
fications for construction: temple of Athena Nike (JG 13 35), the arsenal of Philo
(JG 22 1668), and a temple at Lebadia (JG 7 3073; Martin 1965: 189-90). Equally
common, inscribed accounts of a building project were put on public display, the
most famous being those for the Parthenon (JG I3 436-51; Pollitt 1990: 185-205).
GEOTECHNICAL ENGINEERING
The built structures of mainland Greece are in various conditions of preservation.

It is widely assumed that the natural phenomenon of earthquakes was responsible
for leveling the many buildings of ancient Greece, but this is not the case. High
winds toppled more columns than did earthquakes, and those buildings in the
worst shape were demolished in the late antique period to recover the metal clamps
and dowels. Had human destruction not intervened, much ancient Greek archi-
tecture would still be standing today, as do the temple at Bassae, the Hephaisteion
in Athens, and a number of Greek temples in Italy and Turkey. Along with the
Parthenon, these structures were engineered far beyond what any contemporary
engineer or architect would call the test of time, putting them into a special class
"of the best final design and construction" (Ambrayses 1971: 379).
The ancient Greeks held the sanctity of site, prominence on the skyline, and
situation within a sanctuary to be of greater importance than the quality of soil and
rock on which a temple or building was constructed. Unavoidable and unpredict-
able soil conditions at building locations led to geotechnical engineering, sophis-
ticated even by modern standards. The fifth-century B.C. architect did not have at
his disposal tests for rock mechanics, formulae for load distributions, or tables of
presumptive standards for soil types. In fact, until the 1950s, architects and engi-
neers could only hope that rock conditions would not cause trouble, and "most
foundation designs were based on assumed properties deduced from surface ob-
servation" (Harvey 1982: I).
Foundation Design
The accumulated evidence at most Greek sites points to architectural components
designed or laid to maximize the geotechnical properties of the stone: foundations
that double the bearing capacity of the poor bedrock underneath. The design rep-
resents as well the highly successful drainage of water off the foundations. Fre-
quently, limestones were selected that had nearly the same appearance as the local
limestone, but double the bearing strength. There was also systematic quarrying
and coursing of ashlars and drums, keeping the planar structure of the stone hor-
izontal when in compression, and vertical for suspension in unsupported spans.
The list is impressive and consistent: the architect knew and grasped geotechnical
engineering and the mechanics of his materials.
The archaeological evidence for a survey of ancient Greek engineering practice
in sub-foundation preparation is limited by the failure of early excavators to
excavate properly ancient backfills against the perimeters of buildings, a difficult
undertaking. Nonetheless, in those cases where excavation reports are sufficiently
detailed, there is evidence for a mat foundation, consisting of large rock fragments
and boulders contained by cribbing walls for lateral stabilization, which isolates the
platform from bedrock. Often there was a bed of silty clay or soil, alone or in con-
junction with the stone layer. A mat foundation with a spread footing placed within
it represents an excellent solution to the problem of a low bearing capacity of the
terrain, but its function goes far beyond this single purpose. This type of foun-
dation also incorporates an aseismic design called "base isolation," an alternative
to "conventional methods [that] seek to protect buildings against earthquake
attack by increasing the strength of structures and their capacity to dissipate en-
ergy. Seismic regulations require that the earthquake attack be absorbed by in-
elastic action of the structural system" (Merritt 1976: arts. 6-99, 6-116; cf. Merritt
1975: arts. 3-84). Kelly (1982: 17) sets forth the principles of this alternative system,
aseismic base isolation, in which the "base reduces the transmission of horizontal
acceleration into the structure."
In the mid-sixth century B.C., the sculptor and all-around inventor Theodorus
of Samos acquired additional fame for his substructure beneath the temple of Ar-
temis at Ephesus. Pliny (HN 36.95-97) gives credit to Theodorus' expertise: "[The
temple of Ephesus] was constructed on marshy ground so that it would be neither
subject to earthquakes nor in danger of earth-slips; on the other hand in order that
such massive work would not be placed on shifting and unstable foundations, they
were underlain with a layer of packed charcoal, and after that with a layer of sheep-
skins with the fleece." Diogenes Laertius (2.103) gives an abbreviated account in
which Theodorus "advised the placing of a layer of ashes beneath the foundations
of the temple in Ephesus."
In 1870, John T. Wood sank four trenches at the temple of Artemis, three in the
cella and one in the northern peristyle, where he found a four-inch-thick (ca. 0.10
m) layer of "putty" (a form of mortar?) underneath a layer of charcoal, and be-
neath that another layer of "putty" (Hogarth 1908: 10-11). The layers have not been
detected in subsequent excavations, nor have the wooly sheepskins mentioned by
Pliny, but the deposits of charcoal alternating with other materials, including
marble chips, appear in cross-sections and hint at an engineering solution of "base
isolation" (cf. Bammer 1984: figs. 78, 79, 82, 113) . Bammer's so-called Hekatempedos
Figure 9.1. Delphi, massive settlement of walls. (Photograph by F. Cooper.)

0 I 5 1011
Figure 9.2. Polygonal foundation stonework, Demetrias temple. (F. Cooper, adapted
from Bequignon 1937: pl. 3.)
looks more like a layer of packed stone fill (Bammer 1984-85: 13-28, figs. 1, 17, 20; cf.
Vetters 1986: 78-79, figs. 1-4).
The terrace that supported the temple of Apollo at Delphi quite likely retained
large unhewn stones and reused building blocks in a mat foundation comparable
to that at Bassae and elsewhere. A mass of large stones still lies behind the southeast
inside corner of the polygonal wall at Delphi (Algren-Ussing and Bramnaes 1975:
N180-N190; W130-W170); the area, however, had been used as a dump for miscel-
laneous excavated material (Amandry 1981: 679) . A gap or an interlayer of earth be-
tween the Acropolis bedrock and the foundation course of the Parthenon also in-
dicates a possible design of base isolation, as do photographs and drawings of the
sub-foundations along the south side. Here thick layers of huge stones, both unhewn
blocks and discarded building material, were interlayered with earth (Kawerau et al.
1974: figs. 38, 59-62, 82; Tschira 1972: pls. 1-2; Korres 1983: 11-12, 667; Orlandos 1977:
vol. 2, fig. 75). A mat foundation similar to those just described underlies the temple
of Nemesis at Rhamnus. "Temporary" or cribbing walls inside the outer retaining
walls stabilized the large rock fill. Petrakos (1983: 10) exposed the fabric of "an
artificially made terrace, supported on the north and east by strong isodomic re-
taining walls erected in the fifth century B.C. before the large temple was built."
Merritt's engineering handbook (1976: art. 7-7) states that "A successful design
distributes concentrated loads in order to inhibit differential settlements." A spread
footing, running underneath the peristyle and common nearly to all Greek designs,
achieves this very end with the additional advantage of transmitting the concen-
trated loads at the intercolumniations over as wide an area as possible. This design
reduces the weight-to-area ratio appreciably and ensures that the maximum loads
do not exceed the bearing capacity of a usually unstable rock terrain (Herubin and
b.
Figure 9.3. a. Shifted column drums, Hephaisteion, Athens; b. Resistance of Doric col-
umns to seismic stress. (F. Cooper, adapted from Ambraseys 1985.)
Marotta 1987: 43-7). Sometime in the fourth century B.C., soft, brown conglom-
erate or a red breccia became a standard material for sub-foundation construction,
especially in Attica. Where the conglomerate was not placed underground, it was
encased by a durable limestone facing, often of Peireus poros. A crumbly volcanic
rock also became a common foundation material (Thompson and Wycherly 1972:
37, 80, 104, 129, 160-61, 167; Kousell and Dimou 2006: 29). The conglomerate or
breccia consists of a coarse, reddish material, tending to degrade when exposed
to the weather, but, like a rubble stone core, absorbing stresses from settlement
or seismic waves by means of shifts brought about by micro-fracturing. The con-
glomerate base served the same engineering design principle as the aseismic mat
foundation mentioned earlier.
The skewed and undamped jointing of blocks in basement courses (biased
jointing) retards differential settlement in the area where settlement is most likely
to happen (F. A. Cooper 1992-1996: vol. 3, pl. 39.b--d; vol. 4, pl. 9). Polygonal
jointing was not uncommon in the construction of foundations. It was used, for
instance, in the later temple of Aphaea on Aegina and in the Tholos at Delphi. In
all cases, the joint work in the substructure does not mesh with the jointing in
the platform, thus distributing subsidence evenly over a wide area rather than
through a seam of vertical joints. All too often, this polygonal stonework in
foundation courses has been associated with archaic masonry technique (Shear
1971: 243-55), but this design was structural, not stylistic. The technique was used
for foundation jointing into the Hellenistic period (Stahlin et al. 1934: pls. 47-48),
e.g., for the foundations to the temple of Artemis Agoria at Demetrias (Bequignon
1937: pl. 3) for the temple of Zeus at Pherae, and for the temple at Hermione
(figures 9.1-2).
Hydraulics and Building Construction

Drainage is crucial to foundation stabilization procedures because the strength of
soil generally decreases with an increase in water content. Retention of ground
water produces several adverse effects: pore water reduces the cohesiveness of soil
particles, increases plasticity, and thereby increases the potential for slippage
(Merritt 1975: art. 4-11; 1976: art. 7-15). When loaded, soil consolidation takes
place through the voiding of water, often resulting in unpredictable settlement
and deformation. Moreover, saturation of certain rocks, such as limestones, can
reduce compressive strengths (Vutukuri 1974: 50-57). Tests have demonstrated
that the temple podia at Bassae and Messene were constructed so as to dissipate all
ground water rapidly and completely. At both sites and elsewhere the subfounda-
tion consists of alternating layers of earthy clay and small cobblestones. Tests con-
sisted of pumping water under pressure into borings and a systematic observation
of the effects. The mat foundation design of other structures provided this benefit
as well.
Comparison of an ancient solution with contemporary geotechnical proce-
dures suggests a remarkable sophistication in ancient engineering knowledge, de-
spite the absence of most formulas in the fifth century B.C. The juxtaposition
highlights the formidable feats of engineering on the part of a temple's architect,
who gained insight into the problems of good foundation design through empir-
ical observation and experiment. These observations make it clear that the ancient
builder had a practical, scientific knowledge of geotechnical engineering. Empirical
experience over time was the guide, rather than mathematical proof.
PERISTYLE AND CELLA AS

INDEPENDENT UNITS
Segmented drums form an elastic design of tightly fitted, undowelled drums and
capitals that makes the Doric shaft intrinsically responsive to tremors and thereby
especially stable (figure 9.3; Ambraseys 1985). Engineering research into the dynam-
ics of seismic forces on Doric columns, in general, proves the remarkable stability
of the peristyle. Forces that overcome the shear strength of columns sitting on their
stylobates must be greater than those that will set the column rocking. That is,
severe storms with high gusts can subject columns to greater stresses than can high-
velocity earthquakes. "We find that the [earthquake] 'intensity' required to over-
turn a column is so high that before a column is overturned most other man-made
structures around it should have collapsed" (Ambraseys 1985: 214).
The Greek architect's treatment of peristyle and sekos (the cella, or central
building containing the cult statue) as two separate entities follows a logic consis-
tent with the overall aseismic design of a building. Nowhere throughout the lower
levels of construction does a block or wall bind peristyle to sekos. Paving in the
pteromata (circumferential corridor) was never bonded to the stylobate (platform
on which the columns stand) and toichobate (series of blocks at the base of a wall)
in Greek temple architecture, because exposed clamps were considered unsightly.
Less usual are detached sleepers embedded on earth fill that brace the paving stones.
Only at ceiling level with coffer spans did the physical bond between peristyle and
sekos begin. Clamps were not all dispersed at this level; sometimes an infrequent
dowel had the single purpose of being a temporary prop while intermediary cof-
fers were slipped into place (F. A. Cooper 1992-1996: vol. 3, pl. 32.c).
BUILDING MATERIALS: LIME-BASED

SUBSTANCES
Masonry
The Greeks developed a variety of limes and cements appropriate to a number of
applications: grouting, waterproofing, a matrix for pebble flooring, and different
types of wall plastering. Through the Hellenistic period there was an increasing use
of grouting, using slaked lime mixed with earth, for the stabilization of foundation
fill. Conspicuously absent is the use of the mortars to bond masonry and the
mixture of an aggregate with cement to form concrete-the latter a building tech-
nique that arrives in the Roman period (chapter 10). Glues were applied in wood
construction (Theophrastus, Hist. pl. 5.6.2, 5.7.2; see chapter 17). But Theophrastus
(Hist. pl. 5.7.2) emphasizes the fact that glue will not bind woods of different
fibers-whether mismatched grains or dissimilar wood species-or altogether
different materials. He goes on to assert that "one piece should be of similar

character, and not of opposite character, like wood and stone." Vitruvius (De arch.
7.3.9, 7.6.1) alludes to a plaster admixture, a substance added to plaster to al-
ter its properties, which may date to the classical period: "[A plaster of] sand
mortar and of marble, thinly and completely applied, [is] brilliant after being
subjected to repeated polishing," This is comparable to modern Keene's cement, a
dead-burned gypsum product that yields a hard, high-strength plaster (Merritt and
Rickets 2000: sec. 24).
Plaster and Stucco Wall Surfacing

A water-resistant hydraulic cement was used to line the interior of cisterns from the
sixth century B.C. onward (Thompson and Wycherly 1972; Camp 1984). Water
basins at fountain houses were treated in much the same manner, but with the
additional precaution of cutting the vertical joints of the orthostate slabs into a
V shape and sealing them with poured hydraulic cement or lead. Otherwise, the
ashlar or dressed masonry was treated with a layer of plaster. The concern was an
engineering problem: the permeability of stone. Stone fabric can suffer from dis-
integration due to thermal variation and the passage of water and water vapor, due
to both the porosity of the stones and the interstices between blocks. The appli-
cation of plaster to some surfaces shows an engineering awareness of this property
of stone, that is, a knowledge of rock mechanics.
The use of a plaster coating for the protection of adobe walls against weath-
ering can be seen as early as the Neolithic period. This simple type of protection
against weathering was universal and continues today. Monumental Greek ar-
chitecture in stone starts in the late eighth to early seventh centuries B.C. in the
Corinthia. It begins a long tradition of finely dressed ashlar masonry construction,
especially for sacred and public buildings meant to be beautiful and imposing. One
of the earliest examples of an ashlar temple is that of Isthmian Poseidon (ca. 700-
670 B.C.). A block preserves a patch of stucco with traces of a painted design,
making it probable that frescoes decorated the entire exterior walls. The procedure,
both as protection and decoration, most likely was inherited from earlier mud-
brick design.
The practice of coating finely-dressed and jointed masonry walls continued to
evolve. Much of the architecture at Olympia throughout its building history was
constructed of the local shelly limestone, a friable material that easily weathers.
Typically, it was protected by a thin coat of stucco, nearly all of which has weath-
ered away, subjecting the exposed stone surfaces to further erosion. The marl
limestone used for the fourth-century B.C. temple of Zeus at Nemea likewise was
given a thin coat of stucco, a variety of dead-burned gypsum, a hard, high-strength
plaster with an admixture of marble dust. The plaster was painted in Egyptian blue
and red across appropriate areas of the frieze (triglyphs) and entablature (guttae).
The remaining plaster was polished to a gleaming finish in imitation of marble, as
prescribed by Vitruvius (De arch. 7.3.9). One obvious reason for the plaster was its
brilliant appearance, but the protection afforded by the stucco was another, a sign of
a developed engineering practice for the preservation of stone.
The unusual treatment of vertical joints in the Hephaisteion in Athens might
be explained in this regard. The ashlar joint faces have the same sort of vertical,
slotted groove described above. In all probability, the purpose of the grooved joint
faces in the Hephaisteion was the same as found in water basins, namely, a form of
waterproofing to reduce the sweating of the walls during atmospheric condition of
high humidity coupled with the sometimes radical daily thermal changes. Not only
does this procedure keep the interior dry, it also retards the disintegration of the
stone. By contrast, the temples of the Parthenon and Zeus at Olympia had shallow
pools of water in slight recesses in their interior paving to provide the opposite
effect (Pausanias 5.11.11); the Phidian chryselephantine statues needed the height-
ened humidity for preservation of the ivory fabric.
Rubbing stone or a stuccoed surface with a chamois helps seal the porosity of
the material and increases, sometimes appreciably, the resistance to water and
hence to weathering. Decorated frescoes of the Bronze Age, Hellenistic, and Roman
periods show signs of polishing. The marble-compound plaster over porous stone
architecture noted above also shows traces of polishing, indicating this preserva-
tion technique was widespread (cf. Vitruvius, De arch. 7.3.9-10). Evidence that
marble architectural surfaces were polished in the same manner and for the same
purpose comes from marble sculpture. The subject of polishing Greek marble
sculpture is controversial but deserves closer study than it has been afforded to
date (Haynes 1975). It is widely held that Roman marble statuary was commonly
polished, but not Greek. Cult statues and other works displayed within a building
were not subject to any degree of weathering. I have, however, observed a few cases
of well-preserved marble pedimental statuary where the outside portions have a
matte or weathered surface, but there are patches of polished surface on the
backside of the same pieces, where the marble had been sheltered by the corona of
the geison and the tympanum wall. This appraisal of the scant evidence leads to a
conclusion that the resilience of the exterior of marble buildings was improved by
an understanding of the long-term benefit of increased water-resistance attained
through the chamois polishing of both plaster and stone surfaces.
House Flooring and Plastered Walls

The earliest Greek temples (late eighth century B.C.) had tamped, earthen floors, as
did the contemporary houses. Temples acquired stone flooring very quickly, while
house floors continued as earthen; that is, the floors of temples and houses came to
be treated in quite different ways. In the earliest temple examples, stone paving
often was laid as decorative, crazy-quilt flagging. In the sixth century B.C. and on-
ward, temple floors consisted of undamped rectangular pavers laid on top of a gril-
lage of free-floating slippers. This design reduced heaving from frost and settlement
and reduced cracking of the stonework. Not until the late fourth century B.C. and
the Hellenistic period did house floors acquire a more durable surface. A layer of
terracotta chips, usually fragments of roof tiles, were embedded in the earth base.
Increasingly, colored river pebbles, set in a matrix of cement or grout, became the
norm (Thompson and Wycherly 1972: 46, 77, 179, 197). In the Hellenistic period,
decorative pebble mosaics appear in finely finished residences and public buildings
(Dunbabin 1999: 5-17).
Prior to the late fifth and fourth centuries B.C., the Greek house was a crude
affair, for the most part adobe on top of a rubble or roughly coursed stone wall so-
cle, with packed earth floors. House architecture became increasingly sumptuous
during the late classical and Hellenistic periods. Walls increasingly were built of
field stones or even masonry, although not always coursed. There are examples at
Eretria and classical Thira and in Ionia that also preserve surfaces of painted plaster
shaped like drafted ashlar masonry, a forerunner of the Pompeian First Style of
wall painting. Fashion may have been in play, but the engineering technology of
stucco insulation also made the interior spaces more comfortable and the walls
more durable.
BUILDING MATERIALS: BAKED CLAY

ARCHITECTURAL UNITS
Architectural terracottas-roof tiles and decorated revetments-appear early, with

the rise of the first monumental Greek temples at Isthmia and Corinth in the later
seventh century B.C. (Coulton 1977: 34-35; Hellmann 2002: 298-326; Rhodes forth-
coming). Coinciding with the technical developments of stone carving, the firing of
very large roof tiles made a permanent impact on subsequent building practice and
design. Chronologically, roof tiles undergo a cycle of technological development,
starting with complicated design and high-quality clay and firing in the archaic
period, then moving to a less fastidious production in the late classical and Helle-
nistic periods. The proto-Corinthian roof tiles were handmade, but the late archaic
and classical period Corinthian and Laconian tiles were molded. In the Hellenistic
period, tiles were mass-produced by fabricants, and the quality became indifferent to
poor. By Roman times terracotta roof tiles were mass-produced, but the quality
could be quite high.
The earliest proto-Corinthian tiles were fashioned for double-sided hipped
roofs. By the second generation, the proto-Corinthian tile was made into a com-
bination pan and cover tile, a very complicated design, so much so that modern
scholars struggle to identify the fragments and how they work together (Rhodes
1984, forthcoming). In the first half of the sixth century B.C., the Laconian type of
Figure 9.4. Typical Corinthian style tiled roof, reconstruction. (F. Cooper.)
tile was introduced, along with an akroterion disk that adorned a single gabled
facade. At the heart of the development of the Greek terracotta roof tile (in the
classical period imitated in marble, see below) is a realization that the roof "costs
10% of the building program but causes 90% of the problems" (Merritt and Rickets
2000). The pan and cover tile appear to be simple in design but, in fact, are a
complex of interlocking lips and overlaps, sized to fit a two-foot building module,
laid bottom to top from ends of building to the center. Ridge cover tiles, antefixes,
and spouts often were adorned with finials, but at the same time were fashioned so
that the hundreds of pieces for any given roof worked together for a watertight
covering (figure 9.4; N. K. Cooper 1989).
Clay preparation for roof tiles, even firing in the kiln, and application of
protective slips in red and black continued to develop in pace with the sophisti-
cated pottery-making technology of the sixth to fourth century B.C. The earliest
roof tiles sometimes were given alternating red and black slips for a bright check-
erboard pattern. In the fourth century B.C. the color slips seem to disappear; in-
stead, the top sides of pan tiles and sometimes cover tiles have finger-impressed
swirls. The patterns may be decorative; more likely, the technique served to slow
the force of water runoff during heavy storms.
Roof tiles were always popular for practical secondhand use; for example, as
drain walls and grave covers. When broken into small pieces they were packed as
flooring, sometimes in a cement matrix. Flat ceramic tiles were produced only for
use in suspended hypocaust floors in baths.
ROOF SLOPE
Greek roof pitch varied little after the early archaic period; it became one of
adjustment rather than experimentation. The earliest Greek temple roofs can only
be reconstructed through interpretation of the interior arrangement of stone
plinths for columns, along with their building plans. These sources give rise to their
restoration as a gabled facade with pitched apsidal rear or the alternative rear gable
as well. Confirming this, terracotta models of houses and shrines of this period
have the same styles of roofs as temples and also show that the roofs were steeply
pitched. A single- or double-hipped roof comes with the advent of terracotta tiles
for the proto-Corinthian roof design. These were low-pitched and maybe crooked
at the eaves to make a "Chinese roof" (Hellmann 2002: fig. 405). The low-pitched,
hipped roof was given up shortly thereafter, and by the sixth century B.C. the
double-gabled roof had become the standard. This linear development follows that
of the roof tile, the engineering design of the one affecting the other. Some sur-
viving early- and mid-sixth century pediments on the Athenian Acropolis have a
relatively steep pitch, p (horizontal to vertical at peak height). By the early fifth
century B.C., the roof pitch normalized at 4:1, or about a 15 degree rise in the angle
at the horizontal vertexes. This pitch is optimal for water runoff (Merritt 1976: art.
11-27; Merritt and Rickets 2000: art. 3-4), and allows for a roof covering of unat-
tached tiles to stay in place rather than sliding down the slope or tearing away
under high winds. These roof tiles were not fastened to the underpinning rafter; a
pierced hole to accommodate a hook or nail attachment in the plate of the pan tile
or a slot at the edge is rarely found in terracotta or in stone tiles.
Flat roofs may have existed but cannot be proved either way. Flat roofs are
given to failure from the ponding of water (American Institute of Timber Con-
struction 1974: arts. 4-165 and 4-166). Shed roofs covered buildings that demanded
unusual design, such as the Stoa of the Athenians at Delphi (ca. 470 B.C.).
BUILDING MATERIALS: METAL
Iron
The Greek blacksmith rendered a variety of grades of iron, from pig iron, which is
"weak brittle but very hard" (Herubin and Marotta 1987: 209) to wrought iron
(very low in carbon and slag; see chapter 4). Wrought iron was the principle
material for architectural hardware, as "it is highly resistant to corrosion, highly
ductile, and readily machined. It is easily worked" (Herubin and Marotta 1987:
209). Iron was used invariably for clamps and dowels; the former secured blocks
across the tops of joints and the latter the bottoms of blocks to the course below.
Iron pins stationed vertical struts underneath the ridge beam and purlins to the
wooden members below (Arsenal of Philo, JG 22 1668; Pollitt 1990: 199-202; Meiggs
1982: 213). In classical Athens, Thucydides (1.93.5) boasts of "stones of large size
hewn square, closely laid together, bound to one another on the outside by iron
clamps and lead." Iron nails, however, were not commonly used for the joinery of
wood members in Greek architecture. This is indicated in part by the rarity with
which they are found in archaeological contexts, with the exception of librar-
ies, where they may have been used for the attachment of cabinets to a wall. Pro-
tective grates and grills over window and door openings were made of iron bar.
Hinges were also rare, but used for folding doors on several West Greek temples.
Doors swung on pivots, made of wood (see below). A few preserved locks and keys
also were wrought in iron. Nails and certain tools may have been smithed from
carburized iron, whereas wrought iron went into the clamps and dowels.
Theophrastus tried to come to terms with the paradox of "hardness" in dif-
ferent types of iron, stone, and wood; poorer grades of iron cut some grades of
wood but not others (Hist. pl. 5.5,1, 5.6.2-4). The same paradox applied to iron on
stone and stone on stone. "It was not until the time of Mohs (1773-1839) and other
modern mineralogists that differences in hardness became important criteria for
classifying and identifying minerals" (Caley and Richards 1956: 147). Iron clamp
and dowel design evolved over time, responding to an optimization of the ductile
and strength properties of the materials (Dinsmoor 1950: fig. 64).
Experiments in Structural Iron

There are two instances where a Greek designer/engineer experimented with a
structural use of iron bars. The examples are separated by geography and by time:
the temple of Zeus at Agrigento, Sicily (ca. 490 B.C.), and the Propylaea at Athens
(ca. 438 B.C.). In both cases, a linear cavity was cut along the axis of a horizontal
stone architrave for the insertion of iron bars (figure 9.5). In 1823, traces of iron
were found inside the slot at Agrigento (Broucke 1996: 115), a cutting 0.31 m deep
and 0.10 m wide that runs along the longitudinal center line of the soffits to pairs of
adjoining exterior architrave blocks. Both ends of the architrave pairs are sup-
ported above interaxial and engaged piers, while the vertical joint is unsupported.
The slot at the Propylaea at Athens differs in several details: it runs along the
tops of girders that support the ceiling beams. Some distance from either end of the
block the slot terminates; also the cutting deepens along the middle portion of its
length. Dinsmoor (1922: 148-58; 1950: 102) and Coulton (1977: 83) interpret the slot
and iron bar feature in both buildings as "a direct support" between the inter-
columniation spans. Heyman (1972) correctly understands the purpose and value
of the iron bar at Agrigento: a temporary brace that "only had a function during
construction." The iron bar inserted across the top of the girder in the Propylaea
served the same function. In this case the bar was cambered upward, rising above
b.
♦ ♦
J
C.
d.
Figure 9.5. a. Iron support bar, Agrigento; b. compression iron bar, Propylaia; c. simple
beam; d. fixed end beam. (F. Cooper.)
the top of the bedding surface and opposite the anticipated deflection movement.
When loaded by the weight of the ceiling beams, the iron bar pressed horizontally
outward, while the countersunk portion of the slot received the flattened iron bar.
This stress along the upper portion of the marble girder increased its tensile
strength (i.e., the load-bearing capacity of the beam; Merritt and Rickets 2000: art.
5-99; American Institute of Timber Construction 1974: art. 4-65) by improving the
stone's bending moment and thereby reducing the amount of deformation and
possible failure as the superimposed block settled into place (see figure 9.5).
Figure 9.6. Delphi, ductile iron clamp and setting. (Photograph by J. P. Oleson.)
Lead
Iron clamps and dowels were embedded in cuttings below the surface of stone
blocks and sealed into place by lead, for the most part poured in a molten state but
sometimes pounded around the iron fastening. Although inevitably tight, such
clamp and dowel assemblies were never left exposed, but were sealed in by adjoining
blocks. The lead inhibited water penetration and consequent corrosion of the iron,
but that was not the primary reason for use of the softer material. The lead added to
the flexibility of the overall clamp or dowel fastening. The ductile properties of the
iron fasteners encased in lead allowed for a fractional and organic movement of the
structure during times of stress, particularly that caused by earthquakes (figure 9.6;
see also below). In the Hellenistic period, "blind dowels," those embedded within
the bottom side of a block rather than at the end, were rammed into position
without lead. The vertical, protruding dowel head slipped into a rectangular cutting
in the block underneath. The slot was large enough to allow for maneuvering the
block into exact position; lead then was poured around the insertion. Now and
then, sheets of lead served as pliable shims for a final leveling of a block or capital.
Bronze
The properties of bronze are more appropriate to sculpture than to architecture: a
low enough melting point to allow casting, relatively low tensile strength, capacity
for decoration by coloring, chasing, and inlay (see chapter 16). The material was
more resistant to weathering than iron, but it was far more expensive. In buildings,
bronze bosses and medallions decorated faces of wooden doors and disguised the
seams between wooden panels. Bronze plates have been found in pivot sockets in
sills, as a simple device to ease the rotation of the door.
BUILDING MATERIALS: STONE
Wall Construction
Generally, walls in temples were made of solid stone ashlars that passed through the
thickness of the wall, while house walls were built of adobe, a sun-dried brick.
Fabrication of adobe goes back to Neolithic times in Greece and the Near East and
represents a continuous technology, summarized by Vitruvius (De arch. 2.3.2). Stone
or rubble wall socles about 0.35 m wide supported single-story domestic buildings;
socles double that width suggest two-story construction. All home-building de-
pended on vernacular technology. Wall engineering came into play in the seventh-
century B.C. generation of temples called hekatompeda ("one-hundred-footers" in
length, ca. 30 m), where an adobe wall had to be reinforced by engaged stone or
wooden piers. These strengthened the flank wall as a structural precaution for the
load of the new type of permanent roof incorporating painted terracotta tiles. The
temple of Hera at Olympia, built around 600 B.C., is the best known and studied
example of a design where engaged columns reinforced an adobe superstructure
above stone orthostates. With the advent of all-stone architecture, the interior,
engaged supports moved toward the center of the cella, positioned as a pair of free-
standing colonnades. They supported not only the longitudinal purlins directly
above, but also crossbeams that carried the struts in support of the ridge beam.
The cella flank wall and the outer peristyle provided two additional supports
for the untrussed roof system. In the late fifth and fourth centuries B.C., interior
design changed, and interior colonnades pushed back against the cella wall
(Norman 1980, 1984). The interior colonnade all but disappeared in the Hellenistic
period, most likely due to a greater confidence in materials engineering, namely
that of wood spans.
Rock Mechanics
Merritt's analyses of the forces acting within stone blocks are relevant to ancient
Greek architecture (1976: art. 6-23): "The unbalanced moment of the external
forces. About a vertical section through a beam in equilibrium, there is an un-
balanced moment dure to external forces. It is called the bending moment. Thus,
when the bending moment is positive, the bottom of the beam is in tension and the
top is in compression." At both Agrigento and the Propylaea, the ambitious spans
Figure 9.7. Fissures along stone bedding planes, Bassae Temple. (F. Cooper).
were engineering achievements, though the success never entered into standard
Greek architectural practice. Greek engineering expertise in rock mechanics em-
braced building stone and extended to the geochemical engineering of foundation
preparation. From my studies at Bassae, Nemea, and Messene, and from published
research by others (e.g., Germann 1988), it is clear that the ancient architect,
perhaps in consultation with a quarryman, selected outcrops of stone having the
highest tensile and compressive strength, values available through simple analysis
at a given source. Preferential judgments were based on assessments of the evenness
and minimal frequency of bedding planes and the related question of concentra-
tions of impurities such as iron oxides contained in the rock.
In numerous visits to Greek (and Roman) building sites over the years, I have
observed a consistent feature of building stone: blocks were extracted from quar-
ries with bedding planes running horizontally in cases of wall blocks (maximizing
compressive strength) and vertically in cases of spans (beams and epistylia) to
maximize tensile strength (figure 9.7). This geotechnical selectivity did not happen
by chance; most outcrops in Greece are thrusted rock formations, and thus bed-
ding planes rarely run horizontally. Marble deposits contain bedding planes that
also dip from the horizontal; but the crystal fabric of marble, for the most part, is
more homogeneous and structurally uniform than limestone. It is also obvious
that thick seams of rock were reserved for the larger blocks laid across wide spans,
while thinner bedded stone blocks went toward orthostate and wall blocks.
Today's geotechnical engineer uses a classification of rock type by strength. The
greater the homogeneity and cohesiveness of stone fabric, the greater the strength
and suitability for architectural and sculptural purposes. Building stone in ancient
Greece was limited to limestone (mainly) and marbles, and occasionally serpen-
tines and gneiss. Quality limestone has a toughness and modulus of rupture greater
than that of marble, while the toughness and resistance of granite is about two to
six times greater than that of marble. Granite, however, appears not to have been
used for building material in the classical Greek world. Had it been available, it
would have been uneconomic to shape it into blocks.
The ancient Greek architect's sense of building materials extended to the
selection of superior grades of limestone or marble in quarry faces (Theophrastus,
Lap. 5-7). Other signs indicate the Greek architect possessed a knowledge of the
mechanical properties of stone: he invariable placed blocks within the temple so as
to use the bedding planes of the rock to optimal advantage. Preferential selection of
stone does not require a background in geotechnology, which, after all, is a young
science. A first-rate sculptor, ancient or modern, predicts certain physical prop-
erties of stone by inspection and thereby makes qualitative selections based on
judgments of personal and accumulative experience. Personal experience, however,
cannot account for the awareness that one stone has a compressive and tensile
strength that a stone of similar appearance does not have. Perhaps engineering
experience led to a realization that concentrations of iron oxides not only reduced
the load-bearing capacity, but also caused detrimental weathering. Only routine
geochemical testing will show how extensive this knowledge of stone properties
was in antiquity, but a few comparable analyses of stone from other sites indicate
that this ancient awareness existed. "Analysis of frequency distribution of iron
contents in Thessalian marbles reveals a distinct shift of the distribution curves for
stelae marbles toward lower values as compared to marble samples taken from
ancient quarries. This means that the ancient quarrymen selectively mined marbles
of high quality and purity which today cannot be sampled adequately, a fact which
should be considered carefully not only in geochemical analysis" (Germann et al.
1980: 103-4). Elsewhere, Peschlow-Bindokat and Germann (1981: fig. 96) publish a
variation diagram of iron/magnesium ratios of southwest Anatolian marbles, again
showing concentrations of iron content in quarry samples as being higher than
concentrations in samples from ancient buildings.
DESIGN BENDING MOMENT
The colonnade, whether as a peristyle around a temple or as a row of columns along

a stoa, depended on the fixed-end beam system. On occasion, the Greek architect
experimented with a shaped profile of a stone member as a means to counteract
stress when a stone member was put into suspension. The ceiling coffers of the
Bassae temple were shaped with cambered backs; their profile follows the convex
curve of bending moments, rising in a direction opposite to the deflection of the

stone put into suspension. Engineers today draw up a bending moment diagram
determined from mathematical calculations. These graphs assume a curvature of a
beam subjected to loading. A predicted moment of rupture increases when the
deflection is decreased by means of a design reinforcement; on the Propylaea on
the Acropolis this involved the thickening of the back side of the coffer panel or the
placement of the tension bar in the ceiling girder. This economical solution in-
creased the strength of the coffer; it did not, however, increase the stiffness, as is
sometimes supposed. The ceiling above the porch of the temple at Sangri, N axos,
also carried cambered beams (F. A. Cooper 1992-1996: vol. 4, pl. 38.b 4; Coulton
1977: 143-47, figs. 63-64), a situation that continues the engineering experiment of
the girder tension rod in the Propylaia design noted above.
STONE ROOF TILES
Beginning with the Periclean building program at Athens in the mid-fifth century
B.C., major temples and public buildings were outfitted with an all-stone roof,
usually of marble: on the Acropolis, the Parthenon, Propylaea, Nike temple; in the
agora, the temple of Hephaestus, and the third-century B.C. Stoa of Attalus. Other
stone roofs may have existed, but, as today, roof coverings were the first to be lost
as the ancients hauled the stone and terracotta tiles away as building material and
for a variety of other purposes. Marble and limestone stone roofing almost dis-
appears, but there are surviving examples; for example, the original limestone roof
of the heroon at Messene (ca. 170 B.C.). In this later period, replacement and new
roof covering were increasingly made by commercial and often indifferent man-
ufacturers.
For both early and classical stone roofs, an ingenious interlocking system for
the tile joints is indicative of an engineering design to eliminate leakage. The early,
proto-Corinthian terracotta roofs at Isthmia and Corinth (late eighth to early
seventh centuries B.C.) have covers attached to pans. As Rhodes (forthcoming)
shows in his experimental recreations of these tiles, the laborious manufacturing
process proved successful in the reduction of the number of open joints between
rows of tiles. The combination tile went out of fashion, but then returned by the
time of the Periclean Acropolis program. The Parthenon sports a marble, double-
pan tile with a fake cover running down midway. An attached cover tile at one edge
further reduced the open jointing for a watertight roof, a necessity given the ivory
statue of Athena Parthenon underneath. The marble roof over the temple at Bassae
reflects the Parthenon design, but elaborate, tightly-fitted marble roofs were too
expensive for routine use. As mentioned above, repairs at Bassae and the heroon at
Messene and elsewhere were made in terracotta. The earlier translucent marble
gave these roofs a secondary purpose: a diffused illumination of the interior space.
Bassae had an all-marble interior ceiling, and it could be that the Parthenon had no
woodwork between the cross-beams at ceiling level, exposing the interior of the
marble roof.
BUILDING MATERIALS: Wooo
Evidence of the exploitation of timber for building purposes in ancient Greece

depends almost entirely on written testimonia rather than surviving physical and
archaeological evidence-in contrast to the other engineering techniques and prac-
tices. Hodge (1960 ), Hellmann (1986: 237-47; 2002), and Liebhart (1988) have inter-
preted the sockets for beams, rafters, and doors in extant buildings in combina-
tion with citations from epigraphical evidence. Meiggs (1982) presents the overall
picture, including a particularly acute overview of the written sources.
Wood played a prominent role in the engineering of the rooftree of temples
and other large structures. Rafters lodged in cuttings spaced along the top and back
sides of the horizontal geison (crowning molding above entablature) and were
underpinned by a pair of purlins (longitudinal roof beams), which divide the slope
into two nearly equal parts. Ordinarily, rafters were spaced two building feet apart
and carried the edges of two-foot-long roof tiles on top. A set of three longitudinal
beams, a ridge, and a pair of intermediate purlins, one on either side of the pitched
roof, supported each rafter at three positions. The purlins ran above the flank wall,
while the bottom end of the rafter was slotted into a sloping cutting along the
horizontal geison course or into a substitute rafter rail resting on the geison. The
structural design was that of a horizontal beam supported at three positions
(geison, purlin, and ridge), and having minimal force of a lateral thrust.
There are no examples in which horizontal beams tie directly into the rafters,
for example by a pivot connector or gusset plate, devices that allow movement
within the joints of a simple truss design (Salvadori and Heller 1986: 112). Fur-
thermore, in Greek building design, cross beams supporting roof struts always have
an interval greater than the two-foot interval of the rafters. Even in those cases
where the interval between the rafters coincides with that between the cross beams,
the members are quite different in thickness. In other words, the simple truss had
not been devised for Greek architecture, although the argument against it is too
complex to summarize here (White 1984: 74, n. 111; cf. chapter 17). The cross section
of rafters usually measured 0.16 m-0.32 m, and they were usually square or nearly
square in section. Purlins were larger (ca. 0.60 m sq.), but they are also trapezoidal
to coincide with the slope of the rafters. Ridge beams were quite a bit stouter and
ranged around 0.60 m ± 0.10 m in width and 0.80 ± 0.10 m in height. The cross
section sizes of rafters, purlins, and ridge beams are standard and do not vary a
great deal between large temples and small.
Wood Mechanics and Theophrastus

The topic of ancient Greek engineering in wood deserves a far more expansive
treatment than can be provided here. Indeed, Theophrastus' overriding thesis-
that wood is a valuable commodity which serves a variety of building
applications-was based on a quasi-scientific understanding of particular prop-
erties such as knots and direction of grain and has the same commercial purpose
as the tables in modern handbooks (e.g., Forest Products Laboratory, U.S. Depart-
ment of Agriculture 1990). As late as 1975, Rhude (in Merritt 1975: art. 8-o) could
declare that "practice, not engineering design, has been the criterion in home
building. As a result, the strength of wood often was not fully utilized." This anal-
ysis does not apply to the situation in Theophrastus' time, since book 5 of Theo-
phrastus' Historia plantarum contains statements comparable to those in a con-
temporary engineering handbooks such as Herubin and Marotta 1987 and Merritt
and Rickets 2000. The section headings in Merritt and Rickets 2000 are closely
comparable with those in Theophrastus: "Mechanical Properties of Wood" (4-33;
Hist. pl. 5.1.5-4.1), "Effects of Hygroscopic Properties of Wood" (4-34; Hist. pl.
5,4.2-4.8), "Commercial Grades of Wood" (4-35; Hist. pl. 5.5,1-7.8), "Destroyers and
Preservatives" (4-36; Hist. pl. 5,4.2-4.8), "Glues and Adhesives for Wood" (4-50;
Hist. pl. 5.7.2, 4). Issues presented in Merritt and Rickets supplement this list: for
example, the seasons for felling timber (Hist. pl. 5.1.1-1.4). In other words, the topics
considered by Theophrastus, and often their ordering, resemble the organization
of the modern engineering handbook: the proper season to fell a particular tree, the
need to season the timber, the deficiencies of certain woods, the mechanical prop-
erties of the different species, the paradoxes observed between strength, hardness,
and bending, the rules of wood production, and the effect of growth locality on
these properties.
Theophrastus, followed by Vitruvius (De arch. 2.9.13-10.2) and Pliny (HN 12-
17 passim), envisioned boundaries, sometimes corresponding to geographical and
political ones, demarcating land where one species of tree would grow and where
another would not. In other words, the pine tree was rare in Arcadia, but common
in Elis (Theophrastus, Hist. pl. 3.9.4), whereas the Holm oak was abundant in
Arcadia, but did not grow anywhere in Elis and Sparta (Pliny, HN 16.34). To these
ancient authors, many individual species of trees could be found only in particular
sections of the world. The Old Oligarch (Ath. pol. 2.12) states the same notion: "But
no other town possesses two of these commodities; for instance, the same town has
not both timber and flax." In building contracts and commercial agreements at
major Greek sanctuaries, inscriptions give the source of wood as Macedonia,
perhaps because it was subject to politically-based price fluctuation (Eleusis, 329/8
B.C., JG 22 1672.66; Delphi, Bourguet 1932: 3.5.41, line 7; Delos, JG 22 199A.57). In a

treaty of 389/83 B.C. between Amyntas III (King of the Macedonians) and the
Chalcidians, silver-fir was the single material for ship-building and house-building
materials permitted for export by the Chalcidian League (Tod 1946-48: no. 111).
Theophrastus ( Char. 23) also seems to refer to the restrictive policies of the Mac-
edonians regarding the exportation of timber from their country.
Wood Mechanics in the Fifth Century B.C.

This scientific approach to the mechanics of wood can be documented more than a
century earlier, during the Periclean building program on the Acropolis at Athens,
by three fragmentary inscriptions. An Athenian decree (JG I3 1454, ca. 430 B.C.)
bestows honor on an Eteocarpathian for a dedication and contribution of a tall and
venerated cypress tree for the temple of Athena. The donation's recipient Athena is
ambiguous; it has been argued both that it was meant to serve as the wooden mast
for the chryselephantine statue of Athena Parthenos by Phidias, and that it con-
stituted some part of the entrance to the opisthodomos (rear, treasury room) of the
Parthenon. An ordinary, mature cypress tree would have fit the needs for the
wooden mast of the Phidian statue, which can be estimated as about 13.4 m tall,
including a sinking of 1.8 m into a socket 0.75 m x 0.45 m in the Parthenon floor
(Stevens 1955: 244-46). An even shorter cypress tree could yield the planks for the
pieced panels of a door. Yet the cypress was used for neither the mast nor the door
panels. Instead, the probable purpose for the Carpathus cypress was the ridge beam
of the Erechtheum. This roof member can be restored as having measured about 75
feet (ca. 23 m), the height of a very tall cypress. A table giving an economical span
range for main wood-framing members lists 10-25 feet for a solid-sawn, contin-
uous span, a value that increases to 10-50 feet for a laminated beam (Merritt 1975:
table 10.2). According to these figures, the ridge beam for the Erechtheum was a
successful experiment, testing the bearing limits of the cypress timber.
The second inscription, JG i3 461, is attributed to the accounts of the con-
struction of the Parthenon. The fragmentary text specifies elm and cypress, but the
context is lost. This combination of woods, as Meiggs (1982: 200) rightly notes,
suggests they were destined for the east entrance of the Parthenon. A third in-
scription from the Parthenon account, JG I3 439.107, dated to 442/ 441 B.C., spe-
cifies pine, but the quantity, price, and purpose do not survive. A logical assump-
tion is that, five years into construction, the pine was destined for scaffolding to
build walls and/or to start the roof. Theophrastus suggests that elm and cypress
were coveted woods. This is reflected in the high value attached to doors, windows,
and framing of elm and cypress in the inscribed accounts from 414 B.C. of the
confiscated properties of those convicted of the mutilation of the Herms. In JG 13
421-39, the prices are much above those given to other types of household goods
and property (Pritchett 1956).
Timber is likewise a specified commodity in inscriptions. What these inscrip-

tions tell us, in connection with Theophrastus, is that the designer, Ictinus in the
case of the Parthenon, specified species of wood optimally appropriate to struc-
tural function and placement. In other words, there already existed by the mid-fifth
century B.C. an engineering basis for the mechanics of wood as well as stone.
FIRE PROTECTION ENGINEERING
Greek architecture follows a linear development from adobe to thatch and wood,
and finally to stone construction. Literary references to catastrophic fires in tem-
ples include Delphi, Argos, and Ephesus, along with many others. For the most
part, sacrifice burning took place at outdoor altars, not in the temple. This ar-
rangement may be in part due to tradition, but it was also a means to reduce the
hazard of conflagration within the temple. The danger of fire came from roof tim-
bers, textiles, and inflammable dedications inside the temple. The stone rafters of
the temple at N axos were perhaps intended to avert the danger of fire (Gruben
1982). Modern building code requires ventilation in attics and enclosed areas,
especially in hot climates where temperatures can reach the combustion point. In
antiquity, a common solution was ventilation of the attic by an opaion roof tile; this
piece has an oculus at the center ringed by a raised lip. Opaion roof tiles along with
interior staircases and open pedimental frames to attic fronts were some of the
engineering solutions employed by the ancient Greek architect. The latter two de-
signs are found typically in West Greek architecture, for instance in Temple F at
Agrigento. Examples are known at the Parthenon and Bassae.
During his travels in Greece in 1911, Le Corbusier realized that ancient Greek
architectural engineering was practiced in classical times as an interdisciplinary
profession. Coulton (1977: 16) expresses a more usual view: "Until the development
of theoretical mechanics in the late fourth and early third centuries, there was no
distinct concept of engineer as opposed to architect." My use of the term for
engineer derives from mechana, a term Herodotus uses to mean both "contriv-
ance" (3.83; 3.152; 8.57) and "a lifting machine" (2.125), the latter in his description
of the erection of pyramids. Had Aristotle's treatise Mechanica survived, it could
have led to a much clearer appreciation today of the role of architectural engi-
neering in the history of Greek architecture. Aristotle and Theophrastus wrote at a
time when siege warfare and elaborate theatrical productions came of age and a
mechanikos designed both war and theater machines, as in modern day mechan-
ical engineering. Yet, Le Corbusier recognized the genius of the Greek architect-
engineer. In his Towards a New Architecture (1987: 11) he not only interweaves
Greek architecture, especially monuments on the Acropolis, with automobiles and

American grain elevators, but also explicitly states his comprehension in his intro-
duction: "The engineer's aesthetic and architecture-two things that march to-
gether and follow one from the other."
Although often portrayed as an impractical, idealistic artist, the Greek architect
in fact was well aware of the demands and opportunities of the available materials,
and prepared to confront the challenges of constructing very demanding designs in
a seismic environment. The focus of classical Greek society on the communal
temple or sanctuary, which helped to define the city state identity, and the innate
conservatism of the Doric and Ionic orders limited the scope of design innovation.
At the same time, this intense concentration of talent and funding fostered the
development of exquisitely appropriate refinements in detail, and lasting solutions
to the problems of stability and endurance.
REFERENCES
Algren-Ussing, G., and A. Bramnaes 1975. Fouilles de Delphes. Vol. 2, Atlas. Paris: De
Boccard.
Amandry P. 1981. "Chronique delphique (1970-1981)," Bulletin de Correspondence Helle-
nique 105: 673-769.
Ambrayses, N. 1971. "The value of the historical records of earthquakes," Nature 232:
375-79.
Ambraseys, N. 1985. "On the protection of monuments and sites in seismic areas," in
Restoration of the Acropolis monuments. Athens: Ministry of Culture and Sciences,
Committee for the Preservation of the Acropolis Monuments, 207-28.
American Institute of Timber Construction 1974. Timber construction manual. 2nd ed.
New York: Wiley.
Bammer, A. 1984. Das Heiligtum der Artemis von Ephesos. Graz: Akademische Druck- und
Verlaganstalt.
Bammer, A. 1984-85. "Plinius und der Kroisostempel," 6sterreichisches Jahrbuch 55:
13-28.
Bequignon, Y. 1937. "Recherches archeologiques aPheres de Thessalie." Diss., Universite de
Strasbourg.
Bourguet, E. 1932. Fouilles de Delphes. Vol 3.5, Les comptes du ive siecle. Paris: E. de Boccard.
Broucke, P. 1996. "The temple of Olympian Zeus at Agrigento." Ph.D. diss., Yale Uni-
versity.
Caley, E. R., and J. F. C. Richards 1956. Theophrastus: On stones. Columbus: Ohio State
University Press.
Camp, J. McK. II, and W. B. Dinsmoor Jr. 1984. Andent Athenian building methods.
Princeton: American School of Classical Studies at Athens.
Cooper, F. A. 1983. The temple of Zeus at Nemea: The reconstruction project. Athens: Benaki
Museum.
Cooper, F. A. (ed.) 1992-1996. The temple of Apollo Bassitas. 4 vols. Princeton: American
School of Classical Studies.
Cooper, F. A. 1999. "Curvature and other architectural refinements in a Hellenistic heroon

at Merrene," in Haselberger 1999: 185-97.
Cooper, N. K. 1989. The development of roof revetment in the Peloponnese. Studies in
Mediterranean Archaeology, Pocketbook. Partille: Paul Astroms Forlag.
Coulton, J. J. 1977. Ancient Greek architects at work: Problems of structure and design. Ithaca,
NY: Cornell University Press.
Dinsmoor, W. B. 1922. "Structural iron in Greek architecture," American Journal of Ar-
chaeology 26: 148-58.
Dinsmoor, W. B. 1950. The architecture of ancient Greece. 3rd ed. New York: Batsford.
Dunbabin, K. M. D. 1999. Mosaics of the Greek and Roman world. Cambridge: Cambridge
University Press.
Forest Products Laboratory, U.S. Department of Agriculture 1990. Wood engineering
handbook. Englewood Cliffs, NJ: Prentice Hall.
Germann, K., et al. 1988. "Provenance characteristics of Cycladic (Paros and Naxos)
marbles, a multivariate geological approach," in N. Herz and M. Waelkens (eds.),
Classical marble: Geochemistry, technology, trade. Dordrecht: Kluwer, 251-62.
Germann, K., G. Holzmann, and F. J. Winkler 1980. "Determination of marble limits of
isotope analysis," Archaeometry 22: 99-106.
Gruben, G. 1982. "Der Burgtempel Avon Paros. Naxos-Paros, Vierter Vorlaufiger Bericht,"
Archiiologischer Anzeiger 1982: 197-229.
Harvey, J. C. 1982. Geology for geotechnical engineers. Cambridge: Cambridge University
Press.
Haselberger, L. 1997. "Architectural likenesses: Models and plans of architecture in classical
antiquity," Journal of Roman Archaeology 10: 77-94.
Haselberger, L. (ed.) 1999. Appearance and essence: Refinements of classical architecture-
curvature. Philadelphia: University Museum.
Haynes, D. E. L. 1975. "A question of polish," in Festschrift Ernst Homann-Wedeking.
W aldsassen: Stiftland-Verlag, 131.
Hellmann, M.-C. 1986. "Apropos du vocabulaire architectural dans les inscriptions de-
liennes: Les parties portes," Bulletin de Correspondence Hellenique no: 237-47.
Hellmann, M.-C. 2002. L'Architecture grecque. Vol. 1, Les principes de la construction. Paris:
Picard.
Herubin, C. A., and T. W. Marotta 1987. Basic construction materials: Methods and testing.
3rd ed. Englewood Cliffs, NJ: Prentice-Hall.
Heyman, J. 1972. '"Gothic' construction in ancient Greece," Journal of the Society of
Architectural Historians 31: 3-9.
Hodge, A. T. 1960. The woodwork of Greek roofs. Cambridge: Cambridge University Press.
Hogarth, D. G. 1908. Excavations at Ephesus: The archaic Artemisia. London: The British
Museum.
Kawerau, G., J A Bundgaard, and W. Dorpfeld 1974. The excavation of the Athenian
Acropolis, 1882-1890: The original drawings, 2 vols. Copenhagen: Gyldendal.
Kelly, J. M. 1982. "Aseismic base isolation," Shock and Vibration Digest 14.5 May: 17-25.
Korres, M. 1983. Melete apokatastaseos tou Parthenonos [Study for the restoration of the
Parthenon]. Athens: Upourgeio politismou kai epistemon. Epitrope suntereseos
mnemeion Akropoleos.
Korres, M. 1999. "Refinements of refinements," in Haselberger 1999: 79-104.
Kousell, K., and E. Dimou 2006. "Building materials (except for Pentelic marble) used in
ancient Athens," ASMOSIA, Berne Colloque International, Aix-en-Provenye, France,
12-18 June 2006. Abstracts: 29.
California Press.
Le Corbusier 1987. Towards a new architecture. Trans. F. Etchells. London: The Architecture
Press.
Liebhart, R. F. 1988. "Timber roofing spans in Greek and Near Eastern monumental
architecture during the early Iron Age." Diss., University of North Carolina at
Chapel Hill.
Martin, R. 1965. Manuel d'architecture grecque, I: Materiaux et techniques. Paris: Picard.
Mccredie, J. R. 1979. "The architects of the Parthenon," in G. Kopcke and M. B. Moore
(eds.), Studies in classical art and archaeology: A tribute to Peter Heinrich von
Blanckenhagen. Locust Valley, NY: Augustin, 69-73-
Meiggs, R. 1982. Trees and timber in the ancient Mediterranean world. Oxford: Oxford
University Press.
Merritt, F. S. (ed.) 1975. Building construction handbook. 3rd ed. New York: McGraw-Hill.
Merritt, F. S. (ed.) 1976. Standard handbook for civil engineers. 2nd ed. New York: McGraw-
Hill.
Merritt, F. S., and J. T. Rickets (eds.) 2000. Building design and construction handbook. 6th
ed. New York: McGraw-Hill.
Norman, N. J. 1980. "The 'Ionic' cella: A preliminary study of fourth-century B.C. temple
architecture." Diss., University of Michigan.
Norman, N. J. 1984. "The temple of Athena Alea at Tegea," American Journal ofArchaeology
88: 169-94.
Orlandos, A. 1966-1968. Les materiaux de construction et la technique architecturale des
anciens grecs. 2 vols. Paris: De Boccard.
Orlandos, A. 1976-1978. He architektonike tes Parthenonos. 3 vols. Athens: Athens Ar-
chaeological Society.
Pakkanen, J. 2004. "The Temple of Zeus at Stratos: New observations on the building
design," Arctos 38: 95-121.
Pakkanen, J. 2006. "The Erechtheion construction work inventory (IG 13 474) and the
Dorpfeld temple," American Journal of Archaeology no: 275-82.
Peschlow-Bindokat, A., and K. von Germann 1981. "Die Steinbriiche von Milet und
Herakleia am Latmos, mit einem Beitrag, Lagerstatteneigenschaften und herkunfts-
typische Merkmalsmuster von Marmoren am Siidwestrand des Menderes-Massivs
(Siidwestanatolien)," Jahrbuch des deutschen archiiologischen Instituts 96: 157-235.
Petrakos, B. 1983. A concise guide to Rhamnous. Athens: Archaeological Society of Greece.
Petronotis, A. 1972. Zum Problem der Bauzeichnungen bei den Griechen. Athens: Dodona-
Verlag.
Pollitt, J. J. 1990. The art of ancient Greece: Sources and documents. 2nd ed. Cambridge:
Pritchett, W. K. 1956. "The Attic Stelai, Part II," Hesperia 25: 178-317.
Rhodes, R. F. 1984. "The Beginnings of monumental architecture in the Corinthia." Diss.,
University of North Carolina.
Rhodes, R. F. forthcoming. "The Manufacture of early Corinthian rooftiles," forthcoming
in Symposium on Issues of Architectural Reconstruction, Notre Dame University, 20-23
January 2006.
Robertson, D. S. 1964. Handbook of Greek and Roman architecture. 2nd ed. Cambridge:
Salvadori, M. G., and R. A. Heller 1986. Structure in architecture: The building of buildings.
3rd ed. Englewood Cliffs, NJ: Prentice-Hall.
Shear, T. L. 1971. "The Athenian agora: Excavations of 1970," Hesperia 40: 243-55.
Stahlin, F., E. Meyer, and A. Heidner 1934. Pagasai und Demetrias: Beschreibung der Reste
und Stadtgeschichte. Berlin/Leipzig: de Gruyter.
Stevens, G. P. 1955. "Remarks upon the colossal chryselephantine statue of Athena in the
Parthenon," Hesperia 24: 240-76.
Tassios, T. P., and C. Palyvou (eds.) 2006. Ancient Greek technology: Proceedings of the 2nd
International Congress. Athens: Technical Chamber of Greece.
Thompson, H. A., and R. E. Wycherley 1972. The agora of Athens: The history, shape, and
uses of an ancient city center. Excavations in the Athenian Agora 14. Princeton:
American School of Classical Studies at Athens.
Tod, M. N. 1946-1948. Greek Historical Inscriptions. 2 vols. Oxford: Clarendon Press.
Tschira, A. 1972. "Untersuchungen im Siiden des Parthenon," ed. S. Sinos, Jahrbuch des
deutschen archiiologischen Instituts 87: 158-231.
Vetters, H. 1986. "Ephesos: Vorlaufiger Grabungsbericht fiir die Jahre 1984 und 1985,"
Anzeiger der 6sterreichischen Akademie des Wissenschaften in Wien, Philosophische-
Historische Klasse 123: 75-110.
Vutukuri, V. S., R. D. Lama, and S. S. Saluja 1974. Handbook on mechanical properties
of rocks. Vol 1. Bay Village, OH: Trans Tech Publications.
CHAPTER 10
ROMAN ENGINEERING
AND CONSTRUCTION
LYNNE LANCASTER
THE ROMAN-BUILT ENVIRONMENT
Strabo (5.3.8), writing in the Augustan period, notes that the Greeks were adept at
founding beautiful and strategically located cities that took advantage of natural
harbors and productive soils, whereas the Romans focused on practicality with the
construction of roads, aqueducts, sewers, tunnels, and viaducts. With the benefit of
hindsight we can also add man-made harbors to his list. Each of these is treated in
other chapters of this volume. This chapter, however, presents developments that
affected actual buildings, many of which occurred after Strabo wrote. The study
of building technology involves two main issues: the available materials and the
methods of putting them in place so that they stay put, which in modern parlance
translates to "materials technology" and "structural engineering" (White 1984: 73).
However, advances were also affected by the reorganization of work practices and
the combination of existing technologies (Greene 1992: 101), a skill at which the
Romans were most adept. The Romans drew on many technological innovations
made in the Greek world, so with this in mind, I focus here on the aspects of Roman
building technology that go beyond what had been accomplished earlier and thus
provide insight into broader trends in the Roman world.
ROMAN ENGINEERING AND CONSTRUCTION 257
LABOR SUPPLY AND CONTRACTS
The acquisition of manpower for large projects, the legal framework in which it
occurred, and the social institutions involved all played a role in the development
of building technology. During the Republic, public projects were let out to private
contractors (redemptores), usually by the censors or aediles using a contract called
locatio conductio, which stipulated certain obligations for each party (Martin 1989:
29-40). A similar use of contractors for major public works continued during the
imperial period, as attested by Frontinus (Aq. 2.119) and the funerary inscriptions
of the redemptores who worked on such projects (Anderson 1997: 108-13). The use
of multiple contractors within the same project may go some way in explaining
the differences in building techniques found within a single project, such as at the
Colosseum, Trajan's Markets, and the Baths of Caracalla (Rea et al. 2002: 370-74;
Lancaster 1998: 305-8; Delaine 1985: 196). Both architects and builders used mea-
sured drawings to communicate ideas, as indicated by Vitruvius (De arch. 1.2.2,
3.5.8) and demonstrated by examples preserved on papyrus, carved on walls and
pavements, and depicted in mosaics and stone reliefs (Wilson Jones 2000: 49-58
with bibliography), but the differences in details within single projects suggest that
there was a greater degree of flexibility in executing the instructions than one would
expect today. Unfortunately, evidence for how large projects were overseen is scant.
An inscription from Tunisia notes the use of an exactor to oversee four redemptores
on a project in Thuburbo Maius (AB 1940.16), but for large imperial projects
presumably an architect was involved. Social institutions, such as clientela, whereby
persons of different social strata were bound by certain obligations for each other's
welfare (especially in the case of freed slaves), also played a vital role in the way
labor was acquired. Delaine (2003: 723-32) has traced such connections between
freedmen and former masters in the building industry at Ostia. This type of in-
teraction was often financially beneficial for both parties and could provide in-
centives for technological innovations.
THE ARCH, VAULT, AND

TRABEATED ARCHITECTURE
The Romans were known for their creative exploitation of the arch, and research
has shown that it appeared in Rome by the sixth century B.C. (Cifani 1994: 194).
The success of the arch relies on the radiating joints between the wedge-shaped
voussoirs to direct the load toward the sides and away from the opening under-
neath. It is the optimal form for stone construction because the material is subjected
to compressive rather than tensile stresses. Stone is very strong in compression but
weak in tension, so a stone arch can span a larger distance than a flat stone beam,
which undergoes both compressive and tensile stresses. Moreover, builders in
Rome were relying on the local volcanic tuff, which has much less tensile strength
than marble. The arch became a basic element for Roman architectural types such
as the freestanding theater and amphitheater. It was translated into concrete con-
struction, and the resulting curvilinear forms transformed the architectural vo-
cabulary in the Mediterranean.
Once travertine and marble became available in Rome in the late second and
early first centuries B.C., the builders were also faced with lifting heavier stones of
much larger dimensions than earlier. Lifting technology using cranes, ropes, pul-
leys and capstans came from the Greek world (Coulton 1974; see also chapter 13),
but the Roman contribution was to use these elements on a scale far beyond what
had occurred earlier (figures 2.4, 13.1). For example, the appearance of multiple
lewis holes on single blocks, such as the 58-ton architrave of the Temple of Jupiter
at Baalbek (first century A.D.), suggests that multiple lifting devices were coor-
dinated and used together (Wiegand 1921: 66-67). From a much later period, the
base of the obelisk of Theodosius at Constantinople (A.D. 390) depicts the use of
multiple capstans to move the obelisk itself. No doubt similar techniques were used
in the early second century in Rome to lift the blocks (25-77 tons) of Trajan's
Column (Lancaster 1999). This procedure contrasts sharply with Pliny the Eider's
(HN 35.96) description of the use of ramps to lift the massive architrave of the
Temple of Artemis at Ephesus in the fourth century B.C.
A significant difference between Greek and Roman stone architecture was in
the construction of columns: the Greeks using stacked drums and the Romans
preferring monolithic shafts. The Roman preference for monoliths had a great ef-
fect on the technological requirements of supplying the stone, and it also demon-
strates the importance of reorganized work practices mentioned above. The period
from the late first century B.C. to the early second century A.D. was one in which
the imperial administration increasingly took control of quarries supplying marble
for imperial projects (chapter 5). Wilson Jones (2000: 147-56) notes that this was
also the period when a new type of standardized design procedure developed:
what he calls the 6:5 rule, whereby column shafts were typically five-sixths the total
height of the column (i.e., base+ shaft+ capital). This phenomenon appears
from the Augustan period and is related to the production in imperial quarries of
standard-sized roughed-out Corinthian capitals and of monolithic column shafts
in lengths of multiples of either 4 Roman feet or 5 Roman feet (= RF). This rule of
thumb resulted in a 50-RF column with a 40-RF shaft, a 48-RF column with a 40-
RF shaft, a 36-RF column with a 30-RF shaft, and so on. Such standardization was
only possible once the technology existed for the quarrying, transporting, and lift-
ing of large monolithic shafts.
An analysis of the Pantheon porch facade illustrates the types of issues that
arose when building with the largest of these monolithic shafts. Wilson Jones
(2000: 199-212) has suggested that the Pantheon porch, which is supported on 40-
RF shafts of Mons Claudianus and Aswan granite, was originally intended to be
taller and to be supported on 50-RF shafts. The argument is based on various

constructional anomalies in the building as well as an analysis of the relationship of
the facade to the rotunda, which shows that the taller version would have yielded a
more proportionally pleasing porch. One explanation for why the change occurred
is that the 50-RF shafts (ca. 100 tons) were twice as heavy as the 40-RF ones (ca. 50
tons) and would have been much more difficult to quarry and transport. Two other
tantalizing pieces of evidence regarding this problem have come to light. Carved
onto the pavement in front of the Mausoleum of Augustus have been found full-
scale working drawings, one of which matches the pediment of the Pantheon and
was likely used by the workers to lay out the blocks (Haselberger 1994). Another of
the drawings depicts a capital for a 50-RF column shaft; Wilson Jones therefore
suggests that these were designs for the original, taller porch that was ultimately
lowered. Moreover, a papyrus (A.D. 118-119) records an urgent request for more
barley to feed the large number of draft animals transporting a 50-RF shaft (almost
certainly from Mons Claudianus) to the Nile, thus highlighting the complexity of
the logistics of transporting a 100-ton stone (Pefi.a 1989). Given the date and the
rarity of such large columns, the column mentioned in the document was likely
destined for either the Pantheon or the Temple of Divine Trajan. The true sequence
of events leading to the ultimate porch design at the Pantheon may never be known,
but the issues raised above demonstrate the technological and logistical problems
that could be encountered in transporting and lifting such large stone elements.
With the continual development of larger structures, the establishment of a
secure base on which to build became critical, especially in areas of inherent in-
stability. One method was simply to build very deep foundations, reaching to bed-
rock when possible. At Sardis, which was completely destroyed by an earthquake in
A.D. 17, one of the characteristics of the post-earthquake reconstructions is that
they have very deep (8 m) mortared rubble foundations (Hanfmann 1983: 142).
The enormous Temple of Jupiter at Baalbek, also in an earthquake zone, has stone
foundations that extend to a depth of up to 17 m in places (Ragette 1980: 105).
The geological nature of the terrain also affected the stability of structures. At the
Basilica of Maxentius in Rome, the depth of the foundations (up to 8 m) was
regulated according to the nature of the subsoil and the degree of stress to which
they were subjected (Calabresi and Fattorini 2005: 77-79). The builders of the
Colosseum also took special precautions to avoid settling by building a 6-m deep
solid concrete foundation ring that spanned soils of different densities (Funiciello
et al. 2002: 161-67; Lancaster 2005b: 57-59). In marshy areas, a common technique
was to drive wooden piles (often iron-tipped) into the soil to create a stable plat-
form on which to build (figure 10.1A), as at the Theater of Marcellus in Rome
(Ciancio Rossetto 1995: 99) and the Circus at Arles (Sintes and Arcelin 1996: 78).
Another method to stabilize the soil and to facilitate drainage was to bury am-
phoras underneath the foundations (figure 10.1B; Pesavento Mattioli 1998). On
sites where a new building was constructed over the remnants of an older one, the
builders often included relieving arches in the new walls to channel the loads to
particular points in order to avoid different settling patterns within the building
1~c-+-- wooden
piles
Figure 10.1. Methods of stabilizing foundations. A: wood piles; B: amphoras; C: relieving

arches. (L. Lancaster.)
(figure 10.1C as in the Domitianic walls built over Neronian foundations along the
Clivus Palatinus (Tomei and Morganti 1986: 521).
CONCRETE AND ITS FACINGS
One of the most renowned of the Roman contributions to building technology was
the development of concrete, which affected not only vaulted roofing schemes but
also bridge and harbor construction. By the late third century B.C., the builders
were beginning to experiment with a type of mortared rubble construction, opus
caementicium, which consisted of fist-sized pieces of stone, or caementa, set in
mortar to create a solid mass. The main differences between opus caementicium and
modern concrete are the size of the stones used and the method of putting the
mixture in place. Modern concrete typically consists of a mix of cement-based

mortar and an aggregate of small stones that is poured into place, whereas Roman
concrete uses larger stones that are hand laid in the mortar and tamped into place.
This type of construction belongs to what might be called a "wet technology," in
that it requires great quantities of water to mix the mortar, and in fact the earliest
examples occurred only after aqueducts began supplying a continuous flow of
water into Rome, the earliest of which was the Appia (312 B.C.).
Roman concrete was different from earlier attempts at mortared rubble con-
struction in that it contained a volcanic ash called pozzolana, which both increased
the strength of the mortar and gave it hydraulic properties so that it could set
under water. A simple lime mortar made of quartz sand, slaked lime, and water
hardens and gains strength through its contact with carbon dioxide in the air, so
the mortar at the center does not develop the same degree of strength as that closer
to the outer surface. In contrast, when pozzolana is added, it plays an active role in
the chemical transformation of the mortar throughout the mass. It contains both
silica and alumina, which through the heat of the eruptive process are converted
into soluble forms. The resulting chemical reaction creates internal formations
that bind the elements together and strengthen the cohesiveness of the mortar
(Lechtman and Hobbs 1987: 96-99; Oleson et al. 2006); studies show that the re-
sistance to compression of pozzolana-lime mortar is five to eight times that oflime
mortar (Ferretti 1997: 70). Roman builders also recognized that adding crushed
terracotta to mortar would create a strong hydraulic mortar (Vitruvius, De arch.
2.5.7; Pliny, HN 36.175). This mixture (often referred to as opus signinum or, in
Italian, cocciopesto) was used mainly as a waterproofing material for cisterns and
elements exposed to the weather and as a setting bed for stone revetment, but in
nonvolcanic areas outside of central Italy, crushed terracotta came to be added to
the mortar of structural elements as a strengthening agent; the best known, al-
though late, example is Hagia Sophia in Constantinople (Mainstone 1988: 70).
Pozzolana was produced by a number of different volcanic systems in central
Italy from Rome to the Bay of Naples, and over time Roman builders learned that
most varieties produced strong hydraulic mortar. The most famous sources of
pozzolana are near Pozzuoli (ancient Puteoli), from which the modern term takes
it name. The Colli Albani volcanic system south of Rome, however, produced vast
quantities of pozzolana, which was used by builders in Rome. Scientific methods of
analysis are being used to identify the constituent elements within Roman mortar,
and future results should help clarify the patterns of use of the different types of
pozzolana throughout the empire (Jackson and Marra 2006, Oleson et al. 2004,
2006; Saturno 2001; Chiari et al. 1996; Lamprecht 1987: 41-69).
Walls built of opus caementicium were faced with stone or brick, and the
changes in the facing methods provide some insight into the relationships among
building technology, society, and the economy. The facing elements of concrete
walls formed a smooth outer surface, while the inner core consisted of caementa
laid in abundant mortar. Both the facing and the core were laid and rose together
as a single unit, usually with two masons working on opposite sides of the wall, as
shown in a painting from the tomb of Trebius Justus outside Rome (Rea 2004: figs.
82-83). Vitruvius (De arch. 2.8.1) mentions two types of facing, opus incertum and
opus reticulatum. The earlier of the two, opus incertum, appeared by the late third
century B.C. and consisted of small blocks of stone placed in a random pattern
(figure 10.2A). In Rome, the blocks were of soft volcanic tuff, whereas outside the
volcanic zone around Rome, limestone was used. By the late second century B.C.,
the facing had become more regularized, so that it eventually consisted of square,
pyramid shaped blocks of tuff set in a diagonal grid pattern, opus reticulatum
(figure 10.2B). Two other modern terms, opus quasi reticulatum and opus mixtum,
have been coined to describe different types of opus reticulatum. Opus quasi re-
ticulatum refers to a rough type of reticulatum that does not form a regular grid,
but it is a subjective term that should be used with caution. Opus mixtum refers to
the technique that became common in the first and second centuries A.D. in which
panels of reticulatum were separated by bands of brick facing (figure 10.2D).
The origins of opus reticulatum are inherently connected to Rome and its
environs by both the building materials used and the social institutions involved.
The earliest examples of the technique (Lacus Iuturnae, 116 B.C., and Temple of
Magna Mater, 110-100 B.C.) employ pieces of the local, easily carved volcanic tuff.
Vitruvius (De arch. 2.8.1, 2.8.5) makes clear that soft stones were used, and the
availability of the tuff was no doubt a factor in the development of the technique.
In spite of the decorative potential of opus reticulatum, one of the early polychrome
examples, the Capitolium in Terracina (mid-first century B.C.), was apparently
covered with plaster (Kammerer-Grothaus 1974: 229). Thus, aesthetics was unlikely
to have been a major impetus for the initial development of the technique, al-
though it was used later for decorative purposes (A. Wilson 2003). Structural
superiority also can be eliminated as a reason for its development, as Vitruvius (De
arch. 2.8.7) specifically states that it was inferior to opus incertum because of the
tendency to develop cracks along the diagonal joints. He also stresses, however,
that the mortared techniques were developed as a means of building more quickly
than was possible with cut-stone construction, which he considered more stable
and long lasting. Some scholars have argued that the development of opus re-
ticulatum was part of a move toward more efficient work practices that took
advantage of the increased slave pool after military conquests during the mid-
second century B.C. (Coarelli 1977; Torelli 1980). The separation of the fashioning
of the stone from the laying of the wall would have created a division of labor that
allowed for newly acquired slave labor to be employed in the tasks requiring less
training, and for the highly skilled labor (slave or free) to be used exclusively for
putting the preformed blocks into place more efficiently. So the initial develop-
ment of opus reticulatum was likely influenced by the availability of the easily
shaped volcanic tuff of central Italy coupled with the military conquests that
provided an increased labor force.
The use of opus reticulatum in the provinces is not common and often held
special significance. The earliest datable example occurs in the retaining wall for the
monument celebrating Octavian's victory over Mark Anthony at Actium in 31 B.C.
Figure 10.2. Wall facings used at Rome. A: opus incertum; B: opus reticulatum; C: opus
testaceum; D: opus mixtum. (L. Lancaster.)
In this case, the choice probably carried a political message of Rome's dominance
(Malacrino 2004: 119). Another Augustan example with possible imperial con-
nections appears in the winter palace of Herod the Great at Jericho, where the
workers may have been provided by Agrippa in 15 B.C. (Netzer 1990: 45). Some-
times the use of the technique can be directly related to a military presence, as in
the Flavian city walls at Samosata in Syria, or to imperial intervention, as in the
aqueduct at Antioch on the Orontes funded by Caligula after an earthquake in A.O.
37. At other times, the choice to use opus reticulatum stems from a desire to express
a connection with the power base of Rome itself, as at Emesa (Syria) in the tomb of
a prominent citizen, C. Julius Samsigeramos (A.O. 78-79) (Dodge 1990: 112; Spanu
1996: 923-39). Medri (2001) provides the most recent methodical examination of
opus reticulatum in the provinces and concludes that it was often used to convey
meaning but that the meaning changed according to the context. It is a building
technique that provides an instructive example of the role played by technology in
the transmission of messages of power, pride, and central authority.
During the first century A.O., opus testaceum (figure 10.2C), or brick-faced
concrete, superseded opus reticulatum in Rome, although the two continued to be
used together in opus mixtum until the late second century A.O. Fired brick is a
latecomer to the building technology of ancient Rome-not for lack of knowledge

but rather for lack of incentive. Fired terracotta was well known from the seventh
century B.C., as is shown by roof tiles and decoration the Romans borrowed from
the Etruscans (Cifani 2004: 220). The intensive production of flat bricks, however,
developed only after the Augustan period. The earliest examples of brick/tile facing
on concrete walls in Rome (first century B.C.) consist almost exclusively of roof
tiles, usually with the flanges knocked off. Facing made with a mix of bricks and
tiles then appears in the walls of the Castra Praetoria (A.D. 21-23) under Tiberius.
Only under Claudius in the mid-first century A.D. were the walls faced predom-
inantly with small bricks (bessales = ½ Roman foot square) cut or sawed into tri-
angles (Lugli 1957: 588-90 ). One advantage of creating bricks specifically for walling
rather than for roofs is that fuel costs could be reduced, because bricks for walling
can be fired at a lower temperature than roof tiles, which require higher temper-
atures to ensure impermeability (Lugli 1957: 545; Olcese 1993: 124). The most critical
change in brick production was the introduction of large bricks, sesquipedales (1½
Roman feet square) and especially bipedales (2 Roman feet square). Significantly,
the increased use of brick coincides with the period after the fire under Nero in
A.D. 64 that destroyed much of Rome and left in its wake a great concern for
building fireproof structures in the city (Tacitus, Ann. 15.42). Within half a century,
the brick industry had grown significantly and was providing income both to the
wealthy who owned property with clay beds and to the workers who fashioned the
bricks (Steinby 1993: 139-43).
The production of bricks in imperial Rome is a prime example of technological
advance that resulted from increased organization rather than innovation. Bricks
(as opposed to roof tiles) were produced in both southern and northern Italy at the
sites Velia and Aquileia by the third and second centuries B.C., long before they
became common in Rome (Righini 1970; Johannowski 1982: 23-26). The significant
change in the first century A.D. in Rome is that brick manufacture reached an
industrial scale never before achieved (cf. chapter 15). The appearance of bricks in
Republican Rome was delayed largely because the abundance of tuff left little
incentive to use them; only once the scale of building grew during the early im-
perial period was there sufficient demand to warrant dramatically increased pro-
duction. Once it occurred, evidence shows that the bricks were often made in the
same production units as other coarse wares such as dolia, mortaria, and terracotta
sarcophagi (Gasperoni 2003: 39, 145, 176-80); thus the brick industry grew from a
previously existing infrastructure.
The large-scale adoption of brick in Rome during the first century A.D. was
certainly a factor behind its increased use in the provinces by the second century
A.D.; nevertheless, brick was often used in the provinces in a different manner and
for different reasons than in Rome. In the Greek east, brick tended to be used in
bands of a few courses that extended the full thickness of the wall or as the main
structural material composing the wall, whereas in Rome it was used only as a fac-
ing for concrete walls (Dodge 1987: 106-16). For example, compare the brick bands
of facing in opus mixtum at Rome (figure 10.2D) and the solid brick bands in
brick arch
stone key----tt-+-+-+--l-+-f-+++-t---+t----cfl-H-f-f-ffl'm-:,'?i'"'3~~Wrr~
mortared
rubble fill
solid
setting bed
of mortar
bricks
J
hypo"""
Figure 10.3. A: Pier construction with bands of brick and rubble infill (modeled on
Gymnasium Complex at Sardis); B: Typical hypocaust construction in imperial Rome
(adapted from Yegiil 1992: fig. 442). (L. Lancaster.)
mortared rubble walls in Asia Minor (figure 10.3A). Yegill (1986: 124) suggests that
the solid brick bands would have served to solidify the mortared rubble walls by
compartmentalizing them so that uneven settlement and cracks could be better
controlled. He points to a method still used by local builders in the earthquake-
prone zone around Sardis, in which timbers are built into the fabric of the mor-
tared walls to arrest vertical cracks. In Rome the builders sometimes used courses
of bipedales that ran through the entire thickness of the wall (figure 10.3B), but
these seem to have served more as a constructional aid by providing flat surfaces at
critical points in the structure (top of foundations, spring of arches, stages of
scaffolding), by aiding in keeping the wall true as it rose, and by marking critical
levels within the building process, to name a few uses (Delaine 1997: 143-45;
Lancaster 1998: 285-99).
LARGE-SCALE VAULTED STRUCTURES
The development of large-scale vaulted structures is one of the most significant

Roman contributions to the history of building technology and engineering. One
issue debated since the nineteenth century is whether large concrete vaults acted
monolithically without exerting lateral thrusts on the walls. Some scholars have
asserted that the pozzolana mortar created a concrete strong enough to resist in-
ternal tensile stresses so that it did indeed act monolithically (Blake 1959: 163; Ward-
Perkins 1981: 101). Structural analyses have shown, however, that this claim is not
true except for very small vaults where stresses remain low enough so that cracks
do not develop (Mainstone 1975: 118; Mark et al. 1992: 120-31). For large concrete
vaults, including the dome of the Pantheon, the existing cracks demonstrate that
the material was not strong enough to remain monolithic. Moreover, the various
building methods developed by the Roman builders to counter these lateral thrusts
indicate that by the imperial period they understood that the outward thrusts had
to be controlled.
The mastery of large and complex wooden centering was an important factor
in building the most imposing concrete vaulted structures, and the use of the
triangular truss was critical in this development. A truss has the advantage over a
simple beam of spanning great distances using a number of smaller timbers that
employ the principle of diagonal bracing. That trusses were in use by the time of
Augustus is suggested by comments by Vitruvius (De arch. 4.2.1, 5-1.9; figure 17.8)
and as implied by large spans achieved in structures like the Odeum of Agrippa at
Athens (24.5 m span). The earliest graphic representation of the principle occurs on
a relief found under the Palazzo della Cancelleria in Rome that depicts a wooden
amphitheater with triangular trusses, dating from either the first (Coarelli 2001: fig.
3) or second century A.D. (Rodriguez-Almeida 1994: 215-17). Diagonal bracing was
probably used much earlier in the Hellenistic siege towers of the fourth and third
centuries B.C., as implied later on by Apollodorus's second-century A.D. treatise
that describes the construction of such towers (Lendle 1983: 77-107). In fact, similar
structures may well have been used to form the central towers for the centering of
large domes like that of the Pantheon (Rasch 1991: 365, 369-70; Lancaster 2005a:
45), although some scholars maintain that large domes were built without the aid
of a central tower (Adam 1994: 175-76; Taylor 2003: 195-208).
Studies using photogrammetry on domes have shed light on the development
of wooden centering structures. Rakob' s (1988) photogrammetric documentation
10
~ - - - - step-rings
i +---_-_=- -;::-_ ~ / - - - - bipedalis course
~----'t~~\ ~,"+~-~~""" amphora
~; ~~~-\;\\~~~~~
~~ - R i- _:_; _ :_-~~~\ ~
_ _ _, ;
--a--_
- -
-~\-\
-~~---~--+~
~--=
:'?_ :
lattice ribbing
Figure 10.4. Section through dome of "Temple of Minerva Medica" in Rome (first half of
fourth century A.O.). (L. Lancaster.)
of the dome of the "Temple of Mercury" at Baiae (21.6 m span) revealed that the
dome has significant deformations in profile from one part to another suggesting
that the technology used to build the wooden structure was not yet perfected. A
complementary study is included in Rasch's (1991: 370-79) documentation of a
series of fourth-century A.O. domes around Rome, including the Tor de'Schiavi
(13 .2 m span), the "Temple of Minerva Medica" (24.0 m span), and the Mausoleum
of Helena (20.2 m span). These domes show no sign of deformation, and the details
of construction, such as the use of courses of bipedales that coincide both with the
ends of the formwork board imprints and with the tops of the step-rings sur-
rounding the extrados of the dome (figure 10.4) show that the construction process
had become much more systematic and regularized.
Along with the ability to build large wooden centering structures came a
greater understanding of the behavior of vaults once the centering was removed.
An advantage of concrete is that the weight of the material in different parts of the
building can be controlled by regulating the type of caementa used. This proved to
be a critical factor in the success of the largest and most daring vaults. The local
volcanic geology of central Italy provided a variety of different types of stones from
leucititic lava, or selce, (2,800 kg/m 3 ) to pumice (600-700 kg/m 3 ), which were used
to regulate the weight of the concrete with lighter stones at the top and heavier
stones lower down. The move toward using lighter stones in vaults had already
begun in the second half of the first century B.C., but the technique was used
systematically only from the latter part of the first century A.O. It was later used
outside of central Italy in isolated areas with access to lightweight volcanic stones,
such as in the volcanic region of the Hauran in Syria in the baths at Bostra and
Philippolis of the late second and third centuries (Ward-Perkins 1958: 343-46) and
near a volcanic zone in southeast Turkey in baths at Hieropolis Kastabala, Ana-
zarbus, and Tarsos (Spanu 2003: 25).
The most advanced use of lightweight caementa began to appear in Rome in
the early second century A.D. under Trajan, when Vesuvian scoria from the Bay of
Naples was imported specifically for the purpose of use in vaults at the Baths of
Trajan and the Basilica Ulpia. It was then used later in the dome of the Pantheon
and various other imperial structures through the third century A.D. The scoria is
somewhat heavier (750-850 kg/m3 ) but more durable than the pumice (600-700
kg/m3 ) that was available closer to Rome. Structural analyses of the Pantheon have
shown that its use at the crown of the dome significantly reduced the tensile stresses
within the material and the outward thrust on the walls (Mark and Hutchinson
1986: 24-34; Lancaster 2005a: 158-61). The scoria appears almost exclusively in
imperially sponsored buildings in Rome, suggesting that the builders were well
enough aware of its advantages to cause the imperial administration to import it
for the most daring vaulted structures. It was one of the only nondecorative con-
struction stones imported from outside the immediate environs of Rome.
The Pantheon dome also illustrates another method of controlling the mass of
vaults: the use of step-rings. The finite element structural analysis by Mark and
Hutchinson (1986: 24-34) revealed that the addition of the step-rings along the
haunch of the vault played a critical role in reducing the tensile stresses in the dome
once it cracked. The combination of lightweight caementa together with the step-
rings suggests that the builders understood that simply making the dome lighter
was not sufficient; the critical issue was to make the crown of the dome as light as
possible while adding weight, or surcharge, to its haunch. Thrust-line analyses
show that if the entire dome had been made as light as the crown, the lateral thrust
actually would have increased (Lancaster 2005a: 158-61).
One of the basic methods of countering the lateral thrusts in vaults was to
juxtapose the vaults so that they balanced each other. This technique was used in
the earliest concrete vaulted structures in Italy from the second century B.C., such
as the Sanctuary of Fortuna Primigenia at Palestrina, the vaulted structure in Rome
traditionally associated with the Porticus Aemilia (Tucci 2006), and the Stabian
Baths at Pompeii. Later, this technique was honed and used as a way of providing
buttressing for very large vaults. A particularly impressive example was the "Ser-
apeum" (second/third century A.D.) at Ephesus (17 m span), where the cut stone
vault of the central room was buttressed by narrow vaulted passages on either side
(Walters 1995: 295-304).
From the principle of juxtaposing vaults grew the idea of the buttressing arch,
which was developed in the context of the imperial thermae. The control of the
outward thrusts of vaults became critical once bath buildings grew in size, because
the central cross-vaulted frigidarium had to be elevated above the surrounding
rooms to allow for natural light. In fact, Seneca (Ep. 86.8, 90.25) notes that the
lighting of the new baths of his day (mid-first century A.D.) was one of the major
anchor block
Figure 10.5. Methods of countering lateral thrusts in vaults. A: Buttressing arches at Baths
of Diocletian; B: Typical tie bar construction found in palaestra vaults of imperial bath
buildings in Rome. (L. Lancaster.)
differences from the old-fashioned baths, and he relates this phenomenon directly
to the recent introduction of window glass. Once the cross vaults of the frigidarium
were elevated, the lateral thrusts were contained and directed down to the lower
walls by placing buttressing arches along the sides of the elevated cross vault (figure
10.5A). The buttressing elements are isolated at the corners of the cross vaults,
thereby allowing for light to enter the lunettes through the glass panes that filled
them. These buttressing arches represent the first step toward the development of
the flying buttresses of such medieval masterpieces as Chartres Cathedral.
THE STRUCTURAL REINFORCEMENT OF

VAULTS AND ARCHES
Another means of controlling the lateral thrusts of vaults was the iron tie bar,
which was used in conjunction with vaults supported on colonnades, usually made
up of monolithic shafts of foreign colored stones. The growing taste for colored
stones during the first century A.O. occurred together with the development of ad-
vanced vault forms, so that a natural progression occurred toward the integration
of two building methods: the post and lintel system, which showed off the colored
stone columns, with the concrete vaulted system, which allowed for curvilinear
forms with flat terraces above. There were probably early attempts at the use of tie
bars under Augustus, but the fully developed form appeared in the early second
century A.D. in the Baths of Trajan and possibly the Basilica Ulpia. The techno-
logical conundrum encountered by the builders was how to stabilize a row of
columns that supported a concrete vault, which could produce lateral thrusts both
during and after construction. At the palaestrae of the Baths of Trajan, and later the
palaestrae of the Baths of Caracalla (Delaine 1985: 198-202) and of Diocletian, they
solved it by embedding a stone anchor block into the concrete wall for the at-
tachment of an iron tie bar that extended through the crown of the vault to one of
the entablature blocks above the colonnade (figure 10.5B). In this way, they were
able to stabilize the colonnades and also to provide a flat concrete terrace from
which to watch the activities in the palaestra. This use of the tie bar is different from
that in later Byzantine and Renaissance architecture because the tie bar was con-
cealed in the crown of the vault rather than exposed at the springing of the vault.
Presumably this was an aesthetic decision by the designers, who did not want the
tie bars to intrude on the space created by the curving form of the vaults (Lancaster
2005a: 113-25 with bibliography).
The shift toward the use of exposed tie bars can ultimately be related to the
decline of the marble trade and with it a transformation of aesthetic values. During
the fourth century, flat architraves were often replaced by arcuated lintels, which
were typically created with brick arches supported on columns taken from earlier
structures, as in Santa Costanza (Guidobaldi 1995: 419-41). The new style became
popular, at least in part, because it eliminated the necessity for marble architraves,
which were no longer easily available unless taken from earlier structures (Pen-
sabene 2001: 122-23). By the time arcades became the preferred style in fourth-
century Rome (they had been popular in the east for some time), concrete vaults
were not as common as they had once been, and therefore, the tie bars were not so
necessary. However, once the imperial court moved to Constantinople in A.D. 330
and the resources were concentrated in the Byzantine world, vaulted architecture
combined with arcades became the dominant style, and metal tie bars and wooden
tie beams became exposed design elements (Ousterhout 1999: 210-16). The tran-
sition to the use of exposed ties ultimately represents the different aesthetic values
of a culture that no longer had access to the same material resources that the im-
perial capital had once commanded; therefore, new styles were adopted that in-
corporated the former technology but in a different form.
The Roman interest in controlling the forces within their vaulted structures is
demonstrated by the development of vaulting ribs. The earliest concrete vaults
were built with the stone caementa set radially in imitation of the voussoirs of an
arch, but by the imperial period in Rome the caementa were set in horizontal
courses, thus indicating an increased understanding of the strength of pozzolana
mortar. In places of extraordinary loads, however, the builders continued to rein-
force the concrete vault with radially-set stones or bricks that acted as ribs. Sig-
A: solid radial B: lattice ribbing

brick ribbing
C: pitched mud D: vault with

brick vault edge-to-edge
bricks at crown
Figure 10.6. Types of brick vault construction. A: Solid radial brick ribbing; B: Lattice
ribbing; C: Pitched mud brick vaulting; D: Vaulting with edge-to-edge bricks at crown.
(L. Lancaster.)
nificantly, the earliest example occurs at the Sanctuary of Hercules Victor at Tivoli
(mid-first century B.C.), where the reinforcement consists of arches of the locally
quarried travertine. The earliest and only example of the use of such travertine
vaulting ribs in Rome occurs over a century later in the substructures at the
Colosseum, to reinforce concrete vaults that supported walls above. All other ribs
in subsequent buildings in imperial Rome employ large bricks, bipedales, in one
form or another (figure 10.6A; Lancaster 2005a: 86-112). The idea of reinforcing the
vault developed in cut-stone construction and was then translated into brick
during the first century A.D. The concept of reinforcing the vault at particular
points grew from a need to create larger and more complicated structures in which
walls did not necessarily align from one level to the next, so the use of concrete
reinforced with ribs ultimately allowed for a greater freedom of design.
The subsequent development of the vaulting rib in Rome followed the vicis-
situdes of the brick industry and demonstrates the way in which technological
development goes hand in hand with economic fluctuations. From the late first
century A.D., brick stamps often give the name of the landowner of the clay bed as
well as the brick maker, so we have an indication of the number and types of people
involved in the brick industry and how they changed over time. Indeed the brick
industry in Rome involved the gamut of social classes including slaves, freedmen,
and members of the senatorial class (both men and women), who all benefited
financially from the production of bricks. The period during the late first and early
second centuries when the greatest number of people was involved is the period
when the bipedalis ribs were most common. During the second half of the second
century, imperial building in Rome declined and with it so did the production of
bricks. The brick ribs then changed in both form and function. The bipedalis ribs
were eventually superseded by ribs built of a latticework of bipedales separated by
smaller bricks, so that fewer large bricks were used (figure 10.6B). Whereas the solid
bipedalis ribs had once been carefully placed in specific load-bearing situations, the
lattice ribs commonly extend the entire length of a barrel vault. In some cases they
appear to be intended to reinforce the vaults against the loads of walls above, as in
the Baths ofMaxentius on the Palatine (Lancaster 2005a: 104-5), but in others, such
as the Basilica of Maxentius, the lattice ribbing was used to help evenly distrib-
ute the loads within the enormous (25 m span) barrel vaults (Amici 2005: 134-36).
The change in the form and use of brick ribbing is one example of the effect that the
local economy had on building technology.
The use of radially-laid brick ribbing in Rome can be contrasted with the brick
vaulting in the provinces, where the material and technique were applied in a dif-
ferent manner and for different reasons. Outside of Rome, radially-laid brick vaults
had different roots and were not used primarily as reinforcement. Early examples
occur in a mid-first-century B.C. tomb (ca. 1.0 m span) at Sarsina, Italy (Ortalli
1987: 166, pl. 23a) and in Asia Minor at a first-century A.D. tomb (2.2 m span) at
Sardis (Hanfmann and Waldbaum 1975: 59-60), but they both covered very small
spaces that were built into the ground so that lateral thrusts were not an issue. Later
examples of brick barrel vaults in large standing structures, which began to appear
in the late first or early second century A.D. in Asia Minor (Hume-i Tepe Baths at
Miletus), differ from the typical brick ribbing in Rome in that the entire intrados of
the vault was built of radially-laid brick with an infill of mortared rubble at the
haunches. Barrel vaults built entirely of radially-laid brick occur in Rome, but they
are not common and usually relate to special structural circumstances (Lancaster
2005a: 91-98). Bricks offered the advantage of acting as substitutes for stone vous-
soirs, thus avoiding the meticulous carving of the wedge-shaped blocks, and in
doing so avoiding the need for high quality pozzolana mortar. Moreover, bricks
(1,350-1,600 kg/m 3 ) were much lighter than the typically available stones, often
limestone (2,600 kg/m3 ). Bath buildings, which were particularly well suited to
vaulted construction for reasons of fireproofing and resistance to the moist envi-
ronment, tended to be the harbingers of new vaulting techniques outside of Rome.
Examples of large-scale radial brick barrel vaulting (12-18 m spans) can be seen
from the second century A.D. at Ephesus in the Harbor Baths, East Baths, and
Theater Baths, and at Sardis in the Bath-Gymnasium Complex (Yegiil 1986: 127-28;
1992: 258).
The large bath complexes at Ephesus and Sardis also provide instructive ex-
amples of the way the provincial builders, working within a tradition of cut stone
masonry, were thinking about transferring loads in large structures. The vaults of
radially-laid brick (or mortared rubble) were often supported on stone piers
connected by brick or stone arches, whereas the partition walls between the piers
typically consisted ofless solid mortared rubble or brick (figure 10.3A). The system
implies a desire to channel the load to the stone piers, which were perceived as
more stable but took greater effort to build, while the mortared rubble walls were
left largely non-load bearing (Yegiil 1992: 266-70). Similarly, in Republican ar-
chitecture in Rome the builders had been careful to use the stones with the greatest
resistance (typically travertine) at points of greatest stress, such as under columns
and at the springing of arches. The reinforcing ribs in imperial vaults reflect a
developed form of the same principle.
HEATING SYSTEMS
The spread of the bathing habit and consequent desire for bath buildings led to
technological advances in heating methods, such as invention of the hypocaust for
heating the floors (figure 10.3B) and later the related innovations for heating the
walls and vaults. The fundamental idea for a raised heated floor was probably
already developed in the Greek world, as is shown by the third-century B.C. baths
at Gortys on Crete, but the systematic exploitation seems to have occurred in
Campania in the late second century B.C., where there was a thriving tradition of
fired architectural terracottas (Delaine 1989: 124). Terracotta was advantageous
because it was fireproof and bricks made of it were produced in modular sizes so
that the four corners would fit onto the regularly spaced pilae that supported them
(Webster 1979: 287). The next step, to heat the walls, seems to have occurred in the
first century B.C., when one finds tegulae mammatae, bricks with bosses at the
corners of one side, used to create a space for hot air along the walls at Pompeii in
the Forum Baths and Stabian Baths (Yegiil 1992: 363). The most advanced form of
wall heating came with the introduction of tubuli, rectangular hollow tubes that
lined the walls (figure 10.3B). The combination of the hypocausts with the tubuli
created a much higher percentage of warm surfaces in a single room so that larger
rooms could be adequately heated. Only after such innovations had occurred could
bath buildings reach the scale of the imperial thermae and maintain the appro-
priate temperature while also allowing glass-filled openings for light.
PROVINCIAL CONTRIBUTIONS TO
CONSTRUCTION TECHNIQUES
New building techniques were often added to the Roman repertoire by builders in
the provinces. An example is pitched brick vaulting, a technique that had devel-
oped as early as the third millennium B.C. in Egypt and Mesopotamia. Such vaults
were constructed by placing mud bricks edge-to-edge so that they formed the curve
of the vault (figure 10.6C). The bricks were leaned at an angle against the wall so
that each ring of brick could be "glued" onto the preceding ring by means of a fast-
drying gypsum mortar. Slanting the bricks helped ensure that the succeeding rings
did not slip off before the mortar dried, so that the vault could be built with
minimal, if any, wooden centering, which was a critical issue in areas where wood
was scarce. In Roman Egypt this traditional building method continued into the
Roman period, as is shown by the pitched mud brick vaulting in the first-century
A.D. grain warehouses at Karanis (Ward-Perkins 1958: 91-93).
A variation on pitched brick vaulting appeared in the second century A.D. in
Greece at Argos (Aupert and Ginouves 1989), Eleusis (Durm 1905: fig. 275), and
Isthmia where fired bricks were set edge-to-edge, albeit upright (not pitched). The
examples at Eleusis and Isthmia are small drain vaults that could have been built
without centering, but the largest of the Greek examples, the Theater Baths at Ar-
gos, has a span of almost 11 meters. Given the size of this vault, the fact that the
bricks are not pitched, and the use of lime-based rather than gypsum-based
mortar, the motive for using the technique is not likely to have been to eliminate
the need for centering. A somewhat different variation appears in barrel vaults in
Asia Minor from the mid-second century A.D. where radial brick vaulting at the
haunch is combined with edge-to-edge brick vaulting in the crown (figure 10.6D)
Examples can be seen in Slope House 2 at Ephesus and the late second to early
third-century substructures of the basilicas at Aspendos and at Izmir (Dodge 1987).
These barrel vaults with vertically set, edge-to-edge bricks are typically listed
among examples of "pitched" brick vaulting, but the bricks are not actually pit-
ched. They may well be influenced by military contact at this time with Parthia,
where the bricks were already set in similar fashion in the first-century A.D. palace
at Assur (Reuther 1938: fig. 100).
True pitched brick vaulting occurs particularly in shapes of double curvature,
such as the dome, semidome, and sail vault (domical vault set on a square base).
Small sail vaults can be seen at Ephesus in the third-century A.D. Slope House 2,
EM6 (ca. 2 m span), and by the early fourth century the technique was used at a
larger scale in the dome (ca. 13.5 m span) of the mausoleum of Diocletian at Split
(Hebrard and Zeiller 1912: 89-90, 93-94). That this type of construction was in-
tended to reduce the amount of centering needed is implied by a fourth-century
letter of Gregory of Nyssa (PG XLVI, 1097) to the Bishop of Iconium requesting
workmen skilled in constructing vaults without centering because of a dearth of
wood in the area. Earlier large domes had been built of radially-laid brick, as can be
seen in the two rotundas to either side of the mid-second-century A.D. "Serapeum"
at Pergamon (12 m span). Fragments of radially-laid brick found in the excavation
of the Temple of Aesclepius at Pergamon suggest that it too had a massive dome
(23.9 m span) of radially-laid brick (Ziegenaus 1981: 46). The technique of using
pitched brick does not appear in Rome until the early fifth century in the resto-
ration of some of the towers of the Aurelian walls under Honorius. It was likely
brought to Rome by eastern craftsmen, if not by military personnel directly (Cozza
1987: 42, fig. 4).
Other terracotta vaulting techniques were also developed in the provinces. At
the legionary baths at Chester (third century) in Britain, the barrel vaults were built
with hollow terracotta voussoirs placed end to end (figure 10.7A). These elements
have been found at numerous nonmilitary sites in southeast England as well,
usually in bath buildings (Brodribb 1987: 80-81; Johnston 1978: 82). Some (but not
all) of the hollow voussoirs have holes cut in the sides indicating that they were
intended to have hot air circulating within them (Brodribb 1987: 79-83). The most
notable examples, however, occur at the Great Baths at Bath (second half of second
century A.D.), in two large barrel vaults (15.3 m and 10.5 m spans), and there they
did not have holes in the sides and were clearly not used to form a heated ceiling
(Cunliffe 1969: 98, pl. 20.b and 1). In this case, hollow tiles would have provided a
means of reducing the weight of such large vaults in a region with no access to
lightweight volcanic stones. This technique was probably introduced into Britain
by the military. The legions had their own terracotta manufacturing units that
would have facilitated technological experimentation. Both the bathing habit and
the technology that accompanied it then spread to civilian contexts.
Another innovative means of constructing vaults, also connected to bath build-
ings, is the so-called armchair voussoir, which is a voussoir (typically of terracotta
but occasionally of stone) that has notches to support flat bricks that form hollow
channels between each arched rib (figure 10.7B). The earliest example (terracotta)
of the concept has been found in excavations at Fregellae in a bath building dating
to the early second century B.C., but there the voussoirs supported only one layer
of bricks such that there were no hollow channels between the ribs (Tsiolis 2001:
106-8, figs. 6-7; Coarelli 1998: 61). This example represents an early attempt at
creating a vault without having to carve the individual voussoirs at a time before
concrete vaulting had been fully developed. The more advanced form with hollow
channels appeared during the second and third centuries A.D. in various parts of
the western Mediterranean, including Britain (Brodribb 1987: 46), France, southern
Spain, Portugal, and northern Morocco (Torrecilla Aznar 1999; Pinker 1986 both
with bibliography).
A third vaulting technique employing terracotta is the use of vaulting tubes, or
tubi fittili (figure 10.7C). Vaults using this system consist of a series of interlocking
terracotta tubes often connected by quick-setting gypsum mortar that held them in
place to form the curved intrados. The earliest examples occur in a third-century
B.C. bath building at Morgantina, Sicily (Allen 1974: 376-79) in both a domed and
A: hollow box tile voussoirs
B: "armchair" voussoirs
C: vaulting tubes (tubi fittili)
Figure 10.7. Methods of terracotta vault construction. A: Hollow terracotta voussoirs;

B: Armchair voussoirs; C: Vaulting tubes (tubi fittili). (L. Lancaster.)
a barrel vaulted room (spans ca. 6 m). The tubes are about three times larger (60-
70 cm long) than the tubes that became common four centuries later in North
Africa (12-20 cm) and are clearly in an experimental phase. Another early example
has been found in a second-century B.C. bath at Cabrera del Mar in southern Spain
(Martin 2000: 160). R. J. A. Wilson (1992: 107-8) cites later examples of a method of
kiln construction in which waster pots were interlocked mouth to foot to create the
vaulted superstructure of the firing chamber and suggests that the original idea for
the vaulting tubes may be an adaptation of this type of kiln construction. The
vaulting tubes at Morgantina are quite similar in size and manufacture to the
drainage tubes found throughout the site. The existing infrastructure for creating
the drainage tubes probably provided the impetus for this creative adaptation. As
with the early example of the armchair voussoirs at Fregellae, the creation of these
vaulting tubes was likely a response to the functional needs of a bath structure at a
time before concrete vaulting had developed.
The greatest proliferation of the tubi fittili occurs in Tunisia in the late second
century A.D., once concrete vaulting was well established. As with pitched brick
vaulting, a motive for their use may have been to avoid the use of wooden centering
(Storz 1994: 67). Tunisia was the center of various terracotta industries, including
transport amphoras for agricultural products, especially olive oil. Finds of Africana
1 Tunisian oil amphoras suggest that oil production in this area had begun to
increase in the late second century A.D. (Mattingly 1988: 51, 6), precisely when the
vaulting tubes began to appear. The proliferation of the technique in North Africa
at this time must have been in part inspired by the existing terracotta infrastructure
(Rakob 1982: 114). The initial spread of the technique outside of North Africa
during the early third century tends to be focused on military sites. Examples in-
clude Bath F3 at Dura Europus and baths at British legionary forts at Caerleon
(Zienkiewicz 1986: 336), Chester (Mason 1990: 215-22), and York (Whitwell 1976:
45). The date of Bath F3 at Dura Europos, which contains the vaulting tubes, was
originally considered to be the first century A.D., but more recent reevaluations
put it in the latter part of the second century A.D. (Deichmann 1979: 483-85; Storz
1994: 83 n. 56; Downey 2000: 165). The use of vaulting tubes offered a variety of
advantages, but it is also a technique that was intimately connected to the wider
terracotta industry that in turn was related to both agricultural production and
military production.
A method of vault construction often confused with the vaulting tubes is the
technique of embedding amphoras into concrete vaults (figure 10.4), but the two
techniques are quite different in origins and application. The placing of amphoras
within concrete vaults has typically been seen as an independent phenomenon, but
it is best viewed as part of a long history of the reuse of amphoras for land recla-
mation projects going back to at least to the fourth century B.C. (see above, figure
10.1B). Embedding the amphoras into concrete seems to have been modeled on this
practice, and the earliest examples at Aosta and Pompeii occur in foundations and
walls rather than vaults. The earliest examples in vaults occur during the Hadrianic
period outside of Rome proper and usually consist of the globular Dressel 20 oil
amphoras from Spain (at the Villa alla Vignaccia in the suburbs of Rome, at a
horrea building at Ostia, and in the Casa de la Exedra at Italica, Spain). Another
second-century example has been found in a barrel vault at the Harbor Baths at
Elaeussa Sebaste on the southern coast of Turkey, employing Anemurium A type
amphoras (Borgia and Spanu 2003: 308, fig. 260). After these early sporadic oc-
currences, the technique reappears within the city of Rome in the late third to early
fourth century A.D., with a concentration during the reign of Maxentius (A.D.
306-12), when as many as 6,000 or more were used in the vaults of the circus at his
villa on the Via Appia. The common explanation for the use of the amphoras in
vaults is that they were intended to lessen the weight and therefore the lateral
thrusts of the vaults, but a calculation of the weight reduction shows that it was
negligible and often occurred in the wrong part of the structure to have been
effective (Lancaster 2005a: 68-85, 160-64 with bibliography). The examples in
fourth-century Rome coincide with a period of decline in Italy after the economic
crisis of the mid-third century when building materials were not as easily attainable
as earlier. The initial use of the amphoras was probably an attempt to save on
materials and labor following the practice of using them for land reclamation.
Both the amphoras and the vaulting tubes became common in early Christian
architecture in Italy and in the Byzantine east (Ousterhout 1999: 229), but both
were used in a different manner than they had been earlier. The vaulting tubes,
which in the second and third centuries were used as permanent centering for
concrete vaults, were later used in early Christian architecture as a type of light-
weight vaulting in its own right, often without significant fill on top (Storz 1994:
68). The lightweight vaults exerted less horizontal thrust than a solid concrete
vault, so thinner, more economical walls could be used while also reducing the
amount of wooden formwork needed. The use of amphoras in Byzantine vaults
also reveals changing motives. In some cases, rather than embedding the amphoras
into the concrete, the builders simply stacked them up above the haunches of
the vault as a means of providing a lightweight filler to support a wooden roof
structure or to create a level surface above (Ousterhout 1999: fig. 194). By the early
medieval period the vessels were not limited to transport amphoras as earlier but
often included kiln clinkers and household pots (Mazzucato 1970: 339-41), a sit-
uation that reflects the changing economic conditions of the times.
Technological prowess, today as in antiquity, is equated with cultural superiority,

and the Romans certainly promoted it as such. The ability to amass the resources to
fund and to provide the materials, expertise, and manpower for large construction
projects was a major display of power that became evident throughout the Empire.
Technological innovation occurs in response to a number of conditions including
accumulated knowledge, evident need, economic possibility, and social/cultural ac-
ceptability (White 1984: 21), and a brief summary of the conditions that influenced
Roman innovation provides some additional insight into the motivating forces
within Roman culture. For the builders in Rome during the Republic, the accu-
mulated knowledge was partially inherited from the Greek world and partially a
result of their own experience with the wide variety of materials available in the
volcanic region of central Italy. Later, during the imperial period, technological
knowledge was transferred back and forth throughout the Roman world with ex-
tended trading patterns and military movements as the primary agents. The unifi-
cation of Roman territories during the imperial period was a critical factor in
providing financing as well as establishing various production infrastructures that
paved the way for technological change. The terracotta industry in particular pro-
vided the economic possibility for new materials and methods such as hypocausts,
hollow terracotta voussoirs, vaulting tubes, and even the advanced uses of fired
bricks for vaulting ribs. The evident need for advances in previously existing build-
ing technologies both in Rome and in the provinces most often accompanied the
desire for public amenities, namely places of entertainment such as theaters and
amphitheaters and, most importantly, bath buildings. Theaters and baths, which
provided a place to display the latest materials and architectural forms, were at the
forefront of building technology in every part of the Roman world. Moreover, the
natural resources and traditions of particular regions within the empire often re-
sulted in creative modifications of techniques developed in Rome or in entirely new
solutions better suited to the local materials. Social/cultural acceptability also played
a role in the development of new techniques. The sporadic use of opus reticulatum
in the provinces provides an example of the way in which technology could be used
as used as a means of expressing Roman identity. The marble trade and the avail-
ability of colored stones from conquered territories was an expression of power that
resulted in changing aesthetic preferences, which in turn led to technological de-
velopments such as the use of hidden metal tie bars in vaulted colonnades. The
various industries that supplied large building projects throughout the empire relied
on the social institution of slavery and with it manumission of slaves into freedmen;
these saw the building industry as a means of social and economic advancement,
which provided impetus for innovation in building techniques and in the manu-
facturing, production, and supply processes on which they were dependent. With
the rise of Christianity in the fourth century, the situation changed with the redi-
rection of funds and resources toward the building of ecclesiastical structures, the
most impressive of which were no longer centered in Rome itself. So, by the sixth
century, the advanced building techniques of the day, such as exposed tie bars/
beams, pitched brick vaults, and thin and lightweight vaults of tubi fittili, were more
often found further east in Byzantine centers such as Constantinople.
REFERENCES
Adam, J.-P. 1994. Roman building: Materials and techniques. Translated by Anthony
Mathews. London: Batsford.
Allen, H. L. 1974. "Excavations at Morgantina (Serra Orlando)," American Journal of
Archaeology 78: 361-83.
Amici, C. M. 2005. "Construction techniques and processes," in Giavarini and Amici 2005,
125-60.
Anderson, J. C. 1997. Roman architecture and society. Baltimore: Johns Hopkins Univer-
sity Press.
Aupert, P., and R. Ginouves 1989. "Une toiture revolutionnaire a Argos," in S. Walker and
A. Cameron (eds.), The Greek renaissance in the Roman Empire. Papers from the 10th
British Museum Classical Colloquium. London: University of London, Institute of
Classical Studies, 151-55.
Blake, M. E. 1959. Roman construction in Italy from Tiberius through the Flavians.
Washington, D.C.: Carnegie Institution of Washington.
Borgia, E., and M. Spanu 2003. "Le terme del porto," in E. Equini Schneider (ed.), Ela-
iussa Sebaste II. Un ponte tra Oriente e Occidente. Rome: L'Erma di Bretschneider,
247-335.
Brodribb, G. 1987. Roman brick and tile. Gloucester: Alan Sutton Publishing.
Calabresi, G., and M. Fattorini 2005. "Subsoil and foundations," in Giavarini and Amici
2005, 75-91.
Chiari, G., L. Cimitan, G. della Ventura, M. G. Filetici, M. L. Santarelli, and G. Torraca
1996. "Le malte pozzolaniche del mausoleo di Sant'Elena e le pozzolane di Tor-
pignattara," Materiali e Strutture 4.1: 1-36.
Ciancio Rossetto, P. 1995. "Indagini e restauri nel Campo Marzio meridionale: Teatro di
Marcello, Portico d'Ottavia, Circo Flaminio, Porto Tiberino," Archeologia Laziale 12.1:
93-101.
Cifani, G. 1994. "Aspetti dell'edilizia romana arcaica," Studi Etruschi 60: 185-226.
Cifani, G. 2004. "Recenti approcci e metodi per lo studio dell'edilizia antica: 11 caso della
Roma arcaica," in E. De Sena and H. Dessales (eds.), Metodi e approci archeologici:
L'industria e il commercio nell'Italia antica. Oxford: Archaeopress, 219-25.
Coarelli, F. 1977. "Public building in Rome between the Second Punic War and Sulla,"
Papers of the British School in Rome 45: 1-23.
Coarelli, F. 1998. Fregellae. Vol. 1, Le fonti, la storia, il territorio. Rome: Quasar.
Coarelli, F. 2001. "Gli anfiteatri a Roma prima del Colosseo," in A. La Regina (ed.), Sangue
e arena. Milan: Electa, 43-77.
Coulton, J. J. 1974. "Lifting in early Greek architecture," Journal ofHellenic Studies 94: 1-19.
Cozza, L. 1987. "Osservazioni sulle mura aureliane a Roma," Analecta Romana 16: 25-52.
Cunliffe, B. W. 1969. Roman Bath. Oxford: Society of Antiquaries.
Deichmann, F. W. 1979. "Westliche Bautechnik im ri:imischen und rhomaischen Osten,"
Romische Mitteilungen 86: 473-527.
Delaine, J. 1985. "An engineering approach to Roman Building techniques: The Baths of
Caracalla in Rome," in C. Malone and S. Stoddart (eds.), Papers in Italian archaeology
IV, Part iv: Classical and medieval archaeology. Oxford: British Archaeological Reports,
Intl. Series S246. Oxford: BAR, 195-206.
Delaine, J. 1989. "Some observations on the transition from Greek to Roman baths in
Hellenistic Italy," Mediterranean Archaeology 2: m-25.
large-scale building projects in Imperial Rome. Journal of Roman Archaeology Suppl. 25.
Delaine, J. 2003. "The builders of Roman Ostia: Organization, status and society," in S.
Huerta (ed.), Proceedings of the First International Congress on Construction History,
Vol. 2. Madrid: Instituto Juan de Herrera, Escuela Tecnica Superior de Arquitectura,
723-32.
Dodge, H. 1987. "Brick construction in Roman Greece and Asia Minor," in S. Macready
and F. H. Thompson (eds.), Roman architecture in the Greek world. London: Society
of Antiquaries of London/Thames and Hudson, 106-16.
Dodge, H. 1990. "The architectural impact of Rome in the East," in M. Henig (ed.),
Architecture and architectural sculpture in the Roman Empire. Oxford: Oxbow, 108-20.
Downey, S. 2000. "The transformation of Seleucid Dura-Europos," in E. Fentress (ed.),
Romanization and the City: Proceedings of a conference held at the AAR to celebrate the
50th anniversary of the excavations at Gosa 14-16 May 1998. Journal of Roman Ar-
chaeology Suppl. 38. Portsmouth, RI: JRA, 155-72.
Durm, J. 1905. Handbuch der Architektur: Die Baukunst der Etrusker und der Romer.
Stuttgart: A. Kroner.
Ferretti, A. S. 1997. "Proposte per lo studio teoritico-sperimentale della statica dei mon-
umenti in opus caementicium," Materiali e strutture 7.2-3: 63-84.
Pinker, M. 1986. "Les briques claveaux: Un materiau de construction specifique des
thermes romains," Aquitania 4: 143-50.
Funiciello, R., L. Lombardi, and F. Marra 2002. "La geologia della Valle dell'Anfiteatro," in
R. Rea (ed.), Rota Colisei: Lavalle del Colosseo attraverso i secoli. Milan: Electa, 161-67.
Gasperoni, T. 2003. Le fornaci dei Domitii: Ricerce topografiche a Mugnano in Teverina.
Viterbo: Diadalos.
Giavarini, C., and C. M. Amici (eds.) 2005. The Basilica of Maxentius: The monument, its
materials, construction and stability. Rome: L'Erma di Bretschneider.
Greene, K. 1992. "How was technology transferred in the eastern provinces?" in M. Wood
and F. Queiroga (eds.), Current research on the Romanization of the western provinces.
Oxford: Tempus Reparatum, 101-5.
Guidobaldi, F. 1995. "Sull'originalita dell'architettura di eta costantiniana," Corsi di Cul-
tura sull'arte ravennate e bizantina 42: 419-41.
Hanfrnann, G. M. A. 1983. Sardis from prehistoric to Roman times. Cambridge, MA: Harvard
University Press.
Hanfrnann, G. M. A., and J. C. Waldbaum 1975. A survey of Sardis and the major monu-
ments outside the city walls. Cambridge, MA: Harvard University Press.
Haselberger, L. 1994. "Ein Giebelriss der Vorhalle des Pantheon: Die Werkrisse vor dem
Augustusmausoleum," Romische Mitteilungen 101: 279-308.
Hebrard, E., and J. Zeiller 1912. Spalato: Le palais de Diocletian. Paris: Ch. Massin.
Jackson, M., and F. Marra 2006. "Roman stone masonry: Volcanic foundations of the
ancient city," American Journal of Archaeology no: 403-36.
Johannowski, W. 1982. "Considerazioni sullo sviluppo urbano alla cultura materiale di
Velia," La Parola del Passato 37: 225-46.
Johnston, D. E.1978. "Villas of Hampshire and the Isle of Wight," in M. Todd (ed.), Studies
in the Romano British villa. Leicester: Leicester University Press, 71-92.
Kammerer-Grothaus, H. 1974. "Der Deus Rediculus im Triopion des Herodes Atticus:
Untersuchungen am Bau und zu polychromer Ziegelarchitektur des 2. Jahrhunderts n.
Chr. in Latium," Romische Mitteilungen 81: 130-252.
Lamprecht, H.-O. 1987. Opus caementicium: Bautechnik der Romer. Diisseldorf: Beton-Verlag.
Lancaster, L. C. 1998. "Building Trajan's Markets," American Journal of Archaeology 102:
283-308.
Lancaster, L. C. 1999. "Building Trajan's Column," American Journal of Archaeology 103:
419-39.
Lancaster, L. C. 2005a. Concrete vaulted construction in Imperial Rome: Innovations in
context. New York: Cambridge University Press.
Lancaster, L. C. 2005b. "The process of building the Colosseum: The site, materials, and
construction techniques," Journal of Roman Archaeology 18: 57-82.
Lechtman, H. N., and L. W. Hobbs 1987. "Roman concrete and the Roman architectural
revolution," Ceramics and Civilization 3: 81-128.
Lendle, 0. 1983. Texte und Untersuchungen zum technischen Bereich der Antiken Poliorketik.
Wiesbasden: F. Steiner.
Lugli, G. 1957. La tecnica edilizia romana con particolare riguardo a Roma e Lazio. Rome: G.
Bardi.
Mainstone, R. J. 1975. Developments in structural form. Cambridge, MA: M.I.T. Press.
Mainstone, R. J. 1988. Hagia Sophia-Architecture, structure and liturgy ofJustinian's Great
church. London: Thames and Hudson.
Malacrino, C. G. 2004. "L'approvvigionamento idrico di Nicopoli e l'acquedotto presso
Haghios Georghios: Una nuova attestazione di opus reticulatum in Grecia," Rivista di
Archeologia 28: 107-24.
Mark, R., A. S. <;:akmak, and M. Erdik 1992. "Preliminary report on an integrated study of
the structure ofHagia Sophia: Past, present, and future," in R. Mark and A. S. <;:akmak
(eds.), Hagia Sophia: From the age of Justinian to the present. New York: Cambridge
Mark, R., and P. Hutchinson 1986. "On the structure of the Pantheon," Art Bulletin 68:
24-34.
Martin, A. 2000. "Las termas republicanas de Cabrera del Mar (Maresme, Barcelona),"
in C. Fernandez Ochoa and V. Garcia Entero (eds.), Termas romanas en el ocddente del
Imperio II (Gij6n): 157-62.
Martin, S. D. 1989. The Roman jurists and the organization of private building in the late
Republic and early Empire. Brussels: Latomus.
Mason, D. J.P. 1990. "The use of earthenware tubes in Roman vault construction: An
example from Chester," Britannia 21: 215-22.
Mattingly, D. J. 1988. "Oil for export? A Comparison of Libyan, Spanish, and Tunisian
olive oil production in the Roman Empire," Journal of Roman Archaeology 1: 33-56.
Mazzucato, 0. 1970. "Ceramica medievali nell'edilizia laziale," in Atti del Convegno
Internazionale della Ceramica (Albisole): 339-70.
Medri, M. 2001. "La diffusione dell'opera reticolata: Considerazioni a partire dal caso di
Olimpia," in J.-Y. Marc and J.-C. Moretti (eds.), Constructions publiques et pro-
grammes edilitaires en Grece entre le Ile siecle av. J.-C. et le Ier siecle ap. J.-C.: Actes du
colloque organise par l'Ecole Franraise d'Athenes et le CNRS, Athenes 14-17 mai 1995.
Paris: CNRS, 15-40.
Netzer, E. 1990. "Architecture in Palestina prior to and during the days of Herod the
Great," in Akten des XIII Internationalen Kongresses fur Klassische Archiiologie. Mainz:
von Zabern, 37-50.
Olcese, G. 1993. "Archaeologia e archeometria dei laterizi bollati urbani: primi resultati e
prospettive di ricerca," in W. V. Harris (ed.), The inscribed economy. Journal of Roman
Archaeology Suppl. 6. Ann Arbor, MI: JRA, 121-27.
Oleson, J.P., C. Brandon, S. M. Cramer, R. Cucitore, E. Gotti, and R. L. Hohlfelder 2004.
"The ROMACONS Project: A contribution to the historical and engineering analysis
of hydraulic concrete in Roman maritime structures," International Journal of Nau-
tical Archaeology 33.2: 199-229.
Oleson, J.P., C. Brandon, L. Bottalico, R. Cucitorre, E. Gotti, and R. L. Hohlfelder 2006.
"Reproducing a Roman maritime structure with Vitruvian pozzolanic concrete,"
Journal of Roman Archaeology 19 (2006) 29-52.
Ortalli, J. 1987. "La via dei sepolcri di Sarsina: Aspetti funzionali, formali e sociali," in H.
von Hesberg and P. Zanker (eds.), Romische Griiberstrassen: Selbstdarstellung-Status-
Standard. Munich: Verlag der Bayerischen Akademie der Wissenschaften, 55-82.
Ousterhout, R. 1999. Master builders of Byzantium. Princeton, NJ: Princeton University
Press.
Pefia, J. T. 1989. "P. Giss. 69: Evidence for the supplying of stone transport operations
in Roman Egypt and the production of 50 foot monolithic column shafts," Journal
of Roman Archaeology 2: 126-32.
Pensabene, P. 2001. "Criteri di reimpiego e nuovo mode architettoniche nella basilica
paleocristiana di Roma," in M. Cecchelli (ed.), Materiali e techniche dell'edilizia
paleocristiana a Roma. Rome: De Luca, 103-26.
Pesavento Mattioli, S. (ed.), 1998. Bonifiche e drenaggi con anfore in epoca romana: Aspetti
tecnici e topografici. Modena: Franco Cosimo Panaini Editore.
Ragette, F. 1980. Baalbek. Park Ridge, NJ: Noyes Press.
Rakob, F. 1982. "Ri:imische Architektur in Nordafrika, Bautechnik und Bautradition,"
150-Jahr-Feier, Deutsches Archiiologisches Institut Rom. Romische Mitteilungen Supp. 25.
Mainz: von Zabern, 107-115.
Rakob, F. 1988. "Ri:imische Kuppelbauten in Baiae," Romische Mitteilungen 95: 257-301.
Rasch, J. J. 1991. "Zur Konstruktion spatantiker Kuppeln vom 3. bis 6. Jahrhundert,"
Jahrbuch des deutsches Archiiologischen Instituts 106: 311-83.
Rea, R. 2004. L'ipogeo di Trebio Giusto sulla Via Latina. Vatican City: Pontificia com-
missione di archeologia sacra.
Rea, R., H.-J. Beste, and L. C. Lancaster 2002. "11 cantiere del Colosseo," Romische
Mitteilungen 109: 341-75.
Reuther, O.1938. "Parthian architecture. History," in A. U. Pope (ed.),A Survey ofParthian
art from prehistoric times to the present. Oxford: Oxford University Press, 411-44.
Righini, V. 1970. Lineamenti di storia economica della Gallia Cisalpina: La produttivitd fittile
in eta repubblicana. Brussels: Latomus.
Rodriguez-Almeida, E. 1994. "Marziale in marmo," Melanges de l'Ecole franraise de Rome et
d'Athenes 106.1: 197-217.
Saturno, P. 2001. "Analisi minero-petrografiche di alcune malte antiche (IV-VII secolo
d.C.) da edifici romani," in M. Cecchelli (ed.), Materiali e techniche dell'edilizia pa-
leocristiana a Roma. Rome: De Luca, 159-79.
Sintes, C., and P. Arcelin 1996. Musee de l'Arles antique. Arles: Actes sud.
Spanu, M. 1996. "L'opus reticulatum e mixtum nelle province asiatiche," in M. Khanoussi,
P. Ruggeri, and C. Vismara (eds.), L'Africa romana 11.2. Atti dell'XI Convegno di studio,
Carthage 15-18 december 1994. Ozieri: 11 Torchietto, 923-39.
Spanu, M. 2003. "Roman influence in Cilicia through architecture," Olba 8: 1-38.
Steinby, E. M. 1993. "L'organizzazione produttiva dei lateri: Un modello interpretivo per
l'instrumentum in genere?" in W. V. Harris (ed.), The inscribed economy. Journal of
Roman Archaeology Suppl. 6. Ann Arbor: JRA, 139-43.
Storz, S. 1994. Tonrohren im antiken Gewolbebau. Mainz: Philipp von Zabern.
Taylor, R. 2003. Roman builders: A study in architectural process. New York: Cambridge
University Press.
Tomei, M. A., and G. Morganti 1986. "Ambienti lungo il Clivo Palatino," Bollettino
Comunale 91.2: 514-22.
Torelli, M. 1980. "Innovazioni nelle techniche edilizie romane trail I sec. aC e il I sec. dC,"
in Technologia, Economia, e Societa' nel Mondo Romano. Como: Banca Popolare,
139-61.
Torrecilla Aznar, A. 1999. "Materiales de construcci6n en las termas de la Hispania romana:

A prop6sito de los materiales hallados en la villa de El Saucedo (Talavera la Nueva,
Toledo)," in Actas del XXIV Congreso Nacional de Arqueolog{a 1997, vol. 4. Murcia:
Instituto de Patrimonio Hist6rico, 397-416.
Tsiolis, V. 2001. "Las termas de Fregellae. Arquitectura, tecnologia y cultura balnear en
el Lacio durante los siglos III y II a.C," Cuadernos de prehistoria y arqueolog{a Uni-
versidad autonoma de Madrid 27: 85-114.
Tucci, P. L. 2006. "Navalia," Archeologia Classica 57: 175-201.
Walters, J.C. 1995. "Egyptian religions in Ephesos," in H. Koester (ed.), Ephesos metropolis
of Asia: An interdisciplinary approach to its archaeology, religion and culture. Vally
Forge, PA: 281-310.
Ward-Perkins, J.B. 1958. "Notes on the structure and building methods of early Byzan-
tine architecture," in D. Talbott Rice (ed.), The Great Palace of the Byzantine emperors,
2nd Report. Edinburgh: Edinburgh University Press, 52-104.
Ward-Perkins, J.B. 1981. Roman imperial architecture. 2nd ed. Harmondsworth: Penguin.
Webster, G. 1979. "Tiles as a structural component in buildings," in A. McWhirr (ed.),
Roman brick and tile. British Archaeological Reports, Intl. Series S68. Oxford: BAR,
285-93.
White, K. D. 1984. Greek and Roman technology. London: Thames and Hudson.
Whitwell, J.B. 1976. The Church Street sewer and an adjacent building. The archaeology of
York: The legionary fortress. London: Council for British Archaeology.
Wiegand, T. 1921. Baalbek: Ergebnisse der Ausgrabungen und Untersuchungen in den Jahren
1898 bis 1905. Berlin/Leipzig: Walter de Gruyter.
Wilson, A. 2003. "Opus Reticulatum panels in the Severan Basilica at Leptis Magna,"
Quaderni di archeologia della Libia 18: 369-79.
Wilson, R. J. A. 1992. "Terracotta vaulting tubes (tubi fittili): On their origin and distri-
bution," Journal of Roman Archaeology 5: 97-129.
Wilson Jones, M. 2000. Principles ofRoman architecture. New Haven: Yale University Press.
Yegiil, F. K. 1986. The bath-gymnasium complex at Sardis. Cambridge, MA: Harvard
University Press.
Yegiil, F. K. 1992. Baths and bathing in classical antiquity. Cambridge, MA: MIT Press.
Ziegenaus, 0. 1981. Altertumer von Pergamon 11.3. Das Asklepieon. Die Kultbauten aus
romischer Zeit and der Ostseite des Heiligen Bezirks. Berlin: Walter De Gruyter.
Zienkiewicz, J. D. 1986. The legionary fortress baths at Caerleon. Vol. 1. The buildings.
Cardiff: C.A.D.W., Welsh Historic Monuments.
CHAPTER 11
HYDRAULIC
ENGINEERING AND
WATER SUPPLY
ANDREW I. WILSON
SINCE the beginnings of agriculture and urbanization, the control and management
of water has been vital to all societies. Natural rivers and springs were exploited
wherever possible, but where they were lacking, alternative means of water supply
were needed. In the highly urbanized Greco-Roman world, artificial techniques of
water supply not only supported large city populations but also shaped the ame-
nities that defined urban living. Many of the achievements of Greek and Roman
civilization would not have been possible without the infrastructure based on skills
in hydraulic engineering.
WELLS
Wells were the earliest and simplest form of artificial water supply, consisting in
their most basic form of a simple hole dug down to the water table. Neolithic wells
are known from Cyprus (ninth and eighth millennia B.C.), Crete (Peltenburg et al.
2001; Manteli 1992), and Israel (Garfinkel et al. 2006), and until the invention of
rainwater-collection cisterns, wells remained the sole artificial technique of water
supply; later, they continued to be important alongside other technologies. Wells
were nearly ubiquitous throughout the ancient world, wherever groundwater could
be reached at moderate depth by sinking a shaft (Jarde 1907; Hodge 1992: 51-58;
2000).
Most domestic wells, and many public wells, were either circular or square, just
large enough for the digger-about 0.80-1.0 m across. Where the geology was suf-
ficiently stable, the shaft would simply be cut into the rock, but in unstable sands,
earth or gravels the sides would need lining (called "steining"); this was frequently
the case in the upper part of the shaft above bedrock. In the Athenian agora from
the fourth century B.C. onward, some wells were lined with terracotta rings made of
three curved sections, each with a handle that served as a foothold when put in place
(Lang 1968: 6-7). More often, however, the well was stabilized with a lining of mud
bricks or steined with masonry. In northern Europe during the Roman period,
wooden barrels with their ends knocked out were sometimes used to line wells dug
in clay or soft earth, as at Roman Silchester or numerous sites along the Rhine limes
(Marliere 2002: 40-89). Alternatively, wooden shoring or framework was used.
Depths varied considerably; wells up to 15 m deep are not uncommon, and
although wells of more than 25-30 m are rare, examples are known of Roman wells
about 60 m deep in the Aures region of Algeria, and even 80 m in the case of a
Roman well near Poitiers in Gaul. For really deep wells, perhaps the biggest dis-
advantage was the sheer effort required to haul the water up (Hodge 2000: 30). To
prevent people falling in accidentally, the top of a well was usually surrounded with
a wellhead or puteal-either of terracotta or stone; the latter in some cases elabo-
rately decorated. Some stone well-surrounds show deep scoring marks from the
friction of ropes where the jars were pulled up by hand. A stone or wooden frame-
work might support a pulley to ease the task of hauling up the water jars. Cuttings in
stone well-surrounds sometimes indicate that the well could be covered with a
metal or wooden cover (e.g., on the Odeon Hill at Carthage).
Larger wells are occasionally encountered. In some cases they might take the
form of a deep and wide excavation with a staircase leading down to the water table,
to enable people to descend into the well and draw water. More often, an excep-
tionally wide well indicates the use of some kind of mechanical water lifting device
(chapter 13). Rectangular wells with an opening about 3 m long and 1 m wide seem
designed to house a chain of pots strung over a wheel and moved by draft animals
through an angle gear drive (saqiya). A Roman example was found in the center of
the Square of the Cisterns at Ptolemais (over 11 m deep), and probable Byzantine
examples are known from Andarin and Qasr Ibn W ardan in Syria (A. I. Wilson
2004: 120-21). In London, a well 2.6 m square and some 5 m deep housed a me-
chanical water-lifting device with a series of wooden containers connected by iron
chain links, and probably driven through a saqiya gear (now lost); it may have
served a set of public baths (Blair and Hall 2003). The well itself was lined with wood
and cross-braced with timbers that could also have served as a kind of ladder to
enable descent for maintenance or cleaning.
Domestic wells were usually located centrally in the house, often in peristyle
courtyards; sometimes a well shaft placed between two rooms might have an arched
HYDRAULIC ENGINEERING AND WATER SUPPLY 287
opening within the wall to allow access from both sides. The quality of well water
depended on the local geology, and also on population density and the drainage
technologies in use. Urban wells were at risk of contamination if sewage or refuse
was discharged back into the aquifer via cesspits; this potential problem did not
deter their use, and the consequent risk of disease may have been little appreciated
in antiquity. It was understood, however, that overextraction from a well could
have an effect on neighboring wells. One of Solon's laws decreed that if there was an
existing public well within 4 stadia (720 m) one had to use it rather than dig another;
otherwise people could dig their own wells. But if water was not found after digging
to a depth of 60 ft, the party could draw water from a neighbor's well twice a day
(Plutarch, Sol. 23.5-6).
CISTERNS
Cisterns were developed as an alternative to wells, storing rainwater collected from

the roofs, courtyards, or other paved areas of buildings. Domestic cisterns were
widespread around the Mediterranean in the Greek and Roman worlds, although
less common in the wetter regions of northern Europe.
Among the earliest rainwater-collection cisterns are examples from the MM III
period at Zakro, and from the LM III period at Tylissos. Both are open and cir-
cular, with steps for access; the Tylissos example had a rectangular settling tank
(Biernacka-Lubanska 1977: 27-28). Later cisterns were usually covered, to reduce
contamination and evaporation; a common form in the Greek and Roman worlds
was the bottle- or carafe-shaped cistern (depending on proportions), with a nar-
row neck or shaft at the top and a wider body below. Depths commonly range
between 3 m and 7 m, and volume was achieved by widening the body of the cistern
as far as the solidity of the rock into which it was cut would allow. In Athens, bottle
cisterns make an appearance perhaps in the early fourth century B.C., and from
the middle of that century onward increasingly came to replace wells in the agora
(Camp 1982: 12). In the Hellenistic period a different form of cistern became com-
mon in North Africa and the western Mediterranean-long and narrow, often with
rounded ends, its width was limited by the constraints of roofing with cover slabs
either laid flat or pitched against each other in pairs. Depth compensated for the
narrow width to achieve sufficient storage capacity. Examples are found in Hel-
lenistic Cyrenaica (Berenice) from the second century B.C. onward, in the Punic
world (e.g., Carthage and Ampurias), Cosa in the third century B.C. (Capitolium
cistern), and even in southern Gaul (Enserune, first century B.C.). Rectangular
cisterns roofed with slabs carried on a series of arches first make their appearance
in the Hellenistic period (Delos), and later spread to the Nabataean kingdom and
other parts of the Near East (Oleson 1991: 57-58; 1995).
The introduction of cisterns was made possible by the invention of waterproof

lining mortars, allowing the cistern to retain water even where the rock it was cut
into was permeable. With the exception of the large, open circular cisterns asso-
ciated with the fourth-century B.C. ore washeries in the mining region of Laurion,
where the mortars have been shown to contain high proportions oflead litharge, a
by-product from the silver extraction process (Conophagos 1980, 1982), very little
analysis has been done on ancient Greek waterproof mortars. Better understood
are the Punic hydraulic linings-often a bluish-gray mortar derived from marl
days, mixed with ash, and bulked out with a gravel aggregate. The Carthaginians
used a pinkish cement, containing crushed terracotta, for cement floors, and al-
though this material was later to become standard Roman cement for waterproof
linings (often referred to in modern archaeological literature as cocciopesto or opus
signinum), the Carthaginians themselves seem to have preferred the ash and gravel
mortars as waterproof linings.
At Carthage in the third century B.C., many domestic wells were replaced
by cisterns, at considerable effort and expense, and disused wells were reused as
soakaways for cistern overflow (A. I. Wilson 1998: 65-67). This change may have
been the result of either climatic change (lowering of the water table) or increasing
population pressure (leading to greater abstraction, or to contamination of the
water table). A similar switch from wells to cisterns in the Athenian agora in the
fourth century B.C. has been attributed to drought (Camp 1982), but this hypoth-
esis encounters the objection that increased reliance on rainfall collection seems a
poor response to drought. In the case of Carthage, at least, increasing population
pressure on groundwater resources seems a more likely explanation.
Other forms of cistern include tunnel cisterns, where a well-like shaft leads
down to two (or occasionally more) rock-cut tunnel chambers branching off the
base. The depth of such cisterns depends on what depth a sufficiently solid rock
stratum could be reached in which to excavate the lateral chambers. An early
Hellenistic cistern at Euesperides (Benghazi, Libya) with four arms was barely over
2 m deep. Tunnel cisterns are common in late Republican and early Imperial Italy,
especially in Latium where they could be dug into the soft volcanic bedrock, and in
Campania, as at Pompeii; they are also found at Roman sites in North Africa such
as Lepcis Magna (Walda 1996: 126). These tunnel cisterns, like the other types
described above, stored collected rainwater; but in Italy a variant type might be dug
down into the aquifer, collecting groundwater seeping through the sides of the
chambers. The greater the surface area of the walls, the greater the collecting ability
of such a cistern; consequently, some were developed as networks of parallel or
intersecting tunnels, particularly in Latium and Campania (Devoti 1978; Doring
2002a).
The Roman period saw an important technological advance in cistern con-
struction, with the introduction of mortared rubble building techniques and
vaulting, which allowed much larger covered cisterns to be constructed. The vol-
ume of earlier cisterns had been constrained by the limitations of roofing meth-
ods or rock-cut techniques, but now barrel and cross vaults allowed much greater
widths to be spanned, and multiplying the number of chambers could increase
volume further. The strength and durability of mortared rubble construction meant
that, with thick walls and external buttresses to counter the lateral pressure exerted
by the water inside, cisterns could now be built partly or wholly above ground,
rather than necessarily being sunk underground. This not only allowed their con-
struction in areas where groundwater was close to the surface, but also meant that
water could be run out by gravity flow from large aboveground cisterns, allowing
irrigation of gardens or orchards, rather than having to be lifted out arduously by
hand.
The most common type of Roman cistern is, therefore, rectangular and barrel-
vaulted, with one or more chambers, and a draw-hole about 0.50 m square in the
center of the vault. Terracotta supply pipes or cement-lined built channels from the
roof or courtyard enter above the spring of the vault; tide marks on the lining
mortar suggest that most cisterns were rarely filled above the spring of the vault.
Like wells, cistern mouths were usually protected by a surround or puteal, and a
frame with a pulley was often set over them.
Cisterns of all kinds were waterproofed by the application of mortar linings;
the most widespread type used in the Roman world was the pinkish opus signinum
containing crushed terracotta, but in parts of North Africa the Punic ash-based
bluish-gray mortars remained in use alongside this formula. The joints between
wall and floor were sealed with a quarter-round molding, and the corners were
rounded internally to assist cleaning; sometimes a circular sump was provided in
the floor to assist removal of sludge.
Domestic cisterns collected water from the roof and courtyard of the house in
which they were located; the amount of water they collected was therefore deter-
mined by the local rainfall combined with the roof area or impluvium available
for rainwater collection. Rural cisterns might have specially constructed paved or
plastered areas acting as collection impluvia, as for example in Hellenistic and
Byzantine cisterns on Yeronisos Island, Cyprus (Connelly and Wilson 2002).
The so-called astynomoi inscription from Pergamon in Asia Minor (a Trajanic
regulation repeating a Hellenistic decree) illustrates the importance of domestic
cisterns to the overall water supply of a town; among other duties the town mag-
istrates were required to keep a register of all private cisterns in the town and check
annually that they were kept in good order (Klaffenbach 1954). The importance of
rainwater collection as a complementary water supply technique alongside wells,
springs, and aqueducts is highlighted not only by the virtual ubiquity of domestic
cisterns in the Mediterranean region, but also by the fact that many public buildings
with a large footprint were equipped with large rainwater-collection cisterns. In
Delos, the Hellenistic theater supplied runoff water to a large cistern roofed with
slabs carried on cross arches (Oleson 1995: 716-17). In Roman North Africa, theaters
and odea at Bararus, Thugga, and Carthage were provided with cisterns under the
stage, fed by runoff from the cavea and orchestra, while the third-century A.D.
amphitheater at Thysdrus in Tunisia acted as an impluvium for cisterns some 100 m
to the north. Large paved public spaces, such as the forum at Pompeii, the precincts
of the Temple of Saturn at Thugga, or the South Forum Temple at Sabratha, also
collected water for large cisterns (A. I. Wilson 1997: 59).
The quality of water stored in rainwater collection cisterns was acknowledged to

be inferior to that from wells or sources of running water (Pliny, HN 31.21.34; Oleson
1992: 887-88). Standing water became stagnant and less aerated, and although the
underground placement of most cisterns helped to keep the water cool and away
from sunlight, inhibiting somewhat the growth of algae and insects, ancient authors
often recommended boiling cistern water before drinking it, or adding salt. The
water collected in large cisterns fed from public spaces must have been nonpotable,
and indeed it is likely that, where other sources were available, cistern water was
used primarily for purposes other than drinking, such as washing floors or watering
gardens. Basic methods of water purification were employed, such as metal grilles
on the inlets of feed pipes, and settling tanks, and large quantities of charcoal found
in domestic cisterns in Thysdrus and Leptiminus in North Africa may suggest the
addition of charcoal to cistern water to remove odors by the absorption of colloids,
a practice still current in the nineteenth century (A. I. Wilson 1997: 81). Vitruvius
(De arch. 8.6.15) recommends adding salt to cistern water to purify it.
AQUEDUCTS AND LONG-DISTANCE

SUPPLY
The Near Eastern Background

The Assyrian empire saw the development of ambitious water supply schemes
involving long-distance artificial canals, usually derived from perennial rivers, and
led through intervening ridges by means of tunnels several hundred meters, or even
several kilometers, long. The longer tunnels were dug between pairs of vertical
shafts, which enabled the overall course of the tunnel to be surveyed on the surface.
This shafts-and-gallery tunneling technique broke up the problem of underground
surveying into short, manageable sections, and the shafts were set at sufficiently
close intervals that the tunneling gangs could find each other underground. They
served also for the removal of spoil, and to assist ventilation (cf. chapter 12).
The Assyrian king Assurnasipal II (884-859 B.C.) dug a canal 19.5 km long from
the Upper Zab river to supply Nimrud; it passed under a rock ridge at Negoub via a
tunnel 7 km long dug between pairs of vertical shafts. The system both supplied the
city of Nimrud and irrigated the fields around it. Two later tunnels in the region
seem to be repairs or additions to this scheme: one was cut by Tiglath-Pileser
III (744-727 B.C.), but later blocked, and the other by Esarhaddon (680-669 B.C.),
whose construction inscription records that the earlier tunnel of Assurnasipal II
had silted up. Renovation and maintenance are constant themes in both ancient
and modern water-supply systems.
Similar techniques were used for the water supply for Arbela in eastern Iraq.
This tapped water from three rivers at Bastura and included a tunnel dug between
vertical shafts; an inscription on the entrance of the tunnel records construction
by Sennacherib (704-681 B.C.). Traces of what may be a comparable system were
recorded at Babylon by Rassam in the late nineteenth century; he mentions four
wells lined with rings of red granite, in a straight line, which communicated with a
subterranean aqueduct. The source of the aqueduct was not established with cer-
tainty, but it was thought to have come either from the Euphrates or from a canal
northeast of the main mound at Babylon. Suggested dates for the system range from
the late eighth to the mid-fifth century B.C. (Dalley 2001-2002).
Sennacherib's aqueduct to Nineveh, completed in 690 B.C., was a long and
wide canal used for both irrigation and (probably) urban supply, tapping the
waters of the Gomel River at Bavian. It crossed a wadi at Jerwan on a bridge over
280 m long and 22 m wide, built in ashlar masonry with five corbelled arches
(Jacobsen and Lloyd 1935). Hezekiah's approximately contemporary tunnel (715-
696 B.C.) that brought water into the city of Jerusalem from an underground
spring is a 537-m winding spring-capture tunnel (chapter 12).
The Assyrian water supply schemes involving tunnels dug between sets of ver-
tical shafts have a bearing on the vexed question of the origin of qanats. Qanats are
one of the most important developments in the history of water engineering; they
have enabled human habitation and agriculture in some of the most inhospita-
ble, arid areas on earth. A qanat is an underground gallery that collects groundwater
and leads it by gravity flow to emerge where the ground surface level is lower,
hundreds of meters or even many kilometers from the source ( figure 11.1). The long
underground tunnels were dug in relatively short sections between pairs of vertical
shafts, again using the shafts-and-gallery tunneling technique, just as in the tunnels
on the large Assyrian aqueducts. The difference lies in the water source, which is not
a flowing river or spring, but groundwater, usually tapped at a considerable depth
in a piedmont zone in an arid region. Qanats are found from China to the Sahara
and North Africa; they were even introduced to the New World by the Spanish, and,
despite the increasing use of motorized pumped wells, they are still used today in
many arid regions.
It was long held that the qanat must have originated in Persia, largely because
that is where the largest numbers of qanats are found; and their origin has fre-
quently been attributed to the Achaemenids. More recently, several archaeologists
working in the Arabian Peninsula have argued that qanats (afiaj) originated there,
claiming to have discovered pre-Achaemenid qanats of the early first millennium
B.C. (Magee 2005). To date the arguments are inconclusive; some qanats in south-
eastern Arabia seem to be spatially associated with sites dated by pottery to the Iron
Age II period, but that period extends from about 1000 to 600 B.C., and there is no
conclusive proof that the qanats in question were created at the beginning rather
than the end of it. A late date within this bracket would still allow the possibility of
Achaemenid introduction to the region. It is true, however, that no qanat in Iran
has yet been shown to be of Achaemenid date-although the significance of this
Section
',
...................
',
',_ Water table
Plan
·····O··-·O·····o·····o·····o·····O·····O---
O ----- --.--- ----- ----- ---- ...... ....... ·--
Figure 11.1. Diagram of a qanat or foggara. (Drawing by Alison Wilkins, reproduced by

permission.)
observation diminishes when one remembers that almost no qanats in Iran have
been investigated archaeologically.
The earliest securely dated qanats belong to the Achaemenid period at Ayn-
Manawir in the Kharga Oasis in Egypt, where remains of 22 ancient qanats are
known. They are dated both by stratigraphic finds and by associated ostraka re-
cording the sale of water rights from 443 B.C. onward (Wuttman 2001). Polybius
(10 .28), describing campaigns in Media in 210/209 B.C. between the Seleucid king
Antiochos III and the Parthian king Arsaces II, dearly refers to qanats, whose con-
struction he says was encouraged by the Achaemenid kings by tax concessions on
the land thus brought under cultivation.
In his fundamental study of qanats, Go blot (1979) proposed that the technique
was spread to regions around the Mediterranean by invading peoples or refu-
gee groups; thus the Romans would have introduced qanats to North Africa, while
refugees from Persia introduced them to the Sahara in the early Middle Ages. More
recent work provides a very different model. The spread of the technique to Africa
was due not to transfer through the Mediterranean, but to diffusion along Saharan
trade routes. In the late first millennium B.C., qanats (foggaras) formed the agri-
cultural basis of the emergent Garamantian state in Fazzan (Libyan Sahara); the
technology had probably been introduced from the oases in the western desert of
Egypt, where, as we have seen, it was already in use in the Achaemenid period. From
Fazzan it later spread to the Algerian Sahara, and diffused northward to Roman
North Africa (A. I. Wilson 2005, 2006a), where some Roman aqueducts display a
blend of traditional qanat technology with Roman techniques of reinforcement of
the water channel. Very similar qanat-aqueducts are found in the region around
Trier and Luxembourg and may have been introduced there from North Africa
(Luxembourg: Kohl and Faber 1990; Kohl et al. 1995; Faber 1992; Schoellen 1997;
Germany: Kremer 2005; Grewe 1988: 93). Elsewhere in the Roman world, however,
the use of the qanat was largely limited to Syria, Judaea, and Egypt. Further east, its
spread from Iran to Afghanistan at an unknown but probably early date may also
have been facilitated by trade contacts along the Silk Road.
The development and spread of qanats is important because-as the Achae-
menid examples at Ayn Manawir show-it demonstrates the availability and use at
an early date of complex and difficult engineering water schemes in communities
outside the main civic centers. In many parts of the ancient world the qanat was the
only long-distance water supply technology available prior to the Hellenistic or
Roman periods.
As we have seen, already in the early ninth century B.C. Assyrian water engi-
neers were able to dig tunnels several kilometers long using the same principle as
for the tunnels of qanats. The evidence currently available suggests two possibilities.
Either the Assyrians developed the shafts-and-gallery tunneling technique, which
was later applied to the realization that groundwater found at depth in one place
may nevertheless lie above the ground surface elsewhere and could be led out to the
surface via a tunnel that could be dug only by the shafts-and-gallery technique; or
the qanat was invented as a package both of the shafts-and-gallery technique and
the realization that groundwater found at depth could be transported elsewhere.
Until further early qanats have been reliably dated, certainty is impossible, but a
priori the first of these two possibilities seems easier to believe. Furthermore, if the
Assyrians invented the shafts-and-gallery technique and it was only later applied to
the qanat, the separate diffusion trajectories of the qanat and the shafts-and-gallery
tunneling technique become easier to understand: the shafts-and-gallery tunneling
technique spread to the Greek and Etruscan worlds while the qanat did not.
GREEK WATER SUPPLY SYSTEMS
The archaic period saw the first public water projects of the Greek world, with the
development of springs with fountain houses at Megara by the tyrant Theagenes
(ca. 640-620 B.C.; Pausanias 1.40.1), and perhaps the monumentalization of the
spring of Lower Peirene at Corinth, although this is not closely dated (Hill 1964).
For Samas and Athens, long-distance aqueducts were constructed. The projects at
Megara, Samas, and Athens were all initiated by tyrants wanting to secure popular
favor by public works engineering schemes that were both grandiose and useful.
The aqueduct at Samas, 2.5 km long and built by the tyrant Polycrates (ca. 550-522
B.C.), is best known for the famous tunnel named after its designer, Eupalinus of
Megara (Herodotus 3.60), a 1-km-long gallery just before the aqueduct enters
the city (Kienast 1995). The height of the overlying hill prevented construction
by the shafts-and-gallery technique, so the tunnel was surveyed from opposite
ends, meeting near the middle (chapter 12). The aqueduct itself tapped water from
a spring, and apart from the tunnel of Eupalinus it largely runs along the contours;
significantly, for considerable stretches immediately up- and downstream from
the tunnel of Eupalinus the structure takes the form of an underground tunnel dug
in the shafts-and-gallery technique. Given the cultural and technological contacts
between the Persian Empire and eastern Greek states at the time, it seems that this
must represent an adoption of shafts-and-gallery tunneling methods from the
Achaemenid practitioners of this Assyrian technique.
The Enneakrounos seems to have been a monumental fountain house for the
local Callirhoe spring, sponsored by Pisistratus (Thucydides 2.15.2-3; Pausanias
1.14.1); its location is disputed. In the late sixth century the Pisistratids also built a
long-distance aqueduct, bringing water from over 8 km away in the Ilissus valley in
a conduit made of ceramic pipes laid, for much of its length, in a tunnel. Calcareous
deposits extending only halfway up the pipe walls show that the pipeline did not
run full or under pressure. In any case, the openings in the top of each pipe section,
which allowed the workmen to insert a hand to plaster the inside of the joints, and
were closed by ceramic covers that were not watertight, made that impossible
(Tolle-Kastenbein 1994, 1996).
The archaic-period aqueducts seem to have fed only public fountain houses,
but already in the fifth century B.C. it seems that people had begun to tap them
illegally for private uses; Themistocles as hydaton epistates at Athens in 490 B.C.
fined those who diverted water from the public system (Plutarch, Them. 31.1). How
quickly, and when, aqueduct networks developed to serve uses other than public
fountains remains unclear, since accurately dated Greek aqueducts are rare between
the archaic and Hellenistic periods, and relatively few residential areas in Greek
cities have been systematically investigated. The aqueduct of Priene in Asia Minor
may be contemporary with the refoundation of that city on its present site in the
late fourth century B.C.; the aqueduct consisted of a terracotta pipeline laid in a
ditch covered with stone slabs. It fed a reservoir inside the walls in the upper part of
the city, which in tum supplied public fountains and numerous private houses
through a network of terracotta pipes. At Syracuse, the three earliest aqueducts
probably belong to the reign of Hieron II (270-215 B.C.); all take the form of rock-
cut tunnels with access shafts, although curiously with a second gallery excavated
just above the one carrying water (R. J. A. Wilson 2000: 12-14).
Although Greek architects occasionally employed the arch (Boyd 1978; Hell-
mann 2002: 266-77), they did not use it to create arcades to carry aqueducts across
natural depressions in the terrain. Greek aqueducts, therefore, largely followed the
contours of the landscape, either in rock-cut channels or, more commonly, ter-
racotta pipelines, which were not primarily intended to run full under pressure
HE.I.DEii
TANK
---------------=,,,_.-.......,,,:-=--=--=-=-=-::-:::::--:~--:.t114111
~ I
V£NtER (-.~-'thl 17· 40m)
~270 ,n-Jj
- - - - - - - - - - - - - 2 ·111,n - - - - - - - - - - - - ) ' i l ' t
Figure 11.2. Diagram of an inverted siphon (h: loss of head). (After Hodge 1992: 148, fig.
102,by permission of A. T. Hodge.)
but nevertheless allowed a greater latitude in surveying accuracy than did an open
channel. Large-gauge pipelines of about 15 to 25 cm internal diameter were used
for main long-distance conduits; a double or triple line allowed for greater ca-
pacity. Greek aqueducts were, however, of relatively limited scale. The development
of more ambitious, long-distance projects with complex engineering works was
pioneered by the twin centers of Rome and Pergamon. Republican Rome (dis-
cussed below) led the way with the channel aqueducts of the Aqua Appia (312 B.C.)
and the Anio Vetus (272-269 B. C.), while Hellenistic Pergamon developed large-
scale and long-distance pipeline systems, although there was probably mutual in-
fluence between the two centers (Lewis 1999).
A key development that helped liberate aqueducts from the tyranny of the
contours by enabling the crossing of deep valleys was the invention of the inverted
siphon. Some of the earliest known large-scale examples are at Pergamon, although
there is a small inverted siphon 10 m deep on the 8 km pipeline supplying Olynthus,
dating between the sixth and fourth centuries B.C. (Lewis 1999: 157. n. 49). This
structure employed the principle that water could be conveyed across a depression
in a dosed pipe and would rise at the other end to nearly the same level at which it
entered the pipe ( figure 11.2). Five Hellenistic pipelines are known at Pergamon, all
later than the two earliest aqueducts at Rome. The first two Pergamene systems
supplied the lower city: the Attalus line (before 200 B.C.; a single day pipeline ca. 20
km long); the Demophon (early second century B.C.; a double pipeline also ca. 20
km long), both with inverted siphons of between 20 and 25 m depth. To bring water
to the royal palace on the citadel, the engineers working for Eumenes II (197-158
B.C.) constructed an aqueduct from a spring on the Magra Dag hill that ran for 42
km, initially in a single pipeline, but then in a double pipe and after 15 km in three
parallel terracotta pipes (16-19 cm internal diameter)-perhaps as a result of adding
tributary branches. The system had to cross a depression 3.5 km long and ap-
proximately 200 m deep just before it reached the citadel. The considerable pressure
that the siphon had to withstand precluded the use of terracotta pipes, so it was
built instead oflead pipes held in place by pierced stone framing blocks set at close
intervals. Broadly contemporary with the Madra Dag aqueduct was Pergamon's
Apollonius line, followed in the later Hellenistic period by the Selinous West
pipeline, both without siphons. The Madra Dag aqueduct may well have inspired
the 150 m deep inverted siphon on the Karapinar aqueduct at Smyrna, which
probably belongs to the second century B.C. (Lewis 1999).
ROMAN AQUEDUCTS
Although aqueducts are often considered a quintessentially Roman engineering

accomplishment, we have seen that they have a long tradition in the ancient Near
East and in the Greek world, including gravity flow conduits, tunnels, pipelines and
inverted siphons. The Roman contribution consisted of innovations that together
enabled a much wider and more effective uptake of the basic technology: the use
of arcades to carry channels over valleys and low-lying terrain; the availability of
concrete as a cheap and adaptable building material; the adoption of waterproof
cement linings from the Punic world or Hellenistic Sicily; the expanded use of
lead piping and of bronze stopcocks on distribution systems; and the introduction
of settling tanks and storage and regulation reservoirs on the network. Increasing
levels of state and personal wealth helped fund the construction of numerous aq-
ueducts, for which increasing urbanization generated demand. Aqueducts also fos-
tered the growth of a culture of public bathing, doubtless with a feedback effect
whereby the increased popularity of public baths in turn generated a need for more
copious urban water supplies; by the early Principate, aqueducts and public baths
had become linked features of Roman urbanism. The ornamental use of water
in public fountains and nymphaea also became a striking feature of Roman civic
architecture-coexisting with the practical function of these structures.
Bathing and ostentatious display were certainly not the driving forces, however,
behind the three earliest Roman aqueducts, which were constructed long before
public baths or display fountains. Rome's first aqueduct, the Aqua Appia, was built
in 312 B.C.; it ran for about 16 km from springs east of Rome, and seems to have run
entirely underground in a vaulted conduit until it reached the city. We do not know
what structures it fed; presumably a few public fountains, and it provides indirect
but suggestive evidence for the growing needs of the city. It was followed by the
Anio Vetus (272-269 B.C.), a vastly more ambitious project, running originally for
81 km from the Anio River along the contours of the landscape, although later
modifications, with tunnels and arcades, shortened its course considerably by cut-
ting across valleys. The next aqueduct, the Aqua Marcia (144-140 B.C.), was funded
with the booty obtained from Rome's defeat of Corinth and Carthage in 146 B.C.
and ran for about 91 km, mostly underground, but with the final approach into
Rome for some 10 km on substructures and arcades. In all, some 1.5 km were on
substructures and 9.5 km on arcades (Aicher 1995: 36-37). For the Anio Vetus and
the Marcia, the irrigation of suburban horti may have been a subsidiary motive, but
this can hardly apply to the Aqua Appia, which ran underground and, in the time of
Frontinus, distributed very little of its water outside the city.
In the later second century B.C. (126-125 B.C.), the Aqua Tepula was added,
drawing water from warm springs in the Alban hills; for much of its course it
piggybacked on the arcade of the Marcia. Three further aqueducts were added by
Octavian/Augustus-the Julia (33 B.C.), again superimposed on the Marcia/Tepula
arcade; the Virgo (22-19 B.C.) and the Alsietina (2 B.C.). The latter brought poor
quality lake water from the Lacus Alsietinus northwest of Rome, which was used for
irrigation and Augustus' naumachia. The continued development of Rome, with
growth in population and an increase in the public bathing habit, prompted the
construction of four more aqueducts: the Aqua Claudia and the Anio N ovus (both
begun by Caligula and finished by Claudius in A.D. 52), the Aqua Traiana in 109,
and the Aqua Alexandrina in 226. The network not only became larger, but in-
creasingly complex, with cross-links between aqueducts within Rome, to enable
continued delivery of water if one aqueduct was under repair.
Rome's political and economic dominance ofltaly enabled her to lead the way
in developing ambitious water-supply schemes on a scale unmatched in the Greek
east except by Pergamon. But in the second half of the second century B.C., other
towns were beginning to follow suit. The censor L. Betilienus Varus was honored
around 130/120 B.C. by his home town of Alatri for various gifts, including an aq-
ueduct, whose final stretch carried water to the hilly citadel in an inverted siphon of
lead pipes, 3 km long and 100 m deep. It has similarities with the Hellenistic inverted
siphons of Madra Dag at Pergamon and Karapinar at Smyrna, in that it follows rises
in the intervening depression to reduce the length run at maximum pressure, and
Lewis has suggested a Pergamene link. But at Alatri the two deepest points along the
course of the inverted siphon were crossed by arcades, and the channel leading to the
inverted siphon was a built conduit (Lewis 1999: 153-54, 161-62).
By the first century B.C., the use of aqueducts was spreading, perhaps slowly,
throughout Italy. Research on the Pompeii aqueduct suggests that its initial phase
dates from the colony founded by Sulla around 80 B.C. (Ohlig 2001)-it ran almost
entirely underground from the vicinity of Avella to the north of Pompeii, feed-
ing some public fountains and, presumably, public baths. In a further four towns of
central Italy, allegedly Republican inscriptions (but not closely dated) refer to aq-
ueducts or water supply systems built by local magistrates or decurions: Firmum
Picenum (fermo, Marche), Interamna Nahars (Terni, Umbria), Praeneste (Pales-

trina), and Superaequum (ILLRP 594, 615, 659, 671).
We begin to see Roman aqueduct construction outside Italy with Caesar's aq-
ueduct at Antioch on the Orontes (Lassus 1983: 211; Downey 1961: 153), but the real
proliferation of the technology began in the Augustan period. This development
might be attributable to a concatenation of factors: increased prosperity after the
cessation of civil war, the spread of the public bathing habit, and perhaps most
importantly, the foundation of numerous veteran colonies which exported the
Roman urban model, with its bathing and amenities, to the provinces-although
not all Augustan colonies built aqueducts immediately. Augustan aqueducts at new
colonial foundations within Italy include Venafro (CIL X.4842 = ILS 5743), Lucus
Feroniae, Minturnae, Bononia, and Brixia (Keppie 1983: 114-16; Potter 1987: 144-
45). The largest was the Serino aqueduct, a complex network running for 96 km,
drawing water from sources at Serino and Avella and delivering it to several towns,
including Naples, Puteoli, Nola, Cumae, Pompeii, Atella, Acerrae, Baia, and Mis-
enum (Sgobbo 1938; Doring 2002a; 2002b ), a striking example of regional planning.
Augustan water-supply schemes overseas include the Cornalvo aqueduct at Merida
(Grewe 1993) and possibly the vast, doubtless aqueduct-fed La Malga cisterns at
Carthage. Other projects of the Augustan period include Sextilius Pollio's aqueduct
at Ephesus and the pipeline "Aqueduct C" at Berenice in Cyrenaica, probably
privately funded (Alzinger1987; Lloyd 1977: 199 ). The technical aid given by Rome to
client kings to assist in developing showpiece Roman-style towns seems also to have
included the services of water engineers: the aqueduct at Iol Caesarea (Cherchel,
Algeria) was probably built by Juba II (Leveau and Paillet 1976: 151-53; Leveau 1984:
61-62), while Herod's building projects in Judaea included aqueducts at Jerusalem
(also using Roman military labor) and possibly Caesarea, and his own palace/
fortresses at Jericho and Herodion (Amit 2002; Porath 2002; Patrich and Amit
2002). During the first century A.D. numerous aqueducts were built in Italy, and
increasingly in the provinces, but the bulk of dated provincial aqueducts date from
the late first through second centuries. New aqueducts were still being built in the
Severan period, but very few new constructions postdate A.D. 230, although ex-
isting aqueducts were repaired. There was some resurgence in aqueduct construc-
tion in the Byzantine period, especially under Justinian and especially in the east. In
the west, the maintenance even of existing aqueducts seems increasingly to have
been beyond the resources of most cities from the fifth century onward. In Rome,
only four of the city's eleven aqueducts were repaired after they had been cut and
blocked in the Gothic siege of 537, and their restoration and (intermittent) main-
tenance was a burden assumed by the popes.
The interaction between Hellenistic and Roman aqueduct traditions is shown
by the fact that book 8 (on water) of Vitruvius' De architectura is largely derived
from Greek sources, probably in particular a source from Hellenistic Pergamon.
But for this reason Vitruvius' text is a poor reflection of current aqueduct tech-
nology in Roman Italy at the time he wrote. Although he has added some of his own
material on water distribution, his account of conduit design omits all mention of
arcades, the most distinctive feature of Roman aqueducts for a century before he
wrote (Lewis 1999).
For most of their course Roman aqueducts usually took the form of built
masonry or concrete channels, vaulted or roofed with flat slabs or gabled slabs or
tiles (figure 11.3) (Hodge 1992; Bodon et al. 1994; R. J. A. Wilson 1996; A. I. Wilson.
1997, 2000a). As far as possible, the course would follow the contours of the terrain,
running just below the surface in a back-filled trench. This was the simplest and
cheapest method of construction, while at the same time minimizing obstacles to
movement at surface level. Legal regulations protected its course from planting, to
avoid the channel being fissured by roots or damaged by agricultural activity.
Where the land surface rose significantly above the level of the channel, the channel
had to be tunneled, using techniques described in chapter 12. In other places the
aqueduct might cross lower-lying terrain, out of necessity or because this afforded a
shortcut, instead of winding along the contour up one side of a valley and back
down along the other. In these cases, the channel would run on a low wall or sub-
structure, up to a height of about 2 m; beyond this height, an arcade was used, since
it required less material and allowed movement through the structure. These aq-
ueduct arcades-the modem image of the Roman aqueduct-were frequently am-
bitious affairs, either in their length (Aqua Claudia from Romavecchia to the city, 10
km; Carthage at Oued Miliane, 2 km; figure 11.4), or their height, achieved either
through bracing arches between piers (Merida, Los Milagros, 28 m; Cherchel;
Dougga) or by stacking arcades at two or even three levels (Tarraco, 26 m; Segovia,
28 m; Pont du Gard, 49 m). The Pont du Gard on the Nimes aqueduct is the tallest
surviving aqueduct arcade, made of three levels of different sized arches; it ap-
proaches the limits of stability for an arcaded structure. As Roman engineers grew
more confident in the use of the audacious arcade, in the later first or during the
second century A.D., some aqueducts were shortened by replacing a section that
ran upstream along one side of a valley and back down the other with an arcade
straight across (e.g., at Chabet Ilelouine on the Cherchel aqueduct, and the Ha-
drianic Ponte S. Gregorio on the Anio Novus). Such shortening of a route reduced
maintenance costs.
For reasons of both stability and cost, Roman engineers employed inverted
siphons to cross valleys deeper than 50 m, or a combination of inverted siphon
and arcade bridge (possibly termed venter), as at Aspendos and Lyon. The inverted
siphons usually employed lead pipes, frequently in a battery of multiple pipes laid
side by side to achieve the same capacity as the open channel and to achieve re-
dundancy in case of pipe failure. The transition between open channel flow and the
pipeline was usually achieved by means of header and receiving tanks, of which
good a example can be seen near Soucieu on the Gier aqueduct at Lyon. On the
aqueduct of Termini Imerese (Sicily) the siphon seems to have started as a large-
diameter concrete tube dropping vertically from the base of the channel (Belvedere
1987: 61-62). In the eastern Mediterranean region, inverted siphons were frequently
Figure 11.3. Channel of the Aqua Traiana in Rome. (Photograph by A. I. Wilson.)
built as stone pipelines, constructed of squared blocks with a hole bored through
and fitted together with male-female joints. These were once thought to be Hel-
lenistic; they are now nearly all recognized as Roman in date. Many of the blocks
have small holes bored through their walls, stoppered by small plugs set in mortar,
probably to facilitate cleaning. Examples have been reported from other parts of the
Roman world, too: Dalmatia, Italy, Gaul, North Africa ( figure 11.5), and Spain
(Stenton and Coulton 1986: 45-53; Hodge 1992: 33-41; A. I. Wilson 2000a: 599). The
reasons for the preponderant use of lead pipes in the west as opposed to stone pipe
blocks in the east remain unclear but may be related to the greater availability of
lead in the western Mediterranean (chapter 4).
Small-scale aqueducts, especially those serving rural estates, might consist
entirely of terracotta pipes. In the provinces of northwest Europe, some aqueducts
consisted of timber pipes made from hollowed out tree-trunks and jointed to-
gether with iron rings (R. J. A. Wilson 1996: 21-23; A. I. Wilson 2000a: 602).
Ideally, aqueducts maintained a fairly consistent gradient along their course,
although in practice gradients varied considerably along the course of a single aq-
ueduct. Recorded extremes vary between falls of 0.07 m and 16.4 m per km. Where
an aqueduct needed to lose height rapidly, a series of cascades or drop-shafts was
used to dissipate energy (Grewe 1985; Hodge 1992: 171-97; Chanson 2000).
Figure 11.4. Arcade of the second-century A.O. Zaghouan-Carthage aqueduct, crossing

the valley of the Oued Miliane. (Photograph by A. I. Wilson.)
Roman aqueducts often tapped sources of water with high levels of dissolved
calcium carbonate and other minerals. These precipitated out of solution in the
aqueduct, leaving calcium carbonate deposits (sinter) on channel walls and floor;
turbulence and heat increased the rate of precipitation. As Frontinus mentions
(Aq. 122), such deposits needed to be cleared if flow was not to be reduced and
ultimately blocked, and piles of sinter fragments around shaft openings of the
aqueducts in the Roman Campagna (Ashby 1935: 44) and along the course of the
Nimes aqueduct testify to such maintenance. But equally, the massive thickness of
the sinter deposits elsewhere on the Nimes aqueduct (at the Pont du Gard) and on
numerous other systems, suggests that such maintenance was neither universal nor
perfect.
Efforts to improve or protect water quality were made, but they took the form
of mechanical rather than chemical purification. Settling tanks were sometimes
provided along the course of an aqueduct, where sediment could settle out (e.g., on
the Aqua Virgo and the Anio Novus; Hodge 1992: 123-24), or small basins were
provided in the floor of the channel (on the Zaghouan-Carthage aqueduct), or as an
afterthought added to the aqueduct (as at Siga in Algeria; Grewe 1998). Mesh filters
might also be provided on distribution tanks (see below), or on the outlet pipes of
reservoir cisterns (e.g., Carthage, Bordj Djedid, and Dar Saniat; A. I. Wilson 1997:
81-82) and there are occasional instances of water being filtered through sandbags
(at Cirta) or through a bank of potsherds and shells (Dar Saniat, Carthage). Some
improvement in the taste and odor of water could also be achieved by aerating it,
for example by running it over a stepped cascade as it entered a storage or distri-
bution cistern (A. I. Wilson 1997: 81-82).
Figure 11.5. Stone pipeline aqueduct at Bled Zehna, near Dougga in Tunisia, possibly part
of an inverted siphon. (After L. Carton, "Etudes sur les travaux hydrauliques des romains
en Tunisie," Revue Tunisienne 3.12 [October 1896): 544, fig. 17.)
THE URBAN DISTRIBUTION NETWORK
Distribution to the end user almost always occurred through closed pipes. In some
larger cities areas of the town were served by branch channels, but ultimately water
was delivered to most fountains and all houses through pipes (Last 1975; Eschebach
1983; Lassus 1983; Hodge 1992; A. I. Wilson 1998, 2001; Jansen 2000). The transition
from main channel to a multiplicity of pipes was often made by means of distri-
bution tanks or basins (castella divisoria). The earliest known such castellum is at
Pompeii, consisting in its first phase (ca. 80 B.C.?) of an open circular basin into
which the aqueduct discharged, and from which three lead mains led to different
parts of the town. In its second, Augustan phase, the basin was enclosed within a
roofed building and arrangements were added to allow manual regulation of flow
into the different pipes; screen filters were also provided to trap debris (Ohlig 2001).
At Nimes, a circular distribution basin with five pairs of pipes leading off, pre-
sumably to different quarters of town, was also equipped with filter grilles. At
Rome, the only terminal distribution tank whose form is known seems to be the
elevated terminal basin of the Aqua Claudia, drawn by Piranesi before its destruc-
tion by fire; numerous lead pipes departed from a rectangular tank, elevated on tall
vaulted substructures, into which the main aqueduct discharged. Smaller circular
distribution chambers on the Aqua Marcia near Porta Viminalis and on the Aqua
Traiana seem to belong to subsidiary branches or to have been placed upstream of
the main terminus of the aqueduct (Ashby 1935: 149; Lanciani 1881: 459-61). In none
of the archaeologically known castella is there any support for Vitruvius' principle
of delivery to different classes of user being handled differently, and his proposed
arrangements for distribution of water from a castellum (De arch. 8.6.1-2) must be
treated as idiosyncratic suggestions that may never have been realized in practice
(Ohlig 2001).
At Pompeii, the city whose distribution network is best understood, the pipes
from the main castellum led to a series of 12 subsidiary distribution points, which
took the form of small lead tanks (now lost) of about 1 m 3 capacity, elevated on
brick towers. Pompeii is on a sloping site, the lowest parts of the town some 20 m
below the castellum, and if the piping system had gone straight from the castellum to
the houses in the lowest part of the city the head of pressure would have been 20 m,
inconvenient for users and making operation of stopcock taps difficult. The use of
intermediate open tanks, elevated on towers about 5 m high, broke the system up
into segments where the head on the section to the final user was no more than 5 m,
reducing the problem of excess pressure while retaining enough pressure to supply
the upper stories of houses in the vicinity of each tower. Herculaneum had a similar
system, but such a systematic use of intermediate pressure towers is not paralleled
elsewhere. One suspects, however, that the dendritic schema of this distribution
system, with a primary castellum from which a few pipes led to subsidiary distri-
bution centers, from which connections led to multiple end-users, was found at
other sites. This was not, however, the only model. Some aqueduct networks, such
as the second-century A.D. Zaghouan aqueduct at Carthage, had distribution net-
works which-in their primary divisions-consisted of a binary division or bifur-
cation of a subsidiary channel off the main channel, with a small chamber in which
sluice gates could regulate the flow into one or the other channel (A. I. Wilson
1998). Despite Frontinus' insistence that all connections to an aqueduct had to be
made from a castellum (Aq. 103.4, 106.1; Rodgers 2004: 135-36, 289; cf. Vitruvius, De
arch. 8.6.1-2), this is an administrator's ideal and did not always happen in prac-
tice. At Carthage, Volubilis, Pompeii, and elsewhere, numerous piped connections
were made directly from the main channel, either inside or outside the city (A. I.
Wilson 1998, 2006b).
In the eastern Mediterranean, the Greek tradition of ceramic pipe distribution
networks persisted in the Roman period, and ceramic pipes were normal for urban
distribution systems. By contrast, in the Roman central and western Mediterranean
the use of lead pipes for distribution systems was much more prevalent. Although
lead pipes on extramural lines had been used in the east at Ephesus in the archaic
period (Alzinger 1987: 180) and in the inverted siphon of the Madra Dag line at
Pergamon, it is unclear when their use became common, especially for intramural
distribution. The use of lead piping in the west was doubtless facilitated by the
opening of large silver and lead mines in Spain in the late Republic, although lead
sources are not lacking in the east. Lead piping was probably a feature of the
Pompeian distribution system even in the original phase, around 80 B.C.
As with lead-pipe networks, the date of invention of stopcocks or taps, and the
process and rate at which their use spread, remains unknown. The stopcock may be
either a Hellenistic or a Republican invention (Hodge 1992: 322-26). Roman taps
consisted of a bronze cylinder set vertically across the pipeline, in which sat a
Figure 11.6 . Bronze stopcock on a lead pipe supplying an ornamental fountain in the
peristyle of the House of the Vettii, Pompeii. (Photograph by: A. I. Wilson.)
bronze plug pierced by a hole ( figure 11.6) . When the hole was aligned with the
pipe, water could flow; a quarter-tum either way would shut the flow off; the
action is easy, precise, and can handle fairly high pressure. Such taps are found
both as stopcocks along the course of a pipe and as discharge taps to turn a fountain
or spout on and off at the point of use. The widespread use of taps on urban
distribution systems in the western Mediterranean seems to be a feature of the first
centuries B.C. and A.O. Taps were common at Pompeii by A.O. 79, one house
having as many as 33 (Hodge 1992: 322-26; Fabio and Fassitelli 1992[?]; Jansen 2002:
50-53) .
Piped water was distributed to a minority of elite private houses-at Pompeii,
some 10 percent of households-and was evidently a sign of status, as well as-or
even more than-utility. Where domestic piping systems are traceable, there seems
to be more emphasis on feeding ornamental fountains or pools rather than sup-
plying water to kitchens (A. I. Wilson 1995; Jansen 2 0 02; Jones and Robinson 2005).
RESERVOIR CISTERNS
Very large cistern complexes are sometimes found on aqueduct networks, which
served as reservoirs. These reservoir cisterns are found in the Mediterranean re-
gion, and not in northern Europe, because of patterns of use and of precipitation.
They fall into two categories: terminal reservoirs, into which the entire aqueduct
discharges, and which therefore govern the entire distribution system downstream
from them; and reservoirs on branches or subdivisions of the urban distribution
network, which regulate supply only to their particular branch. These latter are
frequently connected with large bath complexes.
Terminal reservoir cisterns are found in North Africa and very occasionally in
other arid regions (Aptera, Crete; Sepphoris, Israel; Resafa, Syria), where seasonal
fluctuations in aqueduct delivery affected the whole distribution network. They
usually consisted either of multiple parallel, barrel-vaulted chambers communi-
cating through arches pierced in the division walls so that they formed a single
hydraulic space, or of a large cross-vaulted chamber with the arches supported on
piers (figure 11.7). At the outlet, one or more lead pipes passed from the main
reservoir through the wall into a chamber where the outflow could be controlled by
means of stopcocks on the pipes (e.g., Hippo Regius; Rusicade; A. I. Wilson 1998,
2001). Terminal reservoir cisterns in North Africa could have a capacity of up to
50,000 m 3 ( Carthage, La Malga), although most fell in the range of 3,500-13,500 m 3
(A. I. Wilson 1997: 79-80). The large reservoir cistern at the end of the Serino
aqueduct, at Bacoli on the Bay of Naples, held some 12,600 m 3 and probably sup-
plied the fleet stationed at Misenum (Hodge 1992: 276, 279). Because ofrainfall on
the karst aquifer from which they drew their water, the delivery of the Byzantine
aqueducts of Constantinople was subject to such considerable seasonal fluctuations
that several large reservoirs, so vast they could not be roofed over, were constructed
within the Theodosian walls to provide a seasonal reserve capable of supplying
much of the city; the reservoirs were named for Aetius (A.D. 421; 270,000 m 3 ),
Aspar (pre-471; 230,000 m 3 ), and St. Modus (A.D. 419-518; 250,000 m 3 ). In the
reign ofJustinian, these were supplemented by several very large, covered reservoir
cisterns, closer to the imperial palace, with brick roofs supported on numerous
columns reused from classical buildings. The most famous is the Yerebatan Saray
(A.D. 532; 83,500 m 3 ) and the Binbirdirek cisterns (57,800 m 3 ) (Procopius, Aed.
I.11.10-15; Freely and yakmak 2004: 55-56, 146-51).
Large reservoir cisterns specifically associated with large bath complexes are
found throughout a wider area, and illustrate a solution to supplying these water-
thirsty complexes while minimizing the impact on the rest of the distribution
network. If these cisterns were filled overnight, the baths could run on the stored
reserve during the following day, meaning that the full delivery of the aqueduct was
available for distribution to fountains and private users during the day. Reservoir
cisterns for baths are particularly common in North Africa and Italy, where they
are associated with baths of the imperial thermae type: the Sette Sale, for Trajan's
Baths (7,000 m 3 ); the cisterns of the Baths of Caracalla (11,500 m 3 ); the Botte di
Termini, serving the Baths of Diocletian in Rome; the Bordj Djedid cisterns for the
Antonine Baths at Carthage (20,000 m 3 ). Large supply cisterns are found elsewhere
in Italy, at Chieti, and in Crete at Aptera. Aqueduct-fed cisterns on a smaller scale,
100-400 m 3 capacity, are commonly found associated with medium-sized baths
(e.g., the Baths ofJulia Memmia at Bulla Regia, Tunisia).
Figure 11.7. One of seven parallel vaulted chambers in the large reservoir cisterns at
Oudhna (ancient Uthina) in Tunisia. (Photograph by A. I. Wilson.)
FOUNTAINS AND NYMPHAEA
The most important destination within the urban water network was public foun-
tains. For many people, perhaps the majority of urban dwellers, these represented
their main source of drinking water. We have seen that fountains were the major
destination of the Greek and Hellenistic aqueducts, and of the earliest aqueducts at
Rome. Many of the street fountains at Pompeii probably belong to the early phase
of that system (ca. 80 B.C. ); simple and functional, they consist of a stone basin with
a spout, often emerging from a decorative plaque or head. Over the course of the
first century A.O., however, first in Rome and then in the provinces, fountains grew
more elaborate and decorative, and they were increasingly ornamented with stat-
ues, sculpted motifs, and colonnades. Nero's colossal nymphaeum along the east
side of the base of the Temple of Divus Claudius, whose overflow fed the lake of the
Dom us Aurea on the site of the later Colosseum, was a precocious example of the
trend. By the early second century, nymphaea with elaborate facades were appearing
in provincial cities, initially in coastal Asia Minor and later in North Africa. These
often incorporated basins in front of a curved, colonnaded backdrop with statues of
deities, the imperial family, and local dignitaries who had paid or helped to pay for
the aqueduct. There was every conceivable gradation along a range from functional
to ostentatious, ranging from the simple street fountains of Pompeii, Herculaneum,
and Saepinum to the elaborate nymphaea of Lepcis, Perge, or Olympia. But we
should not allow the lavish architecture of the largest nymphaea to blind us to their
utility. An inscription from Cirta (Constantine, Algeria) makes the point. Part of
an inventory of the city's wealth, it lists in the nymphaeum 40 gilded letters and
10 punctuation points (doubtless from the inscriptions honoring the emperor
or private person who paid for the aqueduct), six chained goblets with gold inlay, a
fountain basin inlaid with gold, six bronze statues and a Cupid, six marble statues,
six bronze Silenus-head water spouts, and six hand towels ( CIL 8.6982 = !LS 4921b).
This account reflects a level of wealth and display (the gold inlay and other metal
ornaments) lacking in the surviving archaeology. While the lavish use of colored
marbles is attested by surviving remains, precious metals have invariably been
looted. The inscription also hints at the ways in which the fountain was used on a
daily basis: people went to drink from the chained goblets, splashed their faces and
hands from the basin, and then dried themselves on the towels, doubtless handed
out and retrieved by a public slave, in whose absence they would rapidly have dis-
appeared. Doubtless, the fountain was also a place of gossip and social exchange. At
Pompeii, the importance of the unostentatious public fountains to the city's life is
indicated by the fact that most houses are within 50 m of a street fountain. By the
late first century A.D., Rome had 39 display fountains (munera) and 591 street
fountains (lacus) ( frontinus, Aq. 78 ), and the fourth-century Regionary Catalogues
list between 1,216 and 1,352 fountains (lacus) and 15 nymphaea. (Jordan 1871-1907:
539-74).
BATHS
From the late Republic onward, public baths were among the most demanding
consumers of water on an urban aqueduct network. Greek public baths had con-
sisted of many individual hip-baths in the same room, in which bathers would have
water that had been drawn from a well poured over them by attendants. Roman
baths differed in that the bathers shared large pools, and the water could be heated
in large lead boilers mounted over furnaces, whose exhaust gases also heated rooms
and pools by means of the hypocaust. In this arrangement, pools and floors were
supported on a series of (usually) brick piers, allowing hot air from the furnace to
circulate beneath before escaping up flues in the walls. The temperature of the water
in hot or warm pools could be maintained by a device called a testudo, a bronze half-
cylinder open at one end, which communicated with the pool near its base, so that
it was full of water. The other end of the testudo was mounted over the furnace,
heating the water within it; the heated water escaped upward into the pool, being
replaced by cooler water so that circulation was achieved by convection. To the
frigidaria, tepidaria, and caldaria (cold, warm, and hot rooms) of early baths, later
baths sometimes added sweating rooms and indeed suites of exercise areas, lecture
halls, and libraries to form veritable leisure centers. While smaller public baths were
sometimes supplied from wells, usually with the aid of water-lifting machinery
(chapter 13), the larger ones had to be fed by aqueducts, and the largest elaborately
decorated imperial thermae of Rome and other major cities could accommo-
date hundreds or even thousands of bathers, and demanded special arrangements,
such as large reservoir cisterns to minimize their impact on the rest of the network
(Delaine 1988; Nielsen 1990; Yegiil 1992; Fagan 1999). The cutting of Rome's aq-
ueducts by the Goths in the siege of A.D. 537 put the city's baths out of action, and
although four of the aqueducts were subsequently repaired, public bathing never
regained its former importance in the early Middle Ages. Elsewhere, public bathing
also declined as the aqueducts fell out of use.
THE PURPOSE OF ROMAN AQUEDUCTS
Were aqueducts necessary to Roman cities, or were they a luxury? Early aqueduct
studies assumed that these elaborate systems were primarily useful, and a marker
by which Rome's civilization could be judged. From the 1970s onward, the tide of
scholarly opinion began turning in favor of a view that saw them as useless lux-
uries, built primarily to advertise Rome's control over natural resources, to use
water as an ostentatious symbol of conspicuous consumption, and to support an
urban lifestyle in which public bathing played a major role-in other words,
aqueducts were ideological symbols of empire (Leveau and Paillet 1976: 167, 181;
1983; Leveau 1987; Shaw 1984, 1991; Corbier 1991: 222). But today, with more at-
tention paid to the role of public fountains in the overall pattern of water supply
and to the regulation of delivery using terminal reservoirs and bath-related res-
ervoir cisterns, this binary opposition between utility and luxury appears too stark
(A. I. Wilson 1997, 1998: 89-93; 1999, 2001). The priorities of early aqueduct systems
were not baths and display. Even as civic and social ostentation became important
(with private display developing before civic display), the very utility of aqueducts
increased their ideological impact. Much more detailed research is needed on the
chronology of the spread of aqueducts throughout the Roman world, and on the
chronology of the spread and development of cities in the western provinces of
the empire, but at least for Rome, Pompeii, and North Africa it appears that urban
growth and the building of aqueducts went hand in hand. Cause and effect cannot
be completely disentangled, but we have no reason to reject the idea that aqueducts
enabled cities to grow to larger sizes and support more people than would have
been possible with reliance solely on water from wells, rainwater cisterns, and local
springs.
IRRIGATION SYSTEMS
Although the major riverine civilizations of the ancient Near East had relied on
complex state-controlled systems of irrigation from major rivers, most irrigation in
the Greco-Roman world took other forms (Hodge 1992: 246-53; Oleson 2000). The
exception was Egypt, where the Nile flood was the main source of water for agri-
culture, and irrigation canals derived from the river. Shadufs (tip-beams with buckets
lifted by a counterweight) were used to raise water out of the river or its canals, and
they were joined in the Ptolemaic period by other, more effective forms of water-
lifting machine (see also chapter 13 on water-lifting devices).
One of the most common forms of passive irrigation was agricultural terrac-
ing, which inhibited erosion and retained water in the soil. In arid zones such as
the Negev desert or the Tripolitanian pre-desert, stone and earth walls built on the
hamada or rock-desert plateaus concentrated on a small area runoff water from a
catchment zone many times the size, thus increasing the effective rainfall equivalent
on the cultivated plot. Small check dams built across the wadi floors retained water
and soil from flash floods, creating fertile parcels of soil in the wadi beds that could
be farmed; excess water was directed into cisterns and stored (Barker 1996; Glueck
1959; Evenari et al. 1982; Avner 2001-2002). Such systems supported a considerable
expansion of settlement into these pre-desert or desert areas in the Roman and late
Roman periods. On the more arid fringes of the Greco-Roman world-in the oases
of the Egyptian western desert, on the southern flank of the Aures range in Nu-
midia, and in the Garamantian lands of the Sahara-qanats or foggaras provided
water for both settlements and irrigation.
Much irrigation, everywhere, was small-scale and has left little archaeological
record. The exceptions we can trace are the large-scale schemes, often representing
massive investments of time, labor, and money. In the Roman period, large irri-
gation systems shared much with aqueduct technology. The main differences were
conditioned by the requirement for larger quantities of water than for urban sup-
ply, but with less concern for quality; hence there was a greater propensity to take
irrigation water by diversion from rivers, as in the Ebro valley (Spain) or the Safsaf
plain (Algeria). Diversion dams are found both on permanent water courses and,
in northern Tunisia, in wadis which run for only part of the year, or in spate
after occasional rainfall (du Coudray de la Blanchere 1895). Other, larger dams, as
at Kasserine in Tunisia, may have retained water for irrigation purposes; large
concrete dams in Tripolitania may have served a similar purpose, or, as at Lepcis
Magna, protected sites downstream from the erosive force of spate floods.
Irrigation channels frequently were not covered over, and they were often wider
than urban aqueducts-the Alcanadre channel in the Ebro valley is 2.5 m wide with
a water depth of1.1 m (Dupre 1997: 726-29). Indeed, the Ebro valley was home to a
number oflarge irrigation schemes; the first-century B.C. Tabula Contrebiensis is a
bronze table recording the settlement of a dispute between two communities about
building an irrigation canal over land purchased from a third (Richardson 1983;
Birks et al. 1984), and a bronze inscription from Ag6n, northwest of Zaragoza,
discovered in 1993, is a Hadrianic law regulating the use of a large irrigation system
(Beltran Lloris 2006). The Ag6n system drew water from the river Ebro and was
evidently a large-scale irrigation system involving several communities (of which
two, Ag6n and Gallur, were between 5 and 10 km apart) that contributed to its
upkeep. The central organization and planning of this system, with legal regulations
and a local court, anticipates the much better known hydraulic communities of
Islamic Spain. While this inscription describes the general legal framework for the
use and upkeep of the system, another (stone) inscription from Lamasba (Algeria)
records the specific irrigation schedule for a local community, in the reign of
Elagabalus, with details of which estates were to receive water between which hours
(CIL 8.18584; De Pachtere 1908; Shaw 1982; Meuret 1996).
One of the underground aqueducts of Carthage, from La Soukra in the Ariana
plain, was equipped with large wells or shafts (3 m x 1 m) at intervals along the
channel. These shafts are much larger than normal access shafts on Roman aque-
ducts, and the most likely explanation is that water-lifting machinery was installed
over the shafts to raise water from the aqueduct to irrigate the intensively cultivated,
centuriated land northwest of Carthage. The size, shape, and chronology of the
shafts suggest the use of bucket-chains, coupled with the animal-driven angle-gear
(saqiya; chapter 13). In the early twentieth century, French colonists mounted wind-
pumps over these shafts and used them once again for irrigation (Renault 1912: 471-
75; Fornacciari 1928-1929). The archaeological evidence, while not closely dateable
within the Roman period, confirms Tertullian's evidence for the pre-Islamic use of
the saqiya in North Africa (de Anima 33.7; mules and donkeys driving aquilegas
rotas).
While some systems were constructed specifically for irrigation, urban aque-
ducts were also sometimes tapped for the irrigation of estates along their course.
The evidence for this practice is clearest for the network of aqueducts serving Rome
(frontinus, Aq. 76, 92, 97; A. I. Wilson 1999). Two inscriptions from Rome or its
suburbs record the scheduling of rights to draw water to particular estates (CIL
6.1261; 14.3676), but the practice was more widespread. The aqueduct serving
Amiternum seems to have provided off-take castella for several estates through
which it ran, and scattered evidence indicates the use of aqueduct water by rural
villas and estates in other provinces (A. I. Wilson 1999). Some villa estates in Italy,
Spain, and North Africa constructed their own small-scale aqueducts to supply
water for drinking, bathing, and the irrigation of vegetable plots-the latter often
regulated from aqueduct-fed cisterns of 100-1,000 m 3 capacity (Thomas and

Wilson 1994; A. I. Wilson forthcoming).
SEWERS AND DRAINAGE
The emergent city-states of archaic Greece made use of much the same range of
drainage elements as the urban civilizations of the ancient Near East-soak.away
drains for foul water, and terracotta pipes for overflow water from fountains. Until
the classical period, street drains were rare in Greek cities. Athens' "Great Drain"
began as a storm-drain for the agora in the fifth century B.C., but even in the
fourth century B.C. much domestic waste water was discharged straight into the
streets (Aristotle, Ath.Pol. 50.1). Although many cities (e.g., Smyrna, Euesperides)
did without street drains well into the Hellenistic period, the classical era saw the
adoption at some sites of integrated drainage networks, in which domestic drains
fed into street drains that united in main collector drains under the principal
streets, eventually discharging outside the city. In the Roman world, the foundation
of numerous new colonies often allowed a more systematic adoption of planned,
integrated drainage networks, with under-street drains often built of masonry.
Rainwater entered the drains through gratings, either functional or with holes in a
decorative petal design. Overflow from aqueduct-fed fountains helped flush the
sewers, but the gradient was usually insufficient to prevent accumulation of solid
wastes, and street drains periodically had to be cleared manually by public slaves or
contractors. The need for access and maintenance determined a relationship be-
tween the size of drains and the manner in which they were roofed. Drains large
enough to walk through might be vaulted, while smaller ones were covered with flat
slabs that could be removed. Yet despite the relative sophistication of Roman
hydraulic engineering, some cities still continued to have open drains until the sec-
ond century A.D. Ancient drainage systems are related to increasing urban com-
plexity, but drainage follows urban development with a considerable time lag, due
to the need for strong political control, a developed legal framework of property
law, and access rights prior to implementation of integrated drainage networks
(A. I. Wilson 2000b).
Large land-drainage schemes in the Mediterranean region predate the archaic
period; the Bronze Age drainage of the Kopais basin in Boeotia channeled surface
waters to natural swallow-holes in the limestone. By the reign of Alexander the Great
(336-323 B.C.) this system had failed, and an engineer, Crates, was engaged to drain
the area again, a project left unfinished because of strife among the Boeotians
(Strabo 9.2.18). His scheme has been identified with a partially completed outlet
tunnel, with 16 vertical shafts descending to stretches of an unfinished horizontal
gallery. The same kind of shaft-and-gallery tunneling technique, already familiar
from Assyrian aqueducts and the aqueduct of Samas, is implied in a contract be-
tween the city of Eretria and the engineer Chaerephanes for a scheme to drain the
marshes of Ptechae (JG 9.9.191.1; Argoud 1987). Chaerephanes was to drain the marsh
by means of surface channels discharging into a collecting reservoir; this could then
be drained to the sea by means of an outlet tunnel, controlled by a sluice gate that
enabled the water in the reservoir to be used for irrigation in the spring. The contract
stipulates what land Chaerephanes can and cannot use; he must purchase the land
necessary for the access shafts to the outlet tunnel.
The same kind of shaft-and-gallery tunneling technique was used for the
Etruscan cuniculi, underground drainage galleries that either diverted one stream
under a ridge into an adjacent valley, or drained the marshy bottoms of valleys in the
easily-eroded tufa landscape of South Etruria, where natural surface drainage was
poor (chapter 12; cf. Judson and Kahane 1963; Bergamini 1991). The large Roman
lake drainage schemes, which relied on tunneling an outlet channel (emissarium)
through a ridge, are dealt with in chapter 12.
Extensive Roman land drainage and wetland reclamation schemes have been
identified in numerous areas of Roman centuriation (in the Ager Falernus, the
Campania, and the Po Valley), and also in Britain, with wetland reclamation in the
Severn estuary and the partial drainage of the East Anglian fens. Here, a large drain,
the Car Dyke, acted as a collector for smaller land drains from the high ground to
the west, and discharged the water through a series of canalized natural streams into
a roughly parallel system, the Midfendic, several miles further east. The Midfendic,
protected from the sea by a sea bank, stored the water until it could be released into
the sea at low tide. The backflow of water into the Car Dyke at high tide could be
prevented by sluice-gates on the cross-streams. The whole system anticipates the
ringvart systems of the medieval and later Netherlands; it was probably connected
with the establishment of salt workings on imperial estates in the Fenland under
Hadrian (Simmons 1979).
Notable advances were made in hydraulic engineering during the last six centuries
B.C., involving not only new inventions-waterproof cement compounds, the in-
verted siphon, lead piping, taps, and means of conserving and distributing water-
but also their increasingly widespread uptake. In the sixth century B.C. major hy-
draulic engineering works were exceptional projects undertaken by a handful of
rulers with the necessary resources, but under the Roman Empire long-distance
water supply lines and public water distribution networks became very common, so
that many Roman cities had a water-supply system that was not equaled again until
the nineteenth century. Ancient concepts of hygiene and public health were defi-
cient, but the spread of organized water supply enabled a level of urban growth and
development that could hardly have been achieved or sustained without reliance on
long-distance aqueducts that tapped remote sources. The level of hydraulic de-
velopment achieved in the ancient world is one of the key reasons why the regions
making up the Roman Empire saw a greater degree of urban settlement at that
period than at any subsequent time before the eighteenth century.
REFERENCES
Aicher, P. J. 1995. Guide to the aqueducts of andent Rome. Wauconda: Bolchazy-Carducchi.

Alzinger, W. 1987. "Ephesos," in G. Garbrecht (ed.), Die Wasserversorgung antiker Stiidte,
vol. 2. Mainz: von Zabern, 180-84.
Amit, D. 2002. "New data for dating the high-level aqueduct and the Wadi el Biyar
aqueduct, and the Herodion aqueduct," in D. Amrit, J. Patrich, and Y. Hirschfeld
(eds.), The aqueducts of Israel, 253-66. Journal of Roman Archeology Suppl. 46.
Portsmouth, RI: JRA, 253-66.
Argoud, G. 1987. "Eau et agriculture en Grece," in P. Louis, F. Metral, and J. Metral (eds.),
L'homme et l'eau en Mediterranee et au Proche Orient. Travaux de la Maison de l'Orient
14. vol. 4, L'eau dans l'agriculture: seminaire de recherche 1982-1983 et journees des 22 et 23
octobre 1983. Lyon: Maison de l'Orient, Presses Universitaires de Lyon, 25-42.
Ashby, T. 1935. The aqueducts of ancient Rome, ed. I. A. Richmond Oxford: Clarendon
Press.
Avner, Uzi. 2001-2. "Ancient water management in the Southern Negev," Aram 13-14:
403-21.
Barker, G. W.W. (ed.) 1996. Farming the desert: The UNESCO Libyan Valleys Archae-
ological Survey. vol. 1. Paris, Tripoli, and London: UNESCO Publishing, Department
of Antiquities (Tripoli) and the Society for Libyan Studies.
Beltran Lloris, F. 2006. "An irrigation decree from Roman Spain: The Lex Rivi Hiberiensis,"
Journal of Roman Studies 96:147-97.
Belvedere, 0 (ed.) 1987. Aquae Corneliae Ductus: L'approwigionamento idrico di Termini
Imerese in eta romana. Palermo: Universita di Palermo.
Bergamini, M. (ed.) 1991. Gli Etruschi maestri di idraulica. Perugia: Electa Editori Umbri.
Biernacka-Lubanska, M. 1977. "A preliminary classification of Greek rainwater intakes,"
Archeologia (Warszawa) 28: 26-36.
Birks, P., A. Rodger, and J. S. Richardson. 1984. "Further aspects of the Tabula Con-
trebiensis," Journal of Roman Studies 74: 45-73-
Blair, I., and J. Hall. 2003. Working water: Roman technology in action. London: Museum of
London.
Bodon, G., I. Riera, and P. Zanovello (eds.) 1994. Utilitas necessaria: Sistemi idraulichi nell'
Italia romana. Milano: Progetto Quarta Dimensione.
Boyd, T. D. 1978. "The arch and the vault in Greek architecture," American Journal of
Archaeology. 82: 83-100.
Camp, J. M. 1982. "Drought and famine in the 4th century B.C," in Studies in Athenian
architecture, sculpture and topography presented to Homer A. Thompson. Hesperia
Suppl. 20. Princeton, NJ: American School of Classical Studies at Athens, 9--17.
Chanson, H. 2000. "Hydraulics of Roman Aqueducts: Steep chutes, cascades and drop-
shafts," American Journal of Archaeology 104: 47-72.
Connelly, J.B., and A. I. Wilson. 2002. "Hellenistic and Byzantine cisterns on Geronisos
Island," Report of the Department of Antiquities, Cyprus: 269-92.
Conophagos, C. E. 1980. Le Laurium antique, et la technique grecque de la production de
l'argent. Athenes: Ekdotike Hellados s.a.
Conophagos, C. E. 1982. "Concrete and special plaster waterproofing in ancient Laurion
(Greece)," in T. A. Wertime and S. F. Wertime (eds.), Early pyrotechnology: The
evolution of the first fire-using industries. Washington, D.C.: Smithsonian Institution
Press, 117-23.
Corbier, M. 1991. "City, territory and taxation," in J. Rich and A. Wallace-Hadrill (eds.),
City and country in the ancient world. Leicester-Nottingham Studies in Ancient So-
ciety. London and New York: Routledge, 211-39.
Dalley, S. 2001-2. "Water management in Assyria from the ninth to the seventh centuries
B.C," Aram 13-14: 443-60.
DeLaine, J. 1988. "Recent research on Roman baths," Journal ofRoman Archaeology 1: 11-32.
De Pachtere, F.-G. 1908. "Le reglement d'irrigation de Lamasba," Melanges de ['Ecole
franraise de Rome 28: 373-400.
Devoti, L. 1978. Cisterne del periodo romano nel Tuscolano. Frascati: Associazione Tuscolana
"Amici di Frascati."
Doring, M. 2002a. "Wasser fiir den Sinus Baianus: Romische Ingenieur- und Wasserbauten
der Phlegraeischen Felder," Antike Welt 33: 305-19.
Doring, M. 2002b. "Die romische Wasserversorgung von Pozzuoli, Baia und Miseno
(Italien)," in C. Ohlig, Y. Peleg, and T. Tsuk (eds.), Cura Aquarum in Israel. Siegburg:
Deutsches Wasserhistorisches Gesellschaft, 253-65.
Downey, G. 1961. A history ofAntioch in Syria from Seleucus to the Arab conquest. Princeton:
Princeton University Press.
du Coudray la Blanchere, Rene M. 1895. L'amenagement de l'eau et ['installation rurale dans
l'Afrique ancienne. Paris: Imprimerie Nationale.
Dupre, N. 1997. "Eau, ville et campagne dans l'Hispanie romaine: Apropos des aqueducs
du bassin de l'Ebre," in R. Bedon (ed.), Les aqueducs de la Gaule romaine et des regions
voisines. Caesarodunum 31. Limoges: Presses Universitaires de Limoges, 715-43.
Eschebach, H. 1983. "Die innerstadtische Gebrauchswasserversorgung dargestellt am
Beispiel Pompejis," in J.-P. Boucher (ed.), Journees d'etudes sur les aqueducs romains:
Tagung ilber romische Wasserversorgungsanlagen. Lyon 26-28 mai, 1977. Paris: Les
Belles Lettres, 81-132.
Evenari, M., L. Shanan, and N. H. Tadmor. 1982. The Negev: The challenge of a desert. 2nd
ed. Cambridge, MA: Harvard University Press.
Faber, G. 1992. "Der Romische Qanat von Helmsingen (le Qanat romain de Helmsange),"
Vie Souterraine 2: 13-15.
Fabio, E., and L. Fassitelli n.d. [1992?]. Roma: tubi e valvole. 13th ed. Milan: Petrolieri
d'Italia.
Fagan, G. G. 1999. Bathing in public in the Roman world. Ann Arbor: University of Michigan
Press.
Freely, J., and A. <;:akmak. 2004. Byzantine monuments of Istanbul. Cambridge: Cambridge
University Press.
Fornacciari, C. 1928-1929. "Note sur le drain romain de la Soukra [communication a la
seance de la Commission de l'Afrique du Nord, 10 decembre 1929]," Bulletin arche-
ologique du Comite des travaux historiques et scientifiques: 413-15.
Garfinkel, Y., A. Vered, and 0. Bar-Yosef 2006. "The Domestication of water: the Neolithic
well at Sha'ar Hagolan, Jordan Valley, Israel," Antiquity 309: 686-96.
Glueck, N. 1959. Rivers in the desert: The exploration of the Negev. An adventure in ar-
chaeology. London: Weidenfeld and Nicolson.
Goblot, H. 1979. Les qanats: Une technique d'acquisition de l'eau. Industrie et artisanat 9.
Paris and New York: Editions Mouton/Ecole des Hautes Etudes en Sciences Sociales.
Grewe, K. 1985. Planung und Trassierung romischer Wasserleitungen. Schriftenreihe der
Frontinus-Gesellschaft, Supplementband 1. Wiesbaden: Verlag Chmielorz.
Grewe, K. 1988. "Romische Wasserleitungen nordlich der Alpen," in G. Garbrecht (ed.),
Die Wasserversorgung antiker Stiidte, vol. 3. Mainz: von Zabern, 43-97.
Grewe, K. 1993. "Antike Welt der Technik V: Augusta Emerita/Merida-Eine Stadt ro-
mischer Technikgeschichte," Antike Welt 24: 244-55.
Grewe, K. 1998. "Antike Welt der Technik IX: Der Aquadukt von Siga (Algerien)," Antike
Welt. 29: 409-20.
Hellmann, M.-C. 2002. L'architecture grecque. vol. 1,: Les principes de la construction. Paris:
Picard.
Hill, B. H. 1964. Corinth vol. 1.6, The springs: Peirene, Sacred spring, Glauke. Princeton:
Hodge, A. T. (ed.) 1991. Future currents in aqueduct studies. Leeds: Francis Cairns
Hodge, A. T. 1992. Roman aqueducts and water supply. London: Duckworth.
Hodge, A. T. 2000. "Wells," in Wikander 2000: 29-33.
Jacobsen, T., and S. Lloyd 1935. Sennacherib's aqueduct at Jerwan. Chicago: University of
Chicago Press.
Jansen, G. C. M. 2000. "Urban water transport and distribution," in Wikander 2000:
103-25.
Jansen, G. C. M. 2002. Water in de Romeinse stad. Pompeji-Herculaneum-Ostia. Leuven:
Peeters.
Jarde, A. 1907. "Puteus," DarSag, vol. 4.1, 779-81.
Jones, R., and D. Robinson 2005. "Water, wealth and social status at Pompeii: The House
of the Vestals in the first century," American Journal of Archaeology 109: 695-710.
Jordan, H. 1871-1907. Die Topographie der Stadt Rom im Altertum. 2 vols. in 4. Berlin:
W eidmannsche buchhandlung.
Judson, S., and A. Kahane. 1963. "Underground drainageways in Southern Etruria and
Northern Latium," Papers of the British School at Rome 31: 74--99.
Keppie, L. 1983. Colonisation and veteran settlement in Italy, 47-14 B.c. Rome: British School
at Rome.
Kienast, H. J. 1995. Die Wasserleitung des Eupalinos auf Samos. Samos 19. Bonn: Deutsches
Archaologisches Institut.
Klaffenbach, G. 1954. Die Astynomeninschrift von Pergamon. Abhandlungen der deutschen
Akademie der Wissenschaften zu Berlin, Klasse fur Sprachen, Literatur und Kunst,
Jahrgang 1953, 6. Berlin: Akademie-Verlag.
Kohl, N., and G. Faber 1990. 25 Jahre Raschpetzer Forschung, 1965-1990. Walferdange.
Kohl, N., G. Waringo, and G. Faber 1995. Raschpetzer: Die Ausgrabungschronik der Jahre
1991-1995. Walferdange.
Kremer, B. 2005. "Antike Wassergewinnung an der Mosel. Der romische Qanat von Po-
lich," Schriftenreihe der Frontinus-Gesellschaft 26: 127-55.
Lanciani, R. 1881. "Topografia di Roma antica: I comentarii di Frontino intorno le
acque e gli acquedotti. Silloge epigrafica aquaria," Memorie della Reale Accademia dei
Lincei, Serie 3, no. 4: 215-616.
Lang, M. 1968. Waterworks in the Athenian agora. Excavations of the Athenian Agora,
Picture Books 11. Princeton, New Jersey: American School of Classical Studies at
Athens.
a
Lassus, J. 1983. ''L'eau courante Antioch," in J.-P. Boucher (ed.), Journees d'etudes sur les
aqueducs romains: Tagung uber romische Wasserversorgungsanlagen, Lyons, 26-28
mai, 1977. Paris: Les Belles Lettres, 207-29.
Last, J. S. 1975. "Kourion: The ancient water supply," Proceedings of the American Philo-
sophical Society 119: 39-72.
Leveau, P. 1984. Caesarea de Mauretanie: Une ville romaine et ses campagnes. Collection de
l'Ecole Franyaise de Rome 70. Rome: Ecole Franyaise de Rome.
Leveau, P. 1987. "A quoi servaient les aqueducts romains?" L'Histoire 105 (Nov.): 96-104.
Leveau, P., and J.-L. Paillet 1976. L'alimentation en eau de Caesarea de Mauretanie et
l'aqueduc de Cherchell. Paris: Editions l'Harmattan.
Leveau, P., and J.-L. Paillet 1983. "Alimentation en eau et developpement urbain aCaesarea
de Mauretanie," in J.-P. Boucher (ed.), Journees d'etudes sur les aqueducs romains:
Tagung ilber romische Wasserversorgungsanlagen, Lyons, 26-28 mai, 1977. Paris: Les
Belles Lettres, 231-34.
Lewis, M. J. T. 1999. "Vitruvius and Greek aqueducts," Papers of the British School at Rome
67: 145-72.
Lloyd, J. A. 1977. "Water supply and storage systems," in J. A. Lloyd (ed.), Excavations at
Sidi Khrebish, Benghazi (Berenice). Libya Antiqua, suppl. 5. Tripoli: Department of
Antiquities, vol. 1, 199-209.
Magee, P. 2005. "The chronology and environmental background of Iron Age settlement in
southeastern Iran and tlie question of the origin of tlie qanat irrigation system,"
Iranica Antiqua 40: 217-31.
Manteli, K. 1992. "The Neolitliic well at Kastelli Phournis in eastern Crete," Annual of the
British School at Athens 87: 103-20, pls.1-2.
Marliere, E. 2002. L'outre et le tonneau dans ['Occident romain. Monographies In-
strumentum 22. Montagnac: Editions Monique Mergoil.
Meuret, C. 1996. "Le reglement de Lamasba: Des tables de conversion appliquees a!'irri-
gation," Antiquites africaines 32: 87-112.
Nielsen, I. 1990. Thermae et Balnea: The architecture and cultural history of Roman public
baths. 2 vols. Aarhus: Aarhus University Press.
Ohlig, C. P. J. 2001. De aquis Pompeiorum. Das Gastellum Aquae in Pompeji: Herkunft,
Zuleitung und Verteilung des Wassers. Circumvesuviana 4. Nijmegen: Books on De-
mand Norderstedt.
Oleson, J.P. 1991. "Aqueducts, cisterns, and tlie strategy of water supply at Nabataean and
Roman Auara (Jordan)," in Hodge 1991, 45-62.
Oleson, J.P. 1992. "Water works," in D. N. Freedman (ed.), Anchor Bible dictionary, vol. 6.
New York: Doubleday, 883-93.
Oleson, J.P. 1995. "The origins and design ofNabataean water-supply systems," in Studies
in the history and archaeology ofJordan, 5. Amman: Department of Antiquities, 707-19.
Oleson, J.P. 2000. "Irrigation," in Wikander 2000: 183-215.
Patrich, J., and D. Amit 2002. "The aqueducts of Israel: An introduction," in D. Amit,
J. Patrich, and Y. Hirschfeld (eds.), The aqueducts of Israel. Journal of Roman Ar-
chaeology Suppl. 46. Portsmoutli, RI: JRA, 9-20.
Peltenburg, E., P. Croft, A. Jackson, C. McCartney, and M.A. Murray 2001. "Well-
established colonists: Mylouthkia I and the Cypro-Pre-Pottery Neolitliic B," in S.
Swiny (ed.), The earliest prehistory of Cyprus: From colonization to exploitation. Cyprus
American Archaeological Research Institute Monograph Series 2. Boston: American
Schools of Oriental Research, 61-93.
Poratli, E. 2002. "The water-supply to Caesarea: A re-assessment," in D. Amit, J. Patrich,
and Y. Hirschfeld (eds.), The aqueducts of Israel. Journal of Roman Archaeology Suppl.
46. Portsmoutli, RI: JRA, 104-29.
Potter, T. W. 1987. Roman Italy. London: British Museum Press.
Renault, J. 1912. "Les bassins du trik Dar-Saniat a Cartliage," Revue Tunisienne 19.95: 471-
98.
Richardson, J. S. 1983. "The Tabula Contrebiensis, Roman law in Spain in the early first
century B.C," Journal of Roman Studies 73: 33-41.
Rodgers, R. H. 2004. Frontinus De aquaeductu urbis Romae, edited with introduction and
commentary. Cambridge: Cambridge University Press.
Schoellen, A. 1997. "Des surprenants ouvrages hydrauliques romains," Archeologia Paris
332: 62-66.
Sgobbo, I. 1938. "Serino-L'acquedotto romano della Campania: 'Fontis Augustei
Aquaeductus,'" Notizie degli scavi di antichita: 75-97.
Shaw, B. D. 1982. "Lamasba: An ancient irrigation community," Antiquites africaines 18:
61-103.
Shaw, B. D. 1984. "Water and society in the ancient Mahgrib: Technology, property and
development," Antiquites africaines 20: 121-73-
Shaw, B. D. 1991. "The noblest monuments and the smallest things: Wells, walls and
aqueducts in the making of Roman Africa," in Hodge 1991, 63-91.
Simmons, B. B. 1979. "The Lincolnshire Car Dyke: Navigation or drainage?" Britannia 10:
183-96.
Stenton, E. C., and J. J. Coulton. 1986. "Oinoanda: The water supply and aqueduct,"
Anatolian Studies 36: 15-59.
Thomas, R. G., and A. I. Wilson 1994. "Water supply for Roman farms in Latium and
South Etruria," Papers of the British School at Rome 62: 139-96.
Tolle-Kastenbein, R. 1994. Das archaische Wasserleitungsnetz fur Athen und seine spiiteren
Bauphasen. Mainz: Philipp von Zabern.
Tolle-Kastenbein, R. 1996. "Das archaische Wasserleitungsnetz fiir Athen," in N. de
Haan and G. C. M. Jansen (eds.), Cura Aquarum in Campania. BABESCH Suppl. 4.
Leiden: BABESCH, 129-36.
Walda, H. M. 1996. "Lepcis Magna excavations Autumn 1995: Report on surveying, ar-
chaeology and pottery," Libyan Studies 27: 125-27.
Wikander, 6. (ed.) 2000. Handbook of ancient water technology. Leiden: Brill.
Wilson, A. I. 1995. "Running water and social status in North Africa," in M. Horton and T.
Weidemann (eds.), North Africa from antiquity to Islam. Bristol: Centre for Medi-
terranean Studies (University of Bristol), 52-56.
Wilson, A. I. 1997. "Water management and usage in Roman North Africa: A social and
technological study." D. Phil. Thesis, University of Oxford.
Wilson, A. I. 1998. "Water supply in ancient Carthage," in Carthage papers: The early
colony's economy, water supply, a private bath, and the mobilization of state olive oil.
Journal of Roman Archaeology Suppl. 28. Portsmouth, RI: JRA, 65-102.
Wilson, A. I. 1999. "Deliveries extra urbem: Aqueducts and the countryside," Journal of
Wilson, A. I. 2000a "The aqueducts of Italy and Gaul,'' Journal of Roman Archaeology 13:
597-604.
Wilson, A. I. 2000b. "Drainage and sanitation," in Wikander 2000: 151-79.
Wilson, A. I. 2001. "Urban water storage, distribution and usage in Roman North Africa,"
in A. 0. Koloski-Ostrow (ed.), Water use and hydraulics in the Roman city. Boston:
Kendall Hunt, 83-96.
Wilson, A. I. 2004. "Classical water technology in the early Islamic world,'' in C. Bruun and
A. Saastamoinen (eds.), Technology, ideology, water: From Frontinus to the Renaissance
and beyond. Acta Instituti Romani Finlandiae 31:115-41.
Wilson, A. I. 2005. "Foggara irrigation, early state formation and Saharan trade: The
Garamantes of Fazzan," Schriftenreihe der Frontinus-Gesellschaft 26: 223-34.
Wilson, A. I. 2006a. "The spread of foggara-based irrigation in the ancient Sahara," in
D. J. Mattingly, S. McLaren, E. Savage, Y. Fasatwi, and K. Gadgood (eds.), Natural
resources and cultural heritage of the Libyan desert. London: Society for Libyan Studies,
205-16.
Wilson, A. I. 2006b. "Water for the Pompeians," Journal of Roman Archaeology 19: 501-8.
Wilson, A. I. forthcoming. "Villas, horticulture and irrigation infrastructure in the
Tiber Valley," in F. Coarelli and H. Patterson (eds.), The Tiber Valley in antiquity.
Rome: British School at Rome.
Wilson, R. J. A. 1996. "Tot aquarum tam multis necessariis molibus ... : Recent studies on
aqueducts and water supply," Journal of Roman Archaeology 9: 5-29.
Wilson, R. J. A. 2000. "Aqueducts and water supply in Greek and Roman Sicily: The
present status quaestionis," in G. C. M. Jansen (ed.), Cura Aquarum in Sicilia. BA-
BESCH Suppl. 6. Leuven: Peeters, 5-36.
Wuttmann, M. 2001. "Les qanats de 'Ayn-Manawir," in P. Briant (ed.), Irrigation et
drainage dans l'antiquite: Qanats et canalisations souterraines en Iran, en Egypte et en
Grece. Paris: Editions Thotm, 109-35.
Yegiil, F. 1992. Baths and bathing in classical antiquity. Cambridge, MA: MIT Press.
CHAPTER 12
TUNNELS AND CANALS
KLAUS GREWE
TUNNEL AND CANAL ENGINEERING
Even in the early period of human cultural development, it was impossible to build
streets and roads, or aqueducts and canals, without planning the route ahead.
Roads, aqueducts, and canals have in common that they extend across the land-
scape in a long linear course, and in the process they often have to negotiate
topographical obstacles and resort to structural solutions. Among the buildings
that are included in these structures, valley crossings are conspicuous and often
even spectacular. Bridges 50 m high, such as the road bridge across the Tajo near
Alcantara in Spain and the aqueduct bridge Pont du Gard near Nimes in France, are
testaments to the flourishing state of civil engineering in Roman times. If, instead of
a valley, a mountainous obstacle was in the way of the planned structure, a tunnel
was often the only possibility for providing passage. Since tunnels are naturally less
conspicuous than bridges, they have remained somewhat in the background of the
study of history of technology. The great technological achievement inherent in the
tunnels realized in classical antiquity, from the Eupalinus tunnel on Samas to
the tunnel Nonius Datus supervised in Saldae, has never been denied, but the
accomplishment deserves discussion. Grewe (1998) provides a comprehensive treat-
ment (table 12.1).
It would be ideal to be able to compare a surviving tunnel with its original plan,
allowing reproduction of the ancient thought processes during the planning stage
and the work procedures. The directional accuracy achieved below ground could be
easily evaluated if any ancient plans or diagrams had survived, but we are deprived
of this evidence. Contemporary descriptions of the tunnels give little information
about the amount of technological know-how that went into them. Rather, these
)n of important tunnels in the classical world.
Building Patron
Place Country Purpose Length Technique (Director, if known)
Jerusalem Israel Water supply 533 m Counter- King Hezekiah
excavation
Island of Greece Water supply 1,036 m Counter- Polycrates,
Samos excavation (Eupalincs)
Tivoli Italy River diver- ca. 60 m Counter- Etruscans
sion for road excavation
construction
Veii Italy River diver- ca. 70 m Counter- Etruscans
sion excavation
Ariccia Italy Lake drainage ca. 650 m Qanat Etruscans
Nemi Italy Lake drainage 1,600 m Counter- Etruscans
excavation
Castel Italy Lake drainage 1,400 m Counter- Etruscans/Romans
Gandolfo excavation
with shafts
Cumae Italy Road traffic ca. 1,000 m Counter- Agrippa
excavation (L.Cocceius Auctus)
with shafts
Naples/ Italy Road traffic 705 m Counter- (L.Cocceius Auctus ?)
Pozzuoli excavation
Naples/ Italy Road traffic 780 m Counter- (L. Cocceius Auctus ?)
Pozzuoli excavation
with side shafts
Sernhac / France Water supply 65 m Qanat Claudius
Nimes
$ Sernhac / France Water supply 6om Qanat Claudius
Nimes
Avezzano Italy Lake drainage 5,642 m Qanat Claudius (Narcissus)
Furlo-Pass Italy Road traffic 37 m Counter- Vespasian
excavation
<;.:evlik / Turkey River diver- 185 m Counter- Vespasian and
Antakya sion for flood (86 m and 31m) excavation with Titus
control rock-cut shafts
Petra Jordan River diver- 90 m Counter- (Nabateans)
sion for flood excavation
control
Montefur- Spain River diver- Counter- Romans
ado sion with excavation
gold trap
Briord France Water supply 197 m Qanat Romans
Bologna Italy Water supply ca. 20 km Qanat with vertical Romans
and side shafts
Chagnon/ France Water supply Som Counter- Romans
Lyon excavation
Saldae / Algeria Water supply 428 m Counter- City of Saldae
Bejaia excavation (Nonius Datus)
Walfer- Luxem- Water supply ca. 600 m Qanat Romans
dange burg
Duren Germany Water supply 1,660 m Qanat Romans
sources relate to the circumstances that led to the construction of a tunnel, and to
its patrons and their motivation for the commissioning and finally the financing of
the structure.
Even the inscription of Nonius Datus regarding a tunnel built in the second
century A.D. (see below), while giving an insight into the organization of an ancient
tunnel building site, leaves many open questions about the actual route planning.
Hence Mommsen (1871: 9), one of the first scholars to work on the Nonius Datus
inscription, hoped that "perhaps a competent engineer of our period will be able to
solve from the structure itself what is incomprehensible to us from the report of his
Roman predecessor." As early as the late nineteenth century, then, Mommsen
touched on one of the basic problems of research in the history of ancient engi-
neering. In the absence of authentic technical reports,, or design plans, the infor-
mation about the technology that lies within a structure can often be read only from
the structure itself-an interesting task, of course, but not always without peril in
terms of interpretation and conclusion.
EARLY TUNNELS AND THE QANAT
With the emergence of ironworking around 1000 B.C., humans had at their disposal
a new material for the manufacturing of tools. As a result it was possible to revo-
lutionize building construction, since now there were new possibilities for working
stone. It is generally assumed that the small tunnels of the Israelite royal cities owe
their appearance to the introduction of iron tools. In the cities that are sited on tels,
access was provided to springs gushing from the outer slope by means of stairs and
underground walkways built from the occupation level to the original level of the
spring (Grewe 1998: 41-44).
The history of tunnel construction in the circum-Mediterranean region begins
with the development of a special type of water catchment from underground
resources, the qanat, an early masterful engineering accomplishment that exerts its
influence even in our time as a particular method of tunnel construction (chapter
11). In order to supply oases in arid and semiarid zones with water, underground
aqueducts were built into water-bearing strata, which were usually located at the
foot of the slopes of distant mountains. After the presence of water was ascertained
by means of a trial shaft, known as the mother shaft, the planned line to the oasis
was staked out. A chain of shafts was dug in close proximity, and underground
tunnels were excavated between them, finally forming the qanat. Surviving ex-
amples conducted water over distances of up to 70 km. Al-Karaji (1973; cf. Grewe
1998: 33-40), an Arabian mathematician from the eleventh century A.D., wrote a
handbook on qanat building that presents this ancient technique with great clarity.
He explains the tools, the building methods, and also the method ofline layout. His
TUNNELS AND CANALS 323
description gives a complete picture of the process and is therefore a unique re-
source for the history of this technology, most likely reflecting his ancient sources.
In the qanat technique of tunnel construction (also called the manhole technique
because of the regular pattern of vertical shafts), the entire line is subdivided into
many individual units to keep the distances to be excavated underground short and
the consequent directional error small. For that reason it is necessary to sink a
number of shafts. The planned route of the tunnel is laid across any intervening
elevations in a line that allows the shafts to be as shallow as possible; therefore they
are found mostly in saddles, shifted to the side by a few meters to prevent the shafts
from flooding if it rained during construction. In most cases, the line surveyed
across the mountain curved to follow the topography. Next the shafts were staked
out and sunk. After reaching a precisely calculated depth below ground, a straight
connection was driven between the central shaft and the two adjoining shafts. The
excavated spoil was transported to the surface in leather bags or baskets lifted by
winches. Seen as a plan, the line of the tunnel consists of many short, straight
sections. This qanat technique was later applied in numerous situations by the
Etruscans and the Romans.
But by the sixth century B.C., a second method had appeared, the counter-
excavation technique. In this procedure the tunnel is driven from two sides in an
attempt to meet in the middle of the mountain to be perforated. This method
required the builder to have a much greater knowledge of the basics of surveying
and geometry. The projected meeting point could be reached only by constantly
controlling and correcting the direction of advance. As a result, thorough planning
and layout work were necessary for such a tunneling project, and the excavation
generally was carried out towards a central point, in two separate sections on the
same alignment, or occasionally side-by-side and parallel. The line staked out
across the mountain on the surface was transferred underground to provide the
proper orientation. Once the excavation had made some progress, the builder could
get his bearings by looking back at the light that penetrated through the tunnel
mouth in order to verify if his course was straight. In counter-excavation, devia-
tions from the planned course could be discovered and subsequently corrected only
through continuous control measurements. Geological problems might necessitate
a departure from the projected line; necessary deviations within the mountain
would require additional control measurements. On top of all these difficulties
related to surveying, tunnel construction by counter-excavation over long stretches
was rendered even more difficult by problems of water drainage and ventilation.
The decision by an ancient builder whether to construct a tunnel by the qanat
method or by counter-excavation depended, on the one hand, on his experience
with these respective techniques; on the other hand, the height of the mountain to
be tunneled would have been important, as the effort required to sink the shafts was
dependent on the extent of overburden above the tunnel. Another important factor
was that application of the qanat technique could shorten construction time sig-
nificantly, since the mountain could be excavated not only from the tunnel mouths
but also from shafts. The average rate of advance through solid rock, derived from
construction marks on a tunnel in Bologna, was 30 cm per day (Giorgetti 1988: 180-
85; Grewe 1998: 139-41), although of course the rate would depend on local geo-
logical conditions. Suetonius ( Claud. 20) gives a construction time of 11 years for
the almost 6 km long tunnel Claudius built to drain the Fucine Lake, with shafts up
to 120 m deep. These dimensions set it apart from all other ancient tunnels, but
given the use of the qanat method with multiple work faces, the statistics from
Bologna are within the realm of possibility.
SPECIAL ACHIEVEMENTS IN ANCIENT

TUNNEL CONSTRUCTION
Both the Hezekiah Tunnel in Jerusalem (705-701 B.C.) and the Eupalinus Tunnel
on Samos (end of the sixth century B.C.) were built by counter-excavation. The
plans of the Hezekiah tunnel as well as the Eupalinus tunnel show the difficulties in
transferring directions from the surface to the excavation below ground. Com-
pletion of the Hezekiah tunnel in Jerusalem in 701 B.C. was possible only after many
corrections in the advancing counter shafts. This event caused great rejoicing, since
not only did it assure the water supply of Jerusalem at a time of threat from the
Assyrians, but it also prevented the enemy's access to the water. The Old Testament
enthusiastically acknowledges Hezekiah's achievement (2 Chron. 31), and an in-
scription describes the tunnel breakthrough (Grewe 1998: 45-52). The hypothesis
that the Hezekiah tunnel was not a planned structure, but that the builders were
simply following crevices in the mountain during construction, cannot be sustained
after examination of the tunnel plan. Several corrections in direction that are
evident can only be the results of controlled measurements, which presuppose that
the tunnel was built according to a plan.
The Eupalinus tunnel on the Greek island of Samos can be considered the first
large-scale tunnel that was planned as an engineering structure in the modern sense,
since a strategy of planning and layout is evident in the execution (Grewe 1998:
58-69). The tunnel, which was built by Eupalinus during the reign of Polycrates
(535-522 B.C.), was planned as a counter-excavation tunnel with a meeting point in
the middle. (figures 12.1-2) The hypothesis that the meeting point of the two
sections was planned by Eupalinus to occur not in the middle of the line but rather
underneath the ridgeline of the mountain it perforates cannot be sustained (Kienast
1995). Eupalinus succeeded in making the two sections meet despite an undetected
error in direction in one of the two galleries, and despite changes in plan that were
possibly necessary because of geological problems. For that purpose he used a grid
that he laid over his plan, which allowed him to control the tunnels in real time. It
was possible for him to calculate a deviation within the mountain through a change
in the angle of advance in the order of magnitude of tangent 1:3. This can be
Figure 12.1. Tunnel of Eupalinus, Samos. Access tunnel and trench for pipeline.
(Photograph by K. Grewe.)
ascertained from the existing line of the tunnel. Eupalinus' strategy for the final
stage of advance before the breakthrough was ingenious. He halted excavation in
one section and determined through measurements the exact position of its end
point. Then, to ensure success, he directed that a curve be excavated in the opposing
tunnel, a bend that would eliminate all detected and undetected errors. ( figure 12.2)
The tunnel proper served for excavation and later access to the terracotta pipeline,
which was laid in a narrow, more easily leveled trench beside the walkway. ( figure
12.1) The achievement of Eupalinus is unique in antiquity, and not without reason
did Herodotus (3.60) count the tunnel among the greatest structures in the entire
Greek world. In the case of these first two historically large-scale tunnels, we know
King Hezekiah to be the patron in one instance, and Eupalinus to be the builder in
the other. But we still know nothing about the methods of planning and laying out
the structures. It is a challenge for a modern engineer to work out these methods,
since in neither case do we have plans or survey sketches transmitted from antiquity.
Historians of technology have paid little attention to the fact that almost simul-
taneously with the great achievements in Jerusalem and Samos, significant tunnels
were being built in central Italy by the Etruscans, the cultural and technological
teachers of the Roman Republic. Countless cuniculi-drainage and water-supply
tunnels designed like qanats-were built in the area northeast of Rome (Grewe 1998:
70-77). Large tunnels in the Alban Mountains, all serving to drain lakes, are either
Etruscan structures or, if they were built under the Romans, at least structures
influenced by Etruscan technology. The tunnels at Ariccia, at Lago Albano, and at
Figure 12.2. Tunnel of Eupalinus, Samos. Plan of breakthrough point, showing error
correction gallery. (Drawing by K. Grewe.)
Lago di Nemi must also be counted among the large-scale tunnels. The last two were
built by counter-excavation. Two Etruscan structures at Veii and Ponte Terra, built
to divert rivers, could be termed large-capacity tunnels. The achievements in Roman
tunnel construction cannot be evaluated without including the prior achievements
and models of the Etruscans in the study. The technology of tunnel construction
spread with the Romans over those parts of Europe, North Africa, and Asia Minor
that they controlled (Grewe 1998: 78-192). Tunnels were cut for roads, for river
diversion, for lake drainage, and even for gold extraction. Most often, however,
tunnels were built in connection with aqueducts. The preferred method was the
qanat technique, since short allotments at the construction site reduce the risk of two
excavations missing each another.
One tunnel built by counter-excavation almost resulted in the failure of the
entire aqueduct project, because the opposing galleries missed each other under-
ground. As we will see, an inscription left behind by Nonius Datus, the engineer in
charge of the project at Saldae in North Africa, tells us a great deal about the
planning and organization of an ancient large-scale construction site. The status of
a Roman engineer is clarified as well, since Nonius Datus was a legionary officer,
and was assigned to the construction of the aqueduct. The workmen, too, were
detailed from their military units to build this civilian structure. Nonius Datus even
quotes the correspondence that was exchanged during his employment as the
engineer in charge of the project. This commission must have been the greatest
challenge in the professional life of this engineer, but presumably that was not the
only reason he set up his report in a public inscription; it is very likely that he was
responding to attempts to blame him for the mistakes made during construction.
Tunnel construction was a type of engineering accomplishment very well suited
for demonstrating both the benevolence and the sheer power of a ruler. Therefore it
is not surprising that an inscription seems to have been placed at every substantial
tunnel, as at the river diversion at <;::evlik near Antakya in Turkey, where the names
of the patrons, in this case the emperors Vespasian and Titus, appear (figure 12.3).
This was a substantial diversion tunnel designed to carry floodwater around the
basin of the harbor of Seleucia Pieria, near Antioch on the Orontes. It is conse-
Figure 12.3. Tunnel of Titus, near <;:evlik (Antakya, Turkey). (Photograph by K. Grewe.)
quently understandable that the greatest tunnel executed in the ancient world-the
tunnel cut through Monte Salviano by the emperor Claudius to drain the Fucine
Lake, 80 km east of Rome-attracted great attention even while it was still in the
construction stage. Pliny the Elder (HN 36.124), who seems to have been an eye-
witness, describes the construction in significant detail and terms it one of "the
most memorable achievements" of that active emperor: "The expense was in-
describable and the workers numberless over so many years because, where the
mountain was earthy in character, the spoil from the channel had to be cleared out
through vertical shafts by means of winches, and elsewhere the bedrock had to be
cut away. All of this great effort ... took place in darkness." Even contemporary
court gossip is recorded, as several ancient authors suggest that Claudius' wife,
Agrippina, was scheming against the palace bureaucrat Narcissus, who was in charge
of the financial administration of the project (Tacitus, Ann. 12.57; Cassius Dio 60.11;
Suetonius, Claud. 20). Inaugurated in A.D. 52, the tunnel allegedly took 30,000
workmen 11 years to complete. With a length of 5,595 m, and shaft depths of up to
122 m, it was supposed to drain the bed of the Fucine Lake for agriculture and keep
the water level constant. Despite the careful planning and execution, the tunnel was
only partially successful, and proper drainage of the lake was managed only in the
nineteenth century.
Figure 12.4 . Drover Berg Tunnel near Diiren, Germany. Covered water channel and clay
fill. (Photograph by K. Grewe.)
In the German Rhineland and neighboring Luxemburg, a large number of

aqueduct tunnels were built, presumably from the second century A.O. on (Grewe
1998: 176-92; Frontinus-Gesellschaft 2003). Here only the qanat technique was
applied. It is striking that exclusively small settlements, generally affluent villae
rusticae, were supplied in this manner, rather than the big cities. A particularly
impressive example is the Roman Raschpetzer-Tunnel in Walferdange (Luxem-
burg), since more extensive archaeological investigation has revealed several of the
ancient shafts.
The longest Roman tunnel north of the Alps was driven under the Drover Berg
near Duren (Germany) . Surprisingly, this structure, with a length of 1,660 m, seems
to have supplied only one single villa rustica with water ( figure 12.4) . The water
channel, covered with roof tiles, turned from a sloping course and entered the
tunnel. After construction was completed, the tunnel and its more than 100 con-
struction shafts were filled with clay. Archaeological investigation of this tunnel
provided valuable clues about building techniques, including determining that the
individual shafts were as far apart as they were deep. The only likely explanation of
this detail is that the surveyor could not always be present at the site, or could make
only a single visit, and he therefore had to find an easy way to give the builders a
guideline for the depth of the shafts. He simply surveyed and staked out locations
for the shafts as far apart as they had to be deep. The builders merely had to measure
this distance by means of a string and then transfer it to the vertical shaft. Once the
weighted string was taut in each shaft, the required depth had been reached and the
tunneling linking the foot of each shaft could start. As a consequence of this method
the distance between shafts at the Drover-Berg tunnel increases sequentially from 6
to 8 m at the slopes at each end of the tunnel to 26 m toward the middle.
NONIUS DATUS AND THE AQUEDUCT

TUNNEL AT SALDAE
Ancient Saldae (today Bejafa, Algeria) was a coastal town whose origins reach back
to the Carthaginian period. Under Augustus a settlement was founded there for the
veterans of the seventh legion, but the town was not supplied with drinking water
by means of an aqueduct until the first half of the second century. The aqueduct was
designed to bring water from abundant, clear springs 17 km west of the city at the
foot of Djebel Arbalou, the westernmost peak of the great Kabylei; at present the
springs discharge 10,000 m 3 /day. The line chosen for the course of the aqueduct,
21 km in total, contained two major obstacles that could be overcome with the
technological means of the period. In its upper stretch the line followed an east-
west ridge with a slight gradient but a deep saddle at one point. Since the aqueduct
could not traverse this depression at ground level without losing the head of water
that was needed for coping with subsequent topographical obstacles, the dip had to
be crossed by means of an aqueduct bridge 300 m long. The central pillars of this
massive bridge were up to 15 m high. Further on, the line met a second obstacle in
the form of a mountain range that crossed its path obliquely from northwest to
southeast and could be overcome only by building a tunnel (Grewe 1998: 135-39).
Considered as a whole, the Saldae aqueduct system therefore corresponds to
the common image of the water supply system of a Roman city. A 300 m bridge and
a 428 m long tunnel hardly made this aqueduct stand out from significantly more
technologically challenging aqueducts built elsewhere. What makes it interesting to
the modern observer is an inscription that came to light as an isolated find in the
nearby Roman city of Lambaesis in 1866 ( GIL 8.2728, ILS 5795). The inscription had
been cut on a three-sided semicolumn 1.7 m high, with a matching hexagonal base;
both pieces had been built as spoils into a wall and had suffered some damage
( figure 12.5). Fortunately, the inscription has remained largely legible, but the three
heads at the upper part of the inscribed areas-personifications of virtues-were
destroyed. Mommsen (1871) pointed out that the inscription is incomplete and that
Figure 12.5. Inscription ofNonius Datus (CIL 8.2728), in Lambaesis, Algeria. (Photograph
by K. Grewe.)
the stone that was found may have belonged to a group of several columns that
stood side by side.
The surviving portion of the inscription is of outstanding importance for the
history of ancient engineering, since it not only provides technical details about the
construction of the tunnel at Saldae, but also provides details about the background
to the project. This contemporary report informs us who carried out the planning
and who created the technical basis for the construction of the tunnel. By means of
this practical information, we learn that the technical staff necessary for the exe-
cution of such construction works was available only in the army. Following a
request by the governor of the province of Mauretania Caesarensis to the legate of
Numidia, Nonius Datus, a military librator (surveyor) of the Legio III Augusta, was
assigned to the project. It was apparently this engineer who set up the inscribed
monument, which includes both background documents and personal narration of
the progress of the project. Noni us Datus explains the main problems involved with
tunnel construction by counter-excavation, as well as his solution. He also describes
how the two teams of builders missed each other in the mountain, and how he
remedied the earlier mistakes in his successful attempt to complete the tunnel. This
report, therefore, goes well beyond the usual contents of a building inscription,
even including drama and personal emotion. Since the very person in charge of the
construction of the tunnel composed the text, we have a nearly unique primary
source for the history of technology (translation adapted from Grewe 1998: 136-37,
Lewis 2001: 345-46) ( GIL 8.2728, !LS 5795):
PATIENCE, VIRTUE, HOPE

(or PATIENCE, ENERGY, CONFIDENCE)
[A.D. 153 Varius Clemens to M. Valerius] Etruscus. Saldae is a most splendid city,
and I, together with the citizens of Saldae, ask you, Sir, to request Nonius Datus,
retired surveyor of the Legio III Augusta, to complete the remainder of his work.
I set out and on the way was attacked by bandits; my staff and I were stripped
and injured, but escaped. I came to Saldae and met Clemens the [then] procu-
rator, who took me to the hill where they were bemoaning the poor quality of
workmanship on the tunnel. It seemed that they were considering abandoning it,
because the length of tunnel that had been driven was longer than the width of the
hill. Their bearings had evidently diverged from the straight line, so that the
bearing on the upper side had deviated to the right, the south, while that of the
lower side had likewise deviated to the right, toward the north, and thus the two
sections both deviated from the straight line. But the straight line had been staked
out across the hill from east to west. To save the reader from confusion about the
headings which are called "upper" and "lower," the "upper" is understood to
be the part of the tunnel where water is admitted, the "lower" that where it is
discharged. When I allocated the labor, I organized a competition between the
marines and the Alpine auxiliaries, so that they could learn each other's methods
of tunneling; and so they began driving through the hill. Therefore, since it was
I who had first taken the levels, marked out the aqueduct, and had a second copy
of the plans made which I gave to Petronius Celer the procurator [around
A.D. 137], [I completed] the task. When it was done and the water admitted,
Varius Clemens the procurator dedicated it.
[The capacity of the aqueduct is?] 5 modii.
To clarify my work on this aqueduct at Saldae, I append a few letters.
[A.D. 147/149] Porcius Vetustinus to [L. Novius] Crispinus. Sir, you acted
most generously and from your humanity and kindness in sending my Nonius
Datus, retired, so that I could use him on the works whose supervision he un-
dertook. So, though pressed for time and hurrying to Caesarea, yet I dashed to
Saldae and saw that the aqueduct was well begun, though of great magnitude,
and that it could not be completed without the superintendence ofNonius Datus,
who handled the job both diligently and faithfully. For this reason I would
have asked you to second him to us, so that he could stay for a few months deal-
ing with the matter, had he not succumbed to an illness contracted [ ... ]
The course of events around the construction of this tunnel, as they are related
in this text, can be reconstructed as follows: In the first half of the second century
A.D., the inhabitants of Saldae planned the building of an aqueduct from Djebel
Toujda to increase the water supply to the city. They turned to the responsible
official in Mauretania, the procurator Petronius Celer, with the request to help
them find a capable engineer. The procurator in turn asked his colleague, the legate
ofNumidia, to provide an expert for technical problems from the headquarters of

the Third Legion Augusta to evaluate the scheme. The surveyor Nonius Datus was
chosen. He traveled to the area, did some surveying, and subsequently gave his
plans for the building of the aqueduct to the procurator Petronius Celer. All of this
must have taken place by the year A.D. 138, since Celer was procurator of Maur-
etania only until the end of Hadrian's rule.
Work on the project, however, did not commence until ten years later, when
Porcius Vetustinus was procurator (A.D. 147-150). The long delay is very interesting
in itself. Crispinus was governor ofNumidia at the time, and he instructed Nonius
Datus to contact the procurator Vetustinus to discuss the project once again.
Crispinus was in office in the years 147 and 148, so the start of the aqueduct's con-
struction has to be dated to this time. Nonius Datus got in touch with the builders
who had taken over the execution of the construction, organized the work, and for
some time directed the operation. He would have stayed in Saldae for several
months had illness not forced him to return to Lambaesis. The work crew was
composed of military personnel.
Even during the engineer's absence work progressed well, and the project
probably would have been finished in time had there not suddenly been problems at
the tunnel site. Nonius Datus had staked out the line of the tunnel for the builders
across the mountain, and it was planned to excavate the tunnel from two sides. In
his absence, however, the builders had difficulty transferring the directions un-
derground, and each of the tunnel sections drifted off the planned bearing. Since
the length of the allotments was known from the prepared plans, at some point it
was noticed that the work had advanced far beyond the intended meeting point.
The realization that the teams had missed each other brought the builders to
despair, and they were about to abandon the project.
The procurator of Mauretania, now Varius Clemens (A.D. 151 or 153), turned
once again to the current legate of Numidia, M. Valerius Etruscus, asking him to
dispatch Nonius Datus a second time. Nonius Datus, by now a discharged veteran,
set off to Saldae but was waylaid by bandits. He and his escort managed to escape
and reached Saldae. From there, he was immediately taken to the mountain and
confronted with the tunneling problems. By taking measurements, he determined
the consequences of the erroneous advance and had a lateral link established be-
tween the two allotments. The aqueduct was completed and ready when the break-
through succeeded shortly after, and the procurator Varius Clemens inaugurated it.
When the French authorities cleared the tunnel to use it for a water pipe in the later
nineteenth century, its length was measured as 428 m, and the maximum depth
beneath the mountain summit as 86 m.
Nonius Datus used the construction of the Saldae aqueduct and its problem-
plagued tunnel to present himself on his tombstone as an engineer to whom an
outstanding building project had been entrusted. In the inscription he emphasized
his achievement in the construction and successful completion of the aqueduct
tunnel. The recording of the correspondence between the procurators of Maur-
etania and the legates of Numidia put the project on the highest official level and
removed any doubt with respect to his report. As an ancient technical source, the
inscription of Nonius Datus is unique.
Nonius Datus placed his report under three key words, Patientia, Virtus, and
Spes, each virtue personified by a female head. Presumably he was not so much
calling on three deities but rather stressing the three virtues required of the engineer
entrusted with the difficult task of constructing such a tunnel. The literal translation
would be "Patience, Courage, and Hope," but given the context, there might be a
better rendering. Mommsen replaced "Courage" with "Valor," but one might go
even further. The degree of difficulty in the task of tunnel construction requires not
courage or valor, but rather energy. Hope would be better interpreted as the
confidence of the expert who trusts on the basis of his know-how that the difficult
job will be accomplished. Perhaps Nonius Datus was commemorating his own
virtues of "Patience, Energy, and Confidence."
While quite a respectable number of tunnels were completed in the Greco-
Roman period, the medieval accomplishment is much more modest. In the period
from the fifth century A.D. to the beginning of the modern era, only a handful of
tunnels were built north of the Alps, for example. In the modern era, tunnel con-
struction flourished once again, particularly as a result of the spread of railway
networks in the nineteenth century. These systems required numerous tunnels due
to the special problems of low gradient and gradual turns. It is astonishing, how-
ever, that the nineteenth-century methods of tunnel construction had hardly
changed from those of Roman times. The first two railway tunnels in Germany, for
example, were built like qanats.
CANALS
Even today, water transport is the most economical method for the exchange of
goods between two places. This of course was especially true in antiquity, when the
overland routes were often relatively badly developed and maintained. For heavy
goods, such as building materials for cities, no other means of transport was viable.
The preferred waterways were navigable rivers. If there were no such rivers imme-
diately nearby, the loads were transported on the shortest land route, or an artificial
waterway was built to ease the transport. One reason for building canals was therefore
to link a production site or a destination to an existing navigable waterway. But it was
much more important to link different bodies of water with one another in order to
open up large regions with respect to water traffic. Many canals in the ancient world
were intended primarily for irrigation or drainage (Oleson 2000; Wilson 2000), but
these on the whole were technically less impressive than the navigational canals.
The history of the construction of canals for shipping reaches back to
the Bronze Age (C. Wikander 2000). The earliest canals can be found in Egypt,
Mesopotamia, and China. They are important testimony to the cultural impor-
tance of these countries: the canal from the Nile to the Red Sea, built around 600
B.C. by the Pharaoh Necho, was renovated repeatedly, by Darius I, the Ptolemies,
and Trajan, among others (Redmount 1995). The canal was wide enough to allow
two ships to pass each other. Herodotus (2.158; cf. Diodorus 1.33.9-12) reported
that the entire journey took four days.
Another early canal, cut by the Persian king Xerxes through the narrow strip of
land at the north end of Mount Athos between 483 and 480 B.C. (Isserlin 1996), is
described by Herodotus (7.23-24) as an undertaking intended more for prestige
than practical use. Herodotus gives a partial account of the methods of construc-
tion of this canal, which was wide enough to allow two warships to pass each other.
Xerxes had the structure subdivided into several allotments and apportioned the
work to the various nations and ethnic groups that formed his empire and his army.
All but one of the allotments experienced difficulty:
The foreigners, dividing up the ground by ethnic origin, dug in the following way.
They made a straight line by the town of Sane, and when the trench became deep,
some took positions at the bottom and dug, while others handed over the earth
that was constantly being excavated to another group positioned higher up on
the stepped construction face. The latter, in turn, as they received it, handed the
earth on to yet another group, until they came to those at the top, who carried
it away and dumped it. Everyone except the Phoenicians had double the work
when the steep sides of their excavations collapsed. Since they made the trench the
same width at the top and bottom, this was bound to happen. But the Phoenicians
showed their skill in this, as in so many other matters. Taking in hand the por-
tion that was allotted to them, they excavated the upper portion of the canal to
double the width that the channel itself was supposed to have, and narrowed it
as the work proceeded. At the bottom their excavation was equal in width to that
of the others.
Herodotus' attention to the details of the canal cross-section and dimensions is

atypical of ancient descriptions of canals, and it hints at the enormous physical
effort involved.
The Isthmus of Corinth, another narrow (6 km) and low (75-80 m high) ob-
stacle separating two important shipping lanes, early on attracted solutions to
allow the passage of ships. By the sixth century B.C. the diolkos (the word means
"haul across") had been constructed, a paved overland track over which smaller
ships were dragged (see chapter 23). Serious attempts to build a canal began with
Nero, whose project finally failed (Suetonius, Ner. 19; Dio Cassius 63.16). The ex-
tensive traces of his attempt were obliterated by the construction of the modem
canal (1881-93) and are no longer visible.
In Europe, the first shipping canals were built under the Romans, but not all of
these projects were successful. A project to build a canal linking the Saone and
Moselle rivers, and thus allow passage by ship from the mouth of the Rhone to the
mouth of the Rhine, is mentioned by Tacitus (Ann. 13.53), who alludes to the intent
to link the two seas. Although this canal was presumably conceived as a diversion
canal without locks, the project proved not to be feasible with the technology
available. The connection was completed only in 1882, as the Canal de l'Est.
No visible traces remain of a canal built by Drusus on the Lower Rhine, but two
ancient passages refer to the structure (Suetonius, Claud. 1; Tacitus, Ann. 2.8).
Suetonius praises it: "the Drusus Canal, as they still call it, a remarkable engineering
work which connects the Rhine with the Ijssel." This shipping route was probably
only 3 km long, but it saved the Romans the perilous passage of the North Sea off
the mouth of the Rhine. A similar purpose was served by a second canal that created
a 35 km link on the other side of the Rhine between the Rhine and the river Maas,
although Tacitus (Ann. 11.20) refers as well to a need to keep the soldiers busy. This
may have been a partial motive for many of these extremely labor-intensive engi-
neering projects. The construction of the Fossa Mariana (Marius Canal) between
Arles ( france) and the Mediterranean Sea east of the Rhone satisfied the need for an
alternative shipping route to the Rhone delta, which was subject to continuous
silting (Strabo 4.1.8; Pliny, HN 3.34), and it made Arles an easily accessible harbor
town. This canal undoubtedly featured in the abortive scheme to connect the
Mediterranean Sea with the North Sea via the planned Moselle-Saone canal. Both
Horace (Sat. 1.5.9-23) and Strabo describe the famous canal that crossed the
Pontine marsh (Geography 5.3.6; cf. chapter 22): "Near Terracina, as you go toward
Rome, a canal runs alongside the Appian Way, fed at numerous points by the flow
from marshes and rivers. People navigate it mostly at night, so that embarking in
the evening they can disembark at dawn and continue the rest of the way by the
road-but it is also navigated by day. A mule tows the boat."
The feasibility oflarge-scale engineering projects was at all times dependent on
a number of essential prerequisites: in addition to availability of appropriate ma-
terials and labor, and the technical knowledge to complete the project, funding was
essential, and the power to conceive the project and see that it was brought to
completion. As the example of Nonius Datus shows, experts had to be borrowed
from the military for the most difficult part of the works in Roman tunnel con-
struction. It will have been no different for the construction of canals, since these
works, too, were difficult and required a high level of technological skill. The cor-
respondence between Pliny the Younger, at the time governor of the province of
Bithynia in Asia Minor, and the emperor Trajan amply confirms the need for
imported expertise (Pliny, Tra. 10.41-42). Pliny had been persuaded of the com-
mercial value of a canal linking a lake near Nicomedia with the sea (Moraux 1961),
and he asked Trajan for a librator, that is, an expert in planning and surveying, such
as Nonius Datus. In his reply, Trajan advised Pliny to contact the governor of Lower
Moesia, "since in the provinces there is no lack of such experts." He furthermore
suggested that the engineer ought to make certain through measurement that the
lake would not be completely drained by the planned canal.
Despite the lack of explosives and machine-operated earth-moving equipment, all

the great states and empires in the ancient Mediterranean world managed to create
tunnels and navigational canals. The major prerequisites were the availability of
large quantities of cheap manpower, along with a certain level of surveying skill.
The difficulties involved meant that such projects were generally undertaken only
for purposes of enormous importance to the state, such as urban water supply,
flood control, or the facilitation of trade.
REFERENCES
al-Karaji, M. 1973. La civilisation des eaux cachees. Trans. A. Mazaheri. Nice: Universite
de Nice.
Frontinus-Gesellschaft e.V. 2003. Wasserversorgung aus Qanaten: Qanate als Vorbilder im
Tunnelbau; Symposiumsberichte. Bonn: Frontinus-Gesellschaft.
Giorgetti, D. 1988. "Die antike Wasserversorgung von Bologna," in Die Wasserversorgung
antiker Stiidte. Frontinus-Gesellschaft Geschichte der Wasserversorgung 3. Mainz:
von Zabern, 180-85.
Grewe, K. 1998. Licht am Ende des Tunnels: Planung und Trassierung im antiken Tunnelbau.
Mainz: von Zabern.
Isserlin, B. S. J. 1996. "The Canal of Xerxes," Annual of the British School of Archaeology
at Athens 91: 329-40.
Kienast, H.J. 1995. Die Wasserleitung des Eupalinos auf Samos. Samos XIX. Bonn: Habelt.
University Press.
Mommsen, T. 1871. "Tunnelbau unter Antoninus Pius," Archaologische Zeitung n.s. 3: 5-9.
Moraux, P. 1961. "Die Plane Plinius des Jiingeren fiir einen Kanal in Bithynien," in G. Radke
(ed.), Aparchai: Untersuchungen zur klassischen Philologie und Geschichte des Alter-
tums, vol. 4. Tiibingen: Niemeyer, 181-214.
Oleson, J.P. 2000. "Irrigation," in 6. Wikander 2000, 183-215.
Redmount, C. 1995. "The Wadi Tumilat and the 'Canal of the Pharaohs,'" Journal of
Near Eastern Archaeology 54: 127-35.
Wikander, C. 2000. "Canals," in 6. Wikander 2000, 321-30.
Wikander, 6. (ed.) 2000. Handbook of ancient water technology. Leiden: Brill.
Wilson, A. 2000. "Land drainage," in 6. Wikander 2000, 303-17.
CHAPTER 13
MACHINES IN GREEK
AND ROMAN
TECHNOLOGY
ANDREW I. WILSON
A MACHINE is a device in which a combination of fixed and moving parts are

used to perform work either by multiplying an effort to exert greater force than
can human or animal muscle power alone, or by changing the direction of a force,
or by automating simple or complex repetitive operations. Hero of Alexandria
lists as simple machines (or mechanical powers) the wheel and axle, lever, pulley,
winch, wedge, and screw (Mechanica 2.1-5; cf. Pappus 8.52); one should also
add the inclined plane and the gear wheel (which Hero considered as belonging to
the winch). All these, except the gear wheel and the screw, which are Hellenis-
tic inventions, were known before the archaic period. Complex machines com-
bine and multiply simple machines, with additional improvements or variations
like the cam, crank, or rack-and-pinion. Sources of power in the ancient world
included human muscle power, animal traction, heat, wind power, and water power
(chapter 6). On ancient mechanics, the mechanical writers, and machines, see es-
pecially Drachmann (1948 and 1963), and Lewis (1997). Gille (1980) is an accessible
introduction but unreliable in some details, notably chronology. General accounts
such as Hodges (1970: 209-41) and Landels (2000) are outdated on, for example,
the introduction of the water-mill and the question of technological stagnation.
Hill (1984) is useful on the relation between Greek and Islamic automata (but cf.
Lewis 1985).
Greek and Roman machines were used primarily in construction, water-lifting,
mining, the processing of agricultural produce, and warfare. The use of machines
in manufacture was relatively limited, although the loom mechanized complex

weaving procedures (chapter 18), and in the Roman and late Roman periods we
find some evidence for the diversification of water-power for productive ends.
Other uses included medical implements and entertainment devices.
Sources of evidence are not straightforward. Representations of known devices
are often lacking, and where they do survive may be inaccurate. Because machines
tend to be composed of materials of varying survivability (e.g., wood, metal, sinews,
rope), only fragments may survive; more usually, only those machines requiring
durable fixing or infrastructure in masonry leave traces (presses, mills). Many
devices are not described in surviving texts or are mentioned only in passing. Our
main sources are several writers on mechanics, but many treatises are now lost, and
even some of those that survive do so only in medieval Arabic translations.
The earliest writings on mechanics dealt largely with engines of warfare. The
creation around 305 B.C. of the Museum at Alexandria, a scholastic research center
under the patronage of the Ptolemies, brought about major advances through the
concentration of scholars working on many different aspects of science. Several of
these wrote on machines or mechanical problems. Although in the Peripatetic
tradition, the pseudo-Aristotelian Mechanical Problems, concerning levers and
circular motion, was probably written at Alexandria between about 280 and 260
B.C. Around the same time Ctesibius (fl. 270 B.C.) invented the force-pump and
hydraulic organ; his writings are lost, but some of their content may be deduced
from surviving works of Philo of Byzantium, Vitruvius, and Hero. Archimedes
(287-212 B.C.) visited Alexandria around the mid-third century B.C.; his work
Equilibriums of Planes is important for mechanics.
In the later third century B.C. Philo of Byzantium spent time at Alexandria,
although his partially surviving Mechanice syntaxis dates from after his visit. Of its
nine original books we have books 4 (Belopoeica, "Artillery"), 8 (Poliorcetica, "Siege
Techniques"), and parts of 7 (Parasceuastica, "Preparation for Sieges") in Greek,
and book 5 (Pneumatica, "Pneumatics," or the "Interaction of Air and Water") in
Arabic translation and partial Latin translation. The lost books included works on
levers and mechanics, harbor construction, automata, and water supply (Lewis
1997: 20). A slightly younger contemporary was Apollonius of Perge (fl. ca. 200
B.C.), who besides writing on conic sections produced a now lost work on a flute-
player driven by compressed air released by valves controlled by the operation of a
water wheel (see Gille 1980: 79; Lewis 1997: 49-57).
Apart from some works on engines of war, we have no further surviving works
on mechanics until Vitruvius (ca. 25 B.C.): book10 ofhisDeArchitectura is devoted
to machines (cranes, water-lifting devices, the water-mill, force pump, and water
organ, and siege engines). Hero of Alexandria, one of our most important sources,
wrote during the later first century A.D. (his Dioptra mentions an eclipse datable to
A.D. 62). His Pneumatica andAutomatopoieca ("On Automata-making") survive in
Greek, and three books of his Mechanics in Arabic translation, along with excerpts
preserved in Pappus of Alexandria's Mathematical Collection 8 (ca. A.D. 362). The
Arabic translation is prefaced by the Baroulkos, "On the Lifting of Heavy Objects,"
MACHINES IN GREEK AND ROMAN TECHNOLOGY 339
originally a separate work that became incorporated also into his Dioptra. Frag-
ments survive of other treatises relevant to mechanics, for example, one of four
books of his Water Clocks, quoted by Proclus and referred to by Pappus. Pappus
provides a useful source for and control on the parts of Hero's work that do not
survive in Greek, but adds little new.
The surviving literature shows that while the lever, winch, inclined plane, and
wedge were known in the Greek world during the classical period, considerable
advances in mechanical technology occurred in the early Hellenistic period under
the stimulus of Ptolemaic royal patronage at Alexandria. Philo of Byzantium ex-
plicitly refers to the discovery of formulae governing the power and range of
artillery discovered by repeated experiments financed by the Ptolemies (Belopoeica
50.3); but the Alexandrian school seems also to have been responsible for the
invention of gearing, and the multiple combinations of several mechanical powers
to produce a variety of complex machines and automata.
SIMPLE MACHINES OR
MECHANICAL POWERS
The mechanical advantage (MA) of a machine is the factor by which it multiplies

the input force or effort into an output force. A lower input force must be applied
over a greater distance than the greater output force travels; the ratio of the dis-
tances is the velocity ratio (VR). Theoretically, MA= VR, so that in a machine with
a mechanical advantage of 2, the input force is half the output force but must be
exerted over twice the distance. In practice, friction reduces the ideal mechanical
advantage of a machine. A machine can be used to convert an input force into a
greater output force applied more slowly, or into a smaller output force that
operates more quickly; in this case a mechanical disadvantage is converted to speed
advantage.
The simplest mechanical power is the inclined plane. To raise a weight a certain
distance it is easier to drag or roll it up a slope than to lift it vertically (MA= length/
height of the slope). Before the invention of cranes, dragging or rolling architectural
elements up temporary earthen ramps was almost the only way to place large blocks
in elevated positions on a structure. The lever was also used at a very early date,
although the theory behind its operation was not fully understood until Archi-
medes, who described it in Equilibriums ofPlanes 1. The MA of a lever is the distance
between the fulcrum and the point where the force is applied divided by the dis-
tance between the fulcrum and the weight to be lifted. Crowbars were used in con-
struction from very early times for levering large blocks into place. Levers were also
used in balances: the equal arm balance was known to the Egyptians (Skinner 1954:
779-84), and the unequal arm balance (steelyard) is referred to in the third-century
B.C. pseudo-Aristotelian Mechanical Problems, which also uses the nutcracker and
dental forceps as illustrations of the lever. It was used in Roman dental and surgi-
cal forceps and in anal specula (Jackson 1990: figs. 3, 5-6).
The simple pulley was known to the Assyrians and was used from at least the
ninth century B.C. for drawing water from wells. It confers no mechanical ad-
vantage in itself but changes the direction of pull; it is easier to pull down than up.
The triple pulley, however, with two pulleys attached to the crane and a free pulley
suspended from them, has a mechanical advantage of 3; for compound pulleys of n
pulleys in a similar arrangement, MA= n, although because of friction ancient
pulley systems were probably limited in practice to five sheaves (Vitruvius' pen-
taspaston; see below on cranes). The surviving ancient pulley blocks-for the most
part found on shipwrecks-include a single-sheaved pulley from the first-century
A.D. Lake Nemi barge, and remains of third- and fourth- or fifth-century A.D.
double-sheaved pulley blocks from London and Cenchreae, respectively, presum-
ably the upper blocks of triple-pulley systems (Shaw 1967; Oleson 1983).
The windlass/winch and capstan (the winch has a horizontal axle and the
capstan a vertical one) use handspikes or levers inserted into slots on a drum to gain
a mechanical advantage in circular rotation, given by the radius of the handspike to
the radius of the drum or axle. The winch was known by the fifth century B.C., used
by surgeons for traction when setting limbs (Drachmann 1963: 184, and below).
Simple windlasses were used for horizontal haulage, for example the overland
haulage of boats (and on occasion light warships) along parts of the diolkos or
trackway across the Isthmus of Corinth, built by Periander (625-585 B.C.). Boats
were dragged on large trolleys running, in places, in guide ruts, and rope-wear
marks and sockets cut in a guide wall on a long incline suggest that capstans may
have been used to haul trolleys up the slope (Lewis 2001b; cf. chapter 23). Wind-
lasses are discussed further in combination with other powers in the sections on
cranes and presses below.
Gearing allows the multiplication of force or speed, or changing the direction
of rotation, by a number of different means. Mechanical advantage can be gained by
using a smaller wheel to drive a larger one, which will rotate more slowly and in the
opposite direction. Alternatively, a speed advantage can be gained by making the
driven wheel smaller than the driver. Gears meshing in the same plane will rotate in
opposite directions and can thus reverse rotation; but direction can also be changed
from vertical to horizontal rotation with right-angle gearing. Gearing was probably
developed in the early third century B.C. at Alexandria. Ctesibius' anaphoric water
clock, developed around 270 B.C., used a rack-and-pinion driven by a descending
float as water flowed out of the clock. The contemporary Mechanical Problems
refers to the opposite rotation of adjacent wheels, but it does not explicitly mention
gearing and may refer to effects produced by friction of touching, untoothed
wheels. This detail may hint that the principles of gearing were just beginning to be
explored at this date, but subsequent development was evidently rapid. Archimedes
invented the worm drive around 250 B.C., which presupposes the prior existence of
simpler cogwheel gearing. Indeed, Apollonius of Perge's treatise On the Fluteplayer
and other evidence implying the invention of the vertically-wheeled water-mill and
the saqiya drive between 270 and 240 B.C. also indicate the existence of right-angled
gearing at this date (Drachmann 1963: 200-205; de Solla Price 1974: 53-54; Lewis
1993, 1997: 49-57).
Gears used in practical machines had wooden teeth projecting radially as
spokes from the disc, or at right angles as in a lantern pinion. These were adequate
for applying force, but a different shape of tooth was used in precision instruments
such as planetaria, and preserved in the Antikythera Mechanism (figure 29.1). These
devices used gears cut from bronze discs, with small teeth filed into the shape of
equilateral triangles. Rack-and-pinion gearing was used in Hero's self-trimming
lamp to adjust the length of the wick as the oil level fell (Pneumatica 33).
Very few ancient gears survive. A lantern pinion of six iron bars between two
oak disks on a mill-spindle was found at the Saalburg Roman fort (before A.D. 270;
Jacobi 1912: 89-90; Moritz 1958: 123-28). For fine gearing, apart from the Anti-
kythera Mechanism, there are finds from Bolsena and possibly Timgad, as well as
several devices of unknown origin. At Bolsena, a screw and cogwheel were found in
a hoard of metalwork in the villa ofLaberius Gallus (Oleson 1984: 193). The Timgad
finds come from a bronzeworker's shop destroyed by fire in the fourth or fifth
century A.D. They are described briefly (but not illustrated) as fragments of a gear
wheel ("de roue de machine"), one of which had a diameter of 52 cm, and a
fragment of what the excavator describes as a circular saw ("scie circulaire") (Ballu
1911: 23, 169). Since these objects are all of bronze, not iron, the alleged circular saw
(internal diameter 32 cm) may in fact be part of a toothed wheel. Perhaps they were
intended for a planetarium or the jack-work of an anaphoric clock (chapters 29 and
31), but without further details this can only be speculation. A Byzantine geared
sundial calendar without provenance is now in London ( figure 31.3; Field and
Wright 1985).
The screw works like an inclined plane applied in a rotary sense, either to exert
pressure, or in combination with gearing as a worm drive. Although high levels of
friction reduce its efficiency, it has a considerable mechanical advantage. The worm
drive first appears between 241 and 239 B.C. as part of Archimedes' device for
launching Hieron's ship; it was then used in two medical stretching racks of the
later third century (Lewis 1997: 54-56). Early screws had a male thread with a square
cross-section. Initially, this groove simply engaged with a peg thrust through the
side of the hole into which the screw was inserted. In the first century A.D. Hero of
Alexandria described a screw-cutting machine that allowed the cutting of female
screws (Mechanica 3.21), enabling the wooden screw to be used in presses, both for
drawing down the arm of a lever press and in a direct screw press (below). A thread
with a lenticular cross-section produced less friction and was less prone to jam-
ming, and was consequently used in bronze precision gearing and in wooden screws
with a female nut rather than a peg (Drachmann 1936, 1963).
Bronze screw gearing was incorporated into devices in Hero's Dioptra to im-
prove existing models of the instrument by using the velocity ratio, rather than
mechanical advantage, to allow fine adjustment. One turn of the handle connected
to the screw produced only a small rotation of the geared sighting disc in the
horizontal or vertical plane, allowing precise setting. A groove along the length of
the screw acted as a release mechanism; if aligned with the geared disc, it could be
rotated rapidly by hand (Hero, Dioptra 3; cf. Lewis 2001b: 83). Gynecological
specula (also called dioptra) used the screw's mechanical advantage and high ve-
locity ratio, coupled with precision adjustment, for vaginal dilation. Made of
bronze (sometimes with a brass screw), they had three or sometimes four vanes
adjusted by a handle attached to a screw (Jackson 1990: 9, 21, figs. 8-9; Longfield-
Jones 1986).
Power sources applied to large machines might include animals or flowing
water (chapter 6); in both cases this usually involved a wheel or gearing producing a
rotary motion. From the third century B.C., however, the principle was understood
of converting rotary to reciprocating linear motion by means of the cam on an axle
(chapter 31), and it seems that this device was applied not only in automata but
probably also in larger machines by the first century A.D.
CRANES AND TRACTION
The earliest complex machines were cranes that employed pulleys in combination
with a windlass (figure 13.1b). Power output could be increased by compound
pulleys and geared windlasses, driven by men in a treadmill ( figures 2.4, 13.1c and d;
Vitruvius, De arch. 10.2.5; Hero, Mechanica 2.3, 23, 3.1-5; Drachmann 1956; Cotterell
and Kamminga 1990: 89-93). There is no evidence for the use of cranes or hoists in
architecture before the late sixth century B.C., and until then heavy blocks must
have been raised by pulling them up earth ramps, as the Egyptians had done. The
introduction of cranes eliminated the labor-intensive building and progressive
heightening of such ramps during construction and their subsequent removal on
completion, and enabled the more widespread lifting of heavier blocks than had
previously been possible. One immediate effect however, was a reduction in the
size of the very largest blocks used. The first cranes are implied by the use of lifting
tongs and of lewis irons, the characteristic holes for which appear in Greek
structures in the late sixth century B.C. (Coulton 1974). Cranes using pulleys or a
block and tackle might consist of one pole held in position by stays, with the pulley
attached at the top; this arrangement was unstable but allowed the load to be
swung in an arc by adjusting the angle of the pole using the stays. A more stable
version used two poles fixed together in an inverted V, allowing the load to
be moved back and forth by raising or lowering the frame (figure 13.1). More secure
frames with three or four uprights only allowed the weight to be lifted vertically
(Vitruvius, De arch. 10.2.1-10; Hero, Mechanica 3.2).
(a) (b)
theoretical diagram actual mechanisms
upper block
and tackle
F = _r_
2
(c) (d)
Figure 13.1. Ancient cranes: (a) theoretical diagram and mechanism of pulley block and
tackle; (b) reconstruction of a crane shown on a wall-painting from Stabiae (Villa of S.
Marco); (c) reconstruction of a crane with separate tread-wheel drive shown on a relief
from Capua; (d) crane with multiple pulleys and axle-mounted tread-wheel, based on the
Haterii relief. (J. P. Adam, La Construction Romaine, figs. 91, 89, 93, and 95.)
Before the invention of the compound pulley, early cranes must have used a
rope passed over a simple pulley and wound around a windlass or capstan to
provide mechanical advantage. Capstans winding ropes running over pulleys are
shown on the first- or second-century A.D. relief found near the drainage outlet of
the Fucine Lake, and representing the extraction of spoil from the underground
drainage tunnel via vertical shafts ( Faccenna 2003; Giuliani 2003) . Extra mechanical
advantage could be achieved using a capstan in series with the axle of the crane
(Vitruvius, De arch. 10.2.5).
The invention of triple and compound pulleys ( trispaston and polyspaston) was
a major advance in the design and mechanics of cranes. This has been ascribed,
probably erroneously, to Archimedes, as a result of ancient authors' usage of tris-
pastos and polyspastos to mean both three (or multiple) pulleys and three (or
multiple) winches. Some clues, in fact, point to a date before Archimedes for this
invention. An inscription of 329/328 B.C. (JG 1672.156) mentions multiple pulleys,
although the situation might involve simple pulleys in parallel (Coulton 1974: 1).
The earliest secure mention of the compound pulley appears in the pseudo-Aris-
totelian Mechanics 18 (ca. 270 B.C.). Coulton (1974) argues on the basis of weights
lifted in archaic and classical-period architecture that the compound pulley may
have been known from the late sixth century B.C., although it should be observed
that the required mechanical advantage might have been produced from multiple
ropes each attached to simple pulleys and a windlass.
Several ancient sources report that Archimedes invented a method of dragging
enormous weights, and that he demonstrated this to Hiero II of Syracuse, allegedly
by singlehandedly launching a large ship laden with passengers. The details are
confused by uncertainty over whether our sources are using trispaston and poly-
spaston to mean pulleys or winches. Vitruvius clearly uses them to mean pulleys
when he is describing cranes; Hero, in his Baroulkos, uses them to mean a geared
winch, with three or multiple axles.
Plutarch (Marcellus 14.8) says Archimedes used a polyspaston to move Hieron's
ship, with his left hand only and while sitting down. John Tzetzes (ca. A.D. 1150)
refers to a trispaston (Ghil. 2.107-8). Oribasios (Coll.Med. 49.6, 23), following Galen
(A.D. 129-99), ascribes the invention of the polyspaston to Archimedes and says that
Archimedes invented the trispastos for hauling ships. The machine he goes on to
describe as the doctors' reduction of this instrument has no pulleys, but rather
multiple winches (Drachmann 1963: 178-80), and it bears a generic resemblance to
Hero's baroulkos (below). Since Athenaeus (Deipnosophistae 5.207a-b), in what
seems to be a version of the same tradition, says Archimedes used a helix (screw or
worm drive) to launch Hieron's vast grain freighter, we may conclude that Ar-
chimedes' device involved geared winches driven by a screw. Drachmann (1973: 50-
51) considered that Plutarch's description of singlehanded operation implies a
crank handle. This mode of operation sounds like Hero's Baroulkos and may be
implied by Pappus: "The same theory applies to the work of moving a given weight
by a given power; this was found out by Archimedes, who is said thereupon to have
exclaimed: 'Give me somewhere to stand, and I shall move the earth.' Heron of
Alexandria has shown quite clearly in his so-called Baroulkos how this may be
contrived ... " (Mathematical Collection 8.19; trans. Drachmann 1963: 28). Indeed,
Tzetzes implies that Hero's Baroulkos was derived from one of Archimedes' writings
(Ghil. 12.964-71; cf. Lewis 2000: 332).
Hero's Baroulkos survives in Greek (Dioptra 37), and in Arabic as the opening of
his Mechanics, where it evidently does not belong, since another version exists
within the Mechanics proper as chapter 2.21 (Drachmann 1963: 22-32). The passage
describes how to use a windlass with a series of gears to multiply the power input to
a crane or winch. The first gear is turned by a screw with a cheirolabe (apparently a
cross-handle or windlass handle rather than a crank-Drachmann 1973: 43-44).
Drachmann (1963) thinks the baroulkos was a theoretical device only. But it is very
similar to some medical benches for surgical traction (below); could it be Hero's
practical revival of Archimedes' polyspaston? Who was the engineer who proposed
to Vespasian (reigned A.D. 69-79) a method for hauling columns up to the Capitol
(Suetonius, Vesp. 18)? It is tempting to wonder whether this was either Hero or
someone who had read his description of the baroulkos, which must date from
around this time-the Dioptra, in which it appears, was written after A.D. 62.
A variation on the crane was the pile driver, which raised a heavy weight with
pulleys and then suddenly dropped it onto a pile to drive it into wet ground. Pile
drivers are mentioned by Vitruvius (De arch. 3.4.2) and Caesar (BG 4.17), but not
described. The framework must have consisted of three or four timbers joined near
the top. Usually the weight was dropped vertically onto the head of the pile, but for
the sloping support piles used in Caesar's bridge over the Rhine it must have slid
down an inclined ramp to drive them in at an angle. Caesar may also have used pile
drivers mounted on tethered rafts ( Frau 1987: fig. 69). How the weight was released
suddenly is unclear, but may have involved some kind of trigger mechanism like
those of catapults, onagers, or perhaps the starting gates in a circus.
SURGICAL TRACTION
Oribasius, writing around A.D. 362, describes several surgical traction machines for
resetting broken limbs, some of which go back to the fifth century B.C. (Oribasius
49; Drachmann 1963: 171-84; cf. 1973). These machines spawned a progeny of later
torture implements, notably the rack. The original was the bench of Hippocrates
(ca. 420 B.C.), a rack with simple winches at each end, winding ropes that stretched
the patient. The bench ofNeileus was essentially a winch with handspikes that could
be attached either to a bench, or, if carried to treat the patient away from the
doctor's surgery, to a ladder. The bench of Andreas (court doctor at Alexandria,
murdered 217 B.C.) used a screw thread instead of a winch. The chest of Nym-
phodorus ( fl. ca.250-200 B.C. at Alexandria) was more compact and powerful,
turned by a screw engaging a gear wheel on the winch (Lewis 1997: 55). Later
improvements to Hippocrates' bench used either pulleys or geared winches to gain
a mechanical advantage. In the mechanical chest of Galen, ropes attached to one
end of the patient's limb were wound around two pulleys and taken back to each
end of a windlass axle, while another rope attached to the other end of the limb was
wound around the center of the axle (Oreibasios 49.7, from Galen, On Engines;
Drachmann 1963: 173). Other machines, including the trispaston of Apellis or Ar-
chimedes, and its modifications by Pasicrates, Heliodorus, and Aristion, used
multiple axles with ropes wound around them for mechanical advantage, like the
geared winches in Heron's Baroulkos (Oreibasios 49.23; Drachmann 1963: 178-84).
ENGINES OF WAR
Constraints on the how's effectiveness and range were the power that could be
applied to draw it, limited by human strength to about 45 lb, and the length through
which it could be drawn, limited by the length of an archer's arms. For the con-
struction oflarger artillery, one needed to create a bow or springs with considerable
power and apply sufficient traction to pull back the arms of the machine. In all
artillery pieces except the gastraphetes, traction for spanning the bow was achieved
by variants on the means already described for cranes and surgical traction benches.
Marsden 1969 and 1971 remain fundamental for this topic, although written before
the identification of much of the archaeological material; a number of his ideas are
now challenged. An accessible overview including the archaeological finds is
Campbell 2003; more specialized accounts appear in Baatz 1978, 1994, 1999.
In 399 B.C., Dionysius of Syracuse offered high rates of pay to craftsmen from
around the central Mediterranean to come to Syracuse and produce for him nu-
merous weapons, in preparation for war against Carthage. According to Diodorus
Siculus, "The catapult was invented at that time in Syracuse, since the most skillful
craftsmen from everywhere had been gathered in one place" (14.42.1). Artillery was
first used in anger at Dionysius' siege ofMotya in 397 B.C. This first artillery weapon
was probably the gastraphetes or belly shooter (Marsden 1969: 48-49). Like a large
crossbow, it had a large bow attached to a stock, in which ran a slider carrying the
bolt; the back of the slider engaged with the bowstring, so that before the bow was
spanned the slider projected well forward beyond the end of the stock. A curved
crosspiece at the end of the stock fitted the archer's stomach, allowing him to span
the weapon by leaning the end of the slider against the ground or a wall, and
pushing against the stock with his body. Ratchets on the stock and a pawl on the
slider prevented the slider moving forward again. The trigger consisted of an iron
bar with a claw at one end hooked over the bowstring, fired by pulling a hinged bar
from underneath the back end of the trigger, so that the claw tilted up and released
the bowstring. During the fourth century B.C., larger versions were developed,
mounted on a base and spanned by a windlass. One version shot two bolts at once
(Marsden 1969: 5-16; Campbell 2003: 3-7).
The gastraphetes was more powerful than an ordinary bow, but further de-
velopment was limited by the size of composite bow that could be made from horn,
wood, and sinew. In the catapult, therefore, the single bow was replaced with two
separate arms, each thrust though a twisted skein of sinews or hair held in a frame.
Metal levers placed over circular metal washers at the top and bottom of the spring
frames allowed the skeins to be tightened. The bowstring was strung between the
free ends of the arms, and the slider was pulled back by a windlass, or, in more
powerful catapults, by a windlass and multiple pulley system as in the larger cranes
or medical benches. The catapult was mounted on a tilt-and-swivel base so that it
could be rotated and elevated or-in the case of arrow-shooters-depressed to fire
from city walls (Marsden 1969; Campbell 2003).
Torsion machines of this kind, in which the spring frame formed a straight line,
were called euthytonoi (straight spring) and were used for shooting bolts. More
powerful variants, in which the springs were turned away from each other so that
the arms could be bent back through a greater arc, were called palintonoi (back-
spring devices). These could shoot heavy bolts but were more commonly used for
stone shot, using a wider slider and a sling rather than a bowstring. An inscription
of 306/5 B.C. from Athens mentions a catapult that could shoot either bolts or
stones, presumably by changing the slider (Marsden 1969: 70).
Philo (Belopoeica 50) says that a program of systematic research under the
Ptolemies developed formulae for the optimal proportions for the different parts of
euthytone and palintone machines in relation to the missiles they were intended to
fire. These took as a module the spring diameter, which in the arrow-shooter was to
be one-ninth of the arrow's length. In the palintone stone-shooter the formula was
more complicated, based on the weight of shot:
Diameter of spring in dactyls = 1.1 y' (100 x weight in minae)
Dimensions of all the parts of the machines are expressed as multiples of the spring
diameter (Marsden 1969: 24-41).
Philo (Belopoeica 73-78) describes a repeater catapult built by Dionysius of
Alexandria, in which bolts were fed one at a time from a hopper into a groove on a
cylinder revolving inside a case, which dropped them into the firing groove on the
stock. The machine was spanned by a chain of wooden pallets and iron links fitted
around a pentagonal cogwheel turned by the winch. This cog also rotated the
magazine in a synchronized fashion so that each bolt dropped into the groove after
the arms were pulled back, and released the trigger; and then drew back the arms
and cocked the machine for the next shot. Reconstructions have proved the device
fast and accurate, but Philon criticized it for firing all of its shots at the same target
and thus wasting ammunition when one shot would do: "The missiles will not have
a spread, since the aperture has been laid on a single target and produces a trajectory
more or less along one segment of a circle, nor will they have a very elongated
dropping zone" (Belopoeica 76).
Philo presents several alternative catapults: a wedge catapult, whose torsion
springs are tightened by wedges; a catapult in which twisted sinews are replaced by
bronze springs (Bel. 59-73), and the pneumatic catapult of Ctesibius in which, as the
machine is spanned, pistons are driven into tight-fitting cylinders; the air thus
compressed provides the motive power for the arms when the trigger is released
(Bel. 77-78). These machines may have been built as experimental prototypes, but
were never adopted in general production (Marsden 1969: 41-42).
Some subsequent design improvements can be traced (Marsden 1969: 42,
174-98). The catapult on a second-century B.C. balustrade relief from the sanctuary
of Athena at Pergamon has curved arms, allowing a greater arc of travel than in
straight-armed euthytonoi or palintonoi, and hence a more powerful shot. By
around 25 B.C., when Vitruvius wrote, more efficient oval washers had been de-
veloped, allowing more sinew to be put in the springs. Philo's dimensional for-
mulae probably no longer held good: the improvements allowed smaller, lighter
machines to be constructed with the same firepower, facilitating their use by field
armies. Indeed, Trajan's Column and the column of Marcus Aurelius show catapults
mounted on small carts, which must be the carroballistae mentioned by Vegetius
(Mil. 2.25, 3,14, 24). Heron's Cheiroballistra (possibly not by Heron of Alexandria)
describes parts of a handheld ballista. There is controversy over its reconstruction
(Wilkins 1995, 2000; Iriarte 2000), but the term cheiroballistra (hand-ballista) and
its Latin equivalent, manuballista, surely imply a handheld, portable weapon
(figure 13.2).
Further debate rages around the design of late Roman catapults. In 1972 the
bronze frame of a stone-throwing catapult of the mid third century A.D. was
excavated at Hatra in Iraq (Baatz 1978: 3-9). Its exceptional width (2.40 m) and two
notches on the inside of the spring frames suggested to several scholars that the
arms at rest projected forward and were drawn inward through the frame when the
weapon was spanned. This would allow a travel of some 120 degrees as opposed to
about 67 degrees achievable in a palintone machine with outward-swinging arms,
giving a performance increase of about 50 percent. The inward-swinging design
seems to have been introduced in the late first century A.D., and despite some
disagreement (Wilkins 2000 ), inward-swinging arms were probably also used in the
cheiroballista (figure 13.2) (Anstee 1998; Iriarte 2003; Anon. 2004).
Torsion machines were capable of considerable range. Agesistratus, in the first-
century B.C., built an arrow-shooter that could fire a three-span arrow 700 yards,
and a larger palintone engine that shot a four-cubit bolt 800 yards. Effective ranges
would have been shorter. Modern reconstructions have not achieved such ranges,
but most use hair or nylon rope for the springs instead of the much more elastic
sinew preferred in antiquity. Stone catapult balls from Carthage, Pergamon, Rhodes,
and Tel Dor range in weight between 3 minae (1.3 kg) and 180 minae (78.6 kg).
Josephus says that at the siege of Jerusalem (A.D. 69) the Roman stone-throwers
shot stone balls of one talent some 400 yards; but this may have been against
personnel, and Marsden thinks the effective range against a wall might have been
less than half that.
Trajan's architect Apollodorus of Damascus mentions "one-armed stone-
throwers" in his Poliorcetica, possibly the machines referred to by Vegetius as
onagri. Ammianus Marcellinus (23.4.4-7) attempted a technical description of the
onager (wild ass, so called perhaps because of its back-kicking recoil), a one-armed
ballista for hurling stones from a sling. Most authors (e.g., Marsden 1969; Campbell
(a)
XV>M d
I P
,u
,:1,::
1"111'
(b)
X X\'X \ ' lrl

I p
N
~
0 5 10 1!!i 20 215 JOcm
Figure 13.2. Reconstructions of the cheiroballistra (hand-held ballista) as (a) an outswinger

and (b) an inswinger. (Reconstruction by Aitor Iriarte, reproduced by permission.)
2003: 42-43) reconstruct this as similar to the medieval mangonel. An arm with a
sling is inserted in a torsion bundle stretched horizontally across the frame of the
machine, and the arm is drawn back to the horizontal by a windlass. A trigger
mechanism releases the arm, which swings upward, hurling a stone from the sling,
and a padded buffer in front of the machine absorbs the impact of the arm. A
completely different reconstruction has been proposed (Cherrette 2001-2002), but
has yet to be tested as a full-size working model.
Archaeological finds, representations, and especially the design of fortifica-
tions, in which provision for stationing torsion artillery is common in the Helle-
nistic and Roman periods, suggest that between the fourth century B.C. and the
fourth century A.D. Greek and Roman armies made frequent and widespread use of
powerful artillery in both offensive and defensive roles in siege warfare. They made
more use of artillery, apparently, than did medieval western European armies.
Moreover, the Roman army deployed massive firepower in the field, with each
legion possessing some 55 mobile ballistae, and the in-swinging design allowed for
the construction of lighter and more powerful engines and even hand-held
weapons. This ballistic technology, although available also to some of Rome's
enemies, particularly those in the East, gave the Roman army a key technological
edge and must have been one of the factors in Rome's ability to withstand for so
long repeated attacks on the frontiers with a comparatively small army. In the
eastern Mediterranean, arrow-shooting catapults (including the carroballista and
cheiroballista) survived in use into Byzantine times (Wilkins 2000). The onager
seems to have become more popular in late antiquity and to have survived into the
Middle Ages. Although it was less maneuverable and probably less accurate than
other devices, it was simpler to build and maintain than the two-armed Greco-
Roman palintone stone throwers.
WATER-LIFTING DEVICES
An important early class of machines was for lifting water, to supply places above
where it naturally occurred or to drain underground workings (Schi0ler 1973;
Oleson 1984, 2000). The earliest water-lifting machine was the shaduf, in which a
horizontal lever arm that pivoted on a vertical support had a bucket on one end and
a counterweight on the other. The operator pulled the bucket down into the water
and the counterweight lifted it up so the water could be emptied into a channel or
container. The shaduf conferred no mechanical advantage, but enabled the operator
to pull downward, a more efficient and less tiring motion using stronger muscles
than lifting a weight up. It is shown in an Egyptian wall painting from Thebes of
about 1300 B.C. (Hodges 1970: fig. 112) and mentioned by several classical authors
(White 1975: 45-47; Oleson 2000: 225-29). Lifts are low, but discharge volume is
Figure 13.3. Water-powered water-lifting wheel (noria) at Hama, Syria. (Photograph by

A. Wilson.)
relatively high. The well windlass or pulley also allows a conversion of direction of
force for ease of operation; both systems were common in the Greco-Roman world.
Lifts can be high (limited only by the tensile strength of the rope) but discharge
volume is low, limited by the size of container and the time taken to pull the rope up.
More complicated machines appear in the Hellenistic period, using rotary
motion for continuous operation. The wheel with compartmented body (tympa-
num) was powered by men treading the outside of the wheel; it had a high discharge
but relatively low lift, as the water discharged through a hole near the axle. The
compartmented rim wheel had a higher lift-closer to the diameter of the wheel-
and might also be powered by men treading the rim, or by animals yoked to a
capstan driving the wheel by right-angled gearing (saqiya), or by water turning
paddles on the exterior of the rim (noria) ( figure 13.3). The traditional water-lifting
wheel shown here is very similar to one represented on a fourth-century A.D.
mosaic from nearby Apamea. For higher lifts, a chain of buckets or (later) pots
could be looped over a wheel, powered either by men in a treadmill or by animals
via a geared saqiya drive. An interesting, early example was discovered in London in
2001 ( figure 13.4). A cogwheel on the vertical shaft of a capstan engages a drive wheel
on one end of an axle; on the other end of the axle is a wheel over which is looped a
set of wooden containers held together by cranked iron links. As the containers
come over the top of the wheel, they discharge their contents forward into one of
eight compartments on the outside of the wheel, from which the water then dis-
charges into a trough or launder through side-ports in the compartments. The
original device was probably driven by an animal yoked to a bar on the vertical axle.
Philo of Byzantium's Pneumatics indicates that many of these new machines
were invented in the mid-third century B.C. This work survives only in Arabic, but
Figure 13.4. Working reconstruction of the first-century A.D. water-lifting device dis-
covered at Gresham St, London, on display at the Museum of London. (Photograph by
A. Wilson.)
Lewis (1997: 26-36) shows that letter sequences referring to details on the illus-
trations demonstrate that the chapters on the bucket chain and overshot wheel
certainly, and the noria and the saqiya drive possibly, are in fact translations of the
original Greek. Furthermore, in another clearly genuine section Philo refers to
animal-powered and water-powered lifting machines among the devices with which
one might expect engineers to be familiar, and opposed to which his description of
a siphon system is a novelty (Pneumatics 5; Lewis 1997: 32). More generally, Philo's
Parasceuastica (91.43-44) which survives in Greek, refers to "wheels driven around"
for raising water to flood siege mines. Lewis also gives strong reasons for seeing
Philo as the source for Vitruvius' descriptions of the hodometer, screw, tympanum,
rota, bucket chain, noria, and water-mill (1997: 46).
These water-lifting devices and other machines were therefore already known
when Philo wrote in the later third century B.C. While his Hydragogia and Me-
chanics may not have been written at Alexandria, the subsequent importance of
rotary water-lifting devices in Egypt strongly suggests that these machines, at least,
were known at Alexandria around the mid-third century B.C., and may have been
invented there.
Water-lifting devices like the saqiya and the noria were widely used in later
Hellenistic and Roman Egypt, and although they are usually associated with the
east, there is growing evidence for their use in the western provinces, including
North Africa and probably Spain (Wilson 2003), Italy, Dalmatia, Gaul, and Britain
(Bouet 2005). A saqiya-type bucket-chain was used at Cosa in Italy in the later
second century B.C., and another, using wooden buckets, in the later first century
B.C. (Oleson 1984: 201-2). The earliest of four bucket chains from London, built in
A.D. 63 and used for about 10 years, shows that this technology reached Britain
within 20 years of the Roman conquest (Blair and Hall 2003; Blair 2004; Blair et al.
2005). The London devices used wooden box-buckets linked by a double iron
chain, which must have turned over a facetted axle (figure 13.4). In the use of a
double iron chain the device is similar to Vitruvius' description (De arch. 10.4.4),
with the difference that the buckets were of wood and not bronze; the facetted axle
recalls Philo's bucket-chain (Pneumatica 65) or the repeater catapult (above). The
Cosa and London devices may have been powered by men in a treadmill or by
animals via a saqiya drive. At some point, perhaps between the first and third
centuries A.D., terracotta pots replaced wooden buckets on the saqiya, making the
machines more affordable in regions like Egypt, where wood was scarce and ex-
pensive. Such pots form the most durable and recognizable archaeological evidence
for the use of the saqiya, and document its use in Israel from the late second or
early third century A.D. onward (Ayalon 2000). Papyri and finds of saqiya pots in
Egypt and elsewhere suggest that the use of the saqiya increased during the early
fourth century A.D., possibly as a result of Diocletian's decree of tax relief on
irrigated land (Oleson 1984: 379-80). It seems therefore that uptake of the tech-
nology was first encouraged by design modifications using cheaper materials, and
then by state incentives for investment in irrigation technology with the goal of
increasing agricultural output.
Two other devices are attributed to named individuals-the force pump to
Ctesibius, and the water-lifting screw to Archimedes, although a more recent
theory argues (controversially) that this was an invention of the ancient Near East
which Archimedes merely refined and popularized (Dalley 1994; Dalley and Oleson
2003). The water screw's discharge is high, but the height lifted is low-a fraction
of the length of the barrel. The device was used in irrigation, particularly in Egypt,
and in mine drainage, where batteries of them in series have been found to achieve
the requisite lift (Oleson 1984: 291-301; 2000: 242-51).
The force pump is the most complex ancient lifting device. In its original
(bronze) form, pistons working in a pair of cylinders connected by transverse pipes
pushed water up a central delivery pipe; the water was prevented from returning
under the force of gravity by one-way bronze flap valves at the entrance to a
chamber at the base of the delivery pipe ( figure 13.5). The pistons could be con-
nected by a rocker at the top, and worked reciprocally. The main uses attested in
literary sources were firefighting and spraying perfumed water in theaters and
amphitheaters and during banquets; the bronze pumps found at several archae-
ological sites, often portable and of fairly small capacity, may have been suited to
these purposes. The firefighting version had a nozzle that could swivel and be
raised for directing against a blaze (Hero, Pneumatics 27), and may have been
mounted on a cart. Several bronze pumps from mines may have been used in
fire-setting-spraying cold water on a rock face heated by fire, to make it crack.
In the first century A.D., the force pump was re-engineered in wood, making it
more affordable and easier to produce or repair. It begins to appear at rural villas
and farms, for raising water from wells, and this is the most commonly attested
archaeological use, followed by bilge pumps in wrecks (Oleson 1984; Stein 2004; cf.
'1w,
Figure 13.5. Drawing of Roman bronze force pump from the copper mine at Sotiel
Coronada (Domergue 1990: fig. 35, by permission of C. Domergue.)
Fleury 2005, whose list of pumps is incomplete). At least 32 Roman force pumps are
known: 11 bronze, 1 lead, and 20 wood (Stein 2004, plus the imprint of the wooden
pump recorded by Jones 1980: 95; Oleson 2000: 272-85). At Alet (St. Malo) a late-
first-century A.O. pumping installation included a block with eight cylinders,
worked by two or three men, to pump spring water from just above the shore to a
basin higher up (Langouet and Meury 1976; Oleson 1984: 261-63).
The chain pump was used mainly or exclusively for bilge-pumping in ships,
from at least the start of the first century B.C. and probably earlier. A series of
wooden disks strung on a chain or rope was looped round an upper slotted drive
wheel, and passed over a roller in the bilge (Oleson 2000: 263-67). The upward-
moving disks entered a wooden tube mounted inside of the ship, whose lower end
was immersed in the bilge water. Water entered the tube between disks and was
pulled up the tube after each disk entered, discharging through the gunwales at the
top of the tube. This machine could handle silty water, was self-cleansing, and
could operate in a pitching and rolling ship in a way that other devices could not.
Uncertainty remains, however, over the drive mechanism; a crank-handle on the
upper drive wheel would be the obvious solution, but this has been doubted given
the rarity of evidence for the use of cranks in antiquity. The best evidence for a
possible Roman crank handle, however, is a slotted disk from one of the Nemi
barges (ca. A.D. 40 ), which also produced evidence consistent with a wooden chain
pump (incorrectly reconstructed as a bucket chain at the time), and the crank
handle therefore seems the most likely solution (Drachmann 1973: 37-38; cf. Oleson
1984: 230-33; 2000: 263-67).
The new Hellenistic water-lifting devices revolutionized opportunities for
irrigating land that lay above running water courses. Roman and Byzantine pa-
pyri from Egypt commonly record the hire, construction, or repair of irrigation
devices-although frequently the type is unspecified-and such machines, which
could be expensive to construct, were often an indispensable part of an estate
(Rathbone 1991: 223-24). Although much of the evidence for saqiyas comes from
Egypt, treadmill-driven wheels with compartmented rims have long been known
from bath buildings in Pompeii and Ostia. As a result of recent publications, more
examples are increasingly being recognized and reported by excavators, and the
body of available evidence is growing rapidly (cf. Bouet 2005: 21-26; Bedello Tata
and Fogagnolo 2005; Coadic and Bouet 2005; Conche et al. 2005; Morvillez et al.
2005). The ancient chain pump, previously thought to be a Renaissance invention,
was only recognized during the 1980s, and took a decade or so to become accepted
in the literature on water-lifting devices. Generalized statements about uptake of
water-lifting technology must therefore remain provisional, but already it is clear
that such devices were widely used earlier than previously thought.
WATER-MILLS
The rotary hand quern is an early application of the crank principle to facilitate
rotary or oscillating motion. The rotary mill allowed animals to be yoked to the
mill and to turn it using a lever. The water-mill capitalizes on the automation of
rotary motion by using a flowing stream, and it represents one of the earliest
successful attempts to use natural forces for productive mechanical work. Since the
mid-198os our understanding of the origins and spread of water-milling has been
transformed by a growing body of archaeological evidence from across the Roman
Empire (Wikander 1985, 2000c; Brun and Borreani 1998; Wilson 2001a: 234-36,
2001b, 2002: 9-15.). It now seems that the water-mill was invented in the mid-third
century B.C., with vertical and horizontal versions appearing almost simulta-
neously (Lewis 1997). By the first century A.D. water-mills had spread widely
through the Roman world, with the full variety of vertical wheel types-undershot,
overshot, and breastshot-in use by the late second century A.D. (chapter 6).
Vertical mills require right-angled gearing to convert the vertical rotation about
the water-wheel axle into the horizontal rotation of the spindle passing through the
lower millstone, which supported and turned the upper millstone.
Horizontal mills, by contrast, require no gearing. Water delivered down a chute
or via a nozzle from a tall reservoir (called a drop-tower) struck the inclined blades
of the horizontal wheel almost tangentially and turning it with its spindle, which
supported the upper millstone. This design is suited to channels or streams with a
small or slow flow, as it utilizes the pressure of water in the reservoir rather than the
speed or gradient of a stream. Horizontal mills are attested in third-century B.C.
works of Philo of Byzantium (Lewis 1997), but archaeological evidence is lacking
before the fourth century A.D. Curiously, the earliest identified surviving remains
of horizontal mills, at Chemtou and Testour in Tunisia, indicate the most complex
type of mill, in which water in a tapering wheel race entered a circular wheel shaft
tangentially at the top and exited at the base; a horizontal wheel with angled blades
ran fully submerged at the base of the shaft in a rotating column of water. As a true
turbine wheel, this design was relatively efficient, avoiding dissipation of energy
around the sides of the wheel. The Chemtou and Testour mills were relatively large,
with three wheels each (Roder and Roder 1993; Rakob 1993; Wilson 1995). Together
with the developed design, this sophistication suggests a previous period of de-
velopment for which no archaeological evidence remains. This heritage probably
included the so-called arubah or drop-tower mill, common in the Near East from
the late Roman period onward, and also used probably in the Roman or late Roman
period in North Africa. Water stored in a cylindrical reservoir (drop-tower) several
meters high emerges under pressure through a nozzle at the base, spraying onto the
spokes of a horizontal wheel.
Whatever the drive arrangement, the millstones required careful adjustment.
To control the fineness of the meal produced, and to compensate for wear, the
clearance between the millstones could be adjusted by a tentering mechanism. The
lower end of the spindle was supported on a horizontal beam (bridge tree), pivoted
at one end; a vertical rod attached to the other end allowed the bridge tree to be
raised or lowered slightly, raising or lowering the spindle and thus also the upper
millstone. Archaeological evidence for this arrangement has been found at
Chemtou (Roder and Roder 1993). Grain was delivered to the millstones from a
hopper suspended over them (Vitruvius, De arch. 10.5.2), and by analogy with later
milling practice it is usually assumed that delivery from the hopper was regulated by
a device attached to the upper millstone, which vibrated the hopper as the millstone
rotated, ensuring a continuous and even flow of grain and preventing blockage.
The physical evidence for ancient water-mills usually consists of a millrace
(rock-cut, stone-, or timber-lined), a wheel-pit and a gear-pit; sometimes wooden
elements of the wheel or the wheelrace survive. Nearly all Roman water-mill
structures recognized to date in the Mediterranean are of stone or brick-faced
concrete. North of the Alps, timber construction is more common, with evidence
from Britain, Germany, Switzerland, and northern Italy indicating a developed tra-
dition of engineering headraces and tailraces, with timber linings and revetments,
between the first and fourth centuries A.D. (Castella et al. 1994; Volpert 1997).
Two mill spindles from ironwork hoards at Silchester and Great Chesterford
may suggest the presence of water-mills nearby. But the largest body of evidence is
stray finds of large, power-driven millstones. In Britain alone large millstones of
Roman or probably Roman date have been found on at least 57 different sites
(including seven different locations in London, but excluding the Blackfriars
wreck), and were probably water-powered.
The boat mill or floating mill, in which the machinery is mounted on a boat,
allowing the wheel mounted on the side of the boat to rise and fall with the river
level, was a solution to the problem of operating undershot wheels on large rivers
whose levels fluctuate. According to Procopius (Goth. 5.19.18), Belisarius invented
the floating mill during the siege of Rome in A.D. 537, after the Goths cut the
aqueducts on which Rome's water-mills depended. Whatever the truth of this
account, the invention spread rapidly, and river mills at bridges, probably float-
ing mills, were operating in Paris, Geneva and Dijon from the 550s to the 570s
(Wikander 2000c: 384).
OTHER APPLICATIONS
OF WATER POWER
Besides the water-mill and water-powered noria, water power was also used for
other purposes: dough-mixing, sawmills (figure 6.5), pestles for pounding grain,
and recumbent trip-hammers or vertical ore stamps for crushing ore in mining
regions (figure 4.2; chapter 6; cf. Wikander 1981, 1989, 2000b; Lewis 1997). A relief
discovered at Pammukale in Turkey provides a clear representation of a pair of
stone-cutting saws powered by a water-wheel by means of a crank and connecting
rod (Ritti et al. 2007). Other evidence for the early diversification of water power is
more circumstantial. Lewis (1997: 95-99) argues that the fullers' canal at Antioch,
dated to A.D. 73/74 and known only from inscriptions, could only have been
intended to drive water-powered fulling stocks; and the use of water-powered
bellows has been proposed at Marseille in the fifth century (Guery and Hallier 1987,
273 fig. 5, 274-75; Varoqueaux and Gassend 2001).
Pliny mentions a fractaria machina (breaking machine) incorporating 150 lb
weights, used for attacking outcrops of hard flint in the adits of gold mines in Spain
(HN 33.21.72). It is difficult to see how this could have operated in underground
adits, or how it might have been powered, and it would be tempting to believe that
Pliny has confused this with ore-stamps, were it not for the fact that it is firmly
embedded in his description of workings at the face of the adit. What form the
device took is unknown.
Dough-Kneaders
Roman dough-kneading machines consisted of a cylindrical stone tub with a metal
spindle rotating in a socket in the base; paddles or "arms" projecting from the
spindle passed between rods stuck horizontally through holes in the side of the tub.
The spindle was rotated by a man or a yoked animal pushing a lever, kneading the
dough as it was caught between the paddles and the rods (Mau 1886; Besnier 1877-
1919: 496; Curtis 2001: 363-65). Dough-mixers are common in bakeries at Pompeii
and Ostia, and are found from Volubilis in Morocco (Luquet 1966) to Tomis
(Constanta) in Romania on the Black Sea; none is yet known north of the Alps,
although wooden versions may perhaps have been used there. Their origin remains
uncertain, although a depiction on the Augustan Tomb of Eurysaces gives a rough
terminus ante quern (figure 2.2). Broadly contemporary is Vitruvius' reference to
water-powered dough mixing in the context of a water-mill (De arch. 10.5.2).
Systematic collation of published examples might help to refine the date of the
dough-mixer's introduction, although frequently they remain unpublished despite
their common occurrence on many Mediterranean sites.
TRANSPORT AND PADDLEWHEELS
Vitruvius and Hero describe a hodometer in which gears attached to the axle of a
wagon are arranged to drop a stone ball into a receptacle every mile ( figure 31.4); its
invention perhaps by Archimedes in the third century B.C. may have been con-
nected with the Roman road-building program in Italy at this period (Lewis 2000:
332; see chapter 31). Both Vitruvius (De arch. 10.9) and Hero (Dioptra 38) also
describe a hodometer for ships, driven by a paddle wheel rotating as the ship goes
through the water; this may have been another of Archimedes' inventions (Tzetzes,
Ghil. 12.967-71; Lewis 2000: 332). Drachmann (1963: 159, 165-68) considers this an
armchair invention, but it may be this device that, via a reversal of principles like
that linking the saqiya and the water-mill, inspired one of the most extraordinary
inventions of the Roman world. The anonymous fourth-century author of De rebus
bellicis proposes a ship propelled by three pairs of paddle wheels driven by oxen via
a geared capstan (chapter 17). This proposal is usually dismissed as impractical, and
certainly there is no evidence for its having been used in the Roman world. In-
triguingly, though, from probably the fifth century A.D., and certainly from the
eighth century onward, the Chinese had boats driven by treadmill-powered pad-
dlewheels. Their practical efficacy is shown by their importance in twelfth-century
warfare, when a kind of naval arms race between government forces and southern
rebels led to the construction of boats with between 8 and 24 paddle wheels
(Needham and Ling 1965: 414-35). Needham and Ling consider the transmission of
the ideas of the anonymous author of De rebus bellicis to China as unlikely, and
prefer to see the Chinese paddleboats as independent inventions of about 50 years
later. Similar horse-driven boats were in common use as ferry boats on Lake
Champlain in the eighteenth and nineteenth century A.D. (Crisman 2005), under-
lining the practicality of the concept.
OIL AND WINE PRESSES
To produce olive oil or wine, olives or grapes had to be pressed-in the case of
olives, after a preliminary crushing in trapetum or edge-runner mills (chapter 14).
Although for small quantities this could be done by piling weights directly on the
basket of fruit, usually a mechanical press was used. Ancient oil and wine presses
were based on different simple machines (lever, screw, and wedge), combined in
some cases with the winch or compound pulleys. The lever press was the earliest,
and in its simplest form consisted of a long wooden beam fixed at one end, resting
on the stack of baskets of fruit-either olives that had previously been pulped in a
mill or the grapes and skins that remained after the initial treading. Stone weights
were hung off the free end (an Attic black-figure vase shows a man swinging on the
end of the beam instead). Greater force could be exerted with the lever and
windlass press, using a windlass fixed to the floor of the press-room to winch down
the free end of the beam. Variations had the windlass attached to the beam, or to a
large stone weight; operating the windlass drew the beam down until the resistance
of the fruit stopped it and the weight began to rise off the floor; once the weight had
been winched fully up its weight compressed the stack of fruit further and it began
to settle to the floor.
The lever and screw press was in many respects an improvement using the
mechanical advantage of the screw to apply greater pressure. Pliny the Elder (HN
18.317), writing in the A.D. 70s, mentions this as a Greek invention of the last 100
years (i.e., late first century B.C.); it probably operated with a male screw, but the
female element was probably a simple hole in the press beam with a peg pushed
through the side to engage the male screw. Variants existed in which the screw was
attached to the floor, or was attached to the counterweight to raise it off the floor as
in the lever-and-windlass press. Pliny also says that the direct screw press for
grapes, which is much smaller and lacks the lever, had been invented in the last
22 years, that is, since A.D. 55. This sounds like the small olive-press with one
screw described in Hero's Mechanics 3, whose construction seems to have been
made possible by the invention (by Hero ?) of the screw-cutter to cut female
screws (Mechanics 3.21; Drachmann 1936, 1963). The direct screw press was par-
ticularly favored in urban environments, where space was at a premium, and was
also used by fullers to press clothes. Its drawback was that its operation was more
labor-intensive than the lever-and-windlass press, which once the counterweight

had been raised could run unattended until the stack was fully compressed.
The wedge press was another compact press suited to urban use; it seems to
have been used particularly by perfume makers, who needed high-grade oil as the
base for their scents. Boards held between a pair of uprights over the fruit to
be pressed were forced apart by knocking in wedges between them, compressing the
stack (Mattingly 1990). Besides Philo's wedge catapult, this is one of the few in-
stances of the use of the wedge in ancient complex machines, although the Chinese
had wedge presses too (Cotterell and Kamminga 1990: 100).
ENTERTAINMENT
The water organ (hydraulis) is the first keyboard instrument; invented by Ctesibius,
it is clearly related to his force-pump. Two cylinders worked by levers forced air into
an inverted funnel in a chamber partly filled with water, forcing some of the water
out of the funnel. Because of the constant force of ambient air pressure, the sur-
rounding water maintained a fairly constant pressure within the air space of the
funnel. From the top of the funnel, air was distributed to the organ pipes via a
number of channels, each of which could be opened or closed by valves attached to
a keyboard (Vitruvius, De arch. 10.8; Hero, Pneumatics 76; Drachmann 1948: 7-21).
The hydraulis is represented in Roman terracotta models, mosaics, and reliefs, and
surviving parts have been found at Aquincum, from where there is also a tombstone
of a female singer and hydraulis player (CIL 3,10501), and at Dion. Mosaics from
Nennig (Germany) and Zliten (Libya) show the water organ providing accompa-
nying music at gladiatorial shows and executions of prisoners or criminals by wild
animals. A wind-driven version used cams on a wind-driven wheel to pump a
cylinder up and down (Hero, Pneumatica 77). Apollonius' Fluteplayer (ca. 240 B.C.;
Lewis 1997: 49-57), adapts the principle of water maintaining air pressure but uses a
water-powered wheel to operate valves controlling the water flow into the air
chamber, and tipping spoons to operate the valves that release the air into the flute's
pipe. This device may have been inspired partly by Ctesibius' hydraulis, but it is a
more complex development.
Besides musical instruments, a variety of mechanical devices was used in public
entertainment, especially in Greek and Roman theaters, and Roman amphitheaters
and circuses. The Greek theater machines, which from the fifth century B.C. al-
lowed actors to appear suddenly high up as a deus ex machina (and which are
parodied by implication in Aristophanes' Thesmaphoriazusae, 1009-15), probably
involved little more than a winch and crane. More complex was the mechanism
used in circuses, hippodromes, and stadia to open the starting gates (carceres)
simultaneously. Each stall had two gates held closed by a bar, and when this was
removed the gates flew open instantaneously. The speed with which the gates
opened suggests that they were sprung; the word hysplex, used in an inscription of
280-272 B.C. from Pergamon but referring to the hippodrome of Olympia, and by
Dionysius of Halicarnassus (3.68.3) referring to the Circus Maximus, seems to
imply the use of twisted rope or sinew for the hinges, as in the torsion catapult.
Humphrey reconstructs the arrangement in Roman circuses such as Lepcis Magna,
with ropes running down from a gallery over the carceres to each gate and attached
to a latch to draw the bolt that held the gates closed, and suggests that the ropes to
pull the latches of all 12 gates were released simultaneously by attaching them to a
small catapult (Humphrey et al. 1972-1973: 82-87; Humphrey 1986: 133, 157-70).
In amphitheaters, lifts operated from below by capstans and ropes over pulleys
were used to hoist animal cages up from cells below the arena to near the arena level,
so the animals could emerge through trapdoors. Numerous square shafts in the
arenas of the Colosseum and the amphitheaters at Capua and Pozzuoli illustrate
this arrangement; these and several other amphitheaters also have longer slots in the
arena floor that may have been used for raising scenery by similar mechanisms.
Elements of revolving circular wooden platforms were discovered during the
excavation of Caligula's barges on Lake Nemi. One, nearly a meter in diameter,
revolved on eight small bronze spheres each fitted with an axle; the other was
smaller and turned on wedge-shaped conical wooden rollers (Ucelli 1950: 186-90).
Their function is unknown; bases for cranes or revolving statues have been sug-
gested, but it is also tempting to wonder if they might have been for rotating seats
such as those described in Pertinax' auction of Commodus' property: "There were
also carriages of an unusual technique of construction with meshing and branch-
ing gear trains and seats accurately designed to rotate, sometimes to face away from
the sun, sometimes to take advantage of the breeze, and hodometers, and clocks,
and other items appropriate to his vices." (HA, Pertinax 8.6-7; translation Lewis
2001b: 331).
The role of machines in the social and economic development of antiquity has
perhaps been excessively downplayed in the past. The idea that slavery made ma-
chines unnecessary, or that animism discouraged the exploitation of water power, is
now rejected. Mechanical automata (discussed in chapter 31) were not mere toys:
they elaborated, developed and illustrated mechanical principles, of which the most
significant include early instances of feedback control (Ctesibius' float regulator for
the water-dock and Philo's lamp for constant level: Mayr 1970a, 1970b); pro-
gramming (Hero's theatrical automata), and auto-propulsion (Hero's stages,
whose wheels were driven by a descending weight). Their development was closely
related to the jack-work of clocks and planetaria (Drachmann 1948: 70-71).
In a world without patents, some of the biggest mechanical advances occurred
under royal patronage-the invention of artillery at Syracuse under Dionysius in
399 B.C., the Ptolemaic experiments on ballistae, and also of course the develop-
ment of gearing, pneumatics, and mechanics at the Museum of Alexandria. For the
Roman world, both the anecdote of the engineer and Vespasian (Suetonius, Vesp.
19) and the dedication of De rebus bellicis to the emperor suggest that there was an
expectation that the emperor might reward invention, even if that was not always
realized.
How widespread was the use of machines? For cranes, at least, we can attempt
an answer: from their beginnings in the late sixth century B.C. they had become
common in the Greek world by the late fifth century, and were ubiquitous in the
Hellenistic and Roman worlds. Every public building with monolithic columns or
large blocks with lifting lugs or Lewis holes is indirect but certain testimony to the
use of cranes. In this regard, the use of cranes to lift material in construction appears
more common in the Roman world, where nearly every city possessed several
buildings whose construction and repair required them, than in medieval Europe,
where in many towns the cathedral and the castle might be the only structures
needing them. There is the impression that oil and wine presses dating from the
Roman and late Roman periods are much more frequent than those of subsequent
times until the early modern period; the archaeological (as opposed to docu-
mentary) evidence for water-mills and millstones also appears scarcely less abun-
dant for the Roman than for the high medieval period.
Arguably, the use of machines was in fact more widespread in ancient Greece
and Rome, together with ancient China, than in any other civilization until cer-
tainly the twelfth or perhaps even the fourteenth century A.D. in western Europe.
While most of the simple machines were known to ancient Near Eastern civiliza-
tions, it was the Classical and Hellenistic Greek and the Roman worlds that pro-
duced a variety of complex machines and spread their use both outward geo-
graphically and downward through levels of society, sometimes by re-engineering a
complex design in cheaper materials. There is still much work to do on the question
of the relationship between the early appearance of mechanical technology in
Greece and Rome and in China: independent invention or technology transfer; and
if so, in which direction? Our understanding of many classical machines has ad-
vanced considerably since Needham and Ling addressed the question of the priority
of Chinese and classical mechanical inventions (1965: 366-70, 405-8). Much evi-
dence now points, especially for water-powered machines, to diffusion from the
classical Mediterranean to China, presumably along the Silk Road. The first device
that can be shown to have come from China to the Mediterranean world is the
traction trebuchet, introduced by the Avars, who used it at the siege ofThessaloniki
in A.D. 597 (Nicolle 2003: 10-11).
Older notions of Greek and Roman technological stagnation contrasting with
the technological vibrancy of the early Middle Ages, which saw a kind of "medieval
industrial revolution," must now be rethought. Not only does it appear that, for
example, in the Roman world water power was more widespread, earlier, and used
for more varied purposes than was believedeven in the 1970s, but reassessment of
the medieval evidence now suggests that the diversification of water power for uses
other than grain-milling and fulling was a relatively late (fourteenth- or fifteenth-
century) phenomenon (Lucas 2005). The technological gap between the Roman
Empire and the early Middle Ages looks smaller than it did before. Indeed, there are
areas (pressing, artillery), where Roman machine use may have been in advance of
early medieval use. The implications for our views of progress and of the rela-
tionship between technology and economic development are profound.
REFERENCES
Anonymous 2004. "Trajan's artillery: The archaeology of a Roman technological revolu-

tion," Current World Archaeology 3: 41-48.
Anstee, J. 1998. "'Tours de force': An experimental catapult/ballista," Studia Danubiana
1998: 131-39.
Ayalon, E. 2000. "Typology and chronology of water-wheel (saqiya) pottery pots from
Israel," Israel Exploration Journal 50: 216-26.
Baatz, D. 1978. "Recent finds of ancient artillery," Britannia 9: 1-17.
Baatz, D. 1994. Bauten und Katapulte des romischen Heeres. Mavors Roman Army Re-
searches 11. Stuttgart: F. Steiner.
Baatz, D. 1999. "Katapulte und mechanische Handwaffen des spatromischen Heeres,"
Journal of Roman Military Equipment Studies 10: 5-19.
Ballu, A. 1911. Les ruines de Timgad: Sept annees de decouvertes, 1903-1910. Paris: Neurdein
Freres.
Bedello Tata, M., and S. Fogagnolo 2005. "Una ruota idraulica da Ostia," in Bouet 2005:
115-38.
Besnier, M. 1877-1919. "Pistor, pistrina," in DarSag 4.1: 494-502.
Blair, I. 2004. "The water supply and the mechanical water-lifting devices of Roman
Londinium-'New evidence from three archaeological sites in the City of London,'"
in F. Minonzio (ed.), Problemi di macchinismo in ambito romano: Macchine idrauliche
nella lettatura tecnica, nelle fonti storiografiche, e nelle evidenze archeologiche di eta
imperiale. Como: Commune di Como, 113-24.
Blair, I., and J. Hall 2003. Working water: Roman technology in action. London: Museum of
London.
Blair, I., R. Spain, and T. Taylor 2005. "The technology of the 1st- and 2nd-century Ro-
man bucket chains from London: From excavation to reconstruction,'' in Bouet 2005:
85-114.
Bouet, A. (ed.) 2005. Aquam in altum exprimere: Les machines elevatrices d'eau dans
l'Antiquite. Scripta Antiqua 12. Pessac: Ausonius.
Brun, J.-P ., and M. Borreani 1998. "Deux moulins hydrauliques du Haut Empire romain en
a
Narbonnaise: Villae des Mesclans La Crau et de Saint-Pierre/Les Laurons aux Arcs
(Var),'' Gallia 55: 279-326.
Campbell, D. B. 2003. Greek and Roman artillery, 399 BC-AD 363. Oxford: Osprey Pub-
lishing.
Castella, D., E. Beza, and P.-A. Bezat 1994. Le moulin hydraulique gallo-romain d'Avenche en
Chaplix. Aventicium 6. Lausanne: Cahiers d'archeologie romande.
Cherrette, M. 2001-2002. "The onager according to Ammianus Marcellinus," Journal of
Roman Military Equipment Studies 12-13: 117-33.
a
Coadic, S., and A. Bouet 2005. "La chaine godets des thermes de Barzan (Charente-
Maritime): Une premiere approche," in Bouet 2005: 31-44.
Conche, F., E. Plassot, and C. Pellecuer 2005. "Puiser, elever et distribuer l'eau dans la villa
a
de Careiron et Pesquier Milhaud (Gard): Premiers commentaries," in Bouet 2005:
69-84.
Cotterell, B., and J. Kamminga 1990. Mechanics of pre-industrial technology. Cambridge:
Coulton, J. J. 1974. "Lifting in early Greek architecture," Journal ofHellenic Studies 94: 1-19.
Crisman, K. 2005. "A horse-powered ferry: Burlington Bay, Lake Champlain," in G. F. Bass
(ed.), Beneath the seven seas. New York: Thames and Hudson, 218-19.
Curtis, R. I. 2001, Ancient food technology. Leiden: Brill.
Dalley, S. 1994, "Nineveh, Babylon and the hanging gardens: Cuneiform and classical
sources reconciled," Iraq 56: 45-58.
context of invention in the ancient world." Technology and Culture 44: 1-26.
de Solla Price, D. 1974. "Gears from the Greeks: The Antikythera Mechanism, a calendar
computer from ca. 80 B.C.," Transactions of the American Philosophical Society 64: 1-70.
Drachmann, A.G. 1936. "Heron's screwcutter," Journal of Hellenic Studies 56: 72-77.
Drachmann, A.G. 1948. Ktesibios, Philon and Heron: A study in ancient pneumatics.
Copenhagen: Munksgaard.
Drachmann, A. G. 1956. "A note on ancient cranes," in C. Singer, E. J. Holmyard,
A. R. Hall, and T. I. Williams (eds.), A history of technology, vol. 2. Oxford: Oxford
Drachmann, A. G. 1963. The mechanical technology of Greek and Roman antiquity: A study
of the literary sources. Copenhagen: Munksgaard.
Drachmann, A. G. 1973. "The crank in Graeco-Roman antiquity," in M. Teich and R. Young
(eds.), Changing perspectives in the history of science. London: Heinemann, 33-51.
Faccenna, D. 2003. "I rilievi," in A. Campanelli and S. Agostini (eds.), La Collezione
Torlonia di antichita del Fucino. Pescara: Carsa, 69-79.
Field, J. V., and M. T. Wright 1985. Early gearing: Geared mechanisms in the ancient and
medieval world. London: Science Museum.
a
Fleury, P. 2005. "La pompe pistons dans l'antiquite," in Bouet 2005: 139-51.
Frau, B. 1987. Tecnologia greca e romana. Pubblicazioni dei gruppi archeologici d'Italia
Monografie 29). Rome: Gruppo archeologico romano.
Gille, B. 1980. Les mecaniciens grecs: La naissance de la technologie. Paris: Seuil.
Giuliani, C. F. 2003. "La rappresentazione degli argani," in A. Campanelli and S. Agostini
(eds.), La Collezione Torlonia di antichita del Fucino. Pescara: Carsa, 81-82.
Guery, R., and G. Hallier 1987. "Reflexions sur les ouvrages hydrauliques de Marseille
antique retouves sur le chantier de la Bourse," in A. De Reparaz (ed.), L'eau et les
hommes en Mediterranee. Paris: CNRS, 265-82.
Hill, D. 1984. A history of engineering in classical and medieval times. London: Croom Helm.
Hodges, H. W. M. 1970. Technology in the ancient world. New York: Alfred A. Knopf.
Humphrey, J. H. 1986. Roman circuses: Arenas for chariot racing. Berkeley: University of
California Press.
Humphrey, J. H., F. B. Sear, and M. Vickers 1972-1973. "Aspects of the circus at Lepcis
Magna," Libya antiqua 9-10: 25-97.
Iriarte, A. 2000. "Pseudo-Heron's Cheiroballistra, a(nother) reconstruction: I. Theoretics,"
Journal of Roman Military Equipment Studies n: 47-75.
Iriarte, A. 2003. "The inswinging theory," Gladius 23: m-40.
Jackson, R. 1990. "Roman doctors and their instruments: Recent research into ancient
practice," Journal of Roman Archaeology 3: 5-27.
Jacobi, H. 1912. "Romische Getriedemiihlen," Saalburg-Jahrbuch 3: 75-95.

Jones, G.D. B. 1980. "11 Tavoliere romano: L'agricoltura romana attraverso l'aerofotografia
e lo scavo," Archeologia classica 32: 85-100.
Landels, J. G. 2000. Engineering in the ancient world. Rev. ed. Berkeley: University of
California Press.
Langouet, L., and J.-L. Meury 1976. "Les elements de la machinerie gallo-romain d'Alet,"
Les Dossiers du Centre regional archeologique d'Alet 4: 113-25.
Lewis, M. J. T. 1985. "Review of D. Hill, A history of engineering in classical and medieval
times," Journal of Roman Studies 75: 260-62.
Lewis, M. J. T. 1993. "Gearing in the ancient world," Endeavour 3: 110-15.
Hull.
Lewis, M. J. T. 2000. "Theoretical hydraulics, automata and water clocks," in Wikander
2000a: 343-69.
Lewis, M. J. T. 2001a, "Railways in the Greek and Roman world," in A. Guy and J. Rees
(eds.), Early railways: A selection of papers from the First International Early Railways
Conference. London: Newcomen Society, 8-19.
Lewis, M. J. T. 2001b. Surveying instruments of Greece and Rome. Cambridge: Cambridge
University Press.
Longfield-Jones, G. M. 1986. "A Graeco-Roman speculum in the Wellcome Museum,"
Medical History 30: 81-89.
Lucas, A. R. 2005. "Industrial milling in the ancient and medieval worlds: A survey of
the evidence for an industrial revolution in Medieval Europe," Technology and Culture
46: 1-30.
Luquet, A. 1966. "Ble et meunerie aVolubilis," Bulletin d'archeologie marocaine 6: 301-16.
Marsden, E.W. 1969. Greek and Roman artillery: Historical development. Oxford: Oxford
University Press.
Marsden, E. W. 1971. Greek and Roman artillery: Technical treatises. Oxford: Oxford
University Press.
Mattingly, D. J. 1990. "Paintings, presses and perfume production at Pompeii," Oxford
Journal of Archaeology 9.1: 71-90.
Mau, A. 1886. "Su certi apparecchi nei pistrini di Pompei," Mitteilungen des Deutschen
Archiiologischen Instituts, Romische Abteilung 1: 45-48.
Mayr, 0. 1970a. The origins of feedback control. Cambridge, MA: MIT. Press.
Mayr, 0. 1970b. "The origins of feedback control," Scientific American 223: 110-18.
Moritz, L.A. 1958. Grain mills and flour in classical antiquity. Oxford: Oxford University
Press.
Morvillez, E., P. Chevalier, J. Mardesic, B. Pender, M. Topic, and M. Causevic 2005. "La
noria decouverte a proximite de 'L'oratoire A,' dans le quartier episcopal de Salone
(mission archeologique franco-croate de Salone)," in Bouet 2005: 153-69.
Needham, J., and W. Ling 1965. Science and civilisation in China. Vol. 4. Cambridge:
Nicolle, D. 2003. Medieval siege weapons (2): Byzantium, the Islamic World and India, AD
476--1526. Oxford: Osprey Publishing.
Oleson, J.P. 1983. "A Roman sheave block from the harbour ofCaesarea Maritima, Israel,"
International Journal of Nautical Archaeology 12: 155-70.
Oleson, J.P. 2000. "Water-lifting," in Wikander 2000a: 217-302.
Rakob, F. 1993. "Der Neufund einer romischen Turbinenmiihle in Tunesien," Antike Welt
24.4: 286-87.
Rathbone, D. W. 1991. Economic rationalism and rural society in third-century AD Egypt: The
Heroninos Archive and the Appianus Estate, Cambridge: Cambridge University Press.
Ritti, T., K. Grewe, and P. Kessner 2007. "Stridentes trahens per leviam marmora serras: A
Roder, J., and G. Roder 1993. "Die antike Turbinenmiihle in Chemtou," in F. Rakob (ed.),
Simitthus, vol. 1. Mainz: von Zabern, 95-102.
Schi0ler, T. 1973. Roman and Islamic water-lifting wheels. Odense: Odense University Press.
Shaw, J. W. 1967. "A double-sheaved pulley block from Kenchreai," Hesperia 36: 389-401.
Skinner, F. G. 1954. "Measures and weights," in C. Singer, E. J. Holmyard, and A. R. Hall
(eds.), A history of technology, vol. 1. Oxford: Oxford University Press, 774-84.
Stein, R. 2004. "Roman wooden force pumps: A case study in innovation," Journal of
Ucelli, G. 1950. Le navi di Nemi. Rome: Libreria dello Stato.
Varoqueaux, C., and J.-M. Gassend 2001. "La roue a aubes du grand bassin de la Bourse a
Marseille," in J.-P. Brun, P. Jockey, M.-C. Amouretti (eds.), Techniques et sodetes en
Mediterranee. Paris: Maison mediterraneenne des sciences de l'homme, 529-49.
Volpert, H.-P. 1997. "Die romischer Wassermiihle einer villa rustica in Miinchen-Perlach,"
Bayerische Vorgeschichtsblatter 62: 243-78.
White, K. D. 1975. Farm equipment of the Roman world. Cambridge: Cambridge University
Press.
Wikander, 6. 1981. "The use of water-power in classical antiquity," Opuscula romana 13:
91-104.
Wikander, 6. 1985. "Archaeological evidence for early water-mills-an interim report,"
Wikander, 6. 1989. "Ausonius' saw-mills-once more," Opuscula romana 17: 185-90.
Wikander, 6. (ed.) 2000a. Handbook of andent water technology. Leiden: Brill.
Wikander, 6. 2000b. "Industrial applications of water-power," in Wikander 2000a: 401-10.
Wikander, 6. 2000c. "The water-mill," in Wikander 2000a: 371-400.
Wilkins, A. 1995. "Reconstructing the cheiroballistra," Journal of Roman Military Equip-
ment Studies 6: 5-59.
Wilkins, A. 2000. "Scorpio and cheiroballistra," Journal of Roman Military Equipment
Studies 11: 77-101.
Wilson, A. I. 1995. "Water-power in North Africa and the development of the horizontal
Wilson, A. I. 2001a. "Water-mills at Amida: Ammianus Marcellinus 18.8.11," Classical
Quarterly 51.1: 231-36.
Wilson, A. I. 2001b. "The water-mills on the Janiculum," Memoirs of the American Academy
in Rome 45: 219-46.
92: 1-32.
Wilson, A. I. 2003. "Classical water technology in the early Islamic world," in C. Bruun
and A. Saastamoinen (eds.), Technology, ideology, water: From Frontinus to the Re-
naissance and beyond. Acta Instituti Romani Finlandiae 31: 115-41. Rome: Institutum
Romanum Finlandiae.
PART IV
SECONDARY
PROCESSES AND
MANUFACTURING
CHAPTER 14
FOOD PROCESSING
AND PREPARATION
ROBERT I. CURTIS
THE Mediterranean Sea and surrounding lands provided the Greeks and Romans
with a wide variety of foods, both plant and animal. Ancient agricultural writers,
satirists, physicians, and "food specialists" record the types of foods consumed
(Andre 1981; Humphrey et al. 1998: 147-71). Archaeological evidence of food re-
sources, increasingly supported by their scientific examination, augments much of
what ancient literary sources tell us. Also informative are modern studies that look
at Greco-Roman food from different perspectives, such as food and politics and
food in daily life. Food technology, however, apart from a few specialist studies, has
rarely formed the object of serious investigation, usually receiving attention only as
part of broader studies of agriculture or of ancient technology generally (e.g.,
Forbes 1955-1964; Moritz 1958; White 1984; Curtis 1991; !sager and Skydsgaard 1992).
Not until 2001 did a comprehensive study of the subject appear (Curtis 2001), now
complemented by Thurmond (2006).
Food technology, as the term applies today, is the practical application of the
principles of science and technology to the preparation, processing, preservation,
storage, packaging, and transportation of food. Although agriculture is closely
allied to food processing, they are distinct activities. All actions associated with
growing, tending, and harvesting produce fall clearly under agriculture. Some foods,
such as grapes and olives, often receive further treatment after harvesting, and
cereals are only selected for their various uses after they have been separated from
the stalk through threshing. Food technology, therefore, begins at the point when
these plant foods undergo specific processes to change their form in some way (such
as milling grain into flour) or to so alter them as to create a different food prod-
uct altogether (for instance, transforming the grape into wine). In other words, for
370 SECONDARY PROCESSES AND MANUFACTURING
the grape and olive the division between agriculture and food technology lies in
the press room; for grain, it is on the threshing room floor. The care and feeding
of animals belong to animal husbandry, as does the milking of cows, sheep, and
goats. Producing butter, cheese, or yogurt from milk, however, falls into the realm
of food processing. So also does butchery of animals preparatory to preservation of
their meat. And, finally, methods and tools for catching fish are not part of food
processing, but what one does to preserve the fish for later consumption clearly
qualifies.
In antiquity, as today, foods fell broadly into two categories. First were those
that could be eaten fresh as they came to hand. These included both wild and
cultivated plants, such as fruits, nuts, and certain vegetables, and animal by-
products, including milk and honey. Certain of these foods might, as necessary,
require some preparation, such as washing, cutting, and cooking-be it roasting,
boiling, baking, frying, and so forth-before consumption. The second group re-
quired more sophisticated processing to make them edible, processes so extensive
as to result in significant changes in physical appearance or chemical composition
from its original state. So, for instance, cereals, especially wheat and barley, in some
places, like northern Europe, underwent extensive processing to convert the grain
into a fermented beverage (beer), but Greeks and Romans more often ground them
into flour to make porridge and, especially, to produce bread (figure 14.1). Neither
the Greeks nor the Romans liked or produced beer, considering it much inferior to
wine and the drinking of it a characteristic of barbarians, such as Thracians and the
Celtic populations of Spain, northern Europe, and Britain. In the imperial period,
Roman soldiers, especially those of Celtic background, stationed in northern
Europe and Britain consumed it as part of their rations, and beer was apparently a
major product traded in the northern provinces (Curtis 2001: 131-34, 370-71; Nelson
2001: 95-178; 2005: 9-65).
In a few instances, certain foods fell into both broad groups; that is, consumers
had the choice of eating the item raw or cooked or of processing it to create a
distinctly different product. These include the grape, which could be eaten fresh or
dried (as raisins) or, after treading and pressing, consumed as a drink either directly
in nonalcoholic form or fermented into wine. The olive could be eaten only after
undergoing curing with salt water to remove its bitter taste but often underwent
crushing and pressing to extract its oil (figure 14.2). These three agricultural
products-cereals, grape, and olive-all requiring extensive processing before be-
ing consumed in their most popular form, were the most important staple foods in
the classical diet and formed what is commonly called the Mediterranean Triad
(Sarpaki 1992b; !sager and Skydsgaard 1992: 19-40). Certain animals or animal by-
products also required processing before being eaten. This included butchery to
reduce complete animals, such as pigs, cattle, and fish, into smaller pieces of meat
that could be cooked or processed further in some fashion for later consumption
and the processing of dairy products to produce butter and cheese.
Plants and animals underwent processing for several reasons. First, some foods
in their natural state, such as olives, some legumes, and certain tubers, are inedible,
FOOD PROCESSING AND PREPARATION 371
CEREALS
I \
HUSKED FREE THRESHING
(barley, emmer) (bread wheat)
l
Parching
Pounding
monar-pesue
a)msnua/
b) mechanical (?)
:~:; muscle
! Saddle quern
Olynlhus mill
(Del1an mill?)
Grinding Malting
r,. , .
Mashing
Rotary hand mill
animal power { Rotary donkey mill I

Baking Brewing
power of nature { Water mill
Sieving
/ ~
Kneading Cooking Fermentation
a)manual a) sponlaneous
b) mechanical
Baking
l
Filtration (?)
b) artificial
- PORRIDGE
Figure 14.1. Cereal processing, diagram. -

indigestible, or even unsafe for human consumption. Second, some foods, although
capable of being eaten in their natural state, were more popular in a completely
different form. Bread from grain and wine from grapes have already received
mention. Third, many foods in their primary state have a short shelf life and must
be preserved in some fashion to extend their keeping potential. Such is the case with
raisins from grapes, cheese from milk, and cured olives. And, finally, many foods
received processing for social, economic, and political reasons. Not every level of
Greco-Roman society had equal access to every kind of food. Furthermore, not all
foods were available at all times of the year, nor were all foods available in every
market throughout the Mediterranean basin. Additionally, recurrent food crises
characterized life in classical antiquity. Farmers grew crops and raised livestock,
Pressing
(Tota1lartum)
/
1. Must left in presence of 1. Levar-and-weight
the mare yiekls red wine
2. "Improved" lever-and-weight
2. Must separated from
3. Levar-and-drum
mare yields white wine
4. "Improved" lever-and-drum
5. Levar-and-screw
I 6. Screw press
7. Wedge Jllff•
\ Separation
l oil
Water
(amurca)
l
Bottling Bottling
l l
Amph«ae.
Skins, Baffels
Storage and Transportation Storage and Transportation
Figure 14.2. Grape and olive processing, diagram.
consuming some products, preserving others, and sending the surplus to local
markets. The potential for both short- and long-term storage, therefore, was critical
to ensure that some foods were available during periods of the year when others
were not, and that a reserve was on hand to offset crop failures. On a larger scale,
political motivations of the state and economic incentives of individuals or cor-
porate entities also contributed to the decision to process food. Natural foods could
travel short distances, but long-distance transportation, whether for reasons of state
or private enterprise, required food in stable form.
Through the 1970s, a rather pessimistic picture of ancient technological ad-
vancement, best articulated by M. I. Finley, carried the day among historians
( Finley 1959, 1965; cf. Reece 1969, Pleket 1973). This vision attributed factors in-
hibiting progress to the effects of slavery, a general contempt for labor, especially
among the elite, a natural propensity to conservatism, and lack of adequate energy,
raw materials, financial resources, and technical curiosity (Oleson 1984: 398-403;
A. Wilson 2002: 2). Since the 1908s, however, this view has undergone significant
revision. Innovation has replaced invention as a more important indicator of
technological progress; the absence of an "ancient industrial revolution" no longer

implies stagnation; and gradual improvements to machines and application of new
technology create over time a range of choices, arrayed, as it were, on a "tech-
nological shelf" available for use depending on the economic resources of the user
and the type and scale of the task envisioned (White 1984; Houston 1989; Greene
1990, 1993, 2000; A. Wilson 2002). This change in approach is seen particularly
clearly in the area of ancient food processing, a field that even Finley recognized, at
least in regard to bread-making, was an exception to his otherwise negative view of
ancient technology (finley 1959: 120; 1965). More recent studies of archaeological
material, much of it available only since Finley's day, have shown this to be the case
in other areas of food technology as well.
Ancient food technology does not lend itself well to discussion of its sepa-
rate principles, since several technological applications are often relevant to a
single product. So, for instance, grain milling could involve four distinctly differ-
ent processes: parching or roasting, pounding, grinding, and sieving (figure 14.1).
Wine production entails not only treading and pressing of grapes but also fer-
mentation of the grape juice (figure 14.2). To concentrate on each activity or
principle separately would, I believe, result in a loss of understanding of the whole,
thereby giving a false impression of the technology itself. Consequently, a better
approach is to focus on foods that required processing to distinguish what tools and
machines Greeks and Romans used to process them, to determine what principles
these tools reflect, and to evaluate their successes and failures.
CEREAL PROCESSING
Cereals, principally barley and wheat, provided an estimated 70 to 75 percent of the

calories in the classical diet (foxhall and Forbes 1982: 74). Both required processing
before consumption, but the degree varied according to the species. Husked cereals,
such as barley and the glume wheats, emmer and einkorn, require special prepa-
ration beyond threshing before they yield their grain. This includes either pounding
in a stone mortar with a wooden pestle or roasting or parching to render it brittle
before pounding to release the kernel. This extra work explains the popularity,
especially among Romans, of free-threshing cereals, such as durum and bread
wheat, that release their grain with threshing alone. For these, once threshed and
sieved, grinding proceeded directly. Regardless of the advantage of wheat, barley
remained an important cereal for Greeks and Romans, since it grows more easily
and has a higher yield under adverse conditions of decreased rainfall and increased
heat. Consequently, the mortar and pestle, the basic pounding tools, served their
purpose so well that they remained in use almost unchanged throughout antiquity,
even while grinding methods underwent significant development (Moritz 1958:
xx-xxv; Foxhall and Forbes 1982: 75-81; !sager and Skydsgaard 1992: 21-23). The
assertion that Pliny (HN 18.97) refers to a water-powered pestle acting in the
manner of a trip hammer remains debatable (Lewis 1997: 101-5; Curtis 2001: 280).
Operation of the Neolithic saddle quern to grind grain involved moving a hand
stone over a lower stone in a somewhat circular motion to separate off the chaff and
to crush the endosperm. Exceedingly labor intensive, this method produces a meal
that, once sieved, yields flour to make porridge or bread. Like the mortar and pestle,
the saddle quern remained a basic home processing tool well into the Hellenistic
period, undergoing only slight alteration. A thinner, flatter lower stone with stri-
ations cut into the grinding surface made it more portable and efficient, while a
larger hand stone altered the grinding motion to a more back-and-forth, or re-
ciprocal, pattern (Curtis 2001: 264, 280-81).
The first significant innovation in milling was the introduction of the hopper
or Olynthus mill. Known in the Greek world by the late fifth century B.C., it spread
rapidly into the western Mediterranean, especially Sicily and southern France
(Curtis 2001: 282-87; Frankel 2003: 7-8) Developed from the saddle quern, the
Olynthus mill consisted of rectangular-shaped lower and upper stones placed on
an elevated platform. The upper stone had cut into it a hollow cavity, or hopper,
at the bottom of which was a narrow slot through which the miller fed the grain.
The miller crushed the grains falling between the upper and lower stones by mov-
ing the top stone in a reciprocal motion from side to side. He did this with a long
handle attached in a groove on the upper stone and anchored at its short end ei-
ther by a pivot or by insertion into a slot in a wall. The Olynthus mill, with move-
able parts, was the first milling machine, although its motive power remained
human (Moritz 1958: 42-52; Amouretti 1986: 140-42) Its operation approached that
of modern mills in that it controlled its own feeding by means of the hopper and
automatically discharged the processed meal by way of grooves cut into the mill-
ing surfaces of both stones.
Advantages of the hopper mill were twofold. First was its labor-saving benefit.
The mill, placed on a flat, raised surface and operated by a fixed handle moved from
side to side in wide sweeps, was ergonomically easier to operate by one person or
two individuals working in unison. Second, the hopper mill increased production
per unit of time expended and did so more efficiently. The handle not only in-
creased the mechanical advantage and reduced the effort needed to grind the grain,
but also allowed for use of a larger, heavier upper stone with wider grinding area,
resulting in greater output of milled grain. Feeding kernels through the hopper
meant that the miller did not have to stop as often to add more kernels. Overall, the
result was greater output more rapidly achieved with less energy expended than
could be done with the saddle quern. Increasing output through mechanical in-
novation (addition of handle and hopper) constituted a significant departure from
the approach characteristic of the Near East, Egypt, and Bronze Age Greece that had
sought higher output through increasing the number of workers using traditional
technology (Sarpaki 1992a; Curtis 2001: 131, 201-2, 265) The hopper mill, however,
had its limitations. It was less mobile than the saddle quern and lacked provisions
for adjusting the gap between the stones and so for varying the fineness of the meal
produced.
The application of rotary motion to milling was a landmark innovation in food
processing technology. Although its use in olive processing almost certainly pre-
dated its employment in grain milling, rotary motion saw its greatest extension in
cereal processing. How the idea developed is unknown, but it may have had its
origins in observation of the potter's wheel and the development of iron tools that
significantly increased the ability of stoneworkers to fashion round millstones
(Moritz 1958: 103-21; Alonso Martinez 1997; Curtis 2001: 287-88, 336-41). Perhaps
applied first to hand mills in Spain in the fifth century B.C., rotary milling spread
from there into other parts of the western Mediterranean and northern Europe,
probably by Roman soldiers, reaching Greece only in the first century B.C. (Runnels
1990 ). A fragmentary stone discovered in Carthage and identified as part of a round
hand mill, however, may imply a specifically Carthaginian origin for hand-powered
rotary mills, dating perhaps to the sixth century B.C. (Morel 2001). The earliest
rotary grain mill in Greece was possibly the rather large third-century B.C. Delian
mill whose operational design looked more to the trapetum, used to crush olives,
than to other known rotary grain mills, human or animal powered. It seems to have
been an evolutionary dead end, lacking both predecessors and descendants (Brunet
1997; Curtis 2001 288-89).
In its most developed form, the hand mill had a lower stone (meta) of slightly
convex shape with a socket in its center to accommodate a wooden or iron spindle.
On top of this sat the upper stone (catillus) whose bottom was slightly concave. This
stone had a hole through its center and a rynd to accommodate the spindle of the
meta, a shallow, flat-bottomed hopper at top center, and a hole, or metal band, at or
near the periphery for placement of a vertically inserted wooden handle. The miller
fed grain through the hopper and turned the handle of the catillus to grind the
kernels as they fell between the two stones (cf."Vergil," Moretum 16-29).
The advantages of the rotary hand mill over previous mills are fourfold.
Smaller, portable like the saddle quern that it replaced, and probably cheaper than
the hopper mill, the rotary hand mill was no doubt more readily available to people
of all social and economic classes. Second, operation of the rotary hand mill from a
sitting or standing position by grasping the handle with extended arm proved easier
and more efficient than the equivalent operations for either the saddle quern or
hopper mill. Third, the rynd-and-spindle arrangement permitted the miller to ad-
just the distance between the meta and ca till us, thereby allowing production of meal
of varying fineness, which, after one or more sievings, produced flour of vary-
ing grades. The advantages of the rotary hand mill probably account for the lack
of spread of the Olynthus mill into Spain and northern Europe, and the rapid
replacement of the latter by the former throughout the western Mediterranean
generally. Finally, rotary motion, unlike reciprocal motion, allows for continuous
action and ultimately permits the development of animal and water-driven devices.
It is not known what precipitated a shift in milling from the simple rotary hand
mill with small stones to a mill of much larger size and capacity incorporated into a
wooden frame and operated by animals, but a relationship with rotary mills used in
Spanish lead and silver mines has been suggested (Domergue et al. 1997). Although
perhaps ultimately of fourth-century B.C. Punic origin, the animal-driven rotary
mill, termed "Pompeian mill" from the location of the best known examples,
appeared in Italy a little over a century later. Its adoption in Italy may be associated
with the rise of commercial bakeries at about the same time (Moritz 1958: 74-96;
Curtis 2001: 341-48)
The Pompeian mill ( figure 14.3), driven by two donkeys harnessed to a wooden
frame attached to the mill, comprised the usual two parts, meta and catillus, both
made of rough vesicular basalt. The former, with a rounded top cut to accom-
modate a spindle at its apex, was embedded in a rubble foundation. The monolithic
catillus had the form of two cones stacked apex to apex, creating an hourglass shape.
The interior of each cone was hollow, and a hole ran through the part where the
apexes met to accommodate the spindle affixed in the top of the meta. The upper
cone of the catillus served as a hopper while the lower cone fit over the meta. When
rotated by animals, the exterior surface of the meta and the interior surface of the
cone fitting over it ground the grain between them. After a period of operation, the
grinding surfaces would rub smooth. At that point the catillus could be reversed to
make the upper cone the grinding surface, thereby doubling its working life (Moritz
1958: 77).
The major advantage of the animal-driven rotary grain mill was that it released
humans almost completely from the drudgery of the milling process. Since the
animal performed the heavy work, the miller had only to feed grain into the hopper
as necessary and collect, then sieve, the meal to produce the flour. In addition, since
donkeys, with their greater stamina, could drive the mill for longer periods of time
than a human, flour output increased significantly over what previous mills pro-
duced. Given the scale and expense of the Pompeian donkey mill, it is not sur-
prising that most of them are found in the context of commercial bakeries ( figures
14.3, 15.1). Nevertheless, in addition to the health and condition of the donkey,
technological limitations restricted output. Inadequate harnessing of the animals
may have not only shortened the amount of time the animal could work but also
curbed the speed of the catillus in grinding the grain, thereby reducing the flour
output from what would otherwise have been possible. The second-century novelist
Apuleius provides a vivid picture of the poor conditions in milling establishments
(Met. 7.15, 9.10-13).
The most dramatic advance in milling came with the application of water
power, an event that even Finley called "radical." Seeing no evidence that its use was
anything more than sporadic, however, he dismissed it as ultimately an opportunity
lost (Finley 1965: 29). More recent studies on ancient water technology, however,
have shown that in both scale and quantity water-powered grain mills played a
much more important role in ancient cereal processing than previously thought
(Leveau 1996; R. J. A. Wilson 1996; Wikander 2000: 371-410; Curtis 2001: 348-58; cf.
Figure 14.3. Pompeian donkey mills and oven (furnus) in bakery at Reg. VII.2.22. (Pho-
tograph: Sommer 1224).
chapters 6 and 13). Although the water mill may have arisen in the East, the earliest
clear evidence of it comes from first-century A.D. Aventicum in Switzerland, where
carbonized grains of barley, wheat, and other cereals found in the vicinity of a mill
site indicate its use to process grain (Castella 1994; Lewis 1997: 58-61; Wikander
2000: 394-97, 644-45). The best preserved examples include the early-second-
century A.D. vertical overshot water-mill complex at Barbegal in southern France
and the undershot mills constructed during the following century in Rome on the
Janiculum Hill (Moritz 1958: 122-39; A. Wilson 1995; Leveau 1996; Wikander 2000:
373-78). Vitruvius (De arch. 10.5.2) provides the only ancient technical description
of a water-mill, in his case undershot. The number of identifiable water-mills dating
between the first and third centuries A.D. clearly argues for its early integration into
the Roman economy (figures 6.1-6.3). As of 2002, more than 60 water-mills had
been identified from archaeological remains, and epigraphic sources record others
that operated in the eastern Mediterranean (Leveau 1996; A. Wilson 2000, 2002: 9-
15; Wikander 2000: 397-98). For a full discussion of water-mill design, see chapters 6
and 13.
The major breakthrough in milling technology derives from the use of water, a
natural force, to operate the mill; the actual grinding continued to be accomplished
through rotary motion applied to the upper millstone. The force of the water
striking the paddle of the wheel either directly operated the spindle to drive the
catillus (horizontal water-mill) or indirectly transferred that power "geared up"
through a system of shafts and cogwheels to the spindle that turned the catillus at a
faster speed (vertical water-mill). The advantages were immediately clear. First,
water not only applied a stronger force than could be achieved by human or animal
power and so allowed for use of much larger millstones, but also the "gearing-up"
arrangement of vertical water-mills attained a higher rate of milling. Large mill-
stones moving more rapidly significantly increased product output over that
possible with either hopper mill or donkey mill. Second, although most mill
buildings were small and utilized a single water wheel, the use of a stream, river, or
aqueduct allowed for the placement of a series of mills along the water course, such
as was done on the Janiculum Hill and at Barbegal. On one level, this method of
increasing product output constitutes a reversal in conceptual thinking, as it par-
allels the eastern practice of merely increasing the number of saddle querns within a
single location. What is different is the huge scale of production made possible
through use of water power. And, third, as long as water flowed continuously and
the mechanical parts of the mill did not break down, the miller could theoretically
grind cereal indefinitely, although in practical terms this was impossible. Never-
theless, subject to wear on moving parts, time lost due to breakdowns and repairs,
and interruptions of water or grain supply, the potential for labor savings and
increased product output offered by water-mills greatly exceeded what was possible
with human and animal-powered grain mills.
Water-mills were expensive to build and to operate and were subject to
practical limitations on location. Nevertheless, the large number of water-mills now
shown to have been operating between the first and seventh centuries A.D. in the
western Mediterranean indicates that Romans recognized the economic potential
of water-driven grain mills. It remains unclear who owned and financed their
construction and operation, whether as small individual mills or mill complexes.
Individuals or small corporations may have operated small mills, while large
complexes, such as the 16-mill installation at Barbegal driven by water diverted
from an aqueduct, may have had municipal or imperial ownership. So, for example,
the Janiculum mills in Rome probably served the imperial annona.
Once milling was complete, the baker moved directly to produce bread, a process
involving three steps: sieving the meal to obtain flour, kneading the flour with water
to produce dough, and baking in an oven. Technological innovation attending these
steps is important but less impressive than for milling. Sieving, for example, ad-
vanced little beyond improvement in quality of the mesh and so ability to improve
the fineness of the flour (Moritz 1958: 151-94; Curtis 2001: 361-62). The major con-
tribution to the technology of kneading was the Roman application of rotary motion
to the development of a mechanical kneading machine for use in commercial estab-
lishments. Nowhere clearly described in literary sources (cf. Vitruvius. De arch.
10.5.2), a mechanical kneading machine may be represented in the sculpted relief of a
bakery on the first-century B.C. tomb of Eurysaces, a professional baker and gov-
ernment contractor, in Rome, and archaeological remains have been discovered in
Italy and Tripolitania (figure 2.2; Bakker 1999: 78; Drine 2001; Curtis 2001: 362-65).
The one major innovation in baking was the development by Romans of large
beehive ovens (figures 2.2, 14.3) for use in commercial bakeries that utilized animal-
driven mills to process substantial amounts of flour to bake bread for sale or dis-
tribution to urban populations (Curtis 2001: 366-70). Animal and water-driven mills,
the mechanical kneading machine, and large ovens allowed for a significant increase
in the capacity to produce bread on a large scale.
GRAPE AND OLIVE PROCESSING
Before the classical period, wine produced from grapes was an alcoholic beverage
reserved for the wealthy, while oil from olives was used less as a food item than for
producing expensive perfumes. By the fifth century B.C. wine became the basic
drink for Greeks of all social levels, and oil served a number of purposes, including
as a food. Although wine assumed a central place in Greek social life and played an
important role in commerce in the classical and Hellenistic periods, oil, although
produced abundantly in many local areas and perhaps exported regionally, was
generally not the object of long-distance trade (Ault 1999; Jameson 2001; Lund
2004). During the Roman period, however, many areas, particularly Spain and
North Africa, produced wine and oil on a large scale and actively engaged in long-
distance trade to make these products available to markets throughout the Med-
iterranean (Chic Garcia 2000).
Although grapes are edible whether fresh or dried as raisins, the olive can only
be eaten after curing the pulp through fermentation in salt water to remove the
bitter taste caused by phenolic compounds. The primary form of consumption for
both fruits, however, was a processed product, wine from the grape and oil from the
olive. In each case, the fruit had to be crushed and then pressed. Following these two
procedures, grapes underwent fermentation into wine, while workers separated oil
from the liquid expressed from the olive pulp. Although the manner of crushing
differed, the same kind of apparatus served to press both fruits. For this reason,
distinguishing a winemaking installation from an olive oil-extraction emplacement
can be difficult. Consequently, building on Drachmann's technical treatise on
ancient oil mills and presses, many studies have focused on recognizing wine and oil
installations, classifying the types of equipment utilized, tracing the development of
the production apparatus, and determining the technological principles involved in
their processing (Drachmann 1932; Amouretti 1986; Mattingly 1988; Frankel 1999;
Brun 2003, 2004a, 2004b).
Crushing grapes, a simple and easy process requiring only human feet, re-
mained unchanged throughout antiquity. Paintings, mosaics, and sculpted reliefs
frequently represented scenes of individuals treading grapes in large containers
such as terracotta pithoi or masonry vats. One sixth-century B.C. Attic black-figure
amphora shows a satyr treading grapes in a basket sitting in a spouted trough placed
on an incline, so the expressed must flowed into a nearby collecting vat (Curtis
2001: pl. 22). This scene may represent a method to keep the mare separate from the
must in order to produce white wine. If so, it is the earliest clear evidence for the
process used to make a white as opposed to a red wine, types clearly distinguished in
Greco-Roman literary sources (Sparkes 1976).
Recognizing archaeological evidence for treading is difficult, since portable
apparatuses are not likely to have survived, and permanent masonry treading vats
can be difficult to identify as such. The fifth-century A. D. agricultural writer
Palladius, in his description of a treading floor with a system of channels for
conducting the must to tanks (Agr. 1.18 ), does not mention mechanical presses. It is
unknown how common completely manual processing was, since most excavated
wineries had a press room (torcularium). Those who could not afford a press no
doubt availed themselves either of portable treading vats or, like Palladius, per-
manent treading floors, either cut in the rock or made of masonry. Such installa-
tions are difficult to connect with wine without some corroborating evidence.
Although some wineries may have had a separate, portable treading apparatus, no
longer extant, in most cases the press room, if the floor had a waterproof covering
(opus signinum) that sloped toward a channel that led to a collection vessel,
probably doubled as a place both to tread and to press grapes. Cato (Agr. 112.3)
mentions this practice, and there is physical evidence for it at the second-century
B.C. winery at Mirmekion in the Strait ofKerch on the Black Sea, as well as the first-
century A.D. Roman press room in the Villa of the Mysteries at Pompeii (Curtis
2001: 302, 385-88; Brun 2003: 53-58; 2004a: 122-23). Workers could transfer the mare
directly from the floor into the "frail" (pressing basket) for use with the press. The
must from treading and from pressing ran directly into vats or similar containers.
The first step in the production of olive oil is crushing to reduce the fruit to a
pulp. The earliest methods used a pestle to pound them in a rock-cut or free-
standing stone mortar, a stone roller to mash them in a shallow basin, or wooden
sandals worn on the feet to tread them (Curtis 2001: 381; Brun 2003: 146-49). For
small-scale production, these techniques sufficed. When growing economic activity
during the Hellenistic period, and particularly later during the Roman period, gave
rise to large-scale production and long-distance trade, however, a radically new
method was developed, the mechanical olive crusher. It remains unclear whether
the rotary olive crusher originated in the Levant or Anatolia as early as the mid-
eighth century B.C. and spread west (Frankel 1999: 75), or arose somewhat later on
mainland Greece, the Greek islands, or Macedonia (Sagiv and Kloner 1996; Brun
1997). The Roman trapetum, the best known rotary olive crusher (figure 14.4),
consisted of a stone receptacle (mortarium) having a concave inner surface and a
central post ( miliarium) at the top of which was fixed an iron pivot (columella). The
pivot secured a horizontal wooden beam (cupa) that on each end penetrated
through an upright crushing stone (orbis) a truncated sphere with the convex side
facing out. Olives placed in the mortarium were crushed between its inner, concave
surface and the outer, convex surface of the orbes when two workers pushed the
horizontal beam, causing the stones to revolve on the pivot around the central post.
Figure 14.4. Roman olive crusher (trapetum) from Boscoreale. (Photograph by R. Curtis.)
Although its operation is clear, many questions of detail remain unanswered. For
instance, what distinguished the trapetum, described by Cato (Agr. 20-22), from the
mola olearia, according to Varro (Rust. 1.55.5) virtually the same device but men-
tioned as a separate apparatus by Columella (Rust. 12.52.6-7)? Only three sculpted
relief representations are known. Since Columella (Rust. 12.55.6) asserts that
crushing the pit imparted a bitter taste to the oil, olive crushers must have been
adjustable, but how was this accomplished (White 1975: 226-29; Frankel 1993; Tyree
and Stefanoudaki 1996; Curtis 2001: 381-82; Brun 2004a: 8-9)?
Grapes and olives, although differing in many respects, share an essential step
in their processing, pressing ( figure 14.2). The scene of a lever-and-weight press
painted on a sixth-century B.C. Attic black-figure skyphos provides a terminus ante
quern for the earliest Greek use of a mechanical press ( figure 14.5). One end of the
beam was usually fixed into a wall that served as a fulcrum. The free end of the
beam, when pulled down, crushed the fruit gathered up in a frail or wrapped in
cloth and placed at the point below the beam that allowed the most pressure to be
exerted. The use of a beam significantly increases the mechanical advantage in
applying force to the fruit over that possible by treading alone. This increased
efficiency and achieved greater output, while reducing the labor required ( foxhall
1993; Curtis 2001: 227-33).
The principle of the lever remained a constant in press construction until
invention of the direct screw press in the first century A. D. Technological inno-
vation in beam presses followed two lines of development: improvements in the
method of pulling the beam down to crush the fruit and in the mode of anchoring
the rear end of the beam. As shown in the vase painting cited above ( figure 14. 5), use
of a large stone or bag of stones was the simplest method. The difficulty in ma-
nipulating a single large stone may have made a bag more common for small
- µ •
,,,,1'• ~ If:>-.
Figure 14.5. Lever-and-weight press. Attic black-figure skyphos, ca. 520-510 B.C. Museum
of Fine Arts, Boston, Henry Lillie Pierce Fund, inv. No. 99-525. (Courtesy of the Museum
of Fine Arts.)
enterprises. The amount of force applied by the beam was directly related to the
weight of the attached stones. To increase the pressure, as engagingly shown on the
skyphos, humans could also apply additional force to the beam. Later, stones
shaped and cut to permit easy attachment of a rope allowed for increased weight.
The first development to the lever-and-weight press was the addition of a winch,
operated with a rope attached to the beam and wrapped around a drum that could
be turned by means of handspikes. As the drum rotated, the rope wrapped more
tightly around it and, so, pulled the beam down. This obviated the need for direct
muscle power to assist the weights. It is a short step from this to the lever-and-drum
press without weights. The drum was secured beneath the free end of the beam by
two wooden uprights (stipites) anchored in the ground. The second development to
the lever press altered the manner by which the rear end of the beam was fixed.
Rather than anchoring the rear of the beam (prelum) in a wall, such as is found in
Halieis in the Argolid and Praesos on Crete, of the late fourth and third centuries
B. C., respectively, by the second century B. C. it was inserted into a single wooden
upright (arbor), or slotted pier, or between two wooden uprights firmly seated in
the ground, in a manner not entirely clear, so as to offset any upward thrust caused
by the winch drawing the beam down onto the fruit. This arrangement, typified by
the lever-and-drum press described Cato (Agr. 18-19), may have had an origin in
the Hellenistic Levant ( frankel 2001). North African versions of this development
utilized large, heavy upright stone arbores with a drum attached to a counterweight
operating the beam. In both cases, the beam could be adjusted in the arbores to
allow for the compression of the frail containing the fruit (Drachmann 1932: 50-55;
Mattingly 1988; Sagiv and Kloner 1996; Curtis 2001: 307-11, 382-90; Jameson 2001:
283-84; Brun 2003: 58-60).
The second major innovation in pressing technology was the application of the
screw. This improvement was made possible by the invention of the screw and
screw nut, or female screw, by Hellenistic scientists in the third century B.C.
(Kiechele 1967; Drachmann 1968). It appeared in practical use in Greece by the first
century B. C., either as an assist to the drum of the lever-and-drum press or as the
principal method of drawing down the beam. By the mid-first century A.D., the
beam was being discarded altogether in favor of the screw alone (Pliny, HN18.317).
The purpose of the screw was to extract even more juice from the grape or olive,
even if that meant increasing the bitter liquid from the skins that lowered its
quality (Drachmann 1932: 55-59, 69-82, 96-99; White 1975: 230-32; Curtis 2001:
390-93; Lawton 2004: 15-21; Brun 2003: 60-63; 2004b: 48-49).
Pressed oil flowed out of the frail onto a circular stone press bed etched with a
groove that conducted the liquid to a spout where it emptied into a vat or other
container. In the absence of evidence for an oil crusher, this spouted stone bed
constitutes the clearest indication of an oil installation. One other construction that
can serve to identify an oil installation leads us to the final step in olive oil pro-
duction, separation. Since oil and water do not mix, the usual way to separate the
two is to skim the oil off the top of an oil/water mix with a ladle and to transfer it to
another vessel. Workers at Hellenistic Praesos apparently used a more efficient two-
handled jar with a spout at the bottom, not unlike Bronze-Age terracotta tubs
found on Crete. The purpose of this Minoan artifact remains uncertain. Sugges-
tions include oil separation, grape treading, or even activities not connected with
food preparation (Curtis 2001: 269-70). Once filled, the worker unplugged the
spout, drained the water, and then plugged it again to retain the oil. Romans also
used ladles, a labor-intensive and time consuming method when working with large
amounts of liquid obtained with the mechanical press (Cato, Agr. 66; Columella,
Rust. 12.52.8-12). Although archaeologists in Italy, Sicily, North Africa, and else-
where have uncovered sites at which a series of vats was used to separate oil from
water and other waste material (amurca), either by the principle of overflow or of
underflow (the principle at work in the spouted jars at Praesos) or a combination of
both, in classical antiquity skimming remained the most widespread method of oil
separation (frankel 1999: 174-76; Curtis 2001: 269-72, 393-94; Brun 2003: 156-58;
Brun 2004a: 114-15; Brun 2004b: 38-42). Wine and oil installations used similar
presses, although not the same press, as one product would taint the other and, for
wine production, necessitate the additional step of separation.
Following pressing, grape must (mustum) was set aside for various uses, in-
cluding processing into different nonalcoholic beverages, but the greatest portion
underwent fermentation into wine (Brun 2003: 63-83). No significant advancement
can be discerned in the fermentation or aging processes over methods used in
earlier periods, nor is there any evidence that Greeks or Romans knew distilled
spirits as a drink distinct from wine. Aristotle (Mete. 2.3) discusses distillation, but
does so in the context of experimentation in the evaporation of wine and the
condensation of what he viewed only as water with a specific taste. Finally, although
Greeks and Romans employed ice or snow to cool wine, and the Romans even had
shops reserved specifically for storing snow (Seneca, Ep. 78.23), neither society
developed the use of ice or snow into a method of long-term food preservation
(Curtis 2001: 254-55, 419)
VEGETABLES, FRUITS, AND NUTS
Greeks and Romans consumed many vegetables, fruits, and nuts, and, in the case of
the first two, preserved them by various methods for short- and long-term storage.
Legumes, and certain fruits such as apples, figs, dates, cherries, and grapes, were
dried. Fruits, olives, herbs, onions, and leafy vegetables were often pickled in brine
or vinegar, while other plant foods, such as apples, peaches, quinces, and roots, were
preserved in a variety of liquids, such as wine, grape juice (defrutum), and honey.
Dehydration, fermentation, pickling, and preservation in various liquids were
conservation methods practiced well before the classical period and show little or
no technological improvement; they are mentioned as routine procedures by an-
cient authors (e.g., Apicius 1.12.1-11; Cato, Agr. 116; Varro, Rust. 1.58-59; Columella,
Rust. 12; Pliny, HN 15.59-67, 102-4; see also Andre 1981: 45-48, 86-91; Thurmond
2006: 165-87). The techniques of food processing could affect the dietary quality of
plant foods for better or worse (Stahl 1989).
ANIMAL BY-PRODUCTS
The classical diet included fish and various wild and domesticated fowl and meat
animals, including cattle, pigs, sheep, and goats. Butchery is a basic process re-
quiring a variety of simple tools, some that are heavy and sturdy for breaking bones,
and others that are sharp for cutting through muscle and tendons. Little is known of
butchery tools in the Greek Bronze Age, and much of what we know of butchery in
the classical and Hellenistic periods centers on butchery for ritual purposes. For the
Roman period, archaeological material and sculpted reliefs, such as the tomb relief
showing a butcher holding a cleaver and standing next to a rack of meat in a butcher
shop, or the funerary altar of Atimetus that shows a cabinet filled with variously
shaped knives and cleavers offered for sale (Zimmer 1982: 93-106, 180-84), illustrate
the types of instruments used to butcher animals. A quick comparison of butchery
tools used during the Greco-Roman period with those from earlier periods,
however, shows no marked improvement beyond changes in the material making
up the tool and in the variety of shapes and sizes. Indeed, butchers today use almost
identical tools to cut up meat as did the Greeks and Romans-with the exception of
electric saws (!sager and Skydsgaard 1992: 66; Curtis 2001: 313-14, 395-99).
Once butchered, meat, particularly from cattle and pigs, underwent preser-
vation by various methods: drying, smoking, pickling, preserving in honey, or a
combination of these methods. Most meat, however, was preserved with salt. Lit-
erary sources frequently refer to salted pork, probably the most popular meat,
preserved as sausages or specific cuts, either deboned and filleted or with the bone,
particularly hams and the head (sinciput) (e.g., Cato, Agr. 162; Columella, Rust.
12.55.1-4). A first-century B.C. shipwreck off Provenc;e yielded bones of a boar's
head apparently bisected and possibly salted for long-distance transport, and ham
bones are frequently found on Roman wrecks in conditions favoring the preser-
vation of organic remains (Tchernia et al. 1986: 231-56; Frost 1999: 241-52; Le-
guilloux 2001: 411-14; Curtis 2001: 396-99; MacKinnon 2004: 173-75, 184-86.).
Animals maintained on farms and villas provided various food by-products,
such as milk from cattle, sheep, and goats. Milk, because it spoils quickly, often
underwent processing to produce cheese, a longer-lasting food item. Finds of
bronze cheese graters in Lefkandi in Euboea indirectly affirm the knowledge of hard
cheeses as early as the ninth century B.C. Greek literary sources of the classical
period mention many types of local and imported cheeses, including fresh and
cream cheeses and hard cheeses that could be crushed, grated, and sliced. None,
however, describes the production in more detail than the comment by Aristotle
(Hist. an. 522a22-b6) that milk could be curdled by fig juice or rennet. Roman
agricultural writers, on the other hand, provide specific detail on cheese produc-
tion. The three-stage process they describe no doubt paralleled the Greek method
and differs little in its essentials from cheesemaking today: curdling by warming and
the addition of rennet, separating the liquid whey from the curds, which were
pressed and salted, and ripening. The value of cheese, especially hard cheese, was its
extended shelf life and consequent marketability as an item of long-distance trade
(Curtis 2001: 315-16, 400-2).
Two other animal by-products received attention from the Greeks and Ro-
mans, but their production required little or no processing. Butter, produced by
shaking warm milk, was considered primarily a fatty substance similar to oil and
was used more as an additive for medicine than as a food. Columella (Rust. 9.15.12-
13) explains the procedure for gathering and pressing honeycombs to extract the
honey, the most important sweetener available to the classical cultures (Curtis 2001:
399-400, 417-19).
Although Greeks and Romans consumed fresh fish when and where available,
they most often processed fish meat into dried, smoked, or salted form (salsa-
mentum ). Fish innards and small whole fish went into producing fish sauces, es-
pecially garum, used as a condiment. Columella (Rust. 12.55.4) indicates that pro-
cessing salt fish paralleled the method used to salt pork. Indeed, sheep and cattle
bones excavated from fish-salting vats in Kerobestin and Telgruc on the Atlantic
coast of Gaul imply that salting of meat in the off season kept fish-salting instal-
lations active year round. Fish preservation involved placing alternate layers of fish
meat and salt in a container until reaching the top. Added weights applied pressure
on the mass and facilitated the expression of the moisture from the fish meat and its
replacement through osmosis with salt. The species of fish used and time allowed
for the process determined the shelf life of the final product. In the second century
Athenaeus, in his Deipnosophistae ("Sophists at the Dinner Table"), records names
for varieties of salted fish products that betray specific details of the production
process. So, for example, certain terms indicate that salt-fish were cut into various
shapes and preserved fully or partially salted with or without scales. Certain parts of
the fish, such as the dorsal part or stomach, were particularly selected for preser-
vation (Curtis 1991: 6-15; 2001: 397 ). Although numerous variables came into play in
producing fish sauce, the basic process involved mixing salt in a container of small
fish or fish innards in a specified ratio. This mixture, covered and allowed to sit in
the sun for a period of time up to three months with occasional stirring, underwent
autolysis, during which most of the fish material dissolved, creating a top layer of
liquid (garum) and a bottom layer of undissolved fish matter (allec). This pro-
duction method parallels almost exactly the process for producing modern fish
sauces, such as Vietnamese nuoc-mam (Curtis 1991: 15-26, 191-94).
Recognizing ancient drying and smoking facilities today is nearly impossible,
but large-scale salting establishments have left considerable remains (figure 14.6).
Greeks imported salt-fish products from the Black Sea and from the western
Mediterranean, particularly Punic areas of Sicily and Spain, a source confirmed by
salt-fish pieces adhering to Punic amphoras found in the fifth-century B.C. "Punic
Amphora Building" at Corinth (Curtis 1991: 115). Archaeologists have uncovered
Punic salteries that operated at the same time or a little later in the area of Cadiz.
The best evidence for processed fish, however, comes from excavations of Roman
salteries in the western Mediterranean in Spain, Portugal, North Africa, and the
Atlantic coast of France, and in the eastern Mediterranean in the northern Black Sea
areas of the Strait of Kerch and Chersonese, all of which date primarily between the
first century B. C. and the fourth century A. D. (Ponsich 1988; Curtis 1991; Etienne et
al. 1994; Etienne and Mayet 2002; H0jte 2005; Trakadas 2005). A comparison of
salting apparatuses at all of these Roman sites with earlier Punic installations in-
dicates that methods of salting fish remained practically unchanged over this time
period. There was, however, a significant increase in the volume of production and
trade in these products. Studies of salt-fish amphoras found in shipwrecks or at
inland sites are yielding a clearer picture of the importance of this processed food to
specific areas and to the Roman economy generally (Curtis 1991; Chic Garcia 2000;
Martin-Kilcher 2003).
Food technology during the Greco-Roman period achieved significant advances in

some areas, small incremental improvements in others, and remained stagnant in a
few. Milling of cereals saw the greatest strides. The first mechanical application to
milling, the Greek hopper mill, permitted automatic feeding of grain between
millstones, an innovation that characterized all subsequent mills. Most important
was the employment of rotary motion to drive the millstones. Knowledge of this
Figure 14.6. Roman fish-salting vats from Baelo, Spain. (Photograph by B. Lowe.)
principle brought new inventions, such as the rotary hand mill, the animal-driven
mill, and, most impressive, the water-mill. Additionally, knowledge of rotary mo-
tion led to the invention of the mechanical kneading machine and the trapetum to
crush olives. Except in one instance, advances in grain storage during the Greco-
Roman period differed from earlier periods primarily in quantity and scale rather
than in principles applied. The one exception is the employment of subfloor
ventilation for large, aboveground buildings, providing cool, dry, dark storage for
"clean" grain awaiting milling (Curtis 2001: 276-79, 325-35; figure 7.5). Incremental
improvements appeared in pressing technology as new ways were developed to in-
crease efficiency of the press beam, first by weights, then with a winch, and finally
through use of screw technology. Other aspects of food technology, however, saw
little or no improvement over earlier periods. In this category must be listed
butchery, fermentation and distillation of alcoholic beverages, and refrigeration.
Innovation and invention in Greco-Roman food processing tended in two di-
rections. The first trend led to more efficient, labor-saving machines requiring fewer
operators per unit of output. Unlike earlier cultures where augmented capacity
came from iteration of basic tools, Greeks and Romans turned to larger size and
improved functional efficiency. Most new or improved apparatuses were directed
toward commercial production and consequently tended toward machines bigger,
more technically sophisticated, and costlier than the simpler forms prevailing be-
forehand. Chief among innovations illustrating this theme was the development of
four distinct types of grain mill (hopper mill, rotary hand mill, animal mill, and
water-mill), and of four basic olive and grape presses (wedge, lever-and-drum,
lever-and-screw, and direct screw presses) with intermediary forms. Because of its
low production output, the wedge press, possibly used in perfumeries in Hellenistic
Delos and first-century A.D. Paestum and illustrated in four paintings found in
Pompeii and Herculaneum, was probably used only to press olives for particularly
refined oil used for making perfume. It most likely represents a dead-end tech-
nology in that the faulty structural design of its wooden frame inadequately con-
trolled the thrusts caused by pounding in the wedges (White 1975: 232; Mattingly
1990; Brun 2000: 282-302). Although the concept of improved workplace comfort
may have played no role in labor savings, it often was an unintended consequence
of innovation in both milling and pressing technology.
The second trend in Greco-Roman food processing brought increased avail-
ability of food to a larger number of consumers and contributed to economic
growth. Increasing populations, growing foreign markets, and heightened gov-
ernmental concerns for food shortages and the political and social unrest poten-
tially arising from them, particularly during the Roman period, constituted pri-
mary incentives to increase production capacity and long-distance trade. The
presence of animal-driven mills, water-mills, olive and grape presses, and large-
scale fish-salting installations implies significant capital investment to increase
surplus so as to satisfy immediate needs, to establish a reserve, and to maximize
profits (A. Wilson 2002). Consequently, new technologies led to enlarged food
resources for consumers and to fresh employment opportunities for workers in-
volved at various stages along the route from point of processing, to loading and
offloading conveyances, to transportation by land and sea, and to sale in the market.
Yet, with all this innovation, the Greco-Roman period displayed a persistent
conservatism. The implements for butchery showed no marked improvement, nor
did the processing of husked cereals improve beyond mortar and pestle. This
apparent stagnation most likely resulted not from any lack of imagination but from
recognition that the technology had already achieved a level of efficiency sufficient
for its purpose. Most revealing in this respect, however, is the contemporaneous
availability of simple tools for food processing alongside new, more efficient ap-
paratuses. In the early fourth century A.D., for example, the Edict of Diocletian
simultaneously listed maximum prices for hand mills, animal mills, and water-mills
in a veritable "technology shelf" from which to choose (Edictum Diocletianum
15.54). The consumer had the option to select which technology best suited his
needs and available resources.
Invention and innovation in food technology brought significant benefits to
inhabitants of the Greco-Roman world. First, increased production, larger storage
capacity, and longer shelflife augmented food supply and provided greater access to
foods year round, especially in times of want. Second, transportability of food
preserved in a form more stable than in its natural state provided better access to
food in greater variety to a larger number of people over a wider geographic area.
Finally, in addition to providing more food at a higher quality and in greater variety
than was necessary for survival, food processing, particularly in the form of alco-
holic beverages such as wine and beer, increased the enjoyment oflife. Although the
Greco-Roman period may not have seen an industrial revolution, the technology
necessary to create machines operated first by human power, then by animals, and
finally by water power to achieve greater output more efficiently with fewer workers
cannot be called stagnant.
REFERENCES
Alonso Martinez, N. 1997. "Origen y expansion molino rotativo bajo en el Meditemineo

occidental," in Meeks and Garcia 1997: 15-19.
Amouretti, M.-C. 1986. Le pain et l'huile dans la Grece antique. Paris: Les Belles lettres.
Amouretti, M.-C., and J.-P. Brun (eds.) 1993. La production du vin et de l'huile en Medi-
terranee. Bulletin de Correspondance Hellenique Suppl. 26. Paris: Ecole Frarn;:aise
d'Athenes.
Andre, J. 1981. L'alimentation et la cuisine a Rome, 2nd ed. Paris: Les Belles Lettres.
Ault, B. A. 1999. "Koprones and oil presses at Halieis: Interactions of town and country and
the integration of domestic and regional economies," Hesperia 68: 549-73,
Bakker, J. T. 1999. The mills-bakeries of Ostia: Description and interpretation. Amsterdam:
J.C. Gieben.
Bekker-Nielsen, T. (ed.) 2005. Ancient fishing and fish processing in the Black Sea region.
Aarhus: University of Aarhus,
Brun, J.-P. 1997. "L'introduction des moulins dans les huileries antiques," in Meeks and
Garcia 1997: 69-78.
Brun, J-P. 2000. "The production of perfumes in antiquity: The cases of Delos and
Paestum," American Journal of Archaeology 104: 277-308.
Brun, J.-P. 2003. Levin et l'huile dans la mediterranee antique: Viticulture, oleiculture
et procedes de transformation. Paris: Editions Errance.
Brun, J.-P. 2004a. Archeologie du vin et de l'huile de la pre'histoire a l'epoque hellenistique.
Paris: Editions Errance.
Brun, J.-P. 2004b. Archeologie du vin et de l'huile dans l'Empire romain. Paris: Editions
Errance.
Brun, J.-P., and P. Jockey (eds.) 2001. Technai: Techniques et societes en Mediterranee. Paris:
Maisonneuve et Larose.
Brunet, M. 1997. "Le Moulin Delien," in Meeks and Garcia 1997: 29-38.
Castella, D. 1994. Le moulin hydraulique gallo-romain d'Avenche "en Chaplix." Aventicum
VI. Lausanne: Cahiers d'archeologia romande.
Chic Garcia, G. (ed.). 2000. Congreso Internacional Ex Baetica Amphorae: Conservas, aceite
y vino de la Betica en el Imperio Romano. Sevilla-Ecija, 17 al 20 diciembre de 1998.
4 vols. Ecija: Editorial Graficas Sol.
Curtis, R. I. 1991. Garum and salsamenta: Production and commerce in materia medica.
Leiden: Brill.
Curtis, R. I. 2001. Ancient food technology. Leiden: Brill.
Domergue, C., D. Beziat, B. Cauuet, C. Jarrier, C. Landes, J.-G. Morasz, P. Oliva, R. Pulou,
and F. Tollon 1997. "Les moulins rotatifs dans les mines et les centres metallurgiques
antiques," in Meeks and Garcia 1997: 48-61.
Drachmann, A. G. 1932. Ancient oil mills and presses. Copenhagen: Levin and Munksgaard.
Drachmann, A.G. 1968. "The screw of Archimedes," inActes du VIIIe Congres International
d'Histoire des Sdences, Florence-Milan 1956. Florence/Paris: 1968, 3:940-43.
a a
Drine, A. 2001. "Meules grain et petrins autour du lac El Biben et Gigthi," in Brun and
Jockey 2001: 251-60.
Etienne, R., M. Yasmine, and F. Mayet 1994. Un grand complexe industriel a Troia (Por-
tugal). Paris: de Boccard.
Etienne, R., and F. Mayet 2002. Salaisons et sauces de poisson hispaniques. Paris: de Boccard.
Finley, M. I. 1959. "Technology in the ancient world," Economic History Review 12: 120-25.
Forbes, R. J. 1955-1964. Studies in ancient technology. 9 vols. Leiden: Brill.
Foxhall, L. 1993. "Oil extraction and processing equipment in classical Greece," in
Amouretti and Brun 1993: 183-200.
Foxhall, L., and H. A. Forbes 1982. "Sitometreia: The role of grain as a staple food in
classical antiquity," Chiron 12: 41-90.
Frankel, R. 1993. "The Trapetum and the Mola Olearia," in Amouretti and Brun 1993:
477-80.
Frankel, R. 1999. Wine and oil production in antiquity in Israel and other Mediterranean
countries. Sheffield: Sheffield Academic Press.
Frankel, R. 2001. "Cato's Press a Reapraisal [sic]," in Brun and Jockey 2001: 313-25.
Frankel, R. 2003. "The Olynthus mill, its origin, and diffusion: Typology and distribution,"
Frost, F. 1999. "Sausage and meat preservation in antiquity," Greek, Roman and Byzantine
Studies 40: 241-52.
Greene, K. 1990. "Perspectives on Roman technology," Oxford Journal ofArchaeology 9:
209-19.
Greene, K. 1993. "The study of Roman technology: Some theoretical constraints," in Theo-
retical Roman Archaeology: First Conference Proceedings. Aldershot: Avebury, 39-47.
M. I. Finley re-considered," Economic History Review 53: 29-59.
H0jte, J.M. 2005. "The archaeological evidence for fish processing in the Black Sea region,"
in Bekker-Nielsen 2005: 133-60.
Houston, G. W. 1989. "The state of the art: Current work in the technology of ancient
Rome," Classical Journal 85: 63-80.
Isager, S., and J. E. Skydsgaard 1992. Andent Greek agriculture: An introduction. London:
Routledge.
Jameson, M. H. 2001. "Oil presses of the late classical/Hellenistic period," in Brun and
Jockey 2001: 281-99.
Kiechle, F. 1967. "Zur Verwendung der Schraube in der Antike," Technikgeschichte 34:
14-22.
Lawton, B. 2004. The early history of mechanical engineering. 2 vols. Leiden: Brill.
Leguilloux, M. 2001. "La boucherie et l'artisanat de sous-produits animaux en Gaule
romaine," in Brun and Jockey 2001: 411-21.
Leveau, P. 1996. "The Barbegal water mill in its environment: Archaeology and the eco-
nomic and social history of antiquity," Journal of Roman Archaeology 9: 137-53.
Lewis, M. J. T. 1997. Millstone and hammer: The origins of water power. Hull: University
of Hull.
Lund, J. 2004. "Oil on the waters? Reflections on the contents of Hellenistic transport
amphorae from the Aegean," in J. Eiring and J. Lund (eds.), Transport amphorae and
trade in the eastern Mediterranean. Acts of the International Colloquium at the Danish
Institute at Athens, September 26-29, 2002. Athens: Aarhus Universitetsforlag, 211-16.
MacKinnon, M. 2004. Production and consumption of animals in Roman Italy: Integrating
the zooarchaeological and textual evidence. Journal of Roman Archaeology Suppl. 54.
Martin-Kilcher, S. 2003. "Fish-sauce amphorae from the Iberian peninsula: The forms and
observations on trade with the north-west provinces," Journal of Roman Pottery
Studies. 10: 69-84.
Mattingly, D. J. 1988. "Megalithic madness and measurement-Or, how many olives could
an olive press press?" Oxford Journal of Archaeology 7: 177-95.
Mattingly, D. J. 1990. "Paintings, presses and perfume production at Pompeii," Oxford
Meeks, D., and D. Garcia (eds.) 1997. Techniques et economie antiques et medievales: Le
temps de l'innovation. Colloque international (C.N.R.S.) Aix-en-Provence 21-23 mai
1996. Paris: Editions Errance.
Morel, J.-P. 2001. "Aux origins du moulin rotatif? Une meule circulaire de la fin du Vie
a
siecle avant notre ere Carthage," in Brun and Jockey 2001: 241-50.
Moritz, L. A. 1958. Grain-mills and flour in classical antiquity. Oxford: the Clarendon Press.
Nelson, M. 2001. "Beer in Greco-Roman antiquity." Ph.D. diss., University of British
Columbia.
Nelson, M. 2005. The barbarian's beverage: A history of beer in andent Europe. London:
Routledge.
Pleket, H. W. 1973. "Technology in the Greco-Roman world: A general report," Talanta 5:
6-47.
Ponsich, M. 1988. Aceite de oliva y salazones de pescado: Factores geo-economicos de Betica y
Tingitania. Madrid: Universidad Complutense.
Reece, D. W. 1969. "The technical weakness of the ancient world," Greece andRome 16.1:
32-47.
Runnels, C. 1990. "Rotary querns in Greece," Journal of Roman Archaeology 3: 147-54.
Sagiv, N., and A. Kloner 1996. "Maresha: Underground olive oil production in the Hel-
lenistic period," in D. Eitam and M. Heltzer (eds.), Olive oil in antiquity. Padua:
Sargon, 255-92.
Sarpaki, A. 1992a. "A palaeoethnobotanical study of the West House, Akrotiri, Thera,"
Annual of the British School at Athens 87: 219-30.
Sarpaki, A. 1992b. "The palaeoethnobotanical approach: The Mediterranean triad or is it a
quartet?" in B. Wells (ed.), Agriculture in ancient Greece. Proceedings of the Seventh
International Symposium at the Swedish Institute at Athens, 16-17 May 1990.
Stockholm: Astroms, 61-76.
Sparkes, B. A. 1976. "Treading the grapes," Bulletin Antieke Beschaving 51: 47-56.
Staltl, A. B. 1989. "Plant-food processing: Implications for dietary quality," in D.R. Harris
and G. C. Hillman (eds.), Foraging and farming: The evolution of plant exploitation.
London: Unwin Hyman, 171-94.
Tchernia, A., M. Girard, and F. Poplin, 1986. "Pollens et ossements animaux de l'epave
romaine 3 de Planier (Provence)," in La mer comme lieu d'echanges et de communi-
cation. Vol. 2 of L'exploitation de la mer de l'antiquite a nos jours. Viernes rencontres
interntionales d'archeologie et d'histoire d'Antibes. Valbonne: Association pour la

Promotion et la Diffusion des Connaissances Archeologiques, 231-56.
Thurmond, D. L. 2006. A Handbook offood processing in classical Rome: For her bounty no
winter. Leiden: Brill.
Trakadas, A. 2005. "The archaeological evidence for fish processing in the western Med-
iterranean," in Bekker-Nielsen 2005: 47-82.
Tyree, E. L., and E. Stefanoudaki 1996. "The olive pit and Roman oil making," Biblical
Archaeologist 59: 171-78.
White, K. D. 1975. Farm equipment of the Roman world. Cambridge: Cambridge University
Press.
Wikander, b. 2000. Handbook of ancient water technology. Leiden: Brill.
Wilson, A. 1995. "Water-power in North Africa and the development of the horizontal
Wilson, A. 2000. "The water-mills on the Janiculum," Memoirs of the American Academy in
Rome 45: 219-46.
Wilson, A. 2002. "Machines, power and the ancient economy," Journal of Roman Studies
92: 1-32.
Wilson, R. J. A. 1996. "Tot aquarum tam multis necessaries molibus: Recent studies on
aqueducts and water supply," Journal of Roman Archaeology 9: 18-29.
Zimmer, G. 1982. Romische Berufdarstellungen. Berlin: Mann Verlag.
CHAPTER 15
LARGE-SCALE
MANUFACTURING,
STANDARDIZATION,
AND TRADE
ANDREW I. WILSON
PRINCIPLES
"Industry," in the sense of large-scale mechanized production of nonagricultural

products, is a term rarely applicable to the ancient world, although there are oc-
casional instances where it might be used. Nevertheless, the material culture of
Roman world, perhaps more than any other society until the Industrial Revolution,
was characterized by the widespread availability of mass-produced artifacts. Pottery,
glassware, bricks, coins, plate, and humble metal objects such as nails were pro-
duced in enormous quantities to standard shapes and sizes, and widely traded
around the Roman Mediterranean and northern Europe and even beyond the
frontiers of the empire, reaching India and the Sahara. The apparent sameness of
Roman statuary and sarcophagus sculpture, long contrasted unfavorably by art
historians with the artistic creativity of the Greeks, may now be recognized not only
as a significant facet of a culture in which replication was ideologically important,
but also as an economic phenomenon, a response by a finite body of artisans and
sculptors to booming market demand. Most of the phenomena discussed in this
chapter relate to the Roman period, but some of the elements of Roman mass pro-
duction are traceable as far back as the Hellenistic world, if not before. This chapter
examines the chief features and limitations of mass production in the ancient
world, and attempts to explore what the phenomenon might imply about the
nature and performance of the ancient economy. It will examine technologies not
only of large-scale manufacturing, but also technologies of organization.
By "mass production," I mean the production of very large quantities of the
same artifact, or of essentially similar artifacts, by the same production means.
"Large-scale production" extends the concept to include also the creation of large
quantities of bulk produce (olive oil, salted fish) as well as replicated individual
artifacts. Mass production usually, and large-scale production frequently, involves
one or more of the following factors: division oflabor, standardization of sizes and
forms, and sometimes the creation of standardized, interchangeable parts.
Adam Smith began his Inquiry into the Nature and Causes of the Wealth of
Nations (1776) with the observation that the key to economic growth lies in the
division oflabor. This is true both for the differentiation of trades, which is found in
any complex society (in which farmers, blacksmiths, spinners, weavers are all dif-
ferent individuals), and for the division oflabor within a single trade or profession.
The latter is simply an extension of the former principle, but it is our main concern
here. If a complex task within a single trade is split into a series of smaller and
simpler operations, each performed by one worker, per capita productivity is greatly
increased. Each worker, repetitively performing the same limited sequence of ac-
tions, acquires a dexterity and speed that a worker performing a wider range of tasks
could not achieve; furthermore, time is not lost by switching between the different
parts of a job. The greater the number of separate steps a manufacturing process is
broken into and divided among different workers, the more people must be em-
ployed, but their per capita production increases. The downside is that division of
labor leads to deskilling of the workforce, since the greater the division oflabor, the
more limited is the job each worker does.
Division of labor is one of three main factors enabling per capita economic
growth, the others being improved techniques for doing the same things, and
mechanization, which is a logical progression of division of labor when a manu-
facturing process has been split into tasks so simple that individual steps can be
performed by a machine. The identification of evidence for the increased division of
labor in a particular economic sector over time therefore may reflect per capita
economic growth in that sector, and may be brought to bear on larger arguments
over the possibility and achievement of per capita economic growth in antiquity.
Two ancient texts explicitly address the division of labor. The first is Xeno-
phon's Cyropaedia (8.2.5), which anticipates Adam Smith's observation that divi-
sion oflabor is limited by the size of the market. Writing in the fourth century B.C.,
Xenophon accounts for the excellence of the cuisine at the Persian court by the
multitude of specialized chefs employed:
Nor should it surprise us; for if we remember to what a pitch of perfection the
other crafts are brought in great communities, we ought to expect the royal dishes
to be wonders of finished art. In a small city the same man must make beds and
MANUFACTURING, STANDARDIZATION, AND TRADE 395
chairs and ploughs and tables, and often build houses as well; and indeed he will
be only too glad ifhe can find enough employers in all trades to keep him. Now
it is impossible that a single man working at a dozen crafts can do them all well;
but in the great cities, owing to the wide demand for each particular thing, a
single craft will suffice for a means of livelihood, and often enough even a single
department of that; there are shoemakers who will only make sandals for men
and others only for women. Or one artisan will get his living merely by stitch-
ing shoes, another by cutting them out, a third by shaping the upper leathers,
and a fourth will do nothing but fit the parts together. Necessarily the man who
spends all his time and trouble on the smallest task will do that task the best.
The second relevant text appeared some 800 years later. In the fifth century
A.D., St. Augustine mocked pagan polytheism, with different gods having different
spheres of influence, in terms that draw on the division of labor in silversmithing
(De civ. D. 7.4).
Indeed we laugh when we see them [the pagan gods] distributed by the fancies of
human opinions to particular works, like contractors for tiny portions of the
public revenue, or like workmen in the street of the silversmiths, where one vessel,
so that it may go out perfect, passes through many craftsmen, when it might have
been finished by one perfect craftsman. But the reason why a multitude of
workmen was thought necessary was so that individual workers could learn in-
dividual parts of an art quickly and easily, rather than that they should all be
compelled to be perfect in one entire art, which they could only attain slowly and
with difficulty.
Both authors stress the increased quality brought about by division of labor,
rather than the efficiency gains and lowered costs, although in both passages this
emphasis is based on their function as an analogy. Interestingly, Augustine does
comment on the reduced cost and effort of training, picking up on the fact that,
as we have noted, division of labor leads to deskilling. Classical authors, who were
not trying to write Smithian economics, can only take us so far. To proceed any
further in the study of large-scale production, we need to turn to the evidence of
archaeology.
PRODUCTION SITES
Study of the layout of production sites can sometimes shed light on the scale of
labor employed and the division of labor achieved, while the built structure of the
premises in which a group of workers were gathered under the same roof, any water
supply or other production infrastructure, or even the provision of machinery such
as presses or mills, can sometimes hint at the nature or degree of investment made
for the production of particular goods.
In his study of Roman pottery production, Peacock distinguished several types

of production (Peacock 1982: 9-10 ), which are relevant also to the production of
goods other than pottery. In studying the technologies of organization for mass
production, we are here concerned with his categories of workshops, nucleated
workshops, manufactories, and factories. His other categories, of household pro-
duction, household industry, estate production, and military production, are not
discussed here, the first two because they operate on too small a scale to be relevant
to mass production, and the latter two because they are simply variants of the other
types, in one case located on a villa estate and in the other under military control,
but not necessarily showing any different physical or technical characteristics. A
workshop is a building or part of a building in which an artisan, perhaps with several
assistants, makes products. The nucleated workshop is a cluster of several individual
workshops forming a production complex with shared access to raw materials,
labor, or markets. A manufactory is a dedicated production facility in which nu-
merous artisans "co-operate in producing a single and often highly specialized
product.... Archaeologically the manufactory will be distinguished by the size of
premises, the degree of specialization in the products, by the scale of output, and by
evidence of worker specialization" (Peacock 1982: 10). For the pottery industry,
Peacock made a largely arbitrary distinction, based on ethnographic parallels from
twentieth-century Tunisian pottery production, between workshops, with up to
twelve workers, and manufactories, with more than twelve. The manufactory is not
usually equipped with machinery powered by other than human or animal means,
but it does employ division of labor. True factories, by contrast, operate on a large
scale, employ division of labor, and utilize powered machinery.
MASS-PRODUCED ARTIFACTS
The component elements of mass production began to appear in the sixth century
B.C., with the use of molds and separately made parts in the production of Co-
rinthian "plastic vases," in which mold-made heads of different kinds of animals
might be applied to essentially similar wheel-thrown bodies (Biers 1994), although
the relatively small quantities in which these were made hardly merit the term "mass
production." During the fifth to third centuries B.C., Corinthian mortaria were
made in molds and widely exported, and from the Hellenistic period onward mold-
made wares become increasingly common for relief-decorated bowls, enabling
elaborate but standardized vessels to be turned out quickly and relatively cheaply.
This tradition carried through into the tradition of red-slipped Italian terra sigillata
(ITS), and its later Gaulish derivative, "Samian ware." Roman lamps were also
usually mold-made in two parts (upper and lower) that were then fitted together.
Mass-produced high-quality tablewares, whether decorated and mold-made,

or wheel-thrown and plain, are ubiquitous in the Roman world, even reaching
Britain and India; mass-produced and standardized cooking wares are also com-
mon (figure 20.4; chapter 20). Their distribution implies an integrated pan-
Mediterranean trading network, in which the aggregate size of the market
stimulated producers with access to export distribution networks to enormous
quantities of production. Many ITS and Samian vessels are stamped with the
potter's name on the underside of the base, often understood as relating to the
organization of production or for identifying batches during firing in shared kilns.
Fiille (1997) has argued that these stamps mark the products of officinatores or
workshop managers, who may have been renting potteries or space in a pottery
from a landowner under a contract of locatio conductio rei. Such contracts are
known from third-century A.D. Egypt; a potter leasing a pottery or a part of a
pottery and undertaking to produce a certain quantity of amphoras (in one case,
15,000) in return for cash payment in staged installments (Cockle 1981). The
landowner provided the facilities and materials, while the potter provided the labor
(including assistants). Presumably the potter could sell any excess production
above the stipulated amount at his own profit.
The mass production of Roman tablewares has led many scholars to see the
organization of Arretine production in terms of manufactories, from the numbers
of personnel recorded on stamps in relation to particular individuals, and a set of
large tanks, interpreted as being for clay levigation, found at Arezzo in association
with numerous shards bearing the stamp of C. Perennius and dependent workers.
Neither piece of evidence, however, is sufficient: not all people appearing on
stamps as dependants of the same individual need have been active simultaneously,
and the large vats might equally be clay preparation tanks, or evaporation tanks to
produce fine red slip by slow evaporation from fine mud, neither of which need
imply exceptionally large operations ( fulle 1997: 134-35; Pasqui 1896; cf. Peacock
1982: 54). Fiille therefore argues that Arretine production is that of nucleated
workshops rather than manufactories.
But if stamps cannot be used to estimate the numbers of people simultaneously
active in the same workshop, neither can they be used to estimate the maximum
sizes of such workforces-we do not know what fraction of the total number of
people who originally stamped their products have been recovered. Furthermore, if
the people recorded on stamps are officinatores, they could each have had several
people working for them whose names are not recorded. Moreover, this approach
relies on the epigraphic evidence from stamped products. But the production of
African Red Slip ware surpasses Arretine and Gaulish Samian production both in
chronological range and in extent of distribution, and very few of these products
are stamped. If the stamps on Arretine and Samian ware reflect locatio-conductio
contracts between potters and the owners of production facilities, then the im-
plication is that a different system may have operated in Africa-perhaps one in
which the landowner directly employed or owned the potters. Classic manufactory
production would be easier to envisage in these circumstances, but would be

epigraphically invisible.
The physical evidence for the number and scale of production facilities, and for
possible division oflabor, helps us here. While the African production centers have
not been excavated-the center whose products we know best, El Mahrine, has
been investigated only by surface collection (Mackensen 1993)-there is other
evidence for large-scale production in Italy and Gaul.
The derivative Gaulish productions involving potters attested at Arezzo, which
appear at Lyons from 10 B.C. and then at La Graufesenque in Southern Gaul, are
now thought to reflect migration of expert potters rather than the establishment of
branch workshops (Fiille 1997: 141-44; cf. Oxe et al. 2000: 49-50). Some speciali-
zation and division oflabor can be seen in the case of the South Gaulish potteries at
and near La Graufesenque. The separately prepared foot ring of molded bowls of
form Dragendorff 37, produced from the 6os to about 120, would have facilitated a
basic division oflabor in the preparation of these vessels (Mees 1994: 19). A labor
account from La Graufesenque mentions pueri (slaves, or perhaps apprentices)
preparing the clay, fetching fuel, and assigned to the task ad samiandum, probably
applying the glossy slip most characteristic of Samian ware (Marichal 1988: 226-27,
no. 169; Bulmer 1980: 29). Clearly, these slaves were not the actual potters but were
assigned to a variety of ancillary tasks. Most importantly, firing was in the hands of
specialist kiln-masters operating massive kilns used by groups of potters. Bilingual
graffiti in Gaulish and Latin on broken or rejected dishes list vessels loaded into
kilns for firing; each charging-list gives the names of several potters with hundreds
or thousands of vessels in that firing-up to 12 potters and over 30,000 vessels in a
single firing (Hermet 1923, 1934: 291-355; Albenque 1951; Aymard 1952, 1953; Duval
and Marichal 1966; Marichal 1988; Vernhet and Bemont 1991). The largest number
of vessels recorded in a firing is 33,845 (with a ready-reckoning total of 33,500;
Hermet 1934: 309; on the language of these documents, see Adams 2003: 687-724).
One kiln excavated in 1980 confirms the capacities given by the charging lists: it
measured about 4 m square internally, and about 4 to 5 m high, giving a volume
approaching 80 m 3 (Bemont 1996: 125).
The charging lists from La Graufesenque also seem to provide evidence for the
simultaneous use of up to ten large kilns there. The tally lists are headed either with
the Gaulish word tu0os followed by a Gaulish ordinal number up to ten (e.g., tu0os
decametos, "tenth tu0os") or with the Latin word furnus ("kiln") followed by a
Latin ordinal (e.g., furnus secundus). It has been argued that the Gaulish tu0os
means a kiln-load rather than the kiln itself, and that the documents therefore refer
to a firing cycle throughout the year and not to the number of physical kilns
(Marichal 1988: 97; Flobert 1992: 105; Vernhet and Bemont 1991, 1993; Fiille 2000:
72-76). However, Marichal 83, which starts"[ ... ]bres incepit furnus pri[mus],"
must refer to a month ending in -bres (Duval and Marichal 1966: 1349), and even
September would be far too late in the year for the firing season to start. This detail
appears to exclude the use of ordinal numbers to denote a particular firing in an
annual cycle; they must instead refer to the number of the kiln. The uniform
structure of the lists and the consecutive numbering of the kilns suggest that the
kilns were under single ownership or operation. As each firing might take perhaps
three weeks, it is likely that the ten or more kilns were fired in rotation - perhaps
loading and commencing the firing of a different kiln every couple of days. This
would indicate considerable investment in plant, so that the whole operation could
be managed with a small firing crew and the potters could continue to produce
vessels without having to wait long for another kiln to become available.
The La Graufesenque documents thus give a picture of nucleated workshops
operating with communal firings in massive shared kilns; the size of the operation
and the nature of the tally lists imply that the kiln-masters were separate from the
potters. This practice operated for a considerable time-one series of tally lists is
dated A.D. 40-60, and another to between the reigns of Trajan and Antoninus Pius
(Albenque 1951: 80). Communal firings are also known from other large pottery
production centers: Arezzo, Les-Martres-de-Veyre. and Rheinzabern, for example
( CIL 11.6702, nos. 2, 5, 23a; Duval and Marichal 1966: 1341, n. 3).
Important evidence for the actual layout of pottery workshops comes from two
sites: Le Rozier, about 20 km from La Graufesenque, and Scoppieto in the upper
Tiber valley. At Le Rozier, part of a workshop has been excavated, a building at least
22 m long, with a store for prepared clay at the east end, and a workroom to the west
(Thuault and Vernhet 1986; Thuault 1996). The workroom was at least 17 m long by
7 m wide, with eleven surviving clay storage tubs about 1.1 to 1.2 m square in a row
along the south wall, and a water channel between the tubs and the wall. Although
no potters' wheels were found in situ, the arrangement of the channel and the tubs
excludes these being elutriation or decantation tanks, and parallels from Scoppieto
confirm that these are tubs to hold the clay used by the potters while working. The
arrangement thus provides evidence for a common workroom with at least eleven
potters working side by side. The west end of the room had been eroded by the river,
so the original number may have been more. The Le Rozier workshop was active
between A.D. 50 and 80, during which period 21 potters are attested by stamps on
the pottery, of whom 18 are also known at La Graufesenque.
Scoppieto is a rural site in the upper Tiber valley watershed, much better placed
than Arezzo for supplying the market of Rome. Excavations have revealed a large
workroom with a row of working places for 27 potters, evident from the emplace-
ments of the potters' wheels and clay bins similar to those at Le Rozier (Bergamini
2003, 2004; Nicoletta 2003; cf. also Mascara 2003). More than 50 different potters
are attested over the life of the establishment (Augustan period to early second
century A.D.); some may have worked there only for a limited time (Nicoletta 2002:
213-15). Here it is not the stamps but the operational arrangement that gives us the
number of potters working simultaneously: 27.
At Scoppieto, mass production was achieved by the replication of similar work
units, but each worker carried out many of the stages of production of the vessel.
Economies of scale would have been produced, though, in the organization and
supply of raw materials to several craftsmen all working in the same location, in the
sharing of common facilities and services, notably the kilns but also probably clay
preparation and levigation tanks, and in the marketing and distribution of the
products. Although each potter probably made an entire vessel from the forming
of the wet clay to air-drying ready for firing, it is likely that the initial prepara-
tion and levigation of the clay, and the slipping and firing of the vessels, were carried
out by other workers (as at La Graufesenque), thus implying a basic division of
labor.
At both Le Rozier and Scoppieto, we cannot be sure whether the potters were all
the employees of a single owner or independent potters renting space and facilities
in the workshop. If the stamping of Arretine and Samian pottery reflects leasing of
production facilities under contracts of locatio-conductio, then this implies a more
complex and cooperative model of production than is usually associated with the
concept of manufactory, where the workers are normally employees of the owner.
Le Rozier and Scoppieto, therefore, do not fall neatly into Peacock's categories. On
the one hand, they display the physical characteristics and layouts of basic man-
ufactories, with replication of identical work units in the same place, and some very
elementary division of labor; they qualify for the term "manufactory" chiefly on
grounds of scale. Yet the organization, the relationship between owner and workers,
is very different from the manufactories of immediately preindustrial eighteenth-
century Europe. This would suggest that the categories we commonly use to classify
production units with an implied sense of progression-from household pro-
duction to household industry to individual workshops to nucleated workshops to
manufactories and finally to factories-do not represent the full available range.
Put another way, this is a teleological perspective assuming that all development
builds in a similar manner toward an industrial revolution culminating in factories.
The production facilities of Le Rozier and Scoppieto may have stood somewhat to
one side of the linear trajectory of workshop development that, retrojecting from
the late medieval and early modern periods, we like to see as paradigmatic. In the
Roman world, some landowners appear to have invested heavily in the infra-
structure for large-scale pottery manufacture, building facilities that could ac-
commodate numerous workers with economies of scale through shared access to
raw materials, resources, clay preparation and firing facilities, and marketing and
distribution structures. Yet rather than employing the potters who used the facil-
ities, they may have preferred to engage them via relatively short-term contracts.
This could have worked in one of two ways. The landowners could have sub-
contracted to the potters the production of a certain number of vessels, for which
the landowner would pay the potter and then sell himself. Here the landowner
assumes the final risk in terms of market demand and price. Alternatively, the
potters might have rented the facilities from the landowner-who was thus assured
of a fixed income-and the potter would retain ownership of the vessels until sale.
This latter strategy would limit the landowner's risk, so long as there was sufficient
demand by itinerant or mobile potters to use well-organized facilities.
While pottery is one of the most obvious categories of mass-produced artifact
in the ancient world, others are also very common. Roman sites produce ex-
traordinarily large quantities of blown-glass vessels as well, and glass output did not
reach similar levels again until Venice began large-scale exports in the fifteenth
century. Coins were struck in enormous issues, and quotidian metal objects were
also produced in vast numbers. When the partially completed fortress at Inchtuthil
in Scotland was abandoned as the Romans withdrew, over a million iron nails
(weighing nearly ten tons) were buried at the site to prevent them falling into
enemy hands (Manning 1985: 289). Imperial arms factories existed in the late Ro-
man empire, but we know nothing of their layout or detailed operation. Never-
theless, the production of arms and armor is one obvious area where the large-scale
demand may have stimulated methods of mass production, and some preliminary
indications hint in that direction. The design of spearheads, arrowheads and ballista
bolts enabled their batch production from strips of iron, cut into triangular blanks
and forged into shape by hammering in a curved "bottom swage." The iron strips
for these blanks, and also for scale armor, may have been produced using rollers,
which would have enabled a division of labor between semiskilled artisans who
forged and rolled the sheets and cut the blanks, and a skilled smith who shaped the
projectile heads (Sim 1995). The scale of iron smelting and smithing sites on Ex-
moor suggests mass production of iron tools and weapons (Juleff 2003).
Many different types of object were stamped-ceramic vessels, bricks, lead
pipes, barrels, loaves of bread-in what sometimes looks like a frenzy of bureau-
cratic control but often relates to production contracts, water rights, or food en-
titlement, depending on the object concerned (Harris 1993). Among these stamped
objects, bricks are worth examining in more detail. The large imperial building
projects erected in the city of Rome, coupled with the demand for private ac-
commodation, required many millions of bricks. Mass production of fired bricks
started in the reign of Augustus and lasted throughout the empire; it was partic-
ularly intense in Italy, especially around Rome. Many bricks were stamped with the
names of estate owners or workshop managers, and numerous prosopographic
studies have used these brick stamps to examine the scale and organization of the
industry (Bloch 1938-1939, 1947; Helen 1975, 1976; Setala 1977; Steinby and Helen
1978; Steinby 1974-1975, 1977, 1993). Such was the scale of the Roman building
industry, and of the brick manufacture that supplied it, that the term "industry"
becomes justified here. The demands of the building industry at Rome, Ostia, and
Portus, and in the suburbs of Rome, were met by the brick industry of the Tiber
valley, which was dominated initially by senatorial families. Through intermarriage
and the confiscation of senatorial estates, the emperor or the imperial family in-
creasingly came to control it. As with large-scale pottery production, these are not
standard capitalist industries; exploitation was carried out principally by landown-
ers letting locatio-conductio contracts for brickmaking to individuals who carried
out the work via dependants; these individuals are the officinatores or workshop
managers named on the stamps (Helen 1975).
Bricks were made in standard sizes, thus allowing builders to mix bricks from
different suppliers in the same building without changing working patterns. This
removed a supply bottleneck inherent in masonry construction, where an impor-
tant factor controlling building rates is the output capacity of the quarry supplying
the stone (Wilson 2005). The use of brick facing also allowed concrete walls to be
constructed without formwork, which speeded up the process of construction and
lowered costs (chapter 10). These two factors enabled an enormous project such as
the Baths of Caracalla, which used approximately 6.9 million bricks (DeLaine 1997:
124, table 10,126, table 11), to be finished in five years. The brick producers produced
to stock, for an indefinite market, rather than to order. While the massive demand
generated by imperial building projects and private construction meant that most
of the bricks made in the Tiber valley were destined for Rome or its ports at Ostia
and Portus, some were exported, and Italian bricks are present at most of the major
North African ports from Caesarea Mauretaniae (Cherchel) to Lepcis Magna, ar-
riving perhaps as salable ballast (Tomber 1987; Wilson 2001b). Bricks from the
region of Aquileia were also exported to Dalmatia. The standardization of bricks, as
part of the technology of concrete construction, represents one of the most im-
pressive indicators of the scale of the Roman building industry.
MARBLE TRADE
While brick-faced concrete was increasingly used in many areas for the structural
core of a building, quarried stone, and particularly marble, remained necessary for
architectural ornament and also for structural elements, especially columns, in
trabeated architecture. Accordingly, the marble trade evolved over the first to third
centuries A.D. in response to increasing demand. Architectural marbles, such as
column shafts, capitals, bases and entablatures, begin to be produced to standard
shapes and perhaps also lengths, and shipped in near-final form, for finishing and
polishing on arrival at the importing site; columns in particular were left with rough
collars at the end to protect against damage in transport (Asgari 1978, 1990; Dodge
and Ward-Perkins 1992; Waelkens et al. 1988: no; Waelkens 1990: 64-70). Such
prefabricated marble architectural elements from Proconnessus and Docimium
have been found in several shipwrecks off Sicily (Pensabene 2003), and also in one
at Sile in the Black Sea alongside roughed-out sculptural marbles. By the second
century, the major marble quarries had moved from a quarry-to-order to a quarry-
to-stock regime, a sign that the aggregate market had become so large that a whole
range of different architectural elements was virtually guaranteed to find a buyer-
or to be needed by an imperially-funded building program (chapter 5). This pro-
cedure also enabled the quarries to fulfill orders more quickly. Systems of quarry
marks carved on blocks and columns, both at the quarries and at major delivery
yards at Ostia and Rome, facilitated accounting and stock-taking (Dodge and
Ward-Perkins 1992; Fant 1989; Hirt 2004).
Concurrently with the standardization of architectural marbles, we can also
observe high levels of standardization in the sarcophagus trade. The white marble
quarries of Proconnesus, Ephesus, Aphrodisias, and an as yet unidentified site in

Caria roughed out garland sarcophagi to a basic shape before export, as is shown by
unfinished sarcophagi both in the quarries themselves and in shipwrecks. Each
quarry had its own roughed-out design, with garlands, roundels, and other
blocked-out motifs-pendants or plaques for an inscription-that could be refined
later into a variety of decorative shapes. Hollowing and roughing out minimized
the weight and added value, but was not carried to the point where detailed carving
might be damaged in transit. The sarcophagi were finished by workshops in the city
to which they were shipped, where the same basic rough design might be turned
into swags between bukrania, or Nikai or putti holding garlands, while the blocked-
out roundels might become rosettes, busts, or lion heads. Sometimes sides were left
unfinished if they would not be seen because of the sarcophagus's position in the
tomb; or the entire sarcophagus might even be used in its roughed-out, quarry-state
form, for reasons that remain uncertain.
This standardized production of sarcophagi probably began in the late first
century A.D., and lasted into the third century. By the late second century, the
practice of shipping and even using quarry-state garland sarcophagi had become so
common that the red granite quarries of Assos in the Troad actually produced
sarcophagi in a form imitating the quarry state of the white marble quarries, but
designed to be used without further finishing. They were decorated with a basic
swag motif and a panel for an inscription but, apart from the occasional addition of
an inscription to the central panel, were never further finished at the importing
centers. Assos sarcophagi are found mainly in the eastern Mediterranean in Syria,
Palestine and especially Alexandria; but also at Thessalonica, Dyrrhachium, Ni-
copolis, Ravenna, and some western sites, and in a wreck off Methone. Their main
production runs from the late second century into the early third century (Koch
1993: 172-73).
Also standardized were strigillated sarcophagi from the Saliari quarries on
Thasos, whose products were exported to Italy and the central Mediterranean, and
several types produced at the quarries of Docimium in Asia Minor. These included
columnar sarcophagi exported finished in every respect except for the portrait
details of the deceased, added by workshops where the products were imported
(Waelkens 1982). Simpler funerary monuments, fashioned as an aedicula with a
false door, are found throughout Asia Minor and were roughed out at several
quarries including Docimium and finished in urban workshops elsewhere
(Waelkens 1982; 1986: 20 and index s.v. "Halbfabrikat"). Even the most luxury end
of the market shows an extension of the same phenomenon: sarcophagus lids with
reclining couples, in Thessalonica and in Cyrene, are finished in every respect
(including drapery and patterns on the mattress), except their heads, which are left
rough. This clearly shows manufacture of common types to stock, with the in-
tention that client-specific portrait details would be added later (or not at all!)-
apparently confirmed by quarry-state rough-outs of such sarcophagus lids in the
quarries at Docimium (Thessalonica: Koch and Sichtermann 1982: ill. 421; Cyrene
Museum: personal observation; Docimium: Waelkens 1990: 69, figs. 35-36).
Sarcophagi and the false-door funerary monuments share, of course, a large

potential market. But the vast building activity of the first to early third centuries
A.D. also created in its wake a huge demand for marble statuary-not only to
adorn niches of nymphaea or the halls of public baths, but also for the reciprocal
gestures by which towns honored benefactors with statues. One might expect that
the combination of this considerable demand, from several hundred or even
thousand cities around the Mediterranean, coupled with the limited number of
quarries able to supply statuary-grade white marble, would have led to more
efficient ways of producing statues. Indeed, there is evidence to show that precisely
the systems of standardization identified for architectural marbles, sarcophagi,
and-crucially-the full-length reclining portrait figures on some sarcophagus
lids, were extended also to freestanding portrait sculpture during the second and
early third centuries A.D. The evidence may be grouped into three categories:
roughed-out statues found abandoned in the quarries; rough-outs found in
wrecks; and pieces found elsewhere that show signs of standardized basic working
and but with certain details left unfinished.
The numerous statues ofDacian prisoners in Trajan's Forum were roughed out
in the quarries at Docimium, as a rejected piece left there shows (Waelkens 1990:
69-70, fig. 39). A group of roughed-out statues found in a quarry near Xylophagon
on Cyprus (in hard white limestone rather than marble) included a cuirassed
emperor, a draped female figure, and a nude male, all over life-size, and a less than
life-size statue said to be the Small Herculaneum Woman type (Karageorghis 1969:
494-99). A late imperial, cuirassed emperor statue has been found in quarry state
on Proconnesus (Asgari 1978: 480, pl. 142; 1990: 125-26); roughed-out statues are
reported from the Saliari quarries on Thasos (Kozelj et al. 1985: 75), and the prac-
tice has also been suggested in the case of the Thasos white marble quarries at Cape
Vathy, in particular for a type, common in the second and early third centuries, of a
figure of Eros sleeping on a torch (Herrmann and Newman 1995: 79-80; 1999: 296;
Tykot et al. 2002: 189).
While it could be argued that rough-outs abandoned in the quarry might have
been intended to be finished completely before export, but some flaw prevented
their completion, rough-outs from wrecks prove that some sculptures were ex-
ported partially finished. A wreck excavated at Sile (Black Sea) had been carrying a
cargo of Proconnesian marble including capitals, a sarcophagus lid, a female bust,
and an imperial cuirassed statue, all roughed out in quarry state to be finished at the
final destination (Asgari 1978, 1990). The quarry and wreck evidence suggests that
the export of roughed-out, or even nearly finished, statues, whose portrait details
would be added at the importing centers in consultation with the clients, may have
been common. This hypothesis is supported both by the general likeness of com-
mon stances and poses in full-length Roman portraits, both male and female, and
by specific examples from Cyrene, in which the accomplished details and finishing
of the portrait heads contrasts with the schematic rendering of the drapery (e.g.,
Alfoldi-Rosenbaum 1960: nos. 43, 95; personal observation in Cyrene Museum,
2006; Ben Russell oral communication). In these cases, it seems, the statues have
been exported from the quarries nearly finished, with the drapery schematically
rendered as a guide for the receiving workshop, but not taken to the point where
damage in transit could not be rectified, and the heads either left for finishing or
indeed separately worked. The heads have been finished by portrait sculptors who
have ignored the other details; indeed, the schematized drapery quickly became an
accepted style in its own right.
Cumulatively, the evidence suggests increasing specialization, with the quarries
in the second and early third centuries A.D. producing increasingly finished statues
rather than exporting raw blocks of marble to be worked elsewhere. The advantage of
doing as much as possible of the carving at the quarry was twofold: first, it reduced
the weight of the item to be shipped; and secondly, craftsmen used to working only
on the marble from a particular quarry, familiar with its grain, could work it faster
than generalists who might have to work with a variety of different marbles. The
disadvantage was that finely finished fluted drapery or other details might be chipped
in transit; rough-finishing the statue, with the head left for portrait details to be
added, was therefore a good compromise. The Roman practice of replication in
sculpture, originally developed in part for cultural reasons, naturally lent itself to
standardization and shipping of quarry-state statues. Trimble (2000) presents a
study of the social and political context of the standardized representation of two
female statue types, the so-called Large and Small Herculaneum Women. Indeed, the
shipping of quarry-state rough-outs in the second century A.D. goes some way
toward explaining the widespread diffusion of common poses in which details of
drapery vary somewhat. There was, however, no single uniform system; some
sculptures were clearly exported after simple roughing out of the pose, others with
the drapery schematically worked, and others, like the reclining couple sarcophagi,
even with the drapery finished. In addition, and particularly for statues that did not
conform to the common types, raw blocks were shipped for working from scratch at
the receiving workshop. Practice probably varied by quarry, and over time. The
important point is that by the early second century the market for honorific and
portrait statuary (both of private citizens and the imperial family), had grown to such
a size that it could support standardized production to stock of common types, and
an articulation of production between, on the one hand, the quarries producing
standardized poses and drapery treatment, and, on the other, the importing work-
shops where the portrait details and final finishing were carried out.
DIVISION OF LABOR
The marble trade as discussed above shows a basic division of labor, spatially
separated between quarries and final workshops. But the staged division oflabor in
a single workshop, described by Augustine in his analogy of the silver vessel passing
through the hands of numerous silversmiths, can in fact be detected in the layout
of certain production facilities. Two examples will be examined here: the pro-
duction of bread, and the manufacture of decorative marble objects.
Baking and Milling

Increasing subdivision of labor over time can be traced in the baking trade, a
mundane and therefore important example. Improvements in the technology or
organization of bread production would have far-reaching effects on activities
affecting every household. The baking trade also illustrates the introduction of
rudimentary mechanization.
Bakeries as a specialized workshop are known from the sixth century B.c., but
the introduction of animal-driven grain mills in the second century B.c. boosted
the specialization of the miller-cum-baker as a separate trade, progressively re-
placing laborious domestic bread production using hand querns (chapter 14;
Curtis 2001: 280-93, 358-69; Moritz 1958: 62-74). Animal-driven mills required
capital investment, but had a much larger output, allowing centralized grain
grinding and bread production by miller-bakers. Such mills became particularly
common in urban contexts where large market demand existed. By A.D. 79,
Pompeii had about 35 bakeries, indicating near-universal consumption of shop-
bought bread by town-dwellers.
By the Augustan period if not earlier, Roman bakers were also using me-
chanical dough-mixers (chapter 13). The layout of two very large bakeries at Ostia
also hints at a basic specialization of labor. In the purpose-built bakery in Insula
I.XIIl.4 (ca. A.D. 100-125), grain was brought into a long rectangular courtyard,
and processed in an aisled hall to the north, with seven animal mills in the southern
aisle and five kneading machines in the northern aisle (Meijlink 1999). The basalt
paving of the hall (to withstand the wear of the animals continually walking in
circles) indicates that both the mills and the kneading machines were animal-
powered. Although the layout is tidy, it is not optimal. The only available space for
sieving the flour after milling is in the south range; the sieved flour must then have
been carried back across to the north aisle for kneading in the machines, and the
dough must then have been transferred back again across the courtyard to room 10,
on which it was shaped into loaves on tables, and then baked in the oven at the end
of the room.
In the Via dei Mulini, a large bakery was installed in a commercial building
apparently in the Severan period, and functioned, with modifications, until around
280 (Heres 1999; Bakker 1999b: esp. 57-60.). Here the plan is more efficient,
showing a clear division of the stages of grinding grain, kneading dough, and
baking the bread ( figure 15,1). A large vaulted hall housed at least ten animal-driven
mills paved with basalt slabs; the flour was probably sieved in the northern part of
this hall. In the adjacent room are five dough-mixing machines, also animal-
powered on the evidence of the basalt-paved floor. Beyond the dough-mixing
CAS A DI DIA NA
191B}
' '
0 k~
0
........ . ! □ ; 100
© ; !
B
0 u
I
L
D
i::~l:=~====Jl ~ G
6
.!"!,
--- --.• .- ·-.- :- . -----. --. --.... -----.-.-.-' -. -' -' ..... ' -.... -.- .-. ---. - . -. - --. --. _:; -~ --.-.---.- .-.-. -.-.--.. -- . - . -. --- . -. ~ -~ . ~- . -.. -. -... --.-.-. -. -. -.: :: :: .: .: .: .: :: .: :: ': '
VIA DEi M□LINI
[L
□ □ GRANDI H □ RREA
gure 15.1. Plan of the second-century A.D. bakery at Casseggiato dei Molini, Ostia. (Courtesy of Jan Theo
tkker.)
room, two vast, domed bread ovens were sited at the end of the building in the first
phase of the bakery (one was later removed). Overall this bakery displays a degree
of rational planning, with the different stages of production separated into dif-
ferent rooms and laid out in a rudimentary production line.
The large bakeries at Ostia may reflect arrangements at Rome, where much less
is known about bakeries since few have been excavated or published. The ico-
nography of the tomb of the baker Eurysaces at Porta Maggiore seems to imply
operations on a similar or even larger scale. This travertine monument (dated
stylistically between 40 B.C. and A.D. 5) is decorated in its lower zone with sets of
cylinders stacked three deep, surmounted by the funerary inscription naming
Eurysaces as a contractor (redemptor), and above this is a zone with three layers of
circular holes. Around the top, a frieze shows scenes of bread production ( figure
2.2). The curious circular holes which form the most characteristic and striking
element of the monument are in fact dough-mixing machines laid on their sides
(the spindle sockets are visible at the end of each void), and the stacked cylinders
below them represent the same dough-mixers seen from the side (Ciancio Rossetto
1973: 33-34; Brandt 1993).
The frieze shows the different stages of bread production (Ciancio Rossetto 1973:
41-61; cf. Zimmer 1982: 106-8). On the south face, from right to left, is shown the
delivery of grain, with the state officials who had contracted out large baking orders
to Eurysaces recording their receipts; grain being milled in two donkey-driven mills,
and the flour being sieved by two men at a table; finally an official in a toga is sam-
pling it for quality. On the north face dough is kneaded in an animal-powered mixing
machine, with a man scooping out the dough; two groups of four men roll out dough
into loaves at two tables under the eye of an overseer; on the left-hand side a man
loads an oven with a long bread shovel. The frieze is broken to the left, but part of a
leg may indicate another man loading or unloading an oven in the missing part. On
the west face, reading from left to right, loaves are carried to be weighed under the
supervision of officials whom Eurysaces, as a contractor to the state, is supplying;
they make records in tablet-books, and the loaves are taken away in baskets.
The frieze illustrates exactly the sort of production-line operation implied by
the physical layout of the larger bakeries at Ostia a century and a half later. The
different numbers of men engaged in the different operations show that this is not
the same worker or set of workers performing all the operations in sequence; the
action is simultaneous and we focus on different steps. The exactitude of such
representations cannot be insisted upon, but the ratio of eight men rolling out
dough to one operating the animal-driven dough mixer, and one (or two) loading
the oven is at least suggestive of the scale of manpower required to keep pace with
the mechanized stages of production and with the large-batch baking. The number
of workers shown also conveys the idea that this is a large-scale operation; similarly,
the repeated motifs of stacked dough-mixers stress the size of Eurysaces' business
and its command of mechanical technology. The production-line narrative of the
frieze and the dough-mixers convey the message of a large and rationally organized,
partially mechanized bread manufactory.
The Chemtou Marble Workshops

The imperial quarries at Chemtou (Simitthus) in Tunisia, which produced the
pinkish-yellow marmor Numidicum, seem to have been given over to private ex-
ploitation in the early third century A.D. Large-scale quarrying of architectural
marble ceased and activity concentrated on the production of small but often
decorated marble objects-bowls, plates, and statuettes, often reworked from
unused blocks extracted under imperial ownership. The six parallel halls, originally
a secure barracks for the quarry workers condemned ad metallas, were converted
into workshops. The spatial distribution of production waste and rejected artifacts
in various stages of production when the workshops were abandoned suggests a
production-line process, with the products moving through different operations,
each carried out at a different place in the halls ( figure 15.2) (Rakob 1994: 66-89,
107-31; 1995: 42-43; 1997: 10-17; Mackensen 2000: 493; Ruger 1997). Excavations of
an area to the east of the six-halled building have also shown that the seriated
production of marble cups and statuettes extended well beyond the limits of the
former barracks building (Mackensen 2000: 500-502).
Two phases were identified. In the first, blocks were sawed in the westernmost
hall and then rough- and fine-dressed before polishing on marble benches in the
central halls. The final polishing, with river cobbles or small marble polishing stones
and wet sand, happened at the north end of the halls, over the drain of what had
been the barracks latrine. Relief-decorated bowls and statuettes, many of Venus,
were produced to a limited number of standardized patterns.
The second phase, from the mid-third century until the 280s, followed the
collapse of part of the vaulted roof. Production continued on a reduced scale, but
although the spatial layout changed, the production-line principle did not. Sawing
took place along the south and west sides of the workshops, with rough- and fine-
dressing close by. Products then moved through the central halls for polishing, with
final polishing again being carried out in two of the former latrine bays at the north.
In both phases, the scale and the production-line nature of the operation allow us to
consider this a manufactory.
Large-Scale Production in the Food Sector

The urban populations of the Roman world, particularly those distributed around
the Mediterranean, but also some towns in northwest Europe, and army garrisons
stationed on the imperial frontiers, were usually too large to be fed entirely from
local resources and depended considerably on long-distance trade. Together with a
more distributed rural population, they provided a vast aggregate market for
various foodstuffs. The size of this market, coupled with particular production
factors in some regions, facilitated the development of some installations pro-
ducing certain foodstuffs on a remarkable scale-principally salted fish and de-
rivative products, olive oil and wine, and milled flour.
Polishing with pebbles and
marble polishing stones
Rough and fine chifflling
Polishing on marble polishing benches
\
N
Sawing
Accommodation and lavacra
. ,. ••
~~~-~---~----'--------'----~ >O ...
Figure 15.2. Chemtou (Simitthus, Tunisia), plan of the Fabrika building in the first half of
the third century A.O., with distribution of waste from different stages of the production
of small stone objects. (After Rakob 1994; reproduced by permission.)
In the western Mediterranean, especially around the Straits of Gibraltar, but

also along the modern Tunisian coast and, to a lesser extent, the coasts of Sicily and
southern France, the seasonal migrations of fish led to the potential for enormous
catches at particular times of the year. These far outstripped the demands of local
populations, but needed to be preserved by gutting and salting before they could be
shipped to more distant markets packed in amphoras. The guts and small fry could
be processed to make various kinds of fish sauces-garum or liquamen, used widely
in cooking and appreciated as a delicacy (chapter 14). Roman fish sauces have
captured the modern imagination, but they should be seen as merely epiphe-
nomena! to a much larger trade in salted fish that provided a substantial protein
source for both coastal and inland communities. From the second half of the first
century B.C. onward, we see the emergence oflarge salting factories, with concrete
salting vats lined with waterproof mortar arranged around a courtyard, and areas
for gutting and fish preparation. A particularly clear example is at Cotta in Mor-
occo, with a long fish-gutting hall and 16 vats around a courtyard, with a total
capacity of 258 m 3 ; an area with hypocaust heating was probably designed for
producing the necessary salt by lixiviation from beach sand, rather than for pro-
ducing garum (Ponsich 1988, 150-59; Hesnard 1998). The Cotta factory, apparently
an isolated coastal establishment, though possibly linked to an undiscovered villa,
was by no means the largest. One of the estimated 50 factories at Troia (Portugal)
had a total capacity in excess of 606 m3, while another at Iulia Traducta (Algeciras
Bay) had a capacity of 150 m 3 • While these larger units were usually outside urban
centers, substantial factories were also sometimes located within towns, as at Baelo
Claudia in Spain, Lixus in Morocco, and Tipasa in Algeria (Wilson 2006).
In the western Mediterranean region, the majority of fish-salting factories were
built between 50 B.C. and A.D. 100, and operated until the fifth century, although
usually on a reduced scale after around 200 or 250. Fish were salted and exported
both before and after these dates, but in lesser quantities that could be accommo-
dated in smaller salting containers (pithoi and dolia); but from the first century B.C.
to the fifth century A.D. the economic conditions of Mediterranean trade were such
that the considerable investment in salting plant could be justified by the colossal
market one could reach. In the first century A.D., similar factories were also built in
Amorica (northwest Brittany), another area with marked seasonal shoaling pat-
terns. These included the second largest single installation in the Roman world, at
Plomarc'h in the Bay of Douarnenez, with a total salting capacity of over 466 m 3 •
Epigraphic evidence suggests that these factories were built by investors from
the Mediterranean, presumably in response to the emergence of new markets in
northwest Europe, as increasing numbers of troops were stationed along the Rhine
frontier and in Britain, and the inhabitants of the newly conquered province of
Britain began to acquire Roman dietary habits, including a taste for salted fish.
Salting factories are also found from the first to third centuries A.D. in several towns
of the Crimea, especially on the Kerch peninsula at Tiritace and Chersonesus,
taking advantage of migration routes through the Cimmerian Bosphorus (Curtis
1991; H0jte 2005; Wilson 2006).
In North Africa, the combination of the universal demand for olive oil as a food
source, lamp fuel, massage oil, and multipurpose lubricant, with the availability of
cheap seasonal labor from transhumant tribes who came north to summer pasture
around the olive harvest, encouraged the development of cash-crop olive growing.
Numerous rural press sites are found throughout parts of Numidia, Africa Pro-
consularis, and Tripolitania, and some are very large, well-built functional estab-
lishments with multiple presses. Sites with four presses are common, and some have
six, eight, or even more-the largest are Henchir Sidi Hamdan (nine presses) and
Senam Semana (17 presses). These veritable factories must have processed the olives
from enormous orchards, and imply massive landowning wealth, as well as the
resources to set up such batteries of press machinery. Yet these functional buildings
lack any decoration (mosaics, wall paintings, elaborate residential architecture)
commensurate with such wealth, and were probably administered by a slave bailiff
(vilicus) on behalf of an owner residing elsewhere. The scale of olive oil production
suggested by some of these estate centers is unparalleled in later societies before the
nineteenth century. It is frequently difficult to distinguish between presses for olive
oil and for wine, especially on unexcavated sites, and some of these sites may in fact
have been for wine. This is certainly the case for the massive factory at Kherbet el-
Agoub, a site in inland N umidia with ten pairs of presses and treading vats, arranged
around three sides of a courtyard with a storage room on the fourth (Meunier 1941;
Brun 2004: 233-38). But whether for wine or oil, these huge factory sites with
multiple presses represent major investments by landowners in the processing plants
necessary for cash-crop production for export. Large multiple press sites are also
found in some other Mediterranean provinces outside North Africa-up to four
presses in Spain, six in southern Gaul, and ten double presses at Barbariga in !stria
at the head of the Adriatic.
Large investments in processing machinery are also shown by the construction
of water-powered grain milling factories with multiple wheels (chapter 13). The
earliest example is the early second-century A.D. complex at Barbegal in southern
France near Arles, with 16 wheels; in the third century, if not before, a series oflarge
water-powered mills was built along the course of two branches of the Aqua
Traiana, and possibly also the Aqua Alsietina, on the slopes of the Janiculum hill in
Rome (Wilson 2001a).
The early Hellenistic period saw the beginnings of mass ceramic production for
export in both the Greek and Punic worlds. But it is the Roman period that is
remarkable for the spread of mass production in a variety of sectors, and for the scale
of some operations. Large-scale production is particularly apparent in certain kinds
of food production and the processing of agricultural produce-fish-salting, oil and
wine production, and bread-making; in some of these, the scale of production was
achieved by limited use of machines (presses, mills, and dough-mixers). For man-
ufactured goods, mass production largely took the form of replication of individual
items (ceramics, statues); in some cases simple production-line technologies were
employed. In certain areas, this mass production reached levels unparalleled until
probably seventeenth-century Britain in the half century leading up to the Industrial

Revolution. This is an impressive achievement. At the same time it is important also
to recognize the limits of Roman mass production: there is no hard evidence of the
mass production of complex items using standardized interchangeable parts, in a
manner foreshadowing the so-called American System that evolved out of eighteenth-
century French small arms production. It is not impossible that interchangeable parts
were used, for example in vehicles or siege engines, but there is as yet no evidence for
them. Nevertheless, there is much more research to be done, particularly on the mass
production of arms and armor, and we are only just beginning to see how to read
clues about organization and division of labor from the archaeology of production
sites. Strikingly, most of the sectors in which we can detect mass production relate to
the manufacture of goods or produce from raw materials (clay, iron, marble) or
agriculture, and represent either modes by which landowners (including the state)
could maximize return on their properties, exploiting resources either by agents or
through contracting out. Either way, the arrangement ensured that a large share of
the profits flowed to the landowner, thus limiting the ability of those directly involved
in the manufacturing to emerge as a separate entrepreneurial class.
REFERENCES
Adams, J. N. 2003. Bilingualism and the Latin language. Cambridge: Cambridge University
Press.
Albenque, A. 1951. "Nouveaux graffites de La Graufesenque," Revue des etudes anciennes 53:
71-81.
Alfoldi-Rosenbaum, E. 1960. A catalogue of Cyrenaican portrait sculpture. London: British
Academy.
Asgari, N. 1978. "Roman and early Byzantine marble quarries of Proconnesus," in
E. Akurgal (ed.), The proceedings of the Xth International Congress of Classical Ar-
chaeology. Ankara: Turk Tarih Kurumu, 467-80.
Asgari, N. 1990. "Objets de marbre finis, semi-finis et inacheve du Proconnese," in
M. Waelkens (ed.), Pierre eternelle: Du Nil au Rhin: Carrieres et prefabrication. Brus-
sels: Credit Communal, 106-26.
Aymard, A. 1952. "Nouveaux graffites de La Graufesenque II," Revue des etudes anciennes
54: 93-101.
Bakker, J. T. (ed.) 1999a. The mills-bakeries of Ostia: Description and interpretation. Dutch
Monographs in Ancient History and Archaeology 21. Amsterdam: J.C. Gieben
Bakker, J.-T. 1999b. "The Caseggiato dei Molini, Part III: Interpretation," in Bakker 1999a,
39-60.
Bemont, C. 1996. "Les comptes de potiers," Dossiers d'Archeologie 215: 122-27.
Bergamini, M. 2003. "Una produzione firmata da Marcus Perennius Crescens a Scop-
pieto," Rei Cretariae Romanae Fautorum Acta 38: 133-44.
Bergamini, M. 2004. "Scoppieto e il commercio sul Tevere," paper delivered at the con-
ference "Mercator Placidissimus: The Tiber Valley in Antiquity," British School at
Rome, February 28, 2004.
Biers, WR. 1994. "Mass-production, standardized parts, and the Corinthian plastic vase,"
Hesperia 63: 509-16.
Bloch, H. 1938-1939. I bolli laterizi e la storia edilizia romana: Contributi all'archeologia e alla
storia romana. Rome: C. Colombo.
Bloch, H. 1947. I bolli laterizi e la storia edilizia romana: Contributi all'archeologia e alla
storia romana. Indici analitici. Rome: Comune di Roma, Ripartizione antichita e belle
arti.
Brandt, 0. 1993. "Recent research on the Tomb of Eurysaces," Opuscula romana 19: 13-17.
Brun, J.-P. 2004. Archeologie du vin et de l'huile dans l'Empire romain. Paris: Editions
Errance.
Bulmer, M. 1980. An introduction to Roman Samian ware, with special reference to collections
in Chester and the North West. Chester: Chester Archaeological Society.
Ciancio Rossetto, P. 1973. Il sepolcro del fornaio Marco Virgilio Eurisace a Porta Maggiore.
I monumenti romani. Rome: Istituto di studi romani.
Cockle, H. 1981. "Pottery manufacture in Roman Egypt: A new papyrus," Journal of Roman
Studies 71: 87-95.
Curtis, R. I. 1991. Garum and salsamenta: Production and commerce in materia medica.
Leiden: Brill.
Curtis, R. I. 2001. Ancient food technology. Leiden: Brill.
Delaine, J. 1997. The Baths of Caracalla: A study in the design, construction and economics of
large-scale building projects in imperial Rome.Journal of Roman Archaeology Suppl. 25.
Dodge, H., and B. Ward-Perkins (eds.) 1992. Marble in antiquity. London: British School at
Rome.
Duval, P.-M., and R. Marichal 1966. "Un 'compte d'enfournement' inedit de La Graufe-
senque," in R. Chevallier (ed.), Melanges d'archeologie et d'histoire offerts aAndre
Piganiol. Paris: SEVPEN, 1341-52.
Fant, J.C. 1989. Cavum antrum Phrygiae: The organization and operations of the Roman
imperial marble quarries in Phrygia. British Archaeological Reports, Intl. Series S482.
Oxford: BAR.
Flobert, P. 1992. "Les graffites de la Graufesenque: Un temoignage sur le gallo-latin sous
Neron, III," in M. Iliescu and W. Marxgut (eds.), Latin vulgaire-latin tardif: Actes du
IIIeme colloque international sur le latin vulgaire et tardif (Innsbruck, 2-5 septembre
1991). Tiibingen: Niemeyer, 103-14.
Fiille, G. 1997. "The internal organization of the Arretine terra-sigillata industry: Problems
of evidence and interpretation," Journal of Roman Studies 87: m-55.
Fiille, G. 2000. "Die Organisation der Terra sigillata-Herstellung in La Graufesenque: Die
Topfergraffiti," Munsterische Beitriige zur antiken Handelsgeschichte 19.2: 62-98.
empire in the light of instrumentum domesticum. Journal ofRoman Archaeology Suppl.
Helen, T. 1975. Organization ofRoman brick production in the first and second centuries A.D.:
An interpretation of Roman brick stamps. Acta Instituti romani Finlandiae 9. Helsinki:
Academia Scientiarum Fennica.
Helen, T. 1976. "A problem in Roman brick stamps: Who were Lucilla n(ostra) and
Aurel(ius) Caes(ar) n(oster), the owners of the figlinae Fulvianae?" Arctos 10: 27-36.
Heres, T. 1999. "The Caseggiato dei Molini, Part I: The building history," in Bakker 1999a,
16-33.
Hermet, F. 1923. Les graffites de La Graufesenque pres Millau (Aveyron). Rodez: Imprimerie
Carrere.
Hermet, F. 1934. La Graufesenque (Condatomago). 2 vols. Paris: Ernest Leroux.
Herrmann, J. J., Jr., and R. Newman 1995. "The exportation of dolomitic sculptural marble
from Thasos: Evidence from Mediterranean and other collections," in Y. Maniatis, N.
Herz, and Y. Basiakos (eds.), The study of marble and other stones used in antiquity.
London: Archetype, 73-86.
Herrmann, J. J., Jr., and R. Newman 1999. "Dolomitic marble from Thasos near and far:
Macedonia, Ephesos and the Rhone," in M. Schvoerer (ed.), Archeomateriaux: Mar-
bres et autres roches. Bordeaux: CRPAA, 293-303.
Hesnard, A. 1998. "Le sel des plages (Cotta et Tahadart, Maroc)," Melanges de l'Ecole
franraise de Rome 110.1: 167-92.
Hirt, A. M. 2004. "Mines and quarries in the Roman Empire: Organizational aspects, 27
BC-AD 235." D.Phil. thesis, University of Oxford.
H0jte, J.M. 2005. "Archaeological evidence for fish processing in the Black Sea region," in
T. Bekker-Nielsen (ed.), Ancient fishing and fish processing in the Black Sea region.
Aarhus: Aarhus University Press, 133-60.
Juleff, G. 2003. "Mass iron production on Roman Exmoor, England," Minerva 14: 4-5
Karageorghis, V. 1969. "Chronique des fouilles a Chypre en 1968," Bulletin de Corre-
spondance Hellenique 93: 431-569.
Koch, G. 1993. Sarkophage der romischen Kaiserzeit. Darmstadt: Wissenschaftliche Buch-
gesellschaft.
Koch, G., and H. Sichtermann 1982. Romische Sarkophage. Munich: Beck.
Kozelj, G., A. Lambraki, A. Muller, and J.-P. Sodini 1985. "Sarcophages decouverts dans les
carrieres de Saliari (Thasos)," in P. Pensabene (ed.), Marmi antichi: Problemi d'im-
piego, di restauro e d'identificazione. Rome: L'Erma di Bretschneider, 75-81.
Mackensen, M. 1993. Die spatantiken Sigillata- und Lamptopfereien von El Mahrine (Nord
Tunesien). Studien zur nordafrikanischen Feinkeramik des 4. bis 7. Jahrhunderts. Vol. 1.
Munich: Beck.
Mackensen, M. 2000. "Erster Bericht iiber neue archaologische Untersuchungen im sog:
Arbeits- und Steinbruchlager von Simitthus/Chemtou (Nordwesttunesien)," Romische
Manning, W. H. 1985. "The iron objects," in L. F. Pitts and J. K. St Joseph (eds.), Inchtuthil,
the Roman legionary fortress: Excavations 1952-65. London: Society for the Promotion
of Roman Studies, 289-99.
Marichal, R. 1988. Les graffites de La Graufesenque. Paris: Editions du CNRS.
Mees, A. 1994. "Potiers et moulistes: Observations sur la chronologie, les structures et la
commercialisation des ateliers de terre sigillee decoree," in L. Rivet (ed.), Actes du
Congres de Millau, 12-15 mai 1994. Marseille: Societe frarn;:aise d'etude de la ceramique
antique en Gaule, 19-41.
Meijlink, B. 1999. "Molino I.XIll.4," in Bakker 1999a, 61-79.
Meunier, J. 1941. "L'huilerie romaine de Kherbet-Agoub (Perigotville)," Bulletin de la
Societe historique et geographique de Setif 2 (1941): 35-55.
Moritz, L.A. 1958. Grain mills and flour in classical antiquity. Oxford: Oxford University Press.
Moscara, S. 2003. "I motivi iconografici sulle lucerne prodotte a Scoppieto," Rei Cretariae
Romanae Fautorum Acta 38: 153-60.
Nicoletta, N. 2002. "Un vano di lavorazione del complesso produttivo di Scoppieto,"
Rivista di studi liguri 67-68: 209-303.
Nicoletta, N. 2003. "I produttori di terra sigillata di Scoppieto," Rei Cretariae Romanae
Fautorum Acta 38: 145-52.
Oxe, A., H. Comfort, and P. Kenrick 2000. Corpus Vasorum Arretinorum: A catalogue of the
signatures, shapes and chronology of Italian Sigillata. 2nd ed. Bonn: Habelt.
Pasqui, U. 1896. "Nuove scoperte di antiche figuline della fornace di M. Perennio," Notizie
degli scavi di antichita: 453-66.
London: Longman.
Pensabene, P. 2003. "Sul commercio dei marmi in eta imperiale: 11 contributo dei carichi
naufragati di Capo Granitola (Mazara)," in G. Fiorentini, M. Caltabiano, and A.
Calderone (eds.), Archeologia del Mediterraneo: Studi in onore di Ernesto de Miro.
Rome: L'Erma di Bretschneider, 533-43.
Ponsich, M. 1988. Aceite de oliva y salazones de pescado: Factores geo-economicos de Betica y
Tingitania. Madrid: Editorial de la Universidad Complutense.
Rakob, F. 1994. "Das romische Steinbruchlager (Praesidium) in Simitthus," in F. Rakob
(ed.), Simitthus, vol. 2, Der Tempelberg und das romische Lager. Mainz: Philipp von
Zabern, 51-139.
Rakob, F. 1995. "Chemtou/Simitthus: The world oflabour in ancient Rome," in M. Horton
and T. Weidemann (eds.), North Africa from antiquity to Islam. Bristol: Centre for
Mediterranean Studies, 39-44.
Rakob, F. 1997. "Chemtou: Aus der romischen Arbeitswelt," Antike Welt 28.1: 1-20.
Ruger, C. B. 1997. "Zu Marmorschalen in Chemtou," Romische Mitteilungen 104: 379-85.
Setalii, P. 1977. Private domini in Roman brick stamps of the Empire: A historical and
prosopographical study of landowners in the district of Rome. Acta Instituti Romani
Finlandiae 9.2. Helsinki: Suomalainen Tiedeakatemia.
Sim, D. 1995. "Weapons and mass production," Journal of Roman Military Equipment
Studies 6: 1-3.
Steinby, E. M. 1974-1975. "La cronologia delle figlinae doliari urbane della fine dell'eta
repubblicana fino all'inizio del III secolo," Bullettino della Commissione archaeologica
communale di Roma 84: 7-132.
Steinby, E. M. 1977. Lateres Signati Ostienses. Vol. 2, Tavole. Acta Instituti Romani Fin-
landiae 7. Rome: Bardi.
Steinby, E. M. 1993. "L'organizzazione produttiva dei laterizi: Un modello interpretivo per
l'instrumentum in genere?," in Harris 1993: 139-43.
Steinby, E. M., and T. Helen. 1978. Lateres Signati Ostienses. Vol. 1, Testo. Acta Instituti
Romani Finlandiae 7. Rome: Bardi.
Thuault, M. 1996. "Un atelier campagnard: Le Rozier," Dossiers d'Archeologie 215: 18-19.
Thuault, M., and A. Vernhet. 1986. "Le Rozier," in C. Bemont and J.-P. Jacob (eds.), La
terre sigillee gallo-romaine: Lieux de production du Haut Empire: Implantations, pro-
duits, relations. Paris: Editions de la Maison des Sciences de l'Homme, 110-13.
Tomber, R. 1987. "Evidence for long-distance commerce: Imported bricks and tiles at
Carthage," Rei Cretariae Romanae Fautorum Acta 25-26: 161-74.
Trimble, J. 2000. "Replicating the body politic: The Herculaneum Women statue types in
early Imperial Italy," Journal of Roman Archaeology 13: 41-68.
Tykot, R. H., J. J. Herrmann Jr., N. J. van der Merwe, R. Newman, and K. 0. Allegretto
2002. "Thasian marble sculptures in European and American collections: Isotopic and
other analyses," in J. J. Herrmann Jr., N. Herz, and R. Newman (eds.), Interdisciplinary
studies on ancient stone. London: Archetype, 188-95.
Vernhet, A., and C. Bemont 1991. "Un nouveau compte de potiers de La Graufesenque
portant mention des flamines," Annales de Pegasus 1: 12-14.
Vernhet, A., and C. Bemont 1993. "Le graffite des nones d'octobre," Annales de Pegasus 2:
19-21.
Waelkens, M. 1982. "Carrieres de marbre en Phrygie (Turquie)," Bulletin des Musees royaux
d'art et d'histoire, Bruxelles 53: 33-55.
Waelkens, M. 1986. Die kleinasiatischen Tursteine: Typologische und epigraphische Un-
tersuchungen der kleinasiatische Grabreliefs mit Scheintur. Mainz: Philipp von Zabern.
Waelkens, M. 1990. "Technique de carriere, prefa<;:onnage et ateliers dans les civilisations
classiques (mondes grec et romain)," in M. Waelkens (ed.), Pierre eternelle: Du Nil au
Rhin: Carrieres et prefabrication. Bruxelles: Credit Communal, 53-72.
Waelkens, M., P. de Paepe, and L. Moens 1988. "Patterns of extraction and production in
the white marble quarries of the Mediterranean: History, present problems and
prospects," in J. Clayton Fant, Andent marble quarrying and trade. British Archae-
ological Reports, Intl. Series S453. Oxford: BAR, 81-116.
Wilson, A. I. 2001a. "The water-mills on the Janiculum," Memoirs of the American Academy
at Rome 45: 219-46.
Wilson, A. I. 2001b. "Ti. Cl. Felix and the date of the second phase of the East Baths," in L.
M. Stirling, D. J. Mattingly, and N. Ben Lazreg (eds.), Leptiminus (Lamta): A Roman
port city in Tunisia. Report no. 2. Journal of Roman Archaeology Suppl. 41. Ann Arbor,
MI: JRA, 25-28.
Wilson, A. I. 2005. "The economic impact of technological advances in the Roman con-
struction industry," in Elio Lo Cascio (ed.), Innovazione tecnica e progresso economico
nel mondo romano. Pragmateiai. Atti degli Incontri capresi di storia dell'economia antica
(Capri 13-16 aprile 2003). Bari: Edipuglia.
Wilson, A. I. 2006. "Fishy business: Roman exploitation of marine resources," Journal of
Zimmer, G. 1982. Romische Berufsdarstellungen. Deutsches Archaologisches Institut. Berlin:
Mann Verlag.
CHAPTER 16
METALWORKING
AND TOOLS
CAROL MA TTUSCH
WHAT we know about metalworking in the ancient Mediterranean world derives

primarily from literary testimonia and from depictions of bronze- and ironworkers
on Greek vases and Roman grave monuments (Zimmer 1982: 37-40, 179-96). The
authors and artists were probably not practitioners of these arts, and as a conse-
quence they were only superficially familiar with the industry. Today's classicists
and archaeologists also tend to suffer from a lack of direct experience with the
technologies. Furthermore, the major sources of our information have been liter-
ary and pictorial testimonia, because so little archaeological evidence for ancient
workshops has been excavated-none, in fact, until the mid-twentieth century. Such
workshops were not usually located in a city center or a sanctuary, where excava-
tions have focused, but on the periphery, and so relatively few ancient metalworking
establishments have been excavated to date. For example, no signs of production
have been found during the lengthy excavations of the sanctuary at Delphi, a site
that has yielded hundreds of metal dedications of all kinds, whereas about 25
workshops and dumps have been identified around the Acropolis, the agora, and
the Ceramicus of ancient Athens (Mattusch 1977a; Zimmer 1990). Such sites are
found unexpectedly, and they may not be correctly identified, since the finds consist
of dumped debris that may have been shifted several times. Today, we consider both
the evidence of surviving workshop materials and the metal finds from ancient
cities and sanctuaries. Learning to read the surfaces of a bronze for information
about its production has become part of the process as well. But we still rely heavily
on the secondary evidence of literary testimonia and artistic representations in
piecing together the practices associated with ancient metalworking establishments.
This discussion focuses for the most part on the production of bronze sculpture,
METALWORKING AND TOOLS 419
since that field of metalworking, with its expensive, complex, and often very large
products made use of nearly all the techniques applied to the production of more
practical, everyday objects and the use of other metals and alloys.
HEPHAESTUS AND ANCIENT

METALWORKERS
According to Homer, the workshop of Hephaestus, god of fire and of metal-

working, was equipped with an anvil set on a block, as well as with crucible, bellows,
hammer, and tongs (II. 18). Similarly, hammers and anvils are the most common
accessories in vase paintings depicting armorers. The tondo of the early fifth-
century B.C. Berlin Foundry Cup (figure 16.1) shows Hephaestus presenting armor
for Achilles to Thetis, the hero's mother. Hephaestus sits beside his anvil, a double-
headed hammer in hand. Gems are also engraved with similar images depicting an
armorer wielding a hammer over a helmet, a shield, or a metal vase. The scene is
a familiar one, and it appears both on vases, and-later on-in Pompeian wall
paintings (cf. figure 2.3). The scene may include the god's tools-tongs, a bowed
saw, and a file. On one vase, Hephaestus polishes a shield with pumice (Boston,
Museum of Fine Arts: red-figure vase painting no. 13.188), examples of which have
been found in ancient workshops, stained green by corroded bronze. Human
smiths are shown with various rasps, scrapers, files (as on a red-figure kylix in
Oxford: Ashmolean G 267), perhaps a punch or a drill, and hammers, which they
use to fashion helmets, shields, greaves, and knives. Ironworkers are represented
holding tongs and thrusting blooms of iron toward a fire (as on London, British
Museum: black-figure amphora B 507), hammers and anvils nearby. Tall crucible-
furnaces and bellows appear in many such scenes.
Clearly, Greek metalworkers, like their patron deity, were skilled in extracting
gold, silver, copper, tin, lead, and iron, and in forging, hammering, and casting a
wide variety of objects in these metals. Gold and silver were used primarily to make
jewelry, coins, and vessels. Objects were cast, hammered, cut, and soldered. Forms
of decoration included repousse, stamping, filigree, granulation, engraving, niello,
and inlaying. Lead was used to make various utilitarian objects that had to be shaped
by casting, resistant to corrosion, and inexpensive, such as architectural clamps,
water pipes, cooking utensils, and occasional votive plaques and statuettes. The
Greeks and Romans also continued the ancient Near Eastern practice of inscribing
curses on lead tablets (defixiones).
Ironworking was introduced to Greece from Anatolia sometime around 1000
B.C. There were major sources of iron ore in Macedonia, Euboea, Attica, the
Peloponnese, the Aegean islands, and Crete. Objects such as tools, spearheads and
Figure 16.1. Berlin Foundry Cup, interior, early fifth century B.C. Berlin, Anti-
kensammlung F 2294. ( Courtesy of Isolde Luckert, Antikensammlung, Staatliche Museen
zu Berlin, Preussischer Kulturbesitz.)
arrowheads, daggers, knives, axes, architectural clamps, pins, and vessels were
produced by smelting and hot-forging. In some places, there is evidence that iron
was worked together with bronze by the same craftsmen in the same workshops
(Mattusch 1991a: 389).
EVIDENCE FROM VASE p AINTINGS
Representations of ironworkers in Greek vase painting show that the tools used for
working hot iron were anvil, hammer, and tongs, as on an Attic kylix of around 500
B.C. (Berlin Antikenmuseum no. 1980.7). As for bronze workers, the illustrations
on the Berlin Foundry Cup (figures 16.1-3) provide the best ancient description of
Figure 16.2. Berlin Foundry Cup, exterior, early fifth century B.C. Berlin, Antiken-
sammlung F 2294. (Courtesy of Antikensammlung, Staatliche Museen zu Berlin, Pre-
ussischer Kulturbesitz.)
Greek metalworking activities in both smithy and foundry (Mattusch 1982), and
they show the close link between the two. The scenes on this cup also make specific
reference to hierarchies within the foundry, and to two artistic styles that were
available on the market during the early fifth century B.C. In the tondo, Hephaestus
works alone, whereas on the exterior of the cup, foundry workers are engaged in
joining the pieces of one statue and cleaning the surface of another one. Clearly
metalworking activities of all kinds could be carried out in a single shop.
On one side of the cup, a tall metallurgical furnace is being heated, and the
sections of a naturalistic bronze statue of an athlete are being joined. On the other
side, workmen wielding curved rasps are polishing a large archaic bronze statue of
a warrior. Tools and equipment illustrated on the vase include models for body-
parts, hammers, rasps, a two-handled saw, an anvil, and a poker and bellows for the
furnace. The bearded senior metalworkers wear either skullcaps or waistcloths, the
equivalent of aprons, whereas their junior assistants are nude, beardless, and hat-
less. To judge from the two statues represented in different styles on the cup, the
archaic style continued to be in demand even after a more naturalistic style had been
introduced. There are numerous other representations of metalworkers on Greek
vases, most notably helmet makers, depicted as squatting naked and shaping a hel-
met by hammering it over a small anvil (Thaliarchos Painter, 480 B.C.; Paris, Petit
Palais no. 381; Beazley 1963: 81.1, 1610 ), or finishing the surface with a file (Antiphon
Painter, 485 B.C.; Ashmolean Museum, Inv.G.267/V.518; Beazley 1963: 336.22, 1646).
Figure 16.3. Berlin Foundry Cup, exterior, early fifth century B.C. Berlin, Anti-
kensammlung F 2294. (Courtesy of Antikensammlung, Staatliche Museen zu Berlin,
Preussischer Kulturbesitz.)
EARLY GREEK METALLURGY
Bronze, in the classical world an alloy of copper with about 10 percent tin, was
known in the Mediterranean region from approximately 3000 B.C. onward. The
Greek word for both bronze and copper was chalkos; in Latin it was aes. With bronze,
the Greeks and Romans made jewelry, mirrors, pins and needles, tripods, lamps,
vessels, tools, nails and other hardware, medical instruments, candelabra, dinner-
ware, official weights and ballots, coins, weapons and armor, furniture parts, archi-
tectural ornaments, inscriptions, figurines, and large-scale statuary. They shaped it
by casting, beating and annealing, or repousse (beating from behind, over a bed of
pitch) and assembled sections by soldering or riveting. Decoration involved chasing,
inlay, patination, differential alloying, and gilding.
Solid bronze statuettes and other household objects were being cast by the lost-
wax process from the time of their first appearance in ninth-century Greece. Wax
was cut, rolled, pinched, and carved; the wax body parts were stuck together; the
wax model was invested with a clay mold; the mold was baked to burn out the wax;
and bronze was poured into the mold to replace the wax. The base was often cast
along with the figurine, occupying the space used as a funnel during the pour. Each
work produced in such a manner was, of course, unique.
Olympia has yielded the largest number of these early offerings, but the Her-
aion at Argos has also produced many of these generic figurines, including large
numbers of birds, cows, deer, horses, and humans. There is much repetition, par-
ticularly among those bronzes that come from a single site, both because the
demand for them depended on their adherence to the particular types that suited
their votive function, and because they were produced in quantity. Groups of like
figures no doubt come from a single workshop. Various idiosyncratic ways were
found by which to produce an item in demand quickly, again and again, and serially
produced bronzes soon became the norm (Mattusch 1996: 18-23).
EGYPTIAN INFLUENCE ON GREEK

METALWORKING
The sophisticated metalworking industry in Greece owed much to Egypt, where

the Greeks not only learned about metallurgical materials and techniques, but also
about the design and production of monumental stone architecture and sculpture.
Greeks settled at Naucratis in the Nile delta during the seventh century B.C. By that
time, the Egyptians had been using the same basic formulas for representing hu-
man figures, as well as the same techniques for carving and casting them, for more
than 2,500 years. Metal objects of all kinds were being cast and hammered in large
numbers, many of them richly inlaid, others sheathed with gold or silver (Ziegler
1987: 85-101; Bianchi 1990: 61-84).
During the Twenty-Sixth Dynasty, Psammetichus I (664-610 B.C.) encouraged
intercourse between the Greek colonists and the Egyptians (Herodotus 2.154). The
Greeks must have marveled at the technological virtuosity and the richness that they
saw in Egyptian buildings, images, materials, and technology (Bianchi 1990). The
legendary Daedalus was said to have made statues that looked Egyptian (Diodorus
1.97.6); Pausanias thought that his works were "rather odd ... but inspired" (2,4.5).
Diodorus (4.76.3) said that he was the first to bring a degree of naturalism to stat-
ues, "for which men naturally admired him." Diodorus also reports (1.98) that "the
most renowned of the ancient sculptors-Telecles and Theodorus, the sons of
Rhoecus-visited with [the Egyptians]." He also explains that they worked accord-
ing to the Egyptian system, the proportions of a figure's parts being determined
by formula. It is no accident that Daedalus should be associated with Crete, which
lay on the crossroads of Mediterranean trade, and where resources and ideas were
introduced directly from Egypt and the Near East. Daedalus made a lifelike bronze
cow for Queen Pasiphae and the labyrinth for King Minos, before falling out of
favor and having to fly to Athens on wings that he designed and made from feathers
glued together with wax.
Like Crete, the island of Samos in the eastern Aegean was a trading center and
as such also a focal point for the introduction of new ideas. Although ancient
authors tend to confuse the lineage of Rhoecus, Theodorus, and Telecles, they are
unanimous in associating them with Samos and with a wide range of achievements
in the arts, design, and engineering. Egypt is mentioned as well. The innovations
attributed to them were not limited to a single medium. The chief authority for the
introduction of large-scale bronze casting to Greece in the sixth century B.C. is
Pausanias, writing in the second century A.D. He says that the first to melt bronze
and cast statues were the Samians, Rhoecus the son of Philaeus and Theodorus the
son ofTelecles (8.14.8, 9.41.1), but he admits that he has not seen any of their work
(10.38.6). Pliny the Elder's assignment of the introduction of clay modeling to these
two artists, though tentative (HN 35.152), is appropriate to work conducted in a
foundry. Pliny (HN 34.83) refers to a statue that Theodorus cast of himself hold-
ing a file, thereby implying that Theodorus was seen as both an artisan and an
artist. Pliny attributes production of the labyrinth on Samas to Theodorus (HN
34.83), and Herodotus (3.41) mentions a seal ring made by Theodorus, adding that
Rhoecus worked on the construction of the Heraion on Samas (3.60 ). Rhoecus and
Theodorus may or may not be actual historical figures, but these passages clearly
suggest that sixth-century Samas was responsible for a good measure of innovation
by metalworkers, with the obvious link to artistic activity. Like the renowned sixth-
century Samians, the great artists of the Classical period are credited with working
in more than one medium. The fifth-century B.C. artist Phidias, for example, made
bronze statues, chryselephantine (gold-and-ivory) statues, and even a gilded wooden
statue with marble face, hands, and feet (Pausanias 9.4.1). It is likely that most
Greek and Roman metalworkers were similarly adaptable.
In statuary, the market was driven by a demand for new types and styles, which
was followed by the development of the necessary technology. The tensile strength of
bronze and the lighter weight of hollow-cast forms, which might seem to encourage
the casting of statues in active poses, were not immediately exploited. Instead, the
earliest Greek freestanding sculptures in both bronze and stone were frontal stand-
ing figures with their arms close to their sides, very much like the self-contained
sculptures that the Greeks had seen in Egypt. The poses soon became looser, but the
tensile strength of bronze never seems to have driven stylistic change in Greece.
Rather it seems to have been the introduction of athletic victors' statues in the sixth
century that eventually led to more varied and active poses (Mattusch 1988: 99-100,
128-29, 212-13).
TECHNICAL PRACTICES
The lost-wax process of casting (described below) was used throughout the ancient
world for both utilitarian objects and statuary. The advantages are obvious: a model
can be re-used in case of casting failure, or applied to the production of multiple
copies, and production costs are generally lower than with unique objects. Large,
thin-walled objects such as armor and vessels were produced by a combination of
casting and hammering. Production of armor was a major Athenian industry, and
even well-known artists participated, such as Phidias, who was said to have once
also painted a shield (Pliny, HN 35.54). Helmets were first cast in rough form, then
hammered into their final shapes and finished with rasps.
Bronze, which could be cast into complicated shapes and was resistant to cor-
rosion, was particularly suitable for cosmetic instruments, serving vessels, and fur-
niture fittings. Both standing mirrors and hand mirrors were being cast in Greece
during the sixth century B.C. By the fourth century B.C., both types had been
largely replaced by the round box mirror, consisting of a cast disc, polished on the
front, that was hinged to a hammered or cast lid, to which a repousse decoration
might be soldered with lead.
When the Greeks first made bronze vessels, during the Geometric period, they
used them for dedications and for funerary offerings; later, bronze vessels were the
prizes for athletic victories. These vessels could be very large, technically sophisti-
cated products. A bronze tripod cauldron made on Samos during the seventh cen-
tury cost six talents, a tenth of the profit from a merchant trading expedition to
Spain (Herodotus 4.152). The cauldron was decorated with griffin protomes, and it
was supported by a tripod that consisted of three kneeling bronze figures seven
cubits in height (ca. 3.5 m). It is not known whether the cauldron was cast or ham-
mered, but the griffins and the kneeling figures were no doubt cast as multiples,
which was the norm by the seventh century. Herodotus describes several cauldrons
that would have been supported by three or more similar figures. He mentions two
other huge cauldrons with figures around the rim, one made by the Lacedaemo-
nians for King Croesus of Lydia that held 12,275 litres, and one made for Ariantes of
Scythia that held twice as much (Herodotus 1.70, 4.81). Metalwork on this scale,
undoubtedly weighing hundreds of kilograms, represents a remarkable repertory of
skills, even if elements of the monument were cast and worked separately.
Physical evidence from the Orientalizing period suggests that dedications of
colossal tripod cauldrons were not unusual. During the seventh century, the groups
of protomes that were used to decorate these cauldrons might be produced in a
series, or, one might say, in editions-for speed as well as to ensure that they all
looked alike (Mattusch 1990). Griffins' heads were particularly common attach-
ments, first hammered from sheet-bronze, and then hollow-cast. Some particularly
large cast heads were fitted onto metal necks hammered over cores.
Three large seventh-century B.C. griffins' heads from Olympia, all produced by
the same unusual method (Mattusch 1990 ), illustrate the sophisticated nature of the
early founder's craft. They were formed from nearly identical groups of wax slabs,
to each of which were added wax ears, tongue, and knob. Finally, scales were
punched in the surfaces of all three heads with a single set of tools, further proving
that these protomes were produced simultaneously. Each wax head was then in-
vested and cast, after which it was fitted to a neck, probably made of hammered
sheet metal over a core. The proportions of the griffins show that the tripod caul-
dron that they adorned stood between 4.5 and 5.5 m in height. Part of another
remarkably large dedication is preserved in the form of a hollow-cast column
consisting of entwined snakes, now in the hippodrome in Istanbul. This column
originally supported a tripod cauldron in Delphi that approached 6 m in height, set
up to honor the Greek cities that had defeated the Persians at Plataea in 479 B.C.
(Herodotus 9.81; Thucydides 1.132; Pausanias 10.13.9). By the Classical period, Greek
bronze vessels of a more manageable size were being sold throughout the Medi-
terranean world as luxury items. Examples have been found in Greece, southern
Italy and Sicily, Asia Minor, northern Europe, North Africa, and at sites along the
coast of the Black Sea.
Bronze vessels were usually made by hammering discs of sheet-bronze into
hollow shapes that might then be turned on a lathe for further shaping and de-
tailing. Occasionally the bodies were cast. Handles, very often feet, and sometimes
the rims of vessels were separately cast, and were mass-produced. Feet and handles
were soldered or riveted to the body of the vessel, which might be decorated by the
addition of other cast features. Sometimes the walls of bronze vessels were elabo-
rated by embossing (repousse), engraving, inlay, or gilding. Certain types of vessels
were particularly well suited to manufacture in bronze, such as tripod cauldrons,
double-handled buckets (situlae), and bowls with human- and animal-figured
handles (paterae). Other bronze vessels parallel the forms of contemporary pottery,
including craters, hydrias, amphoras, oenochoes, and the like. The range of plastic
decoration that can be realized in bronze is vast and is usually more fully exploited
than in pottery. During the Hellenistic and Roman periods, tastes tended toward
expensive and elaborately decorated elegant vessels, whether in bronze or in silver
or gold. Major centers for the production of later metal vessels were apparently
located in Macedonia, Thessaly, Asia Minor, South Italy, and, of course, Rome.
During the first century A.D., the lathe was used to produce elaborate rim and base
shapes both by folding and shaping the sheet metal with tools as it turned at high
speed ("spinning"), and by trimming and cutting thicker rim or base castings with a
chisel (Mutz 1972; Hauser and Mutz 1973; Cave 1977). The small, elaborate con-
tainers prepared in this manner resemble contemporary blown glassware vessels,
and the metalworking technique probably was developed under the stimulus of the
new glassworking technique-an interesting example of imitation, competition,
and innovation across materials and workshops. The uniformity of the vessels
suggests either the activity of a small number of workshops or the free exchange of
technical information among them.
TECHNIQUES OF STATUARY
PRODUCTION
As large, expensive, usually public monuments, Greek and Roman statues provide
a glimpse at the most sophisticated techniques and accomplishments of metal-
working. Freestanding statues made of metal were being produced in the Greek
world as early as the eighth century B.C., and their technical evolution was not
unlike that of utilitarian objects. According to Pausanias (3.17.6), the earliest tech-
nique involved hammering metal sheets into the shape of a figure and riveting them
together; Daedalus, the legendary master to whom important discoveries and early
works in various media were often attributed, made such a statue. Ancient authors
refer to this method as sphyrelaton. Three such statues, called sphyrelata, were found
in a small late-eighth- or early-seventh-century B.C sanctuary at Dreros in Crete
(Romano 2000). They were made of cylindrical forms of hammered sheet metal,
riveted at the seams, just as Pausanias describes them.
Pausanias (9.10.2) also mentions a sixth-century B.C. image of which there are
two versions, one in bronze, presumably cast, and one carved in wood. "The statue
[of Apollo Ismenius in Thebes] is the same size as the one in Branchidae [Didyma],
and the form is no different; whoever has seen one of these statues and learned
its sculptor does not need much skill to see that the other one is also a work by
Canachus. They differ in that the one in Branchidae is bronze, whereas the one in
Thebes is cedar-wood." If we think of ancient statuary as being produced in edi-
tions, what Pausanias says about sculptures by the archaic-period sculptor Ca-
nachus is perfectly clear. He also implies that buyers choose the medium. Canachus
could have taken molds from the wooden statue to use in making a bronze version
of the same thing, or the production staff could have used molds taken from the
artist's model to produce a working-model for a bronze and a plaster model for a
wooden statue. In the same vein, a passage about a wooden statue that was made
in two halves, one by Theodorus working in Ephesus, and the other by Telecles
working in Samos, also reflects the use of a single original model (Diodorus Siculus
1.98).
Surviving mold fragments, as well as finger marks, brush marks, and tool marks
on the inside surfaces of surviving bronze statues, show that the indirect process, in
which the original artist's model is not destroyed in the production process, is the
form oflost-wax casting that was the most widely used during classical antiquity.
Large sculptures, even some statuettes, were cast hollow and in pieces, which were
then joined by flow welding (Mattusch 2005). The original design for a statue could
be made of almost anything: it could, in fact, have been no more than a drawing by
the artist. Usually, however, the artist's model was made in clay or plaster so as to
allow for addition and subtraction. After the design had been modeled on a large
scale, skilled artisans took piece casts from it in clay. These molds were lined with a
thin layer of wax that could be removed and assembled with others to produce a
hollow wax working-model. The finished wax working-model was cut up, and the
statue was cast in pieces inside clay molds by means of the lost-wax method. The
working-model could be cast in bronze just as it came from the molds, or it could
be reworked before casting, which might involve adjusting limbs or adding or
embellishing particular features, such as ears, free-hanging locks of hair, or a beard.
Should a buyer request a unique piece, a wax working-model made from the molds
could be altered extensively by the artist before casting. After casting, the separately
cast parts of the statue were put together. The original model, like the molds taken
from it, could be re-used to make as many copies as were required by the artist. As is
the case today, each phase of the process-molding, making the wax, pouring the
bronze, cleaning, joining, and patinating-was handled by different individuals.
Horace (Ars P. 32-35) describes a craftsman who can model fingernails and make
wavy locks but is unhappy because he cannot make the whole statue.
Because of the extraordinary difficulty and risk of casting large, complex
forms, statues were cast in pieces. A vivid picture of the problems and dangers
involved in casting a large statue in one piece can be found in Benvenuto Cellini's
account of the casting of his Perseus in 1554 (Autobiography 2.73-78). In antiquity,
human nudes were usually cast in six or seven major pieces, their fabricators taking
advantage of natural anatomical divisions to conceal flow-welded joins (Mattusch
2005: 133-35, 189-337), while draped statues might be cast in fifteen or twenty
pieces, the joins concealed by garments. The joints are often marked by a row of
ovals, sometimes reinforced by inserted rectangular patches. The rectangular pins
or chaplets that stabilized the core within the investment while the bronze was
being poured sometimes remain in place. The division of statues at mid-neck is
characteristic of archaic and classical-period bronzes from the Greek world, while
Roman bronze workers tended to place the seams along the hairline or jawline so as
to conceal the joins more easily. During the archaic and classical periods, the heads
of Greek statues are often thick and irregular castings, indicating that they were
modeled directly and/or worked over in the wax, making them unique works. Thin
regular walls throughout a statue are the norm in the Roman period, along with
evidence on the inner surface of the bronze for application of the wax to piece
molds. Both features are clear indications that a work was cast by the indirect lost-
wax method, with minimal reworking of the wax working-model before it was
invested with clay for the pour.
An artist's direct work on the wax working-model was often what distinguished
one statue from another. The classical-period Riace Bronzes (figures 16.4-5) il-
lustrate the advantages of using the indirect lost wax practice (Mattusch 1996: 62-
67 ). The two statues have almost exactly the same contours and dimensions, and
their profiles, poses, and gestures are nearly identical. In other words, both statues
are likely to have come from the same model. The wax working-models made from
the molds were differentiated by pulling one statue's feet farther apart than the
other's, and by adding individualized wax hair and beards to the two figures. In-
scriptions providing specific identities would have completed the production of
two very different editions of one original model. Statue-types were standardized,
making them easily recognizable as generals, philosophers, or the like. Such statues
were even sometimes re-used, as was the case with one that Pausanias saw in Argos
(2.17.3), which had been cast in Greek times to represent the mythical Orestes but
was renamed in Roman times so that it could serve as the emperor Augustus.
Two epigrams allegedly written in the fifth century B.C. reflect the use of the
indirect lost wax process of casting. "Perhaps Myron himself would say this: I did
not form this heifer, but of her I modeled the statue" (Anth. Pal. 2.9.718). "Myron
feigned a cow with his own hands that was not formed in molds, but that turned to
Figure 16.4. Riace Bronze A, probably mid-fifth century B.C. H 2.05 m. (Courtesy So-
printendenza Archeologica della Calabria.)
bronze through old age" (Anth. Pal. 2.9.716). The implication of the first epigram
is that Myron took molds from a cow and used them to cast his statue. Both epi-
grams, however, may mean simply that Myron's cow looks so realistic that it is
hard to believe it could have been made by man. Had Myron actually developed a
system of molding from life he would no doubt have received the kind of recog-
nition accorded to Lysistratus a century later for using or, more likely, streamlining
the indirect lost wax process.
A passage written by Plutarch ( Quaest. conv. 2.3.2, 636C; De prof virt. 17.86A)
about the mid-fifth century B.C. sculptor Polyclitus has generated much debate,
but it indicates that the artist's model for a bronze was generally modeled in clay:
"As for the arts, first they model the unformed and shapeless (material), and later
they articulate all the details; whence Polyclitus the sculptor said that the work is
the most difficult when the clay is on the nail." This passage probably refers to the
finishing of the artist's model. Molds were then taken from the clay model and
Figure 16.5. Riace Bronze B, probably mid-fifth century B.C. H 1.96 m. (Courtesy So-
printendenza Archeologica della Calabria.)
lined with wax to make a working-model, which could be customized or not, and
from which a bronze would be cast.
Lysippus of Sicyon was said to have produced an astounding total of 1,500
statues during his career in the fourth century B. C. ( Pliny, HN 34.37). He was said to
have used nature as his model (Pliny, HN 34.61). Lysippus differed from most artists
in that he came from a family of bronze workers, and he was obviously familiar with
the technology involved in producing statuary. His brother Lysistratus is said to
have been the first Greek to produce portraits by pouring wax into plaster molds
taken from living models, and to have developed a similar means of taking casts
from statues (Pliny, HN 35.153). It is logical to conclude from this information that
Lysippus and Lysistratus either used human beings as their models or used finished
models which, when molded, produced waxes that needed no additional work, but
were ready to invest and cast. This would mean that the artist did not need to finish
the waxes, and that the production of a statue could be left entirely to technicians.
The savings in time, effort, and money would have resulted in increased produc-
tion, and might explain how Lysippus could have produced so many statues during
his career.
There are other references to use of the indirect process in the fourth century
B.C., one by Plutarch (Mor., De soll. an. 984a) about two men dispatched to Sinope
by Ptolemy Soter (ca. 367-282 B.C.) in order to bring him back a statue: "Of the
two statues, they thought that they could take with them the one of Pluto, but that
they should take a mold of the Kore and leave behind the actual statue." This is
perhaps more interesting as an example of transporting molds of large-scale stat-
uary for the purpose of reproduction.
Alloys and Color

So few Greek or Roman bronzes have been systematically analyzed that questions
about the value of alloys in determining the origins of bronzes cannot be adequately
addressed. Quantitative analyses of alloys, however, have given an important new
dimension to the study of ancient statuary (Mattusch 2005: 136-40, 333-34). Copper,
tin, and lead are the principal intentional ingredients of bronze. Statuary bronzes
tend to contain between 70 and 90 percent of copper and less than 10 percent of tin,
but lead content ranges widely, from less than 1 percent to nearly 30 percent. Lead
lowers the melting point of the alloy and increases the fluidity of molten bronze
during casting, facilitating the casting of large bronzes. A sufficient percentage of
lead also promotes fusion during the joining of cast sections by flow welding. Trace
elements are elements that appear by chance and in very small quantities in an alloy,
because they were present in the ores from which the major elements were smelted.
Unlike the principal elements, they can provide reliable matches between alloys and
can be used to identify objects cast from a single batch of metal, particularly when
the bronzes also match in design, style, and subject matter.
The eyes of Greek and Roman statues were usually inset with bone (or ivory)
and stone (Mattusch 1996: 24, pl. 1), though occasionally they were simply marked
in the wax and reproduced in the bronze. Bronze plates cut with fringed edges
sometimes served as eyelashes, while holding the eyes in their sockets. Eyebrows,
lips, nipples, fingernails, toenails, wreaths, and bracelets could be inset or added in
copper, tin, and even gold or silver (Hellenkemper Salies et al. 1994).
Apart from a number of ancient bronzes that are known to have been gilded,
such as the equestrian statue of Marcus Aurelius on the Capitoline Hill in Rome,
little is known about the patination of ancient bronze statuary beyond what the
ancient authors tell us-and they were not experts in the field. Interest in this subject
is growing, however, and the widespread use of mechanical methods for cleaning
ancient metals has reduced the damage to the metal caused in the past by inva-
sive chemical cleansing. As a result, it now appears that the Riace Bronzes ( figures
16.4-5) originally had black skin (formigli 1985), and other bronzes are showing
signs of having once been patinated. This evidence lends some credence to remarks
by ancient authors that bronzes had different colors, though they tend to ascribe the
color to alloy rather than to patination. Pliny advises us that Delian bronze differs in
color from Aeginetan bronze (HN 34.10), and an alloy that was preferred for
portrait statues was "liver-colored" (HN 34.8). Plutarch (Quaest. conv. 5-1.2)
mentions a statue whose face was a different color from the body, as is often seen
today in sculptural bronzes by such artists as J. Seward Johnson. Corinthian bronze
was said to be whitish because of having some gold or silver in the alloy.
Evidence from Workshops
Olympia
The surviving physical evidence for metalworking in the ancient Mediterranean
world consists primarily of workshop debris, which may be more reliable than what
nonspecialists reveal about the industry in literary testimonia and in ancient il-
lustrations. The excavations at Olympia have uncovered workshops and workshop
debris from the production of statuettes that were dedicated in the Sanctuary of
Zeus (Heilmeyer 1969). The workshops lay outside the sanctuary in a sand-filled
ravine that had once been a streambed. They were temporary installations, con-
sisting of simple furnaces made of stone, mudbrick, and baked clay, from which
canals led to molds that were packed in the sand. Debris tossed back into the casting
area after use and scattered around the edges of the Sanctuary of Zeus consists of
charcoal, bricks with vitrified inner surfaces, lumps of bronze, clay mold fragments,
bellows nozzles, and clay stoppers. There are bronze funnels cut from the pour,
gates that are both round and rectangular in section, and clay crucibles and molds,
as well as failed castings of statuettes. Some of the foundries date from as early as
the Geometric period. Theophrastus (Lap. 16) refers to the fuel used at Olympia:
"Among the materials that are dug because they are useful, those known as coals are
made of earth, and, once set on fire, they burn like charcoal. They are found in
Liguria ... and in Elis as one approaches Olympia by the mountain road; and they
are used by those who work in metals."
Athens
Metalworking shops were located in and around the agora of Athens from the sixth
century B.C. to the sixth century A.D. (Mattusch 1977a). Their centrality suggests
the importance of the metalworking industry in ancient Athens. Some work areas
had hearths, others had casting pits, and they all left behind their debris. The finds
excavated in these establishments and in the dumps associated with them include
broken clay molds, vitrified bricks from furnaces, scraps of waste metal, and slag.
Some of the deposits contained debris associated specifically with bronzeworking
or ironworking, while some deposits yielded iron, bronze, and even lead debris.
The industrial area around the Temple of Hephaestus, overlooking the agora,
has produced many of these installations and artifacts. An inscription of around
421-415 B.C. even records the purchase of copper, tin, lead, wood, and charcoal for
the completion of the (bronze) cult statues and for the pedestal for the Temple of
Hephaestus (JG I2, 370-71). A curse directed at two bronze-smiths and a woman
was inscribed on a lead tablet (Athens: Agora Excavations no. IL 997) and hidden
beneath the floor of the smithy in the fourth century B.C.
Although the traces of ironworking from the area of the Athenian agora are
substantial, only one actual iron smithy has been excavated. The same holds true
for the single bronze smithy, which contained only a hearth, some slag, and the
hidden curse tablet. The trash from bronze foundries provides far more revealing
information about processes. All were temporary establishments, some for statuary
and others for small utilitarian objects. Clay molds from the earlier Greek period
are thick and heavy, with inclusions such as sand, straw, hair, and shells, whereas
molds from foundries of the Roman period are thinner and more lightweight.
Molds were propped up for baking in the pit that would later serve for casting. No
traces of actual furnaces have been found, although curved vitrified bricks can be
identified as having come from the furnace lining.
A fragmentary inscription from the Acropolis of Athens contains the records of
the overseers who were in charge of producing the colossal bronze Athena Pro-
machus designed by Phidias and erected on the Acropolis between 460 and 450 B.C.
The inscription lists the wages paid and the materials purchased at various points
in the project. In one year, the large sum of 78,110 drachmas seems to have been
available for the project. Charcoal and firewood for the furnace are mentioned
repeatedly, as are clay and hair, for which 26 drachmas were spent in one year. The
clay would have been for core or mold material, and the hair was probably mixed
with the clay to reduce shrinkage and/or strengthen the molds. Wax is listed only
once in the text, if the word keros is correctly restored (Mattusch 1988: 166-69).
Copper, tin, and silver are purchased, the latter for the decoration of the statue.
The cost of a talent of copper was only 35 drachmas and one or two obols, but the
average price of tin was 232! drachmas per talent. Wages were paid by the day, by
the prytany (administrative period slightly longer than a month), and by the job. In
one year, 6,600! drachmas were paid in wages; in another year, 10,100 drachmas
were paid. This single large statue was a very expensive undertaking, nearly on the
scale of an architectural project.
Corinth
Corinthian bronze was highly prized for its color, and the Corinthian metalworking
industry was highly respected. Pausanias (2.3.3) wrote that Corinthian bronzes were
quenched in the spring of Peirene, perhaps implying that this was one reason for its
special character. Pliny the Elder asserted that Corinthian bronze was the most
highly praised bronze known from earlier times, and he cited famous individuals
who coveted objects made of it (HN 34.1, 6-8; 34.48). Petronius (Sat. 15.31.50)
reports that the nouveau-riche Trimalchio had a collection of "Corinthian"
bronzes. Athenaeus (5.199e) refers to two famous Corinthian bowls with capacities
of more than 360 liters, each with seated figures on the rim and relief figures on the
neck and body. Pliny listed three special Corinthian alloys that were used specifi-
cally for utensils or vessels (HN 34.8): one was silvery white, and he thought that it
contained silver; the second was tawny like gold, supposedly from the addition of
that metal; and the third alloy contained equal parts of gold and silver. Whatever we
make of Pliny's passage, it is clear that craftsmen deliberately chose and produced
their alloys to obtain particular effects (Mattusch 1991b).
Bronzeworking and ironworking were carried out in the Forum Area at
Corinth from the sixth century B.C. to as late as the twelfth century A.D. (Mattusch
1977b). Casting pits sometimes were no more than small holes in the floor with
vents at the sides for flues or for attaching bellows nozzles. An iron smithy is
attested by corroded lumps of iron and by magnetic, silvery flakes of iron. Bronze
and iron slag and clay crucibles and molds were found, as well as bivalve molds for
such items as belt buckles, spearheads, handles, and furniture legs. The join be-
tween the halves of bivalve clay molds was secured by an outer investment of clay,
which could also have been used to form the funnel. A bronze fulcrum panel
with damascened copper, silver, and niello acanthus-and-flower motif provides
a tantalizing introduction to the sophisticated production techniques that were
standard fare in the workshops of one of Corinth's major industries (Mattusch
1991b).
Numerous small fragments broken from bronze statues were found in the
Forum area, apparently to be melted down for re-use. Only one large-scale foundry
has so far been identified at Corinth, and it lies well outside the heart of the city
near the gymnasium (Mattusch 1991a). It consists of a casting pit and a furnace pit.
The installation, which had been filled with foundry debris after its final use,
contained wood charcoal, a few bronze statue fragments, bronze slag, bronze drips,
pumice, iron nails, investment mold fragments, mold bases, and props to support
the molds during baking, a vivid testimony to the complexity of the process. Visual
representations assist our understanding (figures 4.4, 16.1-16.3). The uniformity
and quantity of the material dumped into the workshop suggest that this was a
temporary installation, constructed for a single large commission and then closed
down, in keeping with the usual practice in ancient foundries.
GREEK AND ROMAN STATUARY

IN CONTEXT
Today, ancient bronzes have an aura about them that marbles do not, because
there are far fewer surviving bronze statues than there are marbles, mainly because
bronze could be melted down and reused. No more than a few hundred Greek and
Roman bronzes left today, even if we count all the surviving heads without bodies
and bodies without heads. But this statistic is not representative of the numbers of
bronze statues that were produced during antiquity. Mummius filled Rome with
the statues that he brought home after defeating Greece in 146 B.C. (Strabo 6.381;
Pliny HN 34.36). In 179 B.C., statues had to be removed from the Forum of Rome,
and the Capitoline Hill cleared of statues that were obstructing the view of the
Temple of Jupiter (Livy 40.51.3). Cicero was one of many homeowners in the first
century B.C. who had antique sculptures shipped to him from Greece. In the first
century A.D., Pliny (HN 34.36) reports that there were still about 12,000 bronze
statues standing in Athens, Olympia, Delphi, and Rhodes. He also lists 160 bronze
sculptors who worked between the fifth and the first centuries B.C. (HN 34.49-93).
Pausanias singles out for description unusual statues he saw in the course of his
journey around Greece-a dozen or more bronzes in Delphi, about 60 in Athens,
and more than 40 in the Sanctuary of Zeus at Olympia, besides mentioning hun-
dreds of victors' statues, although not all of these may have been made of bronze.
In contrast to this abundance, surviving bronze statues are usually found one at
a time, prompting scholars to consider them as unique objects. It is also rare that
more than a single example of a particular type survives, although there was, of
course, a great deal of repetition among classical bronze statues, so many of which
represented athletes, statesmen, and heroes. These were not individualized portraits
for the most part, but standardized public monuments that celebrated the familiar.
Anyone walking through a Greek or Roman city or sanctuary would have been
looking at statues whose meaning was easily recognizable by the body, the gesture,
and the attributes, as well as the inscription.
THE ROLE OF COPIES
Taking molds from statues to produce copies was a common practice in antiquity,
perhaps even more so than it is today, to judge from a remark by Lucian (Iupp. trag.
33): "It is your brother, the Hermes of the Agora, the one beside the Stoa Poecile: he
is covered with pitch from being molded every day by statue-makers." The copying
industry was particularly active during the Late Republic and the Imperial period. A
Roman sculpture shop at Baiae on the Bay of Naples yielded many fragments of
plaster casts that had been used to produce marble copies of Greek bronze statues,
including the Athenian Tyrant-Slayers by Critius and Nesiotes, three types of
Amazons, and, most famous of all, the Doryphorus by Polyclitus (Landwehr 1985).
Theophrastus (Lap. 67) remarks that "For (taking) impressions, (plaster) seems to
surpass greatly the other materials, and it is used for this especially in Greece, owing
to its stickiness and smoothness." Pliny the Elder also mentions plaster models (HN
35,156).
Copies of the Doryphorus in both stone and bronze from the area around the
Bay of Naples indicate that, as we should expect, the buyer could choose whatever
medium he or she preferred for the reproduction. He could also choose the format,
buying, for example, a bronze herm reproducing only the head of the Doryphorus
(Naples, National Museum, herm: no. 4885). A measure of the immense popularity
of this particular statue can be seen in the fact that the bronze herm is inscribed,
not with the name of Polyclitus, but with the name of the copyist, Apollonius, who
also noted his Athenian heritage, a circumstance that would have enhanced his
credentials and increased the marketability of the bronzes he produced (Mattusch
2005: 276-82).
THE MARKET FOR BRONZE STATUARY
The practice of making more than one copy of a bronze statue-in fact, of making
whole editions of statues-applied in antiquity as it does today. By the sixth century
B.C., duplicates and mirror images were being produced for erection in sanctuaries.
In the classical period, public statues of leaders and heroes were produced ac-
cording to standard types so as to be easily recognizable, and by the Hellenistic
period statues were being produced for a growing private market. Workshops had
specialties, works that were of one particular type, or of a certain style, or possessing
some standardized features. A buyer might want a pair of sculptures-identical or
mirror images-or variations on a theme. The buyer might choose a bronze or a
marble version of a particular work. In looking at the many marble examples of, for
example, the Discus-Thrower by Myron, one might wonder how many bronzes
were produced of that type. Was there only one bronze "original" that Myron
claimed as his own, or did he authorize the production of more than one? Did
Myron make changes to his original model after seeing the first bronze statue(s),
thereby creating more than one edition of the Discus-Thrower? As a successful
artist, Myron would surely have sold more than one example of each sculpture that
he designed. He might have had a few bronzes produced, perhaps varying the
patina, then a marble or two, and perhaps a few reduced versions of the Discus-
Thrower. After all, works in all media are reproducible, and they can be repeated if
the market demands it, in whatever medium the buyer wants.
The physical evidence for alloys, workshops, and techniques used in the production
of ancient bronzes reveals that the complete repertoire of mechanical and chemical
techniques was applied to the production of a wide variety of political and religious
monuments as well as everyday objects. Lost-wax casting, in particular, was a highly
sophisticated process in the ancient Mediterranean world, from the time of its
beginnings in the ninth century B.C. Lost-wax casting was essentially an indirect
process with innumerable variations, and it was used for the production of mul-
tiples, of editions, and of unique works.
REFERENCES
Beazley, J. D. 1963. Attic red-figure vase-painters. 2nd ed. Oxford: Oxford University Press.
Bianchi, R. 1990. Egyptian metal statuary of the Third Intermediate period (ca. 1070-656
B.C.), from its Egyptian antecedents to its Samian examples," in M. True and J.
Podany (eds.), Small bronze sculpture from the ancient world. Malibu, CA: J. Paul Getty
Museum, 61-84.
Cave, J. F. 1977. "A note on Roman metal-turning," History of Technology 2: 78-94.
Formigli, E. 1985. "Die Restaurierung einer griechischen Grossbronze aus dem Meer von
Riace/Italien," in H. Born (ed.), Archiiologische Bronzen: Antike Kunst, Moderne
Technik. Berlin: Reimer, 168-74.
Haiiser, K., and A. Mutz 1973. "Wie spannten die romischen vascularii (Dreher) ihre
W erkstiicke." Technikgeschichte 40: 251-69.
Heilmeyer, W.-D. 1969. "Giessereibetriebe in Olympia," Archiiologische Anzeiger 84: 1-28.
Hellenkemper Salies, G., H.-H. von Prittwitz, and G. Bauchhenss (eds.) 1994. Das Wrack:
Der antike Schiffsfund von Mahdia. Cologne: Rheinland-Verlag.
Landwehr, C. 1985. Die antiken Gipsabgusse aus Baiae: Griechische Bronzestatuen in Ab-
gussen romischer Zeit. Berlin: Gebr. Mann.
Mattusch, C. 1977a. "Bronze- and ironworking in the area of the Athenian agora," Hesperia
46: 340-79.
Mattusch, C. 1977b. "Corinthian metalworking: The Forum area," Hesperia 46: 380-89.
Mattusch, C. 1982. "The Berlin Foundry Cup: The casting of Greek bronze statuary in the
early fifth century B.C.," American Journal of Archaeology 84: 435-44.
Mattusch, C. 1988. Greek bronze statuary: From the beginnings through the fifth century B.C.
Ithaca, NY: Cornell University Press.
Mattusch, C. 1990. "A trio of griffins from Olympia," Hesperia 59: 549-60.
Mattusch, C. 1991a. "Corinthian metalworking: The Gymnasium bronze foundry," He-
speria 60: 383-96.
Mattusch, C. 1991b. "Corinthian metalworking: An inlaid fulcrum panel," Hesperia 60:
525-28.
Mattusch, C. 1996. Classical bronzes: The art and craft of Greek and Roman statuary. Ithaca,
NY: Cornell University Press.
Mattusch, C. 2005. The Villa dei Papiri at Herculaneum: Life and afterlife of a sculpture
collection. Los Angeles: J. Paul Getty Museum.
Mutz, A. 1972. Die Kunst des Metalldrehens bei den Romern: Interpretationen antiken Ar-
beitsverfahren auf Grund von Werkspuren. Basel and Stuttgart: Birkhiiuser.
Romano, I. B. 2000. "The Dreros sphyrelata: A re-examination of their date and function,"
in C. C. Mattusch, A. Brauer, and S. E. Knudsen (eds.), From the parts to the whole:
Acts of the 13th International Bronze Congress, vol. 1. Journal of Roman Studies Suppl
39.1: 40-50.
Ziegler, C. 1987. "Les arts du metal ala Troisieme Periode Intermediataire," in Tanis: L'or
des pharaons. Paris: Association frarn;:aise d'action artistique, 85-101.
Zimmer, G. 1982. Romische Berufsdarstellungen. Berlin: Gebr. Mann.

Zimmer, G. 1990. Griechische Bronzegusswerkstiitten: Zur Technologieentwicklung eines
antiken Kunsthandwerkes. Mainz: von Zabern.
Zimmer, G. 2003. "Hellenistische Bronzegusswerkstatten in Demetrias," in Demetrias: Die
deutschen archiiologischen Forschungen in Thessalien, vol. 6. Bonn: Deutsches arch-
aologisches Institut, 9-82.
CHAPTER 17
WOODWORKING
ROGER B. ULRICH
IRON Age inhabitants of the Mediterranean inherited and practiced a tradition of

working with wood that was already thousands of years old. With the exception of a
few arid regions, wood had been a readily available resource with a seemingly lim-
itless array of applications. Exploitation was universal, and the scale and complexity
of technical processes increased dramatically with the introduction of metal tools.
Virtually every part of the harvested tree could be put to good use. Even the leaves of
many species were valued for feeding livestock, and bark, always useful as a fuel,
could serve as sheathing for roofing or, in some cases, provide the raw material for
twines and ropes. The discovery in 1991 of many wooden artifacts-tool handles,
arrow shafts, cylindrical containers, and the frame of a backpack-with the frozen
body of the now famous "Iceman" of the Italian alps has demonstrated that at the
end of the Neolithic period in central Europe, 5,000 years ago, the properties of
diverse species of wood for specific applications were already understood at a
sophisticated level. The yew used for the Iceman's bow, for example, is still favored
by the makers of traditional longbows today ( Fowler 2000).
There is virtually no aspect of ancient life that was not affected by those who
handled and shaped wood. Transportation on land and sea depended on wooden
materials and joinery. Wooden elements, first left in their natural "roundwood"
state and later formed into squared beams and planks, played a dominant, if not
exclusive, role in the first built structures and never lost prominence; most stone
and brick structures of later Greek and Roman cities were covered with wooden
roofs and sometimes even rested on wooden post or pile foundations. Roman
vaults could not have been built without wooden centering (Taylor 2003; Lancaster
2005a). Woodworkers shaped agricultural tools and household objects for every
function. Although coal played a minor role as a fuel source in the Mediterranean
world (Theophrastus, Lap. 16), and was known in Roman Britain (Dearne and
Branigan 1995), wood and charcoal provided the sole significant source of energy
for heating and industrial processes that included fired pottery, roof tiles, bricks,
metal tools, and weapons. Given the ubiquity of wood as a building material, it is no
accident that the Latin word materia can mean wood as a building material, wood
forming a tree, or "any substance of which a physical object is made or composed"
(OLD s.v. "materia").
EVIDENCE
Despite the absence of actual wooden artifacts in many archaeological contexts

(although sophisticated techniques in recovery and analysis have indicated that
finds of worked wooden objects are not as rare as might be assumed), the evidence
for the technology of woodworking is abundant. This evidence can be summarized
under four broad categories: 1) ancient written sources, including inscriptions; 2)
artistic representations of tools, craftsmen at work, and wooden objects; 3) artifacts
connected with the technology of woodworking (e.g., tools), actual wooden objects,
and objects of which wood formed an important component. The fourth and final
category is the important role of ethnographic analogy and experimental archae-
ology. Many ancient woodworking practices and tools exhibit a striking continuity
that exists virtually unbroken up to the time of the modem Industrial Revolution.
Ancient Accounts
The Homeric descriptions of Odysseus blinding the Cyclops Polyphemus or
making his marriage bed in his palace on Ithaca provide us with some of the earliest
passages in western literature (ca. eighth century B.C.) of the carpenter at his trade
(Richter 1966). The mutilation of the Cyclops's single eye with a stake of olivewood
is compared to the bidirectional rotation of a strap drill: "just as a man bores a
ship's timber with a drill while those below him twirl it with a strap they hold at
either end, so the bit spins continuously" (Od. 9: 384-86). In a quieter passage the
poet describes Odysseus as a talented craftsman; to make his bed the hero
cut away the foliage of the long-leaved olive,

and trimmed the trunk from the roots up, planing it with a brazen
adze, well and expertly, and trued it straight with a chalkline,
making a bed post of it, and bored all holes with an auger.
-Od. 23: 195-98; trans. Lattimore (1965), by permission
These early passages attest to the existence and use of relatively complicated
tools, such as the strap drill, at the dawn of historical Greek civilization, and they
WOODWORKING 441
reveal that the art of woodcraft in legend was seen as an appropriate skill for a king.
Here and elsewhere, extant written sources provide vital information about
methods of woodworking and reveal some of the technical parlance used by Greek
and Roman craftsmen (Mark 2005).
Inscriptions, including building contracts and accounts as well as funerary
epitaphs that mention specialties within the woodworker's trade, have proven to be
important sources for information. The contract for the wooden gallery con-
structed on the fortification walls of Athens, for example, and another for the
arsenal, include technical names of roofing components (IG2 2.463; IG 2 2.1668;
Caskey 1910; Hodge 1960). Another inscription dating to the mid-first century B.C.
from Puteoli in Italy records a contract for an ornamental doorway to a sanctuary
(CIL 1.577, 10.1781; ILS 5317). The inscription includes the technical terms used for
the framing of a door and protecting porch and lists the species of wood to be used
for the individual components of the project.
Between the third century B.C. and the first century A.D., Theophrastus, Cato,
Varro, and Pliny the Elder wrote extensively about silviculture and the uses of
various species of trees. Julius Caesar's vivid account of a wooden pile bridge
constructed over the Rhine in 55 B.C. (Caesar, B Gall. 4.17; Saatmann et al. 1938-
1939) includes an important vocabulary associated with large timber-framed pro-
jects: the description includes details about piles (tigna, sublicae) and their instal-
lation, the bracing of wooden trestles (derecta), and the use of wattle-work (cratis)
to form a roadbed. The only surviving architectural treatise from the ancient world,
written by Vitruvius around 25 B.C., is a compendium of both Greek and Roman
building practices that offers a rich array of terms used for the wooden structural
elements of buildings and military equipment, including the roofing elements of a
basilica at Fanum designed by the author himself.
Representations of Woodcraft, Tools, and Wooden Objects

Most ancient depictions of woodworkers and their craft have survived in the form
of painting and relief sculpture, but the media and contexts of these images can
differ widely. For Greece, Attic vase painting from the archaic and early classi-
cal periods (sixth to early fifth centuries B.C.) is the dominant form of evidence. In
both Greece and Rome, artwork associated with private individuals (as opposed to
state-funded commissions) is most the informative and varied, as is exemplified by
a Roman funerary relief from Ravenna (first century A.D.) depicting Publius
Longidienus swinging his adze; the dedicatory inscription identifies him as a Jaber
navalis, or shipwright (figure 17.1; Clarke 2003; Kampen 1981; Zimmer 1985).
Examples less obviously referential to woodcraft per se, but no less important
for understanding technique, include the Athenian funerary relief of Hegeso sitting
on a bentwood wooden chair (klismos, fifth century B.C.; Richter 1966: 33-37, fig.
175), or a carved bed-leg in the eponymous seventh-century B.C. Etruscan tomb at
Populonia that imitates a wooden element that in reality would have been turned
Figure 17.1. Relief of P. Longidienus, first century A.D., Ravenna. Total height: 2.66 m.
Museo Nazionale, Ravenna inv. 7. (Photograph courtesy of the National Museum, Ra-
venna.)
on a lathe (Steingraber 1979: 340-41, pl. 44). A depiction of Daedalus, the

ur-craftsman of Greek myth, still in situ at the House of the Vettii at Pompeii,
features a detail of the master's son Icarus cutting mortises in a plank with a tanged
chisel and mallet (Gaitzsch 1980: vol. 2, 154). A bow and drill lie on the floor; a
simple plane leans against one leg of the workbench. The Pompeian wall-painter
surely had contemporary models in mind when he created the mythological scene
( figure 17.2).
The Roman penchant for recording historical events with a documentary eye
occasionally yields a rich dividend of information regarding ancient technologies.
For woodcraft the premier example is the Column of Trajan (dedicated A.D. 113) in
Rome, of which the sculpted frieze features the exploits of the emperor's legionary
forces in two campaigns against the Dacians. Here is a prominent example of
public art that included depictions of tools in use, wooden buildings, timbered
bridges, wheeled vehicles, and the felling and hauling of trees, from which one
WOODWORKING 443
Figure 17.2. Daedalus and Icarus presenting the decoy heifer to Pasiphii.e, House of the
Vettii, first century A.D., Pompeii. (Photograph by Michael Larvey, by permission of
Michael Larvey and the Soprintendenza Archeologica di Pompeii.)
can compile a broad spectrum of the carpenter's craft. Specific examples are
considered below. The reliefs also underline the ubiquity of wood as a construction
material.
Dedications of altars by pious guilds (collegia) of woodworkers to gods asso-
ciated with woodcraft, while few in number, can include scenes of workshops in
operation and, invariably, tools. We find again the Roman evidence most striking.
A relief from an altar dedicated to Minerva in Rome, probably from the late first
century, offers the most complete picture of a busy furniture shop in operation:
handtools, including a carpenter's square (norma), large calipers (circinus), and a
bucksaw (serra) hang from the walls; a three-legged table nears completion on a
workbench, and two larger machines, perhaps for ripping and turning wooden
blanks, stand in the foreground. A lathe may be depicted on the stand in the center
left of the relief, or before the seated figure to the right ( figure 17.3; Colini 1947).
A final category of representational evidence includes miniature renditions of
tools associated with woodworking. For reasons not fully understood, tiny images
of planes and saw blades were used as mint marks on Roman Republican silver
coinage of the first century B.C. ( Fava 1969). The British Museum holds a collection
of miniature bronzes from a tomb in Sussex that depict, among other objects,
adzes, axes, and saws (Goodman 1964: 24; Manning 1966).
Figure 17.3. Marble relief of a furniture shop, Rome. Late first century A.D.? Length 1.38 m.
Capitoline Museums (Montemartini), inv. 2743. (Photograph by R. B. Ulrich.)
WOODEN OBJECTS
Wooden artifacts ranging from grains of sawdust to heavy squared timbers have
been reported from sites throughout the littoral Mediterranean, the Black Sea,
mainland Europe, Asia Minor, and North Africa. Conservation techniques adopted
from the mid-twentieth century have saved increasing numbers of wooden artifacts
and ecofacts (Mols 1999; Tampone 1989). Wood is preserved under conditions of
extreme dry, wet, or cold. The recovery of structural timbers from ancient shipwrecks
has permitted a highly detailed picture of how Egyptian, Greek, and Roman boats
were constructed; many still reveal tool marks left from the day they were fashioned
(see chapter 24) . Notable assemblages of recovered wooden artifacts include tim-
bered tomb chambers containing finely-crafted wooden furniture from the "Royal
Tombs" of Gordion (eighth century B.C.; Simpson et al. 1992), a group of wooden
sarcophagi with carved wooden moldings from the cemetery of Kertch on the Black
Sea dating from Hellenistic and Roman times (Wasowicz and Vaulina 1974; Wat-
zinger 1905), a large number of tables, benches, beds and other wooden household
furnishings from Herculaneum (prior to A.O. 79; Mols 1999), and wooden floors,
walls, shingles, and pipes from the Roman legionary camp at Vindolanda in Scotland
(Birley 1977). The first-century B.C. wooden ship and its cargo found at Cornacchio,
and those at Pisa still under excavation, have yielded finds of wooden objects in
near-pristine condition (Berti 1990; Bruni 2000). Wooden artifacts preserved in
Egyptian tombs provide important comparative data.
Tools
Modern scholars have exhibited a keen interest in the use and appearance of ancient
tools. For woodworking, prominent studies include those of Gaitzsche, Goodman,
Greber, Manning, Matthaus, Mercer, Petrie, and Zimmer (see bibliography).
WOODWORKING 445
Woodworking tools can be divided into two broad categories: those that were
employed to cut, trim, or bore wood, and those used for marking and measuring.
The latter group, including rulers, calipers, framing squares, and levels, will not be
described here. They were used by all manner of artisans and builders. Variants of
many of the cutting tools associated with woodworking, such as the chisel, were also
used for stone and metalwork. Identification of function in such cases is based on
physical attributes. The thin flaring blade of a paring chisel, for example, would not
have been found in the toolbox of the stonecutter.
The first woodworking tools were the axe and the adze, originally handheld
instruments of stone but better utilized when paired with a wooden handle.
Copper, bronze, and iron blades were successively introduced in step with advances
in metallurgy. The earliest cast copper and bronze blades ( from ca. 5,000 years ago)
lacked sockets and were sometimes hafted to wooden handles with the aid of
attached flanges (Arnold 1982; Petrie 1917). Axe and adze blades could be fashioned
by bending a sheet of metal around a dowel to form a socket for a handle (a method
used by European artisans well into the nineteenth century) or by casting the blade
in a mold; the latter produced a stronger tool but required a more specialized
workshop for production. By the Roman Imperial period the most common form
of axe was characterized by a long, flaring blade opposed by a squarish, heavy poll
(Manning and Saunders 1972). Examples of the double-bladed axe and the
broadhead axe also have been found at both Greek and Roman sites (Ciarallo and
De Carolis 1999).
The mortise (see below) was efficiently cut with the drill and the chisel, the
former to remove the core of the mortise and the latter to square up its sides. Iron
drill bits and chisel blades have been found throughout the Mediterranean world.
With few exceptions ancient drill bits exhibit a flat diamond-shaped point at the
terminus of an iron rod that was inserted into a cylindrical wooden handle. When
held at the top by a cupped "nave," the handle could be spun bidirectionally by the
reciprocal motion of a thong and bow. The wooden elements of the drill rarely
survive, although an example from a Roman-period tomb from Hawara, Egypt, is
preserved intact (Petrie 1917). This and representations on Attic vase paintings
(Boston, MFA 13.200; figure 2.6), Roman frescoes (figure 17.2), grave reliefs, and
minor arts (figure 2.5) leave little doubt about the ancient appearance of the drill
(Ulrich 2007: 30). The diameter of the bore was limited to the effectiveness of a
tool operated by one man and the friction generated by the thong; a more pow-
erful strap drill, which worked on the same basic principle, was operated by at
least two workmen (Gaitzsch 1980: no. 318; figure 2.7). Pugsley (2003: 186) has pro-
posed and reconstructed a simple fixed drill held in a stand for the boring of small
objects.
The hand-operated bow could also power simple lathes of the Greco-Roman
period, at least those machines that were called on to tum workpieces of a relatively
small diameter. The existence and wide distribution of the lathe (Grk. tornos, Lat.
torn us) is known best by the wooden artifacts and their representations in painting
and sculpture that could only have been made by turning. Lathes were used ex-
tensively for furniture (see below) and for small containers such as pyxides, bowls
and goblets, and for the handles of agricultural implements and tools (Pugsley 2003;
Rieth 1940). A third-century B.C. relief from a tomb in Petosiris, Egypt, shows a
workman turning a slender dowel; the workpiece is rotating on a vertical axis
(Grodde 1989; Rieth 1940). A sarcophagus from Roman-period Greece depicts an
incomplete rendition of a bow-driven machine in a more familiar horizontal
configuration (Rieth 1940). Greater power may have been achieved by the use of a
flywheel; a damaged depiction on the furniture-makers relief from Rome (figure
17.3) may offer a unique glimpse of such a machine.
The carpenter's plane may have evolved from the hand-adze; Gaitzsch has
identified a hybrid type (ascia-hobel) from depictions on Roman funerary reliefs: a
short, broad blade is attached to one or two curved handles. The tool is visible
leaning against one leg of the carpenter's bench depicted in the Pompeian fresco of
Daedalus (figure 17.2; Gaitzsch 1980; Gaitzsch and Matthaus 1981). Unknown in
Pharaonic Egypt, the plane appears to be a Greek invention, but no actual speci-
mens have been reported. The Latin runcina is clearly derived from the Greek
rhykane, and Greek cabinetry work, including paneled doors, attests to the use of
this essential woodworking tool. The earliest Roman-period representations appear
as mint marks on silver coinage (above); the oldest actual planes have been found at
Pompeii, and therefore must date to before A.D. 79. Since the carriage, or stock,
that formed the main body of the plane was most easily cut from a solid block of
wood, and thus generally lost from the archaeological record, plane blades (called
"irons") are found in isolation at many Roman sites (Mutz 1980). The nearly two
dozen surviving examples of planes fitted with an iron baseplate ( the sole) are all
medium-sized smoothing planes ranging from 21 to 44 cm in length, with a rela-
tively steep rake (the cutting angle of the blade) of 50 to 66 degrees (Gaitzsch 1980:
113; Gaitzsch and Matthaus 1981; Greber 1956). Depictions indicate that the com-
mon smoothing plane was held by oval handles cut into or through the stock in
front of and behind the blade. The width and profile of the cutting edge indicate
usage: convex or toothed cutting edges could remove wood quickly or provide a
rough surface for gluing veneers. Straight, broad edges were for smoothing oper-
ations, narrow irons were used for cutting rabbets, while others were shaped to
produce ogee and bead moldings.
Saws included two-man crosscut tools, bow saws, large and small frame-saws
with tensioning devices (some strikingly similar to the modem buck saw), and
smaller handsaws similar to modem keyhole and backed saws. The ripping of a
large bole or beam into boards took place in a saw pit. The saw consisted of a long
blade held in tension within a rectangular wooden frame operated by one man
positioned on top of the log while his partner stood below. There is an excellent
illustration in a first-century fresco from a workshop at Pompeii (VI.7.8-9) de-
picting a procession of carpenters (figure 17.4). The decorated float depicts, from
the left, Minerva (only her shield is visible), a man pushing a long bench plane, two
men ripping a plank with a frame-saw, and a rendition of Daedalus, patron of
carpenters, standing over the body of his son Icarus (or nephew, Perdix). Saw-teeth
were often "set" (splayed alternately to one side and the other) and "sloped," or
WOODWORKING 447
Figure 17-4- Roman-period fresco depicting a procession of carpenters carrying a deco-

rated float with depictions of Minerva, a man pushing a long bench plane, two men
ripping a plank with a frame-saw, and Daedalus, patron of carpenters. Pompeii, first
century, now in the Naples Museum, inv. 8991. Height 66 cm; width 75 cm. (Photograph:
Rossa, DAI Rome, Inst. Neg. 75.1536, by permission.)
slanted, to create a primary cutting stroke (Manning in Frere 1972; Theophrastus,

Hist. pl. 5.6.3, Pliny, RN 16.227 ).
The best chisels and gouges were of iron; cutting edges were hardened by
tempering (Tylecote in Zienkiewicz 1986: 195). The metal blade was tanged or
socketed to accommodate a handle. The socket was more difficult to form, but its
handle less likely to split when struck by a mallet. A tanged paring chisel from
Aquileia was found intact with a wooden handle ending in a mushroom-shaped
butt that fit perfectly in the palm of the hand for delicate shaving operations
(Gaitzsch 1980: vol. 2, no. 181). The most common form is the mortising chisel, with
a thick shank and a sharply beveled cutting edge designed for deep penetration in
the hardest of woods.
Evidence from Extant Architecture

At sites where no wood survives it is nevertheless possible to recover significant
information of usage. Most Greek and Roman buildings of the historical period
were constructed with a combination of stone, concrete, and wood. Roofing and
upper flooring framed with wooden beams can be substantially reconstructed by

examining the cavities into which the ends of beams were fitted (Hodge 1960; Klein
1998; Ulrich 1996). Extant terracotta revetments that protected the otherwise ex-
posed wooden beams of Etruscan temple roofs reveal the dimensions of ridgepoles,
main rafters, and architraves (Andren 1940; Colonna 1985). Vanished wooden wall
decorations can be restored by examining the cuttings in adjacent masonry; ex-
amples from the Greek mainland (Olympia) and Sicily (Selinunte) are notable
(Martin 1965). The layout of door and window casings can be understood by
observing scars in surviving wall plaster. The dimensions and arrangement of
wooden planks for the form work of foundations (shuttering) or for framing domes
and vaults (centering) can still be directly studied by examination of the imprints
left by boards in the concrete as it set (Lancaster 2005a, 2005b; Rasch 1991). Rock-cut
chamber tombs built by the Etruscans in central Italy at sites such as Cerveteri,
Tuscania, and San Giuliano feature full-scale carvings of ceiling beams, coffering,
architectural moldings, and even furniture that imitate wooden components em-
ployed in the homes of the living; the best examples date to the sixth century B.C.
(Oleson 1978; Steingraber 1979).
THEOPHRASTUS AND PLINY ON RAW

MATERIALS: SPECIES AND USES
Theophrastus (ca. 370-287 B.C.) composed two extant major studies on trees and
other plants: Historia Plantarum and De Causis Plantarum. Three centuries later,
Pliny (A.D. 23-79) devoted six books (12-17) of his Natura/History to commentary
on the typology and uses of trees and woody plants; much is derived directly from
Theophrastus. Even with such treatises, in the absence of canonical and universal
methods of illustrating and describing plants the transmission of botanical knowl-
edge was severely hampered (Baker 1978: 20). The modern scholar is commonly
frustrated when trying to identify a specific tree when confronted with an ancient
reference (Rackham in Salmon and Shipleyi996: 38; Meiggs 1982: 410) (cf. table 17.1).
The trees most favored by Greek and Roman woodworkers grew in the wild.
With the exception of the north coast of Africa, supplies were plentiful. Fertile,
low-lying land was first cleared of timber for agricultural use, in higher elevations
woodcutters exhausted stands of mature trees and then moved on to new sources
of wood. Trees were cut at different times of year depending on intended use, but
most timber was harvested during the cooler months, when the sap content was at
its lowest (Cato, Rust. 31.1-2; Vitruvius 2.9.1-2). Some woodcutters would girdle
the bark from a tree some time before cutting, allowing the tree to die and begin the
drying process while still standing (Pliny, HN 16.192; Vitruvius, De arch. 2.9.3).
WOODWORKING 449
Table 17.1. Mediterranean wood species and their primary ancient uses.
Latin name Botanical Family English name Use
abies Pinaceae fir general use, roofing
acer Aceraceae maple furniture
alnus Betulaceae alder pilings, pipes
betula Betulaceae birch furniture, tableware
buxus Buxaceae box combs, inlay
carpinus Corylaceae hornbeam yokes, olive-presses
castanea Fagaceae chestnut general use
cedrus Pinaceae cedar roofing, shipbuilding
citrus Cupressaceae thuja/ sanderac furniture, veneer
cornus Cornaceae corn el/ dogwood dowels, spokes, spears
corylus Corylaceae hazel torches, small implements
cupressus Cupressaceae cypress doors, statuary
fagus Fagaceae beech furniture, general use
ficus Moraceae fig inferior-grade applications
fraxinus Oleaceae ash spears, wheels
hebenus Ebenaceae ebony furniture, inlay
ilex Fagaceae holm-oak architraves, tools
iuglans Juglandaceae walnut general use, furniture
iuniperus Cupressaceae juniper furniture
larix Pinaceae larch heavy structural applications
olea Oleaceae olive utensils, wooden hinges
palma Palmae palm veneer
pinus Pinaceae pine general use
populus Salicaceae poplar general use, veneer
quercus Fagaceae oak general use
salix Salicaceae willow baskets, bentwood, trellises
taxus Taxaceae yew bows
terebinthus Anacardiaceae terebinth tableware
tilia Tiliaceae linden, lime furniture
ulmus Ulmaceae elm wheel hubs
The delimbing of the tree was achieved rapidly by the woodcutter's axe; the
transportation of the heavy and bulky trunk was no small feat, especially if long beams
were required. Soldiers on Trajan's Column are shown dragging logs from the forest
on rope slings (Scene XV). Logs were conveyed with the greatest ease by water; Strabo
reports that the Tiber afforded Rome a steady supply of timber and other building
materials (Strabo 5.3.7; Meiggs 1982). Overland transport was inevitable; Seneca writes
of roads "trembling" under the weight of wagons carrying logs (Seneca, Ep. 90.9).
In both Greece and Italy, the most highly prized timbers for large construction
projects were the mountain pines (Pinus nigra) and high-altitude silver firs (Abies
alba). Those slopes that fell to substantial waterways were certainly exploited first of
all. Less accessible areas were bypassed; mature stands of trees in terrain difficult to
access were always part of most cities' greater environment. Other fine forests near
towns and cities were protected as sacred groves, including the large tracts of beech
and oak growing on the Alban Hills near Rome and the Altis of Greece's Olympia.
Species exhibiting coveted properties were traded across the Mediterranean.

Among the cargoes of the oldest shipwrecks discovered to date, such as that from
Uluburun (1325 B.C.), are logs of boxwood (Buxus sempervirens), the dense grain of
which was highly prized for the manufacture of combs and other small luxury
objects (Pulak 1993). Twenty centuries later boxwood logs formed the cargo of
another ship wrecked off the Adriatic coast of Italy near present day Cornacchio
(Berti 1990). Fragrant Lebanese cedar (Cedrus libani) seemed to last forever and
was widely used along the coast of the Levant and in Egypt. Julius Caesar's army
witnessed firsthand the resistance oflarchwood (Larix decidua) to fire (Vitruvius
2.9.15-16); the tall straight trunks that grew in the Alps were later transported to
Rome for some of the empire's most ambitious building projects. The most
coveted of furniture woods, called citrus, was evidently cut from the burls or roots
of the Sanderac tree (Thuja articulata, Callitris quadrivalvis, or Tetraclinis articu-
lata, all of which refer to the same tree) of North Africa, particularly from the
region of Mauritania. Citrus, with its delightful grain (compared by Pliny to the
leaves of parsley), honey color, and high polish, was used for small, round tables
that fetched astronomical prices (Pliny, HN 13.96-7 ). The high value of these exotic
woods spawned a class of woodworkers (citrarii) expert in the art of veneering.
Oak, beech, elm, chestnut, maple, and alder, the more common hardwoods in
the lands of Greece and Italy, were used in a wide range of applications. Literary
and physical evidence indicates that woodworkers were keenly aware of their in-
trinsic properties and best applications. From the commentaries of Theophrastus,
Pliny, and Vitruvius emerges a theory of the composition of wood based on the
four elements of air, earth, fire, and water. It is the relative presence of these
constituents assigned to a given species that provided the ancient mind with a
rationale for describing observable properties such as resistance to water, hardness,
or flammability. The apprentice learned that the wood of the alder, for example,
was highly suitable for piling because of its resistance to rot, or that the wood of the
elm, nearly impossible to split, made it perfect for the hub of a wheel, while the
springiness of ash was widely exploited for the rims (felloes) of wheels. Oaken
strakes protected the hull of a merchant ship from the pounding of waves, but
oak's tendency to warp made it less suitable for flooring. Beech did not impart an
unpleasant flavor to food when turned to fashion cups and plates; tall pines from
Corsica made strong roofing timbers (table 17.1).
JOINERY
Objects both large and small were shaped from single pieces of wood: tableware,
tool handles, stakes, pilings, spars, rafters, and water pipes. The first wheels, often
with an attached hub or "nave," were carved from single thick planks of wood
WOODWORKING 451
(Piggott 1983). Woodworking techniques, however, usually involved joining two or

more pieces of wood to create a composite object. The type of joint employed is a
product of several related factors including aesthetic appeal, species of wood,
available tools, and a given artisan's skill level. But the single most important factor
is the nature of the anticipated stress placed on the wooden joint. A scarf joint used
to connect two timbers end to end for the keel of a ship is not an appropriate
solution for two similar beams employed vertically to support a heavy load. In the
latter the fibers of the wooden are submitted to constant forces of compression,
while the former may undergo equally intense forces of tension (when, for example,
a heavy hull was dragged up a beach). Some joins were planned to be fixed but
flexible (pivots and hinges) others firm but with a certain degree of give. Thus
shipwrights are known to have sewn the planking of ships together even when more
unyielding (and conventional) methods of joinery were the norm (McGrail 1985;
Mark 2005). In addition to factors of stress, the craftsman must consider the nature
of the wooden piece to be joined. A tenon, for example, can only be cut so that the
fibers of the wood (generally referred to as the grain) run parallel to its long axis,
otherwise the tenon itself will quickly break in half. But the cavity, or mortise, into
which the tenon fits can be cut either with or across the grain. Finally, aesthetics
play an important role. The miter joint, used to form comers, hides the end-grain of
wood and is thus indispensable for jointing decorative moldings at right angles, but
without reinforcement the mitre provides little strength.
Today makers of fine furniture and cabinetry take pride in creating wooden
chairs, tables, and boxes without the use of any metal fasteners; whether this sense
of superior craftsmanship was shared by the ancient woodworker is difficult to
know. Certainly nails of all sizes were available, each made by hand from bronze
or iron. But nails were expensive and required the services of a blacksmith. For
projects requiring the assembly of small wooden pieces, such as chairs or small
tables, nails tend to split wood and loosen over time. Heavier beams, such as those
used for roofing projects, might employ spikes and iron straps. Plaques of terra-
cotta that covered the wooden structural elements of early Italic temples were held
in place by bronze nails. Floorboards were fastened to joists by nails. Iron rivets
connected wooden handles to iron tools. Although the principle of the screw was
well known to Greeks and Romans-large-diameter versions were used for agri-
cultural presses and small metal bolts were used by metalworkers for jewelry and
parade armor (Kiechle 1967; Klumbach 1973; Deppert-Lippitz et al. 1995)-there
was no ancient equivalent of the wood-screw or the carriage bolt. Only in the
eighteenth century was it possible to manufacture threaded screws on any scale
( figure 17.5).
The repertoire of joints used by ancient woodworkers will be familiar to
anyone practicing the traditional (based on hand tools) craft today (Milne 1982).
The single most versatile of wooden joints is the mortise-and-tenon. The tenon
may take the form of a cylindrical dowel, or a fan (or swallowtail), also known in
Latin as a sericula (little hatchet), or a rectangular, tongue-shaped piece of wood.
The tenon may be formed by shaping the end of the workpiece to be joined, or it
c;
,odworkingjoints (Roman contexts): A-F: corner joints: A. simple mitre; B. butt-joint; C: half-lap; D: rebate
to side planks; E-F: corner braces with half-lap and saddle joints. G: mortise and tenon; H: tongue and groov~
I glued boards; I: half-lap (saddle). Dovetail joints: J: asymmetrical ("half") dovetail; K: dovetail; L: lap dov,
I: edge-halved scarf with one dovetail butt.
WOODWORKING 453
may be carved from an independent wooden element. The "shell-first" planking

system used by both Greek and Roman shipwrights employed thousands of
mortises and tenons in every hull (figure 24.3). Individual strakes were joined at the
edges to one another with a tenon that was held in place by a dowel (Ucelli 1950; see
also chapter 24). Scarfs were used for the longitudinal joins, perhaps reinforced by
iron nails or treenails. The mortise and tenon joint was used for joining heavy
beams in architectural applications and was essential for assembling wooden
furniture, particularly the rails and legs of tables, chairs, benches, and beds. The
felloes of a wheel rim were attached to one another with tenons, wooden spokes
were tenoned into the hubs (called "naves").
Overlapping pieces of wood were frequently jointed by half-lap or saddle
joints, which involved the use of the saw and chisel to create opposing notches so
that the two pieces could join while maintaining the thickness of one. Such joints
were used to build lattices or grids of wooden slats. These assemblies might be used
to support the mattress of a bed or couch or to protect a window or the upper
panel of a door (Mols 1999). Finds of small boxes and larger chests reveal that
corners were jointed with mitres and box (or "finger") joints, the latter either of a
consistent rectilinear configuration, or of the more elegant pins and tails of the
dovetailed joint (Keepax and Robson 1978: 38; Desantis in Berti 1990: 266).
Strong and durable glue has always been essential to most small-to-medium-
size projects, including household utensils, furniture, and cabinetry. Glue can be
used alone as a binding agent, most often for edge-to-edge connections such as
those that might be used for the boards of a broad tabletop. Without glue the
highly developed arts of veneering and parquetry would not have been possible.
Even with finely-cut wooden joints, glue was employed to strengthen and tighten
the final bond. That the Greeks and Romans credited the mythological craftsman
Daedalus with the invention of glue indicates the antiquity and importance of the
process (Pliny, HN 7.198). Natural glues were made from the cartilaginous parts of
cattle and fish; Pliny considers a type made from the genitalia of bulls especially
efficacious (HN 28.236).
Wooo TECHNOLOGY IN LARGE

BUILDING PROJECTS
Wood was used in all phases of ancient building projects, in both temporary and
permanent installations. Its use can be documented for below-grade foundation
work, the framing of walls, doors, and windows, the installation of floors, and the
construction of roofs (Adam 1994). Large masonry structures regularly employed
all-wooden upper stories or galleries (Delaine 1996; Ulrich 1996). In frontier
military camps of the Roman period wood was the predominant building material
and was even used for the defensive walls and towers of a legionary encampment
(Shirley 2000). On military campaigns siege machines and artillery were largely
constructed of wooden elements. The first theaters of Rome were built entirely of
wood.
Waterlogged wood can survive even in temperate zones for thousands of years.
Wooden piling was considered suitable for foundation work from prehistoric
times. Vitruvius (De arch. 2.9.11) tells us that in his day most of the buildings of
coastal Ravenna were supported on piles. The practice of constructing houses on
piling in northern Italy and the adjacent Alpine regions had been common since at
least the second millennium B.C. (Arnold 1982). Piling was a staple for bridge
construction. Vertical or splayed supports could be driven directly into waterways
without the construction of cofferdams (Caesar, B Gall. 4.17). Rome's earliest
bridge over the Tiber River was always known as the Pile Bridge (Pons Sublicius; cf.
Plutarch, Num. 9.2-3; Dion. 3.45.2). Arched stone bridges of the Roman Imperial
period were seated on grids of tightly-packed piles, many of which have been
recovered from excavations in the northern provinces, their tips sharpened and
reinforced with iron plates to aid the process of ramming them into place (Clippers
1969; Marchetti 1891).
On dry land, wooden posts buried in pits formed the structural uprights of
Greece and Rome's earliest structures; the post-holes of the well known Heroon
from the site ofLefkandi on Euboia (ca. 950 B.C.) or the "Hut of Romulus" on the
Palatine in Rome (eighth century B.C.) provide complementary examples (Pop-
ham et al. 1982; Puglisi 1951). The use of sleeper beams-horizontal wooden
footings into which timber uprights were mortised, was a northern European
tradition of great antiquity later enthusiastically adopted by Roman military
builders (Birley 1977; Rickman 1971).
With the Roman introduction and rapid adoption of concrete as a foundation
material for land and maritime applications in the late second century B.C.,
wooden planks and beams played an important role in shaping and stabilizing the
wet concrete as it cured ( figure 25.2). For the building of harbors, wooden caissons,
some large enough to be considered "single-mission barges," reinforced with
transverse interior beams, were sunk into place and filled with concrete that was
capable of curing underwater; examples from Caesarea in Palestine have been
described in detail (Brandon 1999; Oleson in Raban 1989). On dry land, temporary
formworks of beams and planks were erected; the planked shuttering could be
removed for reuse (Taylor 2003). Similar arrangements of beams and planks or
walls formed of squared timbers placed either horizontally or vertically were used
as retaining walls or to form quays; the waterfront of Roman London has revealed
well preserved examples of such timberwork (Marsden 1994).
Vitruvius (De arch. 2.1.4) writes about a rustic cabin made from interlocking
notched logs that was built by the indigenous peoples of the timber-rich lands
around the Black Sea; the simple technique was used by Roman military engineers
to built abutments for field artillery. A vertical, planar grid formed by roundwood
or squared beams filled in with panels of another material such as wattle-and-daub
WOODWORKING 455
Figure 17.6. Trajan's bridge over the Danube as depicted on the Column of Trajan (cast in
Museo della Civilta Romana), ca. A.D. 113, Rome. (Photograph by R. B. Ulrich.)
or mud-brick was widely employed for domestic architecture during the Medi-
terranean Iron age. The practice was probably well known from Neolithic times and
never fell out of use. Vitruvius (De arch. 2.8.20) refers to the flammability of walls
built from opus craticium, a closely-related method involving an open timber frame
filled with panels of concrete and rubble. While the technique is usually associated
as a cheaper alternative to solid (and thicker) walls of concrete, its lightness makes it
especially suitable for upper floors with minimal support. Furthermore, the com-
bination of a wooden framework and masonry may have been recognized as being
particularly resilient to the frequent tremors that shook the towns along the
mountainous spines of Italy and Sicily (Papaccio 1993). It is notable in this regard
that at both Herculaneum and Pompeii walls of opus craticium were still standing
after the ground tremors and volcanic eruption of A.D. 79. In mainland Europe,
small buildings were apparently commonly built with sawed planks nailed to a
beam frame; these structures feature prominently on depictions of Dacian villages
shown under attack by Roman soldiers on Trajan's Column in Rome.
Among the largest of all timber-framed buildings constructed in the Roman
period were the timbered theaters, amphitheaters, and large bridges known pri-
marily through references in ancient literature and the occasional artistic rendition
(Dio Cassius 68.13-1; Tacitus, Ann. 4.62). In the Greek world the largest wooden
constructions were the warships of the late third century; these super-galleys were
up to 120 m long (Casson 1991). Masonry theaters, amphitheaters, and bridges were
all built around open wooden frameworks designed to bear tremendous loads;
Lancaster (2005a: 34) has identified buildings in Rome where the wooden centering
may have distorted under the immense load of partially-built concrete vaults.
Wooden bridges were either supported by some form of wooden trestle spanned by
horizontal timbers, perhaps cantilevered for larger spans, or stone piers were set in
place across the river and carried a wooden carriageway (Milne 1985). Perhaps the
greatest of Roman wooden bridges, despite its short life, was the one constructed by
Trajan's gifted engineer, Apollodorus, which spanned the Danube on 20 piers for a
total length in excess of 3,500 Roman feet (O'Connor 1993). A simplified repre-
sentation of the Danube bridgework on Trajan's Column reveals a solution realized
by building a wooden segmental arch reinforced by an arrangement of radial struts
(Lepper and Frere 1988). The use of multiple shorter timbers would have aided the
process of construction over the long spans of open water and allowed the pro-
curement teams to cut local trees of modest dimensions (figure 17.6).
While the Greek theater was characteristically built against a natural slope that
directly supported wooden bleachers (ikria) or stone blocks, Roman theaters and
amphitheaters often stood on level ground; the inwardly sloping seating areas
needed to be supported in a way that provided access to the interior of the
structure and was sufficiently stable to support the weight of thousands of spec-
tators. The simple trestle, each consisting of a pair of uprights spanned by a
horizontal beam, would have formed the essential component of each bay. Knee
braces placed to reinforce the architraves would have created a unit that approx-
imates the form of the bays found in the superimposed arcades of stone and
concrete arenas. Unfortunately, there are no detailed literary descriptions of what
these structures looked like or how they were assembled. A rare notation of
structural collapse from an ancient source records the failure of a timber-framed
amphitheater in A.D. 27 at Fidenae in Italy; Tacitus (Ann. 4.62) tells us specifically
that both the footings and the joinery of the project were inadequate. The date of
the event also proves that these colossal wooden places for public spectacle were
still being built long after permanent and fire-resistant masonry versions had been
erected in central Italy (the Pompeian amphitheater of ca. 80 B.C.; Pompey's
Theater in Rome, ca. 55 B.C.). Nero constructed an amphitheater of "interwoven
beams" in A.D. 57, the last such structure to be built in Rome before the con-
struction of the Colosseum a dozen years later (Calpurnius Siculus, Ee. 7.23-4). A
hybrid stone and timbered amphitheater is represented on Trajan's Column
(Lepper and Frere 1988: 151, pl. 73).
Floors and flat roofs were constructed by building a deck of joists and
planking; by the late first century planked floors in multistoried structures were
routinely covered with a thick layer of concrete that supported a paving of tiles or
stone mosaic. Vitruvius (De arch. 7.1) offers specific recommendations; actual finds
indicate that his strictures were only loosely followed. Abundant evidence in the
form of joist-holes left in the standing walls of houses, shops, and public buildings
at Pompeii and Herculaneum indicate that wooden joists, rectangular in section,
were more closely spaced than necessary to support the concrete subfloor (Ulrich
1996). By the high imperial period evidence from buildings in the Roman port
town of Ostia indicate that the framers of upper floors may have preferred em-
WOODWORKING 457
\ _ \
Figure 17.7. Prop-and-lintel roofing system: A: architrave; B: prop; C: lintel; D: purlin; E:

ridgepole; F: primary rafters; G: secondary purlins; H: sheathing.
ploying two or three heavy crossbeams that supported a set of lighter joists, which
in turn carried the planking and masonry floor.
In terms of framing and joinery, wooden roof construction posed the greatest
challenge to the framing carpenter. Large structures in the west were not covered
with masonry vaults until the second-century B.C., and even from that point all-
wooden roofing was preferred for traditional buildings like temples and large
covered public spaces, including the Greek stoa, the Roman basilica, or the meeting
places for Greek and Roman councils (e.g., the bouleuterion and curia, respec-
tively). In the historical period, most wooden roofs were pitched in order to shed
precipitation (Cicero, Orat. 3-180). Pitched roofs also provided basic insulation
from heat and cold by virtue of an enclosed attic that was integral to their design.
The double-pitched, or gabled, roof, oriented to the long axis of a rectangular
ground plan, characterizes the most "classical" of monumental buildings from the
Greco-Roman world. Depending on how the roof was framed, the gabled roof
could cover spaces up to about 30 meters (just over 100 Roman feet) with a free-
span (Gibson et al. 1994). The double-pitched wooden roof was generally framed in
one of two ways: the prop-and-lintel method or the tie-beam truss (Klein 1998;
Hodge 1960; Meiggs 1982). The former involved supporting the ridgepole and
purlins (if used) with a series of vertical supports that were themselves carried by
crossbeams, interior columns or walls. The rafters of the prop-and-lintel system
were not physically connected to the crossbeams (generally defining the ceiling) .
The prop-and-lintel system was widely used in classical Greece and in Italy even
though the principle of the truss may have been understood from archaic times
(contra, see chapter 9). By the mid-second century B.C., the tie-beam truss was
adopted with enthusiasm by Roman builders (cf. chapter 10).
The basic tie-beam truss is composed of three timbers: two diagonal rafters and
a horizontal tie-beam that form a fixed triangle with jointed corners (Lancaster
----
Figure 17.8. Timber-truss, hypothetical arrangement in Vitruvius's basilica at Fanum.
(After Warren in Morgan 1914: 135.)
2005a: 22; Ulrich 2007: 138). The apex of the triangle either supports or is integrated
with the ridge-beam. The tie-beam truss was capable of sustaining great loads and
spanned the broadest of open spaces (up to 30 m), surpassing those covered by
concrete barrel or cross vaults (generally no more than 26 m across) (Lancaster
2005a: 138). Only the dome was capable of a broader clear span. The horizontal tie-
beam, kept under tension by the rafters and their load, could support the dead
weight of a heavy coffered ceiling. For larger spans the interior of the truss could be
reinforced with a kingpost, queenposts, collar beams, or diagonal braces (Giuliani
2006).
To hypothesize the appearance and to reconstruct the language of framed
wooden roofs one must turn to the abundant evidence of cuttings in stone archi-
tectural elements and revetments, to literary references, and to artistic represen-
tations. In both Greek and Etruscan applications, prop-and-lintel roofs tended to
employ a heavy ridge beam. The socket for the ridge beam at the archaic Temple of
Hera I at Paestum was a full 91 cm high by 67 cm wide, perhaps 20 cm higher than
that of the Parthenon (Hodge 1960: 46). In Etruria, the ridge beams of temples may
have been even larger. The terracotta plaques that covered the exposed ends of the
beams have survived in a good enough state to offer an idea of dimensions; at Pyrgi
and Tarquinia the crowning beams may have measured as much as 1.25 m high by
1.80 m wide. Beams of this size running the length of a monumental temple (the Ara
della Regina of Tarquinia was nearly 40 m long) were most likely made up of
composites of smaller timbers, a practice that was certainly used for wooden ar-
chitraves, for Vitruvius writes about the use of such "composite beams" (De arch.
WOODWORKING 459
4.7.4: trabes conpactiles) for epistyles. The emphasis on width over height on these
early ridge beams is probably due to the requirement of creating a broad bedding
for the principle rafters, themselves of hefty dimensions, which in turn supported
the purlins and a series of lighter rafters and sheathing (figure 17.7).
The viability of the timber truss depended in large part on the strength of
the critical join at the outer corners of the triangular frame, where rafters, under the
immense force of compression, met the tie-beam, itself under great tension ( figure
17.8). It appears unlikely that builders employed composite or scarfed beams for the
constituent members of the truss, thus the greatest attempted spans were probably
limited to available timber of sufficient dimension. The broadest spans identified
were achieved in the covered music halls and the palatial audience halls of the late
Roman Republic and early Empire (first century B.C. through first century A.D.);
beyond this point it seems that the tallest procurable firs and larches had been felled
(Cassius Dio 55.8.4; 66.24.2).
FURNITURE: CARVING,BENDING,
AND VENEERING Wooo
Wooden furniture fashioned by Greek and Roman craftsmen provides evidence of

additional woodworking practices. These specialized skills included the ability to
shape and apply moldings, to carve three-dimensional figural images in relief and in
the round, to bend wood and to apply veneers, such as parquetry, to objects of high
value. These arts were not limited to the makers of furniture. In both Greek and
Roman temples, for example, sculptors created images of deities in wood (cf. Pliny,
HN 16.216), and decorative architectural moldings required carving ability. But it
was in the manufacture of common household furniture that the skills listed above
were routinely applied.
The most common of wooden moldings, such as the unadorned bead or the
cyma (the latter known today as a Roman ogee) were rendered with relative ease
using customized molding planes (above). Foliate friezes, the egg-and-dart, or
similar decorative bands known so well from carved stone examples have been
observed on the carbonized wooden chests and upright cabinets from Hercula-
neum and on the Greek and Roman period wooden sarcophagi recovered from the
Kertch region of the Black Sea (Mols 1999; Watzinger 1905).
The lathe played a special role in the furniture-maker's shop. The legs of
couches and beds, for example, were commonly turned, as were those of heavy,
backed chairs. By the Roman period the aggressive turning of legs produced ele-
ments incapable of supporting heavy loads. Such turned "wooden" legs were in
fact composite structures of vertical iron bars on which deeply-turned wooden
elements were separately "strung" (Mols 1999). Yet other popular furniture types,
such as the elegant Greek klismos or round serving tables, were built without any
significant use of the lathe.
Carving in high relief, to the point of being rendered in the round, abounded
on Greco-Roman furniture. Every household had at least one three-legged table
with a round top (mensa delphica) invariably decorated with a carved animal head,
such as a lion or griffin, on the "knee" of each leg, and a "foot" rendered as a claw
or hoof. The universal application of such elements indicates that the rank-and-file
furniture maker was skilled at least in producing these stock figural images.
There exists no ancient literary passage that describes the process of bending
wood by the use of hot water or steam, but there can be little doubt that the
practice existed. Pliny identifies a number of species considered good for bending,
but the use of moist heat is not explicit (HN 16.227). The curved back and flaring
legs of the Greek klismos would have been fashioned best from bending the
components with steam. The same can be said about the light frame of a chariot,
the rim of a shield, the yoke for a draft animal, the bent frame of a wicker chair, or
the thin-walled, cylindrical box (capsa). The outer rims of wooden wheels could be
made from single pieces of wood bent into a circle (Curle 1911).
Veneering was a highly developed skill among Greek and Roman craftsman.
Thin sheets or strips of valuable woods were glued over the surfaces of less valuable
species. In addition to wood veneers, craftsman applied plaques of ivory or tor-
toiseshell to furnishings. Actual examples of veneer-work, including parquetry,
have been documented over a broad chronological and geographical spectrum,
from the furniture found in eighth-century B.C. Gordion to the carbonized fur-
niture of first-century Herculaneum.
The full range of woodworking techniques employed by woodworkers of the an-

cient Mediterranean throughout the historical period and the practical knowledge
concerning species and their most suitable applications attest to the importance of
wood technologies for agriculture and hunting, domestic life, trade, warfare, and
leisure activities. The tools and techniques employed by the end of the Roman
period would show little change until the appearance of the Industrial Revolution
fifteen centuries later. Given twentieth-century developments in preservation and
the science of analysis, the recovery and restoration of ancient wooden artifacts has
enabled modern scholars to document with increasing detail this vital component
of ancient life.
REFERENCES
Adam, J.-P. 1994. Roman building. Bloomington: Indiana University Press.

Andren, A. 1940. Architectural terracottas from Etrusco-Italic temples. Lund and Leipzig:
Gleerup.
WOODWORKING 461
Arnold, B. 1982. "The architectural woodwork of the Late Bronze Age village Auvernier-
Nord," in McGrail 1982: m-28.
Baker, H. 1978. Plants and civilization. Belmont, CA: Wadsworth.
Berti, F. 1990. Fortuna Maris, la Nave Romana di Cornacchio. Bologna: Nuova Alfa.
Birley, R. 1977. Vindolanda:A Roman frontier post on Hadrian's Wall. London: Thames and
Hudson.
Brandon, C. 1999. "Pozzolana, lime, and single-mission barges (Area K)," in K. Holum,
A. Raban, and J. Patrich ( eds.), Caesarea papers, vol. 2. Journal of Roman Archaeology
Suppl. 35. Portsmouth, RI: JRA, 169-78.
Bruni, S. (ed.) 2000. Le nave antiche di Pisa: The ancient ships of Pisa. Florence: Edizioni
Polistampa.
Carver, M. 0. H., S. Donaghey, and A. B. Sumpter 1978. Riverside structures and a well in
Skeldergate and buildings in Bishophill. The Archaeology of York 4.1. York: Council for
British Archaeology.
Caskey, L. D. 1910. "The roofed gallery on the walls of Athens," American Journal of
Casson, L. 1991. The ancient mariners. Princeton: Princeton University Press.
Ciarallo, A., and E. De Carolis 1999. Homo Faber: Natura, scienza e tecnica nell'antica
Pompei. Milan: Electa.
Clarke, J. 2003. Art in the lives of ordinary Romans. Berkeley: University of California Press.
Colini, A. M. 1947. "Officina di Fabri Tignarii nei Frammenti di un Ara Monumentale
Rinvenuti fra il Campidoglio e il Tevere," Capitolium: 21-28
Colonna, G. (ed.) 1985. Santuari d'Etruria. Milan: Electa, Regione Toscana.
Ciippers, H. 1969. Die Trierer Romerbrucken. Mainz: von Zabern.
Curle, J. 1911. A Roman frontier post and its people: The fort of Newstead in the parish of
Melrose. Glasgow: Maclehose.
Dearne, M. J., and K. Branigan 1995. "The use of coal in Roman Britain," Antiquaries
Journal 75: 71-105.
Delaine, J. 1996. "The Insula of the Paintings: A model for the economics of construction
in Hadrianic Ostia," in A.G. Zevi and A. Claridge (eds.), "Roman Ostia" revisited.
London: The British School at Rome: 165-84.
Deppert-Lippitz, B., et al. 1995. Die Schraube zwischen Macht und Pracht: Das Gewinde in
der Antike. Sigmaringen: Jan Thorbecke.
Fava, A. S. 1969. I simboli nelle monete argentee republicane e la vita dei romani. Turin:
Soprintendenza alle Antichita del Piemonte e Museo Civico di Torino.
Fowler, B. 2000. Iceman: Uncovering the life and times of a prehistoric man found in an
Alpine glacier. New York: Random House.
Franchi dell'Orto, L. (ed.) 1993. Ercolano 1738-1988: 250 anni di ricerca archeologica. Atti del
Convegno Internazionale Ravello-Ercolano-Napoli-Pompeii 1988. Roma: L'Erma di
Bretschneider.
Frere, S. 1972. Verulamium Excavations. Vol. 1. Society of Antiquaries of London Research
Committee Reports 28. London: Thames and Hudson.
Gaitzsch, W. 1980. Eiserne romische Werkzeuge: Studien zur romischen Werkzeugkunde in
Italien und den nordlichen Provinzen des Imperium Romanum. 2 vols. British Archae-
ological Reports, Intl. Series S78. Oxford: BAR.
Gaitzsch, W., and H. Matthaus 1981. "Runcinae-romische Hobel," Bonner Jahrbucher 181:
205-47.
Gibson, S., J. Delaine, and A. Claridge 1994. "The Triclinium of the Domus Flavia: A new
reconstruction," Papers of the British School at Rome 62: 67-97.
Giuliani, C. 2006. L'edilizia nell'antichita. Rome: Carocci.

Goodburn, D. 1991. "A Roman timber framed building tradition," Archaeological Journal
148: 182-204
Goodman, W. L. 1964. The history of woodworking tools. London: Bell.
Greber, J.M. 1956. Die Geschichte des Hobels. Ziirich: VSSM-Verlag.
Grodde, B. 1989. Holzernes Mobiliar im vor- und fruhgeschichtlichen Mittel- und Nordeur-
opa. Frankfurt am Main: P. Lang.
Hodge, A. T. 1960. The woodwork of Greek roofs. Cambridge: Cambridge University Press.
Kampen, N. 1981. Image and status: Roman working women in Ostia. Berlin: Mann.
Keepax, C., and M. Robson 1978. "Conservation and associated examination of a Roman
chest," The Conservator 2: 35-40.
14-22.
Klein, N. L. 1998. "Evidence for West Greek influence on mainland Greek roof con-
struction and the creation of the truss in the archaic period," Hesperia 67: 335-74.
Klumbach, H. (ed.) 1973. Spatromische Gardehelme. Miinchner Beitrage zur Vor- und
Friihgeschichte 15. Munich: Beck.
Lancaster, L. 2005a. Concrete vaulted construction in imperial Rome. Cambridge: Cambridge
University Press.
Lancaster, L. 2005b. "The process of building the Colosseum: The site, materials and
construction techniques," Journal of Roman Archaeology 18: 57-82.
Lattimore, R. (trans.) 1965. The Odyssey of Homer. New York: Harper and Row.
Lepper, F., and S. Frere 1988. Trajan's Column: A new edition of the Cichorius plates.
Gloucester: Alan Sutton.
Manning, W. H. 1966. "A group of bronze models from Sussex in the British Museum,"
Antiquaries Journal 46: 50-59.
Manning, W. H., and C. Saunders 1972. "A socketed iron axe from Maids Moreton,
Buckinghamshire, with a note on the type," Antiquaries Journal 52: 276-92.
Marchetti, D. 1891. "Di un antico molo per lo sbarco dei marmi reconosciuto sulla riva
sinistra del Tevere," Bullettino della Commissione archeologica Comunale di Roma ser.
4: 45-60.
Mark, S. 2005. Homeric seafaring. College Station: Texas A&M University Press.
Marsden, P. 1994. The ships of the Port of London: First to eleventh Centuries AD. English
Heritage Archaeological Report 3. London: English Heritage.
Martin, R. 1965. Manuel d'architecture grecque. Vol. 1, Materiaux et techniques. Paris: Pi-
card.
McGrail, S. (ed.) 1982. Woodworking techniques before A.D. 1500. British Archaeological
Reports, Intl. Series S129. Oxford: BAR.
McGrail, S. (ed.) 1985. Sewn plank boats: Archaeological and ethnographic papers based on
those presented to a conference at Greenwich in November, 1984. British Archaeological
Reports, Int. Series S276. Oxford: BAR.
Meiggs, R. 1982. Trees and timber in the ancient Mediterranean world. Oxford: Oxford
University Press.
Mercer, H. C. 1960. Ancient carpenters' tools. Doyleston: Bucks County Historical Society.
Milne, G. 1982. "Recording timberwork on the London waterfront," in McGrail 1982: 7-23.
Milne, G. 1985. The port of Roman London. London: Batsford.
Mols, S. T. 1999. Wooden furniture in Herculaneum. Amsterdam: Gieben.
Morgan, M. H. 1914. Vitruvius: The Ten Books on Architecture. Cambridge, MA: Harvard
University Press.
WOODWORKING 463
Mutz, A. 1980. "Ein fund von Holzbearbeitungs-Werkzeugen aus Augst Insula 31," Jah-
resberichte aus Augst und Kaiseraugst. Liestal: Amt fiir Museen und Archaologie des
Kantons Basel-Landschaft.
O'Connor, C. 1993. Roman bridges. Cambridge: Cambridge University Press.
Oleson, J.P. 1978. "Technical aspects of Etruscan rock-cut tomb architecture," Romische
Papaccio, V. 1993. "11 telaio ligneo (opus craticium) ercolanese: Considerazioni e ricerche
sui requisiti antisismici," in Franchi dell'Orto 1993: 609-16.
Petrie, W. M. F. 1917. Tools and weapons illustrated by the Egyptian Collection in University
College, London. London: British School of Archaeology in Egypt.
Piggott, S. 1983. The earliest wheeled transport. Ithaca, NY: Cornell University Press.
Popham, M., E. Touloupa, and L. Sackett 1982. "The Hero of Lefkandi," Antiquity 56:
169-74.
Pugsley, P. 2003. Roman domestic wood. British Archaeological Reports, Intl. SeriesSm8.
Oxford: BAR.
Puglisi, S. M. 1951. "Gli abitatori primitivi del Palatino attraverso le testimonianze
archeologiche e le nuove stratigrafiche sul Germalo," Monumenti Antichi 41:
1-98.
Pulak, C. 1993. "The shipwreck at Ulu Burun, Turkey: 1992 excavation campaign," INA
Quarterly 20: 4-12.
Raban, A., 1989. The harbours of Caesarea Maritima, vol. 1. J.P. Oleson (ed.). British
Archaeological Reports, Intl. Series S491. Oxford: BAR.
Rasch, J. 1991. "Zur Konstruktion spatantiker Kuppeln vom 3. bis 6. Jahrhundert," Jahr-
buch des Deutschen Archiiologischen Instituts 106: 311-83.
Richter, G. M. A. 1966. The furniture of the Greeks, Etruscans, and Romans. London:
Phaidon.
Rickman, G. 1971. Roman granaries and store buildings. Cambridge: Cambridge University
Press.
Rieth, A. 1940. "Das Holzdrechseln." Ipek: Jahrbuch fur priihistorischer und ethnographische
Kunst 13: 85-107.
Rowland, I., and T. N. Howe 1999. Vitruvius: Ten Books on Architecture. Cambridge:
Salmon, J., and G. Shipley 1996. Human landscapes in classical antiquity. London: Rou-
tledge.
Simpson, E., K. Spirydowicz, and V. Dorge 1992. Gordion wooden furniture. Ankara:
Museum of Anatolian Civilization.
Saatmann, K., E. Jiingst, and P. Thielscher 1938-39. "Caesars Rheinbriicke," Bonner Jahr-
bucher 143-44: 83-208.
Shirley, E. A. M. 2000. The construction of the Roman legionary fortress at Inchtuthil. British
Steingraber, S. 1979. Etruskische Mabel. Rome: Giorgio Bretschneider.
Tampone, G. 1989. Il Restauro del Legno. Atti del 2°Congresso Nazionale Restauro del
Legno. Florence: Nardini.
Taylor, R. 2003. Roman builders: A study in architectural process. Cambridge: Cambridge
University Press.
Ucelli, G. 1950. Le Navi di Nemi. Roma: La Libreria dello Stato.
Ulrich, R. B. 1996. "Contignatio, Vitruvius, and the Campanian Builder," American Journal
of Archaeology 100: 137.
Ulrich, R.B. 2007. Roman woodworking. New Haven: Yale University Press.
Wasowicz, A., and M. Vaulina 1974. Bois grecs et romains de l'Ermitage. Warsaw: Nar-
odowy.
Watzinger, C. 1905. Griechische Holzsarkophage aus der Zeit Alexanders des Grossen. Leipzig:
Hinrichs.
Weeks, J. 1982. "Roman carpentry joints: Adoption and adaptation," in McGrail 1982,
157-68.
Zienkiewicz, J. D. 1986. The legionary fortress baths at Caerleon, vol. 2: The finds. Cardiff:
National Museum of Wales.
Zimmer, G. 1985. "Romische Handwerker," in Aufstieg und Niedergang der romischen Welt
2.12.3. Berlin: de Gruyter, 205-28.
CHAPTER 18
TEXTILE PRODUCTION
JOHN P. WILD
THE metaphorical use of textile terminology is a characteristic of the classical lan-

guages that reflects the ubiquity of textile crafts in contemporary life (Scheid and
Svenbro 2003). To us, however, such textile concepts may seem obscure. Moreover,
textiles and textile implements were mostly of organic origin and hence survive
infrequently in the archaeological record; it is hard to appreciate what is not tangi-
ble. Nevertheless, an array of sources can be tapped in the study of Greco-Roman
textile technology. At first glance they may appear complementary; but rarely can
they be fitted together neatly (Wild 2000, 2004). Written evidence offers detailed
poetic descriptions of weaving contests on the one hand, humdrum lists of com-
modities and prices in private documents on the other. Earlier commentators looked
no further (Yates 1843), but over the past seventy years textile fragments recovered
during scientifically conducted archaeological excavations where specific microdi-
matic conditions have facilitated preservation, provide a new and more immediate
perspective. Nonetheless, the problem of matching an artifact or craft skill to a pu-
tative account of it in a Greek or Roman source is considerable; the precise meaning
of many ancient terms remains ill defined, and they changed over time.
Representations in art-weaving scenes on Greek pottery, for example-are a
valuable, but sometimes coded, resource. The craftsmen and women themselves,
however, did not communicate their experience or skills in any of the above media
(Bender-J0rgensen 2003). They are best approached through the study of still ex-
isting "primitive" practice, a facet of the ethnographic record that has also been the
inspiration behind much modem experimentation in the reconstruction of lost
techniques (Fansa 1990).
Four basic processes characterize ancient textile production: growing and
harvesting the raw fibers, converting them into yam, weaving or interlacing the
yams to make a fabric, and, possibly, fabric finishing. In a developed industrial
context further techniques can be identified within and between these basic stages.
My account will consider the processes in that order, but it will draw somewhat
arbitrary lines between them for clarity's sake. Dyeing, however, does not fit into
the sequence neatly and has been treated at the end of the chapter, where it would
belong today.
FIBERS: HARVESTING AND

PREPARATION
The textile industry was rurally embedded (Wild 1999; Erdkamp 1999). Some aspects
became urbanized through increasing division of labor, but its dependence on do-
mestic plants and animals, and hence the farming community, is undeniable. There
was little bulk transport of raw fibers over long distances (in contrast to medieval and
modern practice), and the most famous textile centers, like that of the Po Valley,
developed where ample quantities of high-quality raw materials were at hand.
Animal Fibers
Sheep yielded the most varied and versatile textile fiber: wool. Present-day relict
populations, such as the Orkney sheep of North Ronaldsay, indicate the general
character of Greco-Roman sheep: small bodied, long necked, with a fleece rarely
weighing more than 1.5 kg (Maltby 1981; Wild 1982: 110-12; Melena 1987: 397-99).
While wool was historically a secondary product after meat and milk (Sherratt
1983), it is clear that Mycenaean flock-masters with their large numbers of wethers
were already concentrating on wool production (Killen 1964). The mountainous
terrain of Greece and Italy saw flocks in seasonal transhumance from lowland to
highland grazing (Whittaker 1988; Small et al. 2003: 189-93), and in Asia Minor the
hinterland of Miletus was noted for its fine-woolled sheep (Carter 1969; Morel
1978). Traditionally, fine-woolled flocks were protected against dirt and damage by
jackets of skin, a classical Greek practice later familiar in the Roman world (Xe-
nophon, Mem. 2,757; Kraemer 1928; Frayn 1984: 35, 164).
Both rich and poor-even the emperors (Frayn 1984: 176-79)-kept sheep
(King 1999). In Egypt, Apollonius, chief finance officer to Ptolemy II, upgraded the
quality of his wool by importing animals from Miletus (P.Cair.Zen. 59195; Loftus
2000: 175). Roman agriculturalists such as Columella were familiar with the prin-
ciples of selective breeding for wool quality (Columella, Rust. 7.2.4-5), as were Iron-
Age farmers in central Europe (Ryder 2001). Fiber diameter measurement of wool
yarns in ancient textiles indicates the existence (often in the same flock) of fleece-
TEXTILE PRODUCTION 467
types ranging from hairy medium wool (typical of the earliest domesticated sheep)
through generalized medium, to medium, semifine and true fine wool, arguably
achieved through selective breeding (Ryder 1983: 45-49; Wild 2002: 2). Since the
gene for white is dominant, selection led to pure white wool, although the "prim-
itive" pigmented light-, dark-, and reddish-brown and gray (mixed white and
brown) wools were still in evidence (Ryder 1990), and were used for decorative
effect without overdyeing.
The presence of white, fine-woolled sheep in the Greek world is confirmed by
surviving yarns (Ryder 1988). Textiles in the Roman west reflect a preponderance
of generalized medium-woolled fleeces (Ryder 1969: 509, table 5; Ryder 1981); in the
eastern provinces more fine wools have been identified, but in Egypt sheep with
unpigmented, hairy medium wool seem to have been the norm (Wild 1994: 34-35,
table 1; Walton Rogers 1994).
Sheep were usually shorn with a pair of sprung iron shears, in spring or early
summer (Varro, Rust. 2.11.6-9; Wild 1982: 116). The older practice of plucking the
wool at the time of the molt persisted into the Roman period (Pliny, HN 8.191); this
method had the advantage of leaving behind most of the coarser outer hair. On
Apollonius' estate in Egypt, the fine-woolled, jacketed sheep were plucked, while
the coarser-woolled Arabian animals were shorn (P.Cair.Zen. 59430). Surprisingly,
we hear nothing about wool sorting or grading (Ryder 1995). Fleeces were beaten to
remove foreign bodies (Aristophanes, Lys. 574-78; Wipszycka 1965: 33-34). The
extent to which they were washed to remove suint (dried perspiration) and excess
lanolin, especially before dyeing, is hard to estimate. At Pompeii a fixed wool-
washing installation consisted of a cauldron over a boiler, flanked by draining sur-
faces (Borgard and Puybaret 2004: 49-52; Frayn 1984: 148-49), and in Verecundus'
workshop portable draining boards are depicted (Wild 1970: 171, fig. 52). Where
sales of fleeces are recorded, as at Dura-Europos (Baur et al. 1933, nos. 247,252,272),
the wool had probably not yet been degreased. Washed (and sorted) wool was sold
by weight (Wild 2002: 5), since cleaning a fleece reduces its weight by 30 percent or
more.
Wool was often-but not always and not of necessity-combed to align fibers
for spinning and to separate short from long fibers. Wool-combing, a low-status
craft, is encountered everywhere in the Greco-Roman world (Wipszycka 1965: 34-
35; Frayn 1984: 150; Wild 2000: 210-11). Roman wool combs were flat rectangles of
iron with teeth cut or welded on both short ends (only one end in eastern Britain);
they were mounted on a post behind which the operative sat (Wild 1982: 118). On a
second-century relief from Ostia ( figure 18.1) the wool is seen being drawn through
the teeth with a flat wooden implement and formed into carefully weighed rovings
(Holliday 1993: 96, fig. 7).
Goat hair was of minor commercial value, except in arid environments where it
served for tentage and sacking (Ryder 1993; Wild 2003a: 38). The yarns were usually
plied to improve tensile strength (Batcheller 2001). Whether the fine underwool of
the "cashmere" goat of central and southern Asia was ever imported (as is claimed
by Moulherat and Vial 2000; Schmidt-Colinet et al. 2000: 11) is debatable.
Figure 18.1. Woolcomber T. Aelius Evangelus seated at his combing-post, Ostia. In the
Medelhavsmuseet, Stockholm. (Drawing by Priscilla Wild.)
Vegetable Fibers
Flax (linum usitatissimum L.), the principal ancient plant fiber, is hardy, but flour-
ished best in well-watered, fertile soils (Korber-Grohne 1987= 366-79). It was grown
in the Peloponnese in classical Greek and earlier times ( forbes 1956: 36-38; Robkin
1979). Egypt's well documented linen industry flourished under Ptolemaic and
Roman control (Wipszycka 1965), but there were other minor centers of repute in
the Roman Empire (Gleba 2004; Forbes 1956: 35-43). Paleobotanical sampling in-
dicates that flax was a regular component of mixed farming regimes (Gleba 2004:
31-32; Wild 2002: 6-7).
The bast fibers of flax have to be separated from the woody core and bark
between which they lie, in a series of traditional operations that involved all mem-
bers of a peasant family (Wild 1970: 27-29). The stalks (ca. 1 m high) were uprooted,
dried, and then retted, allowing bacterial action stimulated by moisture to loosen
the fibers. Retting could be undertaken in ponds or tanks-or simply by long
exposure in the fields (dew retting; Freckmann et al. 1979: 91-102) . The stalks were
dried, then "broken" with a wooden mallet and "scutched" with a wooden blade to
remove any remaining bark and core. Finally, bundles of fiber were drawn across
the multiple teeth of a fixed comb (hackle) to free them from residual cellulosic
matter (tow) and isolate the individual fibers (Wild 1968). Sacking and cordage
were made from the tow.
Hemp (cannabis sativa L.) also contains bast fiber, tougher and coarser than
flax but similar in character. Well established in late prehistoric Europe (Korber-
Grohne 1987: 379-89; Godwin 1967), it was converted in classical antiquity into
cordage, netting, and sailcloth (Wild 1970: 16-17). The harvested stems (ca. 1.5 m
high or more) were processed like flax (Wild 1970: 29-30).
The annual form of cotton (gossypium herbaceum L.) grown in India and the
Persian Gulf region was known to Herodotus (3.47, 106; 7.65) and Theophrastus
(Hist. pl. 4.4.8; 4.7.7-8) and through occasional imports of cotton cloth into
Greece. In the oases of Roman Egypt it became an increasingly popular crop and
was planted under irrigation in the Fezzan and the Jordan valley. Lower Nubia and
India were alternative external sources of cotton textiles (Wild 1997; Wild et al.
forthcoming). How the short fibers (lint) were detached from the seeds on which
they formed in the boll is uncertain-possibly just with the fingers.
Insect Fibers
In a famous passage Aristotle accurately describes how the filament extruded by the
silk worm of Cos in the Aegean to build its cocoon was unwound, reeled up, and
woven into garments prized for their sheen (Hist. an. 5.19.551b; Wild 1984). The wild
silk moth in question was probably Pachypasa otus, now virtually extinct (Good
1995: 965, fig. 3; Panagiotakopulu et al. 1997). Cultivated silk from the Bombyx mori
was available only in China until the sixth century A.D.; its importation as yarn and
woven cloth may not have begun before the Augustan period. Despite its cost, it was
a powerful and enduring status symbol, the basis of some of the most complex
decorative textiles of antiquity.
Miscellaneous Fibers
A group of fibers with special properties were exploited in the Roman period.
Rabbits, reared in enclosed warrens, provided soft hair for expensive clothing
(Edictum Diocletiani 19.73a-73c; Varro, Rust. 3.3,1-3, 8-10). The earliest archaeo-
logical evidence for this fiber dates to the seventh century (Amrein et al. 1999: 93).
The Pinna nobilis, a large Mediterranean mollusk, anchored itself to the seabed
with filaments that could be converted into a smooth, shiny cloth (Wild 1970: 20;
Maeder et al. 2004). Table napkins spun and woven from asbestos were a dinner-
party amusement, since they could be burnt clean (Wild 1970: 21).
SPINNING
Spinning was not just an essential stage in the conversion of processed raw fibers
into yarn for weaving or interlacing, but a social and economic obligation falling on
all female members of a household, whatever their status, from the Homeric Iron
Age to late antiquity (Homer, Od. 4.125-35; St. Jerome, Ep. 130.15; Wild 1970: 31-40;
Larsson-Loven 1998). The technique was learned at an early age, and it is rare to find
anything but the highest-quality yam, sometimes unbelievably fine, in archaeo-
logical textiles. Two implements were required: the distaff, a short rod to the upper
section of which wool rovings or hackled flax bundles were attached, and the
spindle, a carefully profiled rod with a swelling somewhere along its length on
which the spindle whorl was wedged (Wild 2003c: 25-29). The whorl, a disc of wood
or terracotta or a shaped, recycled potsherd with a central perforation, provided
the necessary momentum. In Greece the rovings were prepared on an epinetron, a
ceramic sheath laid on the upper leg while seated (Jenkins 2003: 72-73). The act of
spinning involved the drawing out fibers from the distaff (held in the left hand) and
inserting twist into them with rotating spindle, manipulated by the right hand.
Deeply conservative cultural norms governed the precise ways in which the
spinning operation was performed. The spindle might be rotated in the hand, rolled
down the thigh, supported on a dish, or suspended to spin free on the end of the
drafted yam, the mode most commonly depicted in classical art (Crowfoot 1931). In
Europe the whorl was usually mounted at the lower end of the spindle, but in Egypt
and some adjacent regions at the upper end. The direction of spin-generally
clockwise {Z-direction) in the western Roman world, anticlockwise {S-direction) in
the eastern-also reflected cultural predilections (Wild 2003b: 82,108). "Z" and "S"
refer to the resulting direction of the slant in the fibers of the thread. Spindles
loaded with yam were traditionally stored in a basket (Wild 2003b: 83). The second-
century Roman grave of a young girl at Les Martres-de-Veyre (Clermont-Ferrand)
contained a distaff loaded with wool rovings (held in place by colored ribbons), a
spindle, and remains of a basket (Audollent 1923: 305, no. 31, pl. VI.12).
Two twist-inserting crafts were entrusted to men: rope-making with flax,
hemp, or another coarse vegetable fiber (Pliny, HN 19.18; Gaitzsch 1986: Abb. 34)
and making "gold thread." For the latter, narrow ribbons of gold (less than 0.5 mm
wide), cut from thin foil sheets, were wound around a silk or flax core (Bedini et al.
2004; Rinuy 2000).
LOOMS AND WEAVE STRUCTURE
Weaving was the center of gravity in the textile industry, but there was a potential
risk of a production bottleneck, for it took some ten spinners in continuous em-
ployment to keep one loom supplied with yam. In practice, however, much textile
work was part-time and seasonal, and in highly industrial milieus like Roman Egypt
much yam was bought and sold. The ancient weaver set out to create on the loom a
textile item of predetermined form and function, ensuring minimal wastage of yam
and minimal cutting and sewing (Granger-Taylor 1982). The tailor's main preoc-
cupation was with recycling old clothing. Except for sailcloth sold as yardage, the
marketing of bolts of cloth in a medieval and modern sense was unknown.
While the horizontal loom of dynastic Egypt, spanned between two beams, was
still used by nomadic groups, it cannot be identified in the Greco-Roman world,
where the principal looms were vertical (Kemp and Vogelsang-Eastwood 2001: 310-
12). The warp-weighted loom dominated weaving in classical Greece and most of
the Roman Empire until at least the first century A.D. (figure 2.11; Hoffmann 1964;
Staermose Nielsen 1999: 63-78: Wild 2002: 10-11). The loom was large (up to 2.50 m
wide) with two timber uprights, a cloth beam across the top, and a fixed shed rod
below ( figure 18.2). It inclined at a slight angle to facilitate the natural shed between
the two warp sheets, which were tensioned by two rows of fired (sometimes un-
fired) clay loom-weights-which are the sole archaeological remains that survive.
The front warp sheet with its weights hung over the fixed shed rod, while the rear
warp sheet hung vertically from the cloth beam. The shed was changed by means of
heddles attached to a heddle rod that was supported in the forward position for
the artificial shed on two brackets. The weavers beat weft upward using a wooden
sword-beater inserted into the shed. Warp was kept in due order by passing a pin-
beater (kerkis, radius) across it so that it "sang" (Wild 1967a; Gudj6nsson 1990: 173-
74).
The warp was secured to the cloth beam by means of a flat-woven or tablet-
woven starting-border prepared in a preliminary operation (Hoffmann 1964: 151-
83). It characterizes textiles woven on the warp-weighted loom, which range from
plain weaves (1/1 tabby) of varying weight to complex diamond twills. In classical
Greece, decorative figured bands in tapestry weave ( for which the design was built
up from individual lengths of colored weft) were inserted into the web (Wace
1948), although on a warp-weighted loom this was not straightforward (Granger-
Taylor 1992: 20-21, fig. 3; Staermose Nielsen 1999: 95, fig. 53).
In the Nordic world, where weaving on the warp-weighted loom is a living
tradition, considerable diversity can be documented in the techniques of the weavers
(Gudj6nsson 1990; Hoffmann 1964: 188). In the Greco-Roman world the variety in
shape and weight of loom-weights recorded even on a single site points to similar
diversity. Pyramidal, fired clay loom-weights perforated at the apex are the most
widely distributed; conical weights are common, while discoidal weights (some-
times mold-made) are typically Hellenistic (Davidson 1952: 146-57; Staermose
Nielsen 1999: 41-48). Compared to those from prehistoric Europe, Greek and Ro-
man loom-weights were light (70-250 g), with up to 70 in a set for one loom (Tebar
Megias and Wilson forthcoming).
The two-beam vertical loom, on which a second, lower, horizontal beam
replaced the warp-weights, began to displace the warp-weighted loom in the first
century A.D.; the latter survived, however, in a few contexts (Servius, Ad Aen. 7.14;
Hope and Bowen 2002: 67, 165). The new loom's advantage lay in being readily
adaptable to the devices and techniques required for the patterned fabrics in-
creasingly popular in the eastern Mediterranean. Weavers beat the weft downward
with a wooden comb rather than a sword beater (Thomas 2001: 15, figs. 15, 17, 23)
iii
I
L 1-1 t- ..'...J
iv
c~ :~~-,ii
,-- V
IT
' vi
YI Y( Yi
ii
B/J. /jA
I
Figure 18.2. Structure of the warp-weighted loom. (After Wild 1970: fig. 53, by permission
of the Faculty of Classics, Cambridge University.)
and could sit at the loom; starting-borders are often of twined cords, variously
constructed (Verhecken-Lammens 1992).
A range of decorative techniques can be seen in the surviving ancient textiles,
some of them with a regional or chronological bias. The simplest was a check,
based on bands of yarn in contrasting colors (or contrasting spin-directions) set up
in the warp and repeated in the weft by the weaver (Wild 1964). The so-called
shaded bands with rainbow color-effects were even simpler to weave, for their
progressively changing colors had already been blended by the spinner (Schmidt-
Colinet et al. 2000: 133). The two-beam loom was particularly associated with the
insertion of monochrome or figured tapestry bands or panels into otherwise plain
textiles. For this purpose specific warp-threads could be controlled by extra shed
rods or heddles, and either temporarily eliminated or crossed to allow tighter beat-
ing up of the weft and hence a stronger color effect (Granger-Taylor 1992). Ta-
pestry patterns could also be highlighted in the "flying needle" technique, whereby
obliquely floating ( usually linen) yarns were introduced with separate spools on
the loom itself (easily mistaken for embroidery; De Jonghe 1988: 27, pl. 130).
Designs brocaded with floating weft-threads are also found (De Jonghe 1978).
Surface enhancement-for warmth as well as decoration-by means of rows of
loops added in the weft direction was widely practiced in the Mediterranean region
on wool, linen, and even silk garments and furnishings (van 'T Hoo ft et al. 1994: 145,
fig. 9). Cut-loop pile and classic "carpet knots" are found on eastern Roman textiles,
where they reflect the textile culture of western and central Asia ( Fujii 1980).
It has long been recognized that the complex nature of damask, compound
tabby, and compound twill weaves called for correspondingly complex devices to
open multiple sheds, set up and manipulated by exceptionally capable weavers. A
horizontal loom, it has been argued, would best facilitate such arrangements,
perhaps the forerunner of the medieval and modern draw-loom (Wild 1987). There
is, however, no solid evidence for it. Rather, the zilu loom still in use today in Iran
for compound weaves suggests that the Roman loom for advanced weaves re-
mained a vertical one (Thompson and Granger-Taylor 1995-1996; Ciszuk 2004).
Damask twill carries an overall geometric pattern based on juxtaposed areas of
warp-and weft-faced fabric. Weaving faults in Roman examples indicate that the
design was achieved by raising in a planned order pairs of heddle rods from a set of
as many as 20 rods (De Jonghe and Tavernier 1978: pls. II, V). The most striking
pieces are late Roman silk tunics incorporating tapestry-woven features in purple
wool and gold thread (Bedat et al. 2005); but simpler 2/1 and 3/1 twill damasks in
wool from early Roman Egypt hint that the weave may be a Hellenistic invention
(Schrenk 2001; Cardon 1999a).
Weft-faced, compound tabby can be claimed as another Hellenistic develop-
ment (Pliny, HN 8.196). Its structure allows the interplay of two or more colors of
weft in a near-reversible pattern: weft threads can be raised to the surface of the
cloth or concealed beneath neighboring wefts (De Jonghe 1988: figs 140-45; Des-
rosiers 2004: 14-23, figs. 1-2: Ciszuk 2000). On the zilu loom a set of heddle rods
governs the basic underlying weave, while the numerous sheds for the design are
opened by drawing on combinations of individual heddles (Thompson and
Granger-Taylor 1995-1996). Both work at a right angle to the vertical warp. As-
suming that the Roman loom was similarly equipped, the Roman master weaver
must have had the mental agility of a modern computer programmer, along with a
prodigious memory. Tuck (2006) has proposed that chants and songs assisted the
ancient weaver in the accurate reproduction of complex textile patterns. Most
compound tabbies were furnishing fabrics in wool, a few in silk; but the develop-
ment by the fourth century A.D. of compound twill (see figure 18.3) and the
increasing use of fine silk led to some of the most spectacular textiles of antiquity
(Schrenk 2004: 180-92).
The weaving of bands and narrow fabrics did not necessarily depend on a
framed loom, but on warp spanned between the weaver and a distant fixed point.
For simple 1/1 tabby, the two sheds could be opened by a single heddle-frame of
bone, wood, or bronze (Wild 1970: 72-75). Tablet-weaving, a method of warp-cord
patterning, involved a pack of slim square or triangular tablets with a hole in each
Figure 18.3. Fourth-century A.D. silk tunic in weft-faced compound twill weave showing
erotes in medallions, now in the Abegg-Stiftung, Riggisberg. (Photograph by C. von Virag,
by permission of the Abegg-Stiftung.)
corner for the passage of a warp-thread (Hansen 1990a). There is either direct evi-
dence or a strong presumption that such bands were familiar in Greece and Rome
(Giroire 1997).
INTERLACED AND NON-WOVEN

FABRICS
Naalebinding, a technique of openwork construction with a single needle, was pop-

ular in antiquity for three-dimensional textiles, particularly socks, of which many
examples have come to light in Roman Egypt (Hansen 1990b; Burnham 1972). For
bag-shaped headdresses, fashionable in the Mediterranean world, the technique of
sprang was employed, whereby threads spanned on a frame in a single sheet were
twisted around one another to create a netlike structure (Jenkins and Williams
1985). Ply-split braiding and simpler braiding forms were used for girth-straps and
webbing (Collingwood 1998). The propensity of wool fibers to interlock and shrink
under pressure and with moisture was exploited to make felt, sometimes for
clothing, often for helmet or shield linings (Gervers 1973).
Small-scale textile manufacture with the techniques just described was probably
based in the home, even if it had a commercial aspect. Work on larger looms was
carried out by a team of men and/or women and children in environments ranging
from a cramped sweatshop in Roman Egypt (McGing 1990), an airy courtyard in
Pompeii (Jongman 1991: 162-65), or a well lit villa workshop as designed by Vi-
truvius (De arch. 6.4.2). The respective skills of wool and linen weavers, who took
due note of the physical properties of their materials, are reflected in the differing
structures (e.g., selvedge types) of their products.
FINISHING AND FULLING
Clothing and items of soft furnishing were-or could be-ready to wear or use
virtually as soon as they had been cut from the loom. Texture, surface appearance,
and handle, however, could be greatly improved by "finishing," a task entrusted
in urban contexts to specialists with their own workshops and staff, supplies of
cleansing agents, and ample water. Wool and linen fabrics were treated in different
ways that took note of their respective properties.
The fuller, who handled principally wool textiles, was a well known figure from
the Mycenaean period onward (Chadwick 1976: 43, 71; Pliny, HN 7.57). He treated
items straight from the loom and also acted as laundryman for soiled clothing
(Wild 1970: 82-86; Wilson 2003). The focal point of a fuller, exemplified by some
well preserved establishments in Pompeii (Moeller 1976: 41-51; Jongman 1991: 170-
79) and Ostia (Pietrogrande 1976), was an array of large ceramic vessels with
pointed bases set in a masonry bench against a wall. Elsewhere, slightly different
dispositions have been recorded (Wilson 2001: 274-77), and wooden tubs employed
(Wild 1970: 82, fig. 73). The fulling vessels contained a solution of water and a
cleansing agent, commonly (and most cheaply) stale human urine, the ammonia
component of which reacted with grease in the wool to form a soapy compound
(Bradley 2002: 30-32). Fuller's earth (calcium montmorillonite) was a more ex-
pensive alternative (Robertson 1986: 42-81). The fuller trod the cloth vigorously in
the solution; the result, for loom-state cloth particularly, was to shrink and compact
it. This "waulked" cloth was then rinsed in tanks, and dried. Some towns made
special provision for fullers' water supplies (SEG 35.1483). Optional extras included
having the nap raised to a soft finish with a board set with spines (aena) (Wild 1968)
and then trimmed with shears (see figure 18.4). Provision for this soft finish could
be made at the yarn-spinning stage (Wild 1967b).
Figure 18.4. Fuller's tombstone from Sens (Yonne), showing one operative treading
textiles in a tub and another shearing raised nap. (After Wild 1970: fig. 73, by permission of
the Faculty of Classics, Cambridge University.)
Pure white garments, which might be achieved by bleaching them over a pot of
smoldering sulphur, were a social desideratum (Wild 1970: 83). In the late fourth
century B.C., Theophrastus, whose father was a fuller, lists types of kaolin and chalk
that could be applied to cloth surfaces to enhance the bright effect (Lap. 63, 67-68).
Romans were also impressed by carefully pressed clothes, for which a screw press
and auxiliary devices might be employed (Granger-Taylor 1987). Linen could be
bleached by boiling it with natron (sodium carbonate and sodium bicarbonate),
where this was available, notably in Egypt (Wipszycka 1965: 23-24; Lucas 1932). The
surface of a linen textile, including in particular any seams, might be smoothed and
polished with a glass "slick-stone" (Wild 1970: 85, pl. XII.cl; Walton Rogers 1997:
1775-79).
DYEING
The topic of dyeing is a branch of applied chemistry. It is worth emphasizing,

however, that, given the rich palette of colors the ancient dyer had to offer, his or
her work was integral to that of the spinner and weaver, with whom there was
cooperation at several levels in the textile production sequence.
Wool was most commonly dyed in the fleece to ensure maximum uptake of
dyestuff (Halleux 1981: 44,151; Edict of Diocletian 24). Flax, on the other hand, was
dyed in the hank, as cross-sections of yarns sometimes reveal (Sheffer and Granger-
Taylor 1994: 161). Piece-dyeing seems to have been comparatively rare, except for the
redyeing of faded monochrome textiles. Resist-dyeing of woven cloth was an art of
Roman Egypt: the design was drawn out in a resist-medium, often hot wax, applied
with a brush, so that in the vat only the background took up the dye (Schrenk 2004:
82-93; Cardon 1998). Some textiles were simply painted (Gerziger 1975).
Dyers' workshops in Pompeii, some associated with probable wool-washing
facilities, housed various types of heated vats, variously disposed. The larger lead-
lined vats were arguably for mordanting wool prior to dyeing in smaller vats nearby.
Others may have been for fermentation dyes like woad (Borgard and Puybaret
2004: 52-56).
Advances in the analysis of dyestuff traces in archaeological textiles have shed
new light on the dyestuff sources tapped by Greek and Roman dyers (Cardon
1999b; Cardon et al. 2004; Wouters 1995); to understand how they achieved their
colors has been the objective of experimentation (Edmonds 2000; Cardon 1999b:
46-57).
The fact that only an expert can spot the difference between a Greek or Roman textile
and its modern equivalent is testimony to the consistently high level of skills de-
ployed by ancient textile operatives. Gender divisions are blurred: spinners tend to
be female, professional weavers male, but there are exceptions. The role of children
in the (relatively safe) industry has been vastly underestimated. Technological ad-
vances, arguably driven by change in clothing fashion, are hard to identify; but the
innovative spirit of Hellenistic Alexandria seems to have made itself felt in the textile
industry, as in so many other technologies.
REFERENCES
Alfaro, C., J.P. Wild, and B. Costa (eds.) 2004. Purpureae vestes: Textiles y tintes del
Mediterraneo en epoca romana; Actas del 1 symposium internacional sobre textiles y
tintes del Mediterraneo en epoca romana (Ibiza 8 al 10 de noviembre 2002). Valencia:
University of Valencia.
Amrein, H., A. Rast-Eicher, and R. Windler 1999. "Neue Untersuchungen zum Frauengrab
des 7. Jahrhunderts in der reformierten Kirche von Biilach (Kanton Zurich)," Zeits-
chrift fur schweizerische Archaologie und Kunstgeschichte 56: 73-114.
Audollent, A. 1923. Les tombes gallo-romaines a inhumation des Martres-de-Veyre. Mem-
oires presentes a l'Academie des sciences et belles-lettres 13. Paris: Academie des
Sciences et Belles-Lettres.
Batcheller, J. 2001. "Goat-hair textiles from Karanis, Egypt," in P. W. Rogers and J.P. Wild
(eds.), The Roman textile industry and its influence: A birthday tribute to John Peter
Wild. Oxford: Oxbow, 38-47.
Baur, P. V. C., M. I. Rostovtzeff, and A. R. Bellinger (eds.) 1933. The excavations at Dura-
Europos conducted by Yale University and the French Academy of Inscriptions and
Letters: Preliminary report of fourth season of work, October 1930-March 1931. New
Haven: Yale University Press.
Bedat, I., S. Desrosiers, C. Moulherat, and C. Relier 2005. "Two Gallo-Roman graves
recently found in Nantre (Vienne, France)," in F. Pritchard and J.P. Wild (eds.),
Northern archaeological textiles. North European Symposium for Archaeological
Textiles 7. Oxford: Oxbow, 5-11.
Bedini, A., I. A. Rapinesi, and D. Ferro 2004. "Testimonianze di filato e ornamenti in oro
nell'abigliamento di eta romana," in Alfaro et al. 2004: 77-88.
Bender-J0rgensen, L. 2003. "The epistemology of craftsmanship," in L. Bender-J0rgensen,
J. Banek-Burgess, A. Rast-Eicher (eds.), Textilien aus Archiiologie und Geschichte:
Festschrift fur Klaus Tidow. Neumiinster: Wachholtz, 30-36.
Bender-J0rgensen, L. forthcoming. "Archaeological textiles between the arts, crafts and
sciences," in A. Rast-Eicher (ed.), North European Symposium for Archaeological
Textiles 9.
Borgard, P ., and M.-P. Puybaret 2004. "Le travail de la laine au debut de l'Empire: L' apport
du modele pompeien. Quels artisans? Quels equipements? Quelles techniques?" in
Alfaro et al. 2004: 47-59.
Bradley, M. 2002. "'It all comes out in the wash': Looking harder at the Roman fullonica,"
Burnham, D. K. 1972. "Coptic knitting: An ancient technique," Textile History 3: 116-24.
Cardon, D. 1998. "Textiles archeologiques de Maximianon-Al Zarqa et Didymoi: Exem-
ples precoces de teinture par reserve sur laine," Bulletin de Liaison du Centre
International d'Etude des Textiles Anciens 75: 15-20.
Cardon, D. 1999a. "Les damasses de laine de Krokodilo (100-120 apr. J.-C.)," Bulletin de
Liaison du Centre International d'Etude des Textiles Anciens 76: 6-21.
Cardon, D. 1999b. Teintures precieuses de la Mediterranee: Pourpre-kermes-pastel. Car-
cassonne: Musee des Beaux-Arts de Carcassonne.
Cardon, D., and M. Feugere (eds.) 2000. Archeologie des textiles des origines au Ve siecle:
Actes du colloque de Lattes, oct. 1999. Montagnac: Monique Mergoil.
Cardon, D., J. Wouters, I. Vanden Berghe, G. Richard, and R. Breniaux 2004. "Dye analyses
of selected textiles from Maximianon, Krokodilo and Didymoi (Egypt)," in Alfaro
et al. 2004: 145-54.
Carter, H. B. 1969. "The historical geography of the fine-woolled sheep (1)," Textile In-
stitute and Industry 7: 15-18, 45-48.
Chadwick, J. 1976. The Mycenaean world. Cambridge: Cambridge University Press.
Ciszuk, M. 2000. "Taquetes from Mons Claudianus: Analyses and reconstruction," in
Cardon and Feugere 2000: 265-82.
Ciszuk, M. 2004. "Taquete and damask from Mons Claudianus: A discussion of Roman
looms for patterned textiles," in Alfaro et al. 2004: 107-13.
Collingwood, P. 1998. The techniques of ply-split braiding. London: Bellew.
Crowfoot, G. M. 1931. Methods of handspinning in Egypt and the Sudan. Halifax: Bankfield
Museum.
Davidson, G. R. 1952. Corinth XII: The minor objects. Princeton: Princeton University Press,
3-189.
Desrosiers, S. 2004. Soieries et autres textiles de l'antiquite au XVIe siecle. Paris: Reunion
des Musees Nationaux.
Edmonds, J. 2000. The mystery of imperial purple dye. Little Chalfont: Privately printed.
Erdkamp, P. 1999. "Agricultural underemployment and the cost of rural labour in the
Roman world," Classical Quarterly 44: 556-72.
Fansa, M. 1990. Experimentelle Archiiologie in Deutschland. Archaologische Mitteilungen

aus Nordwest-Deutschland Beilieft 4. Oldenburg: Isensee.
Forbes, R. J. 1956. Studies in ancient technology. Vol. 4. Leiden: Brill.
Frayn, J. M. 1984. Sheep-rearing and the wool trade in Italy during the Roman period.
Liverpool: Cairns.
Freckmann, K., G. Simons, and K. Grunsky-Peter 1979. Flachs im Rheinland: Anbau und
Verarbeitung. Koln: Rheinland-Verlag.
Fujii, H. 1980. Al-Rafidan I: Special edition on the studies of textiles and leather objects from
Al-Tar caves, Iraq. Tokyo: Kokushikan University.
Gaitzsch, W. 1986. Antike Korb- und Seilerwaren. Schriften des Limesmuseums Aalen 38.
Aalen: Limesmuseum Aalen.
Gervers, V. 1973. "Methods of traditional felt-making in Anatolia and Iran," Bulletin de
Liaison du Centre International d'Etude des Textiles Anciens 38: 152-63.
Gerziger, D. 1975. "Eine Decke aus dem sechsten Grab der 'Sieben Bruder,'" Antike Kunst
18: 51-55.
Giroire, C. 1997. "Tissage aux cartons a Antinoe," Bulletin de Liaison du Centre Interna-
tional d'Etude des Textiles Anciens 74: 6-17.
Gleba, M. 2004. "Linen production in pre-Roman and Roman Italy," in Alfaro et al. 2004:
29-38.
Godwin, H. 1967. "The cultivation of hemp,'' Antiquity 41: 42-46, 137-40.
Good, I. 1995. "On the question of silk in pre-Han Eurasia," Antiquity 69: 959-68.
Granger-Taylor, H. 1982. "Weaving clothes to shape in the ancient world: The tunic and
toga of the Arringatore," Textile History 13: 3-25.
Granger-Taylor, H. 1987. "The emperor's clothes: The fold lines,'' Bulletin of the Cleveland
Museum of Art 74: 114-23.
Granger-Taylor, H. 1992. "The grouping of warp threads for areas of weft-faced decoration
in textiles of the Roman period: A means of distinguishing looms?" Vlaamse Ver-
eniging voor Oud- en Hedendaags Textiel Bulletin 1992: 19-28.
Gudj6nsson, E. 1990. "Some aspects of the Icelandic warp-weighted loom, vefstadur,''
Textile History 21: 165-79.
Halleux, R. 1981. Les alchimistes grecs. Vol. 1. Paris: Societe d'Edition 'Les Belles Lettres'.
Hansen, E. H. 1990a. Tablet weaving: History, techniques, colours, patterns. H0jbjerg: Ho-
vedland.
Hansen, E. H.199ob. "Nalebinding: Definition and description,'' in P. Walton and J.P. Wild
(eds.), Textiles in northern archaeology. North European Symposium for Archae-
ological Textiles 3. London: Archetype, 21-27.
Hoffmann, M. 1964. The warp-weighted loom. Oslo: Universitetsforlaget.
Holliday, P. J. 1993. "The sarcophagus of Titus Aelius Evangelus and Gaudenia Nicene,'' J.
Paul Getty Museum Journal 21: 85-100.
van 'THooft, P. P. M.,M. J. Raven, E. H. C. vanRooij, and G. M. Vogelsang-Eastwood 1994.
Pharaonic and early medieval Egyptian textiles. Leiden: Rijksmuseum van Oudheden.
Hope, C. A., and G. E. Bowen 2002. Dakhleh Oasis project: Preliminary reports on the 1994-
1995 to 1998-1999 field seasons. Oxford: Oxbow.
Jenkins, I. 2003. "The Greeks,'' in D. Jenkins (ed.), The Cambridge history of western textiles.
Jenkins, I., and D. Williams 1985. "Sprang hair-nets: Their manufacture and use in ancient
Greece," American Journal of Archaeology 39: 411-18.
De Jonghe, D. 1978. "Met selectieroeden geweven Koptische weefsels,'' Bulletin van de
Koninklijke Musea voor Kunst en Geschiedenis 50: 75-106.
De Jonghe, D. 1988. "Aspects technologiques," in J. Lafontaine-Desogne, Textiles coptes

des Musees royaux d'Art et d'Histoire. Brussels: Musees royaux d'Art et d'Histoire, 22-
33.
De Jonghe, D., and M.Tavernier 1978. "Les darnassees de la proche-antiquite," Bulletin de
Liaison du Centre International d'Etude des Textiles Anciens 47/48: 14-42.
Jongman, W. 1991. The economy and society of Pompeii. Amsterdam: Gieben.
Kemp, B., and G. M. Vogelsang-Eastwood 2001. The ancient textile industry of Amarna.
London: Egypt Exploration Society.
Killen, J. T. 1964. "The wool industry of Crete in the late Bronze Age," Annual of the British
School at Athens 59: 1-15.
King, A. C. 1999. "Diet in the Roman world: A regional intersite comparison of mammal
bones," Journal of Roman Archaeology 12: 168-202.
Korber-Grohne, U. 1987. Nutzpflanzen in Deutschland: Kulturgeschichte und Biologie.
Stuttgart: Theiss.
Kraemer, C. J. 1928. "On the skin-clad sheep of antiquity," Classical Weekly 21: 33-35.
Larsson-Loven, L. 1998. "Lanam fecit: Woolworking and female virtue," in L. Larsson-
Loven and A. Stromberg (eds.), Aspects of Women in Antiquity: Proceedings of the first
Nordic Symposium on womens' lives in antiquity. Sovedalen: Astrom, 85-95.
Loftus, A. 2000. "A textile factory in the third century BC Memphis: Labor, capital and
private enterprise in the Zeno archive," in Cardon and Feugere 2000: 173-86.
Lucas, A. 1932. "The occurrence of natron in ancient Egypt," Journal of Egyptian Archae-
ology 18: 62-66.
McGing, Brian C. 1990. "Lease of a linen-weaving workshop in Panopolis," Zeitschrift
ftlrPapyrologie und Epigraphik 89: 115-21.
Maeder, F., A. Hanggi, D. Wunderlin, and G. Carta Martiglia 2004. Bisso marino: Fili d'oro
del fondo del mare. Basel: Naturhistorisches Museum.
Maltby, J.M. 1981. "Iron-Age, Romano-British and Anglo-Saxon animal husbandry," in M.
Jones, G. Dimbleby (eds.), The environment of man: The Iron Age to Anglo-Saxon
period. British Archaeological Report 87. Oxford: BAR: 155-203.
Melena, J. L. 1987. "On the Linear B ideogrammatic syllogram ZE," in J. T. Killen, J. L.
Melena, and J.-P. Oliver (eds.), Studies in Mycenaean and classical Greek presented to
John Chadwick. Minos 20-22: 389-457.
Moeller, W. 0. 1976. The wool trade of ancient Pompeii. Leiden: Brill.
Morel, J.-P. 1978. "La laine de Tarente (de l'usage des textes anciens en histoire econo-
mique)," Ktema 3: 93-110.
Moulherat, C., and G. Vial. 2000. "Premiere attestation d'un tissu en laine de chevre
cachemire en Gaule," in Cardon and Feugere 2000: 107-13.
Panagiotakopulu, E., P. C. Buckland, P. M. Day, C. Doumas, A. Sarpaki, and P. Skidmore
1997. "A lepidopterous cocoon from Thera and the evidence for silk in the Aegean
Bronze Age," Antiquity 71: 420-29.
Pietrogrande, A. L. 1976. Le fulloniche. Scavi di Ostia 8. Rome: Libreria dello Stato.
Rinuy, A. 2000. "Analyse der Goldfaden," in A. Schmidt-Colinet, A. Stauffer, and K. al-
As'ad (eds.), Die Textilien aus Palmyra. Mainz: von Zabern, 16-19.
Robertson, R.H. S. 1986. Fuller's earth: A history. Hythe: Volturna.
Robkin, A.-L. 1979. "The agricultural year, the commodity SA and the linen industry of
Mycenaean Pylos," American Journal of Archaeology 83: 469-74.
Ryder, M. L. 1969. "Changes in the fleece of sheep following domestication (with a note on
the coat of cattle)," in P. J. Ucko and G. E. Dimbleby (eds.), The domestication and
exploitation of plants and animals. London: Duckworth, 495-521.
Ryder, M. L. 1981. "Wools from Vindolanda," Journal of Archaeological Science 8:

99-103.
Ryder, M. L. 1983. Sheep and man. London: Duckworth.
Ryder, M. L. 1988. "Report on the wool," in V. Tatton-Brown (ed.), Cyprus and the eastern
Mediterranean in the Iron Age: Proceedings of the seventh British Museum Classical
Colloquium, April 1988. London: British Museum, 154-55.
Ryder, M. L. 1990. "The natural pigmentation of animal textile fibres," Textile History 21:
135-48.
Ryder, M. L. 1993. "The use of goat hair: An introductory historical review," Anthro-
pozoologica 17: 37-46.
Ryder, M. L. 1995. "Fleece grading and wool sorting: The historical perspective," Textile
History 26: 3-22.
Ryder, M. L. 2001. "The fibres in textile remains from the Iron-Age salt-mines at Hallstatt,
Austria," Annalen des Naturhistorischen Museums Wien 102A: 223-44.
Scheid, J., and J. Svenbro 2003. Le metier de Zeus: Mythe du tissage et du tissu dans le monde
greco-romain. Paris: Errance.
Schmidt-Colinet, A., A. Stauffer, and K. al-As'ad 2000. Die Textilien aus Palmyra. Mainz:
von Zabern.
Schrenk, S. 2001. "Die spatantiken Seiden in der Schatzkammer des Kolner Domes," Koiner
Domblatt 66: 83-118.
Schrenk, S. 2004. Textilien des Mittelmeerraumes aus spiitantiker bis fruhislamischer Zeit.
Riggisberg: Abegg-Stiftung.
Sheffer, A., and H. Granger-Taylor 1994. "Textiles from Masada: A preliminary selec-
tion," in E. Netzer (ed.), Masada IV: The Yigael Yadin excavations, 1963-1965: Final
reports. Jerusalem: Israel Exploration Society, 151-255.
Sherratt, A. 1983. "The secondary exploitation of animals in the Old World," World Ar-
Small, A., V. Volterra, and R. G. V. Hancock. 2003. "New evidence from tile-stamps
for imperial properties near Gravina and the topography of imperial estates in SE
Italy," Journal of Roman Archaeology 16: 179-99.
Staermose Nielsen, K.-H. 1999. Kirkes Vaev: Opstadvaevens historie og nutidige brug. Lejre:
Historisk-Arkaeologisk Fors0gscenter.
Tebar Megias, E., and A. Wilson forthcoming. "Classical and Hellenistic textile production
at Euhesperides (Benghazi, Libya): Preliminary results," in C. Alfaro Giner (ed.),
Textiles and dyes in the ancient Mediterranean world: 2nd international symposium,
Athens, 24-26 November 2005.
Thomas, T. K. 2001. Textiles from Karanis, Egypt, in the Kelsey Museum of Archaeology:
Artifacts of everyday life. Ann Arbor: Kelsey Museum.
Thompson, J., and H. Granger-Taylor 1995-1996. "The Persian zilu loom of Meybod,"
Bulletin de Liaison du Centre International d'Etude des Textiles Anciens 73: 27-53.
Tuck, A. 2006. "Singing the rug: Patterned textiles and the origins of Indo-European
metrical poetry," American Journal of Archaeology no: 539-50.
Verhecken-Lammens, C. 1992. "Opzetboorden bij Koptische weefsels," Vlaamse Verenig-
ing voor Oud- en Hedendaags Textiel Bulletin 1992: 29-36.
Wace, A. J.B. 1948. "Weaving or embroidery?" American Journal of Archaeology 52: 51-55.
Walton Rogers, P. 1994. "Types of wool in a Roman damask tunic Abegg-Stiftung no.
4219," Riggesberger Berichte 2: 37-40.
Walton Rogers, P. 1997. Textile production at 16-22 Coppergate: The archaeology of York: The
small finds 17/11. York: York Archaeological Trust.
Whittaker, C. R. 1988. Pastoral economies in classical antiquity. Cambridge Philological

Society Suppl. 14. Cambridge: Cambridge Philological Society.
Wild, J. P. 1964. "The textile term scutulatus," Classical Quarterly 14: 263-66.
Wild, J.P. 1967a. "Two technical terms used by Roman tapestry-weavers," Philologus m:
151-55.
Wild, J.P. 1967b. "Soft-finished textiles in Roman Britain," Classical Quarterly 17: 133-35.
Wild, J.P. 1968. "The Roman flax hackle (aena)," Museum Helveticum 25: 139-42.
Wild, J.P. 1970. Textile manufacture in the northern Roman provinces. Cambridge: Cam-
bridge University Press.
Wild, J.P. 1982. "Wool production in Roman Britain," in D. Miles (ed.). The Romano-
British countryside: Studies in rural settlement and economy. British Archaeological
Report 103. Oxford: BAR, 109-22.
Wild, J.P. 1984. "Some early silk finds in northwest Europe," Journal of the Textile Museum
23: 17-23.
Wild, J.P. 1987. "The Roman horizontal loom," American Journal ofArchaeology 91: 459-71.
Wild, J.P. 1994. "Tunic no. 4219: An archaeological and historical perspective," Riggisberger
Berichte 2: 9-36.
Wild, J.P. 1997. "Cotton in Roman Egypt: Some problems of origin," Al-Rafidan 18: 287-
98.
Wild, J.P. 1999. "Textile manufacture: A rural craft?" in M. Polfer (ed.), Artisanat et
productions artisanales en milieu rural dans les provinces du Nord-Quest de l'empire
romain. Monographies Instrumentum 9. Montagnac: Monique Mergoil, 29-37.
Wild, J.P. 2000. "Textile production and trade in Roman literature and written sources,"
in Cardon and Feugere 2000: 209-13.
Wild, J.P. 2002. "The textile industries of Roman Britain," Britannia 33: 1-42.
Wild, J.P. 2003a. "Facts, figures and guesswork in the Roman textile industry," in L.
Bender-J0rgensen, J. Banek-Burgess, and A. Rast-Eicher (eds.), Textilien aus Arch-
aologie und Geschichte: Festschrift fur Klaus Tidow. Neumiinster: Wachholtz, 37-45.
Wild, J.P. 2003b. "The Romans in the West," in D. Jenkins (ed.), The Cambridge history of
Western textiles. Cambridge: Cambridge University Press, 77-93.
Wild, J.P. 2003c. Textiles in archaeology. Aylesbury: Shire.
Wild, J.P. 2004. "The Roman textile industry: Problems but progress," in Alfaro et al. 2004:
23-28.
Wild, J.P. 2005. "Methodological introduction," in M.-L. Nosch and C. Gilles (eds.),
Ancient textiles: Production, crafts and society. Oxford: Oxbow.
Wild, J. P, F. C. Wild, and A. J. Clapham forthcoming. "Roman cotton revisited," in C.
Alfaro Giner (ed.), Textiles and dyes in the ancient Mediterranean world: 2nd inter-
national symposium, Athens, 24-26 November 2005.
Wilson, A. 2001. "Timgad and textile production," in D. J. Mattingly and J. Salmon (eds.),
Economies beyond agriculture in the classical world. London: Taylor and Francis, 271-96.
Wilson, A. 2003. "The archaeology of the Roman fullonica," Journal of Roman Archaeology
16: 442-46.
Wipszycka, E. 1965. L'industrie textile dans l'Egypte romaine. Warsaw: Polskiej Akademii
Nauk.
Wouters, J. 1995. "Dye analysis in a broad perspective: A study of 3rd- to 10th-century
Coptic textiles from Belgian private collections," Dyes in History and Archaeology 13:
38-45.
Yates, J. 1843. Textrinum antiquorum: An account of the art of weaving among the ancients.
Part 1. London: Taylor and Walton.
CHAPTER 19
TANNING
AND LEATHER
CAROL VAN ORIEL-MURRAY
LEATHER is antiquity's plastic, supplying a versatile, supple, hardwearing, and wa-

terproof material. It is also one of the few organic materials to survive intact in
any quantity, providing a wealth of both direct and indirect evidence for the form
and construction of essential military equipment such as saddles and tents, along
with military and domestic footwear and clothing. This class of artifact can also be
a source of social and economic information (van Driel-Murray 1987, 1998a, 1998b,
1999). In function, leather complements textiles and may have provided a cheaper
and longer-lasting alternative to woven cloth, particularly when employed for
wrappings, awnings, and tents. This contribution draws extensively on previous
studies, where further literature can be found (van Driel-Murray2000, 2001a, 2002b,
2003).
Leather is simply the resilient portion (collagen) of animal tissue preserved by
means of drying (resulting in rawhide), or by curing with smoke, fat (chamoising),
or soaking in a mineral bath (tawing), or by tanning with vegetable extracts. Tanning
is often regarded as an industrial process of immense antiquity, with techniques
virtually unchanged until modern times. Archaeological research has shown, how-
ever, that true tanning involving the use of vegetable extracts cannot be traced back
much further than the fifth century B.C., and in many regions it was only introduced
after the Roman conquest. Indeed, in northern Europe outside the Roman Empire,
tanning remained unknown until about the seventh century A.D. Even in Egypt,
artifacts of vegetable-tanned leather only become common under Roman rule: here,
as in Europe, the Romans stimulated major technological developments in lea-
therworking. But these developments were not necessarily permanent, and some
lO00BC S00BC 0 ADS00 AD 1000

I
111111111
Egypt --·•-----1111111111)1
I
I
I
Q. !brim
•:
I
Southern ? Ill 1111

Europe
North
European
Provinces
•
II
I
I
11111111111111111111
11111 Ill
1111 •
•
I
I
I
I 1111
I
Barbarian 1111
Europe
- Cured - Tanned
Figure 19.1. The spread of vegetable tanning based on surviving archaeological evidence.
regions lost all knowledge of tanning following the Roman withdrawal in the fourth
century A.D. and reverted to simpler methods (figure 19.1).
Although the role played by tanners and shoemakers in classical literature is
interesting in itself, references are mainly anecdotal and are rarely helpful in as-
sessing the nature of leatherworking or its scale, particularly as interpretations of
terminology are inevitably colored by modern preconceptions concerning the
technology used (Forbes 1966; Lau 1967; Leguilloux 2004: 60-65). Although the use
of preserved hairy skins and fur is attested outside the Roman frontier, there is little
evidence for the use of furs by the Mediterranean elites before the eighth century
A.D. (Howard-Johnston 1998). This cultural antipathy toward fur-clad barbarians,
as opposed to civilized wearers of processed clothing (Pausanias 10.38.3), extended
to all those concerned with leather working, imparting a subtext of physical and
moral pollution which was gratefully exploited by satirists such as Aristophanes
(Eq. 315-21; cf. Artemidorus 1.51, 2.20).
Archaeological finds have brought new perspectives to the subject, the most
detailed information coming from the large amount of military and civilian lea-
therwork preserved in waterlogged deposits in the northwestern provinces of the
Roman Empire. In the Mediterranean region organic materials do not survive as
well, and although the dry conditions of Egypt have preserved a wide variety of
leather artifacts, their dating is often problematical.
TANNING AND LEATHER 485
SKIN PROCESSING
The processes involved in transforming animal skins into their more permanent
form as leather are well understood (Table 19.1; Forbes 1966; Leguilloux 2004).
Contrary to general assumption, the techniques were not necessarily similar to
medieval and more recent practices; indeed, certain procedures such as liming and
bating are not attested before the fourteenth century or even later. Two basic forms
of skin processing are archaeologically recognizable: curing and tanning. Curing
includes relatively simple methods of delaying the onset of decay, by means of
smoking or applications of fat or mineral earth. These processes are chemically
unstable and reversible, limiting survival. Tanning is a complex process involving
infusions of tannins extracted from tree bark or oak galls. The tannins combine
permanently with the skin collagen, resulting in a chemically stable product that
is water resistant and not susceptible to bacterial decay. Vegetable tanned leather
survives well in waterlogged conditions.
Curing
The earliest methods of preserving skin seem to be smoking and/or rubbing in oils
or fats, followed by vigorous stretching to ensure pliability. In ancient Egypt and
Mesopotamia, sesame oil was the standard dressing, and the importance of ca-
melina sativa (Gold-of-Pleasure) to Iron Age communities may also be linked to
the use of its thin oil in skin processing. The earliest Greek references are also
consistent with curing by means of such methods (Homer, Il. 17.389-93). Dressing
with oil to produce light-colored leather continued in parallel with vegetable
tanning, and its widespread use for footwear may explain the poor survival of
shoe uppers as opposed to soles in Roman find complexes. The red, purple, and
white leathers of classical literature imply the use of cured leather, which is easier to
dye.
In ancient Mesopotamia and Egypt, alum occurs solely as a mordant in com-
bination with madder used to stain leather red. The utility of alum in producing soft
white leather was certainly known to the Romans, even if the term aluta seems to
refer rather to the color of the leather and not to the substance used to produce it.
Pliny (HN 35.183-91) only mentions alum in passing as "giving a finish to skins and
textiles," that is, as a mordant. Since leather tawed with alum is not moisture
resistant, the leather sails of the Veneti, which Caesar describes as aluta (B Gall. 3.13),
must actually have been oil-drenched. The Richborough Type 527 amphoras have
been linked to the trade in alum from the Lipari Islands to Britain, but the relevance
to tanneries is uncertain, since alum has numerous other applications (Borgard et al.
2002; Leguilloux 2004: 31-32).
Cured leathers survive mainly in dry conditions, such as in Egypt, where re-
search on Roman sites is beginning to reveal the extent of the use of untanned
lures associated with tanning.
Procedure Residue Tools or Structures Time Required E
Cut and strip hide None Knives; frame or table A few hours C
from carcass
Trim with knives Horns, hoofs Knives A few hours I-
C
Soaking to loosen Ash deposits, mashed Pits, vats, wood, bone 1-5 days (repeated) I-
hairs in lye or biolog- grains tools
ical agents (ferment-
"
ing bran)
Alkaline soak Lime Pits, vats 2-6 days I-
"
Soaking in acid or Dung, bran mash Pits, vats A few hours to 10 days I'-
alkaline bath
Washing, scraping Reuse of de-hairing Wood, bone tools, A few hours F

agents water-supply t;
d
Immersion in tannins Deposits of finely Permanent vats, pits, Successive baths of F
ground vegetable water channels increasing strength;
matter 9-12 months for hide
Air drying None Covered sheds and Weeks to months I'-

frames
Smoothing and feed- Oil, fat Tables, rubbers, A few hours to a few F
ing leather pounders, clamps days le
S.
leathers for water bags and other equipment (Leguilloux 2004: 142; Winterbottom
1990; Volken forthcoming b ). Cato (Agr. 135.3-5) describes how fresh hides, dressed
with fat, are dried and twisted into cables for wine presses. In Dura Europos,
untanned skins were used for shield surfaces: the moistened rawhide contracts as it
dries, clamping the shield boards together, and providing a good surface for painted
decoration (James 2004: 163, 186-87). No such artifacts would survive in north
European contexts. Although the presence of vegetable tannins in ancient desic-
cated leather can be established by analysis, distinguishing among the various
methods of curing is not yet feasible (van Driel-Murray 2002a).
Vegetable Tanning
The sudden visibility of leather items in the archaeological record of northwestern
Europe following the Roman conquest is the result of the introduction of vegetable
tanning into regions previously unfamiliar with this technology. For the Medi-
terranean region there is no such archaeological evidence, and literary references
are ambiguous, as the substances later used for tanning first appear independently
as dyestuffs or astringents (Aristotle, [Pr.] 32.18; Dioscorides 1.106-11). By the time
of Theophrastus (Hist. pl. 3.8.6, 4.2.8; cf. Pliny HN 16.26, 24.91), however, the
properties of different trees with regard to tanning seem to be thoroughly familiar,
suggesting that the method was already well established by the close of the fifth
century B.C. Iron hobnails in the so-called House of Simon the Shoemaker in the
Athenian agora imply sturdy soles perhaps more likely to be made of tanned hide
(Thompson 1960, dated to ca. 450-410 B.C.; Theophrastus, Char. 4.15), but there is
no further datable evidence for nailed footwear before the mid-first century B.C.,
nor is it depicted in Greek or Etruscan art. The sandals and apparently soft, pliable
shoes attested by Greek shoe-vases and depicted on Etruscan frescoes could all have
been made of cured leather, although when shoes with separate soles are colored
black, tanned leather may be intended. The shoe blacking described by Pliny is the
basis for a modern tanning test (HN 34.123-24; van Driel-Murray 2002a).
Tanning is a complex and time-consuming process, requiring significant in-
vestment in installations and raw material. Its introduction into northern Europe
following the Roman conquest profoundly changed the nature of native crafts in
Gaul and Britain, affecting not only procurement of raw materials but also the
organization of labor. In many nonurban societies leather working is the task of
women, but as hides and skins become a commercial commodity, there is a shift
from domestic production to a male controlled, centralized, fulltime industry, al-
tering women's economic roles in the household economy (van Driel-Murray 2001a:
58 ). Some indication of the amount of capital which might be tied up during tanning
is given by markups of 25% and more between the price of raw and finished hides
listed in Diocletian's Edict (Lauffer 1971: VIIl.9-14). The relatively simple procedures
involved in curing leather could have been accommodated in nondescript struc-
tures; tanning, however, required distinctive and dedicated complexes.
Tanneries
In ancient Athens, a tanning quarter lay well away from the center, and in Rome it
was apparently situated on the right bank of the Tiber, although there is no sup-
porting archaeological evidence for the nature of the installations. Vegetable tan-
ning involves immersion in successive vats of tan liquor of increasing strength,
making a series of interlinked vats a basic requirement, together with drying rooms,
good access to water, and provision for waste disposal. The only tannery identified
in Pompeii (Reg. l.5) contains 15 closely-set vats and various other built-in features.
The contrast with the number of fulleries is striking, although-if the fulleries were
in fact primarily laundries (Bradley 2002)-the disparity only emphasizes the dis-
tinction between servicing and manufacturing trades within the urban fabric. Le-
guilloux (2002: 44-50) reconstructs the route of operations through the building,
revealing a logical and well thought out enterprise, which, considering the inade-
quate water supply, probably specialized in luxurious alum- and oil-cured leathers
or in finishing skins cleaned and depilated elsewhere. A small tannery in the civilian
settlement of Vitudurum (Switzerland) contains six interconnected vats sur-
rounded by drying rooms (figure 19.2; Janke and Ebnother 2001: fig. 4). Most other
sites mentioned in this context can be rejected on grounds of structural unsuit-
ability or incorrect interpretation of the finds assemblage, with the result that-
despite the importance of the leather trade-evidence for tanneries in urban sit-
uations remains remarkably rare. It is a common misconception that dumps of
worn-out leather goods are indicative of tanneries and workshop remains. Vin-
dolanda, for example, was regarded as a tannery in the 1970s, but the leather
deposits have since been recognized as fill in abandoned structures. The deposits
at Catterick and Walbrook (and many other similar sites) are simply leather-rich
rubbish dumps (Hooley 2002; van Driel-Murray 2003: no). Only isolated tanks
were found at Liberchies (Belgium), and the bone refuse indicates glue boiling
(Dewert-Brulet and Vilvorder 2001: 401-5).
In the Roman context, tanning is but one element in the systematic and highly
efficient processing of animal carcasses, in which the waste from one craft activity
forms the raw material of the next (Vanderhoeven and Ervynck 2007). The com-
position of waste products is culturally determined by the demand (or lack of it) for
particular products, and in the Roman period bone refuse-which characterizes
medieval tannery locations-should probably be regarded as an indicator of bone
working rather than a by-product ofleather production. Associations of pits with
horn cores identify horn working rather than tanneries, although the two may be
intimately connected, as at Tongeren (Vanderhoeven and Ervynck 2007). Collec-
tions of foot bones attest to the enormous demand for sinew, the uses of which
range from cobbler's sewing thread to heavy cables for artillery pieces (Rodet-
Belarbi et al. 2002: 323). A text from Vindolanda mentions delivery of no less than
100 pounds of sinew in one consignment (Bowman and Thomas 1994: no. 343.2;
Lewis 2004: 48; James 2004: 209). Waste from the tanneries themselves stimulated
II
lI
·- ·-·-·,
:: OOo~o ci;
--LI_ ~-~~l
ROmerstr. 227/229 \ \_ , I
·1
I I I
i I I I
I 10 m
_._J
I /
L ----·-·'
Hauptstrasse
Figure 19.2. Plan of the tannery at Vitudurum. (After Janke and Ebnother 2001: fig. 4.)
further industries around them (cf. table 19.1), while spent tan could be used as a
fertilizer (Pliny, HN 17.46, 258). The analysis of organic deposits and insect remains
is important in differentiating these interdependent and noisome trades (Hall and
Kenward 2003: 122-23).
Reliance on inappropriate models drawn from more recent practices may also
hamper the identification of tannery structures. Instead of the small, family-run
backyard businesses of medieval times, constrained by guild restrictions and urban
toll barriers, we should perhaps be looking for large complexes at the edges of
settlements or further afield, with access to water the main factor. Conditions in
the Roman Empire seem to have stimulated rural industries, and although rural
potteries are well known, tanneries are less likely to be recognized and excavated. It
is worth noting that in Britain, improvements in transport and the relaxation of
guild restrictions led to the establishment of rural tanneries in the eighteenth
century, and it is perhaps significant that in the second century A.O. Artemidorus
envisages tanners banished to the countryside (1. 51, 2.20).
SCALE OF PRODUCTION
Only anecdotal evidence is available for the nature of Greek tanneries, but for the
Roman Empire the massive scale of operations and the high degree of organization
is attested in the size and variety of leatherwork complexes from northern Europe.
Experimental research gives some insight into both the scale of Roman military
demands and the highly efficient methods of production (Bishop and Coulston
2006: 247-48; Himmler forthcoming; Volken forthcoming a). A legion required a
minimum of 68,ooo goatskins to make tents, and more than 3,000 cattle hides
annually for boots alone. What with bags, equipment covers, and animal harnesses,
equipping the army was a formidable task requiring time and forward planning.
Judging from the quality of surviving military leatherwork, Roman commanders
already had highly successful supply systems in place by the first century B.C. A
fragment of a leather tent from Caesar's siege at Alesia employs the same complex,
waterproof construction as the tents used in Britain 150 years later, while military
footwear found in Qasr Ibrim (Egypt) is virtually identical in construction to that
from northern forts. Techniques perfected in Republican times went virtually un-
changed until the army began to devolve supply into civilian hands as expansion
ceased (van Driel-Murray 1985; 1999; 2001b: 362-64; 2001d).
Efficient tanning is a continuous process dependent on a regular supply of
hides and skins. Despite the prominence of exotic skins in Diocletian's Edict, the
vast majority ofleather artifacts recovered were made of bovine hide (footwear) or
goatskin, with doeskin occasionally used for luxury items. Urban and military
slaughterhouses were best placed to meet this demand, but imports of salted skins
feature regularly in accounts of trade with foreign lands and on Roman tariff lists
(Leguilloux 2004: pl. Ilb). Wholesalers, like those of the corpus pellionum at Ostia
( CIL 14. supp. 1, no. 4549 2 ) form another indication of the scale of the industry. The
army also drew on local sources of tribute: an unreasonable exaction of hides im-
posed on the Frisians caused the revolt that ended Roman aspirations north of the
Rhine delta (Tacitus, Ann. 4.72). Tariff differentiation on the second-century Zarai
list from Numidia (CIL 8.4508) suggest that the trade in raw (presumably salted)
skins was welcomed to serve the needs of the Roman industry, while that of pro-
cessed leather was discouraged.
LEATHERWORKERS AND TANNERS
Cleon of Athens (425 B.C.) seems to be an early representative of a class of wealthy

entrepreneurs investing in tanneries. Since this period coincides with the emer-
gence of distinct words for tanners (bursodepses, skulodepses) and for shoemakers
(skutotomos, neurorraphos, skuteus) in Greece, it may mark the point at which

vegetable tanning, with its complex organizational structure, emerges as a distinct
trade in the classical world (Lau 1967: 56). Despite their wealth, few certain tomb-
stones of tanners are known, even though what seem to be individual shoemakers
appear regularly on monuments in Rome and the provinces, as well as on the much
earlier Greek vases (cf. Zimmer 1982: nos. 47-55; Leguilloux 2004; De Spagnolis
2000: 64-72). An elaborate tomb belonging to a supposed shoemaker at Pizzone,
with its unexpected link to other inscriptions at Ostia, suggest profits from man-
ufactories on a much larger scale than one might guess from the depictions of lone
shoemakers working modestly at their bench (De Spagnolis 2000: 53-55, 71-72). The
discrepancy raises the suspicion that such tombstones might belong to entrepre-
neurs, illustrating the source of their wealth by means of a single representative
vignette, and are not intended to show the deceased in his actual daily work (cf.
chapter 2). Men like C. Iulius Alcimus, a comparator mercis sutoriae from Ravenna,
and C. Iulius Helius, sutor (the caliga on his exceptionally fine tombstone hinting
at military contracts), were controllers of labor, not workmen, even if shoemaking
was the most presentable of their economic activities (GIL 5.5927, 6.33914). The
craftsmen themselves are perhaps attested by control stamps on footwear that
suggest they banded together in corporations, not only in Rome but also in small,
lackluster towns such as Noviomagus Batavorum (Nijmegen) (van Driel-Murray
2001b: 338-39).
Other forms of organization can also be distinguished. In military workplaces
and large establishments such as the Egyptian estates, work was streamlined and
highly controlled. An unpublished text from Oxyrhynchus dealing with the man-
ufacture of harnesses specifies the leftovers, down to a few narrow strips and a couple
of small offcuts (Dominic Montserrat, personal communication). Such meticulous
accounting explains the careful bundles of offcuts of both new and old leather that
are such a feature of excavated military complexes such as Castleford (van Driel-
Murray, 1998b).
TECHNOLOGICAL CHANGE
In the northern provinces of the Roman Empire tanning belongs to a package

of innovations encompassing all aspects of leatherworking, including the sew-
ing and construction of articles, the range of finished products, and the intro-
duction of composite, nailed footwear. Civilian footwear was not usually nailed
in the Mediterranean region, and the popularity of hobnailed shoes in the prov-
inces is probably due to the role of the military in the spread of technical
knowledge. A similar association between the military, hobnailed caligae and veg-
etable tanned leather marks the short-lived Roman occupation of Qasr !brim in
---
Figure 19.3. Fourth-century off-cut with clamp marks, from Cuijk, Netherlands. (Pho-
tograph by A. Dekker, Amsterdam University.)
southern Egypt, a region with strong traditions of fine leatherwork employing

cured skins.
Desirable Roman products such as footwear drew rural communities into
marketing networks, giving all levels of society access to the rapidly changing
fashions. Stylistic dating of footwear, refined for Greece by Morrow (1985), has an
even greater potential in the Roman period, since archaeological contexts provide
securely dated sequences of shoes actually in use (van Driel-Murray 2001b, 2001e).
Despite its initial popularity, by the 330s A.D. nailed footwear was in sharp decline
in the Rhineland, although in southern Gaul it may continue into the fifth century.
Other changes in tanning practice are also discernable, in part to meet the needs of
different shoemaking methods, but also to reduce processing time. More attention
is paid to finishing, and marks of clamps to hold the leather during currying appear
on late Roman offcuts in the Netherlands (figure 19.3; Mould 2003). More pro-
found changes in organization were evidently taking place at the same time, for by
the end of the fourth century leather disappears entirely from the archaeological
record of northwestern Europe. In contrast, tanning continued in Egypt, and the
technology was, indeed, reintroduced into Qasr Ibrim in the so-called Christian
levels. Regional differences suggest that the loss of tanning technology is somehow
linked to the weakening of the urban economy and the breakdown of the system of
supply.
Assessment of the archaeological contribution reveals the limitations of the static

view of ancient technology based on literary sources. In reality, the technology of
skin processing underwent considerable development in scale and technology, linked
to the expansion of the Hellenistic and Roman economic areas. In the absence of
detailed work on the subject in the Mediterranean area, the loss of technological
knowledge after the fourth century A.D. in western Europe cannot be verified else-
where, though the development serves as a reminder that technology is intimately
linked to social and economic factors governing its implementation.
REFERENCES
Audoin-Rouzeau, F., and S. Beyries 2002. Le travail du cuir de la pre'histoire a nos jours.
Antibes: Editions APDCA.
Bishop, M. C., and J. C. N. Coulston 2006. Roman military equipment. 2nd ed. Oxford:
Oxbow.
Borgard, P., V. Forest, C. Bioul-Pelletier, and L. Pelletier 2002. "Passer les peaux en blanc:
Une practique gallo-romaine? L'apport du site de Sainte-Anne aDijon (Cote-d'Or),"
in Audoin-Rouzeau and Beyries 2002, 231-49.
Bowman, A. K., and J. D. Thomas 1994. The Vindolanda writing tablets. (Tabulae Vindo-
landenses II). London: British Museum Press, 217-26.
Bradley, M. 2002. "'It all comes out in the wash': Looking harder at the Romanfullonica,"
Dewert-Brulet, R. J.-P., and F. Vilvorder 2001. Liberchies N Travail de Rivere. Louvain-la-
Neuve: Departement d'archeologie et d'histoire de l'art.
De Spagnolis, M. 2000. La Tomba del Calzolaio: Dalla necropolis monumentale romana di
Nocera Superiore. Rome: Bretschneider.
Forbes, R. J. 1966. Studies in andent technology. Vol. 5, Leather in antiquity. Leiden: Brill.
Hall, A., and H. Kenward 2003. "Can we identify biological indicator groups for craft,
industry and other activities?" in P. Murphy and P. E. J. Wiltshire (eds.), The envi-
ronmental archaeology of industry. Oxford: Oxbow, 114-30.
Himmler, F. forthcoming. "Testing the 'Ramshaw' boot on a long march," Journal of
Roman Military Equipment Studies.
Hooley, A. D. 2002. "Leather from the 1958-9 by-pass excavations (Site 433)," in P.R.
Wilson, Cataractonium: Roman Catterick and its hinterland: Excavations and research,
1958-97 Part II. CBA Research Report 129: 318-80.
Howard-Johnston, J. 1998. "Trading in fur, from classical antiquity to the early Middle
Ages," in E. Cameron (ed.), Leather and fur: Aspects of early medieval trade and
technology. London: Archetype, 65-79.
James, S. 2004. The excavations at Dura Europos conducted by Yale University and the French
Academy of Inscriptions and Letters 1928 to 1937. Final Report VII: The arms and armour
and other military equipment. London: British Museum Press.
Janke, R., and C. Ebnother 2001. "Struktur und Entwicklung des Vicus Vitudurum im 1.
Jahrhundert n. Chr," in G. Precht (ed.), Genese, Struktur und Entwicklung romischer
Stadte im 1. Jahrhundert n. Chr. in Nieder- und Obergermanien. Mainz: von Zabern,
217-26.
Lau, 0. 1967. Schuster und Schusterhandwerk in de griechisch-romischen Literatur und Kunst.
Inaugral-Dissertation Rheinischen Friedrich-Wilhems-U niversitat, Bonn.
Lauffer, S. 1971. Diokletians Preisedikt. Berlin: de Gruyter.
Leguilloux, M. 2002. "Techniques et equipements de la tannerie romaine: L'example de
l'offidna coriaria de Pompei," in Audoin-Rouzeau and Beyries 2002: 267-81.
Leguilloux, M. 2004. Le cuir et la pelleterie a l' epoque romaine. Paris: Editions Errance.
Lewis, M. 2004. "Reconstruction of Heron's cheiroballistra," Current World Archaeology 3:
46-48.
Morrow, K. D. 1985. Greek footwear and the dating of sculpture. Madison: University of
Wisconsin Press.
Mould, Q. 2003. "Man pinches leather?" Archaeological Leather Group Newsletter 18: 1.
Rodet-Belarbi, I., C. Olive, and V. Forest 2002." Depots archeologiques de pieds de

mouton et de chevre: S'agit-il toujurs d'un artisanat de la peau?" in Audoin-Rouzeau
and Beyries 2002: 345-49.
Thompson, D. B. 1960. "The house of Simon the shoemaker," Archaeology 13: 234-41.
Vanderhoeven, A., and A. Ervynck 2007. "Not in my backyard! The industry of secondary
animal products within the Roman civitas capital of Tongeren (Belgium)," in R.
Hingley and S. Willis (eds.), Roman finds: Context and theory. Proceedings of a con-
ference held at the University of Durham, 2002. Oxford: Oxbow, 156-75.
van Oriel-Murray, C. 1985. "The production and supply of military leatherwork in the first
and second centuries AD: A review of the archaeological evidence," in M. C. Bishop
(ed.), The production and distribution of Roman military equipment. British Archae-
ological Reports, Intl. Series S275. Oxford: BAR, 43-81.
van Oriel-Murray, C. 1987. "Roman footwear: A mirror of fashion and society," in D. E.
Friendship-Taylor, J.M. Swann, and S. Thomas (eds.), Recent research in archaeo-
logical footwear. Association of Archaeological Illustrators and Surveyors Technical
Paper no. 8. London [?]: 32-42.
van Oriel-Murray, C. 1998a. "Women in forts?" Geselleschaft Pro Vindonissa, Jahresbericht
1997. Brugg: Vindonissa Museum, 55-61.
van Oriel-Murray, C. 1998b. "The leatherwork from the fort," in H. E. M. Cool and C.
Philo (eds.), Roman Castleford, vol. 1, The small finds. Wakefield: West Yorkshire
Archaeology Service, 285-334.
van Oriel-Murray, C. 1999. "Dead men's shoes," in W. Schluter and R. Wiegels (eds.),
Rom, Germanien und die Ausgrabungen von Kalkriese. Osnabriicker Forschungen zu
Altertum und Antike-Rezeption 1. Osnabriick: Universitiitsverlag Rasch, 169-89.
van Oriel-Murray, C. 2000. "Leatherwork and skin products," in R. T. Nicholson and
I. Shaw (eds.), Ancient Egyptian materials and technology. Cambridge: Cambridge
van Oriel-Murray, C. 2001a. "Technology transfer: The introduction and loss of tan-
ning technology during the Roman period," in M. Polfer (ed.), L'artisanat romain:
Evolutions, continuites et ruptures (Italie et provinces occidentales). Montagnac: Edi-
tions Monique Mergoil, 55-68.
van Oriel-Murray, C. 2001b. "Footwear in the north-western provinces of the Roman
Empire," in 0. W. Goubitz, G. van Waateringe, and C. van Oriel-Murray (eds.),
Stepping through time: Archaeological footwear from prehistoric times until 1800. Zwolle:
Stichting Promotie Archeologie, 337-75.
van Oriel-Murray, C. 2001c. "Vindolanda and the dating of Roman footwear," Britannia
32: 185-97.
van Oriel-Murray, C. 2001d. "Les restes d'une tente Cesarienne en cuir," in M. Redde and
S. von Schnurbein (eds.), Alesia: Fouilles et recherches Franco-Allemandes sur les tra-
vaux militaires Romains autour du Mont Auxois (1991-1997). Memoires de l'Academie
des Inscriptions et Belles-Lettres 22. Paris: de Boccard, 363-68.
van Oriel-Murray, C. 2002a. "Practical evaluation of a field test for the identification of
ancient vegetable tanned leathers," Journal of Archaeological Science 29: 17-21.
van Oriel-Murray, C. 2002b. "Ancient skin processing and the impact of Rome on tanning
technology," in Audoin-Rouzeau and Beyries 2002: 251-65.
van Oriel-Murray, C. 2003. "The leather trades in Roman Yorkshire and beyond," in P.R.
Wilson and J. Price (eds.), Aspects of industry in Roman Yorkshire and the North.
Oxford: Oxbow, 109-23.
Volk.en, M. forthcoming a. "Making the Ramshaw boot: An exercise in experimental

archaeology," Journal of Roman Military Equipment Studies.
Volk.en, M. forthcoming b. "The sarcina and the departure scene from Trajan's Column,"
to appear in Journal of Roman Archaeology.
Winterbottom, S. 1990. "The leather objects," in J. Bingen, "Quatrieme campagne
de fouilles au Mons Claudianus," Bulletin d'Institut Franrais Archeologie Orientale 90:
78-81.
Zimmer, G. 1982. Romische Berufsdarstellungen. Berlin: Gehr. Mann Verlag.
CHAPTER 20
CERAMIC
PRODUCTION
MARK JACKSON
KEVIN GREENE
HISTORICAL, GEOGRAPHICAL, AND

CULTURAL PERSPECTIVES
Applying the labels "Greek" and "Roman" to the study of ceramic technology from
700 B.C. to A.D. 500 involves profound problems of cultural labeling. All areas
brought within Greek and Roman cultural and political hegemony already pos-
sessed distinctive methods of making, decorating, and using pottery-many of
them stretching far back into prehistory. Furthermore, no fundamental technical
changes took place during this 1,200-year period. Thus, what we are studying is less
a question of technology per se than the intersection of diverse political cultures,
artistic styles, trading systems, and forms of consumption.
Greek influence extended to the Mediterranean coasts of Spain and France, and
it rivaled the economic power of the Carthaginians of North Africa long before
Roman expansion absorbed all of these areas. Thus, Italy (including early Rome
itself) was heavily influenced by Greek material culture, while Greek colonies in
France and Spain engaged in trade with "barbarians" to the extent that Greek
pottery is regularly found on Iron Age sites on the Iberian peninsula and north well
into France and Germany (Osborne 1996; Cunliffe 1998: 345). This pattern of trade,
in which amphoras containing wine and oil were accompanied by vessels (metal,
glass, and ceramic) suitable for their consumption, continued into the Roman
period. In the first few centuries A.D., Roman pottery exports can be found beyond
CERAMIC PRODUCTION 497
the borders of the empire from the Scottish Highlands to Scandinavia, from the
Sahara desert to the upper Nile valley, and even across the Indian Ocean in Bengal.
Return cargoes presumably consisted ofluxuries, raw materials, and slaves, forming
a pattern repeated in many post-Roman episodes of colonization and imperial
expansion. Parallels of this kind make it interesting to compare Greek and Roman
ceramics with their medieval and post-medieval counterparts.
In addition to cultural complexity, the study of Greek and Roman ceramics
carries a major historiographical burden. Whereas the technical and aesthetic ac-
complishments of classical architecture and art were closely copied from medieval
and Renaissance times onward, Greek and Roman pottery had no direct successors.
When its forms and finishes were copied in the eighteenth century, potters employed
completely different techniques (e.g., Sevres or Wedgwood "Etruscan" red-figure
kraters in porcelain and stoneware-Charleston 1968: 243, fig. 689, 272, fig. 776).
European potters from the medieval period onward were acutely aware of techni-
cally superior products imported from East Asia-notably Chinese porcelain, with
its high-fired impermeable fabric, which could not be matched in the West until
the early eighteenth century (Freestone and Gaimster 1997: 194-99). The sixteenth-
century Italian writer Pancirolli ranked porcelain second only to the the New World
in his list of discoveries unknown to the ancients (Many Excellent Things Found, Now
in Use among the Modems; 1715). Early Chinese imports were frequently mounted in
precious metals, as if they were rock crystal (e.g., Snodin and Styles 2001: 27, fig. 36).
Even more influential was the Islamic decorated and glazed pottery made as far west
as Spain (Freestone and Gaimster 1997: 110-15). Although classical motifs and myth-
ological scenes were commonplace on European pottery from the Renaissance
onward (Freestone and Gaimster 1997: 116-21), they were achieved by imitating the
appearance of Islamic or Chinese imports, many of which used colorful motifs on a
white background (e.g., Cohen and Hess 1993: 48 [Iznik ware], 59 [Ming porcelain]).
European products-such as Italian majolica and North European delftware-
that copied these elaborately decorated oriental imports retained relatively soft and
permeable earthenware fabrics until the eighteenth century. The only truly im-
permeable pottery made in Europe was stoneware, made from clay that could be
fired at a sufficiently high temperature to fuse the fabric without it melting and
collapsing (Freestone and Gaimster 1997: 122-27). German and Dutch salt-glazed
mugs and bottles were successfully imitated in stoneware in England in the late
seventeenth century as a result of careful experimentation (Snodin and Styles 2001:
128-29). Something very similar had already been achieved by potters making
"purple-gloss" beakers in the New Forest industry oflater Roman Britain (Fulford's
fabric 1a, Fulford 1975: 24-25; Cook and Charleston 1979: pl. 122), but in common
with most Greek and Roman pottery the surface bore a slip, not a glaze. Like some
of its oriental counterparts (e.g., Ding ware; Freestone and Gaimster 1997: 182-87),
the earliest porcelain manufactured in Europe in the early eighteenth century came
from state-sponsored workshops. Royal control guaranteed prestigious use in
aristocratic households, and elaborately decorated Meissen or Sevres porcelain
joined glass and silverware for serving and eating meals (Charleston 1968: 216-24,
236-45). In contrast, British entrepreneurs such as Wedgwood produced robust

and well-glazed earthenware or stoneware for a wider market in the later eighteenth
century. The prestige of British porcelain was maintained at centers such as Chelsea
or Worcester by making ornate hand-painted vases (Snodin and Styles 2001: 294)
and exclusive table services, emblazoned with coats of arms, for fashionable indi-
viduals such as Lord Nelson.
Many aristocrats took part in the Grand Tour in the eighteenth century and
visited Italy, where extensive excavations of ancient cemeteries and other sites were
underway. It became fashionable to buy Greek red- or black-figured pottery to
display in neoclassical houses (Greene 2002: 19, fig. 1.6), although much of it was
actually thought to be Etruscan at the time (Sparkes 1996: 34-63). Thus, by a
historical coincidence, Greek vases first became widely known at a time when
contemporary ceramic products were used by royalty and the aristocracy. This gave
rise to a perception of Greek painted pottery as the artistic and economic equivalent
of expensive porcelain (Vickers and Gill 1994: 19-27).
POTTERY IN GREEK
AND ROMAN SOCIETY
A high regard for Greek vases was sustained by admiration of the comparative
purity of Greek, rather than Roman, artistic and architectural models for Greek
Revival architecture and interior design in the late eighteenth and nineteenth
centuries. Such tastes also embodied respect for ancient Greek democracy (and its
associated philosophy and science), rather than Roman imperialism-a sentiment
heightened by the liberation of Greece from Turkish rule in 1830. By the nineteenth
century it was common to use Greek and Roman pottery as evidence for the decline
of ancient civilization (Greene 2005: 37-39) in a manner epitomized by an expert at
the British Museum:
We have traced the development of painted decoration from monochrome to
polychrome, from simple patterns to elaborate pictorial compositions, and so to
its gradual decay and disappearance under the luxurious and artificial tendencies
of the Hellenistic Age, when men were ever seeking for new artistic departures,
and a new system of technique arose which finally substituted various forms of
decoration in relief for painting. And lastly, we have seen how this new system
established itself firmly in the domain of Roman art, until with the gradual decay
of artistic taste and under the encroachments of barbarism, it sank into neglect
and oblivion. (Walters 1905: 2: 554-55)
An additional negative perception of Roman pottery arose from the influence of
the Arts and Crafts movement, which associated industrial production with poor
craft skills:
Roman vases, in a word, require only the skill of the potter for their comple-
tion, and the processes employed are largely mechanical, whereas Greek vases
called in the aid of a higher branch of industry, and one which gave scope for great
artistic achievements-namely, that of painting. . . . The Romans, who used
metal vases to a far greater extent than the Greeks-at least under the late Re-
public and Empire-did not hold the art of pottery in very high estimation, and
their vases, like their tiles and lamps, were produced by slaves and freedmen,
whereas at Athens the potter usually held at least the position of a resident alien.
(Walters 1905: 2:430, 433-34)
The final element in the elevated perception of Greek pottery was the growth of
connoisseurship in the twentieth century, pioneered by Beazley and modeled on
studies of early Renaissance artists and their workshops (Beazley 1951; Boardman
2001: 128-38). This approach has been challenged since the 1970s, notably by
Vickers, who has questioned the status of individual "artists" who painted vases,
and the monetary (and social) value of pottery in the Greek world. Vickers' position
is that pottery was never anything more than a cheap substitute for metalwork, and
that the art-historical model is inappropriate (Vickers and Gill 1994). Neer has
summarized the strength of the conventional position, using documentary evi-
dence to conclude, "Given that ceramic fine ware was made for an elitist pastime,
adorned with elitist pictures, and described in elitist literature, it seems reasonable
to conclude that it was used by elites" (Neer 2002: 215). Support has come from
archaeological fieldwork in Greece, where more fine pottery associated with formal
dining has been recovered from Greek than from Roman settlement sites (Bintliff
2002: 31). Boardman (2001: 122-27)-like Walters in the passage quoted above-
acknowledges the greater role of metal vessels after the demise of painted vases.
This debate underlines the importance of taking modern preconceptions into
account, for it is impossible to assess the technology of Greek and Roman pottery if
we do not know its social and economic status. Modern consumerism maintains a
contradictory cultural attitude to mass production and individuality; the latter may
be achieved either through making things by hand or by manufacturing them
exclusively according to the ideas of a named designer. A visitor to the ceramic
displays of a museum may see a Roman "Arretine" terra sigillata bowl with relief-
molded decoration and the name-stamp of an Italian workshop (Hayes 1997: pl.
18.2). This is likely to create an impression oflower status and sophistication than
an Athenian red-figure calyx-krater signed by an artist, Euphronius-especially if
the viewer knows that the Metropolitan Museum of New York paid $1 million for it
in the early 1970s (Sparkes 1996: 35, fig. 11.1).
A hard-nosed economist might take a very different view. Who were the cus-
tomers for Greek vases? What were the rival products? How large was the scale of
production? How effectively were the products distributed through a mass market?
Did classical Athens have a small and inelastic distribution of wealth, and did the
very rich display their status by means of exquisite gold and silver tableware,
accompanied perhaps by rock crystal and glass vessels (Vickers and Gill 1994: 33-
54)? Whatever the answer, since the work of skilled artisans was relatively cheap,
modest households would have benefited from the expansion of Greek colonization
and trade, and could buy hand-painted pottery like that of their social superiors.
Potters' workshops could also sell their wares to merchants trading wine into less
sophisticated markets around the Black Sea and the western Mediterranean (Os-
borne 1996). As economic development continued, customers who had previously
bought fine ceramics moved up to metal and glass, while a new (and larger) number
could now afford mass-produced molded pottery and tablewares with simpler
decorative schemes.
While Hellenistic molded pottery and plainer fine wares circulated in some
quantity around the Mediterranean, their distribution was modest compared to
Roman terra sigillata and other red-slipped wares, first from Italy, and then from
centers in northwestern Europe, North Africa, and Asia Minor, which provided
basic and serviceable table vessels in astonishing quantities (Johns 1971; Hayes 1972).
The scale of manufacture and extent of distribution of terra sigillata and red-slip
wares to modest households recalls Staffordshire earthenware or stoneware rather
than porcelain (Charleston 1968: 259-70). They also reveal a high degree of cultural
conformity throughout the Roman world to ideas about the appropriate color and
form of tablewares. Our economist might well suspect the existence of a mass
market, and deduce from the simplicity of decorative styles and techniques that
pottery was aimed at a still lower social level than before. Further economic growth
apparently now allowed even the modestly rich to use blown glass and metal vessels
to a far greater extent than had been possible in classical Greece. A switch to
tablewares with red surfaces that mimicked the coppery sheen of ancient gold may
indicate that the elite had become accustomed to using gold vessels at meals, and
had left behind "black-glazed" pottery that imitated the dark lustrous finish favored
by users of silver (Vickers and Gill 1994: 174-78). Thus, while art historians may
regard long-term changes in the quality and manufacturing technique of ceramics
as symptoms of cultural decline, our fictional economist might interpret them as
indicators of expanding wealth.
Discussion of differing social and economic perceptions of Greek and Roman
ceramics is not an abstract exercise. Broken pottery is rarely useful for secondary
purposes, because conversion of the raw materials to fired clay makes it difficult to
recycle. Thus, it survives in enormous quantities to be excavated in good condition
by archaeologists. Metal and glass vessels only survive when put beyond recovery or
reuse in graves or other ritual deposits; gold and silver vessels are found even less
frequently, usually in ancient hoards that were not recovered after being concealed
in times of trouble. Even finds from Pompeii and Herculaneum are not fully
representative, because many of the inhabitants had time to remove their most
precious possessions. Archaeologists probably underestimate the use of precious
metals, despite the occasional discovery of extraordinary treasures such as the 165
fourth-century B.C. silver and silver-gilt table vessels found at Rogozen in Bulgaria
in 1985 and 1986 (Cook 1989; Vickers and Gill 1994: 52-53). If the Mildenhall treasure
found in Suffolk in the 1940s is a reliable indicator of the kinds of vessels used in late
Romano-British households for serving, eating and drinking (Kent and Painter
1977: 33-39), it is not hard to envisage much greater wealth among the aristocratic
landowners of early Roman Italy. Vickers, Impey, and Allan (1986) illustrated the
consistent way in which pottery followed the forms and surface finishes of metal
and glass in many different periods and areas in the Old World. This kind of
perspective is essential if ceramic technology is to be evaluated meaningfully.
Furthermore, because Greek figured vases have been the subject of art-his-
torical study, less attention has been paid to the everyday undecorated pottery. Few
excavators of Greek sites have quantified the pottery recovered in the manner that
has become commonplace among students of Roman pottery (e.g., Slane 2003). It
cannot be stressed too highly that it is the artistic quality of painted decoration,
especially realistic human figures in narrative scenes, which is the distinctive feature
of Greek pottery from the sixth to the fourth centuries B.C., not its production
technology. Furthermore, although figured vessels have dominated the academic
study of Greek pottery for more than 200 years, excavations on occupation sites
rather than cemeteries show that most pottery in everyday use either had a plain slip
or-more frequently-none at all (Sparkes and Talcott 1970). These important
points govern what is said here about Greek and Roman ceramics, for, unlike glass
production, which was transformed by the invention of blowing in the first century
B.C., there were no major technical changes from archaic Greece to the Byzantine
Empire. Thus, methods of forming and decorating pottery can be described the-
matically rather than chronologically.
PRODUCTION TECHNOLOGY
Clay Preparation and Pottery Forming

Production technology is governed by the physical properties and firing behavior of
clays (Rice 1987: 31-112). In particular the size and shape of its constituent particles,
which vary markedly according to depositional environments, determine a clay's
relationship with water and therefore its plasticity and susceptibility to shrinkage
(Rice 1987: 58). Potters routinely manipulate these properties by mixing different
clays or by including other matter as temper: "Potters assess a clay's plasticity
according to its 'working range,' 'workability,' or 'plastic limits,' characteristics that
can be determined adequately, though non-quantitatively, by a skilled potter
through 'feel' and experience" (Rice 1987: 61). Broken pottery is sometimes crushed
and reincorporated into clay as temper (Britain: Freestone and Gaimster 1997: 59;
Turkey: Poblome 2000). The preparation and modification of clays begins with the
drying of excavated clay, which is then broken down into small lumps before the re-
addition of water, kneading or treading, and the mixing in of temper. Such clay can
be used to make utilitarian pottery, which is often made by hand from coils or on a
slow turntable. No Roman turntables survive, but this forming method can be
deduced from examination of vessels such as black burnished ware cooking pots
that continue Iron Age traditions throughout Roman-period Britain (Peacock 1982:
55; Greene 1992: 49, fig. 21). Ethnographic observations reveal a variety of tools used
in this process, including heavy bent sticks, used to beat the dry clay, and turntables
made from wood (Greene 1992: 54, fig. 26; Ionas 2000: 138). This kind of processing
has been observed in modern communities in a mode of production categorized by
Peacock as "household production" or more typically "household industry" (1982:
13-17). Peacock's attempt "to impose a conceptual framework upon a situation that
in practice may be almost infinitely variable" (1982: 8) was designed to define
"modes of production" (Greene 2005: 40-42). Analogies suggest that for many
Greek and Roman utilitarian wares, this relatively local scale of production would
either be the main source of subsistence or supplement other work such as seasonal
farming (Peacock 1982: 9).
Once pounded into smaller fragments, dry clay will absorb water readily and
can be mixed to form a slurry that may be sieved or allowed to settle in levigation
tanks. The heaviest particles sink first, leaving lighter particles to settle into fine clay
above them. Once the water has evaporated, clay of different consistencies can be
extracted according to the potter's needs. Levigation tanks facilitate the production
of fine-grained fabrics and tend to be found at higher levels of production, for
example at nucleated workshop sites, together with heavy turntables or wheels and
kilns (Peacock 1982: 9, 25, 54). Examples were found on a fifth-century B.C. pro-
duction site at Phari, on the Greek island ofThasos (Blonde et al. 1992: 17) and at the
Roman terra sigillata production site at Rheinzabern in Germany (Reutti 1991;
Peacock 1982: 54, fig. 20 ). Terra sigillata and other fine wares, including Late Roman
red-slip wares, are well levigated and were produced on a large scale on sites of this
kind. Greek potters' wheels are depicted on painted vases (Boardman 2001: 143, fig.
167), and Peacock has collated evidence for wheels (including possible socket
stones) from Oxford, Rheinzabern, and Argonne, and flywheels made from wood
and stone from Speicher and Rheinzabern (1982: 55-57).
Ethnoarchaeological research uses analyses of modern potters to find analogies
for processes used in the past, to identify variables in manufacturing technology,
and to explore the dynamics of production (Stark 2003: 203; Rice 1987: 114). Such
studies help us to evaluate technological decisions made by potters when selecting
and preparing clays (Stark 2003: 211-12). Technology is only one of many topics
addressed by ethnoarchaeologists; others include taxonomy, vessel function, lon-
gevity, recycling and disposal, division of labor, learning, style, ethnicity, distri-
bution, and technological and stylistic change (Rice 1987: 202). A related approach
"involves the experimental manipulation of raw materials and tools of a particular
archaeological context to try to replicate the conditions of their use or production
or both" (Rice 1987: 114; Coles 1973). Ancient visual representations of workshops,
of which there are many, provide invaluable information (figures 2.8-2.10).
Scientific investigation of the production of ceramics has concentrated on
provenance studies, which aim to identify sources of production by comparing
vessel fabrics with clays from sources near their suspected place of manufacture
(Rice 1987: 413-26). Positive geoarchaeological matching is complicated by mixing
and modification of clay by potters during processing and manufacture (Rice 1987:
421-24). The comparison of source clays with final products, however, helps us to
understand these technological processes and to recognize the considerable tech-
nological expertise that potters used in manipulating clay properties to produce
wares with desired characteristics. Whitbread's studies at Corinth have demon-
strated the complex processes used by potters in that city and are a model for
approaches elsewhere. He stresses the need for well-excavated deposits and the
careful collection of samples during excavation (2003: 1):
Ceramic technology, in its own right, should be an integral component of pottery
production studies at centers such as Corinth. For example, variety in ceramic
compositions may reflect workshop distribution, restricted access to raw mate-
rials, or different materials preferences for specific types of ceramics, such as
tablewares, storage vessels, tiles, waterjars, or cooking pots. All aspects of pottery
manufacture, from selection and processing of clays and temper, through forming
and drying, to firing, are intimately dependent upon the nature of the raw ma-
terials. As a result, few conclusions about ceramic technology can be reliably
established without information concerning the raw materials that were used.
Compositional analysis of pottery fabrics can also increase our understanding

of the ancient economy by examining patterns of continuity and change in the
trading of ancient ceramics (Rautman et al. 1999: 377; Rautman 1995: 331; Blonde
and Picon 2000).
Firing
Firing methods were closely related to choices of raw materials so that vessels could
survive changes in their physical properties, including loss of volatiles, weight loss,
and changes in chemical and mineral structure (Rice 1987: 86--94). The most com-
mon inclusions in clay-whether occurring naturally or added by the potter-are
quartz, feldspars, and lime, and they are "important during firing in modifying the
expansion, shrinkage, and microstructure of a clay body" (Rice 1987: 93). Firing
technology varied across the Greek and Roman world, with all types being used from
archaic Greece to Late Roman times. In Britain the processes included bonfires used
to fire black burnished wares in Dorset (Hearne and Smith 1992), simple updraft kilns
(Swan 1984), and more complex structures used for firing tiles (e.g., at the legionary
pottery works at Holt: Grimes 1930: fig. 19). Cuomo di Caprio (1992) has outlined a
broad typology of kiln types, but regional variations are common (e.g., Gaul: Peacock
1982: 68). A number of Greek votive plaques represent kilns (e.g., 2.9-2.10).
Vessels were stacked carefully in kilns to make efficient use of space and fuel
during firing. Occasionally vertical piles of open plates or bowls became fused
together if the firing temperature became too high (Johns 1977); normally, grains of
sand on their foot rings prevented one pot from adhering to the vessel below.
Differences in the external and internal coloring of fired closed vessels also show
that they were stacked closely together, affecting the oxygen content of the at-
mosphere that reached the pot. Skilled clay preparation and kiln control allowed
Greek workshops to produce the glossy surface of plain black-glazed vessels and the
color contrasts of black- and red-figured vases:
In an initial oxidising atmosphere, body and slip fired red, due to the formation of
hematite, ferric oxide. A reducing atmosphere was then induced by closing the
vents of the kiln and adding wood to the fire, which converted the iron oxide of
the slip to magnetite, which is black. At the peak of the firing, in the range 850-
10000C, the potash content of the slip, coupled with its very fine grain size, caused
it to fuse into a dense black glossy layer which was essentially impervious to
oxygen. As the wood burnt down, the vents were opened and the kiln atmosphere
became oxidising once more. Due to its dense, fused condition, the slip remained
black. However, the coarse clay of the body retained an open, porous structure
throughout the firing, which allowed oxygen to diffuse in, converting the reduced
iron oxide to the oxidised form, hematite. Depending upon the precise compo-
sition of the clay and the firing temperature, the body then gave a red or buff
background to the black slip. (D. Williams 1997: 89; cf. Noble 1988)
Numerous pottery manufacturing sites have been excavated. The Greek kiln at
Phari (on the west coast of Thasos), dated to the archaic period by vessels datable
between 525-480 B.C., is one example (Blonde et al. 1992), and it has the same basic
form as Romano-British kilns (Swan 1984). The rural potter's workshop at Phari was
ideally located in terms of raw materials in a wooded coastal region, on a rich clay
source, with its own water supply and near a safe harbor. Different clays were ex-
tracted and processed for making specific wares (Blonde et al. 1992: 19). The site
included clay preparation areas with two settling tanks, for preparing and refining the
clay, and two kilns. The rectangular updraft kilns were well constructed in large
undressed stones with pear-shaped internal chambers and a central pillar to support a
raised floor. Complex terra sigillata kilns separated the smoke from the pottery by
sending it through ceramic flues (figures 20.1, 20.2), an arrangement that allowed the
atmosphere of the kiln to be controlled independently of the fire to ensure that vessels
were fired with an even oxidized red slip. The figures show the Roman pottery kiln
excavated at the east Gaulish terra sigillata manufacturing center at Rheinzabem,
Germany. The fire was lit in an arched flue more than 3 m long. The firing chamber,
almost 3 m in diameter, contains pipes to protect terra sigillata vessels from direct
heat and smoke from the fire; kilns for ordinary pottery dispensed with this feature.
Broadly similar kilns can be found throughout the Greek and Roman periods.
The Range of Ceramic Products

Most of the Greek and Roman world and its neighboring areas had long histories
of ceramic production stretching back into prehistory, and few areas lacked pottery
made on the wheel and fired in kilns. In the first century, Pliny (HN 35-160) states
Figure 20.1. Roman pottery kiln for terra sigillata, Rheinzabern (Germany). (Photograph
courtesy of F. Reutti.)
that "The greater part of the human race uses pottery vessels." Parts of the lower
Rhineland and Britain had not adopted the potter's wheel before the Roman
conquest, but in both cases its introduction came through an expansion of nearby
indigenous wheel-thrown pottery production rather than through the establish-
ment of Mediterranean industries (Greene 1993). Roman military potters did in-
troduce essential new forms such as flagons and mortaria (large bowls with gritted
interior surfaces used in food preparation), but traditional pottery making was not
displaced. The most widely distributed cooking pottery in Roman Britain came
from a region on the southern coast where vessels had been made by hand and fired
in bonfires since the Iron Age (black burnished ware: Tyers 1996: 192-96). Simpler
handmade pottery coexisted with "Roman" wares in northern Gaul for centu-
ries (De Clerq 2005). Even the peripheral Roman province of Britain provided
sufficient markets, consumers, and transport to allow local production to be ex-
panded out of all recognition without changing the technology of manufac-
ture. The distinctive handmade cooking pots from Dorset, whose form ex-
tended back into prehistory, were so successful that potters around the Thames
estuary imitated them using wheel-throwing and kiln technology (Tyers 1996: 182,
186-87).
Archaeologists typically classify ceramics on functional rather than techno-
logical grounds. Vessels may be considered to belong to the same class if they are of
the same general form and are used for the same purpose "irrespective of date,
fabric, and details of form" (Webster 1976: 6). Specific fabrics, forms, methods of
decoration, and technology vary in time and space, but products throughout the
Greek and Roman world employed the full spectrum of technological processes.
0 Sm
Figure 20.2. Roman pottery kiln for terra sigillata, Rheinzabern (Germany), reconstruc-
tion. (Reutti 1991: 23.)
The range includes wheel-made, handmade, and mold-made vessels, as well as

other objects made from clay.
Vessels were produced in both open and closed forms, and those found in
household contexts include thin-walled and slipped wares for the table. Tablewares
such as black glazed ware and terra sigillata were wheel-made and traded over long
distances, as were some plain and painted wares; plain wheel-made vessels were
also used for food storage and preparation. Pots with coarser fabrics-both wheel-
made and handmade-were used in the kitchen for cooking, and some traveled a
considerable distance from their place of manufacture. Other clay products in-
cluded utilitarian items from lamps to beehives.
A greater volume of clay probably went into the production of bricks, tiles,
pipes, and other products than into pottery vessels (Delaine 1995; Freestone and
Gaimster 1997: 158-63; see also chapters 9 and 10). Pliny commented on the
ubiquity of ceramic products in first-century Rome (HN 35.159 ): "Nor are we sated
by the presence everywhere of pottery products, with jars devised to hold wine,
pipes for water, flue ducts for baths, tiles for roofs, fired bricks for walls and
foundations . ... " The conventional image of classical architecture is one of stone
buildings, especially Greek temples made from marble. By the Roman imperial
period, however, even prestigious public building projects such as baths or palaces
were constructed substantially from brick and concrete, which suited their enor-
mous size and elaborate architectural form (Delaine 1997; see also chapter 10). The
combination of brick and concrete allowed the construction of huge vaults for the
first time, and some incorporated ceramic pipes and even amphoras to make them
lighter (Stortz 1993; Lancaster 2005: 68-85) . Brick walls could easily be covered with
a thin decorative cladding of stone, mosaic, or painted plaster. Even when
buildings were constructed in masonry, roofs were made from clay tiles unless a
convenient source of stone slates was available.
A system of ceramic roofing elements developed in the seventh century B.C. in
Greece-notably at Corinth-and was rapidly taken up in Italy and continued into
the Late Roman period (Winter 1993). Roofing systems comprised rectangular pan
tiles with raised edges on their longer sides, and narrow cover tiles with a semi-
circular cross-section, which fitted neatly over the raised edges of two adjacent pan
tiles. Cutout notches and tapered forms allowed the construction of roofs with
overlaps that provided effective weatherproofing (Winter 1993: frontispiece). Greek
buildings had additional tiles to project rainwater away from their walls, while the
visible edges of roofs were decorated with elaborate panels and antefixes that were
sometimes painted as well as molded in relief (Winter 1993: 28, fig. 4); simpler
versions were still made in Roman times. These tegulae were formed in wooden
frames, and careful control over dimensions was necessary to ensure a good fit. The
extensive use of manufacturers' stamps and other control marks (Adam 1994: 64)
would have assisted the selection of tiles of the right form and size for use on an
individual roof. In the Roman period, brickyards around Rome produced thou-
sands of stamped rectangular bricks which give insights into the organization of
the industry (Helen 1975; Steinby 1993). Tiles of more specialized shapes, as well as
pipes, were required in the construction of bathhouses and water supply systems
(Bouet 1999). Large-scale building projects continued to require enormous num-
bers of bricks with makers' stamps well into the Byzantine Empire (Bardill 2004).
Fired earthenware (terracotta) was also a major medium for the production of
sculpture, especially in Greek and Etruscan contexts (Higgins 1954; Merker 2003).
At Corinth, 24,000 fragments of figurines were recovered by excavations in the
sanctuary of Demeter and Kore (Merker 2000: 1):
This number encompasses a great range in scale, from miniatures to statuettes

originally measuring more than ca. 50 cm in height; a corresponding range of
different types; and a surprising differential in technique and concept, from
plaques stamped out in large numbers, which are pale reflections of their original
models, to pieces that, although moldmade, nevertheless preserve something of
the freshness of new creations and probably are close to their archetypes. Such
diversity is a key element in the Corinthian coroplastic industry as a whole.
The makers of terracotta figurines may have worked alongside sculptors

casting bronze statues, since this craft also required knowledge of the working
properties of clay used in the creation of molds and cores (see chapter 16). Figures,
antefixes, or plaques made from plastic or molded clay could of course be deco-
rated with painted slips and even gilding to produce finished works that might be
difficult to distinguish from stone or wooden sculptures in a dimly lit temple.
Although less common than in Greek times, the production of figurines remained
part of the Roman repertoire-for example, molded white pipe-day representa-
tions of deities made in central France, probably for display in household shrines
or for making dedications at temples (Bemont et al. 1993; Rouvier-Jeanlin 1972).
While wheel-thrown open oil-lamps had a long history before classical times,
closed lamps with a projecting nozzle made in two-piece molds were an increas-
ingly common product of Hellenistic and Roman pottery workshops (Bailey 1979,
1980; Bailey et al. 1988; Freestone and Gaimster 1997: 164-69). They shared many of
the decorative features of tablewares, such as relief-molded decoration and glossy
slips; indeed, many were made in the same workshops as tablewares. A workshop
was found within the walls of Pompeii containing two-piece molds, matrices for
making molds, and kilns for firing finished lamps (Cerulli Irelli 1977). Like pottery
fine wares, clay lamps clearly imitated metal prototypes that only survive in small
numbers; metal lamps could even be used as matrices for making molds. Lamps
frequently bear a name under the base, perhaps that of the manufacturer of the
mold or the owner of the workshop in which they were made (Harris 1980). In
addition to providing lighting, lamps commonly bore religious ornament that
suggests ritual functions: Late Roman lamps with Christian crosses, for example,
were produced in enormous quantities. Small containers such as unguentaria and
pilgrim flasks used for carrying perfume or oil also had ritual functions.
Agricultural production of wine and oil not only created long-distance trade
networks but also generated an enormous demand for large ceramic containers;
estates with good clay could organize ceramic production directly or lease its
exploitation to tile makers and potters (Cockle 1981). Enormous spherical dolia
(frequently set into the ground) were used in the storage or fermentation of wine,
which would then be decanted into amphoras for transport and distribution.
Amphoras had been adopted in Greece from the eastern Mediterranean seaboard in
the eighth century B.C. (Whitbread 1995: 3-4), and retained the same basic form
(narrow neck with two handles, globular or cylindrical body, narrow or pointed
base) right through to the Byzantine period (Peacock and Williams 1986). Dolia and
amphoras were wheel-thrown in stages, with intermittent periods of drying to allow
completed sections to support the weight of further additions, and then fired in
large kilns; identical techniques have been documented by ethnographers on Crete
(Hampe and Winter 1962: pls. 36-41). Amphoras were made for transporting both
liquid and dry foodstuffs. The scale of amphora production is reflected in the
Monte Testaccio beside the Tiber in Rome-an artificial hill of amphora shards,
mostly of Spanish origin, representing as many as 53 million amphoras (Blazquez
Martinez et al. 1994: 13). Heavy mortaria, with internally gritted surfaces, thick
flanges, and a pouring spout, were used for grinding and mixing foodstuffs in the
kitchen (Hayes 1997: 80-82). They were distributed over particularly wide areas,
and they frequently have makers' names stamped on the rim that indicate pro-
duction in specialist workshops (like bricks and amphoras). Neither amphoras nor
mortaria show significant technical differences over time. The lack of standard
amphoras of optimum form is consistent with the diversity of their geographical
sources and variety of contents. A merchant observing a ship arriving at Marseille in
the first century A.D. could see at a glance whether the cargo included Spanish olive
oil in the globular Dressel 20 or Italian wine in the cylindrical Dressel 2-4 (Peacock
and Williams 1986: 105, fig. 39, 139, fig. 67). A large sample of amphoras from the
Byzantine shipwreck at Yassi Ada, Turkey (early seventh century A.D.), shows that
their sizes related to Byzantine litra, and that their volumes were more precisely
standardized than those of Roman and Hellenistic amphoras (Van Alfen 1996: 189).
Name-stamps or distinctive regional pottery forms occasionally allow us to
detect the establishment of branch workshops or the movement of potters (Hartley
1977), but such activity is of greater significance to economics than to technology.
"Technology transfer" only took place when a new form or manufacturing method
was introduced into a peripheral area, such as the occasional establishment of Greek
potters in Italy in the seventh to fifth centuries B.C. (Boardman 2001: 74-78) or the
opening of a terra sigillata workshop in Colchester in the second century A.D. (Hull
1963). There is no parallel to the westward diffusion of new glassblowing technology
from the East Mediterranean after 50 B.C, although it may have stimulated the
production of "thin-walled" cups and beakers, which proliferated in the first
century A.D. (Hayes 1997: 67-71; Carandini 1977; Greene 1979)-presumably in
response to the new popularity of glass drinking vessels.
Surface Finishes and Decoration

While no new techniques were invented in the Greek or Roman periods, styles and
methods of finishing and decorating pottery varied in their popularity over time.
Throughout this period of study most forms of pottery were made simply in
undecorated hand-finished clay, frequently in quite coarse fabrics. Much could be
done to produce a smooth surface by burnishing the unfired "leather-hard" pot,
and decorative effects could be achieved by burnishing linear designs onto matt
zones (e.g., black burnished cooking pots from Dorset: Tyers 1996: 183, fig. 227) or
by spiral-burnishing African Red Slip Ware Form 109 (Hayes 1972: 172). While still
soft, pots could be decorated by hand or with simple tools-incising designs,
impressing the surface with decorative stamps (found especially in the Late Roman
period on African and Phocaean red slip wares-Hayes 1972: 217 ff., 346 ff.) or
cutting or gouging to imitate cut glass (Charleston 1955: pls. 14-15). Surfaces of
vessels could be intentionally roughened by the addition of coarse particles to
produce a finish known as roughcasting that made them easier to grip when wet
(Greene 1979: 33-34, figs. 13-14). Ridges are frequently found on the exterior surface
of vessels, made by the potter's fingers or with a tool; such decoration is common
on Late Roman cooking pots, plain wares and on amphoras (Hayes 1997: pl. 1).
The next level of complexity involved the addition of further clay to produce
relief decoration. Liquid clay could be trailed or dropped onto the surface by hand
(Hayes 1997: pl. 4; Greene 1979: 69, fig. 30), or handmade or molded motifs could be
stuck onto the surface (Hayes 1997: 13, pl. I; for molds from a production site, see
Bergamini 2005: 75-76, figs. 2-3). Relief decoration could be formed by pinching
and impressing bands of clay, whether part of the vessel or applied to the surface, to
create images such as faces (Charleston 1955: pl. 92). "Pie-crust" or "rope" deco-
ration is common on pithoi, and also occurs on Late Roman storage vessels
(C. Williams 1989: 53-56). Decoration in relief could be produced more reliably and
rapidly by forming vessels in a single stage in molds rather than decorating them
afterward.
Molding became more common in Greek ceramics in the late third century
B.C., when so-called Megarian bowls were made in hemispherical relief molds that
would have been almost identical to those used in the manufacture of similar vessels
in metal or glass (Hayes 1997: 40, pl. 13; Rotroff 1982, 2006). Relief-molding in
wheel-mounted molds remained standard in the production of decorated terra
sigillata tableware into Late Roman times. Wheel-made clay molds were decorated
by impressing stamps (poirnj:oins) into their interior surfaces to form repeating
designs. Bowls could then be thrown inside the wheel-mounted molds, and their
rims finished by hand. As the bowl dried and shrank, it could be withdrawn from
the mold, which would then be used again to produce another vessel with identical
decoration (Johns 1971: pls. 14b, 15).
Derivatives of terra sigillata made in North Africa and elsewhere around the
Mediterranean from the late first century A.D. (Red Slipped Ware) were no longer
made in relief molds. They were decorated by means of impressed stamps, roller-
stamping (or "rouletting"-rolling a small wheel bearing motifs around a vessel to
impress repeated decoration into the clay) or applied molded motifs (Hayes 1997:
62-63, pls. 21-22; Mackensen 1993). This was not a technical departure-stamps and
applied reliefs had always been used on the plainer Greek and Roman forms
manufactured alongside figure-painted and molded vessels (Hayes 1997: 38, pl. 11;
Sparkes and Talcott 1970: 23). Some large rectangular plates (Form 56) were press-
molded in the African Red Slip Ware industry in Tunisia (Hayes 1972: 293).
Many of the decorative techniques described above were combined with a
slip-a liquid suspension of fine clay particles which can be added to an unfired pot
by painting it onto all or part of the surface or by dipping the pot into it. The
complex technology of slips has been investigated through a combination of ex-
perimental and scientific research for many decades (Peacock 1982: 63). Particularly
fine slips containing suitable minerals (notably illite) can be fired so that the clay
particles fuse to produce very glossy surfaces of the best Greek black glazed ware and
Roman terra sigillata, usefully reducing the permeability of these tablewares. Slips
could be thin and easily worn, and might simply be used to disguise the actual fabric
color of a vessel. Slips containing minerals different from those of the body fabric
were often used as decoration on otherwise plain vessels, while the use of more than
one kind of slip allowed polychrome effects (Hayes 1997: pls. 3-4). In the early
Roman period, delicate painted decoration was found on Nabataean thin-walled
tablewares from Jordan (Schmid 2003), and painting-frequently using more than
one color-continued as a technique into Late Roman times across the empire as
far afield as fourth-century Britain and Coptic North Africa (Bourriau 1981: 92-94).
Greek pottery vessels suitable for use in formal dining (the symposion) usually
displayed a characteristically glossy black slip; Athenian painted pottery produced
from the sixth to the fourth centuries B.C. had a particularly high reputation
(Figure 20.3; Athenaeus, Deip. 1.28c). The column krater (figure 20.3, back right) has
Figure 20.3. Selection of sixth- to fourth-century B.C. Attic pottery vessels. (Photograph
courtesy of Shefton Museum, University of Newcastle upon Tyne.)
black-figure decoration with incised details, while the large cup (front, second from
left) has a red-figured central panel. All were wheel-thrown in fine earthenware,
with handles formed separately and added after the vessels had dried. The small jug
(front left) has a molded face and applied dots of clay forming a textured area
representing hair. One cup (front right) represents the much larger quantity of
Greek tableware that did not bear additional painted decoration. Its black "glaze" is
technically identical to the slips of Roman terra sigillata tableware fired in an
oxidising kiln atmosphere to produce a red rather than black finish.
Roman ceramics also made use of a wide variety of methods of decoration and
surface finish (figure 20.4) . The techniques include colored slips (figure 20.4, back
center and right), painted slip decoration (front left), molding (bottom center and
right) and modeling by hand (flagon with face mask) . The lamp (front center) was
made by pressing leather-hard clay sheets into a two-piece mold. The four beakers
are typical of Roman fine wares, and their decoration illustrates "barbotine" trailed
liquid slip (right), "roughcast" clay particles (center) , indentation (back center),
and "rouletting" (left).
The greatest variety of decorative effects achievable using slips and painting can
be observed on Greek red- and black-figured pottery. It developed from earlier
styles traceable back through Corinthian and Geometric painting to Mycenaean
and Minoan pottery of the second millennium B.C., which drew on a longer
ancestry of pottery decoration in Egypt and other Near Eastern civilizations
(Boardman 2001: 13, figs. 1-3; Freestone and Gaimster 1997: 41, fig. 4). Fine details of
Figure 20.4. Vessels and sherds from the Roman fort and town at Corbridge, North-
umberland. (Photogarph courtesy of Corbridge Museum and English Heritage.)
figures and draperies on Greek black-figure vases could be incised into areas of dark
slip to reveal the contrasting paler clay beneath the surface (Boardman 2001: 44-78).
The transition to the red-figure style allowed greater delicacy, as details could be
added in painted lines of dark slip over a red background; painting with a fine brush
obviously permitted greater subtlety than incision (Boardman 2001: 79-106). Ad-
ditional effects enriched painted decoration even further:
White was used extensively for details of costume like wreaths and especially
for the flesh of female figures and Erotes. Details within these large white fields
were added in a golden brown tone created with dilute glaze. The wings of Erotes
were frequently gilded. Inlays in metal vessels were suggested by beads or lines
of clay that were painted white to evoke silver or tinted yellow with dilute glaze or
gilded to represent gold inlays. Contemporary Attic white-ground vases, on
the other hand, may have imitated wall-paintings; figures were outlined in thin
black glaze, and yellows and browns were created with diluted glaze. Reds, yellows,
blues, and greens were also produced with added earth and vegetable colors.
(Padgett et al. 1993: 18)
White slip motifs contrasted attractively with the surface of Greek and Hel-
lenistic black-glazed pottery. Similar decoration can be found in the Roman period
on a number of categories of table and drinking vessels. Wheel-thrown drinking
vessels made at Trier in Germany in the second to third centuries A.D. were
decorated with white, yellow, and red painted designs over a black lustrous slip
(Kunzl 1997). They are technically identical to Hellenistic West Slope pottery from
Athens or Gnathia ware from southern Italy (Rotroff 1991; Green 1976).
A clear contrast between Greek and Roman ceramics and those of later cen-
turies is the rarity of fully impermeable glazed finishes rather than fused slips
(Greene 2007). Vitreous glazes had been known in Egypt and Mesopotamia for
many centuries (Freestone and Gaimster 1997: 74-89, 104-9). Lead glazes were used
sporadically on Hellenistic Greek and Roman pottery (Hayes 1997: pl. V; Hochuli-
Gysel 1977), including provincial products made as far away as the Danube and
Rhine provinces and in Britain (Cvjeticanin 2001; Hopken 2003; Arthur 1978). A
distinction must be drawn between the alkaline glazes of western Asia and the lead
glazes used in Hellenistic and Roman Asia Minor and the western provinces of the
Roman Empire (Romer-Strehl et al. 2005: 216; Peacock 1982: 65). Glazed vessels
were often fired in "saggars" to protect their surfaces in the kiln (Peacock 1982: 64,
fig. 27). Glazed pottery became increasingly common in Late Roman to early me-
dieval times (Paroli 1992) and dominated Byzantine ceramics (and their early Is-
lamic equivalent) by the eleventh century (Vroom 2003: 58).
CONSUMPTION
The intertwining of history, culture, economics, production, and consumption

is particularly clearly demonstrated by Greek and Roman pottery. State building
programs demanded enormous quantities of ceramic tiles and other architec-
tural fittings, and the industries that produced them also supplied building ma-
terials to individuals constructing town houses or country villas (Martins 2005).
Transport infrastructure and security allowed producers to risk sending goods over
long distances, while coins-invented to perform high-value state functions-
rapidly became a medium for exchange when base metal denominations prolif-
erated. Flows of heavy goods, whether agricultural or industrial, provided trans-
port opportunities for items such as pottery whose low value would not have
justified trade on their own (Greene 1986: 162-64). Pliny (HN 35.159-63) comments
on the import and export of particularly valued ceramic wares around the Med-
iterranean.
The expansion of Greek city states and the city of Rome into territorial empires
spread styles of living and consuming-and the material wherewithal for cooking,
eating, and drinking-from their original urban centers to the fringes of political
control and beyond (Bats 1988). Greek and Roman cuisine and dining conventions
extended the range of vessel forms imported into, and subsequently manufactured
in, newly Hellenized and Romanized regions. In a similar manner, the popularity of
drinking coffee and chocolate-and especially tea-in early modern Europe re-
quired specialized drinking and serving vessels that were increasingly supplied by
local industries rather than being imported (Snodin and Styles 2001: 129, fig. 18, 252-
53). Diffusion of the use, and subsequently the production, of decorated tablewares
was closely related to trade in slaves, wine, oil, and grain.
One vessel form that epitomizes the importance of the diffusion of con-
sumption practices is the plate. In Britain, plates make their first appearance in the
late Iron Age, when examples were imported along with Italian and Gaulish wine in
the first century B.C. (Freestone and Gaimster 1997: 57-59, figs. 2 and 4). After
having been a consistent part of the British potters' output for more than 400 years,
plates disappeared in the fifth century A.D., when Roman rule ended and the
monetized economy collapsed. The only exceptions were rare imports into aris-
tocratic and Christian centers around the Irish Sea that maintained connections
with the Mediterranean. Wheel-thrown pottery was not made in Britain again until
the late Saxon period, and plates did not become a regular feature of either imports
or British production until the Tudor period, when the diversity of trade and a
monetary economy-combined with Renaissance eating habits-regained a level of
complexity approaching that of the Roman period. Even then, plates did not be-
come common until fundamental changes in ceramic production technology oc-
curred in the eighteenth century, when earthenware and porcelain approaching the
quality of Islamic and Chinese imports began to be manufactured.
The mass production, diversity, and wide diffusion of Greek and Roman ce-
ramics coexisted with technological stability (cf. chapter 32). It took a world econ-
omy, a consumer society, and an Enlightenment approach to the implementation
of scientific techniques in industrial production to move ceramics into a different
orbit (Berg 2004; Snodin and Styles 2001: 281-307).
REFERENCES
Adam, J.-P. 1994. Roman building: Materials and techniques. London: Routledge.
Annecchino, M., et al. (eds.) 1977. L'instrumentum domesticum di Ercolano e Pompei nella
prima eta imperiale. Rome: Giorgio Bretschneider.
Arthur, P.R. 1978. "The lead-glazed wares of Roman Britain," in P.R. Arthur and
G.D. Marsh (eds.), Early fine wares in Roman Britain. British Archaeological Reports
57. Oxford: BAR, 293-356.
Bailey, D. M. 1979. A catalogue of the lamps in the British Museum 1: Greek, Hellenistic and
early Roman pottery lamps. London: British Museum.
Bailey, D. M. 1980. A catalogue of the lamps in the British Museum 2: Roman lamps made in
Italy. London: British Museum.
Bailey, D. M., S. Bird, M. J. Hughes 1988. A catalogue of the lamps in the British Museum 3:
Roman provincial lamps. London: British Museum.
Bardill, J. 2004. Brickstamps of Constantinople. 2 vols. Oxford: Oxford University Press.
Bats, M. 1988. Vaisselle et alimentation a Olbia de Provence (v 350-v 50 av J-C): Modeles

culturels et categories ceramiques. Revue Archeologique de Narbonnaise Suppl. 18. Paris:
CNRS.
Beazley, J. 1951. The development ofAttic black-figure. Berkeley: University of California Press.
Bemont, C., et al. (eds.) 1993. Les figurines en terre cuite galloromaines. Documents d'Ar-
cheologie Fran1yaise 38. Paris: Maison des Sciences de l'Homme.
Berg, M. 2004. "In pursuit of luxury: Global history and British consumer goods in the
eighteenth century," Past and Present 182: 85-142.
Bergamini, M. 2005. "Matrici per terra sigillata da Scoppieto: Studio preliminare dei motivi
iconografici," Acta Rei Cretariae Romanae Fautorum 39: 71-79.
Bintliff, J. 2002. "Settlement pattern analysis and demographic modeling," in P. Attema et
al. (eds.), New developments in Italian landscape archaeology. British Archaeological
Reports, Intl. Series S109. Oxford: BAR, 28-35.
Blazquez Martinez, J.M., J. Remesal Rodriguez, and E. Rodriguez Almeida 1994. Ex-
cavaciones arqueologicas en el monte Testaccio (Roma). Madrid: Ministerio de Cultura.
Blonde, F., and J. Y. Perreault (eds.) 1992. Les ateliers de potiers dans le monde Gree aux
epoques Geometrique, Archai'que et Classique. Bulletin de Correspondance Hellenique
Suppl. 23. Athens: Ecole Fran1yaise d'Athenes.
a
Blonde, F., J. Y. Perreault, and C. Peristeri 1992. "Un atelier de potier archa'ique Phari
(Thasos)," in Blonde and Perreault 1992: 11-40.
Blonde, F., and M. Picon 2000. "Artisanat et histoire des techniques: le cas des cer-
amiques," in F. Blonde and A. Muller (eds.), L'artisanat en Grece ancienne: Les pro-
ductions, les diffusions: Actes du colloque de Lyon, 10-11 decembre 1998. Lille: Universite
Charles-de-Gaulle, 1-26.
Boardman, J. 2001. The history of Greek vases: Potters, painters and pictures. London:
Thames and Hudson.
Bouet, A. 1999. Les materiaux de construction en terre cuite dans les thermes de la Gaule
Narbonnaise. Paris: De Boccard.
Bourriau, J. 1981. Umm El-Ga'ab pottery from the Nile Valley before the Arab Conquest.
Carandini, A. 1977. "La ceramica a pareti sottili di Pompei e del Museo Nazionale di
Napoli," in M. Annechino et al. 1977: 25-31; pls. VIII-XIX.
Cerulli Irelli, G. 1977. "Una officina di lucerne fittili a Pompei," in M. Annechino et al. 1977:
53-72.
Charleston, R. J. 1955. Roman pottery. London: Faber.
Charleston, R. J. (ed.) 1968. World ceramics: An illustrated history. Feltham: Hamlyn.
Cockle, H. 1981. "Pottery manufacture in Roman Egypt: A new papyrus," Journal ofRoman
Studies 71: 87-97.
Cohen, D. H., and C. Hess 1993. Looking at European ceramics: A guide to technical terms.
Malibu, CA: J. Paul Getty Museum.
Coles, J. 1973. Archaeology by experiment. London: Hutchinson.
Cook, B. F. (ed.) 1989. The Rogozen treasure. London: British Museum.
Cook, R. M., and R. J. Charleston 1979. Masterpieces of Western and Near Eastern ceramics.
Vol. 2, Greek and Roman pottery. Tokyo: Kodansha.
Cunliffe, B. W. (ed.) 1998. The Oxford illustrated prehistory of Europe. Oxford: Oxford
University Press.
Cuomo di Caprio, N. 1992. "Les ateliers de potiers en Grande Grece: Quelques aspects
techniques," in Blonde and Perreault 1992: 69-88.
Cvjeticanin, T. 2001 Glazed pottery from Upper Moesia. Beograd: Narodni Muzej.
De Clerq, W. 2005. "Shaped by tradition: On the persistence of hand-made pottery tra-
ditions in northern Gaul, c. 100 B.C.-300 A.D.," Acta Rei Cretariae Romanae Fautorum
39: 201-8.
Delaine, J. 1995. "The supply of building materials to the city of Rome," in N. Christie
(ed.), Settlement and economy in Italy, 1500 B.C. to A.D. 1500. Oxford: Oxbow, 555-62.
large-scale building projects in imperial Rome. Journal of Roman Archaeology Suppl. 25.
Dore, J., and K. Greene, (eds.) 1977. Roman pottery studies in Britain and beyond. British
Freestone, I., and D. Gaimster, (eds.) 1997. Pottery in the making: World ceramic traditions.
Fulford, M. G. 1975. New Forest Roman pottery manufacture and distribution, with a corpus
of the pottery types. British Archaeological Reports. 17. Oxford: BAR.
Green, J. R. 1976. Gnathia pottery in the Akademisches Kunstmuseum Bonn. Mainz:
von Zabern.
Greene, K. 1979. Report on the excavations at Usk, 1965-1976: The pre-Flavian fine wares.
Cardiff: University of Wales Press.
Greene, K. 1986. The archaeology of the Roman economy. Berkeley: University of California
Press.
Greene, K. 1992. Roman pottery. London: British Museum Press.
Greene, K. 1993. "Part 1; the fortress coarse ware," in W. H. Manning (ed.), Report on the ex-
cavations at Usk, 1965-1976: The Roman pottery. Cardiff: University of Wales Press,
1-124.
Greene, K. 2002. Archaeology: An introduction. 4th ed., London: Routledge.
Greene, K. 2005. "Roman pottery: Models, proxies and economic interpretation," Journal
of Roman Archaeology 18: 34-56.
Greene, K. 2007. "Late Hellenistic and early Roman invention and innovation: The case
study of lead-glazed pottery, American Journal of Archaeology.
Grimes, W. F. 1930. Holt, Denbighshire: The works-depot of the Twentieth Legion at Castle
Lyons. London: Society of Cymmrodorion.
Hampe, R., and A. Winter 1962. Bei Topfern und Topferinnen in Kreta, Messenien und
Zypern. Mainz: von Zabern.
Harris, W. V. 1980. "Roman terracotta lamps: The organization of an industry," Journal of
Roman Studies 70: 126-45.
Hartley, B. R. 1977. "Some wandering potters," in Dore and Greene 1977: 251-61.
Hayes, J. W. 1972. Late Roman pottery. London: British School at Rome.
Hayes, J. W. 1997. Handbook of Mediterranean Roman pottery. London: British Museum.
Hearne, C. M., and R. J. C. Smith 1992. "A late Iron Age settlement and Black Burnished
ware (BB1) production site at Worgret, near Wareham, Dorset (1986-7)," Proceedings
of the Dorset Natural History and Archaeology Society 113: 55-105.
Helen, T. 1975. Organisation of Roman brick production in the first and second centuries AD.
Helsinki: Suomalainen Tiedeakatemia.
Higgins, R. A. 1954. Catalogue of the terracottas in the Department of Greek and Roman
Antiquities, British Museum 1: Greek, 730-330 B.C. 2 vols. London: British Museum.
Hochuli-Gysel, A. 1977. Kleinasiatische glasierter Reliefkeramik (50 v. Chr. bis 50 n. Chr.) und
ihre oberitalischen Nachamungen. Acta Bernensia 7. Bern: Stampfli.
Hopken, C. 2003. "Die Produktion glasierter Keramik in romischen Koln," Acta Rei
Cretariae Romanae Fautorum 38: 365-66.
Hull, M. R. 1963. The Roman potters' kilns of Colchester. Society of Antiquaries of London
Research Report 21. Oxford: Society of Antiquaries of London.
Ionas, I. 2000. Traditional pottery and potters in Cyprus: The disappearance of an ancient
craft industry in the 19th and 20th centuries. Aldershot: Ashgate.
Johns, C. M. 1971. Arretine and Samian pottery. London: British Museum.
Johns, C. M. 1977. "A group of Samian wasters from Les-Martres-de-Veyre," in Dore and
Greene 1977: 235-46.
Kent, J.P. C., and K. S. Painter (eds.) 1977. Wealth of the Roman world. London: British
Museum.
Kunzl, S. 1997. Die Trierer Spruchbecherkeramik. Trier: Rheinisches Landesmuseum.
Lancaster, L. 2005. Concrete vaulted construction in imperial Rome: Innovations in context.
Mackensen, M. 1993. Die spatantiken Sigillata- und Lampentopfereien von El Mahrine
(Nord-Tunisiens): Studien zur nordafrikanischen Feinkeramik des 4. bis 7. Jahrhunderts.
Miinchner Beitrage zur Vor- und Friihgeschichte 50. Munich: Beck.
Martins, C. B. 2005. Becoming consumers: Looking beyond wealth as an explanation for villa
variability. British Archaeological Reports 403. Oxford: Hadrian Books.
Merker, G. S. 2000. The sanctuary of Demeter and Kore: Terracotta figurines of the classical,
Hellenistic, and Roman periods. Corinth vol. 28, pt. IV. Princeton: The American
School of Classical Studies at Athens.
Merker, G. S. 2003. "Corinthian terracotta figurines: The development of an industry," in
Williams and Bookidis 2003: 233-45.
Meyza, H., and J. Mlynarczyk (eds.) 1995. Hellenistic and Roman pottery in the Eastern
Mediterranean-Advances in scientific studies. Warsaw: Research Centre for Medi-
terranean Archaeology, Polish Academy of Sciences.
Neer, R. T. 2002. Style and politics in Athenian vase-painting: The craft of democracy, ca.
530-460 B.C.E. New York: Cambridge University Press.
Noble, J. V. 1988. The techniques of painted Attic pottery. London: Thames and Hudson.
Osborne, R. 1996. "Pots, trade and the archaic Greek economy," Antiquity 70: 31-44.
Padgett, J.M., et al. 1993. Vase-painting in Italy: Red-figure and related works in the Museum
of Fine Arts, Boston. Boston: Museum of Fine Arts.
Pancirolli, G., and H. Salmuth 1715. The history of many memorable things lost, which were in
use among the ancients, and an account of many excellent things found, now in use
among the moderns, both natural and artificial. 2 vols, London: John Morphew.
Paroli, L. (ed.) 1992. La ceramica invetriata tardo-antica et altomedievale in Italia. Florence:
Edizioni all'Insegna del Giglio.
London: Longman.
Peacock, D. P. S., and D. F. Williams. 1986. Amphorae and the Roman economy: An
introductory guide. London: Longman.
Poblome, J. 2000. Sagalassos red slip ware: Typology and chronology. Turnhout: Brepols.
Rautman, M. 1995. "Neutron activation analysis of Cypriot and related ceramics at the
University of Missouri," in Meyza and Mlynarczyk 1995: 331-51.
Rautman, M., H. Neff, B. Gomez, S. Vaughan, and M. D. Glascock 1999. "Amphoras and
roof-tiles from Late Roman Cyprus: A compositional study of calacareous ceramics
from Kalavasos-Kopetra," Journal of Roman Archaeology 12: 377-91.
Reutti, F. 1991. Neue archaologische Forschungen im romischen Rheinzabern. Rheinzabern:

Terra Sigillata-Museum.
Rice, P. M. 1987. Pottery analysis: A sourcebook. Chicago: University of Chicago Press.
Romer-Strehl, C., A. Gebeland, and G. Frischat. 2005. "Bleiglasurtechnologie in Mitte-
leuropa (1-12 Jh.n.Chr.): Eine Untersuchung zur Glasurtechnik und -rezeptur," Acta
Rei Cretariae Romanae Fautorum 39: 209-16.
Rotroff, S. I. 1982. The Athenian Agora. Vol. 22, Hellenistic pottery: Athenian and imported
moldmade bowls. Athens: American School of Classical Studies at Athens.
Rotroff, S. I. 1991. "Attic West Slope vase painting," Hesperia 60: 59-102, pls. 14-46.
Rotroff, S. I. 2006. "The Introduction of the moldmade bowl revisited: Tracking a Hel-
lenistic innovation," Hesperia 75: 357-78.
Rouvier-Jeanlin, M. 1972. Les figurines gallo-romaines en terre cuite au Musee des Antiquites
Nationale. Paris: Maison des Sciences de l'Homme.
Schmid, S. G. 2003. "Nabataean Pottery," in G. Markoe (ed.), Petra rediscovered: Lost city of
the Nabataeans. New York: Abrams, 75-81.
Slane, K. W. 2003. "Corinth's Roman pottery: Quantification and meaning," in Williams
and Bookidis 2003: 321-35.
Snodin, M., and J. Styles 2001. Design and the decorative arts: Britain, 1500-1900. London:
Victoria and Albert Museum.
Sparkes, B. A. 1996. The red and the black. London: Routledge.
Sparkes, B. A., and L. Talcott. 1970. The Athenian Agora. Vol. 12, Black and plain pottery of
the 6th, 5th, and 4th centuries B.C. 2 vols. Princeton: American School of Classical
Studies at Athens.
Stark, M. 2003. "Current issues in ceramic ethnoarchaeology," Journal of Archaeological
Research 11.3: 193-242.
Steinby, M. 1993. "L'organizzazione produttiva dei laterizi: un modello interpretativo per
l'instrumentum in genere?" in W. V. Harris (ed.), The inscribed economy. Journal of
Roman Archaeology Suppl. 6. Ann Arbor, MI: JRA, 139-44.
Stortz, S. 1993. Tonrohren im antiken Gewolbebau. Deutsches Archaologisches Institut Rom
Sonderschriften 10. Mainz: von Zabern.
Swan, V. G. 1984. The pottery kilns of Roman Britain. London: H.M.S.O.
Tyers, P. 1996. Roman pottery in Britain. London: Batsford.
Van Alfen, P. G. 1996. "New light on the 7th-c. Yassi Ada shipwreck: Capacities and
standard sizes of LRA1 amphoras," Journal of Roman Archaeology 9: 189-213.
Vickers, M., and D. Gill. 1994. Artful crafts: Ancient Greek silverware and pottery. Oxford:
Clarendon Press.
Vickers, M., 0. Impey, and J. Allan. 1986. From silver to ceramic: The potter's debt to
metalwork in the Graeco-Roman, oriental and Islamic worlds. Oxford: Ashmolean
Museum.
Vroom, J. 2003. After antiquity: Ceramics in the Aegean from the ;,th to the 20th century AC:
A case study from Boeotia, Central Greece. Leiden: University of Leiden.
Walters, H.B. 1905. History of ancient pottery: Greek, Etruscan and Roman. 2 vols. London:
John Murray.
Webster, P. 1976. Romano-British coarse pottery: A student's guide. London: Council for
British Archaeology Research Report 6. London: Council for British Archaeology.
Whitbread, I. K. 1995. Greek transport amphorae. Fitch Laboratory Occasional Paper 4.
Athens: British School at Athens.
Whitbread, I. K. 2003. "The study of a basic resource for ceramic production," in Williams
and Bookidis 2003: 1-14.
Williams, C. 1989. Anemurium: The Roman and early Byzantine pottery. Toronto: Ponti-
ficial Institute of Mediaeval Studies.
Williams, C. K., and N. Bookidis. 2003. Corinth: The centenary, 1896-1996. Princeton:
Williams, D. 1997. "Ancient Greek pottery," in Freestone and Gaimster 1997: 86-91.
Winter, N. A. 1993. Greek architectural terracottas: From the prehistoric to the end of the
archaic period. Oxford: Clarendon.
CHAPTER 21
GLASS PRODUCTION
E. MARIANNE STERN
PRIMARY AND SECONDARY

WORKSHOPS
Throughout antiquity, glassmaking and glassworking were two separate crafts.

Glassworkers imported raw glass which they turned into finished products, just as
metalsmiths and other metalworkers imported refined material to be worked into
the final product. Women and men glassblowers are known from the first century
A.D. (Stern 1999b: 456-57; 2004: 115-17), but there is no evidence regarding gender
before this time.
Before the invention of glassblowing in the first century B.C., raw glass was
made in just a few specialized centers, from which it was exported throughout the
ancient world. Glass continued to be made in few primary workshops in the Roman
period. The amazing uniformity of the composition of Roman glass from different
areas and periods has been ascribed to the use of just one source of sand, located
in the eastern Mediterranean (Picon and Vichy 2003: 17-31), but ancient literary
sources of the first century A.D. mention five glassmaking areas that made use of
local sands (Stern 2004: 96-97; Freestone forthcoming). Seen against this back-
ground, the excavation of materials possibly indicative of primary glassmaking at
Rhodes in the Hellenistic period is exciting (Triantaphyllides 2000: 193-94; Rehren
et al. 2003).
Glassmaking took place in the Roman province of Germania in the second half
of the fourth century A.D. (Gaitzsch et al. 2003), but numerous remains of glass
furnaces elsewhere in western Europe and Britain have been identified almost
without exception as secondary workshops (Foy and Sennequier 1991; Foy and
Nenna 2001: 47-60).
GLASS PRODUCTION 521
Very little is known about how the trade was organized to keep a steady supply
of raw glass for glassworkers and other artisans, but it must have been quite complex,
since arrangements involved provisions for transport over land and over water. (On
the structure of the Roman glass industry, commerce and trade, and the usage of
glass in a social context, see Stern 1999b and Nenna 2000.)
The division into primary workshops for making the glass and secondary
workshops for working and shaping it affected not only the structure of the glass
industry; it also influenced early theories about the nature of glass. Plato (Ti. 61b)
expressed the opinion that glass belonged to the class of bodies that are compounds of
earth and water: "And of these substances, those which contain less water than earth
form the whole kind known as 'glass' and all the species of stone called 'pourable.'"
Aristotle associated glass with metals: "gold, silver, bronze, tin, lead, glass, and many
kinds of stone which have no name, for all of these are produced by heat" (Mete.
4.389a; c£ Theophrastus, Lap. 49). These theories fit traditional explanations of the
natural world as a product of transformation of the four elements, air, water, earth,
and fire. Greek philosophers did not recognize that glass was created by a true
transmutation of materials (Beretta 2004b: 10--14, 17). The confusion can be explained,
probably, by the fact that, althougli glass objects were common in classical and early
Hellenistic Greece, the raw glass was imported. The difference between a metal ingot
made by refining and a glass ingot made by transmutation went unrecognized.
GLASS MAKING
Glass is the earliest man-made, artificial material. The basic raw materials of an-
cient glass-soda, lime, and silica-are the same as those used in other areas of
ceramic technology, such as faience and glazes, but the materials are mixed in
different proportions and processed differently. Melted silica (sand, crushed quartz,
or crushed flint) would make the ideal glass, but this material requires such high
temperatures that it has only become possible in modern times. Adding alkali
(soda, natron, or potash), which acts as a flux, lowers the melting temperature but
has the disadvantage of making the mixture soluble in water. Lime counteracts the
solubility and acts as a stabilizer, making the glass durable.
In antiquity, the lime needed for durability was not always recognized as a sep-
arate component. We can only guess how many glass objects have completely dis-
integrated, leaving only those that happened to contain sufficient lime. The absence
of this stabilizer may in part explain the dearth of glass objects from the late second to
early first millennia B.C. (Reade et al. 2003). Ancient glassmaking recipes indicate that
the lime came either with the alkali, because some plant ashes contain lime, or with
the silica, for example in the form of crushed shells mixed with sea sands (Brill 1988:
264-69; Newton and Davison 1989: 54). Most Mesopotamian glassmakers made glass
from a pure silica, in the form of river pebbles, mixed with plant ashes, which
contained lime, while most Mediterranean and later most Roman glassmakers used
sands containing lime with a relatively pure form of mineral soda called natron that
formed naturally, for example in the Wadi Natrun between Cairo and Alexandria
(Pliny, HN 31.46; Shortland 2004). There were numerous recipes for glass. Silica is the
main ingredient, constituting more than half the total amount. Most Roman glass
probably contained about 15 to 20 percent Na2 0 and about 10 percent CaO.
Pliny (HN 36.190-92) mentions the occasional addition of shells to the batch,
which would have been a source oflime, but he does not recognize its importance.
When he describes the "new method of making glass" he names only two ingredi-
ents: sand and soda. The stabilizing lime content explains the renown of particular
sands for glassmaking, such as those from the Belus River in ancient Phoenicia and
from the Volturnus River in Campania.
In antiquity, the process of making glass from basic ingredients involved two or
more stages, each requiring a different furnace (Newton 1980; Newton and Davison
1989: 61-62). Clay tablets with cuneiform texts found at Nineveh in the library of
king Assurbanipal (668-627 B.C.) preserve recipes for making and coloring a vit-
reous material that was probably glass (Brill 1970; Oppenheim 1970; Moorey 1994:
210-14). The recipes show that Mesopotamian glassmakers made a basic glass in a
two-step heating process: first the production of frit and then transforming the frit
into raw glass.
At least three types of installations for preparing raw glass existed. Small round
cakes of colored glass, as found on the Late Bronze Age shipwreck at Ulu Burun,
were made individually in cylindrical pots (Nicholson et al. 1997; Rehren 2005).
Roman-period shipwrecks have produced large chunks of natural bluish green glass
(e.g., Foy and Nenna 2001: 108, fig. 132); these came probably from large tank fur-
naces of the type excavated at Hadera and elsewhere in Israel (Gorin-Rosen 2000;
Freestone et al. 2000 ). Chunks with rounded edges, as retrieved from the late third-
century B.C. Sanguinaires A Shipwreck (Foy and Nenna 2001: 25, fig. 3), were made
by a third method, recognized at Beirut (Foy 2005b: 12) and in the Wadi Natrun
area (Nenna et al. 2000; Thirion Merle et al. 2003). The raw glass flowed from a tank
into a shallow basin with curved walls in which it cooled and stiffened (cf. Nenna
2003: 54). The surviving installations probably date from the first or second century
A.D. (M.-D. Nenna, personal communication 2005).
THE WORKING PROPERTIES OF GLASS
The properties of glass were recognized sporadically, over a long period of time. The
earliest glass objects (first half of second millennium B.C.) were beads and small
ornaments. After the discovery of glass as a material, it took about 500 years before
Strain
14
13 Anneal
12
11
.,
-~
~
10
.5
t8
.,
> Cheese 8
Soften
Honey 4
500° 1000°
Temperatures in Degrees Celsius
Figure 21.1. Viscosities and temperatures of soda-lime-silica glass. (After Stern 1995: 34,
fig. 16.)
the first vessels were produced, and an additional 1,500 years before a glassworker
realized that glass could be inflated, or blown. The most important property of glass
is the fact that the transformation from a solid to a fluid state takes place gradually.
Unlike ice, which melts to water at one specific temperature, glass softens very
slowly. The hotter the glass, the lower its viscosity (resistance to flow). The degree of
viscosity depends not only on temperature, but also on the composition of the glass.
Modern soda-lime-silica glass has the consistency of a hard cheese at approximately
700°C; it is called molten when it is as drippy as honey at approximately 1,050°C.
The graph in figure 21.1 (based on Brill 1988: 270-80) demonstrates the relationship
between temperature and viscosity. The horizontal line of the graph shows tem-
peratures in degrees Celsius, the vertical shows the viscosity. Viscosity is measured
in units known as Poises. The value is so high that these are noted as logarithms:
(log) 7.6 is equivalent to 107 .6 Poise (about the consistency of hard cheese). Ancient
glass probably softened and melted at lower temperatures (R. Lierke, personal
communication 2005).
The difference between the softening point (10 7.6 Poise) and the temperature
for removing molten glass from a furnace or crucible ("gathering") is the working
Cheese Honey
Strain Annealing Softening Melting
Point Point Point Point
14.5 13 7.65 4
I
I I I I I I I I
I I I I I I
15 14 13 12 11 10 9 8 7 6 5 4 3
Chunk Gathering 14.0-11.0
Working Range 7.65-4.0
Sagging 10.0-5.5
Fuse to Stick 7.0-6.0
Flatten to a Disk 5.5-5.0
Chip Casting 5.3-4.3
Marvering 4.5-4. l
Drawing Cane 4.5-4.3
Trail Thread 4.4-4.2
Hot Gathering 4.0-3.2
Figure 21.2. Viscosities required for various glassworking operations. (After Stem 1995: 35,
fig. 17A.)
range. The time during which different glasses can be worked can vary consider-
ably because the temperatures needed to reach a specific viscosity depend on the
composition of the glass. In addition, the heating time required to reach a specific
temperature varies per composition. The working range of most soda-lime-silica
glasses is approximately 1,050°C-700°C ). Ancient glassworkers gauged the tem-
perature and viscosity of the glass by its color and feel (its movement on the tool),
just as glassblowers do today. Color is an accurate indicator of temperature. The
cuneiform texts from Nineveh describe a three-stage sequence of colors: "to glow
red," "to glow green/yellow" (perhaps orange yellow is meant?), and "to glow
golden yellow," the highest temperature they could reach (Oppenheim 1970: 73).
Each glassworking operation requires a specific viscosity (figure 21.2). The graph in
figure 21.3 shows the approximate temperatures required for most soda-lime-silica
glasses (Stern and Schlick-Nolte 1994: 21-24).
Glass shrinks during cooling. To relieve stress and tension between the surface,
which cools first, and the interior, which retains heat for a much longer time, glass
needs to be cooled slowly so that all areas shrink at the same rate: the thicker the
glass, the longer the annealing period. Annealing cracks in many ancient glasses
suggest they were not annealed sufficiently. Ancient glassblowers may have let their
Cheese Honey
Softening Melting
Strain Annealing Point Point
Point Point 700° 1000°
490° 529°
40 oo 11 00° 120 oo
~00° I 600° 700° 800° 900° 1000°
Working Range
700°-1000°
Chunk Gathering
505°-590°
Sagging
D
625°-830° I I
Fuse to Stick
735°-800° L
Flatten to a Disk
830°-875°
Chip Casting
n
850°-965° I
Marvering
930°-985°
Drawing Cane
~J
930°-965°
Trail Thread
940°-970°
Blow
970°-990°
IJ
Hot Gathering
1000°-1150° I
I
Figure 21.3. Temperatures required for various glassworking operations. (After Stern
1995: 35, fig. 17B.)
pieces cool in trays or baskets of ashes, olive pits, or the like. It is thought that "two
black ash heaps on either side of the (glassblowing) furnace, containing olive pits,
and several vessels, some of them distorted" excavated in a glassblower' s shop at
Bet Shean, Israel, served for annealing; the olive pits "were probably used for their
heat-conserving properties" (Gorin-Rosen 2000: 59-60).
Glass adheres to other materials when both materials are hot. This creates a
problem for artisans working with molds. Greek day molds of the late fifth to early
fourth century B.C. preserve on their surface traces of corundum, thought to have
served as a separator (Schiering 1991: 19-21). An experiment at the Toledo Museum
of Art revealed accidentally that bone ash, a substance readily available in antiquity,
can act as a separator (Stern and Schlick-Nolte 1994: 23-24).
A salient difference between ancient and modern glassworking is the limited
use of molten glass in antiquity. The reasons were probably twofold: insufficient
pyrotechnology (the glassblowing furnace, invented in the first century A.O.,
changed this) and unawareness that glass can be melted entirely (Stern 1991; 1999b:
450-54) . Crushed glass and individually heated small chunks sufficed for the
production of most early artifacts. Artisans took care not to overheat, because a
piece was lost if the glass got too hot and melted.
Figure 21.4. Beehive-shaped core-forming furnace. (Drawing by D. F. Giberson, after-

Lierke 1999: 30, fig. 53.)
Before glassblowing, glassworkers probably made do with small mobile fire

pots similar to those used by metalsmiths in Pharaonic Egypt (Stern and Schlick-
Nolte 1994: 24-25, 82, fig. 153). Open at the top, the beehive-shaped firepot allowed
the artisan to work above the heat and control the temperature in the piece (figure
21.4; Giberson 1995, 1996). Similarly-shaped clay bread ovens are still produced
near Antakya (personal observation, 2005). Glassworkers resorted to potter's kilns
for firing glass in molds and preparing flat glass.
Consequently, it should be no surprise that glassworking furnaces predating the
invention of blowing have not been found. The materials from a Hellenistic work-
shop at Rhodes are illuminating. No traces of furnaces were observed, although the
finds include thousands of glass beads made by a variety of techniques, mono-
chrome and polychrome overlay canes, small chunks of raw glass, and waste from
bead and vessel-glass production as well as objects of bone and clay which might
have been tools for shaping the glass and "shallow clay receptacles ... covered with a
thin, uneven layer of glass, either all one color or streaked with several colors ... "
(Weinberg 1969: 149, pis. 85-86; Triantaphyllides 1998: 31-32 adds new finds) .
COLORED GLASSES
Due to impurities in the ingredients (Verita 1995), the natural color of glass is
usually a pale bluish green, often yellowish green in late Roman times (Cool and
Price 1995: 11-12; Foy 1995a: 198-200). Depending on the amount of oxygen in the
atmosphere of the furnace, a wide range of yellowish and bluish greens can be
achieved without the addition of colorants. In a reducing atmosphere, the iron im-
purities, combined with sulphur, can produce yellowish and greenish browns, often
called amber (Brill 1988: 269-76). Beginning in the sixteenth century B.C., when the
first glass vessels were made, glassmakers produced a wide range of bright colors by
adding small amounts of metallic oxides to the frit. Copper produced blues, greens,
and reds; manganese produced pinks and purplish reds; and cobalt produced deep
blues (Newton and Davison 1989: 42-45; Fiori and Vandini 2004: 185-90). Much
depended on the amount added, the combination of elements in the glass, tem-
perature and length of heating, and-last but not least-the atmosphere in the
furnace (oxidizing or reducing).
The opacity or limited translucency of most early glasses resulted from low
melting temperatures that allowed microscopic air bubbles to remain in the glass as
it cooled. Sometimes antimony was added deliberately. Lead and antimony pro-
duced a yellow glass, calcium and antimony opaque white. The addition of a copper
oxide produced opaque red and orange glasses; these were difficult to make because
the glass had to be heated for a long time under reducing conditions (Brill 1970: 119-
21). For black glass, ancient artisans employed strongly colored translucent and
transparent glasses, because these appear black when no light is transmitted. Ptol-
emaic and Romano-Egyptian inlays often use manganese pinks and purples, but
dark blues and dark greens also occur (Stem and Schlick-Nolte 1994: 21). In Roman
glass of western Europe, black glass was made with ambers and greens.
In Pharaonic Egypt, coloring took place in primary workshops (Rehren et al.
1998; Rehren 2005); the trade in ingots attests to this practice. The Ulu Burun
shipwreck produced ingots of two tones of blue, purple, and orange (Pulak 2005:
68-71); opaque red ingots come from Nimrud (Barag 1985a: 108-9, nos. 166-67;
Bimson and Freestone 1985: 119-22). The cuneiform tablets from Nineveh pro-
vide recipes for adding colorants in a third heating process (Oppenheim 1970: 79).
In Pliny's time, however, coloring appears to have been done also in secondary
workshops (HN 36.193; cf. Stem 1995: 67; 2004: 99).
The change in practice would have reduced problems of compatibility, because
glasses with different compositions often have different rates of shrinkage, and
these can cause the glass to crack when it cools. A secondary workshop at Jerusalem,
active in the first half of the first century B.C., produced transparent colors by
adding manganese and/or cobalt (Israeli 2003: 57). Raw glass was colored on the
spot, probably, also at Rhodes (Triantaphyllides 2000: 194).
Pliny mentions several gemstones counterfeited in glass: amethyst (HN 12.72,
37.51), opal (HN 37.83), carbunculi (HN 37.98), callainae (HN 37.112), jasper (HN
37.117), and a saffron-yellow stone (HN 37.128); he also discusses the detection of
fakes (HN 37.199-200). Most glass gems, however, were not deliberate fakes (Krug
1995: 182-83; Stem 2001: 358-60). Counterfeits were traditionally the domain of
alchemists, who studied the transmutation of substances and counterfeited pre-
cious stones. Several collections of ancient alchemists' recipes exist for creating
counterfeits in stone and glass (Halleux 1981; Beretta 2004b: 21-30; Tolaini 2004:
211-19 ). The earliest recipes probably were written down in the fourth century A.D.,
but they call to mind Theophrastus' comments (Lap. 48) on artificial stones made
by melting earth, softening it, and hardening it again.
Many ancient glasses have lost their original colors because moisture and
chemicals in the soil attack the surface of the glass, leach away the alkali, and alter
the composition of the glass in a process called weathering. Weathering can pro-
duce a variety of effects, from a thick creamy white or dull black layer, beneath
which the original color is often preserved, to a metallic or oily sheen known as
iridescence.
COLORLESS GLASS
Pliny (NH 36.194) and Strabo (16.2.5) report technical improvements for making
crystal clear (decolorized) glass. If they refer to the same improvements, these must
have been developed before Strabo's death in 24/25 A.D. The exact nature of the
improvements is not clear. They may have resulted from a new furnace technol-
ogy that allowed for higher temperatures during melting (Stern 1995: 23-25), or
they may have been associated with an intermediate stage in the production of raw
glass when a material Pliny calls hammonitrum, "sandsoda," was formed (Stern
2004: 80--82).
Ancient artisans decolorized glass with manganese or antimony, of which the
latter was more effective (Verita 1995: 292). Colorless glass was not a new invention
of the Roman period, since cast glass vessels of the eighth to seventh century B.C.,
decolorized by the inclusion of antimony in the batch, have been excavated at
Gordion (Saldern 1959). In the fifth to fourth centuries B.C., glass bowls almost
as clear as rock crystal, the so-called Achaemenid vessels, were produced in some
quantity (Oliver 1970; Triantaphyllides 2001; Ignatiadou 2004a). High quality, col-
orless glass vessels continued to be produced throughout the Hellenistic and Roman
periods.
In Greece, the optic qualities of colorless glass were a source of fascination from
the moment the material became available. Aristophanes (Nub. 768) joked about a
glass lens that could be used to burn away a record of debt, but the archaeologi-
cal evidence for such lenses is scant (Sines and Sakellarakis 1987; Plantzos 1997).
Gorgias (483-376 B.C.) and Theophrastus (Jg. 73) mention glass "burning-mir-
rors," but there is no archaeological evidence for mirrors in this period. Pliny (HN
36.193) attributes the invention to Sidon. Small convex glass mirrors, coated with
lead on the interior, were made from the first century A.D. (Amrein 2001: 41-48;
Lazar 2004: 78, no. 93).
In addition to its magnifying, reducing, and reflecting powers, the transpar-
ency of glass was a major attraction. A famous painting by Pausias (ca. 350 B.C.)
depicted "both the glass cup (from which she was drinking) and the woman's face
showing through it" (Pausanias 2.27.3). Transparent glass appeared in the con-
struction of mechanical devices and gadgets, many of which are described by Hero
of Alexandria (Di Pasquale 2004). Ancient medicine and anatomy, in particular
research on the working of the human eye, benefited in a variety of ways from
increased understanding of the properties of clear glass (Beretta 2004c).
The experiments of Claudius Ptolemy, who was active in Alexandria in the
second century A.D., are of interest from the point of view of glass technology.
Ptolemy describes two sets of experiments requiring pure transparent glass (Optics
5,14-20, 67-86). He measures the refraction of the visual ray when it passes between
and through transparent substances of diverse density (air, water, and glass) with a
solid piece of decolorized glass, and he uses thin-walled transparent colorless vessels
to measure the distortions in form and dimension of objects seen through surfaces of
various curvatures (Strano 2004). Previously, Seneca had noted that "fruits are much
larger when seen through glass" (Q Nat 1.3.9; cf. 1.6.5), a phenomenon that must
already have been familiar in the second half of the first century B.C., as is shown by
numerous frescoes in Rome and Campania depicting glass bowls filled with fruit
(Naumann-Steckner 1991; Sabrie and Sabrie 1992). Ptolemy's Optics (ca. 160-170
A.D.; Smith 1996: 3) is our earliest evidence for the decolorized glass that would
eventually make "Alexandrian" glass the generic name for clear glass (cf. Diocletian's
Price Edict 16.1,3: Giacchero 1974; Barag 1985b: 113-16; 2005; Stem 1999b: 460-62).
With regard to optics, we may probably credit glass cutters with two discov-
eries (Stem 1997: 202-4; 2004: 119-20): first, that the close-set, slightly concave
facets with which they began to decorate tableware in the early 70s A.D. created an
optical illusion of countless reflections on the surface, improving, as it were, the
natural shimmer of rock crystal; second, that facets hollowed out on the under-
side of a piece of transparent glass appeared as convex bosses when seen from
above. Artisans exploited this optical illusion to create a new line of cast and cut
colorless glass tableware that "improved" upon the effect of embossed silver.
(Grose 1991: 12-18).
GLASSWORKING IN CLASSICAL GREECE
The most authoritative study of ancient glass in Greece stresses the curious dis-
crepancy between the silence of ancient writers on the subject of glassworking in
Greece and the archaeological evidence which shows clearly that Greece played a
significant role in the development of the craft (Weinberg 1992: 9). More recent
research now allows us to suggest that several important discoveries regarding the
properties of glass and the development of new glassworking techniques originated
in Greece.
The largest numbers of glass vessels produced in the Mediterranean area in the
seventh to first centuries B.C. were small polychrome scent bottles shaped around a
core (or just a rod) that was removed after completion of the vessel. Core-formed
vessels were made in Greece from the late sixth century B.C. to the early first cen-
tury A.D., when they were supplanted by blown vessels (Harden 1981; Grose 1989).
Dating back to about 1500 B.C., the curious technique is one of the earliest known
for making glass vessels. The artisan prepared a core of clay and sand, mounted the
core on a metal working rod (mandrel), coated the core with glass, and scraped out
the core after the vessel was completed. It required the barest minimum of pyr-
otechnology to pack crushed glass against a wetted core. Water made the glass stick
like wet sand. Careful heating caused the glass to glaze over, and subsequent layers
could be added by rolling the core through more crushed glass (Stern 1998: 186-88,
200-203, figs. 1-19; 2002b, 356-59, figs. 8-17). Coils observed near the bottoms of
Greek core-formed vessels (Stern and Schlick-Nolte 1994: 39-41, fig. 22) suggest
these were made by coiling a thick trail of glass around the core, a procedure that
would have speeded up production. The colorful decoration was trailed on either
with prefabricated, thin, monochrome canes or with hot bits of colored glass on
other mandrels. Dragging the threads up and/or down with a pointed tool or hook
created zigzag patterns, festoons, feather patterns, and the like.
While core-forming was suitable for the production of narrow-necked, closed
vessel forms, casting was more appropriate for open vessels. The artisan probably
filled the mold and/or the mold's sprue with one or more chunks of glass at room
temperature and "fired" it in a potter's kiln. Unlike bronze, glass is not a good
conductor of heat. Crucible pouring (called hot-pour casting) would have been
difficult, because it requires temperatures of approximately 1150° C or more to
prevent the glass from cooling and stiffening before filling the mold completely.
Casting was primarily a means of making a hollow object, not of making mul-
tiples. Early cast vessels are precious, individual pieces, made often of transparent,
colorless glass imitating rock crystal (Saldern 2004: 53-62). Both the wax model and
the mold were destroyed in the process: the wax model when the artisan burned it
out of the mold in preparation for casting and the mold after the piece was fired.
Made of plaster that was soluble in water, the mold was broken away to extract the
piece. In the fourth century B.C. casting was largely supplanted by glass pottery (see
below).
In Greece, transparent colorless glass seems to have been reserved at first for
temples and conspicuous private luxury in the form of jewelry and dedications to
the gods. In the fourth century B.C., colorless glasswares began to grace the tables
of the wealthy. The evidence from excavations on Rhodes (Triantaphyllides 2000:
195-201) and in Macedonia (Ignatiadou 2002) indicates that clear glass vessels were
produced in Greece in the second half of the fourth century B.C. Literary evi-
dence and records in the Athenian Inventory Lists suggest vessel production may
have begun already in the first half of the fourth century B.C., with casting in the
Achaemenid tradition (Stern 1999a).
Excavations in Phidias' workshop at Olympia added an unexpected dimension

to our knowledge of glassworking in fifth-century B.C. Greece (Schiering 1991).
Solid ornaments, cast in open clay molds, decorated the throne and/or some other
part of the statue of Zeus. This technique was then already centuries old (Stern and
Schlick-Nolte 1994: 49-50). The most exciting finds were thin, curved scraps of
colorless glass, thought to be waste from glass drapery for a female statue, perhaps
the figure of Nike perched on Zeus' outstretched hand (Schiering 1999). The curved
fragments were not cast, but sagged.
Sagging required flat glass. Preparing large slabs of flat glass with a consistent
thickness and building sturdy molds for sagging them were technical feats in them-
selves. The slabs could not be made by crucible pouring. Their production was
based, probably, on a newly discovered property of glass: if a sizable chunk of glass
is heated sufficiently, it will flatten out into a round disk, about 0.7 cm thick; the
diameter depends on the size of the original chunk (Stern and Schlick-Nolte 1994:
66-67; Stern 2002b: 362-63). Multiple slabs could be prepared in a potter's kiln. The
fragments from Phidias' workshop are the earliest examples of sagging known to-
day (Stern and Schlick-Nolte 1994: 68; Stern 2002b: 361-63), and it is tempting to
credit Phidias with the invention. Sagging was to become one of the most important
techniques for shaping vessels in the Hellenistic period.
Even if Phidias' original goal may have been protection of the statue's gold leaf
decoration, he would have soon discovered the magnifying effect of the glass and
exploited its artistic qualities (on similar modern experiments see Schiering 1999).
The concept of covering gold leaf with clear colorless glass was to become a major
element in the artistic repertoire of Greek artisans. It inspired the great Greek
tradition of gold-glass, which flourished, and possibly originated, at the Mace-
donian court (Stern 1999a: 40-41; Ignatiadou 2001, 2002).
Closely related to the preparation oflarge disks of flat glass is the production of
small roundels with one convex surface. Any small piece of glass will do (Lierke
1999: 22, figs. 34-36). The colored glass inlays in the capitals of the Erechtheion and
the base of the Nemesis statue at Rhamnous were probably made this way (Stern
1985; 1989), as were numerous "buttons" and gaming pieces.
GLASS POTTERY
Casting was a time-consuming process. Rotary mold-pressing could achieve the

same result more quickly. The technique is familiar from the production of mold-
made Hellenistic pottery (Rotroff 1982), thought to have been invented in Athens
in the last quarter of the third century B.C. (S. Rotroff, personal communication
2005). The main difference was that potters employed molds for mass production,
while glassworkers could not reuse molds with decoration in (high) relief be-
cause they had to be broken away from the finished product. There are indica-
tions that the basics of the technique were explored by Greek glassworkers. The
earliest evidence comes from Macedonia in the second half of fourth century B.C.
(Ignatiadou 1998, 2004b), although it can not be excluded that some Achaemenid
vessels were mold-pressed (Lierke 1993b: 324-25; 1999: 36; for cast Achaemenid
vessels, see Stem and Schlick-Nolte 1994: 166-67; Ignatiadou 2004a.). The role of
faience production in the development of mold-pressing needs to be researched.
The potter's wheel is unknown in modem glass production. Lierke (1991)
broke new ground when she discovered that it provided the solution to the riddle
of Hellenistic spiral reticella bowls, which are made from bicolored trails coiled
spiral-fashion as in a basket. Subsequently, she identified several other groups of
wheel-made glass vessels. These include shapes that previously had not been sat-
isfactorily explained, as well as vessels previously thought to have been made by
other methods, such as casting, kiln-sagging, blowing, and grinding from a blank.
Characteristic tool marks of glass pottery are rotary scratches (Lierke and Lindig
1997) and indents in the external rim area (Stem and Schlick-Nolte 1994: 78, figs.
134-36).
There are two main groups of glass pottery. Mold-pressing and its variants
required concave molds, which were usually made of plaster because they had to be
removed after completion of the piece. The other techniques could be executed
with reusable ceramic molds. As a result, most mold-pressed vessels were individual
pieces, while sagging and tooling were used frequently for mass production. Some
ofLierke's proposals are under dispute, such as those concerning the production of
cameo glass vessels and cage-cups (diatreta). Closer attention to tool marks, more
research, and more experiments are needed. Nevertheless, the creativity of ancient
glassworkers in adapting the potter's wheel to their needs is astonishing, and a
striking example of the dose relationship between glassmaking and ceramic pro-
duction. The following selection gives an impression of the wide range of appli-
cation of the wheel in ancient glass production.
Mold-Pressing
A concave mold is mounted on a wheel (figure 21.5). The artisan inserts a gob of
hot glass into the mold, which shapes the future vessel's exterior, and uses a moist
wooden plunger to press the glass against the mold's wall and shape the vessel's
interior. In theory, mold-pressing could have been done with a stationary mold, but
the rotary movement of the wheel helped to distribute the glass evenly throughout
the mold. The combination of two forces-vertical pressing and rotary rubbing-
required lower temperatures than modem mold-pressing. After the invention of
blowing, this technique remained popular for luxury wares such as vessels with
ceramic profiles (Lierke 1993b; 1998; 1999: 55-58), cameo glass (Lierke 1995b; 1999:
67-96), and vessels decorated with high relief (Lierke 1999: 100-107).
(c)
Figure 21.5. Moldpressing. (After Lierke 1999: 82, fig. 210.)
Double Mold-Pressing
Lierke's proposal that cage cups (diatreta) were cut from mold-pressed, double
walled blanks (Lierke 1995a: 60-61; 1999: 116-29) has evoked strong reactions from
scholars who believe the entire vessel was made by cutting (Saldern 2004: 386-87).
Nevertheless, her arguments against cutting the entire vessel, including all the struts
from a thick-walled blank, are convincing (Lierke 2001; 2004). The artisan first
made a solid cup (the future cage) with a mold and a moist wooden plunger. Then a
smaller, perforated mold was placed inside the cup and a second gob of hot glass
was pressed into this mold to form the inside cup. The plunger forced part of the
glass outwards through the perforations until it touched the outer cup and formed
struts joining the two cups. The walls of the outer cup were then cut away between
the struts, the perforated mold ground away, and all the edges (and often the struts)
were ground and polished to create a lacelike openwork cage.
Rotary Pressing
This technique, a variant of mold-pressing not yet verified by experiments, com-
bined elements of mold-pressing with elements of sagging. A convex mold re-
sembling an upside down bowl is mounted on a wheel. The artisan presses down
the glass flowing over the convex mold with a concave mold applied from
above. The convex mold shapes the vessel's interior, the concave mold the vessel's
exterior. Lierke suggests Hellenistic footed bowls were made this way. The foot
was added by mold-pressing, using a separate mold (Lierke 1998: 198-200; 1999: 49,
fig. no).
Winding
Hellenistic spiral reticella vessels are made by winding. The artisan mounts a
convex mold on a wheel and pulls up to five parallel reticella coils from the hot
bit(s) onto the mold by turning the wheel (Lierke 1991; 1999: 39-43; cf. Stern and
Schlick-Nolte 1994: 71-72). The multicolored strands making up each reticella coil
are twisted while they are being pulled off the bit.
Sagging
The artisan mounts a convex mold on the wheel and sags a prefabricated, heated
disk or a hot cake of glass over the mold. By turning the mold slowly the glass-
worker can remain seated while pressing the rim of the disk against the mold to
stem the flow of the glass (Stern and Schlick-Nolte 1994: 70-71). This technique was
used for mass-producing Hellenistic grooved bowls and for early imperial luxury
wares such as mosaic glass. It is closely related to tooling.
Tooling
The glass disk or hot cake is tooled while it sags or flows down over the turning
convex mold. (figure 21.6) The ubiquitous ribbed bowls, whose production began
in the first century B.C., were tooled. Lierke identified the technique with Pliny's
term torno terere "tooling on a potter's wheel" (HN 36.193), one of three glass
production methods for which Sidon had been famous "once upon a time" (Lierke
1993a; 1999: 51-55; cf. Stern and Schlick-Nolte 1994: 72-79).
Free Shaping
This curious technique, apparently invented in Crete in the early second century
B.C. and used exclusively for a small group of vessels known as Cretan pyxides
(Weinberg 1959), consisted of "throwing" the glass and raising the wall by pressing
the glass with a moist wooden paddle against a plunger sunk in the middle of the
glass (Lierke 1999: 37-39; Stern and Schlick-Nolte 1994: 79-81). Plunger and paddle
replaced the potter's hands for manipulating the hot glass while the wheel turned.
The influence of ceramic technology seems obvious.
Figure 21.6. Tooling of ribbed bowls. (After Lierke 1999: 54, fig. 124.)
GLASSBLOWING
The Roman Empire is renowned for many inventions and institutions that shaped
the cultures of the West; among the most lasting is the art of glassblowing. The
almost unlimited possibilities of transforming the shape of glass on blowpipe or
punty are still a source of fascination for glassblower and spectator alike: the glass-
blower gathers a gob of molten glass on a metal blowpipe about 1.0 to 1.5 m long,
blows and shapes the vessel, affixes a solid metal rod, also known as a punty or
pontil, about 0.80 to 1.25 m long, with a wad of glass to the bottom of the vessel,
separates the vessel from the blowpipe, and holds it on the punty while he or she
finishes the mouth of the vessel. During blowing, glass seems to defy the laws of
nature. Even if a piece falls on the ground, it will not shatter. If the blower reattaches
it quickly and reheats, he or she can restore the shape and complete the piece as
planned. Such a workshop accident may be at the root of ancient stories about an
unbreakable glass cup that was dropped on the floor, was dented like a metal vase,
and was hammered back to shape in front of the emperor (Petronius, Sat. 51; Pliny
36.195; Dio Cassius 57.21.6; cf. Stern 1999b: 441-42).
Pliny (HN 36.193) says that ftatu figurare "shaping by breath," was another of
the three production techniques for which Sidon had been famous in the past. The
discovery that glass can be blown revolutionized the entire glass industry. The
invention of the blowpipe meant that hollow objects and vessels that previously
required labor intensive operations could be made in a fraction of the time, and that
less glass was needed per object. Moreover, blowing permitted the production of
new classes of items. Blown glass tableware played an important role in bringing
Roman culture to the provinces of the empire.
Our earliest evidence for inflation comes from Jerusalem, where numerous
glass tubes, some of them inflated at one end, were found in the waste of a workshop
active in the first half of the first century B.C. (Israeli 1991; 2003: 56). It is revealing
that glassblowing began with tube-blowing. Similar tubes were found in a Hellenistic
bead-making shop at Rhodes (Weinberg 1969). The thread hole of a glass bead was
made either by cutting a prefabricated tube into small sections (drawn bead) or by
winding a trail of softened glass around a mandrel (wound bead).
Like other major discoveries that have shaped the course of history, the dis-
covery that hot air will expand softened glass was accidental. Similar accidents must
have happened previously, but it took an observant artisan to notice what went
wrong and why, and-most importantly-to be curious enough to try to duplicate
the event (Stem 2004: 87-88). If the accident took place during bead-making, this
would explain in part why it took so long before inflation was discovered. Ancient
glassworkers were often specialized; they produced either vessel glass, or beads and
jewelry (Spaer 2001), or window glass (Stem 1999b: 462, 464-66; Dell'Acqua 2004;
Foy 2005a). Beads and vessels usually were not made by the same artisan, with the
possible exception of rod- and core-formed vessels, the techniques of which had
much in common with wound beads. The techniques for producing glass vessels-
casting, core-forming, and glass pottery-afforded little opportunity for discov-
ering that glass can be inflated, and bead-makers had little use for inflation.
While the initial discovery took place somewhere along the Syro-Palestinian
coast, where glassworking and glassmaking boasted a centuries-old tradition, glass-
blowing was perfected in Italy (Stem 1999b: 442-50). Over a century of experiments,
inventions, and improvements separate the first trial inflation of a heat-softened
glass tube from full-fledged Roman glassblowing in the second half of the first
century A.D. Most of the tools and techniques now taken for granted as integral to
the craft were invented during this period: the glassblowing furnace with a dosed
heat-chamber, the iron blowpipe, the introduction of molten glass for gathering,
and the pontil technique for fire-finishing the vessel's rim.
Working with molten glass prepared the way for recycling (Stem 1999b: 450-
54), but it was not universal in Roman glassblowing. Many glassblowers began their
piece by picking up a preheated chunk of glass at the tip of the blowpipe instead of
gathering molten glass. Both techniques had advantages and disadvantages (Stem
1995: 36-37; 1999b: 452-54).
Glassblowing Tools and Equipment

Unlike the chance discovery that hot air will expand softened glass, most of the tools
and equipment needed for blowing resulted from purposeful research in antiquity.
The Roman glassblowing furnace has a dosed heat-chamber that allows for evenly
heating the glass on the pipe, necessary in order to expand the glass evenly (Stem
1999b: 446-47). The glassblower shown on a first-century A.D. lamp found in
Slovenia (figure 21.7) sits on a stool in front of the marver, the flat working surface
for shaping and cooling the glass, depicted here as a small ledge sticking out from
the furnace on the right (Lazar 2005). The man squatting on the left is probably
working the bellows. The lamp dates from around A.D. 70, when iron blowpipes
existed (Stem 1999b: 447-48) but probably were not yet used universally. The
Figure 21.7. Glass worker at furnace, Roman lamp from Skolarice, Slovenia, ca. A.O. 70.
Koper, Regional Museum. (after Lazar 2004: 27, fig. 15, with permission).
depicted pipe is too short to be of iron, since it would have become too hot to
handle. A curious narrow strip tied to the underside suggests reinforcement of a
fragile ceramic (?) blowpipe for blowing a large, heavy tall-necked bottle. Prime
reason for the persistence (and invention) of ceramic blowpipes would have been
the ease and low cost of their production (Stern and Schlick-Nolte 1994: 81-85; Stern
1995: 39-42; 2001: 88-89; 2005).
Very few iron blowpipes have been excavated (Sternini 1995: 83-85; Foy and
Nenna 2001: 41, 77, no. 59). It has not yet been established beyond doubt whether a
thin length of clay stuck into a glob of glass excavated in Turkey is hollow or solid
(Lauwers et al. 2005: 27-28, fig. 4). If it is hollow, it may be a small blowpipe; if
solid, it is perhaps a mandrel for adding small handles or decorative threads.
In addition to blowpipe and punty, the most commonly used tools in any glass
shop today are various types of shears. Of these, only the U-shaped iron shears with
parallel blades, also known as jacks, were used in antiquity (Stern 2002a: 162-64;
2004: 109-12). Originally a Celtic invention, the U-shaped shears (Greek, psalis;
Latin, forfex) were the universal cutting tool of the Greco-Roman world, but one
cannot cut semi-viscous glass with them. For that one needs shears with two cross
mounted blades moving on a pivot, like a pair of scissors. There are dubious in-
dications that this tool had been invented in the Roman empire, but finds are
extremely scarce, difficult to date, and all come from Europe, north of the Alps.
Several characteristics of ancient glass vessels can be explained by the lack of pivoted
shears. Roman glassblowers usually began the handle of a vessel at the shoulder,
pulled the glass up, and attached it at the top where they drew the extra glass out
thin and snapped it off. Many glassblowers exploited the fact that they could not cut
the glass: they folded it back and forth to create interesting folds and spurs at the
upper end of the handle.
Today, many glassblowers shape and cool the glass in moist wooden block
molds, shaped like a cup or bowl, while seated at the bench. Although the tool itself
is simple, there was no need for it in antiquity. This tool was invented after the
introduction of the glassblower's bench in the seventeenth century, when glass-
blowers moved away from their traditional seat in front of the furnace with its
marvering ledge (cf. figure 221.7) and created a new workstation at a distance from
the heat (Stern 2002a: 161-62; 2004: 86).
Mold-Blowing
While free or "offhand" blowing was being perfected in Italy, glassblowers on the
Syro-Palestinian coast explored the intricacies of the multipart mold, also known as
piece mold (Stern 1995). Free blown vessels imitated first polychrome and later
colorless precious stones; mold-blowing concentrated on recreating the chiseled
and chased relief of contemporaneous silver vessels. According to Pliny (HN 36.193)
argenti modo caelare ("chasing like silver") was the third glass forming technique for
which Sidon had been famous in the past. Since glass cannot be chased or chiseled,
and vessels decorated with engraving or wheel cutting postdate Pliny, he probably
meant mold-blown glass since mold-blown vessels "look embossed as well on the
outside as on the inside" (!sings 1957: 45; cf. Stern 1995: 67; 2000: 166-67). Appar-
ently, Pliny did not understand how mold-blown vessels were made and reasoned
that they were produced by the same techniques as their metal counterparts. An
ancient glassblower queried on the subject may well have left Pliny in the dark or
even misled him on purpose. The same misconception reappears in Martial (Epigr.
14.94): audacis plebeia toreumata vitri ("plebeian chased cups of dreadnought
glass").
The ancient glassblower who presented the emperor with an unbreakable glass
cup (above) had no luck with his gift. After learning that no one else knew the
secret, the emperor had the artisan put to death and his workshop destroyed because
unbreakable glass "might cause a devaluation of metals such as copper, silver, and
gold" (Pliny, HN 36.195). If Pliny associated the invention of vitrum flexile "flexible,
malleable glass" correctly with the emperor Tiberius, mold-blowing may have ex-
perienced an early setback in Italy under Tiberius' rule (Stern 1999b: 441-42). After
his death, workshops in the west began producing a formidable quantity of mold-
blown tableware and other vessels (Price 1991). In the eastern Mediterranean,
commercial mold-blowing began earlier. The earliest mold-blown glass currently
known was excavated in an Augustan context at Magdalensberg, Austria, but it was
blown by an eastern Mediterranean glassblower, perhaps by Ennion, the most fa-
mous glassblower of antiquity (Stern 2000: 165-67).
The first glassblowing molds were piece-molds that included the base of the
vessel. Mold seams in preserved vessels show how complex the molds were, con-
sisting often of three or four parts (Stern 1995: 29-30, fig. 15). As time went on
construction became simpler, and by the second half of the first century many
molds in Italy and the west had just two parts. The glassblower did not necessarily
create the archetype, which could be a silver vessel or even another glass vessel.
Successful designs were sometimes pirated, with the result that distant workshops
produced seemingly identical vessels.
The production of mold-blown glass decorated with relief fell out of favor just
around the time that Mount Vesuvius erupted in A.D. 79. The reason was probably
the increased use of molds for utility wares (Stern 2001: 41-42). Large storage
bottles, blown in smooth-walled molds, began to flood the market in the third
quarter of the first century A.D. (Cool and Price 1995). Their prismatic shape
facilitated packing in wooden boxes, and the base moldings promoted brand
recognition. The realization that mold-blown vessels could be mass produced
caused their value to drop. In the eastern Mediterranean, particularly Asia Minor,
where mold-blowing did not become a means of mass production, elegant mold-
blown beakers featuring mythological imagery remained in use well into the sec-
ond century A.D. (Weinberg 1972; Wight 1988).
Decorative mold-blowing experienced a revival in the fourth century A.D.,
when glassblowers discovered a new way of using patterned molds. By expanding
and shaping the vessel after it exited the mold, a wide range of shapes could be
blown with just one mold. The technique appears to have been developed by Syrian
glassblowers who began to expand geometric patterns made in full-size piece
molds (Stern 2001: 133-34). Although this particular technique remained a regional
curiosity, it pointed the way to a novel use of the one-piece mold that would
dominate mold-blowing for all times to come: pattern-blowing. Using one-piece
molds, also known as dip molds, Syro-Palestinian glassblowers produced a distinct
group of vessels decorated with barely expanded geometric patterns, known col-
lectively as Honeycomb Bowls. The real success story was to be the straight-walled,
tapered mold with ribbing or fluting on the interior. Easy to make and easy to use,
the mold enabled glassblowers to create a large variety of expanded vertical and
spiral ribs. Immediately popular throughout the Roman Empire, pattern-blowing
with expanded ribs was destined to become a lasting success that returned to mold-
blowing some of its original appeal. Delicately ribbed glass vessels evoked the
luxury of fine fluted silverware, but without the odium of mass production, be-
cause the glassblower shaped each piece individually.
OTHER DECORATIVE TECHNIQUES
Around 200 B.C., polychrome glass objects imitating precious stones became
fashionable and remained in vogue until about A.D. 70. The modern names given
to various types of luxury glass of this period are self-explanatory: variously pat-
terned mosaic glass vessels, reticella or network vessels, and gold band vessels. They
were made with prefabricated polychrome elements, such as mosaic glass bars,
twisted reticella trails, or strips of gold foil encased between two layers of trans-
parent colorless glass. The most sophisticated objects were so-called sandwich gold
glass vessels: two glass vessels fitted one into the other, with a pattern cut out of
gold foil attached to the outside of the interior vessel (Stern and Schlick-Nolte 1994:
54-66). When commercial glassblowing began in the last quarter of the first cen-
tury B.C., glassblowers tried to emulate the effect of polychrome luxury wares.
Decorative techniques are classified as "hot" or "cold," depending on whether
the artisan decorates the glass while it is hot-threads, flecks, ribs, indents-or
cold, in which case the artisan need not be the glassblower in person-painting,
wheel-cutting, wheel-abrading, freehand scratching, and facet cutting. Sophisti-
cated cold decoration, such as painting and figural cutting, probably was executed
by specialized artisans. Simple wheel-cutting in the form of horizontal grooves
took place in the glassblower's shop, as at Jalame in Israel (Weinberg 1988: 87-98).
Winding a colored glass trail around a vessel was one of the oldest techniques
for beautifying glass (cf. the early core-formed vessels). Applying the trail before
inflating the vessel to its full size could create a marbled pattern imitating precious
stone; threads in relief were applied after fully inflating the vessel. Pinching ribs in
the wall of a vessel likewise took place at an early stage. One particularly attractive
group of early blown vessels, the so-called zarte Rippenschalen, combines wound
trails with pinched ribs (Stern 2001: 82-83).
Many new shapes and decorative techniques imitated contemporary pot-
tery, stone, and metal, indicating the remarkable adaptability of this technology.
Glassblowers in Campania and northern Italy copied the colored sand decoration
of thin-walled Augustan pottery by rolling the inflated vessel through crushed glass
laid out on the marver, but they abandoned this technique after discovering they
could do something potters could not. By applying the crushed glass before fully
inflating the vessel, glassblowers created colored flecks that expanded and flattened
out during blowing, imitating the more expensive mosaic glass (Stern 2004: 119,
pls. 9-12). A variant technique made use of failed glass vessels. Broken into small
pieces, the colorful shards were applied individually to the walls of a new vessel
(Stern 2001: 68).
THE ANCIENT GLASSBLOWER's

OUTPUT
A survey of the archaeological vestiges of workshops has shown that in comparison

with other fire-based crafts the total number of workshops producing any kind of
glass was much lower. In Switzerland, remains were found of just 23 workshops
(6%) producing glass objects, as compared to 87 (23%) workshops for copper alloys,
143 (39%) for iron working, and 116 (32%) for pottery. These percentages are in
agreement with those found elsewhere north of the Alps (Amrein 2006).
In view of the particular exigencies of the metier, modem assumptions that
Roman glassblowing was a large-scale industry comparable to the pottery industry-
with hundreds of employees or even slaves laboring in one establishment-are un-
realistic. Every glassblower needs his or her own working port at the furnace, and
ancient pyrotechnology did not permit expanding the glassblowing furnace with a
second working port. Our earliest evidence for a furnace accommodating more than
one glassblower is a late fifteenth-century drawing in the Vatican Library (Stem
2002a: 160, fig. 3). The only way to boost production was by building a series of small
furnaces, as may have been the case in the Hambacher Forst (Gaitzsch et al. 2000)
and elsewhere north of the Alps, where series of round and rectangular furnaces have
been excavated at single sites.
The Roman glassblowing furnace was small (cf. figure 21.7). The interior di-
ameter was approximately 0.45-0.65 m, a measurement confirmed by excavated
remains at Avenches, Switzerland, and elsewhere (Amrein 2001; Foy and Nenna
2001: 62-63). Where the remains of more than one furnace are excavated in close
proximity, it is not always clear to what extent they operated simultaneously.
Multiple furnaces may have been required for blowing on a daily basis. In the mid-
twentieth century, a primitive glassblowing furnace at Herat, in Afghanistan, op-
erated only every other day because the furnace took 24 hours to cool down after a
day's work. When demand suddenly increased, the owner built a second furnace.
Other impediments to expanding into large-scale operations were the size of the
crucible containing molten glass and the amount of space available for annealing.
To gain a sense of the ancient glassblower's output, it is useful to look at the
output of comtemporary low-technology workshops (Stem 1999b: 454-56). Based
on examples from Herat and Cairo, an ancient glassblower probably averaged 100
vessels per day, or 11,000 per year. This works out to 330,000 vessels in 30 years!
Such a long period of activity may not have been common, but it was not im-
possible, given the witness of the tombstone of the opifex artis vitriae ("glass artist")
Julius Alexander who died in Lyon at age 75.
The physical restrictions imposed by size and design of the ancient glass-
blowing furnace on the number of persons who could blow simultaneously in one
shop affected many aspects of the industry, from the number of people blowing in
one shop and their relationships with each other to external transactions such as
buying raw glass and the marketing and selling of the finished product.
While glassmaking benefited from the achievements of metal and faience pro-
duction, glassworkers were closely related to potters. The development of recipes
for coloring and decolorizing glass contributed to ancient chemistry; the ancient
glassblower's need for a new furnace advanced pyrotechnology. The discovery and
exploitation of the properties of glass was an ongoing process that began in the late
third millennium B.C., but the story is not straightforward, as is exemplified by the
rediscovery of glass pottery in the twentieth century. Discovery of a new property
often lead to the development of a new tool or piece of equipment. It usually took
many decades before the full potentialities of a property were realized. As in all
crafts, tools determined what was feasible.
Unlike other fire-based industries, such as pottery, bronze, and metalworking,
glassblowing did not develop into a large-scale enterprise in antiquity. This was
probably due entirely to physical restraints. The division into primary and secondary
glass production affected not only the structure of the industry but also ancient
thought about the nature of glass. Advances in glass technology and the discovery of
the properties of glass were the exclusive domain of artisans. Unlike philosophers
who attempted to explain natural phenomena in a systematic way, glassworkers
were (and still are) primarily practical. They built on experience, did things "on the
spur of the moment," and learned from unforeseen events and accidents, which they
might-at first for fun-try to duplicate.
REFERENCES
Amrein, H. 2001. L'atelier de verriers d'Avenches: L'artisanat du verre au milieu du 1er siecle
apres J.-C. Cahiers d'archeologie romande 87 = Aventicum 11.
Amrein, H. 2006. "Quelques reflexions sur !'implantation et !'organisation des ateliers de
verriers dans les provinces romaines au Nord des Alpes," in G. Creemers, B. Demarsin,
and P. Cosyns (eds.), Roman glass in Germania Inferior. Atuatuca 1.Tongeren: Pro-
vinciaal Gallo-Romeins Museum, 58-63.
Barag, D. 1985a. Catalogue of western Asiatic glass in the British Museum. Vol. 1. London:
British Museum.
Barag, D. 1985b. "Recent important epigraphic discoveries related to the history of glass-
making in the Roman period," Annales de l'Association Internationale pour l'Histoire
du Verre 10: 109-16.
Barag, D. 2005. "Alexandrian and Judaean glass in the Price Edict of Diocletian," Journal of
Glass Studies 47: 184-86.
Beretta, M. (ed.) 2004a. When glass matters: Studies in the history of science and art from
Graeco-Roman antiquity to early modern era. Florence: Olschki.
Beretta, M. 2004b. "Between nature and technology: Glass in ancient chemical philoso-
phy," in Beretta 2004a: 1-30.
Beretta, M. 2004c. "From the eye to the eye-glass: A pre-history of spectacles," in Beretta
200¥: 249-82.
Birnson, M., and I. C. Freestone 1985. "Scientific examination of opaque red glass of the
second and first millennia B.C.," in Barag 1985a: 119--22.
Brill, R. H. 1970. "The chemical interpretation of the texts," in Oppenheim et al. 1970: 105-
28.
Brill, R. H. 1988. "Scientific investigations of the Jalame glass and related finds," in
Weinberg 1988: 257-94.
Cool, H. E. M., and J. Price. 1995. Roman vessel glass from excavations in Colchester,
1971-85. Colchester Archaeological Report 8. Colchester: Colchester Archaeological
Trust.
Dell' Acqua, F. 2004. "Le finestre invetriate nell'antichita romana," in M. Beretta and
G. Di Pasquale (eds.), Vitrum: Il vetro fra arte e scienza nel mondo romano. Florence:
Giunti: 109-19.
Di Pasquale, G. 2004. "Scientific and technological use of glass in Graeco-Roman antiq-
uity," in Beretta 2004a: 31-76.
Fiori, C., and M. Vandini 2004. "Chemical composition of glass and its raw materials:
Chronological and geographical development in the first millennium A.D.," in Beretta
2004a: 151-94.
Foy, D. 1995a. "Le verre de la fin du IVe au VIiie siecle en France Mediterraneenne: Premier
essai de typo-chronologie," in Foy 1995b: 187-242.
Foy, D. (ed.). 1995b. Le verre de l'antiquite tardive et du haut moyen-age: Typologie-
chronologie-diffusion. Association frarn;:aise pour l'archeologie du verre, Se rencontre.
Guiry en Vexin: Musee archeologique du Val d'Oise.
Foy, D. (ed.) 2005a. De transparentes speculations: Vitres de l'antiquite et du haut moyen-age
(Occident-Orient). Bavay: Musee/Site d'Archeologie Bavay-Bagacum.
Foy, D. 2005b. "Une production de bols moules a Beyrouth a la fin de l'epoque hellenis-
tique et le commerce de ces verres en Mediterranee occidentale," Journal of Glass
Studies 47: 11-35.
Foy, D. and M.-D. Nenna. 2001. Tout feu tout sable: Mille ans de verre antique dans le Midi
de la France. Marseille: Edisud.
Foy, D., and G. Sennequier (eds.) 1991. Ateliers de verreries de l'antiquite a la periode pre-
industrielle. Association frarn;:aise pour l' archeologie du verre, 4e rencontre. Rouen:
Musee departemental des Antiquites.
Freestone, I. C. et al. 2000. "Primary glass from Israel and the production of glass in late
antiquity and the early Islamic period," in Nenna 2000: 65-83.
Freestone, I. C. forthcoming. "Pliny on Roman glassmaking," in M. Martinon-Torres and
T. Rehren (eds.), Archaeology and texts.
Gaitzsch, W. et al. 2000. "Spatromische Glashiitten im Hambacher Forst-Produktionsort
der ECVA-Fasskriige: Archaologische und naturwissenschaftliche Untersuchungen,"
Bonner Jahrbilcher 200: 83-241.
Giacchero, M. 1974. Edictum Diocletiani et collegarum de pretiis rerum venalium. Pub-
blicazioni dell'Istituto di storia antica e scienze ausiliarie dell' Universita di Genova 8.
Genoa: Istituto di storia antica e scienze ausiliarie.
Giberson, D. F. 1995. "The volcano dream," The Glass Art Society Journal: 77-84.
Giberson, D. F. 1996. "Ancient glassmaking: Its efficiency and economy," Ornament
(Summer): 76-79.
Gorin-Rosen, Y. 2000. "The ancient glass industry in Israel: Summary of the finds and new
discoveries," in Nenna 2000: 49-63
Grose, D. F. 1989. Early ancient glass: The Toledo Museum ofArt. New York: Hudson Hills.
Grose, D. F. 1991. "Early Imperial Roman cast glass: The translucent coloured and col-
ourless fine wares," in Newby and Painter 1991: 1-18.
Halleux, R. 1981. Les alchimistes grecs. Vol. 1. Paris: Les Belles Lettres.
Harden, D. B. 1981. Catalogue of Greek and Roman glass in the British Museum. Vol. 1, Core-
and rod-formed vessels and pendants and Mycenaean cast objects. London: British
Museum Publications.
Ignatiadou, D. 1998. "Three cast-glass vessels from a Macedonian tomb in Pydna," Annales
de ['Association Internationale pour l'Histoire du Verre 14: 35-38.
Ignatiadou, D. 2001. "Glass and gold on Macedonian funerary couches," Annales de
['Association Internationale pour l'Histoire du Verre 15: 4-7.
Ignatiadou, D. 2002. "Macedonian glass-working in the 4th c. B.C.," in Kordas 2002: 63-70.
Ignatiadou, D. 2004a. "Glass vessels," in J. Zahle and K. Kjeldsen, The Maussolleion at
Halikarnassos, Vol. 6: Subterranean and pre-Maussollan structures on the site of the
Maussolleion: The finds from the tomb chamber of Maussollos. Jutland Archaeological
Society Publications XV.6. Aarhus: Aarhus University Press, 181-201.
Ignatiadou, D. 2004b. "Keramike kai ualourgia sten proime ellenistike Makedonia," in St'
epistemonike sunantese gia ten ellenistike keramike, Volos 2000. Athens, 693-702.
!sings, C. 1957. Roman glass from dated finds. Archaeologica Traiectina 2. Groningen:
Wolters.
Israeli, Y. 1991. "The invention of blowing," in Newby and Painter (eds.) 1991: 46-55.
Israeli, Y. 2003. "What did Jerusalem's first cent. BCE glass workshop produce?" Annales de
['Association Internationale pour l'Histoire du Verre 16: 54-57.
Kordas, G. (ed.) 2002. Hyalos vitrum glass: History, technology and conservation ofglass and
vitreous materials in the Hellenic world. Athens: Glasnet.
Krug. A. 1995. "Romische Gemmen im Rheinischen Landesmuseum Trier," Bericht der
Romisch-Germanischen Kommission 76: 159-218.
Lauwers, V. et al. 2005. "Le verre de Sagalassos: De nouvelles preuves d'une production
locale de verre," Bulletin de ['association franraise pour l'archeologie du verre: 26-29.
Lazar, I. 2004. Rimijani: Steklo, Glina, Kamen. Celje: Pokrajinski Muzej.
Lazar, I. 2005. "An oil lamp depicting a Roman glass furnace: A new find from Slovenia,"
Instrumentum 22: 17-19.
Lierke, R. 1991. "Glass bowls made on a potter's wheel," Glastechnische Berichte/Glass
Science Technology 64: 310-17.
Lierke, R. 1993a. "Aliud torno teritur-Rippenschalen und die Spuren einer unbekannten
Glastechnologie: Heisses Glas auf der Topferscheibe," Antike Welt 24: 218-34.
Lierke, R. 1993b. "It was the turning wheel and not the lathe: Moldpressing and mold-
turning of hot glass in ancient glass vessel production," Glastechnische Berichte/Glass
Science Technology 66: 321-29.
Lierke, R.1995a. "Glass vessels made on a turning wheel in Roman times (survey)," Annales
de ['Association Internationale pour l'Histoire du Verre 13: 55-62.
Lierke, R. 1995b. "Glass vessels made on a turning wheel: Cameo glass," Annales de ['As-
sociation Internationale pour l'Histoire du Verre 13: 63-76.
Lierke, R. 1998. "The ancient glass pottery process," Annales de ['Association Internationale
pour l'Histoire du Verre 14: 198-202.
Lierke, R. 1999. Antike Glastopferei: Bin vergessenes Kapitel der Glasgeschichte. Mainz:
Philipp von Zabern.
Lierke, R. 2001. "Re-evaluating cage cups," Journal of Glass Studies 43: 174-77.
Lierke, R. 2004. "Late antique cage cups and their cutting or grinding marks," In-
strumentum 19: 18-20.
Lierke, R., and M. R. Lindig 1997. "Recent investigations of early Roman cameo glass, Part
1: Cameo manufacturing technique and rotary scratches of ancient glass vessels,"
Glastechnische Berichte/Glass Science Technology 70: 189-97.
McCray, P., and W. D. Kingery (eds.) 1998. The prehistory and history of glassmaking
technology. Ceramics and Civilizations 8. Westerville, OH: The American Ceramic
Society.
Moorey, P.R. S. 1994. Ancient Mesopotamian materials and industries. Oxford: Clarendon
Press.
Naumann-Steckner, F. 1991. "Depictions of glass in Roman wall paintings," in Newby and
Painter 1991: 86-98.
Nenna, M.-D. (ed.) 2000. La route du verre: Ateliers primaires et secondaires du second
millenaire av. J.-C. au Moyen Age. Travaux de la Maison d'Orient Mediterraneen 33.
Lyon: Maison de l'Orient Mediterraneen-Jean Pouilloux,
Nenna, M.-D. 2003. "Les ateliers traditionels d'aujourd'hui: Des modeles pour l'arche-
ologie?" in D. Foy (ed.), Coeur de verre: Production et diffusion du verre antique.
CH-Gollion: Infolio editeurs, 52-59.
Nenna, M.-D., et al. 2000. "Ateliers primaires et secondaires en Egypte a l'epoque greco-
romaine," in Nenna 2000: 97-112.
Newby, M., and K. Painter (eds.) 1991. Roman glass: Two centuries of art and innovation.
Society of Antiquaries Occasional Paper 13. London: Society of Antiquaries.
Newton, R. 1980. "Recent views on ancient glass," Glass Technology 21: 174-78.
Newton, R., and S. Davison. 1989. Conservation ofglass. London and Boston: Butterworths.
Nicholson, P. T. et al. 1997. "The Ulu Burun glass ingots, cylindrical vessels and Egyp-
tian glass," Journal of Egyptian Archaeology 83: 143-53.
Oliver Jr., A. 1970. "Persian export glass," Journal of Glass Studies 12: 9-16.
Oppenheim, A. L. 1970. "Glasses in Mesopotamian sources," in Oppenheim et al. 1970:
4-101.
Oppenheim, A. L et al. 1970. Glass and glassmaking in ancient Mesopotamia. Corning, NY:
The Corning Museum of Glass Press.
Picon, M., and M. Vichy. 2003. "D'Orient en Occident: L'origine du verre a l'epoque
romaine et durant le haut Moyen Age," in D. Foy and M.-D. Nenna (eds.), Echanges
et commerce du verre dans le monde antique. Actes du colloque de l'Association
frarn;:aise pour l'archeologie du verre, Aix-en-Provence and Marseille, 2001. Mono-
grafies Instrumentum 24. Montagnac: Monique Mergoil, 17-31.
Plantzos, D. 1997. "Crystals and lenses in the Graeco-Roman world," American Journal of
Price, J. 1991. "Decorated mould-blown glass tablewares in the first century A.D.," in
Newby and Painter 1991: 56-75.
Pulak, C. 2005. "Das Schiffvon Uluburun and seine Ladung," in Yal<;:in et al. 2005: 53-102.
Reade, W., et al. 2003. "Innovation or continuity? Early first millennium BCE glass in
the Near East: The cobalt blue glasses from Assyrian Nimrod," Annales de ['Association
Internationale pour l'Histoire du Verre 16: 23-27.
Rehren, T. 2005. "Der Handel mit Glas in der Spatbronzezeit," in Yalcin et al. 2005: 533-39.
Rehren, T. et al. 1998. "Glass coloring works within a copper-centred industrial complex
in Late Bronze Age Egypt," in McCray and Kingery 1998: 227-50.
Rehren, T. et al. 2003. "The primary production of glass at Hellenistic Rhodes," Annales de
l'Association Internationale pour l'Histoire du Verre 16: 39-43.
Rotroff, S. I. 1982. Hellenistic pottery: Athenian and imported moldmade bowls. The Athenian
Agora, Vol. 22. Princeton, NJ: American School of Classical Studies.
Sabrie, R., and M. Sabrie 1992. "Un theme decoratif des paintures murales romaines: Le
vase de verre," Revue archeologique Narbonnaise 25: 207-22.
Saldern, A. von. 1959. "Glass finds at Gordion," Journal of Glass Studies 1: 22-49.
Saldern, A. von. 2004. Antikes Glas. Munich: Beck.
Schiering, W. 1991. Die Werkstatt des Pheidias in Olympia. Vol. 2, Werkstattfunde. Olym-
pische Forschungen 18. Berlin: Walter de Gruyter.
Schiering, W. 1999. "Glas fiir das Gewand einer klassischen Kolossalstatue (Nike?) in
Olympia," Antike Welt 30: 39-48.
Shortland, A. J. 2004. "Evaporites of the Wadi Natrun: Seasonal and annual variation and
its implication for ancient exploitation," Archaeometry 46: 497-516.
Sines, G., and Y. A. Sakellarakis 1987. "Lenses in antiquity," American Journal of Archae-
ology 91: 191-96.
Smith, A. M. 1996. Ptolemy's theory of visual perception: An English translation of the Optics
with introduction and commentary. Transactions of the American Philosophical As-
sociation 86.2. Philadelphia: American Philosophical Association.
Spaer, M. 2001. Ancient glass in the Israel Museum: Beads and other small objects. Jerusalem:
The Israel Museum.
Stern, E. M. 1985. "Die Kapitelle der Nordhalle des Erechtheion," Athenische Mitteilungen
100: 405-26.
Stern, E. M. 1989. "Colored glass inlays in architectural ornament: Athens and Rhamnous,"
American Journal of Archaeology 93: 254.
Stern, E. M. 1991. "Glassworking before glassblowing," Annales de l'Association Inter-
nationale pour l'Histoire du Verre 12: 21-31.
Stern, E. M. 1995. Roman mold-blown glass: The first through sixth centuries. (The Toledo
Museum of Art). Rome: L'Erma di Bretschneider.
Stern, E. M. 1997. "Glass and rock crystal: A multifaceted relationship," Journal of Roman
Stern, E. M. 1998. "Interaction between glassworkers and ceramists," in McCray and
Kingery 1998: 183-204.
Stern, E. M. 1999a. "Ancient glass in Athenian temple treasures," Journal of Glass Studies 41:
19-50.
Stern, E. M. 1999b. "Roman glassblowing in a cultural context," American Journal of
Stern, E. M. 2000. "Three notes on early Roman mold-blown glass," Journal of Glass Studies
42: 165-67.
Stern, E. M. 2001. Roman, Byzantine, and early medieval glass. Ostfildern: Hatje Cantz.
Stern, E. M. 2002a. "The ancient glassblower's tools," in Kordas 2002: 159-65.
Stern, E. M. 2002b. "Glass for the gods," in Kordas 2002: 353-65.
Stern, E. M. 2004. "The glass banausoi of Sidon and Rome," in Beretta 2004a: 77-120.
Stern, E. M. 2005. "A la recherche de la premiere canne asouffler," Instrumentum 21: 15-18.
Stern, E. M., and B. Schlick-Nolte 1994. Early glass of the ancient world, 1600 B.C.-A.D. 50.
Ostfildern: Verlag Gerd Hatje.
Sternini, M. 1995. La Jenice di sabbia: Storia e tecnologia del vetro antico. Bari: Edipuglia.
Strano, G. 2004. "Glass and heavenly spheres: Astronomic refraction in Ptolemy's Optics,"
in Beretta 2004a: 121-34.
Thirion-Merle, V., et al. 2003. "Un nouvel atelier primaire dans le Wadi Natrun (Egypte),
et les compositions des verres produits dans cette region," Bulletin de l'Association
franraise pour l'archeologie du verre 11: 21-24.
Tolaini, F. 2004. "De tinctio omnium musivorum: Technical recipes for glass in the so-called
Mappae clavicula," in Beretta 2004a: 195-219.
Triantaphyllides, P. 1998. "New evidence of the glass manufacture in classical and Helle-
nistic Rhodes," Annales de l'Association Internationale pour l'Histoire du Verre 14: 30-
34.
Triantaphyllides, P. 2000. Rodiake Ualourgia, Vol. 1: Ta en thermoi diamorphomena dia-
phane aggeia poluteleias. Athens: Upourgeio Aigaiou: kb' ephoreia proi"storikon kai
klasikon archaioteton.
Triantaphyllides, P. 2001. "Achaemenian glass production," Annales de l'Association In-
ternationale pour l'Histoire du Verre 15: 13-18.
Verita, M. 1995. "Le analisi dei vetri," in Foy 1995b: 291-300.
Weinberg, G.D. 1959. "Glass manufacture in ancient Crete: A preliminary study," Journal
of Glass Studies 1: 10-21.
Weinberg, G.D. 1969. "Glass manufacture in Hellenistic Rhodes," Archaiologikon Deltion
24: 143-51.
Weinberg, G. D. 1972. "Mold-blown beakers with mythological scenes," Journal of Glass
Studies 14: 26-47.
Weinberg, G.D. (ed.) 1988. Excavations at Jalame: Site of a glass factory in Late Roman
Palestine. Columbia: University of Missouri Press.
Weinberg, G.D. 1992. Glass vessels in ancient Greece. Athens: Ministry of Culture, Ar-
chaeological Receipts Fund.
Wight, K. 1988. "Mythological beakers: Questions of provenance and production," Annales
de l'Association Internationale pour l'Histoire du Verre n: 71-76.
Yal<;:in, -0., C. Pulak, and R. Slotta (eds.) 2005. Das Schiff van Uluburun: Welthandel vor
3000 Jahren. Bochum: Deutsches Bergbau-Museum.
PA RT V
TECHNOLOGIES
OF MOVEMENT
AND TRANSPORT
CHAPTER 22
LAND TRANSPORT,
PARTI:ROADS
AND BRIDGES
LORENZO QUILICI
THE STUDY OF ROMAN

ROADS AND BRIDGES
The roads that the civilization of Rome created throughout the ancient world
represented a political event of universal significance. It is sufficient to note that
these structures still today often constitute the basis for the modern road system,
not just in Italy and the countries around the Mediterranean, but also throughout
Europe and the Middle East. Only the invention of modern paving techniques and
the construction of limited-access highways resulted in additions to the ancient
system. The public roads alone totaled an impressive 120,000 km throughout the
Empire. The construction of such a vast, well-organized network, and its main-
tenance over a period of 800 years across the vast and varied territory ruled by
Rome, constituted one of the foundations of public order; hundreds of millions of
travelers made use of it, along with draft animals and vehicles-carrying goods and
projecting ideas and power. The road network assisted the fusion of institutions,
races, and cultures of the most diverse origins into the fertile synthesis that was
imperial Roman culture. The roads are also a symbol of the Roman emphasis on
public order and utility, moles necessariae ("indispensable structures") in contrast
to the Egyptian pyramids, for example: regum pecuniae otiosa et stulta ostentatio
("the useless and stupid display of wealth by the Pharaohs"; Pliny, HN 36.75; cf.
552 TECHNOLOGIES OF MOVEMENT AND TRANSPORT
Frontinus, Aq. 1.16). Strabo (5.3.8) praises the road system as part of the infra-
structure of Roman imperial life: "The Romans gave particular attention to areas
the Greeks neglected: paved roads, aqueducts, and sewers capable of washing the
filth of the city out into the Tiber. Furthermore, they have also paved the roads in
the countryside, adding both cuts through hills and viaducts across valleys so that
wagons can take on a shipload." Dionysius of Halicarnassus, another historian of
eastern Mediterranean origin, echoes these sentiments (Ant. Rom. 3.67.5): "The
extraordinary greatness of the Roman Empire manifests itself above all in three
things: the aqueducts, the paved roads, and the construction of the drains."
PRE-ROMAN ROADS
This does not mean, of course, that the concept of a well-defined, paved road was a
Roman invention. Even in Italy and the west there were well-defined routes of travel
long before the spread of Roman rule. Note the ease and rapidity with which Caesar
moved his forces around Gaul during the conquest of that region, and the use of
preexisting trails and routes during the conquest of Spain and North Africa. Al-
though Egypt could rely on the Nile as a convenient route of travel, already in the
Bronze Age paved roads connected cities, sanctuaries, and necropoleis. Eratosthenes
could not have calculated the circumference of the earth with such remarkable
accuracy if he had not been able to depend on the careful measurement of the route
between Alexandria and Siene (Aswan; Cleomedes 10). The level and treeless plains
of the Tigris and Euphrates river valley facilitated land travel and fostered the earliest
development of wheeled vehicles, but the Babylonian kings also prepared special
processional roads for religious and political functions. The Assyrian kings projected
their power across a large territory by sending their troops out on routes that often
represented significant feats of engineering (Forbes 1965: 133-37; Dorsey 1997; Beitzel
1992). The "King's Highway" that can still today be followed from Damascus to
Aqaba originated in the Bronze Age, as did the "Royal Road" or "Great Trunk Road"
that follows the coastline of the Levant, connecting Egypt with Mesopotamia
(Beitzel 1992: 778-79). In Greco-Roman traditions, however, the Royal Road across
the Persian Empire enjoyed the greatest reputation among the pre-Roman road
systems (Quilici 1989b; Beitzel 1992: 781-82; Chevallier 1997: 25-29). Significantly,
this road alone benefited from an organizational infrastructure similar to what the
Romans later provided. Herodotus described it (5.52-53):
The nature of the road is as follows. All along it are royal rest stops and excellent
lodgings, and the entire road runs through inhabited and safe country.... In
all there are 111 stages with as many rest stops on the road going up from Sardis to
Susa. If the Royal Road has been measured properly in parasangs, and the
parasang is equivalent to 30 stadia, as it certainly is, then from Sardis to the royal
LAND TRANSPORT: ROADS AND BRIDGES 553
palace called Memnon the distance of 450 parasangs makes 13,500 stadia (ca. 2,600
km). And for people travelling at a rate of 150 stadia each day, just 90 days will
be needed for the route.
The road was not paved, however, and Herodotus and Xenophon (An. 1.5.7) note
the need to cross rivers at fords or on ferry boats, and the possibility of vehicles
getting stuck in mud.
An extensive road network united the Mycenaean citadels and territories of the
Greek mainland during the Late Bronze Age, including viaducts and bridges
(Hope-Simpson and Hagel 2006), but after the fall of the palace system the political
and economic factors that had supported the land routes no longer existed. By the
classical period, the dangers and hardships Theseus suffered travelling the 125 km
by land from Troizen to Athens had become the stuff of heroic legend. The worst
part of this route was not fixed until the reign of Hadrian: "Originally, they say, the
road from Megara to Corinth was made for travel by unencumbered and agile
pedestrians, but the emperor Hadrian made it wider, and it was even suitable for
chariots to pass each other when they met." (Pausanias 1.44.6). Typically, Greek
routes for land travel were paths designed for mounted or foot traffic rather than
wheeled vehicles. The one exception were the special "sacred roads" that connected
many sanctuaries with their host city or a port: Athens and Eleusis, Sparta and
Amyclae, and Elis and Olympia, for example (Forbes 1965: 140-45).
Given the opinion current even in antiquity that the Roman road system was a
particularly extensive and well-developed example of that technology, this chapter
will focus on the Roman accomplishment. Furthermore, given the enormous scope
of the system, I will focus on the roads and bridges of the Italian peninsula. The
literary sources and archaeological remains are particularly rich for this part of the
system, and a focus on Italy also allows us to trace the earliest stages of its de-
velopment and spread.
THE VIA APPIA AND THE BEGINNINGS

OF ROMAN ROAD CONSTRUCTION
The Via Appia, built in 312 B.C., represented a completely new approach to the
networking of the landscape, not only because of the technologies involved but also
in large part because of the historic and political aspects involved in the realization
of its construction. In fact, it represented the affirmation of a rational design-even
if at great cost-through areas considered very large at that time, where its route
was impeded by natural obstacles and the territories of foreign states (figure 22.1).
The design of the Via Appia can be compared to that of modem Italian auto-
stradas: the ancient road did not aim at connecting even the important cities along
Sipontum
.,. ..
~ Melapont.im
G~~
Heratle"a
-~~~~~e.o
L.Q .
___ JWOkm
_,/
Figure 22.1. Route of the Via Appia from Rome to Brindisi. (L. Quilici.)
its path (such as Lanuvium and Velitrae), but they were accessed through branch
roads. The road aimed as much as possible at the final, long-distance destination:
Capua, in Appius Claudius Caecus' first realization (132 Roman miles, 196 km)
(Castagnoli 1969: 78-96; Humm 2005: 133-44). For this reason, the route followed a
series of perfectly straight lines, taking account only of narrow passages and passes
enforced by the geomorphology of the territory traversed. There seems to have
been little concern with any technical difficulties imposed by this approach, only to
keep the route as short as possible. It is enough to mention the section from Rome
to Terracina, a single straight line of almost 90 km, where the route crossed the
Roman Campagna, the Alban Hills, and the vast area of the Pontine Marshes
(Quilici 1990). The subsequent section was no less arduous, because of the presence
of the mountains behind Terracina and Itri, the swamps and forests that sur-
rounded Fondi, and the wide Garigliano and Volturno rivers (Quilici 1989a). The
revolutionary character of this engineering undertaking is highlighted by the nu-
merous challenges that were faced and surmounted: the topographic analysis of the
territory to be traversed, the geological evaluation and surveys indispensable for
the definition of the route, the requirement for enormous hydraulic projects for
land reclamation within the region, the drainage and diversion of thousands of
minor creeks, the impressive land excavations involved in opening trenches or
cutting bedrock, the removal of excavated soil and the transport of materials for
backfill and leveling, the completion of the support terracing for the road along
both valley and mountainside, and the construction of bridges to cross large rivers.
With regard to the need for back-sights to verify the accuracy of survey over
long distances, the completion of the straight line that runs between Rome and
Terracina is technically significant. The back-sights were taken from the highest
point along the route, at Colle Pardo near Ariccia, defining a straight line of 23 km
toward Rome to the north, presumably with the temple of Diana on the Aventine
as the target, and to the south a straight line of 61 km to Terracina, with the seaward
cliff of Leana (near the Lucus Feroniae shrine) as the target, since its summit blocks
the view of Terracina behind (Quilici 1990). These two sightlines differed in ori-
entation by 5 degrees. Lewis (2001: 217-45) has reconstructed the instruments,
procedures, and remarkable accuracy of Roman road surveying.
The creation of the Via Appia fostered the foundation of the colonies at
Velitrae (Velletri) and Tarracina (Terracina) between 341 and 318 B.C. and the
associated land reclamation works around those colonies, the Ufens River (Ufento)
and the Ager Falernus in northern Campania (Quilici-Gigli 1983; Cancellieri 1990).
Whether the reclamation and road works were contemporaneous or the road came
later, it is notable how even at this early date the respective engineering projects
could be coordinated. For a distance of almost 20 miles, from Tre Ponti to
Amaseno, the substructures of the Via Appia served as one of the embankments of
the transport and drainage canal fed by the Pontine Marshes. In the late first
century B.C., Horace (Sat. 1.5.9-23; cf. Strabo 6.3.6) described a journey by pas-
senger boat along this canal, which drained water from the flat plain toward the sea
near Terracina (Quilici 1990: 55-57; see also chapter 12).
Appius Claudius' objective in laying out this route was military, focused on
facilitating the rapid deployment of armies to the theater of operations in Cam-
pania, where the Second Samnite War was at its peak. This alternative was shorter
than the tortuous itinerary of the ancient Via Latina inland along the Sacco and
Liris valleys. The opening of the Via Appia, like that of all the roads that ultimately
connected the city of Rome with the most distant borders of the empire, served as
well as a conduit of Roman culture and Roman law, surmounting not only natural
obstacles but also non-Roman ways oflife. In consequence, these roads served as a
technological metaphor for military triumph, and the magistrate who realized each
project had the privilege of giving it his name. Triumphal arches were often placed
at the beginning and end of a road, at the entrance to cities, and on the most
grandiose bridges.
Originally, the Via Appia had only a gravel surface. Livy comments (9.29.6)
that Appius Claudius "protected the road" (viam munivit), which implies some-
thing less than stone paving, and the Elogium to Appius in the Forum of Augustus
simply uses the term stravit ("surfaced"; CIL 11.1827), which is ambiguous. Else-
where, however, Livy (10.23.12) states that the first mile of the Via Appia, from the
Porta Capena to the Temple of Mars, was paved "with dressed stone blocks" (saxo
quadrato) in 296 B.C., and that in 293 B.C. the next 12 miles, from the Temple of
Mars to Bovillae, were paved silice, "with basalt paving stone" (10.47.4). Livy also
reports (38.28.3) that a century later the first mile was repaved silice, suggesting that
the first paving material had been blocks of soft volcanic tuff.
We know in general the design and structure of the roadbed, as it still for the
most part exists today as far as Capua. The paving, composed of polygonal basalt
blocks, is consistently 4.1-4.2 m wide (14 Roman feet), a width that easily allowed two
carriages to pass each other going in opposite directions (cf. Procopius, Goth. 1.14).
The pavement had a camber so that rainwater would flow to the curbs on either side.
These were made of upright slabs and framed sidewalks made of packed gravel, which
varied in width from 1.1 m to 3.0 m depending on the anticipated traffic. Upright
blocks were built into the curb every 3 to 5 m, to prevent wheeled vehicles from riding
up on the sidewalks and to assist travelers in mounting their pack animals. Milestones
were also placed along the road, to record distances and the authorities responsible
for paving or other improvements. The oldest known milestone on the Via Appia,
found at mile LIii ("53," near Mesa) dates to 253/5 B.C. (CIL 10.6838).
The road was built by setting the curb stones, then excavating the earth be-
tween them until a firm base was reached. A layer oflarge diameter gravel was then
laid in the cavity (essentially a long trench), and compacted to create the foun-
dation for the structure (rudus). A layer of finer gravel was then added (nucleus),
and the wedge-shaped bases of the polygonal basalt blocks that formed the
pavement (summum dorsum) were firmly seated in this material. The partial
renovation of the original paving of the Via Appia was recorded on milestones:
those of Nerva between the Colli Albani and the Pontine Marshes, those of Trajan
farther along by the 20-mile-long canal embankment (Di Vita-Evrad 1990), and
one of Caracalla between Terracina and Formia (Quilici 2004: 541-42). It is par-
ticularly interesting that this last milestone specifies that the new basalt pavement
replaced a worn-out limestone pavement (CIL 10.6854). This sequence has been
documented archaelogically also for the stretch through the Pontine Marshes prior
to Trajan and in the section beyond Fondi (Quilici 1990: 58-60; 2002: 139-40).
Even more interesting is archaeological evidence along the Pontine and Itri
segments of a far more ancient phase of the road, where the road was surfaced only
with gravel, in accordance with the literary sources. The first segment of the road
consisted of an embankment with a double, stepped slope 6.6 m wide, reinforced
by pilings in the marsh and contained on the sides by rectangular blocks, for a total
width of 10.5 m (Quilici 1990: 55-60). In the segment by Itri there is a graveled
foundation 9.5 m wide, in between terracing of polygonal blocks (Quilici 2002: 113,
124). These could be traces of the original road work.
The Via Appia, built as far as Capua by Appius Claudius when he was censor in
312 B.C., despite the elimination of the Samnites-to assist whose defeat it had
been conceived-was extended to the Roman colony of Beneventum after its
foundation in 268 B.C. following the Roman victory in the war against Pyrrhus.
The road was then carried on to Venosa, a Latin colony since 291 B.C. After the
final subjugation of Tarentum in 272 B.C. and the conversion of Brundisium
(Brindisi) into a Latin colony around 240 B.C., the road was brought as far as the
latter port, which subsequently took a leading role in the overseas conquests. The
overall length of the road from Rome to Brundisium was 385 Roman miles, 569 km
(Mazzarino 1969: 109-12; Quilici 1989a: 2:42-62).
No Roman road, perhaps no ancient road, had the same celebrity and im-
portance as the Via Appia. It connected the most flourishing cities of central Italy
with southern Italian ports that harbored vessels leaving for Greece, Egypt, and the
Levant. It became, in fact, not only the most important road for military expe-
ditions, but also a route of fundamental importance for commercial and cultural
interchange. By communicating with the seaport of Puteoli (Pozzuoli), the closest
protected anchorage to Rome, which lacked a proper harbor until A.D. 50, the Via
Appia formed an important link with Africa. Because these regions were the most
developed and prosperous in the Mediterranean, the network was always moni-
tored with particular care by the authorities. As its land base, the road received
special attention, and was termed "Queen of roads" (Regina viarum; Statius, Sil.
2.2.11), a paradigm for all the roads that connected Rome with the most remote
regions of the known world.
For obvious reasons, the road was always maintained scrupulously until the
very late Imperial age, and over time various improvements and restorations kept
and enhanced its efficiency. Of particular note are the great terraces and retailing
walls intended to stabilize the course of the road, and the bridges whose varied
designs provide evidence of the commitment to the preservation of the the road's
efficiency. In the countryside near Velletri, the Ponte di Mele is a primitive bridge,
possibly dating back to the original construction of the road. It was partially cut
into the bedrock and has a massive vault built of square blocks of tuff, with a span
of 3.6 m (Quilici 1991a). The Ponte Alto bridge, in the last stretch of the road
through the Pontine Marshes, was built of limestone during the first half of the
second century B.C., 6 m wide with a span of 6.1 m (Lugli 1928: 58). The most
spectacular roadwork is the viaduct at Ariccia, built at the end of the second
century B.C. to carry the road over the steep valley slope at a low gradient. The
structure (8.2-9.0 m wide at the top, 230 m long) consists of two massive walls up
to 13 m high, built of blocks of volcanic tuff laid as headers and stretchers, that
retain concrete fill. Four vaulted openings allowed the passage of transverse traffic
and the flow of the water (Lilli 2002: 235-85).
An example of improvements and repairs can be seen in the mountain gorges
near Itri, where the road is supported by impressive walls hundreds of meters long
and up to 14 m tall, built in various techniques involving stone blocks and concrete
that date from the fourth to the third century B.C. (Quilici 2002). Pullover areas,
sometimes with services for travelers and animals, lightened the traffic along the
difficult uphill stretches, such as the pass at Terracina and in the Aurunci
Mountains (Lugli 1926: 181-209; Quilici 2004). Trajan built a marvelous series of
bridges, in a refined style of elegant limestone blocks with bossing, along the canal
in the Pontine Marshes. At Tre Ponti, in modem Faiti, the Maggiore and Antico
bridges, from 6.35 to 13 m wide, remain very solid and still support the intense flow
of traffic; the third bridge has been destroyed (Quilici 1989a: 2:8-17). The complex
is approximately 45 m long, the central part composed of two segmental arches
with spans of 5.5-5.7 m, their height intentionally kept low to minimize the central
peak; the central pier is only 1.2 m thick, involving techniques that can be seen well
developed in bridges in the Veneto. Ponte Maggiore, which crosses the river
Ufente, was built with a semicircular arch with a height of 10.5 m, to allow effective
navigation beneath. The cutting across the face of Pisco Montano at Terracina,
commonly attributed to Trajan, is a spectacular accomplishment. To avoid the
steep ascent over the crags of Anxur, the limestone cliff face was cut back along the
seashore, leaving a vertical cutting up to 36 m high, and allowing passage of the
road around the obstacle just above sea level (Lugli 1926: 209-12). Large Roman
numerals cut into the artificial cliff proudly record the progress of the excavation
from top to bottom, at intervals ofio Roman feet, finishing with CXX (120).
Other bridges neatly built in stone for Trajan are preserved just before Bene-
vento (Quilici 1989a: 2:46-49). Concurrently with the magnificent series of stone
bridges that Trajan built as he renovated the Via Appia, it is notable that bridges were
constructed as well in a new technique, modeled on the brick arcades of aqueducts,
and the bridge built over the Volturno River by Domitian 15 years before to connect
the Via Appia with Cuma (see below). In order to impede erosion by the river, the
foundations and piers were built completely of stone blocks, while the elevation was
built of concrete, with a double course of bricks over the arches, which had spans of
14-15 m. There were also spectacular viaducts, such as those over the Cervaro and
Carapelle Rivers, which kept the road level while crossing the river valleys; they are
320 m and 700 m long, respectively, with seventeen and ten central arches (Ashby
and Gardner 1916; Quilici 1989a: 2:63-83; Mertens 1993). The design of the Hadrianic
bridge-viaduct Ponte Rotto on the Via Appia over the Calore north of Eclano, is a
further example of the evolution of the techniques of road and bridge construction
during the Imperial age; it was almost 200 m long and up to 13-14 m high, with eight
arches of which the central six had spans ofi3.2-13.7 m (Quilici 1996b: 274-87).
THE MAJOR ROMAN ROADS IN ITALY
Just as the Via Appia represented the main artery for southern Italy, the Via
Flaminia, built by the censor Gaius Flaminius between 220 and 219 B.C., served as
the main artery for the northern part of the peninsula. With the purpose of
colonizing the ager Gallicus, which included territories from Ancona to Rimini, he
made this latter city the outpost for the conquest of the Po Valley (Ashby and Fell:
1921; Luni 1993; Messineo 1991; Bonomi 1993; Esch 1997: 59-90). The Via Aemilia,
marked out around 175 B.C. by the consul M. Aemilius Lepidus, extended the road
system to Bologna, which he had founded in 187 B.C. This urban center was
destined to become the communications node for all northern Italy, and the road
system for the territories at the head of the Adriatic Sea branched off from it. The
Via Aemilia extended from Bologna to Aquileia, founded in 181 B.C. In 148 B.C. the
consul Sp. Postumius marked out the Via Postumia from Aquileia to Genoa,
passing through Verona and Cremona. In 132 B.C. the consul P. Popilius con-
structed the Via Popilia from Rimini through Ravenna to Padova, improving a
connection already marked out by the consul T. Annis in 153 B.C. (figure 22.2; De
Feo 1997; Cera 2000; Postumia 2000; Quilici 2000; Annia 2004; Pellegrini 2004).
E T A
UM
PAN -N ONIA
a.
"1
<
-1?
"1
I'
/
"1
_,
:.
100 200 3 0 0 km
0
Figure 22.2. The roads of ancient Italy. (L. Quilici.)
The inextricable interrelationships among military activity, colonization, trade,

and road building are clear in this region.
During the third and second centuries B.C., other roads penetrated deeply into
the former Etruscan territories north of Rome. Along the seacoast, the Via Aurelia
was built as far as the colony of Cosa, founded in 273 B.C., then on to Luni,
founded in 177 B.C., and finally in 109 it was extended through the territories of
Liguria to Genoa and Vado. Three other roads, the Viae Amerina, Clodia, and
Cassia, penetrated the heartland of Etruria, already thoroughly urbanized in the
sixth century B.C. on routes that date back to at least the same period. The Via
Cassia was particularly important; it lead through Sutri to Volsinii (re-founded
near Bolsena in 264 B.C.), then on to Chiusi and Arezzo, which gave access to the
passes over the Apennines, and finally to Florence, founded in the mid-first century
B.C. (Frederiksen and Ward Perkins 1957; Aurelia 1968; Degrassi 1982-1984, 1984-
1985; Quilici 1989b; Esch 1997: 26-58; Talbert et al. 2000: 39-43).
In southern Italy, after the extension of Via Appia to Brindisi, the Via Popilia
was built through to Reggio Calabria. Initially by T. Annius and finished by P.
Popilius, both of whom were patrons ofroads in the Po Valley (Burckhardt 1988:
41-46; Givigliano 1994: 287-318).
These roads, projected into the regions as the primary vehicle of civic life and
economic progress, came to characterize the regions to such an extent that the
name of the road could be used to indicate the territory itself: "are you departing to
[the region served by] the Via Aurelia, or to Campania?" (utrumque inAureliam an
in Campaniam abisti; Pronto, Ep. 4.4.2, van den Hout ed.). The gloss on this passage
(5,423.26) explains this phrase as "Aurelia is the region or the province." (Aurelia
terra est vel provincia). In this way, referring to the road as synonymous of the places
being crossed, the phrase per Flaminiam meant Picenum (CIL 6.1509), and still
today the region Emilia takes its name from the ancient road (Cicero, Fam. 10.30.4;
Martial 3.4). During the empire, the renovation and expansion of the road network
took place on a grand scale. It is significant that already during the reign of Au-
gustus, the inscription on his arch at Rimini contains the line "the numerous roads
ofltaly, sustained by his planning and authority" (celeberrimeis Italiae vieis, consilio
et auctoritate eius muniteis; CIL 11.365). The emperor was in many ways synony-
mous with the means ofland travel and transport, the network of roads that bound
many parts of the empire together.
DESIGN AND LAYOUT OF ROMAN ROADS
The Via Appia immediately became a model for the construction of new roads and
the renovation of existing routes. The Via Latina, renovated soon after construction
of the Via Appia, departed Rome with a series of amazing straight stretches, the first
one 15 km long, and subsequently passed diagonally across the Alban Hills and
entered the Sacco valley with decisive straight lines (Quilici 1990: 52-56; Caiazza
1995: 13-80). The straight-line configuration became the pattern, seen on stretches
of the Via Aurelia 15 km and 30 km long through the flat land seaward of Cerveteri
and Tarquinia (Aurelia 1968: 121-39 ). The 242-km route of the Via Aemilia from the
Rubicon to Piacenza could even be considered a single straight line, and in the
plains of the Cisalpine region there were many straight stretches along the Via
Popilia, Via Annia, and Via Postumia (Bosio 1991; Calzolari 1992; Esch 1997: 26-58;
Talbert et al. 2000: 39-42; Cera 2000 ). In hilly ground, at least, the straight stretch of
road makes more sense as a pathway for fast-marching foot soldiers, or as a met-
aphor of domination and control, than as a cost-effective route for wheeled vehicles
drawn by animals.
Despite their apparent simplicity, the longer straight stretches of road represent a
triumph of accurate surveying, in that they usually imply an ability to calculate the
direct line between two points out of sight of each other (Lewis 2001: 217-45). The Via
Flaminia, as early as the third century B.C., is an excellent example of a road project
that was planned and marked out rapidly over the very long distance of 213 miles (by
the Augustan calculation; cf. CIL 11.6635): 314 km from the Tyrrhenian to the Adriatic
Sea, through the territories ofVeii, the Faliscans, Umbrians, and Picenes. The route
was coordinated from south to north with a particular geological formation of the
Apennine mountain chain that does not involve the typical parallel ridges that
anywhere else make passage from sea to sea difficult. In Umbria the mountains bend
like an arch, allowing easy communication to the north. The maximum height along
the Via Flaminia, at the pass of Scheggia near Bevagna, is only 572 m above sea level,
allowing descent to the Adriatic. An alternate route through Spoleto involves an
ascent to the pass at Somma, at 869 m above sea level. The most difficult routes of
passage were the valleys of N arni and Furlo, later made accessible by Augustus' bridge
over the Nera, and Vespasian's tunnel (Grewe 1998: 129-35). The Roman road en-
gineers clearly were able not only to evaluate the details of the landscape, but also to
envision it on a very large scale and to choose the most effective routing.
What techniques were used to define these routes? We have already seen their
model in the the Via Appia. The routing of such roads, or rather their potential plan
and elevation, required a variety of skills in geometry and surveying. The Latin terms
commonly used in the context of road construction are viam innovare, or instituere,
munire, sternere, struere, which-depending on the context-might be interpreted as
"to renew," or "lay out," "build," "pave," "construct," or "surface" a road, respec-
tively. We can recognize the criteria for design and construction in the roads
branching off from Rome and in the long straight stretches on well-documented
roads such as the Via Salaria, Flaminia, Aemilia, and Postumia, or the road in the
Valle d'Aosta. The axis of the straight stretches was marked directly on the ground,
establishing on site where the road had to pass, including the intermediate destina-
tions, but without excluding the possibility of corrections or improvements. The
terrain to be crossed was evaluated: the nature of the soil and subsoil, its exposure, and
the obstacles to be surmounted or avoided. The construction of the road required not
only cuttings, bridges, viaducts, and tunnels, but also sewers, sinkholes, ditches, and
drainage canals, along with structures designed to protect the roadbed from rainfall or
ground water, falling rocks, and landslips, and-in higher elevations-from snow.
The mountain routes make particularly evident the effort of the Romans to
overcome so many difficulties, and every element was studied to facilitate safe
transit by man, beast, and vehicles. An amazing example is the Via Salaria in the
high valleys of the Velino and Tronto. The road stays well above the river, which is
very dangerous because of the floods, and passes the steep-walled gorge on a
continuous shelf, partly excavated into the bedrock with cuttings up to 18 m high on
the inside face, and supported on the outside by retaining walls 5-6 m high built of
rectangular or polygonal blocks. The road itself varied in width from 4.5 to 6 m,
with occasional stopping areas from 8 to 12 m wide. Some of these works date to the
Republic, others to the empire, particularly the later first and early second century
(Conta 1982: 337-402; Quilici 1993: 130-44). Also remarkable is the astonishing
itinerary of the Via Flaminia through the valley of Burano, the descent from the pass
over the Apennines to Metauro. The road enters the valley from the top with
extraordinary rapidity and audacity; it descends almost headlong down the slopes
then follows straight along the very tortuous gorge by means of imposing cliffside
terraces, now flanking the river that laps against the lines of spectacular retaining
walls, crossing the river time and again to cut off curves and shorten the distance,
without economizing on bridges and cuttings, until its path exits the gorge ofFurlo
(Luni 1993; Gaggiotti 2004).
The difficult alpine passes have also preserved evidence of the amazing skill and
determination of the Roman engineers. Augustus' road up the Valle d' Aosta,
through the passes of the Small and the Great Saint Bernard, is emblematic: a road of
extraordinary interest for the technical solutions employed, imposing itself on the
conformation of the very tricky mountain slopes with viaducts hundreds of meters
long, the foundations for their walls cut in the bedrock, and cuttings in the
mountainsides up to 12 to 13 m high. The blocks extracted from the cuts were used in
the supporting walls of the terraces on the valley side of the road, up to 16 m high,
polygonal or regular stone blocks facing a concrete core. The walls were reinforced
with buttresses and spur walls, placed at various heights to fit into the irregularities
of the rocks, and blind arches set up over clefts and brooks. Spectacular bridges
carried the road over gorges. The route was planned to keep above the potential
flood level of the Dora and for the most part followed the left side of the valley, to
have a favorable exposure, fundamental in an environment with weather charac-
terized for most of the year by frost and snow. The slope was kept around 7 percent;
10 percent was the maximum. The width of the pavement was 4.4-4.75 m, and
increased to more than 5 m in the most arduous passes, where it was flanked by areas
that allowed stops or permitted vehicles to pass each other. The most difficult passes
had ruts cut in the bedrock road surface to guide the wheels of the carriages, along
with corrugations to provide traction to te animal hooves. The masterpiece of the
road was the spectacular cut at Donaz, 222 m long and almost 13 m high, monu-
mentalized by an arch and by a milestone carved in the rock (Mollo Mezzana 1992).
CONSTRUCTION AND p AVING
Today, the Roman roads on most of the routes through Etruria, Latium, and
Campania typically are paved with polygonal basalt blocks, which can be quarried
in those regions. Elsewhere, considerations of practicality and economy enforced
other sources, such as limestone in the Apennine Mountains. The Via Appia was
paved with basalt as far as Capua, after which it was paved with limestone blocks
(Pagano 1978; Carfora 2001; Zevi 2004: 875-77). Paved roads such as these reveal a
constant width of 4.1-4.2 m (14 Roman feet), a canonical measure for the long-
distance or high-volume roads, but often found as well on the secondary roads
(figure 22.3). This standard dimension dominates road design from the first decades
of the fourth century B.C. (Lugli 1957: 192-93). Even if the paved area of a road had
the canonical width of 4.1 m, in particularly busy sections these dimensions could
be expanded considerably. But the real, complete capacity of the road included also
sidewalks for pedestrians, normally 3 m wide on each side, giving the road a total
width of 10.2 m. While the road pavement was seldom reduced to less than 4.1 m,
the sidewalks could even be decreased notably or even omitted.
Some paved roads, such as the Via Cassia as far as Volsinii, had a width of 3.8-
3.9 m (13 Roman feet?). Although normally around 4.1-4.2 min width, the paving of
the Via N omentana, Via Tiburtina, and Via Latina occasionally present the reduced
dimensions of 3.8-3.9 m. The Via Valeria, in its original ascent of the hill of Arsoli,
had a width of 3.6 m; this dimension seems to be related to a pre-Roman, Osco-
Italic foot (of ca. 0.257 m?), 14 of which add up to the width of this paving. This foot
could have remained in use during the middle Republic even after the diffusion of
the Roman foot system, and it could even have been preserved fortuitously during
renovation of the original paving (Quilici 1989b: 478-82; 1992: 26-30).
Livy's comments concerning the Via Appia in 293 B.C. (quoted above) suggest
that the application of stone paving must have increased starting from the be-
ginning of that century. The change may, however, have been gradual, since Livy
also notes (41.27.5) that between 194 and 174 B.C., the censors let contracts "first of
all for the roads within the city of Rome to be paved with basalt blocks, and those
outside the city to be founded on gravel, and that they be given proper curbing, and
that bridges should be constructed in many locations ... " (vias sternendas silice in
Urbe, glarea extra Urbem, substruendas marginandasque primi omnium locaverunt,
pontesque multis locis faciendos .. .).
During the Augustan Age, when the literary and monumental sources celebrate
the restoration of the great highways, other literary sources and inscriptions inform
us as well about the commitment of private individuals to the paving or repair of
local streets-such as Messalla Corvinus' road in the territory of Tusculum and the
Via Mactorina between Velletri and the sea (Quilici 1992: 31). In this period the use
of stone paving was general on both main and secondary road routes. Throughout
the Roman suburbs, meaning all the region including Monti Tiburtini and Pre-
nestini, Colli Albani, and Sabatini as far as the sea, all the roads, even those less
important, were paved during the Imperial period (Quilici 1974b).
The infrastructure for the road system was, in general, that already outlined for
the Via Appia. Nevertheless, when less firm ground had to be crossed, in particular
sandy or flooded areas, the roadbeds were supported with more powerful fill and
foundations, and by embankments with massive walls of cut stone or concrete. For
example, through the flat land beyond the Milvian Bridge, the pavement of the Via
Figure 22.3. Paving of the Via Praenestina. (L. Quilici.)
Flaminia was laid over a foundation of building debris over 1 m thick and con-
tained between walls tuff blocks or concrete (Messineo 1991: 68, 72, 122). (figure
22-4). Because the Via Tiburtina, when passing the area of the Acque Albule, was
subjected from time to time to the calcium carbonate deposits typical of those
waters, it frequently had to be repaved. Excavation revealed four separate stone
pavements over a height of 2.5 m, contained on the sides by deeply based walls 2.23
m wide, built of stone blocks or concrete.
Via Ostiense
In order to understand better how this kind of technical intervention evolved over
time, it will be useful to examine in detail some roads that went through such
phases. Among the roads best documented by literary and archaeological sources is
the Via Ostiense. After the restoration that followed the First Punic War, docu-
mented by milestone XI ( CIL 6.31585), it was characterized by magnificent straight
lines between the curves of Tiber and the hills to the east (Fischer-Hansen 1990: 35-
38). In this epoch, the road was graveled, with a width of 6.1-6.9 m framed by
blocks of volcanic tuff, and including magnificent engineering works such as the
0 ' 10 1
t---+--+------;-----t--.........--lm '
Figure 22.4. Section of the Via Flaminia between milestones VI and XXXII, from the
Tomba dei Nasoni to the Muro del Peccato on the Treia. (L. Quilici.)
viaducts of Ponte Ladrone and Guardapasso in Acilia. The former, approximately

40 m long, with 12 arches 2.2-2.3 m in height, separated by powerful, buttressed
piers of tuff blocks, crossed a small, steep-walled valley. The Guardapasso viaduct,
300 m long, with sides built of stone blocks pierced by vaulted drains, had arched
buttress walls inside the foundation mound. The Via Ostiense may have been
paved with stone during the Sullan period (early first century B.C.), when notable
bridges were built. Moreover, the viaduct that carried the road through the boggy
salt flats around Ostia can be dated to the second century B.C. This viaduct, 6.4 m
wide, was framed by walls built of four courses of regular tuff blocks and reinforced
with external buttresses. The interior fill consisted of a layer of crushed tuff 1.4 m
thick, over a layer of river gravel 0.50 m thick, on a concrete base that was sup-
ported by oak pilings (Quilici 1996a).
Via Salaria
The very old Via Salaria, which followed the left bank of the Tiber north of
Rome, was originally surfaced with gravel, but it was paved with basalt presumably
during the late Republican or Augustan period, at the canonical width of 4.1 m,
supplemented by sidewalks flanked by drainage ditches. The gravel surface was kept
even during the Imperial age on the portion of the route from Eretum past Rieti and
as far as the pass of the Alto Velino in the Apennines. The width varied from 6.o to
6.6 m between curbs built of one or more courses oflimestone or travertine blocks,
laid lengthwise, with an occasional header block keyed into the road's foundation.
Although there may have been renovations, this configuration of the Via Salaria
should date back to the period after the conquest of the Sabines by M'. Curius
Dentatus in 290 B.C, with the concession to the region of civitas sine suffragio and
then optimo iure in 268 B.C. During these years the road had to be radically ren-
ovated and made more efficient, since it was an important key to the new territorial
arrangements. A precise date can be established for the Ponte Diavolo, which
monumentalizes the entrance of the road into the territory of Trebula Mutuesca,
just after the mid-second century B.C. The Via Salaria also underwent renovations
during the late Republic, in the form of massive viaducts built after the earthquake
of 76 B.C. at Rieti to ease the difficult approaches to the Ponte Sambuco. These
ramps were about 64 m long and 5-7 m high, each perforated by a single arch with a
span of only 3.7 m. Another viaduct, built of stone blocks with a series of arches with
spans 5-6 m wide, gave access to the elevated city of Rieti (Quilici 1993, 1994).
Via Domitiana
Laid out along the coast south of Sinuessa in A.D. 95 to connect the Via Appia with
Puteoli (Pozzuoli), the Via Domitiana is a good example of a major road paved with
stone from its first creation. The road was built over a short period of time, which
is well documented by the surviving remains, and its construction was described
carefully but with poetic drama by Statius (Sil. 4.3.40-55). Remarkably, this is the
only technical description of road construction remaining to us from antiquity. The
road is depicted as a new creation by an imperial patron, through marshes and vast
forests and across large rivers such as the Voltumus. The description begins with
the initial survey work, then the excavation for the foundations of the road, finally
the construction of the roadbed itself. The associated tasks of quarrying, lumbering,
and drainage are described as well, and-significantly-even the amount of time
travelers saved by using the new road:
The emperor Domitian, vexed at the slow journeys of his people and at the plains
that lengthen a trip, banished the long windings and solidified the troublesome
sand with a new construction.... Here the dejected traveler, carried on a single
axle, once swayed back and forth on the swinging carriage pole, while the wicked
earth sucked at the wheels, and the landlocked Latin people shuddered at the
hazards of a sea voyage .... But now a journey that used to wear away a solid day is
completed in scarcely two hours ....
The first labor was to dig ditches and to cut out the borders of the road and to
excavate the ground within to a great depth; next to fill in the emptied ditches
with other materials and to prepare the interior for the high ridge of the road, so
that neither would the soil give way nor a treacherous foundation provide an
unstable bed for the paving stones under the pressure of traffic; finally, to bind the
road with blocks close-set on either side and with packed polygonal pavers. Oh,
how many gangs are at work at the same time! This group cuts down forests and
clears the mountains, another group smoothes stakes and fashions beams with
iron tools; another binds the rocks and weaves together the work with lime and
dusty pozzolana; still another group dries up the thirsty pools and draws the
smaller streams far away by their labor.
The bridge over the Volturnus, also celebrated in Statius' poem (Sil. 4.3.67-100),
was a boldly conceived structure with at least twelve arches, with elegant double
voussoir courses of bricks facing a concrete core, and piers in mixed stone and
brickwork (Crimaco 1991: 39). This bridge may represent the appearance of new
engineering techniques and may have served as a model for the bridges engineers
constructed for Trajan and Hadrian along the Via Appia.
Gravel Roads
Despite the prominence of stone-paved roads in the archaeological record, roads
surfaced with well-packed gravel, sometime with the addition of lime, were the
most common type of local and even long-distance roads. There are archaeolog-
ically well-dated examples linked to the colonial foundations of the second half of
the third and the beginning of the second century B.C. associated with the city/
harbor pairs Caere/Pyrgi, Tarquinia/Gravisca, and Cosa/Heba. These routes vary in
width from 10.2 to 13.6 m and were framed by curbs of stone blocks. The Tarquinia/
Gravisca road was characterized by a line of blocks along its central axis, either to
divide the road into lanes or to reinforce the marked camber. The Cosa/Heba road
was characterized by a central path 3.9 m wide for vehicles, flanked by prominent
sidewalks 4.4 and 5.3 m wide (Quilici 1989b: 460-61). It has been shown above that
even major roads such as the Via Appia and Via Ostiense originally were not paved,
or that stone paving was only laid on the first portion of the road, as on the Via
Salaria. Even the other major roads that radiated from Rome were paved for
varying distances, after which the surface became gravel. The Via Flaminia was not
paved beyond Narni. The Via Aemilia, Via Popilia, Via Annia, Via Postumia, and
even the Via Appia Traiana were gravel-surfaced for their entirety.
Roads did tend to be paved through cities and their immediate vicinity, where
suburbs and cemeteries were concentrated. Roads might be paved when crossing a
vicus (small village), in front of a mansio (rest house), or even in front of a par-
ticularly monumental villa or mausoleum, thanks to the euergetism of the owner.
The Via Appia Traiana, for example, was paved through the cities it crossed, and on
the bridges it was surfaced with small cobbles. The Via Popilia, normally graveled,
was paved with trachite blocks from 4 km before Ravenna and for at least 2 km
beyond it, 10 m and 9 m wide, respectively (Maccagnani 1994: 77-79). This pat-
tern can also be seen on the Via Aemilia. An intermediate stage or economical
alternative around minor population centers involved the construction of parallel

wheel tracks, 0.10-0.30 m wide and 0.20-0.25 m thick, with deposits of stones from
adjacent rivers (Ortalli 2000: 87-88; Quilici 2000: 127-30).
The foundation embankment that can still be traced for much of the Via
Aemilia has been sectioned at several points and has provided evidence even for the
initial road profile of 187 B.C. Originally, the road was laid out directly on the
ground, normally with a width of 6 m, with lateral ditches for drainage: the ground
between was excavated, and a foundation laid down of pebbles compacted with clay
or sand, then raised 1-2 m above the original ground level with alternating layers of
gravel, mud, and sand with lime. During the renovations of the late Republic, the
foundation was rebuilt in more or less with the same technique, but the width was
enlarged from 6 to 7.5 m. The side ditches and drain of the clayey subsoil of crushed
rock or brick were renovated in the Julio-Claudian period, or often rebuilt by
means of lines of broken or intact amphoras in a disposition orthogonal to the
street. In its manifestation during the Empire, the road surface, composed oflayers
of tamped gravel partly cemented with sand and lime, had a marked camber. The
road and foundation embankment reached 2 to 4 m in height and a width on the
top of 10-11 m, even 12 m during Late Antiquity (Marini-Calvani 1992; Ortalli 2000;
Quilici 2000: 130-32).
The Via Aemilia represented the standard of the graveled roads of the Po
Valley and the Veneto. Among those that we know best, the Via Sarsinate must be
noted, which descends from the crest of the Apennines into the valley of Savio as
far as Cesena, then continued through the flat plain toward Ravenna. The structure
has been well documented around Sarsina. The simple layer of gravel laid down in
the Republican period was replaced during the early Empire with a clayey em-
bankment 4.2-4.8 m wide, built over an excavated foundation. The roadway was
3.4 m wide, with a slightly cambered roadbed made of well-seated flat cobbles, fixed
with lime, gravel, sand, and grit. On the valley side there was a compacted earth
sidewalk 0.8-1.4 m wide outlined by a cobblestone border, while the uphill side was
flanked by a shallow trench 2 m wide that protected the road from water running
down the slope (Ortalli 1992: 152-56).
Numerous graveled roads are known in the Veneto, on elevated foundation
mounds and flanked by ditches. The Via Postumia, from Lagonzo north to Altino,
had a base up to 32 m wide supporting a road from 6 to 10 m wide, and up to 7 m
above ground level. The other road, called Arzeron of the Queen, north of Padova,
has a base 30-36 m wide supporting a packed gravel road 18 m wide, up to 4 m above
ground level. This foundation consists of clayey materials laid over gravel and brick
ballast that guaranteed drainage. The practice of providing drainage fields com-
posed of rocks or broken bricks beneath a clay road foundation is well documented
in all of Padania and in Veneto. In particularly swampy areas, vertical and hori-
zontal support posts were used in the sub-foundation, and planks or bundles of
sticks. Near Adria, the Via Popilia was built on a high foundation mound of rocks,
which was placed on a sub-foundation of concrete, laid in tum over thick oak
pilings (Bosio 1991; Rosada 1992; Galliazzo 1994; Basso et al. 2004; Cerchiaro 2004).
BRIDGES
No surviving bridges built of stone in Italy can be dated securely before the be-
ginning of the second century B.C., a period of grandiose and innovative con-
struction activity during which the use of the arch rapidly became commonplace
and was used as well for bridging water courses. The first fundamental reference
point appears to be the Pons Aemilius in Rome, which historical sources record as
being built with a wooden superstructure on stone piers in 179 B.C. In 142 B.C. it
received stone arches. In its final form the bridge was 135 m long, 8.9 wide, con-
structed of tuffblocks with six semicircular arches spanning 14.5-16.5 m (Galliazzo
1994: 18-20). It is likely that a structure on such a scale, over a major river, was the
result of a very long period of experiment and practical experience. Whatever the
background, by the first half of the second century B.C. bridges were being con-
structed on a large scale, completely of stone blocks held together with iron clamps,
the voussoirs and extrados of the arches forming smooth curves (O'Connor 1993).
By the second half of the century the stone-block facing was often backed by a
concrete core. It is notable that these bridges rarely have wide spans, and the piers
are thick in relation to the span of the arches in order to withstand their force.
The oldest bridges, dating back to the first half of the second century B.C., are
those already mentioned along the Via Appia, near Velletri and on the Pontine
Marshes, together with the Archi of S. Lidano. The latter bridge, 4.4 m wide and
constructed of limestone blocks with iron clamps, incorporated three semicircular
arches, the central one spanning 5.3 m, the side arches spanning 4.4 m. (Lilli 1996).
Another early example is the small bridge under Ponte Nona on the Via Prenestina,
4.6 m wide with a span of 5.8 m, which must be earlier than the structure built on
top of it around 100 B.C. (Galliazzo 1994: 40-44). From the beginning of the
second century there were also spectacular viaducts, like those already mentioned
on the Via Ostiense and, in the second half of the century, the large stone block
structure on the Via Aurelia right after the Pons Aemilia that bypassed the bogs
along the right bank of the Tiber. This viaduct was at least 65 m long and 5.9 m
wide, rose more than 5 m above contemporary ground level, and was supported by
11 or 12 semicircular arches with 3 m spans and piers 2.35 m thick. These pro-
portions resemble those of a contemporary arcade of the Aqua Marcia (Gatti 1940 ),
reminding us of the necessary interchange among engineers of information re-
garding designs and materials. The spectacular viaduct on the Via Appia at Ariccia
(mentioned above) can be compared with the Muro del Peccato on the Via Fla-
minia in the valley ofTreia. This structure, 250-300 m long and up to 9.6 m high,
climbing 45 m over its length, had two framing walls of tuff blocks laid as headers
and stretchers and a fill of gravel and earth; the body of the structure tapers in
width from 13 mat the base to 10.8 mat the level of the road (figure 22.4; Ashby and
Fell 1921: 158-60; Quilici 1989b: 499-502).
Two nearly identical bridges on the Via Aurleia in S. Marinella, south of
Civitavecchia, can be dated to just after the mid-second century B.C. They have
powerful courses of bossed limestone voussoirs, originally with iron clamps, and
an independent core of concrete; their widths are 7.2 m and 6.2 m, with spans of 4.4
m and 6.6 m, respectively (Favilla 1996). Another work of the same age is the Ponte
della Catena in Cori (Galliazzo 1994: 81), which, like some contemporary city gates,
has an arch composed of two superimposed voussoir courses. Once again, identical
construction techniques can be seen in very different types of engineering projects.
A particularly impressive bridge from this period is the Ponte Milvio, which
was built in 109 B.C. to carry the Via Flaminia over the Tiber north of Rome. It is
nearly 150 m long and 7.2-7.7 m wide; in addition to the small arches of the access
ramps, there are four major arches with spans of 18-18.5 m. An advanced char-
acteristic is that the arches are segmental (slightly less than semicircles in form),
allowing wider spans with respect to the rise in the structure. The piers, which are
prow-shaped on the upstream side, carry small open arches meant to allow the
passage of water during particularly high flood levels and thus relieve stress on the
structure. The piers and outer courses of the arches are built of travertine blocks
that bond into the tuff blocks forming the body of the structure (Delbriick 1907:
3-11; Galliazzo 1994: 32-36).
This technique of constructing arches with outside courses of voussoirs that
alternately bond with the fabric behind them was particularly widespread in central
Italy in the late second and early first centuries B.C.; examples include the viaduct of
Ariccia, the bridges at Tor di Valle and Malafede of the Via Ostiense, the No-
mentano and Mammolo bridges on the Via Nomentana and Via Tiburtina, and the
Ponte di Nona on the Via Praenestina. This last bridge, 125 m long and 10.2 m wide,
built of tuff blocks and concrete, is much like a viaduct that flattens the passage
across a small valley. The seven vaults in the central portion rested on massive piers
and spanned 6.2-7.1 m (Quilici 1974a: 363-81).
Many other bridges of this period were constructed in massive blocks of
limestone, laid without mortar and reinforced with external buttresses, such as the
Ponte del Diavolo on the Via Salaria (Quilici 1994). The Ponte del Diavolo in
Manziana, a viaduct 90 m long with a single arch about 11 m high, was built of tuff
blocks in the same technique (Galliazzo 1994: 58-59). The Ponte Manlio at Cagli on
the Via Flaminia, was built oflimestone with one arch spanning 10.8 m, after which
the structure continues as a viaduct for over 100 m (Galliazzo 1994: 157-60).
The Ponte di S. Cono in Buccino, on the boundaries of Lucania, is a very
elegant structure, although it is only 3.2 m wide and on a secondary route. The tall
central arch, spanning 17.3 m, rests on bedrock, with a smaller relieving arch on one
side, and prominent buttresses against the force of the water. The structure was
built of travertine blocks with iron clamps, the voussoirs arranged in groups of
headers and stretchers. It is interesting that, in contrast to most bridges, which
were part of a large, imperial plan, numerous inscriptions relate that this bridge
was constructed by the local municipality Volcei, and financed with public money
and private contributions (Quilici 1996b).
A particularly bold bridge along this same design was the Ponte Salario on the
Anio river, with a semicircular arch spanning 25 m, and travertine voussoirs on the
facades that bond with tuff blocks and a concrete core. Inscriptions on the parapets
record a reconstruction by Narses in the mid-sixth century, but the structure
belongs to the impressive accomplishments of the late Republic (Quilici and
Quilici-Gigli 1996: 103-12).
The age of Augustus was an extraordinary epoch for the realization of bridges,
which sometimes appeared consistently along entire itineraries, for example along
the Via Flaminia or the road up the Valle d' Aosta. The uniformity of the designs is
valuable, since it constitutes a verifiable chronological reference for engineering
techniques involved. On one side of the Apennines, the Augustan bridges of the Via
Flaminia were constructed in the traditional materials of tuff or limestone blocks,
with varying degrees of bonding, smooth or with rusticated surfaces, with a core of
dry rubble or concrete, the arches smooth even along the extrados. Examples can
be seen in the beautiful bridges on the Piccino river toward Gallese, or those of
Calamone, Cardaro, S. Giovanni de Butris, and Fonnaia on the path after Narni
toward Bevagna, or in the bridge Sanguinario in Spoleto (Quilici 1991b: 77; Maraldi
1996: 143-46; Galliazzo 1994: 179-80, 182).
The bridge of N arni is one of the most amazing works of ancient engineering:
it was originally 180 m long, 8 m wide, and up to 33 m high, with four semicircular
arches with spans of 19.6, 32.1, 17, and 16 m, framed by very strong shoulders
and piers. It was built of limestone blocks around a core of concrete; the blocks
have elegant bossing defined by cornices that highlight the architectural elements
of the elevation, for example the external curves of the arches (figure 22.5). The
span of 32.1 m is among the widest known for block-built structures in antiq-
uity. Furthermore, one of the arches, different from the others, has a vault
formed by five distinct courses of voussoirs, equidistant and covered by slabs,
similar to a technique seen in the Augustan bridges of the Valle d' Aosta (Galliazzo
1994: 183-90).
Bridges such as this represent another, spectacular approach to bridge engi-
neering: these structures are all seated directly on the bedrock, with well-decorated
square piers built of local stone, buttresses where the river beds appeared more
dangerous, and only one span. The voussoirs can have curved or polygonal ex-
trados, and the most astonishing example, at Saint Martin, has a span of 32 m. This
bridge and others, such as those at Bard, Chatillon, and Saint Vincent, have vaults
built in parallel courses of voussoirs independent of each other, following a
technique that is typical of this valley and which we find again only in the bridge at
N arni and in other Augustan bridges in Provincia N arbonensis. This technique
appears to be characteristic of a single school of engineering, but it appears again in
bridges in Africa in the second century (Lucchese 2004). Other spectacular bridges
of the Augustan period with large spans are those of Solesta and Cecco in Ascoli
Piceno, with semicircular spans of 14.5 and 22.2 m and many structural analogies
with the bridge of Narni (Pasquinucci 1975: 6-18). It is clear that there was com-
munication among the engineers who designed major public works such as roads
and bridges, but the degree of influence and the method of communication are still
difficult to define.
Figure 22.5. Surviving arch of the bridge on the Via Flaminia at Narni. (L. Quilici.)
During the Augustan age, the classical structure of the bridge persists: large
piers and arches with rounded extrados, following a design that remains normal
through the reigns of Trajan and Hadrian, for example in the bridges of the Via
Giulia Augusta (Massabo 1996; Bulgarelli 1996). After Augustus, the arch with
polygonal extrados became widespread as well, for example in the Ponte dell'A-
quoria below Tivoli (Galliazzo 1994: 64-65). This technique, in which the voussoirs
and the wall around them interlock, reaches a greater degree of refinement in the
Pons Antico on the Via Appia Pontina and the Ponte del Diavolo in Civitavecchia,
both Trajanic in date (Aurelia 1968: 91-92), and climaxes with the Pons Aelia in
Rome, whose arches span 18-4 m and hide their polygonal extrados in an elegant
cornice (Galliazzo 1994: 13-17; Ioppolo 1996).
Successful new initiatives can be noted, particularly in the late Republican and
Augustan periods, involving the construction of spectacular arches that span more
than 30 m. We can document a notable school or tradition of bridge architecture in
the limited geographical area of the Veneto, where a series of bridges dating from
the late first century B.C. to the mid-second century A.D has particular structural
characteristics: the arches are segmental, with notable spans and springing from
less massive piers than usual. The ratio between the span of the arch and the
thickness of the piers seems on average to be about 5:1 and ranges from around 4:1
to a maximum of 8:1. The flattening of the arches, calculated as the ratio of the
height above the springers to the span between them, corresponds to values be-
tween 1:3 and 1:5. The motivation for this structural innovation must have been the
desire to avoid, in flat landscapes, the tiring ascent over a bridge with one or more
high arches. Contributing factors were the generally soft sedimentary ground that
would have offered difficulties to erection of the typical massive piers and high,
semicircular arches, along with the generally slow-moving streams. In general these
bridges were constructed of dry limestone masonry with iron clamps as facing for a
concrete core. Bridges built in this manner are found in Verona, Vicenza, Padova,
Concordia, Ceggio, and Fiuminicetto along the Via Annia. The most famous is the
mid-first century B.C. bridge on the Via Postumia at Verona, 91.4 m long, origi-
nally with five arches spanning 13-18 m; two subsequently were joined to make one
span of 32 m. The foundations of these engineering masterpieces are interesting,
placed as in so many other bridges in the Veneto on very unstable swampy ground.
They were frequently founded on horizontal boards laid on piles, often bonded to
each other with transverse beams and filled with stones (Cera 1996).
We have already noted that a real technical revolution started with the con-
struction of bridges in mixed materials, beginning with Domitian's bridge over the
Volturno, which inaugurated the use of arcades with brick arches. This technique,
probably borrowed from the last big aqueduct systems built at Rome, allowed
spectacular viaducts on series of arches to cross the widest valleys while providing a
level roadbed, with great economy of cost for both materials and construction. The
mastery of the use of concrete, particularly from the reign of Hadrian through to
the Late Empire, allowed the application of local stones to the facing of concrete
cores. The final product was bridges composed of ten arches with spans of 10-15 m
(Galliazzo 1994: 63, 121-22; Bendi 1996; Fozzati and Capotti 1996; Luni 1996: 160-63;
Quilici 1997: 126-31; Olivieri 1998: 47-50; Borghi 2000).
ROAD AND BRIDGE ENGINEERING

OUTSIDE ITALY
The engineering patterns of the roads and their infrastructure seen within ancient
Italy can serve as a paradigm for what Roman engineers realized within the rest of
the Empire (Galliazzo 1994; Chevallier 1997). In Italy we have traced the appearance
of roads surfaced with paving, pebbles, and gravel; roads built on an elevated
foundation and those laid in a cutting across a ridge; rock-cuttings, terraces,
galleries, and drainage systems. Other themes were the norms for the selection of
routes, the dimensions of roads, restoration or amelioration to maintain or im-
prove circulation of traffic. While some generalization was possible, we have also
traced variations resulting from differences in topography, geology, climate, pre-
vious or contemporary history, and economic and political factors (Chevallier
1997: 107-18, 155-67, 200-70; Talbert et al. 2000: 7-93). Along these roads no
structure had more significance than the bridge, the design and construction of
which symbolizes the technological capabilities and the political will of Roman
culture. There were, however, local differences. For example, no masonry bridge
was built across the middle or lower course of the Po River, despite the fact that
sophisticated bridges, with piers of stone and wood trusses of extraordinary length,
were built on rivers such as the Rhine (at Cologne; 420 m long, on 19 piers, spans
up to 34.4 m), and the Danube (at Turnu-Severin, 1,135 m long, with 21 arches,
spans of 32.6 m) (Galliazzo 1994: 271-73, 280-84, 320-24). Outside Italy, the Iberian
peninsula possesses the largest number of well preserved Roman bridges in the
ancient world (Alarcao 1988; Galliazzo 1994: 313-19, 325-68; Mendes Pinto 1998).
Out of the many routes that made up the Roman road system, three can be
pointed out as well-known, diverse examples of the technology: the Via Domitia in
Gaul, the Via Egnatia that crosses the Balkans, and the caravan road from Aleppo
to the Euphrates.
The Via Domitia

Laid out in 116 B.C. by Cn. Domitius Ahenobarbus, the Via Domitia went through
Gallia N arbonense, from the Alpine pass of Susa to a pass in the Pyrenees, in support of
Roman military activity in Spain. Further work was undertaken in 77 B.C. by Pompey,
then later by Augustus and subsequent emperors. The main route was 284 km long-
deviations developed from time to time in response to the changing importance of
habitation centers along the route- and it quickly became one of the most important
in the northwest Mediterranean region for communications and commerce. Although
long, the route incorporated an extraordinary number of straight stretches, particu-
larly from Beaucaire to the Pyrenees. In the many stretches where the original roadway
survives, it consists of a fine-grained self-cementing surfacing 6-10 m wide on a
substantial gravel bedding, laid in turn on a rubble foundation. The total thickness
varied from 0.60 to 1.80 m. As usual, the road was paved with stone through urban
centers and some smaller stops; well preserved stretches, consisting of oblong blocks
laid on edge, can be seen at Ambrussum and Castelnau. The passes across the Alps and
Pyrenees were carefully selected to allow passage also in the winter, the track being cut
in the bedrock in some places. Numerous bridges of the Augustan period, built of
stone blocks, can be seen along the road; among the most spectacular are those at
Ambrussum, more than 200 m long with 11 arches, at Sommieres, 190 m long with 17
arches, and at Saint-Thibery and Beziers, more than 100 m long with 9 arches. Here
and elsewhere in the Gallic provinces the bridges usually date to the Julio-Claudian
period and resemble those built during the same period in Italy; for example, those at
Saint-Chamas, Bonnieux, and Vaison-la-Romaine (Laforgue 1985; Galliazzo 1994:
246-66; Talbert et al. 2000: 15-17, 25).
The Via Egnatia

Crossing the whole Balkan peninsula-Illyria, Macedonia, and Thrace-the Via
Egnatia connecting the ports of Dyrrhachium (Epidamnus) and Apollonia with
Thessalonica, then proceeded on via Cypsela in the valley of the Hebrus to the
ports on the Hellespont and to Byzantium. It was considered a continuation of the

Via Appia, which terminated at Brundisium on the other side of the Adriatic, thus
providing a single route from Rome to Anatolia and the Black Sea. As a result, the
Via Egnatia became the most important land passage connecting East and West.
The road was laid out in the mid-second century B.C. by Cn. Egnatius while he was
proconsul of Macedonia, then improved above all by Nero, and next by Trajan in
connection with the completion of the Appia Traiana, by the Severans, and by the
Tetrarchs-all with operations against the Parthians in mind. Finally, with the
foundation of Constantinople, it became the connector between the old and new
imperial capitals.
In creating the road, Cn. Egnatius made use of routes already laid out by the
Macedonian kings, but his road was structured on a unitary plan, according to the
principles of Roman road construction already described for Italy, and on a much
larger scale, reaching 1,030 km in length. The route took the road through regions
that were varied in climate and often with a geomorphology not propitious for
road construction. Despite the difficulties, the Via Egnatia traced a remarkably
regular path, with straight stretches of significant length. The road was for the most
part surfaced with gravel or pebbles, but paved with stone slabs in urban centers or
areas of significant use. The typical width was 6.0-6.7 m, but paved sections were
4.0-4.3 m wide, with sidewalks on either side; this approach is particularly well
preserved near Philippi (Hammond 1974; Talbert et al. 2000; Fasolo 2003).
Roads in Syria
A particularly important route led from Antioch on the Orantes via Beroia
(Aleppo), to the Euphrates. This is a road that passed through an environment very
different from that of the roads previously described: across deserts, following age-
old caravan routes (Beitzel 1992: 778-80). By the Roman period a roadway 4-7 m
wide was paved with pebbles and flanked by sidewalks of beaten earth framed by
dry-stone walls. Particularly striking are stretches paved with solid ashlar stone
blocks across difficult rocky or swampy terrain; after 2,000 years of erosion, these
roads resemble wide, low walls passing across a stony landscape. The construction
of the carriage way (in general 6 m wide) involved blocks laid along the line of the
road, framed by upright blocks along the margins. The foundation consisted of a
layer of mortar directly on the stone, or on pilings. These roads, along with the
associated watch towers, seem to date to the Severan period. In the territory of
Aleppo, three bridges are recorded on the route toward Zeitunak via Nebi Ouri, all
of the Severan period and about 100 m long with three arches (Cumont 1917: 19-23;
Wagner 1985; Bauzou 1989).
Although most bulk cargoes were carried by sea during the Greco-Roman period,
significant local and regional trade went by land, on roads of varying degrees of fin-
ish. Military traffic was always important, along with short- and long-distance pe-
destrian traffic. During the period of Roman domination, the circum-Mediterranean
region enjoyed greater freedom of movement by land than would be seen again
until the mid-nineteenth century. An honorific inscription of A.D. 362 found at
Ankara sums up these accomplishments: "To Julian us Augustus, ruler of the whole
world, who opened up the road through the barbarians, destroying all who opposed
him, and who traveled in only one summer season from the British sea to the
Tigris" (Cagiano De Azevedo 1939: 47).
REFERENCES
Alarcao, J. 1988. Roman Portugal. 2 vols. Warminster: Aris and Philips.

Annia 2004. La via Annia e le sue infrastrutture. Treviso: Antiga Edizioni.
Appia 1990. La via Appia. Decimo incontro di studio del Comitato per l'Archeologia Laziale.
Quaderni del Centro di studio per l'Archeologia etrusco-italica 18. Roma: Consiglio
Nazionale delle Ricerche.
Ashby, T., and R. Gardner 1916. "The Via Traiana," Papers of the British School in Rome 8:
192-271.
Ashby, T., and R. Fell 1921. "The Via Flaminia," Journal of Roman Studies 11: 125-61.
Aurelia 1968. La via Aurelia da Roma a Forum Aureli. Quaderni dell'Istituto di Topografia
antica 4. Rome: De Luca.
Basso, P. et al. 2004. "La via Annia nella Tenuta di Ca' Tron," in Annia 2004: 41-98.
Bauzou T. 1989. "Les routes romaines de Syrie, II," in J. -M. Dentzer, W. Orthmann,
C. Auge (eds.), Archeologie et histoire de la Syrie, II: La Syrie de la l'epoque achemenide a
l'avenement de l'Islam. Saarbriicken: Saarbriicker Druck & Verlag, 205-21.
Beitzel, B. J. 1992. "Roads and highways (pre-Roman)," in D. N. Freedman (ed.), Anchor
Bible dictionary. New York: Doubleday, 5:775-82.
Bendi, C. 1996. "11 ponte di Selbagnone presso Forlimpopoli," ATTA 5: 207-14.
Bonomi, P. L. 1993. "Nuove conoscenze sulle infrastrutture della via Flaminia in Umbria,"
ATTA 2: 155-66.
Bonora, M. G. 1992. "Tecnica stradale nella Regio XI: La via Regina," ATTA 1: 51-55.
Borghi, R. 2000. "Ponte del Romito sull'Arno," Orizzonti 1: 205-14.
Bosio, L. 1991. Le strade romane della Venetia e dell'Histria. Padova: Editoriale Programma.
Bulgarelli. F. 1996. "Ponti romani della Val Quazzola e del Finalese lungo la via Iulia
Augusta," ATTA 5: 231-50.
Burckhardt, L. A. 1988. Politische Strategien der Optimaten in der Spa.ten Romischen Re-
publik. Stuttgart: Franz Steiner.
Cagiano De Azevedo, M. 1939. Civilta romana: Le strade. Roma: Colombo.
Caiazza, D. 1995. Archeologia e storia antica del mandamento di Pietramelara e del Mon-
temaggiore. Vol. 2. Rome: Banca Popolare di Ancona.
Calvani, M. M., R. Curina, and E. Lippolis (eds.) 2000. "Aemilia": La cultura romana in
Emilia Romagna dal III secolo a.C. all'eta costantiniana. Venice: Marsilio.
Calzolari, M. 1992. "La via Postumia da Cremona a Verona: Aspetti topografici," in Itinera.
Scritti in onore di Luciano Bosio. Padova: Societa Archeologica Veneta, 45-58.
Cancellieri, R. 1990. "11 territorio pontino e la via Appia," Appia 1990: 61-71.
Carfora, P. 2001. "Ad Novas: Una stazione della via Appia tra Calatia e Caudium," ATTA
10: 233-42.
Castagnoli, F. 1969. "11 tracciato della via Appia," Capitolium 44: 77-100.
Cera, G. 1996. "Peculiari esempi di architettura strutturale in alcuni ponti delle Venetia,"
ATTA 5:179-94.
Cera, G. 2000. La via Postumia da Genova a Cremona. ATTA Suppl.7. Rome: L'Erma di
Bretschneider.
Cerchiaro, K. 2004. "La tecnica stradale della Decima Regio: Un contributo," in Annia
2004: 241-51.
Chevallier, R. 1997. Les Voies Romaines. Paris: Picard.
Conta, G. 1982. Asculum. Vol. 2. Pisa: Giardini Editori.
Crimaco, L. 1991. Volturnum. Rome: Edizioni QUASAR.
Cumont, F. V. M. 1917. Etudes syriennes, Paris: Geuthner.
De Feo, F. 1997." La via Postumia," in Geographia antiqua, Vol. 6. Firenze: Giunti gruppo
editoriale, 79-105.
Degrassi, N. 1982-1984. "Via Clodia e Via Cassia: Nomi e percorsi," Rendiconti della
Pontifica Accademia di Archeologia 55-56: 155-74.
Degrassi, N. 1984-1985. "La Tabula Peutingeriana e l'Etruria settentrionale tirrenica,"
Rendiconti della Pontifica Accademia di Archeologia 57: 169-90.
Delbriick, R. 1907. Hellenistische Bauten in Latium. Vol. 1. Strassburg: Trubner.
Di Vita-Evrard, G. 1990. "Inscriptions routieres de Nerva et de Trajan sur l'Appia pontine,"
in Appia 1990: 73-93.
Dorsey, D. A. 1997. "Roads," in E. M. Meyers (ed.), Oxford encyclopedia of archaeology in
the Near East. Oxford: Oxford University Press, 4.431-34.
Esch, R. 1997. Romische Strassen in ihrer Landschaft. Mainz: von Zabern.
Fasolo, M. 2003. La via Egnatia. Roma: Istituto Grafico Editoriale Romano.
Favilla, M. C. 1996. "I ponti di Santa Marinella," ATTA 5: 127-42.
Fischer-Hansen, T. 1990. Scavi di Ficana. Vol. 1. Rome: Istituto Poligrafico e Zecca dello
Stato.
Forbes, R. J. 1965. "Land transport and road building," in R. J. Forbes, Studies in ancient
technology, 2nd ed. Vol. 2. Leiden: Brill, 131-92.
Fozzati, L., and L. Capotti 1996. "Nuove scoperte in Piemonte," ATTA 5: 213-21.
Frederiksen, M., and J.B. Ward Perkins 1957. "The ancient road system of the central and
northern Ager Faliscus," Papers of the British School in Rome 25: 67-203.
Gaggiotti, M. 2004. "Le iscrizioni della galleria del Furlo," ATTA 14: 121-33.
Galliazzo, V. 1994. I ponti romani, 2 vols. Treviso: Canova.
Gatti, G. 1940. "11 viadotto della via Aurelia nel Trastevere," Bullettino della Commissione
Archeologica Comunale 68: 129-41.
Givigliano, G. P. 1994. "Percorsi e strade," in S. Settis (ed.), Storia della Calabria antica: eta
italica and romana, vol. 2. Reggio Calabria: Gangemi, 214-62.
Grewe, K. 1998. Licht am Ende des Tunnels: Planung und Trassierung im antiken Tunnelbau.
Mainz: von Zabern.
Hammond, N. G. L. 1974. "The western part of the Via Egnatia," Journal of Roman Studies
64: 185-94.
Hope-Simpson, R., and D. K. Hagel 2006. Mycenaean fortifications, highways, dams and
canals. Studies in Mediterranean Archaeology 133. Savedalen: Paul Astroms Forlag.
Humm, M. 2005. Appius Claudius Caecus: La republique accomplie. Rome: Ecole frarn;aise
de Rome.
Ioppolo, G. 1996. "Ponte Elio: indagini e restauri 1994," ATTA 5: 85-102.
Laforgue, J. 1985. Amenagement de la via Domitia en Languedoc-Roussillon. Montpellier:
Ecole d'Architecture.

University Press.
Lilli, M. 1996. "Gli archi di S. Lidano in Campo Setino," ATTA 5: 45-52.
Lilli, M, 2002. Ariccia: Carta archeologica. Rome: L'Erma di Bretschneider.
Lucchese, L. 2004, "I ponti romani di Pont-Saint-Martin, Bard, Saint-Vincent, Chatillon,
Aosta e Leverogne," ATTA 13: 7-23.
Lugli, G. 1926. Ager Pomptinus, Anxur-Tarracina. Rome: Unione Accademica Nazionale.
Lugli, G. 1928. Ager Pomptinus, Circeii. Rome: Unione Accademica Nazionale.
Lugli, G. 1957. La tecnica edilizia romana. Rome: Bardi.
Luni, M. 1993. La Flaminia nelle gale del Furlo e del Burano. Urbino: Quattro Venti.
Luni, M. 1996. "I ponti scomparsi della via Flaminia nella vallata del Metauro," ATTA 5:
151-63.
Maccagnani, M. 1994. "La via Popilia-Annia," ATTA 3: 69-105.
Maraldi, L. 1996. "I ponti di San Giovanni de'Budris e del Diavolo sulla via Flaminia oltre
Carsulae," ATTA 5: 143-50.
Marini-Calvani, M. 1992. "Strade romane dell'Emilia occidentale," ATTA 1: 187-92.
Massabo, B. 1996. "I ponti romani di Loano lungo la via Iulia Augusta," ATTA 5: 223-30.
Mazzarino, S. 1969. "L'Appia come prima via censoria," Capitolium 44: 101-20.
Mendes Pinto, P. 1998. Pontes romanas de Portugal. Lisboa: Associac;ao Juventude e Pa-
trim6nio.
Mertens, J. 1993. "Les ponts de la via Traiana dans la traversee du Tavoliere de Foggia,"
ATTA 2: 7-18.
Messineo, G. 1991. La Via Flaminia. Rome: Editrice QUASAR.
Mollo Mezzana, M. R. 1992. "La strada romana in valle d'Aosta: Procedimenti tecnici e
costruttivi," ATTA 1: 57-72.
O'Connor, C. 1993. Roman bridges. Cambridge: Cambridge University Press.
Olivieri, E. 1998. "I ponti romani di Claterna," ATTA 7: 43-50.
Ortalli, J. 1992. "La cispadania orientale: Via Emilia e altre strade," ATTA 1: 147-60.
Ortalli, J. 1995. "Nuove fonti archeologiche per Ariminum," in A. Calbi and G. Susini, Pro
poplo Ariminese. Epigrafia e Antichita 14. Faenza: Lega, 469-529.
Ortalli, J. 2000. "Le tecniche costruttive," in Calvani et al. 2000: 86-92.
Pagano, M. 1978. "Note su una localita della via Appia fra Sinuessa e Capua: 11 'pons
Campanus,' "Rendiconti dell'Accademia di Archeologia, Lettere e Belle Arti di Napoli 53:
227-34.
Pasquinucci, M. 1975. Asculum. Vol. 1. Pisa: Giardini Editori.
Pellegrini, G. 2004. "Item ab Aquileia Bononiam: Un itinerario di eta romana tea la via
Emilia ed il Po," ATTA 13: 43-63.
Postumia 2000. Tesori della Postumia. Milano: Electa.
Quilici, L. 1974a. Collatia. Rome: De Luca.
Quilici, L. 1974b. "La Campagna Romana come Suburbio di Roma antica,'' Parola del
Passato 29: 410-38.
Quilici, L. 1989a. La Via Appia. 2 vols. Rome: Palombi.
Quilici, L. 1989b. "Le antiche vie dell'Etruria," in Atti del Secondo Congresso Internazionale
Etrusco, Firenze 1985. Rome: Giorgio Bretschneider, 451-506.
Quilici, L. 1990. "11 rettifilo della Via Appia tra Roma e Terracina: La tecnica costruttiva,"
in Appia 1990: 41-60.
Quilici, L. 1991a. "11 Ponte di Mele sulla via Appia,'' Archeologia Classica 43: 317-27.
Quilici, L. 1991b. Le strade. Viabilita tra Roma e Lazio. Rome: Edizioni QUASAR.
Quilici, L. 1992. "Evoluzione della tecnica stradale nell'Italia centrale," ATTA 1: 19-32.
Quilici, L. 1993. "La via Salaria da Roma all'alto Velino: La tecnica struttiva dei manufatti
stradali," ATTA 2: 85-154.
Quilici, L. 1994. "Ponte del Diavolo sulla Salaria al confine territoriale tra Cures e Tremula
Mutuesca," ATTA 3: 119-30.
Quilici, L. 1996a. "I ponti della via Ostiense," ATTA 5: 51-79.
Quilici, L. 1996b. "Evoluzione tecnica nella costruzione dei ponti: Tre esempi tra eta
repubblicana e alto medioevo," ATTA 5: 267-92.
Quilici, L. 1997. "I Simbruina stagna di Nerone nell'alta valle dell'Aniene," in Uomo, acqua
e paesaggio. ATTA suppl. 2. Rome: L'Erma di Bretschneider, 99-142.
Quilici, L 2000. "Le strade dell'Emilia antica," Orizzonti 1: 115-38.
Quilici, L 2002. "La valorizzazione della Via Appia al valico di Itri," ATTA 11: 107-46.
Quilici, L 2004. "Santuari, ville e mausolei sul percorso della Via Appia al valico degli
Aurunci," ATTA 13: 441-543.
Quilici, L., and S. Quilici-Gigli 1996. "I ponti del basso corso dell'Aniene," ATTA 5: 103-25.
Quilici-Gigli, S. 1983. "Sistemi di cunicoli nel territorio tra Velletri e Cisterna," in Arche-
ologia Laziale, vol. 5. Quaderni del Centro di Studio per l'Archeologia etrusco-italica 7.
Rome: Consiglio Nazionale delle Ricerche, 112-23.
Quiri, P. 1992. "Scavi e restauri lungo la via Flaminia nel tratto marchigiano," in La Via
Flaminia nell'Ager Gallicus. Urbino: Quattro Venti, 327-35.
Rosada, G. 1992. "Tecnica stradale e paesaggio nella decima regio," ATTA 1: 39-50.
Talbert, R. J. A. et al. (eds.) 2000. Barrington atlas of the Greek and Roman world. Princeton:
Princeton University Press.
Wagner, J. 1985. Die Romer an Euphrat und Tigris. Antike Welt Suppl. 16. Feldmeilen: Raggi-
Verlag.
Zevi, F. 2004. "Attivita archeologica a Napoli e Caserta nel 2003," Atti del Convegno di Studi
sulla Magna Grecia, 43, Taranto-Cosenza 2003. Taranto: Societa Magna Grecia,
853-923.
CHAPTER 23
LAND TRANSPORT,
PART 2: RIDING,
HARNESSES, AND
VEHICLES
GEORGES RAEPSAET
HISTORIOGRAPHY AND METHODOLOGY
Transportation constitutes a constant activity in human life: to move, drive, drag,

pull, push, lift, carry are necessities during all moments of living and producing the
means of life. Since invention of the rope or the pole that hunter-gatherers used to
drag along their wild quarry, long before the classical period, innumerable forms of
traction and portage have been developed, and all the forms of human and animal
power have been summoned to drive them. The yoking of animals was mastered
during the long Neolithic period, and since then a great variety of methods of draft
have been put to the service of transport. This chapter will attempt to define com-
prehensively the technical framework and the principle configurations of transport
across the Greco-Roman world. Like this one, many of the more recent studies con-
cerned with the productive technologies of classical antiquity have demonstrated that
these technologies were neither insignificant nor marginal but were part of an energy
and economic dynamism whose importance must not be obscured or devalued
(Greene 1990; 1994; Schneider 1992; Amouretti and Comet 1994; Raepsaet 2002).
The analysis of land transport in antiquity is difficult, however, because the
direct sources of evidence are few and scattered, and because the topic has often
LAND TRANSPORT: RIDING, HARNESSES, AND VEHICLES 581
been approached by historiographers in a tone of controversy and polemic. Until

the 1970s the subject of transport was seldom treated in studies of economic history
or the history of technology, judged as irrelevant or even indicative of the under-
development of technology in the classical world. At that time the accepted opin-
ion, resulting from the convergence of two theories-the one economic, the other
technological-minimized and devalued the importance of land transport. A
conceptual dichotomy in the history of the Western world was proposed by Le-
febvre des Noettes (1931): the classical cultures were "blocked" by a defective system
of harnessing animals, while those of the medieval period liberated themselves and
brought progress through the use of the horse collar. This approach was taken up
and passed on in the "primitivist" vision of ancient culture in Marxist thought,
manifested in particular in the early 1960s by Moses Finley (1965) and post-Finleyan
minimalists-whose arguments have been dissected and refuted by Greene (1990,
1994, 2000). In this primitivist vision, the mediocre development of transport
technology was justified by the absence of commerce in a quasi-subsistence econ-
omy producing little surplus, by the concept of cities that consumed rather than
producing, and by the marginalization of exchange to a few, high-value-added
commodities. In the supposed absence of productive reinvestment, no techno-
logical progress was possible. The allegedly undeveloped socioeconomic structure
of the classical cultures reduced land transport for the most part to relations be-
tween the rural landscape and the nearby market town, in a largely noncommercial
pattern of exchange. Even in the twenty-first century, the orthodox primitivist view,
particularly the equation of the classical cultures with technological stagnation and
blockage and with a minimal and irrational economic system, still has its propo-
nents. To give just one example, J. G. Landels, in the revised edition of his En-
gineering in the Ancient World (2000: 170) writes: "The shortness of this chapter by
comparison with the preceding one reflects-crudely but quite accurately-the
unimportance of land transport by comparison with sea transport in the classical
world." Clearly an epistemological prejudice concerning this topic still exists. A
quantitatively very different evaluation of the technological and economic pro-
ductivity of the classical cultures within and across small and large regions appears
in the works of Laurence (1999) and Mattingly and Salmon (2001) on the one hand,
and of Fellmeth (2002) on the other.
To evaluate properly the technology of land transport in the classical world,
one must start with a close examination of the evidence in its own right and in
the context of its practical application, and an analysis of the technical fact, such as
it is, as part of the global group of preindustrial mechanical systems (Sigaut 1985;
Langdon 1986; Haudricourt 1987; Amouretti and Sigaut 1998; Raepsaet 2002; Comet
2003), and one must start too with a long historical perspective (Jankuhn et al.
1989).
A harnessing system is composed of forces and resistance, the elements of
whose operation must be understood. A technological analysis of transport in-
volving harnessed animals must therefore be the first step, and to create and validate
this analysis, evidence must be drawn from every region and historical period. The
history of technology cannot be subdivided into periods, as can battles or social

institutions. Grinding grain, drawing water from a well, pressing oil, or working
the soil constitute some of the activities and practical know-how which can be at
once extremely individual in technical style and microcultural environment, and at
the same time participate in spatial and functional frameworks of activity that
extend well beyond that specific period (e.g., Astill and Langdon 1997). Contrary to
the persistent received opinion, the history of transport cannot be divided into a
"primitive period" prior to the shoulder collar and a "modern" period subsequent
to its appearance. "Before" and "after" better fit the replacement of human, animal,
wind, and water power with that derived from nonrenewable fossil fuels, a "revo-
lution" that took place in the nineteenth century in certain parts of the Western
world, but one for which at least a third of the human race is still waiting (figure
23.1).
To carry out this attempt at clarification, it is best to use multiple and over-
lapping sources and approaches. The relevant preserved textual sources are few in
number, although an important technical literature on the subject existed in an-
tiquity (Nicolet 1996; Meissner 1999). But apart from the writers on warfare, Vi-
truvius (Fleury 1993), and some agricultural authors, these sources have rarely
survived (Meijer and Van Nijf 1992; Humphrey et al. 1998: 409-42). Epigraphical
sources provide important information, in part through accounts of construction
works (Burford 1960; 1969; Raepsaet 1984) or the epitaphs of traders and transport
personnel (Raepsaet and Raepsaet-Charlier 1987-1988), in part through texts writ-
ten or scratched on objects of daily use. Papyrus documents are particularly valu-
able (Leone 1988, 1992). We owe much to three of the traditional subcategories
of archaeological data as well (e.g., Garbsch 1986; Vierriidriger Wagen 1987; Oexle
1992): artifacts in context (chariot burials, pieces of harness), iconography, and the
diffusion patterns of commercial products. The interpretation of an image, of
course, requires caution, and a map of distribution patterns depends on the ac-
curacy of the quantification of the data.
Recourse has frequently been made to comparative ethnology, used not as
anecdotal illustration, but as a structural framework for the investigation (e.g.,
Fenton et al. 1973; Sigaut 1985; 1987; Mingote Calderon 1996; Raepsaet 2002). The
fundamental importance of the ethnology of contemporary cultures and of ethno-
history as a frame for all research on ancient technologies deserves emphasis, not
only from the materialist and functionalist points of view, but through the new
epistemological concepts and historical topics derived from them: archaeology and
history of skills and knowledge transfer, concepts of innovation and technological
choices, concepts of technical identity, technical behaviors such as socioeconomic
productions, and experimental archaeology, which helps to explain the "chaines
operatoires" (systems of production, or production lines). Emphasis is now placed
on technological choices, behaviors, and styles as social products, the cultural
manifestations of technical facts. These very interesting lines of research are bound
to the deeper, interdisciplinary inquiries carried out on the territory of traditional
villages and founded on an important collection of documentation that permits
Figure 23.1. Horses or mules pulling a wagon. Bas-relief on funeral monument at Gor-
sium in Pannonia (AE 1972: 435). (Photograph by G. Raepsaet.)
significant new conclusions. Unfortunately, the absence of data for classical an-
tiquity only rarely permits one to approach these new horizons with precision.
Nevertheless, it is clear that a harness and a carriage can be studied as much as
cultural facts as mechanical contrivances, and that the interaction of the two
approaches can only enrich our understanding. Today the form and function of a
traditional means of land transport still has importance for the economic life of
many counties around the world (Hopfen 1970; Techniques rurales 1971). The
research of Starkey (2002) concerning the development of animal traction, carried
out under the aegis of the Department for International Development in the
United Kingdom, provides a sharp reminder of this. The point of view defended
by the agronomists and animal husbandry experts of the emerging nations also
concerns us. These experts reject the idea that a simple, so-called primitive tech-
nology must necessarily be rejected and replaced by a Western, so-called modern
technology. The cultural and social values bound up with tools and technological
behaviors cannot be denied, and they can no longer be considered at the start as
something to be discarded. Today, in the dynamic of the economic, social, and
intellectual emancipation of the populations concerned, the optimization of tra-

ditional knowledge is regarded with more favor than its eradication. In the research
of the ethnologists, agronomists, and agricultural economists at work in the Third
World there are conceptual tools and methodological principles of great interest
for the historian, along with an immense reservoir of comparative documentation.
Finally, experimentation constitutes an equally interesting approach, sometimes a
decisive one, for understanding the function of a harness. Nevertheless, this re-
mains above all an interesting trial of a hypothesis, an excellent test bench, but
nothing more. The attempt is nonetheless indispensable in the study of ancient
technical operations, given the absence of equivalent ethnological data.
MECHANICS OF FORCES AND

POTENTIAL ENERGY
If mechanics can be defined as the science of force and movement, the application
of a force on a mass in order to move it in a given direction, along with the nu-
merous associated constraints, are at the heart of the problem posed by transport
(Abeels 1995; Cotterell and Kamminga 1990). From the classical period through the
nineteenth century, mechanics as applied to farming and to transport was largely
empirical, intuitive, or habitual in its applications. Even if Archytas of Tarentum
can be seen to make some efforts at theoretical analysis, and we can detect in the
works of Archimedes and Vitruvius a conscious desire to link theoretical and
practical advances (Fleury 1993; Argoud and Guillaumin 1998; Healy 1999; White-
head and Blyth 2004), one must wait for the end of the eighteenth or the beginning
of the nineteenth century for the birth of an "agricultural mechanics" as an im-
portant part of industrial mechanics. The works of Poncelet, Morin, Coulomb,
Lecamus, and Gasparin (1863) are concerned as much with the capacities of the
animal as with the draft of carriages, and they share a common desire to understand
the action, to construct the theoretical basis, and to perfect it in order to increase
the output. Even if mechanical sophistication, bound up notably with some fun-
damental changes in the sources of energy used, is a relatively recent phenomenon,
the concepts of progress in mechanics, of choice, of innovation are always-along
with the applications-of considerable importance with regard to the level of
output or economy of effort. This fundamental mechanics is inscribed in slow and
discontinuous rhythms, but it lacks neither interesting manifestations nor effec-
tiveness (Cotterell and Kamminga 1990).
Certain obvious points concern us: every mass to be moved implies in reality
some constraints or obstacles more or less difficult to overcome (Raepsaet 2002: 19-
30). The importance of the coefficients of stationary or rolling friction, the effects of
Table 23.1. Comparison of capacity for dragged loads (so-called hook measures).
Maximum dragging traction in kilograms
Type of Animal 2 hours of effort 4 hours of effort
Heavy Horses 260-290 240-270
Light Horses 180 160
Mountain Cattle 160-170 140-150
Lowland Cattle 140-150 120
slope, and their linked and interrelated effect on the effort required show that
transport is a mechanical setup that must be mastered both in its separate parts and
in its overall configuration. The same defined input of force can be sufficient at a
given location but insufficient at another, or at the same location but on a slope.
Likewise, the nature of the surface on which the haulage takes place, and of the
contact between it and the weight, can modify all the assumptions concerning a
method of transport. Easily accommodated on the smooth paving of a Roman road,
a wagonload of the same weight becomes unmanageable on soft ground. In the
same way, the sinuous line of a road can be directly linked to the need to diminish
the effects of slope. To overcome inertia and the effects of rolling friction, it is always
necessary to marshal a potential force superior to the combined forces of resistance.
Between the load (resistance) and the motivating force there is a structure that
intercepts, transmits, applies, and responds to the force exercised on the mass to
move it; in the circumstances of harnessed transport, this is the draft harness.
Did the classical cultures have a concept of the mechanics of forces? It is all a
question of theoretical knowledge as opposed to practical and empirical knowledge,
and of the relation between the two, and-underlying it all-that of the capacity
(sometimes contested) of the classical cultures to move from theory to application
or to develop a theoretical analysis of applied technologies. Gille (1980) is one of the
first to have highlighted the existence of a true technical culture in the classical
world, composed of an accumulation of research, success, setbacks, methods, prin-
ciples, and self-awareness. At Alexandria one can even speak of a school of prac-
titioners of technical arts who in a real way gave birth to technology (Argoud and
Guillaumin 1998; Lewis 2000).
Formerly, Chersiphron, Metagenes, Rhoecus, and Eupalinus appeared to be
engineers rather than architects, men of the earth and practitioners rather than
theoreticians. Archaeology shows that the architecture of archaic Greece clearly
reveals a great empirical knowledge of the application of the mechanics of force, as
well as the mechanisms for multiplying force (see chapter 9). The loss of ancient
literature is often lamented, but technical literature suffered much more than
philosophy, poetry, or rhetoric. The misunderstanding of this reality among con-
temporary historians may in great part be responsible for the stubborn misinter-
pretation of the very existence of technology in the classical world. A snippet from
Table 23.2. Comparison of the potential force and power of the principal portage
and draft animals.
Power Developed
Approximate Mean speed
traction force in kg of work Kgmls Horsepower
Animal Mean Weight
Light Horse 400-700 60-80 1m / sec 75 1
Steer 500-900 60-80 0.6-0.85 56 0.75
Buffalo 400-900 50-80 0.8-0.9 55 0.75
Cow 400-600 50-60 0.7 35 0.45
Mule (male) 350-500 50-60 0.9-1.0 52 0.70
Ass 200-300 30-40 0.7 25 0.35
Vitruvius (De arch. 10.3.7-8) is a revelation not only of the interest an ancient
technical writer could bring to bear on a problem of animal traction, but also of the
wish to carry it to level of mechanical theory. The question concerns the balancing
of weights:
When very heavy loads are carried by groups of four or six porters, the burden is
distributed evenly at the very center of the carrying poles, ... so that each porter
carries an equal portion on his shoulder.... Following the same principles, oxen
draw an equal load when their yoke is fixed (to the draw pole) by binding straps at
its midpoint. When the oxen are of unequal strength, however, and one by ex-
erting more force causes difficulty for the other, the thong is repositioned so that
one side of the yoke is longer and thus assists the weaker animal.
Humans and animals constitute the primary reservoir of energy utilized in the
movement ofloads by road, whether by portage or harnessed vehicle. This is always
the case today in the undeveloped part of the world. We have at our disposal many
comparative studies of the force and harnessed power of animals (Hopfen 1970: 12;
Raepsaet 2002: 31-64; Cotterell and Kamminga 1990: 207). In general, one can as-
sume that the tractive force of an animal is directly proportional to its weight and
corresponds to approximately one-tenth of its weight. For horses, the proportion
rises to 15 percent, and for short periods horses can exert a force close to half their
weight, or more. Animal experts have observed, among other things, that small
animals in general produce an output greater with respect to their size than large
animals, principally because their angle of traction is lower. They also have greater
energy and endurance, which compensates for their lesser weight. This is true for
both small horses (of 400 kg and a height of 1.35-1.40 m at the withers, statistics
close to those of the ancient horse) and bullocks (Hopfen 1970: 8-12). In all tra-
ditional societies bovine draft animals play an important role (Raepsaet 1998-2001).
The ox, with the capacity to exert 300-400 watts of force for six hours, and the
capability of a strong pair of oxen to exert maximum force on the order of 20-25
horsepower over 100 m, the ox constituted the principal motive force in economic
activities involving heavy burdens (tables 23.1, 23.2).
Figure 23.2. Tondo of an Attic red-figure vase representing a mule with pack-saddle (After
Raepsaet 2002: 70, fig. 5.)
In the hierarchy of economic applications, the ass family (asses and mules;
Raepsaet 1998-2001, 2002) ranks immediately after the bovines, and, in some re-
gions for a great number of applications they still constitute the main potential
source of energy, the only source of energy to supplement that of humans. Even in
the temperate parts of Europe in the mid-nineteenth century, there was discussion
concerning the respective advantages and drawbacks of the horse, mule, and ass.
The mule has three excellent advantages under harness, nearly equivalent to those
of a light horse: more constant effort, lesser fragility, and a marked docility (figure
23.2). The skeletal structure of the asses allows use of the yoke instead of the collar,
at a lesser cost in terms of both purchase and maintenance. While the price of a
well-formed mule is nearly the same as that of a country horse, the ass is notably
cheaper, which gives it an enormous role in hilly or dry landscapes. For hours it can
carry its own weight, sometimes more. The ass's capacity for tractive force is no less.
In France, some river barges were drawn by a pair of yoked asses even at the
beginning of the twentieth century.
The strength of the horse has been celebrated frequently and for numerous
reasons, but not necessarily in proportion to its real economic benefit (Raepsaet
1998-2001; 2002: 35-37) . Both Eurocentric historiography and sympathetic appre-
ciation of the animal have had a great effect. From the beginning of the tenth
century A.O., the association of the horse with the shoulder collar optimized its
potential energy, which was utilized in an ever increasing manner in northern and
western Europe. The mean harnessed power of a small or medium-sized horse is
Table 23.3. Capacities for portage.

Distance/Day Total Carried/Day
Load (kg) (km) Speed (kmph) (kg)
London Docker 90-150

Sedan-Chair Porter 70 2.5
Stevedore with Constant 40 20 2.7 800
Load (19th c.)
Stevedore with One-Way 60 11 660
Loads (19th c.)
approximately 40 to 80 kg m/sec for a task carried out over a day. At work, the mean
effort is 40-50 kg m/sec. Gasparin (1863: 3, 88-97) gives as well 45 kg m/sec for a light
horse (360 kg) on a good gravel road harnessed to a load of 1.5 tons, that is, a force
greater than 10 percent of its weight at moderate effort. For an average African horse
of 400-700 kg weight, Hopfen records a force of 60-80 kg m/sec and a power of 75
kg m/sec (which corresponds to 1 HP). The literature concerning the capacities of
the horse is immense (Vigneron 1968; Hyland 1990), along with the research
concerning the anatomy of the draft horse, particularly in the nineteenth century,
and especially in relation with application of new breeds of "carriage horses" and
heavy draft horses. This has little importance for us, because the horse was seldom
put to use for its tractive force either in the classical world or anywhere else later on,
with the exception of modern western Europe.
The capacities that animals offer with respect to carrying and dragging bur-
dens were considerable and were well known from the Neolithic onward. There was
often, however, a considerable discrepancy between the theoretical potential of the
animal and its real capacity, as a result of zoological and technological knowledge,
economic and social contexts, cultural traditions or the capacity for investment, and
the state of veterinary medicine and the medical treatments the animals received.
GENERAL CATEGORIES OF PORTAGE

AND HARNESSED TRANSPORT
Contrary to the opinion expressed by Lefebvre des N oettes and often repeated since,
the techniques and the related functional structures for locomotion and transport
in the Greco-Roman world were not generally specific to the ancient world alone,
but relevant as well to traditions present in Europe up to the nineteenth century and
still active today in some parts of the world. In fact, the sources of energy available
and consumed to move a weight-human, animal, and hydraulic-changed little
Table 23.4. Performance of animals with packsaddle.

Distance Transport per day
Load (kg) (km/day) Speed (km/h) (kg/km)
Horse (walking) 100-120 40 4 4,000-4,800
Horse (at a trot) Bo 60 8 4,800
European Mule 150-180 20-24 3-5 3,600-3,900
European Army Mule 70-80 30-48 2,500-3,500
British Army Ass 80-100 24-30 2,400
Dromedary
Middle East, normal 230 40 9,200
Middle East, short 450
distance
British Army, Egypt 170
India 300-320 32 3-4.5 8,300
Australia 400
Bactrian Camel, 250 48 12,000
Middle East
from the Neolithic period to the invention of the steam engine and, soon afterward,
fossil fuels. This stability did not stand in the way of either a great diversity of
vehicles and harnesses or multiple forms of progress, innovations, and adaptations
to the needs encountered in each type of society or preindustrial environment.
Seen, however, from a global and structural point of view, packsaddle and
portage are universal (Raepsaet 2002: 66-68 ), as is yoked traction by two bovines or
asses, at least in the major and dominant modes of economic transport on land
(table 23.3). In the restricted economies of mountainous regions, portage remains
important even today. On certain steep and winding paths, only portage on the
human back is feasible. Likewise, in many domestic and rural activities, in ports
and in workshops, occasional or professional portage by humans is the best solu-
tion. Ethnography gives us an idea of the capacity. In equatorial Africa, 25 kg can
be carried on the head over 25 km in one day. In the mountains, some Tibetan
porters carry up to 100 kg on their backs, sometimes with the assistance of a staff or
a carrying pole. A pair of porters can make use of a suspension pole, a litter, or a
stretcher, as in the case of transporting game that has been killed. In the classical
cultures there were professional porters, known as saccarii in Roman ports, for ex-
ample, and various forms of the pack bag (pera), pouch ( mantica), and other means
of assisting carriage, all the way up to the comfortable sedan chair, or litter (lectica).
Curiously, although sedan chairs were a common feature of Roman urban life, very
few visual representations have survived.
Portage by animal, and, in its developed form, with the packsaddle (Raepsaet
2002: 68-72) was of fundamental importance for the transport of burdens that
could be subdivided (table 23.4). In economies with brief transactions and a low
volume of exchange, for example from a small rural estate to a nearby market town,
the ass, mule, and to a lesser extent the horse transported between 100 and 200 kg
of goods, according to the terrain and the climate. Columella (Rust. 6.37) terms the
mule "better suited to the packsaddle" (figure 23.3). Dependence on the packsaddle
is still common today in the mountains of Arcadia, in the Troodos Mountains of
Cyprus, and elsewhere in the world. The miserable life of the badly treated ass is the
subject of numerous jokes, traditional sayings, and fables. There were several var-
iations on the packsaddle, from the soft double sac, the mantica with two pouches
that is balanced on the shoulders or the back, to the complex frameworks of wood
covered with cloth or fitted with straps (sagma, stratum). The packsaddle could
even be adjusted for the transport of persons (astraba). The animal thus saddled is a
porter led by a walking merchant, the typical means of transport in long convoys
devoted to foodstuffs. The guild of the utricularii, for example, comes to mind,
which in the Gallic provinces carried and distributed skins of wine and other
foodstuffs by means of pack animals, undoubtedly mules (Deman 2002).
From the Neolithic to our day conveyance involving the harness and the yoke
has been universal, along with its multiple applications in every sphere of activity
involving tractive power: plowing, towing, milling, hauling, and carriage. The ve-
hicle with a draw pole pulled by two horses is traditionally considered the standard
form of harnessing in the Greco-Roman world (Lorimer 1903; Roring 1983; Crouwel
1992; Raepsaet 2002: 72-154). It is the means most prominent in the iconography,
but the least obtrusive in the economic reality. In its military, ceremonial, or sport
version, the small cart with two wheels (biga), low box, draw pole (temo ), and double
yoke (iugum) is well documented in the texts and iconography, but it represents
poorly the everyday reality ofland transport. The structural basis for the traction is
the rigid joining of the draw pole and yoke; the latter is placed on the animals' backs
and held in place by a girth strap or flexible band which exerts its pressure on the
lower portion of the neck, in front (cf. Homer, Il. 24.268). It was the understanding
of this system for the transmission of force that elicited the negative evaluation
by Lefebvre des Noettes. On the basis of a restricted, poorly understood icono-
graphical corpus he concluded that the girth strap strangled the horse and that the
yoke damaged the withers, severely restricting its tractive power. A new reading of
the ancient documents, accompanied by an important set of experimental data,
permitted Spruytte (1977) to propose a new interpretation of the tractive power of
equids in antiquity. According to him, the yoke rested on the curve of the back,
behind the shoulders, and was attached to a girth strap that exerted force on the
shoulder joints and not on the throat. When correctly understood, this harness
could be classified as an "abbreviated breast harness" (bricole ecourtee). The ar-
rangement for exerting force is mechanically correct. Experiment showed that loads
of a ton on a small four-wheeled wagon posed no problem for two horses, even on
difficult terrain. Supplementary, "roped" horses (funales) could be added to a two-
horse arrangement, outside the team yoked to the draw beam, to create the three-
and four-horse teams that the Greek vase-painters were so fond of. There are some
exceptional cases as well, of multiple pairs of horses harnessed along an extended
Figure 23.3. Mule cart with amphoras, detail of Theban Cabiric vase. (After Raepsaet
2002: 172, fig. 94.3.)
draw beam, but these are of lesser technological importance. The problem of the
harnessing of horses with a yoke and breast strap, now well understood for the
Greek and Roman world, appears in nearly identical form in the Near East and
Egypt, where the very rich iconographical evidence presents war and parade chariots
(Littauer 1968; Littauer and Crouwel 1979; Crouwel 1981; Rommelaere 1991).
The only basic harnessing system that had a true economic significance for the
transport of merchandise was the yoking of steers or oxen, asses or mules. Varro
(Rust. 1.20, 2.8) summarizes the situation: "Some use asses, others steers or mules"
(alii asellis, alii vaccis ac mulis utuntur); "For all vehicles on the roads are drawn by
mules yoked in pairs" (hisce [= mulis] enim binis coniunctis omnia vehicula in viis
ducuntur). Concerning asses, Columella (Rust. 7.1.2) writes: "The ass pulls sig-
nificant loads on a wagon" (asellus ... non minima pondera vehiculo trahat). Here
as well, the equipment consisted of a small two-wheeled cart, draw pole, and dou-
ble yoke, but the box, often raised on props or axle-blocks (Greek: hamaxopodes;
Raepsaet 1984), could take a variety of forms adapted to the varied needs of trans-
port and cargo. In this case, the iconography is restricted, but ethnographic evi-
dence is useful in supplementing our evidence, since the size of the ass family and
bovines and their capacity for traction under the yoke is more or less a constant and
has always been utilized in certain regions. The morphology and anatomy of the ass
and of the mule are favorable to providing a good seat for the yoke. As for the ox, it
carries the yoke on the horns or nape of the neck-the advantages of one or the
other arrangement were discussed in the same terms by both Latin agronomists and
those of the nineteenth century: "For the yoke fits better on the neck, and thus this
manner of yoking finds the most approval. But in some areas there is the odd habit
of tying the yoke to the horns" (iugum melius aptum cervicibus indicat. Hoc enim
genus iuncturae maxime probatum est. Nam illud, quad in quibusquam provinciis
usurpatur, ut cornibus illigetur iugum; Columella, Rust. 2.2; cf. Palladius, Ag. 2.3) .
The heavier the loads, the more there was a tendency to make use of oxen. The
transport of large stones for construction was reserved for oxen, yoked in line or
fanned out, both at the Pentelic quarries in 450 B.C. (Raepsaet 1984) and at the
quarries of Carrara in 1920. At this level of load-blocks of eight or ten tons were
Figure 23.4. Yoked cattle in Portugal, ca. 1910. (Photograph by J. Malvaux.)
not rare-the problems of transport are concerned more with the solidity of
the frame or wagon and with the terrain than with the available tractive power.
Argument over the relative importance of the two-wheeled cart or the four-wheeled
wagon is pointless. A situation of broken terrain, with a need for stability and
braking capacity favors the wagon, but the cart had the advantages of lightness,
maneuverability, and speed (figures 23.3-4).
In both the iconographic and textual evidence, the choice of vehicle appears
linked to the means, needs, functions, and circumstances of transport: the hamaxai
have two wheels in the narrow streets of Athens and the vineyards of Corinth
(Raepsaet 1988), but four-wheeled wagons (tetrakukloi) were used to carry column
and architrave blocks up to Epidaurus (Burford 1969) or down to Eleusis. The
situation appears the same at Rome (Adam 1984: 44-53). On a fresco at Boscoreale,
two-wheeled carts are shown at a construction site (Adam 1984: 216). But the
winegrowers of Langres used a solid four-wheeled wagon (Molin 1987-1988). As
for travelers, above all during the Roman imperial period, they had at their disposal
carriages of various designs and arrangements, with either two or four wheels
(figure 23.5).
We must examine in more detail the problems involved with heavy transport
across the Isthmus of Corinth on the diolkos, a paved road by means of which, from
archaic Greece through the Roman Empire, ships or their cargoes were hauled
between the Saronic Gulf and the Gulf of Corinth. This road is a paradigmatic
example that summarizes all the technical difficulties of heavy transport in antiq-
uity and all the solutions brought to bear to solve them (Raepsaet 1993): the length
of the route on land (6-7 km), slopes of up to 6 percent, the weight and volume of
the loads to be transported (even the smallest ships weighed at least 10 tons, while a
trireme surpassed 30 tons), haulage of the boats from the sea to dry land-always a
Figure 23.5. Ox-drawn wagon with large barrel, relief sculpture from tomb monument,
second or third century A.D., Langres. (Photograph by G. Raepsaet.)
difficult task, even for boats with an abbreviated keel. The course of the road has
been badly disturbed by the modern Corinth canal, but the traces preserved, the
arrangements of the quay at the entrance to the Gulf of Corinth, and portions of the
paving with ruts at Loutraki provide enough information to allow a tentative
reconstruction of the transport procedures and the mechanical devices used.
The paved road, like the quay, was constructed with large facing blocks. Be-
tween 3.40 m and 5.60 m wide, the paving is marked with a complex pattern of ruts
caused by wear, in particular in the area of a paved ramp that interrupts the course
of the road 600 m from its west end. Both downhill and at the approach to this
inclined plane, alongside the principal ruts, which are approximately 1.5 m apart,
there are numerous traces that make clear important maneuvers by the carts in-
volved. Grooves cut by the friction of cables are also visible, sometimes nearly
cutting entirely through blocks in the structure. Analysis of the contour lines of the
topography is the key element in the interpretation of the techniques of the rolling
stock put to use. The first 800 m from the coastline along the Gulf of Corinth
correspond to the zone of beach and marine sands, 12 m deep. The slope here is
modest, around 1.5 percent. At the point just below where the ramp is located the
slope increases to 2.5 percent for a few hundred meters, then to 3.5 percent above it,
then very quickly to 6 percent before reaching the summit at a level area 4 km east of
the Gulf of Corinth, at a height of 75-80 m above sea leval. The descent to Kalamaki
was 2.5-3 km long. By taking account of the constants of topography, archaeological
remains, and naval and mechanical technology, it can be seen that the pick-up of
the boats hauled out on the quay was accomplished with strong vehicles, or with a
line of wagons, or by means of multiple pairs of wheels without chassis, in each case
pulled by oxen, doubtlessly more than 20. Along the first stretch of road where the
slope was gentle, the ruts belong to the same type of trace left by traditional carts
elsewhere in the ancient world. After the ramp, where the slope increases signifi-
cantly, the profile of the ruts changes, and they become regular, smoothed, cali-
brated. This was a hodopoiia, a special road track, practically a "railway," executed
with the greatest care in order to lower the coefficient of rolling friction to the
I
I I
I ' I
-- --~
-~:"'~j- ·- ··----~P:~-
~~~-
L . ' - - -
2M
0 1 2 :J 4 ; M
Figure 23.6. Reconstruction of transport carts and windlass location at the Diolkos. (After
Raepsaet 1993: 254, fig. n.)
lowest possible level and to tackle the effect of the increasing slope in the best
conditions. In this way the negative effects of slippage and the possible problems of
guidance were eliminated. The required effort for the heaviest cargoes remained
considerable, however, up to 20,000-35,000 Newtons. The operation, in conse-
quence, must be understood to have involved two phases: a classic haul-out, fol-
lowed by a change in the lading through the transfer of the boat to another support
by means of the ramp. For the most difficult portion of the route, a battery of
capstans would have been necessary; an analysis of the wear lines in the blocks of the
ramp and thus the directions of pulling, even allows location of their positions. A
frame or support mounted on wheels calibrated to the dimensions of the grooves
would have served for the means of transport (figures 23.6-7) .
This complex originated without doubt in the sixth century B.C. In fact, the
functioning of the diolkos is the product of a series of activities that had been
relatively common since the beginning of the sixth century B.C., both in the removal
and transport of monoliths in quarries and workshops and in activities associated
with harbors. The analysis of this project highlights the exceptional quality of the
interrelated chain of technologies put to work at Corinth. One notes all at once the
scope of the installations, the precision of the track way, the knowledge of the effects
of slope, the remarkable care expended on the hodopoiia, and the sagacity shown in
the selection and interplay of the techniques and means of hauling. Did Sicilian
wheat destined for Athens or amphoras of wine and oil from the Aegean on their
way to Greek cities in the west all pass along the diolkos, or around the south end of
the Peloponnese? The literary sources are silent on this subject, but a tax on its use,
and the rental of infrastructure for economic and commercial activities, could have
been advantageous for the city, along with the military application.
Figure 23.7. Military baggage on ox-drawn wagons, Column of Marcus Aurelius, ca.
180-193 A.D., Rome. (Photograph: Fototeca DAI Rome no. 55.1252.)
One of the oddities-and one of the most spectacular innovations-in the

technology of transport during the Roman imperial period was the application of
traction by means of a pair of shafts, or "single-headed" harnessing, in Italy and the
northwest provinces (Raepsaet and Rommelaere 1995). This type of harness ap-
peared in northern Italy and northern Gaul in the first century A.D. and revised the
structure of harnessing by introducing draft with a single animal. The texts are
silent on this subject, but the iconography is rich and shows a whole range of two-
wheeled carts for traveling or mixed use drawn by an ass or a horse with the aid of
a small yoke (also called jouguet, "little yoke" or "single yoke") placed on the neck
(Raepsaet 1982; Molin 1987-1988). The use of this harness seems to have been
restricted to the Celtic and German provinces. The technological value of the in-
novation has been contested, exclusively on the basis of iconographical evidence, by
Lefebvre des Noettes for the same reasons as the alleged strangulation of the har-
nessed draft animal. The discussion has now been completely revised, following the
discovery of a complete and well-preserved single yoke, along with portions of the
harness, in a second-century Roman well in Bade-Wurtemberg (Raepsaet 2002:
266-67). These artifacts spawned an experimental research project led by the Uni-
versity of Brussels.
Draft by means of a single yoke is described by Pliny (HN 18.296) and Palladius
(Ag. 7.2.2-4) with regard to the Gallic reaper (vallus), which the animal pushed
from behind (Raepsaet 2002: 266-67). Among its other advantages, then, this
harness could be used either for pulling or pushing, unlike the double yoke or
Figure 23.8. Wooden single yoke bar, molded copy and reproduction used in the Ro-
chefort Experiment. Found in a second-century A.D. well at Pforzheim. (Photograph by
G. Raepsaet.)
shoulder strap. The iconographical evidence, which had caused such perplexity
among archaeologists and historians of technology (Vigneron 1968: 129-30), can
now be clearly explained. The function of the single yoke is bound up, structurally
and mechanically, with lateral "disks" that have the form of rounded swivel joints
attached to the yoke. Leaning against the shoulder and designed to adjust closely to
the morphology of the animal, these joints prevented the rubbing of the girth strap
and transferred the support and push of the shoulders. This remarkable innovation
relies on a knowledge of the anatomy of the draft animal, the structure of forces of
traction and their transmission, as well as the balance of the harnessing technique.
This harness, which bears some similarities with the "yoke saddles" of the Near
East (Littauer 1968), is well adapted to the use of one animal for draft, particularly
the ass or the mule. This arrangement has been tested on an ass pushing a recon-
structed vallus, and it demonstrated remarkable efficiency in the task of harvest-
ing grain on a slope and in difficult terrain (figures 23.8-9) . The lateral "disks" are
constantly in use during traction; they respond to all the movements of the
shoulder and fore limbs, without hampering them, while transmitting their pro-
pulsive force. In summary, this jouguet with swivel joints represents an interesting
compromise between high and low traction points and prefigures the medieval
shoulder collar specifically designed for the horse.
The diversity of ancient cartage (Tarr 1969; Roring 1983; Treue 1986; Rad und
W agen 2004) reflects the variety of social contexts, economic needs, and geographical
realities. Qualitative progress in the construction of these vehicles is not linear but is
nevertheless perceptible during the Empire, above all in the evidence concerning the
travel coach. In both the Greek and Roman world the transport coach is relatively
Figure 23.9. (Left) Experimentation with single-yoke harness at Rochefort (Belgium).

(Right) Detail of single-yoke harness on mules, possibly at a milestone, Roman relief,
Arion (Belgium). (Photographs by G. Raepsaet.)
simple and adaptable (transformable and polyvalent) in comparison with the pa-
rade, war, or racing chariot. Official inscribed accounts at Eleusis (JG 22 1673),
Epidaurus, and Didyma, however, show that wheelwrights on some occasions cre-
ated heavy versions mounted on framework and disk wheels.The Roman Empire
benefited from the contributions of Celtic technology (Schonfelder 2002), as the
specialized Latin vocabulary tends to show. A rich terminology for land vehicles
borrowed from Celtic-including henna, carpentum, carruca, carrus, cisium, covinus,
currus, petoritum, pilentum, plaustrum, sarracum, and triga-is indicative of the
growing functionality of Roman vehicles, above all for the movement of people. One
must nevertheless point out that at one and the same time and in the same region,
the lovely cart mounted on large 12-spoked wheels of the House of Menander at
Pompeii could coexist with the rustic wagon of the Villa Regina farm at Boscoreale,
which had solid wheels, an axle that turned with them, and a rudimentary box
(Raepsaet 2002: 231). If the platform, box, and draw pole display carpentry skills, the
construction of the wheel was generally a more specialized skill, and one that reveals
regional traditions. Every design was present in the Greco-Roman world, from the
primitive solid, one-piece disk wheel to the large 12-spoked wheel with a metal rim
fastened at only one point. But there were also some common points, in that the
wheel disk was always in the same plane-perpendicular to the surface on which it
rolled. Rim offset and wheel camber did not appear in the west until the end of the
medieval period. The design of the typical Greek wheel-the "cross-bar wheel"-is
very original: four or six rim segments fixed on a central axle by means of per-
pendicular cross-bars (Raepsaet 1988), despite the fact that the Middle East had
known the wheel with radial spokes for millennia and the Roman world of the
Republic also rapidly developed this interesting technology. Another characteristic
of ancient wheels is the thickness of the hub, and its projection. One must, however,
avoid judging the quality of a wheel only on the basis of its degree of technical
sophistication, size, or the number of spokes. It is also a question of competence,

context, and profitability of the technological investment. Ethnographic evidence
from southwestern France and northern Portugal includes transport carts with two
disk wheels (either solid or with small cut-outs) that could carry more than a ton of
cargo without difficulty.
The question of whether the two front wheels of a four-wheeled wagon were
mounted so they could turn to one side or the other has now been solved. It is clear
that from the Hallstatt period (Piggott 1983; Vierri:idrige Wagen 1987), in the early
first millennium B.C., the front wheels of a parade wagon were capable of pivoting
beneath the box, thanks to a suspension working around a kingpin. Nevertheless,
many representations of Roman wagons, if interpreted literally, show front wheels of
a size that would have made such turns difficult, although not impossible. The size of
the wheel made it impossible in many cases to execute a sharp turn, but the distance
between the large-hub of the wheel and the platform of the wagon was sufficient to
allow some pivoting of the front axle. Four-wheeled carts with a completely fixed
front axle still exist in many parts of the world; they slip around the turn, carried by
the team. There is no problem with maneuvering, but the wheels are worn out
rapidly. Both the fixed and the turning axle coexisted, their contemporaneity more a
question of quality of workmanship than of chronological evolution.
CusToMs, CONTEXT, AND CosT
Can we reconstruct the real situations and the practical applications of the trans-
port of people and goods on land (Raepsaet 2002: 178-275)? From the rural estate to
the construction site, from the harbor docks to traveling markets, land transport
was omnipresent in antiquity, and multifarious. Neither marginal nor poorly de-
veloped, this technology was adapted to many aspects of the public and private
economy, in domestic contexts and in commercial transactions, complementing
the transport by sea, river, and canal. The vehicle drawn by animals was already part
of the basic agricultural tool set on Hesiod's farm in Boeotia. By 600 B.C., long files
of oxen yoked together in pairs carried boats over the diolkos at Corinth on trolleys,
or enormous architectural members weighing many tons to construction sites at
Didyma, Epidaurus, and Eleusis (figure 23.10).
The capability for transport is in practice more a capacity for organizing means
of transport than of technological limitations. The small hamaxai of Athens carried
everything: vegetables, wine amphoras, oil jars, three or four persons, a marriage
party, or a funeral bier (Lorimer 1903). In Roman Apulia, caravans of asses with
packsaddles formed long lines on their way to the sea (Varro, Rust. 2.6.5): "From
Brundisium and the surrounding part of Apulia they transport oil or wine, likewise
grain and anything else to the sea by means of asses with packsaddles" (asellis
Figure 23.10. Transport of marble on wagons, Carrara, ca. 1910. (After Raepsaet 2002: no,
fig. 40.)
dossuariis). Even if one misinterprets the ancient economy as having consisted only
of food production on the estate and export from it, the problem of transport
constituted a constant concern: manure on the farm, tree trunks in the forest, parts
for mills or presses, cereals and fodder, various types of ceramic vessels, all carried
for short or long distances on paved roads or winding paths. The ancient agron-
omists testify to this; for example, Varro (Rust. 1.16), for whom the choice of a farm
is bound up with the necessity "to export our produce" (fructus nostros exportare),
to which end it is advisable to have access to "roads and rivers" (viae et fluvii).
Apart from the special cases of racing or ceremonial chariots, the movement of
people often took place by the same means and methods as the transport of mer-
chandise, and sometimes-as the iconography shows-the two were carried to-
gether.
In the Roman imperial world the sophistication, functionality, specialization,
indeed the professionalism of transport and travel began to increase (Kolb 2000;
Laurence 1999). Mounts and vehicles could be rented, with or without driver. In
order to take care of the most varied needs of the state and of some private indi-
viduals, Roman culture developed some true traveling wagons and carts adapted to
special circumstances, needs, or tastes, from the rather rustic raeda to the com-
fortable and expensive "sleeping wagon" (carruca dormitoria) on four wheels, from
the light cisium to the covered carpentum on two.
Armies, of course, were important consumers of transport services. The Ho-
meric chariot and the cavalry troop represent one aspect, but the logistics of supply
and the transport of baggage are more important problems. The settled encamp-
ment in time of peace, the deployment of troops, and maneuvers in time of war all
owe much to the efficiency of the means of transport. The examples are numerous:
the hundreds of baggage carts that took part in Xenophon's march to the sea (An.
3.27), the convoys of thousands of mules that followed Alexander's army (Quintus
Curtius 3,13,16; 8.7.4), the wagon train of 400 vehicles (clitellae) that proved in-
sufficient to provision Capua in 212 B.C. (and were later heaped up as an impro-
vised fortification) (Livy 25.36.7), or, finally, the hundreds of carpenta and vehicula
taken as booty from the Gauls, Boiens, or Romans in various military actions dur-
ing the second century B.C. The sculptured columns of Trajan and Marcus Aur-
elius in Rome provide a veritable compendium of the means of transport used by
the Roman army (figure 23.7), including the movement of siege engines and siege
towers or transport by ships on sea and river.
The notable development of means of transport during the Roman imperial
period is bound up equally with the imperial message service, the cursus publicus
(Kolb 2000; Stoffel 1994). Founded by Augustus, this express message service, at
first carried out by horsemen, became a vast carriage service at the disposal of the
State with an infrastructure of stations and fresh horses, and its own personnel, but
it relied as well on requisitions. Various types of vehicles appear in the Theodosian
Code (De cursu publico angariis et parangariis 5, 8), along with their load limits,
from the birota rated at 200 pounds (66 kg) to the raeda rated at 1,000 pounds (330
kg) and the angaria rated at 1,500 pounds (492 kg). These statistics have often been
considered, following the lead of Lefebvre des Noettes, as a decisive proof of the
inefficiency of ancient harnessing technique. It has been shown above that heavy
transport of ten tons or even more was, if not commonplace, at least perfectly
feasible and had been accomplished already in archaic Greece. The figures in the
Theodosian Code, then, present problems of interpretation, but they can in no way
be considered a true indication of the technical limits of transport loads. Legis-
lation concerned itself as well with the regulation of traffic in urban centers. The lex
Oppia, passed during the period of the Punic Wars, forbade women from traveling
in an animal-drawn vehicle within the walls of Rome (Livy 34.1.3). During the reign
ofJulius Caesar (CIL 1.593) privately owned wheeled vehicles were prohibited from
entering Rome during daylight hours, and later on Claudius forbade travelers from
entering towns in Italy "except on foot, in a sedan chair, or in a litter" (Suetonius,
Claud. 25.2).
In addition to those who-on a casual or professional basis-rented out
mounts or carriages, with or without a driver, at city gates, harbors, and highway
rest stops, the personnel engaged in transport enjoyed an important role in both the
private economy and in public services. They often organized themselves in pro-
fessional associations that assumed an important part in large-scale commerce, in
particular in the western provinces. The corporations of carriers by sea and river
(navicularii, nautae) are well documented in the important cities and ports ofltaly
and the western provinces. The nautae, "entrepreneurs engaged in river transport,"
who specialized in a particular river basin, had to guarantee as well the transfer
of goods by land between rivers, either personally or by subcontract. The com-
plexity and the technical sophistication of long-distance transport justified the
existence of a specialized group of workers and raises questions as to why fewer
collegia (corporations)-apart from the cisiarii (cart drivers, CIL 11.2615)-are
documented for land transport. Journeys by land were often shorter and perhaps
easier to organize than those by water, so negotiatores (traders-well attested in
epigraphical sources for a variety of products) must often have had to arrange their
own means of transport, or have rented these services from the owners of teams,
particularly farmers.
One must conjure up as well an image of the road system, which ought to be
integrated in the statement of the issues involved in land transport (Laurence 1999;
see also chapter 22). Contrary to a widespread misconception, the great Roman
road network, although conceived in the spirit of strategic value and bound up with
the needs stemming from the exercise of centralized authoritarian power, never-
theless constituted an important support for commercial transport. Besides, a paved
road is not essential for movement on land, even with animal-drawn transport.
Aside from traditional sources, archaeological and cartographical studies on the
dispersion and diffusion of artifacts show that the process of commercial distri-
bution, for example of sigillata wares (Raepsaet and Raepsaet-Charlier 1987-1988),
makes use of all routes, from strategic armatures to local byways, and supplemented
by rivers, which were not at all a marginal factor. With regard to the quality and
quantity of the products purchased, there appears to have been no difference in
purchasing power between sites on a river and those on a land route, despite the
marked differences in cost among transport by road, river, or sea. One of the tra-
ditional props for the minimalist theory in matters of land transport is the com-
paratively very high cost of the latter. The figures given by Diocletian's Edict
Concerning Maximum Prices show in effect that in every case of routes within the
empire, transport by land was significantly more expensive than transport by river,
and still more than that by sea. These statistics corroborate some anecdotal reports
in literary texts (e.g., Cato, Agr. 22.3). This fact, which in the orthodox view would
appear as a decisive argument against the importance of the role ofland transport in
the classical world, is in reality a widespread commonplace among preindustrial
societies. In eighteenth-century England, for example, the relative cost of transport
by sea, river, and land was on average a factor of 1, 4.7, and 22.6, respectively, figures
close to those in the Roman Empire. Besides, the ancient system of commerce-in
large part preindustrial in character-functioned by including the overhead costs of
transport into the base cost of the merchandise. To speak of a handicap in the
classical world is both anachronistic and epistemologically irrelevant.
Ancient land transport was based on human portage, the pack saddle, the light two-
wheeled cart with draw pole and a pair of horses, asses, or mules, the heavy, four-
wheeled wagon (or trolley with numerous wheels) drawn by oxen yoked together in
pairs, sometimes several pairs in series, to which may be added the exceptional
western Roman two-pole vehicle with a single draft animal. The Greek hamaxa, a
small all-purpose cart with two crossbar wheels, existed alongside the heavy tet-
rakukloi wagons, just as at Rome the small, fast cisium with spoked wheels com-
plemented the imposing carrucae, veritable predecessors of the heavy mail coach.
The Roman imperial culture was in the process of developing the means of
transport by land, including an excellent control of the techniques for producing

the spoked wheel, a functional range of service vehicles, integrating the interesting
accomplishments of Celtic skill in cart construction. The pax romana favored a
dynamic of development in which the quality of means of transport played an active
role.
In general, in the Greco-Roman world transport in all its forms took part in all
the surrounding economic systems, satisfying the demands made of it without any
particular handicap. It relied on the traditional motive forces (human, animal,
water, wind), which have the inconvenience of a lesser potential in comparison with
modern fossil fuels, but the advantage of being indefinitely renewable. The principal
ancient technique for harnessing is in every case bound up with the double or sin-
gle yoke, the habitual method for bovines and asses. It is less well adapted to the
morphology of horses, but their economic role was marginal. In any case, even with
horses, experimentation has shown that, with a light load of a ton, or even more, the
harness design does not pose any problems for traction. Structurally, with regard to
the human and animal energy available and the principal draft harness used, the
means of land transport in the classical world aligns itself well with the typical
operation of transport in preindustrial societies. Its efficiency, however, reached a
very exceptional peak during the Roman Empire, in terms of both public admin-
istration and private use. The technological innovation of the horse collar in certain
parts of western Europe, associated with the economic grown of the beginning of the
medieval period, finds an interesting prototype in the ancient Gallo-Roman single
yoke. Some other qualitative jumps-with regard to the harness and the technology
of animal draft: soft traces, whiffle-tree; with regard to the vehicle: the offset rim,
wheel camber, suspension on longitudinal straps-mark the long stretch of history,
but without causing any true general quantitative jump prior to the fundamental
changes in the sources of energy that took place in the nineteenth century.
The social value of a technical component depends as much on the effectiveness
and the coherence of the chain of technological links and on the economic envi-
ronment as on its own specific mechanical quality. It constitutes one link in the
means of production, dormant or active. The Greek and Roman cultures had at
their disposal a technical capacity for land transport that was real and varied, even
innovative, inscribing its own rhythms and inflections on the long-term patterns of
preindustrial societies.
REFERENCES
Abeels, P. 1995. "Les configurations de la traction aux brancards: Forces et effets," in

Raepsaet and Rommelaere 1995: 13-29.
Adam, J.-P. 1984. La construction romaine, materiaux et techniques. Paris: Picard.
Amouretti, M.-C., and G. Comet 1994. Hammes et techniques de l'Antiquite a la Re-
naissance. Paris: Colin.
Amouretti, M.-C., and F. Sigaut (eds.) 1998. Traditions agronomiques europeennes: Ela-
boration et transmission depuis l'Antiquite. Paris: CTHS.
Argoud, G., and J.-Y. Guillaumin (eds.) 1998. Sciences exactes et sciences appliquees a
Alexandrie. Saint-Etienne: Publications de l'Universite.
Astill, G., and J. Langdon (eds.) 1997. Medieval farming and technology: The impact of
agricultural change in northwest Europe. Leiden: Brill.
Burford, A. 1960. "Heavy transport in ancient Greece," Economic History Review 13: 1-18.
Burford, A. 1969. The Greek temple builders at Epidauros. Liverpool: University Press.
Comet, G. (ed.) 2003. L'outillage agricole medieval et moderne et son histoire. Toulouse:
Presses universitaires du Mirail.
Cotterell, B., and J. Kamminga 1990. Mechanics of pre-industrial technology. Cambridge:
Crouwel, J. H. 1981. Chariots and other means of land transport in Bronze Age Greece.
Amsterdam: Allard Pierson.
Crouwel, J. H. 1992. Chariots and other wheeled vehicles in Iron Age Greece. Amsterdam:
Allard Pierson.
Deman, A. 2002. "Avec les utriculaires sur les sentiers muletiers de la Gaule romaine,"
Cahiers Glotz 13: 233-46.
Fellmeth, U. 2002. "Eine wohlhabende Stadt sei nahe . .. , " Die Standortfaktoren in der
romischen Agrarokonomie im Zusammenhang mit den Verkehrs- und Raumordnungs-
strukturen im romischen Italien. St. Katharinen: Scripta Mercaturae Verlag.
Fenton, A. et al. (eds.) 1973. Land transport in Europe. Copenhagen: Nationalmuseet.
Fleury, P. 1993. La mecanique de Vitruve. Caen: Presses universitaires.
Garbsch, J. 1986. Mann und Ross und Wagen. Miinchen: Museum fiir Vor- und Friih-
geschichte.
Gasparin Cte de. 1863. Cours d'agriculture. 3rd ed. Paris: Librairie agricole de la Maison
rustique.
Gille, B. 1980. Les mecaniciens grecs: La naissance de la technologie. Paris: Seuil.
Greene, K. 1990. "Perspectives on Roman technology," Oxford Journal of Archaeology 9:
209-19.
Greene, K. 1994. "Technology and innovation in context: The Roman background to
medieval and later developments," Journal of Roman Archaeology 7: 22-32.
M. I. Finley re-considered," Economic History Review 53: 29-59.
Haudricourt, A.-G. 1987. La technologie, science humaine. Paris: Maison des Sciences de
l'Homme.
Healey, J. F. 1999. Pliny the Elder on science and technology. Oxford: Oxford University
Press.
Hopfen, H.J. 1970. L'outillage agricole pour les regions arides et tropicales. Roma: FAO.
Humphrey, J., J.P. Oleson, and A. Sherwood 1998. Greek and Roman technology: A
Hyland, A. 1990. Equus: The horse in the Roman world. London: Batsford.
Jankuhn, H. et al. (eds.) 1989. Untersuchungen zu Handel und Verkehr der vor- und fruh-
geschichtliche Zeit in Mittel-und Nordeuropa. Vol. 5, Der Verkehr. Gottingen: Van-
denhoeck and Ruprecht.
Kolb, A. 2000. Transport und Nachrichtentransfer im Romischen Reich. Berlin: Akademie
Verlag.
Landels, J. G. 2000. Engineering in the ancient world. Rev. ed. Berkeley: University of
California Press.
Langdon, J. 1986. Horses, oxen and technological innovation: The use of draft animals in
English farming, 1066-1500. Cambridge: Cambridge University Press.
Laurence, R. 1999. The roads of Roman Italy: Mobility and cultural change. London: Rou-
tledge.
Lefebvre des Noettes, Ct. 1931. L'attelage, le cheval de selle a travers les Ages. Paris: Im-
primerie nationale.
Leone, A. 1988. Gli animali da trasporto nell'Egitto greco-romano e bizantino. Roma: Editrice
Pontifico.
Leone, A. 1992. Gli animali da lavoro, da allevamento e gli hippoi nell'Egitto greco-romano e
bizantino. Roma: Edizioni Atliena.
Lewis, M. 2000. "The Hellenistic period," in 6. Wikander (ed.), Handbook of ancient water
Littauer, M. A. 1968. "The function of the yoke saddle in ancient harnessing," Antiquity 42:
27-31.
Littauer, M. A., and J. Crouwel 1979. Wheeled vehicles and ridden animals in the ancient
Near East. Leiden: Brill.
Lorimer, H. L. 1903. "The country cart in ancient Greece," Journal of Hellenic Studies 23:
132-51.
Mattingly, D., and J. Salmon (eds.) 2001. Economies beyond agriculture in the classical world.
London: Routledge.
Meijer, F., and 0. Van Nijf. 1992. Trade, transport and society in the ancient world: A
Meissner, B. 1999. Die technologische Fachliteratur der Antike. Berlin: Akademie Verlag.
Mingote Calderon, J. L. 1996. Tecnologia agricola medieval en Espana: Una relacion entre
la etnologia y la arqueologia a traves de los aperos agricolas. Madrid: Ministerio de
agricultura.
Molin, M. 1987-1988. "La faiblesse de l'attelage antique ou la force des idees re1yues
en histoire ancienne," Bulletin archeologique du Comite des travaux historiques 23-24:
39-84.
Nicolet, C. (ed.) 1996. Les litteratures techniques dans l'Antiquite romaine. Geneve: Fon-
dation Hardt.
Oexle, J. 1992. Studien zu merowingerzeitlichem Pferdegeschirr am Beispiel der Trensen.
Mainz: von Zabern.
Piggott, S. 1983. The earliest wheeled transport: From the Atlantic coast to the Caspian Sea.
Itliaca, NY: Cornell University Press.
Rad und Wagen 2004. Rad und Wagen: Der Ursprung einer Innovation. Wagen im Vorderen
Orient und Europa. Mainz: von Zabern.
Raepsaet, G. 1982. "Attelages antiques dans le Nord de la Gaule: Les systemes de traction
par equides," Trierer Zeitschrift 45: 215-73-
a
Raepsaet, G. 1984. "Transport de tambours de colonnes du Pentelique Eleusis," L'An-
tiquite Classique 53: 234-261.
Raepsaet, G. 1988. "Charrettes en terre cuite de l'epoque archaique aCorinthe," L'Antiquite
Classique 57: 56-88.
Raepsaet, G. 1993. "Le diolkos de l'Istlime a Corintlie: Son trace, son fonctionnement,"
Bulletin de Correspondence Hellenique 117: 233-61.
Raepsaet, G. 1998-2001. Der Neue Pauly. s.v. Esel, 4: 130-135. s.v. Landtransport, 6: 1098-
1106. s.v. Maultier, 7: 1043-1047. s.v. Pferd, 8: 692-703. s.v. Rind, 9: 1014-1020.
Raepsaet, G. 2002. Attelages et techniques de transport dans le monde greco-romain. Brux-

elles: Timperman.
Raepsaet, G., and M.-T. Raepsaet-Charlier 1987-1988. "Aspects de !'organisation du
commerce de la ceramique sigillee dans le Nord de la Gaule au Ile siecle den. ere,"
Munstersche Beitrage zur antiken Handelsgeschichte 6: 1-29 and 7: 45-69.
Raepsaet, G., and C. Rommelaere (eds.) 1995. Brancards et transport attele entre Seine et
Rhin de l'Antiquite au Mayen Age. Bruxelles-Treignes: ULB.
Rommelaere, C. 1991. Les chevaux du Nouvel Empire egyptien. Louvain-la-Neuve: UCL.
Ri:iring, C. 1983. Untersuchungen zu romischen Reisewagen. Koblenz: Numismatischer
Verlag.
Schi:infelder, M. 2002. Das spatkeltische Wagengrab von Boe (Dep. Lot-et-Garonne): Studien
zu Wagen und Wagengrabern der jungeren Latenezeit. Mainz: Ri:imisch-germanisches
Zentralmuseum.
Schneider, H. 1992. Einftlhrung in die antike Technikgeschichte. Darmstadt: WBG.
Sigaut, F. 1985. L'evolution technique des agricultures europeennes avant l'epoque industrielle.
Paris: EHSS.
Sigaut, F. 1987. "Haudricourt et la technologie," in Haudricourt 1987: 9--32.
Spruytte, J. 1977. Etudes experimentales sur l'attelage: Contribution a l'histoire du cheval.
Paris: Crepin-Leblond.
Starkey, P. 2002. Moyens de transport locaux pour le developpement rural. London: DFID.
Stoffel, P. 1994. Uber die Staatspost, die Ochsengespanne und die requirierten Ochsenge-
spanne. Bern: Lang.
Tarr, L. 1969. The history of the carriage. New York: Arco.
Techniques rurales 1971. Techniques rurales en Afrique. Vol. 13, Manuel de culture avec
traction animale. Paris, Ministere des Affaires etrangeres.
Treue, W. (ed.) 1986. Achse, Rad und Wagen: Funf Tausend Jahre Kultur- und Tech-
nikgeschichte. Gi:ittingen: Vandenhoeck and Ruprecht.
Vierradrige Wagen 1987. Vierradrige Wagen der Hallstattzeit: Untersuchungen zu Geschichte
und Technik. Mainz: Ph. von Zabern.
Vigneron, P. 1968. Le cheval dans l'Antiquite greco-romaine. Nancy: Universite.
Whitehead, D., and P.H. Blyth 2004. Athenaeus: Mechanicus, On Machines. Stuttgart:
Steiner.
CHAPTER 24
SEA TRANSPORT,
PART 1: SHIPS AND
NAVIGATION
,
SEAN MCGRAIL
SOURCES OF EVIDENCE
Nowhere in the world have technical discussions of shipbuilding or navigation

come down to us from before late medieval times. Today, both technologies are
primarily science based, but in the period under discussion-800 B.C. to A.D.
500-they were more of an art in that reliance was placed on personal experience
and ability. During those centuries rafts, boats, and ships were built by eye and
rules of thumb (McGrail 2003), while navigation-the art of determining a vessel's
position on the open sea and taking her in safety from one place to another-was
undertaken without instruments, using handed-on experience, local knowledge,
and detailed observation of natural phenomena (McGrail 1996a). These ship-
building and navigational skills were acquired by oral instruction and supervised
"apprenticeships." We cannot, therefore, expect accounts of these techniques to
have been compiled before ship design and navigation became science-based and
printing became commonplace-that is, until the fifteenth century A.D.
SEA TRANSPORT: SHIPS AND NAVIGATION 607
Textual Evidence
Nevertheless, some of the written works that have come down to us do provide
clues to early shipbuilding and navigational techniques (see, for example, Hum-
phrey et al. 1998: 442-84; Casson 1989). Furthermore, numerous two- and three-
dimensional representations of boats and ships add to our knowledge, especially of
masts, sails, and rigging (Basch 1987a). Excavated vessels, even if fragmentary, give
us direct evidence of shipbuilding techniques, but archaeology throws little light on
early navigation since, apart from the sounding lead and line, it was carried out
without instruments. Contemporary ethnographic evidence offers some assistance.
Since the main aim of Homer, and of later classical authors such as Caesar,
Strabo, and Pliny, was not to document shipbuilding or navigational techniques
but to record events that had been handed down to them or that they themselves
had observed, the nautical information found in their surviving works is mostly of a
general nature and often inconsequential. Nevertheless, there are passages that alert
us to some special characteristic of a contemporary form of water transport, leading
to deductions about the capabilities of both ship and crew (Mark 2005).
Iconographical Evidence
Representations of rafts and boats are the main source of evidence for the earliest
stages of shipbuilding, prior to the Late Bronze Age. Moreover there are depictions
of rafts and of boats made of hide, basketry, or logs, physical examples of which
have not yet been excavated. Egyptian paintings and models are especially infor-
mative: they not only depict plank boats being built and boatmen with lead and
line ready to check water depth (pace Oleson 2000) but also illustrate all forms
of propulsion and steering. Furthermore, the survival, recovery, and recording of
the Cheops ship of approximately 2600 B.C.-the earliest documented planked
vessel-has provided some control on the interpretation of Egyptian iconography
(figure 24.1). On the other hand, without extensive boat remains in the Aegean
before the first millennium B.C., lack of effective controls has led to some imag-
inative interpretations of early boat depictions (McGrail 2004a: 120-22).
Ethnographical Evidence
In India and other parts of the developing world, boats are still built using inherited
tools and techniques, to a design handed on by the builder's predecessors. Such
boats are built of wood and are propelled by muscle power or by the wind, char-
acteristics that make it possible to use them to interpret clues to ancient technol-
ogy found in excavated boat remains. There are problems in using cross-cultural
analogies, but the more alike in technological, environmental, and economic terms
the ancient and the recent cultures are, the greater the value of ethnographic boat
Figure 24.1. Cheops ship (ca. 2650 B.C.), Giza, Egypt. (Photograph by P. Johnstone.)
studies to the investigation of early nautical technologies and boat use. This is not to
say that there is a one-to-one relationship between ancient technology and "living
tradition." Ethnographic evidence may suggest a range of possible answers to ques-
tions arising from incomplete, fragmented, and distorted boat remains, but "only
archaeology, in conjunction with the various natural sciences, can give the right
answers" (Clark 1953: 357), bearing in mind that, in the present stage of knowledge,
no answer may be possible at all.
Archaeological Evidence
Egypt provides the world's earliest examples of excavated, planked vessels. Never-
theless, such excavated ships and boats do not instantly reveal their design princi-
ples, original form, sequence of construction, or propulsion and steering arrange-
ments. They are incomplete (especially when wrecked rather than entombed): the
sailing rig rarely survives, and individual timbers are distorted and displaced. Before
a comprehensive account of such a vessel can be published, a lengthy multidisci-
plinary investigation has to be undertaken, leading, when sufficient evidence sur-
vives, to a theoretical reconstruction of the original vessel (or as near to that as it is
now possible to get) in the form of a small-scale model or drawing, together with
appropriate calculations (McGrail 2006). Occasionally this process leads to the
building of a full-scale model (called a replica) of this reconstruction. This phase of
post-excavation research (which is not only lengthy but also expensive) has not al-
ways been undertaken rigorously, sometimes leading to questionable results (Coates
et al. 1995; Crumlin-Pedersen and McGrail 2006).
Until the late twentieth century, vessels were usually dated by the archaeo-
logical context from which they had been excavated, or from cargo or other finds
associated with them. These indirect methods were sometimes insufficiently ac-
curate, however, and in more recent times vessels have been dated directly by
radiocarbon assay or by dendrochronology. A clearer understanding of techno-
logical changes within the early Mediterranean will only be possible when the more
precise dates obtainable from tree ring analysis become available. Such investiga-
tions may also suggest the region where the parent trees of the ship's timber grew,
and thus help in the search for origins.
Each early plank boat was unique, since until the end of our period they were
conceived and built individually. Nevertheless, sufficient finds have now been ex-
cavated and published to show that groups of vessels can be identified, each member
with its own individuality yet having several structural characteristics in common
with other members of its group. These traditions are theoretical constructs that
help our understanding; they may or may not be similar to the concepts of the
people who built and sailed these vessels. Two such traditions recognized in our
period are vessels with sewn-plank fastenings (figure 24.2) and vessels with locked
mortise and tenon fastenings (figure 24.3). There are also "hybrid" boats that are
mostly mortise and tenon fastened but with some sewn strakes (for an explanation
of this and other nautical terms, see the glossary at the end of the chapter). Ships
and boats of both traditions were built plank-first-that is, the planking was fas-
tened together to form the shape of hull required; then the supporting framing was
fastened to the planking. It may be that as the definition of individual building
traditions in terms of structural features is refined, it will become practicable to
recognize changes over time with precision and even to identify the region where
the boat that constitutes an individual wreck had been built. Indeed, the long-term
aim must be to narrow origins down to a particular shipyard identified by a personal
touch given to several excavated wrecks by a particular master shipwright.
The identification of the timber species used in the hull, the dendrochrono-
logical examination of the timber, the home port of the ship's crew as deduced from
their personal possessions, and the nature of the cargo and ballast carried have all
been used to investigate a ship's origins. Cargo and ballast may be misleading,
however, since both may be embarked late in the ship's life, far away from her
supposed home port. On the other hand, the origin of the cargo may prove useful
in determining what had been or was to be the vessel's route on the final voyage.
Dendrochronological examination and species identification of hull timbers are
the most promising ways to a ship's origins, followed by personal possessions, if
these can be recognized as such.
Figure 24.2 . Ma'agan Mikhael ship, reconstruction of sewn fastenings. (Photograph by: Y.
Kahanov.)
Cargo found with a wreck can indicate that the original vessel was a merchant
ship. If the reconstructed lines of the vessel show that she had a full transverse section
and a length to breadth (LIB) ratio of the underwater hull of about p, this function is
confirmed. In the classical world a specialized warship had a L/B ratio of about 6:1 or
7:1; examples of the sixth or fifth century B.C. are likely to have had a ram and thus
would have had a full complement of oarsmen. Apart from timbers inside the Athlit
ram (Casson and Steffy 1991) no unambiguous warship remains have so far been
identified. It has been argued (Morrison and Coates 1986: 128; Casson 1991: 82) that
trireme warships needed no ballast and were built of such lightweight timber that
they did not sink when damaged, but lay crippled on the surface until they could be
towed away. If this was in fact the case, their remains are unlikely to be excavated.
A seagoing vessel has to have adequate freeboard and stability. Assessment of
these and other qualities must be based on an authentic reconstruction of the orig-
inal shape, structure, propulsion, and steering of an excavated vessel. An open boat
ofless than a certain size is unlikely to have been seagoing, but a decked vessel could
have been. A rockered bottom, a "boat-shaped" underwater hull, and a sheerline
rising toward the ends suggest a seagoing vessel; a box-shaped vessel, on the other
hand, is likely to have been a river or harbor craft. It is important that such attributes
and features are not worked into a reconstruction without good reason (Crumlin-
Pedersen and McGrail 2006).
Experimental Archaeology
Theoretical reconstructions of an ancient vessel, as well as the effectiveness or
otherwise of deduced, noninstrumental methods of navigation, need to be tested at
sea (Coates et al. 1995; McGrail 1993, 2004a: 7). Twentieth-century Micronesian
11 ''6 )
,, l
~L_j, I
I
'
I
yr~--,~~1111
.r ,
VG
'/ :,
~l···-J17
'1 :I
mFJ
•1 :l ,
I
I
---- --------____ q[~ ~J----------
n
i~r1- -0-i+-----
·1
-- - -ti
'' ,-'<> ~L-------
-i'
--~-'<>~
'i ,rL_ -- - J,,
kc~ l1----··'-, k:=!:l k··•-.i.j ',
J: I :! :: i li '\.
Figure 24.3. Typical locked mortise and tenon fastenings, diagram. (Drawing by S. Haad.)
methods of navigation have been successfully used under controlled conditions in

the South Pacific (McGrail 2004a: 339-45). The findings of such trials are applicable
to voyages undertaken after the initial discovery of islands during exploratory
voyages. It would be much more difficult-probably impossible-to reenact an
exploratory voyage on which the participants did not know what, if any, land lay
ahead, other than the general direction of distant land from the flight line of mi-
gratory birds or the visible deflection of a swell by a faraway island. Nor would these
first explorers have known with any certainty what environmental conditions lay
ahead of them, except as projections of the weather and sea states recently expe-
rienced; this state of doubt would also be difficult to reenact.
Although experimental navigation has not yet been undertaken in the Medi-
terranean, it seems reasonable to expect that the theory and practice of noninstru-
mental navigation as they evolved in a South Pacific context are generally applicable
to the Mediterranean. Indeed, they are probably valid worldwide, since the differ-
ences now noted between the navigational techniques used recently in Oceania, the
China Seas, and the Indian Ocean are principally due to differences in latitude and
other aspects of the maritime environment, rather than to fundamental differ-
ences in techniques (McGrail 2004a: 434). Moreover, there are passages in Homer's
Odyssey from which it is clear that noninstrumental navigational methods used in
the Early Iron Age Mediterranean fit the South Pacific model (McGrail 1996a; 2004a:
100-102; Mark 2005: 138-52).
Full-scale reconstructions of two notable Mediterranean vessels-both with
mortise and tenon plank fastenings and nail-fastened framing-were designed,
built and sailed during the 1980s (McGrail 2004a: 141-45; 151-52), each being an
example of a particular approach to experimental boat archaeology. The first was a

specific experiment, in which the documentation of the excavated remains of the
Kyrenia wreck of around 300 B.C. (Steffy 1985; 1989; 1994: 45-59) was transformed
(probably via a small-scale drawing or model) into a full-scale, seagoing ship
(Katzev and Katzev 1986, 1987; Katzev 1989a, 1989b, 1990; Cariolou 1997).
In contrast, since there are no known remains of a trireme, the most extreme
warship design of the sixth and fifth centuries B.C., the Trireme Project's aim was
to reconstruct a representative example of an Athenian trireme of the fifth century
B.C. It was hoped that this reconstruction would achieve historically documented
measures of performance while complying with evidence derived from icono-
graphic, historical, and archaeological sources illuminated by naval architectural
principles (Coates and McGrail 1984: 11-12, 51-52; Coates et al. 1990; Shaw 1993;
Morrison et al. 2000: 191-206). This full-scale reconstruction was subsequently
commissioned into the Hellenic Navy as Olympias (figure 24.4).
There are many difficulties to be overcome by those who undertake either type
of experiment (Coates et al. 1995; McGrail 1986, 1993, 2004b). Not the least of these
are the addition of missing parts to the incomplete, dispersed, and distorted re-
mains that are excavated, and the design of an appropriate propulsion and steering
outfit. Both experiments have been evaluated (Westerdahl 1992; Tilley 1976, 1992,
1995, 2004; Basch 1987b; McGrail 1992). It is generally agreed that the Trireme
Project has increased understanding of warship building and seafaring in the early
eastern Mediterranean. This experiment was undertaken in a scholarly and scien-
tific manner and progressed in phases, each being presented and/or published
before moving on to the next phase. Planning began in 1982; an "advisory discus-
sion" was held at the National Maritime Museum, Greenwich, in 1983; the pro-
ceedings, including an agreed design, were published in 1984, and the building of
Olympias began in 1985. Reports on the planning, on the building, and on the several
trial seasons were subsequently published (Morrison and Coates 1989; Coates et al.
1990; Shaw 1993). This openness to criticism generated confidence in the recon-
structed ship, which, despite the choice of what was later deduced to be an inap-
propriate length of cubit as a module (Morrison et al. 2000: 245-46, 268), can now
be seen to be as valid a reconstruction of a representative fifth-century B.C. trireme
as can be expected with the evidence available (contra, Tilley 2004).
On the other hand, although the excavated evidence on which the Kyrenia ship
reconstruction (Kyrenia 2) was specifically based has been well argued and pre-
sented by Steffy (1985; 1994: 45-59), a definitive report on the design, building, and
use at sea of Kyrenia 2 has yet to be published. The authenticity of this experi-
ment cannot at present be confirmed, nor can its value to the study of a late-first-
millennium B.C. Mediterranean cargo ship-building and use be determined.
Doubts about the authenticity of reconstructions and the reliability and validity of
experiments could be forestalled if experimenters were to publish their experimental
philosophy and methodology and clearly state their chain of argument from evi-
dence to reconstruction, before they committed to building a full-scale model. This
should allow time for wide criticism so that, in time, an agreed "floating hypoth-
Figure 24.4. Reconstructed trireme Olympias under sail. (Photograph by P. Lipke, Trir-
eme Trust.)
esis" could be built that would be archaeologically and historically valuable

(McGrail 2004a: 5-6, 141-45, 151-52; 2006).
Environmental Contexts
In order to understand the use of early water transport it is necessary to reconstruct
the maritime environment in which it was operated: that is, former sea levels,
coastlines, weather, currents, and tides. Over the past 3,000 years, however, the
effects on Mediterranean seafaring of changes in sea level and climate have gen-
erally been insignificant (Murray 1987; McGrail 2004a: 89-95). In our period, the
seasonal winds, local winds, land and sea breezes (prominent during a Mediter-
ranean summer), and current flows are all thought to have been only slightly
different from those of today. But tectonic movements, coastal erosion, silting, and
soil erosion will have caused significant local changes. With this proviso, today's
Mediterranean maps, charts, weather and tidal data may be applied to the late Iron
Age.
Tidal ranges and streams can be influenced by sea level and coastline changes,
but the Mediterranean's almost closed nature means that it is, and was, almost
tideless: only in the straits of Gibraltar, Messina, and Euripus are there appreciable
tidal flows, and only at the head of the Adriatic and in the Gulf of Gabes (Tunisia)
do tidal ranges exceed one meter.
SHIPBUILDING BEFORE 800 B.C.
Egypt
Herodotus (2.5) described Egypt as "the gift of the River Nile," meaning that
Egyptian civilization would not have existed without the fertile silt regularly carried
northward by that great river. The Nile may also be said to be responsible for
Egypt's nautical development, since it was the principle means of communication
through the land, with no part of the valley more than 16 km from the river.
Compared with the Mesopotamian rivers, the Nile had the advantage that it flowed
north against the predominant northerly wind; thus sail could be used on boats
heading southward against the river current, and was so used before 3100 B.C.
(McGrail 2004a: 16, fig. 2.5; James 1983: 21-22). Excavated tomb paintings and
models show that from the Early Dynastic Period (ca. 3100-2866 B.C.) onward reed
bundle rafts were used on the Nile and in its wide-spreading delta channels. Such
rafts have not been excavated, but there are several examples of excavated planked
vessels from later periods.
In 1954, the Cheops ship was recovered from an underground chamber near the
pyramid of Cheops (or Khufu) and dated to about 2650 B.C. (Lipke 1984; Haldane
1997b; Ward 2000: 45-60) (figure 24.1). As reassembled, the ship measures 43.4 x 5.9
x 1.8 m, with the ends rising to 6-7.5m. This impressive vessel was built plank-first,
the cedar hull planks being positioned next to one another by projections from the
plank's edges which enmeshed with indentations on adjacent planks, by treenails
within the seams, and by unlocked mortise and tenon joints. Another version of
that joint, in which the loose tenon was locked by treenails, was also known to
Egyptian woodworkers, and was used in this vessel's superstructure but not in the
hull. These fittings and features within plank seams not only relocated the planking
during refits after dismantling for stitch renewal, but also resisted the tendency of
adjacent planks to slide relative to one another when afloat, by carrying the shear
stresses imposed on the hull.
The planks were fastened together by widely-spaced, individual lashings of
halfa grass (Demostachya bipinnata), and the whole shell of planking fastened from
sheer to sheer by transverse sewing, the sewing holes being worked within the
thickness of the strakes so that stitches were not exposed externally, where they
would be damaged when the ship was berthed. This Egyptian care for sewn fas-
tenings is a feature of every later sewn-plank technology (McGrail and Kentley 1985;
Mark 2005: 25-69). The planked hull was subsequently stabilized by crossbeams let
into the top strake and by sixteen great floor timbers lashed to the planking. A
central carling, lashed to the beams and supported by stanchions, reinforced the
hull longitudinally.
Several other early Egyptian boats have been found on land. During the last
decade of the twentieth century, a group of fourteen boat graves, provisionally
dated to the period before 2600 B.C., were exposed at Abydos in the vicinity of
second and first dynasty monuments. One vessel has been partly excavated (Ward
2003). The parallel-edged planks of these flat-bottomed boats were fastened to-
gether by lines of sewing over caulking, running transversely across the bottom and
sides of the boat, as in the Cheops ship. Excavations between 1893 and 1895 around
the pyramid of Sesostris (Senusret III) at Dahshur revealed six boats dated to
around 1850 B.C. (de Morgan 1894-1903). The three planks forming the plank-keel
of these boats were butted end to end and fastened together by wooden dovetail
cramps set into the planks' inboard faces. The strakes were fastened together, edge
to edge, by unlocked mortise and tenon joints and by widely spaced dovetail
cramps; the ends of the fourth strakes, which do not run the full length of the hull,
were lashed in position. Ward (Haldane 1996, 1997a) has suggested that the dove-
tail cramps are modern replacements for L- or V-shaped lashing holes within the
planks' thickness. The boats have no framing but are reinforced by thirteen cross-
beams lashed in position and held to the third strake by treenails.
The earliest known use of locked mortise and tenon fastenings in the Levant
region is on a table from a Jericho tomb of the mid-second millennium B.C.
(Wachsmann 1998: 241, figs. 10.28-29). Furthermore, Sleeswyk (1980) has shown
that the Romans subsequently knew such joints as "Phoenician joints" (coagmenta
punicana; Cato, Ag. 18.9). Mid-second-millennium B.C. wrecks off the Turkish
coast at Uluburun and Cape Gelidonya, with their planking joined by locked
mortise and tenon fastenings, have been identified from their cargo as Canaanite/
Syrian or Levantine (Bass 1991; 1997: 269; Pulak 1998: 214; Wachsmann 1998: 206-7),
while Pulak (2003: 29) states that they were "built along the Syro-Palestinian coast
or on Cyprus." The view that these vessels were built on the Levant coast has not
gone unchallenged, and alternative origins in the Aegean have been suggested.
Identifying a wreck's origin is difficult (McGrail 1991: 83-84; 2004a: 10), and a
definitive answer may prove impossible. Nevertheless, the balance of evidence does
seem to point toward an origin in the Levant or Cyprus region rather than in the
Aegean. The locked mortise and tenon plank-fastenings in the Uluburun wreck (ca.
1305 B.C.) are the earliest known in a ship's hull. It is therefore conceivable that early
Phoenician shipbuilders adopted Egyptian woodworking techniques and devel-
oped a locked joint that was watertight (Wachsmann 1998: 241). Such an innovation
would have enhanced the Phoenician reputation for seafaring excellence (McGrail
2004a: 129). On their subsequent overseas trading voyages, the Phoenicians may
well have introduced locked mortise and tenon hull plank fastenings to other
Mediterranean lands (Basch 1981).
Early Egyptian seagoing ships are depicted with girdles at bow and stern, and
with hogging hawsers running from bow to stern. By the reign of Ramesses III
(1198-1166 B.C.), however, hawsers and girdles are no longer depicted, suggesting
that confidence had been increased in the structural integrity of the hull-this may
have been due to the introduction of locked mortise and tenon fastenings subse-
quently seen in the mid-first millennium B.C. ship excavated at Matariya, near
Heliopolis (Wachsmann 1998: 222).
If the conjecture that Phoenicians in the later second millennium inherited
shipbuilding skills from the Egyptians is correct, the Phoenicians would, initially,
have used sewn plank fastenings as well as mortise and tenon fastenings, and frames
would have been lashed to that planking. No frames were excavated from the
Uluburun or the Cape Gelidonya sites, and there were no frame fastenings visible on
the small sections of planking recovered. It is unlikely, however, that such seagoing
vessels were without framing, even though the sturdy tenons used in the Uluburun
ship appear to have acted as internal frames. Furthermore, the small areas of
planking surviving on both sites (the largest section of Uluburun planking mea-
sured 1.8 x 1.0 m) leave open the possibility that, toward bow and stern, the planking
was also sewn. Such a blend of fastenings was not only an early Egyptian charac-
teristic but was subsequently used on several of the first millennium B.C. excavated
ships of the eastern and central Mediterranean (Mark 2005: 50-51, 60-69, 183-85).
A further conjecture is that in the third millennium B.C., before the Phoeni-
cians became wide-ranging seafarers, Mediterranean hulls-of which we have sim-
ple representations (McGrail 2004a: 105-11) but no excavated examples-had la-
shed framing and sewn planking. Such a technological phase would be equivalent to
what is thought to have been the earliest phase of plank-boatbuilding in northwest
Europe and in southeast Asia (McGrail 2004a: 184-94; 297-98, 304-5).
Propulsion and Steering

Oars were excavated with the Cheops ship, and paddles are depicted in tomb
paintings and with models from the Old Kingdom (ca. 2600 B.C.) onward. There are
also many illustrations of rowing in the sit-pull and stand-pull modes, oars being
pivoted in grommets made fast to the vessel's sides. The earliest evidence for sail
appears on a vase from N aqada of about 3100 B.C.: this is a single square sail on a pole
mast stepped toward the bow (McGrail 2004a: fig. 2.5). Later depictions mostly show
a square sail with a boom at its foot on a bipod mast stepped about one-third the
waterline length from the bow, which, generally speaking, favors sailing with a wind
in the stern sector. On the Nile, with a predominant northerly wind, sailing upstream
with a following wind would have been the normal practice; at sea, however, a
requirement to have the wind from astern could have been restrictive. By the mid-
second millennium B.C. the pole mast stepped nearer amidships is almost exclusively
depicted; it may have been possible to sail such vessels in other than following winds.
The standing rigging consisted of a forestay and a backstay, with auxiliary
backstays that may have been moved abreast the mast to act as windward shrouds.
Sheets are sometimes illustrated; braces are shown running from both yardarms to
the helmsman, and there is frequently a bowline from the sail's leading edge to the
bows. Braces and bowline would have been used to trim the sail to a wind that was
from forward of the stern sector. The use of a forked pole (a "tacking" spar) to bear
out the sail's leading edge also indicates the ability to sail on a reach, with the wind
on the beam. No block and tackles are depicted, so halyards and other running
rigging were probably directly hauled and veered, possibly through greased holes in
the mast. During the second millennium lifts were rigged to give yard and sail
further support. The earliest depictions of sail in Mediterranean waters are ships
with mast and rigging engraved on small Minoan stone seals of around 2000 B.C.
(McGrail 2004a: fig. 4.16)
Vessels may be steered by paddle, steering oar, side rudder, or centerline rudder,
and there are Old Kingdom (2686-2160 B.C.) depictions of each method. During
the later third millennium, two side rudders were used, each pivoted on the vessel's
quarter and against a vertical stanchion (Landstrom 1970: fig. 152): they were thus
true rudders that were rotated around their own long axis. By the first millennium
B.C., such paired side rudders had become standard fittings in the eastern Medi-
terranean. Egyptian depictions of a centerline rudder (which does not seem to have
been used in the Mediterranean until late medieval times) became increasingly
common by the end of the third millennium B.C. (McGrail 2004a: 34, fig. 2.18).
SHIPBUILDING 800 B.C. TO A.D. 500
Eastern Mediterranean shipbuilders and seamen of the early eighth century B.C.
inherited much from their Egyptian and Phoenician predecessors: the rivers, coasts,
and seas of the eastern and central Mediterranean, the Black Sea, and the Red Sea
were well known, and the art of navigating without instruments, when out of sight
of land, had been learned. All the basic types of water transport and all forms of
propulsion and steering had been devised. Planked vessels were built in the plank-
first sequence, the planks being aligned by treenails within the plank edges before
being fastened together using two types of fastening-sewing and locked mortise
and tenon. The vessel's framing was then lashed to the planking.
Ships in Homer>s Poems

The method for building a seagoing vessel in Homer's time is described in a passage
in the Odyssey (5.259-87) that is both enigmatic and incomplete, doubtless because
Homer was a poet and not a boatwright. Even if agreement could be reached on the
precise meaning of key words in Homer's text, knowledge of ancient plank fastening
techniques is needed to fill the gaps in his description. Furthermore, the building
methods described have been interpreted as both those of a vessel fastened by locked
mortise and tenon joints (Casson 1964; 1992) and those ofa sewn-plank vessel (Mark
1991; 1996; 2005: 25-39). A third possibility, which could resolve this matter, is that
Homer is describing the construction of a boat with both types of fastening, as in
earlier Egyptian ships (see above), and as shown on the possibly Syrian ships (iden-
tified as such by the dress of crew and merchants) depicted in the mid-second
millennium tomb of Kenamun at Egyptian Thebes (Wachsmann 1998: fig. 3.5).
Early Use of Locked Mortise and Tenon Fastenings

The fragmentary remains of the Uluburun ship's planking, dated to about 1300
B.C., include the earliest known nautical use of locked mortise and tenon fasten-
ings (Pulak 2003). Whether this planking was first positioned by treenails set into
the plank edges, whether sewn fastenings were also used, and whether frames were
lashed to the planking is not known due to the limited extent of the surviving
planking. It is conceivable, however, that the Uluburun ship had sewn fastenings at
her (nonsurviving) ends, similar to those depicted on the near-contemporary
representations in Kenamun's tomb. Locked mortise and tenon fastenings (figure
24.3) are also thought to have fastened the planking of the wreck from Cape Ge-
lidonya dated to around 1200 B.C (Bass 1991), and they were used in the seventh-
century B.C. Mazarr6n I wreck from Playa de la Isla, off southeast Spain (Ne-
guerueala et al. 1995). The frames of the latter ship were lashed to the planking. It is
not yet clear whether she had sewn-plank fastenings in addition to locked mortise
and tenon joints, as research is still in progress, but caulking in the seams appears to
have been held in place by a light cord through sewing holes (Mark 2005: fig. 12). It
seems possible that the combination of sewing with locked, watertight mortise and
tenon fastenings was a second-millennium B.C. Phoenician development of es-
tablished Egyptian practices, but this cannot yet be demonstrated. Locked mortise
and tenon joints, like clenched nails, are positive fastenings on both sides of a seam,
and do not need reinforcing by a second type of fastening. Moreover, when fitted
tightly (in the ship's longitudinal direction) within their mortises, they have a
significant resistance to shear forces (Coates 1996, 2001).
Sewn Planking
A wide range of sewn-plank fastenings, dated from around 3000 B.C. to the present
day, is known throughout the world (McGrail 1996b). Some of these consist of a
series of individual stitches or lashings, as found in the Bronze Age Ferriby boats of
the Humber estuary on the east coast of England (Wright 1990: 65, figs. 2.3, 3,10,
4.8), but the great majority involves continuous sewing. Boats with sewn planking
are mentioned by Homer (II. 2.133; Od. 5. 259-87) in the eighth century B.C., and by
later authors ranging from the fifth century B.C. (Aeschylus, Supp. 134-35) to the
fifth century A.D. (St. Jerome, Ep. 128.3): whether individual stitches or continuous
sewing is unclear. These references are mainly within a Greek context.
No Mediterranean vessels with planking fastened by individual stitches or
lashings have been excavated, but four Mediterranean vessels have continuous
sewing wedged within the sewing holes. This sewing is along plank seams, not
transversely across the hull as in earlier Egyptian vessels. It is found on the ships
from Giglio, off the west coast ofltaly (ca. 580 B.C.; Bound 1991); Bon Porte, off the
south coast of France (540-510 B.C.; Basch 1976; Pomey 1981, 1985, 1997); Cala Sant
Vicen~ off the island ofMallorca (530-500; Kahanov and Pomey 2004: 14-15); and
Place Jules-Verne 9, Marseilles (525-510 B.C.; Pomey 1995, 1996, 1999). Enough of
the Place Jules-Verne Wreck's planking has survived to make it probable that she
was entirely sewn, and the others may have been so as well.
There are also several wrecks with planking mainly fastened by locked mortise
and tenon joints that also had wedged, sewn-plank fastenings in the lower hull, at
bow and stern, or in repairs: Place Jules-Verne 7 and Cesar 1, both of Marseilles and
dated 525-510 B.C.; Gela 1, off southern Sicily of 500-480 B.C. (Kahanov and Pomey
2004: table 2; Pomey 1997; Freschi 1991); Grand Ribaud F off the south coast of
France of 515-470 B.C. (Kahanov and Pomey 2004: 17-18); Gela 2 of 450-425 B.C.;
and Ma'agan Mikhael off the Israeli coast of about 400 B.C. (Linder and Kaha-
nov2003; Kahanov and Linder 2004). Except for Place Jules-Verne 7, Cesari, Gela 2,
and Ma'agan Mikhael, vessels in these two groups had their planking positioned by
shear-resisting treenails within plank edges. Most of the ships with sewn planking
had some form of caulking material associated with their stitching, and the stitches
of Ma'agan Mikhael were wedged and sealed with a resin within their holes.
The frames of those ships solely fastened by sewing were widely spaced, at
around 0.96 m; their undersurfaces were fashioned to fit over the stitching, and they
were lashed to the planking. The floors and futtocks of the Place Jules-Verne W reek
were nailed (the earliest known use); the top timbers were lashed and treenailed; the
frames of the Cesar 1, Grand Ribaud F, and Gela 1 Wrecks were nailed. The Place
Jules-Verne Wreck's nails, and the copper nails used in the Ma'agan Mikhael ship,
were hook-clenched by turning the tip through 180 degrees, back into the inboard
face of the frame. All these ships had a keel or plank-keel, and all had a full,
transversely rounded hull except the Gela 2 and Ma'agan Mikhael Wrecks, which,
like the Kyrenia ship of about 300 B.C., had wineglass-shaped cross-sections (Ka-
hanov 2003: table 31; Kahanov and Pomey 2004: 18-19).
Ten sewn plank boats, dated from the late centuries B.C. to the eleventh century
A.D., have been excavated from lakes and rivers in Croatia and the Po delta (Pomey
1981; Brusic and Domjan 1985; Beltrame 2000; Kahanov and Linder 2004: 66-76).
The planking of these Adriatic vessels was fastened together by continuous sewing;
nails were additionally used in two of the Po boats. Treenails within plank edges
were noted on the Ljubljana boat of around 200 B.C., but not elsewhere; they may
have been unnecessary in boats not intended to be seagoing. The stitches of the Nin/
Zaton boats of Croatia, from the early centuries A.D., were wedged within sewing
holes which were then blocked with resin and the outboard part of the stitch cut
away-a technique used in the twentieth century in seagoing boats from the Lamu
archipelago off east Africa so that stitching would not be damaged when a boat took
the ground or was dragged across the foreshore (Prins 1965).
Warships with Rams

The earliest depictions of warships with rams are found on Greek Geometric pot-
tery of the ninth to eighth centuries B.C. (McGrail 2004a: fig. 4.24). Although there
is some disagreement as to the purpose of these projections, they probably repre-
sent pointed rams intended to hole planking at or below the waterline. Blunt-ended
rams are depicted during the sixth century (Casson and Linder 1991: fig. 5.4), and
trident-headed rams appear about 400 B.C. A 600 kg bronze ram found on the
foreshore at Athlit, south of Haifa, Israel, is of this latter type and is dated by its
decorations to the first half of the second century B.C. (Murray 1991).
John Coates, naval architect of the Olympias project, considers that, since a
warship's retardation when piercing a hole with a sharp ram would not have been
excessive, the early, pointed rams could have been used by sewn-plank ships (per-
sonal communication). Blunt and finned rams, on the other hand, when not used
to break the oars of the target ship, may have been used to thump the planking and
start the seams: the effects of this on an attacking ship are disputed, but it is con-
ceivable that she would have needed the stiffness and strength given by closely
spaced mortise and tenon joints rather than sewn-plank fastenings (Steffy 1991).
Propulsion and Steering

In Homer's time a single square sail on a mast stepped near amidships was probably
the only sailing rig. From the late sixth century B.C. onward, occasional ships are
depicted with a second square sail on a foremast (Casson 1996: fig. 97), or, from the
first century A.D., as an artemon at the bow. A third sail on a mizzen is described in
the mid-third century B.C. (Casson 1996: 240). Fore-and-aft sails are illustrated on
small craft; a spritsail from the second century B.C. (Casson 1996: figs. 175-79), and
a lateen from the second century A.D. (Casson 1996: fig. 181). Sails were probably
made oflinen, since Ezekiel, writing in the late-sixth century B.C., noted that sails of
this material were used on Phoenician ships (27:7), and a second-century B.C. linen
shroud fragment with an attached wooden brail ring, from the temple at Edfu,
Egypt, has been identified as part of a sail (Black 1996).
Rigged with square mainsail and foresail, and in good sailing conditions, the
closest to the wind that the experimental trireme Olympias could sail was with the
wind one point or so forward of the beam, that is, some seven points (ca. 79°) off
the wind (Roberts 2005.). If this observation discounted the leeway of one point
(ca. 11°) that was also noted, it may be concluded that an ancient Mediterranean
ship similar to Olympias would probably have been able to make good a track
across the wind and return on the reciprocal track. Ships were generally fitted with
a side rudder on each quarter (Casson 1996: figs. 147-51). Sea trials of Olympias
showed that the rudders induced such a disproportionate amount of drag that only
one of the pair was generally used and that was partly immersed. It may be,
however, that, a redesign of the trial rudders would have minimized this problem.
TECHNOLOGICAL CHANGES-800 TO
300 B.C.
In the eastern Mediterranean we know of two ships before 800 B.C. (Uluburun
and Cape Gelidonya Wrecks, with possible origins on the Levant coast) that were
built in the plank-first sequence and had locked mortise and tenon plank fasten-
ings, a woodworking technique that was earlier known in Egypt and in Phoenicia.
The seventh-century B.C. Mazarr6n 1 ship has this type of plank fastening and
also lashed framing; both of these features may well have been characteristics of
Phoenician-built ships of the first, and possibly second, millennium B.C. Whether
these ships also had sewn plank fastenings (as did later vessels) is unknown, but
such plank fastenings, with frames lashed to the planking, are likely to have been
used from an early date throughout the Mediterranean. It also seems likely that
seagoing, sewn-plank vessels of those times would have had treenails across seams,
within the plank thickness. These would not only have helped in the positioning of
adjacent strakes, initially and after periodic stitching renewal, but would also have
minimized shearing stresses on the stitching when afloat, especially if the rays of the
treenails were aligned with the plank seams (Coates 1985, 2001).
One interpretation of a passage of Homer, Odyssey 5.259-87, suggests that
Greek ships of around 800 B.C. had sewn planking (Mark 2005); authors in later
centuries appear to confirm that stitching was a characteristic of early Greek-built
ships. An alternative hypothesis is that Homer's ships had both sewn and mortise
and tenon fastenings. Four seagoing ships (the Giglio, Bon Porte, Cala San Vicenc;:
and Place Jules-Verne 9 Wrecks-all of which could have been Greek-built), dated
between 580 and 510 B.C., had sewn planking, as did a dozen or so boats of the late
centuries B.C. to the eleventh century A.D. from Adriatic lakes and rivers. Fur-
thermore, four seagoing vessels dated between 525 and 480 B.C. (Place Jules-Verne
7, Cesar 1, Grand Ribaud F, and Gela 1 Wrecks), Gela 2 Wreck of 450-425 B.C., and
the Ma'agan Mikhael Wreck of around 400 B.C., all with nail-fastened framing, and
planking mainly fastened by mortise and tenon joints, had sewn-plank fastenings at
the ends, in the lower hull, or in repairs.
Leaving aside the somewhat isolated Adriatic boats, the sequential nature of
these two groups suggests that, from about 525 B.C., Greek builders of seagoing
ships changed from sewn-plank fastenings to (Phoenician?) locked mortise and
twJI Fl E-1 Fl Fl f
MIOSHIP SECTION
.STSO PORT
Figure 24.5. Transverse half-section of the Kyrenia ship showing structural details. (After
Steffy 1994: fig. 3.31.)
tenon fastenings, supplemented by sewn fastenings in vulnerable parts of the hull.

They also stopped using treenails across plank seams, decreased frame spacing
(center to center) from 96-84 cm to 75-70 cm, and fastened frames to planking by
hook-clenched nails rather than by lashings. The earliest evidence for hulls with a
wineglass-shaped cross-section appears a century later, in the second half of the
fifth century. Similarly, present evidence shows no ship's planking solely fastened
by locked mortise and tenon joints until around 300 B.C.
The Kyrenia Ship

This merchant ship of about 300 B.C. was excavated off the north coast of Cyprus in
1968/69 (Steffy 1985, 1989, 1994: 42-59; 1995). The remains were lifted and conserved,
and are now on display in Kyrenia. The pine planking was fastened together with
locked mortise and tenon joints; there was no sewing (although sewing holes were
found in one reused ceiling plank) and there were no treenails in the seams. The
framing, closely spaced at 25 cm, consisted of floor timbers (with chocks) spanning
six to nine strakes each side, alternating with paired half-frames that did not quite
reach the keel (figure 24.5). Side timbers were added in line with, but not fastened
to, the floors and half-frames, and there were occasional top timbers. Copper nails,
possibly driven through treenails, and clenched by hooking the point through 180
degrees back into the timber, were used to fasten framing elements to the planking.
This framing pattern became characteristic of Mediterranean hulls until the
changes that took place in the fourth to sixth centuries A.O. as a prelude to the shift
from the plank-first to the frame-first sequence of building.
Like the Gela 2 and Ma'agan Mikhael ships, the Kyrenia ship's underwater hull
had a wineglass shape: that is, the lowest three or four strakes, seen in section,
formed a reverse curve to the keel so that the garboards were nearer the vertical than
in earlier ships, thereby enhancing the leeway-resisting properties of the underwater
hull and keel. The rabbets worked along the length of the keel to take the garboard
strakes, and the lead sheathing of the planking, were innovations.
TECHNOLOGICAL CHANGES
AFTER 300 B.C.
Tentative conclusions may be drawn from reports on a number of excavated

merchant ships dated from the second century B.C. to the second century A.D.,
although the data from them can, in some cases, be challenged (McGrail 2004a:
158). These large ships had cargo capacities ranging up to 600 tons. They had double
planking, keels with complex scarfs and of greater dimensions than earlier ones, and
a greater than proportionate increase in tenon breadths: these features greatly
added to hull strength and integrity. The fitting of prototype stringers and elon-
gated mast-step timbers (effectively keelsons), and the fastening of some frames to
the keel, further added to structural strength.
After the second century A.D., wrecks of such large vessels are rare, but six
smaller vessels, from the fourth to the seventh century, show a shift in emphasis from
plank fastenings toward increased reliance on framing, keelson, ceiling planking,
and decking. The hull of the fourth century Yassi Ada wreck 2 (van Doorninck 1976)
was, in many ways, conventional for her date: she was built plank-first, the planking
was fastened by locked mortise and tenon joints, and the floors alternated with half-
frames. The joints, however, were smaller than in earlier ships of comparable size,
and tenons did not fill mortises, leaving gaps of 6-7 mm at each end and 17-18 mm at
the sides; they were also more widely spaced, intervals being 320 mm in places (Steffy
1994: 79-80, fig. 4.1).
Five other ships, dated to the sixth and seventh centuries, show further signs
of change. The Dor D Wreck off the coast of Israel (A.D. 550-650; Royal and
Kahanov 2005), St. Gervais 2 Wreck of the coast of southern France (A.D. 600-625;
Jezegou 1989), Yassi Ada Wreck 1 off the Turkish coast (seventh century; van
Doorninck 1982), and Pantano Longarini Wreck from Sicily (seventh century;
Throckmorton 1973) still retained some edge-fastened planking and therefore were
probably built plank-first. Nevertheless, their mortise and tenon joints were mostly
not locked, the tenons of those joints did not fill their mortises, and plank
thicknesses decreased (Royal 2002).
Steffy (1994: 77-78, 83-85; 1995: 27) has suggested that these changes were
the result of labor becoming more expensive due to the decline in slavery and to
the growing importance of independent ship owners who preferred smaller, less
costly ships. The classic style of shipbuilding, as seen in the Kyrenia Wreck, was
skilled-labor-intensive, especially in the fashioning, fitting and fastening of the

thousands of locked mortise and tenon joints needed in even a moderate-sized
hull-so fine were woodworking skills that caulking was generally not needed. A
reduction in the need for such joints resulted in fewer skilled shipwrights being
employed and thus in reduced costs.
In retrospect, these changes can be seen as the beginning of a decline in building
plank-first. Indeed, two fifth- to sixth-century A.D. ships from the Israeli coast, Dor
2001/1 and Tantura A, had no edge-fastened planking at all, and were built in the
framing-first sequence (Kahanov 2001, 2003). There was caulking material within
their seams, but mortise and tenon joints, sewing, and treenails were not found
there. Furthermore, frames were nailed to the keel, and planking was butt-scarfed
and nailed only at frame stations. These ships had been built in the following se-
quence: keel, posts, and elements of the framing were first set up and fastened
together; then planking was fashioned and fastened to this framework; more fram-
ing and more planking followed. Although different in detail, this is similar to the
building sequence used to build two ships and a seagoing boat of the Romano-
Celtic tradition in second and third century A.D. northwest Europe (see below).
When using the plank-first sequence, each individual builder fashioned and
fastened together planking so that a hull shape was achieved which rules of thumb
and his experienced eye told him was appropriate to the role the ship was to
undertake. In the framing-first sequence, on the other hand, it was the framework
of posts, keel, and selected frames that determined hull shape. For anything bigger
than a boat, that framework had to be designed. How that was done in the mid-
first millennium A.D. is uncertain, but the method may have been similar to that
subsequently used in Venetian shipyards from the fifteenth century (McGrail 2003:
128; 2004a: 164-65). In these later ships, the main dimensions were defined as pro-
portions of a modular unit, which was usually either the keel length (Bellabarba
1993: 274) or the maximum beam (Steffy 1994: 93). The shape of the master frame
(which was set up near amidships) was given by a traditional rule that specified the
orthogonal coordinates of the required curve at four points (Ballabarba 1993: fig.
3). The shapes of the remaining designed frames were derived from the shape of the
master frame using a wooden tablet (known as a mezza luna, "half moon") and a
measuring stick. As the art and science of generating hull shapes developed and as
understanding increased of the effects of shape changes on ship performance,
improvements could readily be introduced into a design, and a successful design
could be repeated.
This change to non-edge-fastened planking and the framing-first sequence of
building was a fundamental technological shift. Subsequent to the fifth- to sixth-
century Dor 2001/1 and Tantura A Wrecks, this technique is seen in the eighth-
century Tantura F and the ninth-century Tantura B Wrecks (Kahanov 2003), and
in the wreck from Serc;:e Limani off the Turkish coast dated to about A.D. 1025
(Steffy 1994: 85-91). Subsequently, ships of the Iberian-Atlantic tradition of the
fifteenth century and later were built in this manner (Alves 2001).
BOATS AND SHIPS OF NORTHWEST

EUROPE
Mediterranean-Style Vessels
A number of boats and ships of the first few centuries A.D. so far excavated in
northwest Europe had been built in Mediterranean fashion, with planking fastened
by locked mortise and tenon joints. Broad oak washstrakes had been fastened by
such joints to the logboat base of a third/fourth century A.D. boat excavated in 1987
from Lough Lene, Co. Westmeath, some fifty miles northwest of Dublin (6
hEailidhe 1992). The remains of a ship dated to A.D. 290-300 were recovered from
the County Hall site near the River Thames in London in 1910/11. Her fittings and
features are almost entirely Roman, but dendrochronological examination has
shown that she was built of oak from southeast England (Marsden 1994: 109-29).
The elements of a steering oar blade and two boats (2A and 6), all from Zwam-
merdam in the Netherlands, and a boat from Vechten near Utrecht, had some
mortise and tenon fastenings (de Weerd 1988; de Weerd and Haalebos 1973). All
these had been built locally, possibly by indigenous people under the supervision of
Mediterranean shipwrights or influenced by observing their ships.
Romano-Celtic Boats and Ships

During the late twentieth century, approximately 20 vessels dated from the first to
fourth centuries A.D. were excavated at sites in an arc stretching from the Swiss
lakes, along the River Rhine, to the Thames and the Severn, and on to Guernsey off
the northwest coast of France. Most of these vessels are of elongated box shape and
come from the Rhine region: they were probably river and canal barges (McGrail
2004a: 201-5). Seagoing vessels of this tradition, with plank-keels, raked posts, and
a full, round hull but flat in the floor, have been excavated from three sites: the
Thames in London in 1962 (Blackfriars 1 Wreck, ca. A.D. 150; Marsden 1994: 33-95);
the main harbor of Guernsey in 1985 (St. Peter Port 1 W reek, ca. A.D. 275; Rule and
Monaghan 1993); and the northern shores of the Severn estuary, southeast Wales in
1993 (Barland's Farm Wreck, ca. A.D. 300; Nayling and McGrail 2004).
Several characteristics of these vessels echo Julius Caesar's first-century B.C.
description of the ships of the Veneti, a Celtic seafaring people of southwest Brittany
(B Gall. 3.13; cf. Strabo 4.4.1). It seems likely that the ships of the Veneti were
forerunners of the excavated vessels we now know as members of the Romano-Celtic
shipbuilding tradition: Celtic because they were excavated from sites which were
formerly inhabited by Celtic-speaking people; Roman because all known finds are
dated to the late Roman period. The latter term also acknowledges that some of
the techniques used by the Celtic shipwrights originated in the Mediterranean. These
vessels were built from oak timbers that had been sawn to shape-the main, probably
the only significant, use of Roman technology. Some of the planking of two of the
Rhine barges is fastened by locked mortise and tenon joints, another borrowing.
There are other features comparable with those found in contemporary Medi-
terranean vessels, but whether they were derived from Roman technology is un-
clear. Romano-Celtic planking was fastened to relatively massive frames by great
iron nails (the size of Veneti frames and timbers was noted by Caesar), each one
driven through a treenail inserted in the frame, and clenched by hooking the nail
tip back into the inboard face of the framing timber (figure 24.6). This is com-
parable with the Mediterranean technique of fastening framing to planking using
hooked copper nails (see above). Clenching a nail, by hooking the tip through two
90-degree bends and hammering it down on the inside surface of the strake or
frame through which it had been driven, is first known in the Mediterranean on the
Place Jules-Verne 7 vessel of 525-510 B.C. Arnold (1999: 42, fig. 7), however, has
drawn attention to the north European use of turned and hooked iron nails in the
second century B.C. murus gallicus and in Halstatt cart wheels before 500 B.C.
Technological transfer could have taken place in either direction along the Rhone-
Seine or the Po-Rhine routes that connected the North Sea and the Mediterranean.
There are differences between the two technologies: Celtic boat builders plugged
nail holes with treenails before driving nails through them, and they used iron nails to
fasten planking to frames rather than copper nails to fasten frames to planking. No
elements of Celtic framing were joined together, whereas the Mediterranean tradition
from the early sixth century B.C. involved scarfing futtocks to floors (see above); the
form of the Celtic bow and stern posts, and the scarf joints between posts and plank-
keel are unlike anything known elsewhere. Furthermore, the Celts stepped their masts
about one-third of the underwater length of the boat from the bow, whereas Roman
masts were nearer amidships.
The most innovative characteristic of the Celtic seagoing vessels of A.D. 175 to
300 was that they were built in the framing-first sequence: the flush-laid planking,
caulked with macerated wood, was not fastened together, but each individual plank
was fastened to the frames. This is comparable with the framing-first building
sequence of the fifth- or sixth-century A.D. Dor 2001/1 and Tantura A ships from the
Israeli coast (see above). On present evidence, Romano-Celtic use of the framing-
first sequence is at least two hundred and may be four hundred years earlier than its
first known use in the Mediterranean. Was this a case of technological transfer of
what turned out to be a very significant innovation, or was it independent in-
vention? A due to the origins of this technique may lie in prehistoric times. Well
before Caesar noted the Veneti ships, perhaps as early as the second millennium
B.C., frame-first techniques must have been used in Celtic northwest Europe to
build seagoing hide boats, skin over framework (McGrail 1990: 36-39).
In the medieval Mediterranean, the framing-first sequence of building ships
became dominant except, perhaps, in smaller boatyards. After the fourth century
A.D., however, the Romano-Celtic style of shipbuilding does not feature in the
Treenail
Frame---
Plank
Figure 24.6. Method of fastening planking to framing in Romano-Celtic vessels. (After

Nayling and McGrail 2004: fig. 7.3.)
archaeological record of northwest Europe, and the vast majority of early medi-
eval finds are of the Nordic, clinker-built, tradition. It remains to be seen whether
the framing-first technique lived on somewhere in northwest Europe through the
migration period, a time of significant cultural change. The sixth or seventh century
Port Berteau 2 Wreck, coastal and river boat, may be evidence for such survival.
This boat, excavated from the River Charente on the west coast of France, had
flush-laid planking which was not fastened together but treenailed to the frames
(Rieth 2000; Rieth et al. 2001). Unlike the seagoing Romano-Celtic vessels, she had
composite frames with futtocks fastened to floor timbers (Rieth et al. 2001: fig. 112) .
She may, therefore, have been built frame-first, in the strict sense of the term: that
is, the entire framework may have been fashioned and fastened to keel and stems
before any planking was positioned.
The framing-first or frame-first caravella, built by Portuguese shipbuilders in
the early fifteenth century, were the predecessors of the ships in which Bartolomeu
Dias, Vasco da Gama, Christopher Columbus, and Ferdinand Magellan explored
the world. Whether this Atlantic version of frame-first building was inherited from
the Mediterranean or from Celtic northwest Europe is the subject of continuing
research.
THE TECHNOLOGICAL ACHIEVEMENTS

OF THE CLASSICAL WORLD
Mediterranean shipbuilders and seamen of 800 B.C. to A.O. 500 consolidated and
developed their inheritance from Egypt and the Levant coast: the warships and cargo
ships of the fourth century B.C. and later were a technological advance on the ships
of their predecessors. From the sixth century B.C. locked mortise and tenon fas-
tenings began to replace sewn fastenings, at first in the Greek, then the Roman
world. In the early centuries A.D. the use of this distinctive joint spread to northwest
Europe, even to Ireland outside the empire. It was probably also transmitted across
the Indian Ocean around this time, since such fastenings have been excavated in
northern Vietnam on a first-century B.C. plank-extended log boat from Dong Xa,
and on planking reused as a coffin at Yen Bae in the second century A.D. (Bellwood
et al. 2007). Locked mortise and tenon fastenings were also found in 1953 on an
undated wreck excavated at Johore Lama in Malaya (McGrail 2004<1: 302, 436).
In the first few centuries A.D., Mediterranean shipwrights abandoned plank
fastenings and shifted to a framing-first sequence of shipbuilding, Whether or not
these techniques had originated with northwest European shipwrights, this sig-
nificant technological change seems to have been perfected within the medieval
Mediterranean, and may have led to the worldwide seafaring achievements of the
fifteenth- and sixteenth-century shipbuilders and mariners of the Iberian Atlantic
coast. The social, political, and economic effects of these overseas voyages of ex-
ploration are still in evidence today.
NAVIGATION BEFORE A.D. 500
In conditions of good visibility, all islands in the Mediterranean are visible from
high ground on the European, African or Asian mainland, or from another island
which is so visible, or from a boat when land earlier sighted is still in view (Henkel
1901: fig. 1), and most of these islands had been colonized by the end of Neolithic
times. During the midsummer season, a boat under oars could have undertaken
such voyages during daylight-even the longer passages to Cyprus, Malta, and the
Balearic islands. In such conditions, when land, astern or ahead, is always in sight,
the seaman uses pilotage techniques to note his position relative to a succession of
recognizable landmarks. In other seasons, and in less than good weather, with land
no longer visible, it seems likely that these early seafarers evolved simple naviga-
tional techniques such as estimating directions from the pole star, or, with less
accuracy, from the sun at noon, or relative to the wind and swell.
The use of the sounding lead and line on a boat approaching the Nile delta is
described by Herodotus (2.5), and many leads have been found in inshore waters,
the earliest being from the Gela 1 Wreck of 500-480 B.C. (Oleson 2000). Lucan
(8.177-81) tells us that, in his day, the zenith altitude of certain stars was estimated
against the mast (presumably in harbor); such an estimate gave a measure oflati-
tude relative to other harbors where similar estimates had been made (Taylor 1971:
47-48). A unique mechanism, excavated in 1900 from a first-century B.C. wreck off
Antikythera, has bronze gear wheels which appear to have been capable of tracing
out the relative movements of sun, moon, and certain stars (figure 29.1). Rather
than a navigational instrument, this is now interpreted as an instructional aid,
possibly for astrological calculations (de Sola Price 1974; Johnston 1997; see also
chapter 29). It is not until the ninth century A.D. that we learn of instrumental aids
to celestial navigation: an Arabian author, al-Khwarizmi, described a staff for mea-

suring star altitudes at sea (Aleem 1980: 588; Fatimi 1996). The use of this proto-
kamal may well have been based on the Greek angle-measuring device known as a
dioptra (M. J. T. Lewis 2001).
Thus, apart from the sounding lead and line, no navigational instruments were
used during our period. Today, seamen out of sight of land in several regions use
noninstrumental navigational methods based on inherited traditions, personal ex-
perience and detailed observation of the maritime environment (McGrail 1998: 275-
76; Morton 2001). Furthermore, Homer's Odyssey contains descriptions of open sea
passages undertaken without instruments (3,188-95; 14.285-91). Although Odysseus
had no chart, he would have had a mental map in his head on which he "plotted" the
spatial relationships of the coastal lands and islands of the eastern Mediterranean
(McGrail 1996a; Mark 2005: 138-52). For example, he used the Great Bear constel-
lation (Ursa Major) to steer an easterly heading when leaving Calypso's island, and
monitored the rising and setting of Orion, Arcturus and the Pleiades (Homer, Od.
5.298-304). Odysseus' use of this star compass suggests that he had a good working
knowledge of astronomy (figure 24.7). Furthermore, he also used a wind compass
(Homer, Od. 12.313, 350-51; 14.286, 337-38, 520). Eight elements of such a compass
were depicted on the first-century B.C. Tower of the Winds in Athens (figure 29.3;
McGrail 2004a: fig. 4.4). As in many maritime cultures, Odysseus' standard unit of
distance was "a day's sail" (Homer, Od. 4.397), and the passage of time at night was
marked by certain stars reaching their zenith (Taylor 1971: 48).
When on passage out of sight of land, Odysseus would have frequently used
these environmentally based aids, together with his assessment of his boat's speed,
to plot on his mental chart his estimated position relative to his home port or to
his destination. When land was eventually sighted, after such a passage, Odysseus
was able to identify the particular coastal landscape (for example: Pharos island,
Homer, Od. 4.395-96, and the "wooded peak of windswept Neriton," Od. 9.23-24)
and could then deduce the route to his intended destination.
From Homer onward, Greek and Roman writers recognized Phoenician sea-
faring abilities. Strabo (1.1.6) gave them precedence in navigation, and Pliny (HN
7.57) claimed that they were first to apply astronomical learning gained from the
Chaldeans to navigation at sea. Phoenician seamen realized that the constellation
Ursa Minor orbited the celestial North Pole in a tighter circle than did Ursa Ma-
jor; they therefore used Ursa Minor to give them a more precise direction of north.
The classical world came to call this useful, but not easily recognized, constellation
Phoinike (Aubet 1993: 142).
In the fourth century B.C., written sailing directions (periploi) for coastal voy-
ages began to be compiled from oral accounts (Dilke 1985). These were not just for
the Mediterranean, but also for a passage from Atlantic Europe to Massilia (Mar-
seilles; Murphy 1977: 7-9), passages along the coast of northwest Africa (Murphy
1977: 9-11; Oikonomides and Miller 1995), and from the Egyptian Red Sea harbor of
Myos Hormos on one of two routes: either around the Horn of Africa and south to
the Zanzibar region, or past the Persian Gulf and on to India and beyond. This last
compilation, the Periplus Maris Erythraei (Casson 1989), written in Greek by an
I
\ / I
Figure 24.7. Diagram illustrating steering by the stars in the South Pacific. (After Lewis
1994: fig. 14; courtesy of University of Hawaii Press.)
Alexandrian merchant in about 50 A.D. (Reade 1996: 312), was a combination of

what we nowadays would call sailing directions, a regional handbook, and a trad-
ing guide, useful both to pilots and merchants. Early Mediterranean seaman sailed
these waters guided by oral versions of such periploi, and with the human eye and
brain as their only navigational instruments.
After A.D. 500, Arab astronomers and seamen brought to the Mediterranean
navigational ideas and techniques they had earlier learned from the Greeks. Thus
were laid the mathematical foundations of the shift to navigation by instruments
and charts that took place in the Mediterranean from the thirteenth century on-
ward (Taylor 1971: 89-121) .
GLOSSARY OF NAUTICAL AND

NAVIGATIONAL TERMS
bilge: region between sides and bottom of a vessel.

bowline: line used to keep the weather edge of a sail taut.
brace: line used to trim a yard.
brail: ropes used to bundle a sail rapidly.
carling: fore-and-aft timber at deck level.

caulking: material inserted between two timbers to make the junction wa-
tertight.
cramp: wooden fitting which draws together two timbers across a seam.
floor: transverse timber set against the planking and extending from tum of
bilge to tum of bilge.
fore-and-aft sail: sail that is generally set in or near the fore-and-aft line of a
vessel.
frame: transverse member made up of more than one timber.
frame-first: building sequence in which a framework of keel, posts, and frames
is set up before planking is fashioned.
framing-first: building sequence in which keel, posts, and elements of the
framing are set up before planking is fashioned.
freeboard: height of sides above the waterline.
futtock: pairs of timbers which, with a floor, constitute a frame.
garboard: strake next to the keel.
grommet: strands of rope layed up in the form of a ring.
halyard: line to hoist and lower yard and sail.
kamal: navigation instrument consisting of a wooden tablet at the end of a line
that had knots tied in it at graduated intervals. Evolved in the Arab world in
the ninth century A.D. and used to measure star altitudes (angle above
horizon).
keelson: center-line timber on top of the floors, adding to longitudinal strength
and stiffness.
lateen sail: triangular, fore-and-aft sail bent to a long yard.
leeway: the distance or angle a vessel is set down wind.
lift: line running from yardarm to mast.
mast step: fitting used to locate the heel of a mast.
mizzen: aftermost mast in a three-masted vessel
pilotage: techniques used to keep a vessel's reckoning when in sight of land.
plank-first: building sequence in which the planking is (partly) erected and
fastened before framing is inserted.
plank-keel: a keel-like timber with a low cross section.
rocker: for-and-aft curvature of keel or bottom of vessel.
settee: quadrilateral lateen with a short leading edge.
sheer(line): curve of the upper edge of the hull.
sheet: a rope that regulates the angle at which a sail is set to the wind.
shroud: line leading from masthead to the sides of a vessel to support the mast
athwartships.
sprit sail: four-sided, fore-and-aft sail set on a sprit or spar, the lower end of
which is made fast to the mast while its upper end supports the peak of the
sail.
square sail: four-sided sail that is laced to a yard which lies square (i.e., at right
angles) to the mast.
stanchion: an upright brace or support for a ship's deck.

stay: line leading from masthead forward and aft to support the mast.
strake: single plank or combination of planks that extends from one end of a
vessel to the other.
tidal range: the difference between the height of high water and the next low
water.
timber: any piece of wood used in boatbuilding.
treenail: wooden peg or dowel used to join two timbers.
washstrake: additional strake fitted to increase freeboard.
yard: spar suspended from a mast, and to which the head of a square sail is
bent.
zenith: highest point of a heavenly body's trajectory.
REFERENCES
Aleem, A. A. 1980. "On the history of Arab navigation," in M. Sears and D. Merriman
(eds.), Oceanography: The past. New York: Springer Verlag, 582-95.
Alves, F. ( ed.) 2001. Proceedings of a symposium on archaeology of medieval and modern ships
of Iberian-Atlantic tradition. Trabalhos de Arqueologia 18. Lisbon: Instituto Portugues
de Arqueologia.
Arnold, B. 1999. "Some remarks on Romano-Celtic boat construction and Bronze Age
wood technology," International Journal of Nautical Archaeology 28: 34-44.
Aubet, M. E. 1993. Phoenicians and the west. Cambridge: Cambridge University Press.
Basch, L. 1976. "Le navire cousu de Bon-Porte," Cahiers d'archeologie subaquatique 5:
37-42.
Basch, L. 1981. "Carthage and Rome: Tenons and mortises," Mariner's Mirror 67: 245-50.
Basch, L. 1987a. Le musee imaginaire de la marine antique. Athens: Greek Institute for the
Preservation of Nautical Tradition.
Basch, L. 1987b. "Review: J. Coates and S. McGrail (eds.), Greek trireme of the 5th century
BC. Greenwich: National Maritime Museum, 1984; J. S. Morrison and J. Coates,
Athenian Trireme. Cambridge: Cambridge University Press, 1986," Mariner's Mirror
73: 93-105.
Bass, G. F. 1991. "Evidence of trade from Bronze Age shipwrecks," in N.H. Gale (ed.),
Bronze Age trade in the Mediterranean. Studies in Mediterranean Archaeology 90.
Jonsered: Paul Astroms, 69-82.
Bass, G. F. 1997. "Mediterranean Sea," in Delgado 1997: 268-74.
Bellabarba, S. 1993. "Ancient methods of designing hulls," Mariner's Mirror 79: 274-92.
Bellwood, P., J. Cameron, N. Van Viet, and B. Van Liem 2007. "Ancient boats, boat timbers
and locked mortise and tenon joints from Bronze-Iron Age northern Vietnam," In-
ternational Journal of Nautical Archaeology 36: 2-20.
Beltrame, C. 2000. "Sutiles naves of Roman age," in J. Litwin (ed.), Down the river to the sea.
Gdansk: Polish Maritime Museum, 91-96.
Beltrame, C. (ed.) 2003. Boats, ships and shipyards: Proceedings of ISBSA 9 (Venice 2000).
Oxford: Oxbow.
Black, E. 1996. "Where have all the sails gone?" Tropis 4: 103-12.
Bound, M. 1991. Giglio Wreck. Enalia Suppl. 1. Athens: Hellenic Institute of Marine Ar-
chaeology.
Brusic, Z., and M. Domjan 1985. "Liburnian boats: Their form and construction," in
McGrail and Kentley 1985: 67-87.
Cariolou, G. A. 1997. "Kyrenia 2: The return from Cyprus to Greece of the replica of a
Hellenic merchant ship," in S. Swiny, R. L. Hohlfelder, H. W. Swiny (eds.), Res
Maritimae: Proceedings of the 2nd International Symposium "Cities of the Sea," 1994.
Atlanta: Scholars Press, 83-97.
Casson, L. 1964. "Odysseus' boat," American Journal of Philology 85: 61-64.
Casson, L. 1989. Periplus Maris Erythraei. Princeton: Princeton University Press.
Casson, L. 1991. "Ram and naval tactics," in Casson and Steffy 1991: 76-82.
Casson, L. 1992. "Odysseus' boat," International Journal of Nautical Archaeology 21: 73-74.
Casson. L. 1996. Ships and seamanship in the ancient world. Rev. ed. Princeton: Princeton
University Press.
Casson, L., and E. Linder 1991. "Evolution and shape of the ancient ram," in Casson
and Steffy 1991: 67-71.
Casson, L., and J. R. Steffy (eds.) 1991. Athlit Ram. College Station. Texas A & M University
Press.
Clark, J. G.D. 1953. "Archaeological theories and interpretation: Old World," in A. L.
Kroeber (ed.), Anthropology today. Chicago: Chicago University Press, 342-60.
Coates, J. 1985. "Some structural models for sewn boats," in McGrail and Kentley 1985:
9-18.
Coates, J. 1996. "Appendix F," in J. S. Morrison, Greek and Roman oared warships, 399-30
B.C. Oxford: Oxbow, 347-48.
Coates, J. 2001. "Planking tenons in ancient Mediterranean ships built shell-first," Tropis 6:
153-70.
Coates, J., and S. McGrail (eds.) 1984. Greek trireme of the 5th century BC. Greenwich:
National Maritime Museum.
Coates, J., S. McGrail, D. Brown, E. Gifford, G. Grainge, B. Greenhill, P. Marsden, B.
Rankov, C. Tipping, and E. Wright 1995. "Experimental boat archaeology: Principles
and methods," International Journal of Nautical Archaeology 24: 293-301.
Coates, J., S. K. Platis, and T. Shaw (eds.) 1990. Trireme trials 1988. Oxford: Oxbow Books.
Crumlin-Pedersen, 0., and S. McGrail 2006. "Some principles for the reconstruction of
ancient boat structures," International Journal of Nautical Archaeology 35: 53-57.
Delgado, J.P. (ed.) 1997. Encyclopedia of underwater and maritime archaeology. London:
British Museum Press.
de Morgan, J. 1894-1903. Fouilles a Dahchour. 2 vols. Vienna: Adolphe Holzhausen.
de Sola Price, D. 1974. Gears from the Greeks: The Antikythera Mechanism. Transactions of
the American Philosophical Society 64.7. Philadelphia: American Philosophical So-
ciety.
de Weerd, M. 1988. Schepen voor Zwammerdam. Amsterdam: de Weerd.
de Weerd, M., and J. K. Haalebos 1973. "Schepen voor het opscheppen," Spiegal Historiael
8: 386-97.
Dilke, 0. A. W. 1985. Greek and Roman maps. London: Thames and Hudson.
Fatimi, S. 0. 1996. "History of the development of the kamal," in H.P. Ray and J.-F. Salles
(eds.), Tradition and archaeology: Early maritime contacts in the Indian Ocean. New
Delhi: Oxford Univesity Press, 283-92.
Freschi, A. 1991. "Note tecniche sul relitto Greco arcaico di Gela," Atti 4: 201-10.
Haldane, C. W. 1996. "Ancient Egyptian hull construction," Tropis 4: 235-44.

Haldane, C. W. 1997a. "Dahshur boats," in Delgado 1997: 122-23.
Haldane, C. W. 1997b. "Khufu ships," in Delgado 1997: 222-23.
Henkel, X. 1901. "Die sichtbarkeit im Mittelmeergerbiet," Petermann's Geographische
Mitteilungen.
James, T. G. H. 1983. Introduction to ancient Egypt. London: British Museum.
Jezegou, M. P. 1989. "L'epave 2 de l'anse St Gervais a Fos-sur-mer," Tropis 1: 139-46.
Johnston, P. F. 1997. Antikytliera Wreck," in Delgado 1997: 31-32.
Kahanov, Y. 2001. "Byzantine shipwreck (Tantura A) in the Tantura lagoon, Israel," Tropis
6: (1996): 265-71.
Kahanov, Y. 2003. "Dor D wreck, Tantura lagoon, Israel," in Beltrame 2003: 49-56.
Kahanov, Y., and E. Linder 2004. Ma'agan Mikhael Ship, vol. 2. Haifa: Israel Exploration
Society.
Kahanov, Y., and P. Pomey 2004. "Greek sewn shipbuilding tradition and the Ma'agan
Mikhael ship," Mariner's Mirror 90: 6-28.
Katzev, M. L. 1989a. "Kyrenia 2: Building a replica of an ancient merchantman," Tropis 1:
163-75.
Katzev, M. L. 1989b. "Voyage of Kyrenia 2," Institute of Nautical Archaeology Newsletter
16.1: 4-10.
Katzev, M. L. 1990. "Analysis of the experimental voyage of Kyrenia 2," Tropis 2: 245-55.
Katzev, M. L., and S. W. Katzev 1986. "Kyrenia 2," Institute of Nautical Archaeology
Newsletter 13.3: 2-11.
Katzev, M. L., and S. W. Katzev 1987. Kyrenia 2: An ancient ship sails again. Piraeus:
Hellenic Institute for the Preservation of Nautical Tradition.
Landstrom, B. 1970. Ships of the Pharaohs. London: Allen and Unwin.
Lewis, D. 1994. We the navigators. 2nd ed. Honolulu: University of Hawaii Press.
University Press.Linder, E., and Y. Kahanov 2003. The Ma'agan Mikhael ship: The
recovery of a 2400-year-old merchantman: Final report, vol. 1. Haifa: Israel Exploration
Society.
Lipke, P. 1984. Royal ship of Cheops. National Maritime Museum, Greenwich, Archae-
ological Series 9. British Archaeological Reports, Intl. Series S225. Oxford: BAR.
Mark, S. E. 1991. "Odyssey (5.234-53) and Homeric ship construction," American Journal of
Mark, S. E. 1996. "Odyssey (5.234-53) and Homeric ship construction: A clarification,"
Mark, S. E. 2005. Homeric seafaring. College Station: Texas A & M University Press.
Marsden, P. 1994. Ships of the Port of London, vol. 1. London: English Heritage.
McGrail, S. 1986. "Experimental boat archaeology-Some methodological consider-
ations," in 0. Crumlin-Pedersen (ed), Sailing into the past. Roskilde: Viking Ship
Museum, 8-17.
McGrail, S. 1990. "Boats and boatmanship in the southern North Sea and Channel Re-
gion," in S. McGrail (ed.), Celts, Frisians and Saxons. CBA Research Report 71. York:
Council for British Archaeology, 32-48.
McGrail, S. 1991. "Bronze Age seafaring in the Mediterranean: A view from N.W Europe,"
in N. H. Gale (ed.), Bronze Age trade in the Mediterranean. Jonsered: Paul Astroms,
83-91.
McGrail, S. 1992. "Replicas, reconstructions and floating hypotheses," International Journal

of Nautical Archaeology 21: 353-55.
McGrail, S. 1993. "Experimental archaeology and the trireme," in Shaw 1993: 4-10.
McGrail, S. 1996a. "Navigational techniques in Homer's Odyssey," Tropis 4: 311-20.
McGrail, S.1996b. "Studyofboats with stitched planking," in H. Ray and J.-F. Salles (eds.),
Tradition and archaeology: Early maritime contacts in the Indian Ocean. New Delhi:
Manohar.
McGrail, S. 1998. Ancient boats in north-west Europe. 2nd ed. London: Longman.
McGrail, S. 2003. "How were boats designed before the late-Medieval period?" in Beltrame
2003: 124-31.
McGrail, S. 2004a. Boats of the world. 2nd ed. Oxford: Oxford University Press.
McGrail, S. 2004b. "North-west European seagoing boats before AD 400," in P. Clark
(ed.), Dover Bronze Age Boat in context. Oxford: Oxbow Books, 52-66.
McGrail, S.2006. "Experimental boat archaeology: Has it a future?" in L. Blue, F. Hocker,
A. Englert, (eds.), Connected by the sea: Proceedings ofISBSA 10, 2003. Oxford: Oxbow:
8-15
McGrail, S., and E. Kentley (eds.) 1985. Sewn plank boats. British Archaeological Reports,
Intl. Series S276. Oxford: BAR.
Morton, J. 2001. Role of the physical environment in ancient Greek seafaring. Leiden: Brill.
Morrison J. S., and J. Coates 1986. Athenian trireme. Cambridge: Cambridge University
Press.
Morrison, J. S., and J. Coates (eds.) 1989. An Athenian trireme reconstructed. British Ar-
chaeological Reports, Intl. Series S486. Oxford: BAR.
Morrison, J. S., J. Coates, and N. B. Rankov 2000. Athenian trireme, 2nd ed. Cambridge:
Murphy, J.P. (ed.) 1977. Rufus Festus Avienus: Ora Maritima. Chicago: Ares.
Murray, W. M. 1987. "Do modern winds equal ancient winds?" Mediterranean Historical
Review 22: 139-67.
Murray, W. M. 1991. "Provenance and date," in Casson and Steffy 1991: 51-66.
Nayling, N., and S. McGrail 2004. Barland's Farm Romano-Celtic boat. CBA Research
Report 138. York: Council for British Archaeology.
Negueruela, I., M. Pinedo, M. Gomez, A. Minano, I. Arellano, and J. S. Barba 1995.
"Seventh-century BC Phoenician vessel discovered at Playa de la Isla, Mazarron,
Spain," International Journal of Nautical Archaeology 24: 189-97.
6 hEailidhe, P. 1992. "Monk's boat: A Roman period relic from L. Lene, Co. Westmeath,"
International Journal of Nautical Archaeology 21: 185--90.
Oikonomides, A. N., and M. C. J. Miller (eds.) 1995. Hanno the Carthaginian, Periplus.
Chicago: Ares.
Oleson, J. P. 2000. "Ancient sounding weights: A contribution to the history of Medi-
terranean navigation," Journal of Roman Archaeology 13: 294-310.
Pomey, P. 1981. "L'epave de Bon Porte et les bateaux cousus de Mediterranee," Mariner's
Mirror 67: 225-43.
Pomey, P. 1985. "Mediterranean sewn boats in antiquity," in McGrail and Kentley 1985: 35-48.
Pomey, P. 1995. "Les Epaves Grecques et Romaines de la Place Jules-Verne a Marseille,"
Comptes Rendus de l'Academie des Inscriptions et Belles-Lettres 1995: 459-84.
Pomey, P. 1996. "Un example d'evolution des techniques de construction navale antique:
a
De !'assemblage par ligature !'assemblage par tenons et mortaises," in D.Meeks
and D. Garcia (eds.), Techniques et economie antiques et medievales. Paris: Editions
Errance, 195-203.
Pomey, P. 1997. "Bon Porte wreck," in Delgado 1997: 252-53.

Pomey, P. 1999. "Les epaves Grecques du 6e siecle av. J-C de la Place Jules-Verne a
Marseilles," in P. Pomey and E. Rieth (eds.), "Construction navale maritime et flu-
viale," Archaeonautica 14: 147-53.
Prins, A.H. J. 1965. Sailing from Lamu. Assen: Van Gorcum.
Pulak, C. 1998. "Uluburun shipwreck: An overview," International Journal of Nautical
Pulak, C. 2003. "Mortise-and-tenon joints of Bronze Age seagoing ships," in Beltrame
2003: 28-34.
Reade, J. (ed.) 1996. Indian Ocean in antiquity. London: British Museum and Kegan Paul.
Rieth, E. 2000. "Medieval wreck from Port Berteau 2," in J. Liwin (ed.), Down the river to
the sea. Gdansk: Polish Maritime Museum, 225-28.
Rieth, E., C. Carrierre-Dubois, and V. Serna (eds.) 2001. L'epave de Port Berteau 2. Paris:
Maison des Sciences de }'Homme.
Roberts, 0. T. P. 2005. "Windward sailing capability of ancient Greek ships," Mariner's
Mirror 91: 106-7.
Royal, J. 2002. Development and utilization of ship technology in the Roman world in late
antiquity. PhD diss., Texas A & M University.
Royal, J. G., and Y. Kahanov 2005. "New dating and contextual evidence for the frag-
mentary timber remains located in the Dor D site, Israel," International Journal of
Nautical Archaeology 34: 308-13.
Rule, M., and J. Monaghan 1993. Gallo-Roman trading vessel from Guernsey. Guernsey
Museum Monograph 5. St. Peter Port: Guernsey Museum Services.
Shaw, J. T. (ed.). 1993. Trireme project. Oxford: Oxbow.
Sleeswyk, A. W. 1980. "Phoenician joints, coagmenta punicana," International Journal of
Steffy, J. R. 1985. "Kyrenia ship: An interim report on its hull construction," American
Steffy, J. R. 1989. "Role of three-dimensional research in the Kyrenia ship reconstruction,"
Tropis 1: 249-62.
Steffy, J. R. 1991. "Ram and bow timbers: A structural interpretation," in Casson and Steffy
1991: 6-39.
Steffy, J. R. 1994. Wooden shipbuilding and the interpretation of shipwrecks. College Station:
Texas A & M University Press.
Steffy, J. R. 1995. "Shipwreck archaeology: An essential medium for interpreting ancient
ship construction," in V. Karageorghis and D. Michaelides (eds.), Cyprus and the Sea.
Nicosia: University of Cyprus, 23-31.
Taylor, E. G. R. 1971. Haven-finding art. London: Hollis and Carter.
Throckmorton, P. J. 1973. "Roman wreck at Pantano Longarini," International Journal of
Tilley, A. F. 1976. "Rowing the trireme," Mariner's Mirror 62: 357-59.
Tilley, A. F. 1992. "Three men to a room: A completely different trireme," Antiquity 66.252:
599-610.
Tilley, A. F. 1995. "Warships of the ancient Mediterranean," Tropis 3: 429-40.
Tilley, A. F. 2004. Seafaring on the andent Mediterranean. British Archaeological Reports,
Intl. Series S1268. Oxford: BAR.
van Doorninck, F. H. 1976. "Fourth century wreck at Yassi Ada," International Journal of
van Doorninck, F. H. 1982. "Hull remains," in G. Bass and F. H. van Doorninck (eds.),
Yassi Ada, vol. 1. College Station: Texas A & M University Press, 32-64.
Wachsmann, S. 1998. Seagoing ships and seamanship in the Bronze Age Levant. College
Station: Texas A & M University Press.
Ward, C. A. 2000. Sacred and secular: Andent Egyptian ships and boats. Archaeological
Institute of America Monograph Series 5. Philadelphia: University of Pennsylvania
Museum.
Ward, C. A. 2003. "Sewn planked boats from Early Dynastic Abydos, Egypt," in Beltrame
2003: 19-27.
Westerdaltl, C. 1992. "Review: F. Welsh, Building the Trireme. London: Constable, 1988,"
Wright, E. 1990. Ferriby boats. London: Routledge.
CHAPTER 25
SEA TRANSPORT,
PART 2: HARBORS
DAVID J. BLACKMAN
THE HISTORY OF RESEARCH

ON HARBORS
Interest in ancient harbors was aroused in the nineteenth century by the discovery of
wall paintings and mosaics with maritime scenes, mainly in Herculaneum and
Pompeii.* Although ship depictions attracted the most attention, several harbor
scenes were also identified (Blackman 1982). Such scenes were also found on all
manner of other media, including glass panels and engraved glass flasks of the Roman
period, the latter depicting the sites of Baiae and Puteoli. Reliefs, mainly from sar-
cophagi, provided a larger field for depicting a complex scene, such as a ship entering
harbor and another ship made fast to a mooring-stone to enable unloading to take
place. The harbor on the famous Torlonia Relief of about A.D. 200 (figure 25.1) is
usually identified with the Claudian harbor at Portus ("the Port") near Ostia, based
on comparison of the lighthouse with representations of the Portus lighthouse on
local mosaics. Many coins, almost all Roman, depicted harbors, some clearly iden-
tifiable from the inscription or from structural details (Blackman 1982: 80--85).
There was, however, little systematic work on ancient harbors in general. There was
*Given the unavoidably large number of sites mentioned in this chapter, I have decided in
most cases to omit citation in the text of bibliographical references concerning sites mentioned
more than once. The documentation for these sites can be found in a gazetteer placed before the
references for this chapter. Further discussion and bibliographical information for most of these
sites can also be found in Blackman (1982).
SEA TRANSPORT: HARBORS 639
Figure 25-1. Torlonia Relief, stylized representation of Portus, ca. 200 A.D. Rome.
(Courtesy of Bridgeman Art Library.)
discussion about the topography of Rome, which had an important river port, of
Ostia/Portus at the mouth of the River Tiber; of Syracuse; and of Piraeus, the port of
inland Athens. Otherwise few studies of harbor sites were published.
Much valuable material-including plans and descriptions-resulted from the
work of nineteenth-century hydrographers preparing the first Admiralty Charts of
the Mediterranean. For them, the features of ports such as location, water supply,
identification marks, and hazards were of prime importance. But their work did not
receive from archaeologists the attention that it deserved. Early in the twentieth
century, two developments led to a new interest in the remains of ancient harbors and
submerged coastal sites. First, major programs of drainage work were being carried
out, along with the excavation of canals and the dredging and modernization of
harbors. Second, the phenomenon of sea-level change was being studied, particularly
in Greece and Italy. The evidence for earth movements in the Bay of Naples aroused
the interest of geologists such as Gunther (1903a-b) and archaeologists such as
Dubois (1902, 1907), who wrote a book on ancient Puteoli and studied the harbor
construction methods described by Vitruvius (De arch. 5.12) . The first major work on
ancient harbors was written by Lehmann-Hartleben (1923) in the years around World
War I. Its subtitle indicated his particular interest: Contributions to the History of City
Planning in Antiquity. Given the period, much of this monumental work is inevitably
a compilation based on ancient texts and modern publications rather than on his
own personal observation, but it was very thorough and detailed and remains an
essential source for study of the subject. One defect is that Lehmann-Hartleben too
readily assumed that methods of harbor construction developed and progressed
evenly, and that innovations were adopted universally. We can now see that this was
not so, and that small coastal settlements may not have had the resources or the need
to construct sophisticated, major harborworks at any time in antiquity. More recent

studies have emphasized the importance of small ports, and of cabotage (short-range
coastal shipping) (McCann and Oleson 2004: 207-10).
Lehmann-Hartleben's book did not provide any immediate major stimulus to
harbor studies, certainly not in the field. The next major figure was Antoine Poi-
debard (1939, 1951), a Jesuit priest who worked at Tyre from 1934 to 1936 and Sidon
from 1946 to 1950, although his pioneering work was not fully appreciated until the
1960s. Poidebard was the first to recognize the value of air photographs for studying
sites submerged in shallow water. At Tyre, he carried out underwater survey and
photography to check the air photographs and published with remarkable speed a
full survey of the outer roadstead and the inner harbor. At Sidon, using similar
methods, he traced a similar pattern in the remains, and found and explained a
remarkable de-silting system. Military experience during World War II confirmed
the value of air photographs, and experts had much new material to exploit. Thus
Bradford brilliantly deduced the existence of a west harbor at the city of Rhodes,
now silted up and previously unknown to archaeologists. Schmiedt exploited the
Italian air photo archive for the study of coastal sites, revealing many submerged
remains and other sites now buried inland. He also researched evidence for sea-level
change (Schmiedt 1972). Air photographs continue to be a valuable resource, since
so many sites have been threatened with destruction in the industrial and touristic
development of the Mediterranean coastline.
For some ancient cities lying under modern towns, like Rhodes, Corcyra
(Corfu), or Syracuse, it has been possible to piece together ancient topography
gradually thanks to the work of local archaeological services, who scrutinize
modern building developments and investigate ancient remains in deep levels
before the new buildings are started. In an extreme example, in Naples the ancient
harbor was discovered only during Metro construction in 2003. Nevertheless, many
ancient sites have been damaged or built over without proper study or recording.
Fortunately this was not the case with the Claudian harbor at Portus, much of
which now lies under the Fiumicino airport; nor with the shore of the Vieux Port of
Massalia (Marseilles), where important remains of ancient quays and slipways, and
of beached ships, have been carefully excavated since the 1990s, in preparation for
the construction of underground car parks. At Pisa the river harbor was discovered
during work for a new high-speed train line. Other harbors were surveyed in
advance of the arrival of a major coast road, for example Phaselis in Lyda. Other
major sites have been excavated as land sites, even when not under threat: the
Roman harbor at Leptis Magna in Tripolitania, and the harbors of ancient Carthage.
Special techniques have been developed for the difficult task of survey and exca-
vation in shallow water, for example in the Bay of Naples. With the increasing threat
of coastal development, the survey of ancient harbor sites is a major priority, and
much less expensive than excavation. A number of sites have been investigated since
the 1970s with visual survey and subsequently with the latest techniques of geo-
physical prospection. The potential of the latter demonstrated between 1997 and
2004 by dramatic results in the area between the harbor on the River Tiber and the
Trajanic harbor at Portus (Keay and Millett 2006).
Increased public interest in all aspects of marine archaeology, including har-

bors and discoveries on harbor sites, has led to a renewed interest in, for example,
Vitruvius' description of the laying of concrete foundations under water, now the
subject of a major research project (Oleson et al. 2004, 2006). Much new evidence
has been found for sea level changes in historical times, notably from slipways and
fish-tanks, to contribute to geomorphologists' studies of the extent of such changes
and of the processes that affect them (movements of the land and changes in the
level of the sea itself). Sometimes the archaeologist may be able to indicate a fairly
precise date for a particular sea level, and in return the geomorphologist may be
able to help with dating otherwise undatable structures. Dating of dead marine
fauna also has great potential. This is a continuing debate and is largely dependent
on close cooperation among varied scientific disciplines (Blackman 2005). Wide
interdisciplinary cooperation, for example, with sedimentologists, has brought
valuable results in the study of the paleoenvironment of ancient harbors, for ex-
ample, at Marseilles, Citium, Tyre, and Carthage (Marriner and Morhange 2007).
Large-scale excavation under water has been undertaken at only a few sites,
notably Cenchreai, Casa, Amathus, and Caesarea in Palestine, and one could add
more limited work at Apollonia, the port of Cyrene, at Thasos, Halieis, Samas,
Aegina, and Paphos (see the gazetteer at the end of this chapter, Shaw 1972,
Blackman 1982, and entries in Delgado 1998). Many of the main results are referred
to below. Major excavations in progress as this chapter was being prepared, which
cannot be fully incorporated here, are those in the Great Harbor at Alexandria.
Herakleion, 30 km to the east in Aboukir Bay by the Canopic mouth of the Nile, is a
remarkable example of a port with four basins, linked to a sanctuary, and may prove
to have been the predecessor of Alexandria (see Gazetteer). It is now submerged to
uneven depths by catastrophic faulting, soil liquefaction, and subsidence.
In the first years of the twenty-first century there have been several regional
surveys of coastlines, notably the ANSER (Porti antichi del mar Tirreno) project,
which has enabled the study of the relationship between neighboring ports and
their hinterland and offshore islands, from Latium to southern Spain (Gallina Zevi
and Turchetti 2004). We now know much more about, for example, river ports on
the Guadalquivir (ancient Baetis) and the first harbor installations of the pre-
Roman Iberian period (fifth-fourth century B.C.) from Alicante.
HISTORICAL OVERVIEW OF
HARBOR DEVELOPMENT
Documentation for most of the developments mentioned in this chapter can be

found in Blackman 1982. In the Dark Age that followed the collapse of the Bronze
Age kingdoms of the Aegean and Near East at the end of the second millennium
B.C., the main long-distance maritime activity seems to have been carried out by
the Phoenicians. They resumed the use of rock-cut harbors along the Levantine
coast, such as Tyre and Sidon, which had probably already been developed in
the Late Bronze Age; perhaps they were the original builders. It was certainly the
Phoenicians who added the built structures such as quays and jetties at Sidon, the
original harborworks at Athlit and Akko, and a jetty at Tabbat el-Hammam dating
from the ninth century B.C. (Blackman 1982: 90-94; Frost 1971, 1973, 1995; Raban
1995).
Within a century, Greek trade with the Near East revived, and Greek traders
participated in the development of harbor facilities in that region. A number of
Greek cities followed the Phoenicians in establishing trading posts for maritime
trade in the West; this trend was further stimulated from the late eighth century by
the settling of colonies in many parts of the Mediterranean and Black Sea. Colo-
nization created the need for improved harbor facilities, which, like other major
public works, were made easier to finance by the introduction of coinage in the
seventh century.
Many of the known early public works in Greek cities were the work of powerful
tyrants who built up their city's trade and prosperity and had the necessary resources
of finance and labor. The earliest harborworks in the Aegean that are firmly dated are
those built by Polycrates, tyrant of Samos, in about 530 B.C. and described by
Herodotus (3.60.3) as a major feat of engineering. Some remains of his massive
breakwater may survive; by its size it clearly marked a new stage in harbor develop-
ment, but it was not the first built harborwork in the Aegean. A quay and breakwater
on Delos have been dated to the late eighth century; this dating is not certain, but the
late eighth and seventh centuries provide a very likely context for the first large-scale
attempts to improve the available natural harbors in the Aegean. A seventh-century
B.C. date was previously suggested for a rubble breakwater at Eretria, over 600 m
long, running out into depths of over 20 m, diverting a silt-bearing stream and
ending at a natural reef; but a sixth-century date now seems more likely (Walker
2004: 100). Rubble breakwaters were also built in much later periods.
The indented coastline of the Aegean provided many good natural harbors
(Morton 2001), so many cities needed to build no more than a shoreline quay or jetty
against which even large merchant ships could berth. But some cities lay on exposed
coasts and required a different solution. A feature of certain harbors on exposed
coastlines in Greece and Sicily may show Phoenician influence: an artificial harbor
basin (cothon) excavated landward of the coastline. This technique was particularly
important along Phoenician trade routes on the North African coast. The basin
sometimes made use of existing low-lying ground or a lagoon, perhaps joined to the
sea by one or more channels. The prime example of such a harbor is that at Carthage,
a Punic foundation. Although the date of the visible features is much later, there is
no reason to doubt that the harbor originated with an excavated basin. Rock-cut
basins are difficult to date, but there is similarly no reason to doubt that the harbor
of Mahdia also had a Punic origin (Blackman 1982: 93-94; Tusa 2004).
The island settlement at Motya in western Sicily had such a basin, although
apparently not until the sixth century. Here we know from historical sources that
we are dealing with a Phoenician construction, although its identification as a

harbor has been disputed. Phoenician harborworks have also been suggested at
Tharros in Sardinia, preceding the Roman structures. The only harbor on the Greek
mainland that may reasonably be called a cothon is the inner harbor at Lechaeum,
the west harbor of Corinth. The dating is uncertain, but it may plausibly be ascribed
to Periander, the great tyrant of Corinth (reigned ca. 627-587) who was famous for
his contribution to naval developments and who instituted transit tolls on traffic
through the Corinthian ports.
ANCIENT TECHNICAL LITERATURE
Unfortunately, no ancient technical handbooks on harbor construction survive,

although some author's names and even titles survive; for example, Limenopoii'ka
(Harbor Construction) written by the engineer Philon in the late third century B.C.
By this date there were flourishing schools of engineering, notably at Alexandria
and Rhodes. Vitruvius, writing in the first century B.C., was able to draw on a body
of technical literature, and possibly on his own experience, for his one chapter on
harbor construction (De arch. 5.12). Oleson (Oleson et al. 2004: 205-6) plausibly
suggests that there may by then have been a number of subliterary harbor engi-
neering manuals, and perhaps even specialist collegia for those building ancient
Roman harbors. Although we have the names of many specialist laborers who
worked in harbors, we do not have any references to collegia, nor do we have any
reference to a harbor engineer during the classical period. Names of authors of
works titled Harbors are known to us, and some would have been experts: for
example, Timosthenes of Rhodes, a captain in the Ptolemaic fleet in the third
century B.C. (Strabo 9.3.10 ). But from what little we know of these literary works,
they seem to have been geographical rather than engineering handbooks, like the
Coastal Pilots or Periploi, of which a number survive (Blackman 1982: 79-80; Casson
1989; Morton 2001: 180-81).
TECHNOLOGICAL EVOLUTION
From the late sixth century B.C. onward, the main developments in harbor en-
gineering may be attributed to the Greeks and then to the Romans. A continuing
contribution by the Phoenicians is hard to estimate, but it may have been more
significant than we can now detect. Some have suggested a continuing Phoenician
tradition in naval architecture, which may have extended to harbor engineering.

Others, including the author, have wondered whether Greek architects built the
shipsheds at Carthage and earlier at Citium. So far we know little about Etruscan
harbor installations (McCann 1977; Oleson 1977; Bruni 2003).
The Greeks built harborworks much like normal land structures, using blocks
of ashlar masonry joined without mortar, but often with clamps, to form the quays;
breakwaters were built of rubble, but sometimes the upper surface and inner side
were faced with ashlar. In the Roman period, architectural developments on land,
following the invention of structural concrete, led to developments in harbor
construction-notably the invention of the hydraulic form of concrete with a
volcanic ash additive (pozzolana from the Bay of Naples) that would set under
water. Instead of merely using stone blocks, engineers were able to build free-
standing structures in the sea: the foundations for moles and breakwaters, quays,
lighthouses, and other shoreline structures such as piscinae or fish-tanks (Oleson et
al. 2004, 2006). The invention has not yet been precisely dated: it probably resulted
from an accidental discovery or deliberate experiments in the Bay of Naples, per-
haps at Puteoli itself, in the late third or early second century B.C. There are also
possible indications of experiments with lime mortar and wooden forms in the
harbor of Hellenistic Alexandria in the mid-third century B.C. Vitruvius describes
the material and the methods of its use (De arch. 2.6.1; 5.12).
At present, our earliest evidence of a major structure built with hydraulic
concrete are five concrete pilae (large piers or blocks) built on or by an earlier
breakwater in the harbor of Cosa, a Roman colony on the coast of Etruria; the date
is disputed between the late second and the mid-first century B.C., but analysis of
new samples indicates the later date (Oleson et al. 2004). This technology was
famously used in the construction of Herod the Great's harbor at Caesarea in
Palestine (ca. 22 B.C.); analysis of the concrete used there has shown that the crucial
ingredient, pozzolana, was indeed imported from the Bay of Naples (Branton and
Oleson 1992). The same is true of a Roman mole in the smaller harbor of Cher-
sonesus on the north coast of Crete (Brandon et al. 2005).
One could imagine that at first a few teams of skilled engineers traveled the
Mediterranean to apply their (no doubt expensive) expertise in various ports for
rich benefactors; or they may have been sent on an imperial mission at Caesarea and
possibly Paphos, where the harbor needed major repair after a serious earthquake in
15 B.C. (Hohlfelder 1996: 95). But techniques cannot have been kept secret for long;
they must soon have become widely known and copied, and perhaps described in
engineering manuals. In time it was also discovered that some local volcanic sands
could be used instead of pozzolana from Naples. This is the subject of a major
research project, the Roman Maritime Concrete Study (ROMACONS: Oleson et al.
2004, 2006), which is taking samples of hydraulic concrete from a wide range of
sites across the Mediterranean in order to define its structural characteristics and to
elucidate its production technology and the history of its use, along with wider
questions of its transportation, the financing of projects and the composition of the
specialist labor force. Relative cost advantages may be part of the explanation of the
success of this technology and the resulting rapid advances in Roman harbor
construction. The experimental reconstruction of a pila with Vitruvian-type con-
crete has cast light on methods of placing this concrete. Analysis of the samples
shows that Roman hydraulic concrete used in maritime structures is surprisingly
soft and porous, but nevertheless durable (Oleson et al. 2004, 2006)
Crucial for the use of concrete for building structures in the water was the design
and construction of timber formwork to hold the concrete in place and protect it
from the waves until it had set. The forms could then be left in place, or removed.
Vitruvius describes several formwork designs (De arch. 5.12), but evidence has been
found for a number of methods, in timber or ashlar enclosures. The timber forms
often show a high quality of carpentry, as if they had been built by shipwrights;
sometimes single-mission barges built with the same care as ships were used as forms
and sunk by the addition of a load of hydraulic concrete at the construction site
(figure 25.2; Brandon 1999). In the west mole at Portus a huge merchant ship was
sunk containing the concrete foundation of the lighthouse (Suetonius, Claud. 20.3;
Pliny, HN 16.201-2, 36.69-70), and some smaller ships were also found in the con-
crete parts of the mole. Study of remains at many ports besides Caesarea show that
the construction procedures varied, but the quality of the pozzolana concrete was
relatively uniform. At some sites the formwork has not survived, but can be deduced
from the impressions left on the surface of the concrete structure (Oleson et al.
2006), for example, at Side (Knoblauch 1977: figs. 75-77). Sometimes the traces are
horizontal holes across the structure, which held timber tie-beams, such as the two
rows across the concrete mole at Thapsus in North Africa. Study of the holes in the
Roman moles at Anti um has shown the complex structure of the timber armature
(Felici 1993, 2002).
The use of hydraulic concrete combined with their generally sophisticated
engineering skills enabled Roman engineers to build harbors wherever it was de-
sirable for political, military, or economic reasons, and on coasts where, unlike the
Aegean, there were few or no natural harbors: Italy, North Africa, or the Levant, for
example. The large, deep open roadsteads needed for modem tankers would have
been shunned by ancient shipping, which needed cozy, well-protected refuges, with
no great depth, and, in the Mediterranean, no need to allow for tidal ranges. But
engineers now had a range of choices, and harbors did not always make use of
hydraulic concrete (Hohlfelder 1985, 1996; Oleson 1988). The harbor of Cenchreae
was built with methods showing little similarity to the Vitruvian procedures, and a
variety of construction is seen even within the Claudian harbor at Portus.
Grandiose projects like the harbor of Leptis Magna involved the extensive use of
massive ashlar masonry for quay faces and upper surfaces.
One can trace a gradual increase in the size of harbors in the Mediterranean
from the classical Greek through the Hellenistic to the Roman periods, not only the
area available for berthing or anchorage, but also the scale of the facilities, quays,
and dockyard buildings. There was also a greater regularity in plan. But there were
always exceptions: limited, simple harbor facilities built with traditional methods
will have sufficed at many smaller sites (for example, Aperlae in Lyda), even in the
Figure 25.2. Single-use barge caissons being sunk over south breakwater at Caesarea in
Palestine. (Reconstruction by C. Brandon, by permission.)
prosperous period of the early Roman Empire, and beaches remained a frequent
alternative to harbor structures. Recent scholarship has emphasized the important
role of cabotage throughout antiquity, and this kind of trade must have frequently
made use of undeveloped harbors.
The developments in harbor technology are less clear in the Aegean area, where
the Romans often took over existing ports, with the resulting problem that it is
often difficult to distinguish pre-Roman facilities. This is true at Cenchreae or, to
take a less well known example, the small site of Mavra Litharia, the port of Aegira
west of Corinth, where the harbor facilities (built of hydraulic concrete below sea
level and concrete with ashlar revetment above) seem to have been put out of
action by a massive seismic uplift of about 4 m in the third century A.D. (Papa-
georgiou et al. 1993).
Evidence for late Roman harbors is sparse, except for Constantinople (Kingsley
2004: 131-59). Harbor construction by the sixth-century emperors Anastasius and
Justinian is described by a contemporary writer, Procopius (Aed. 1.8.7-9, 11.18-20;
4.10.5-17; Anec. 8.7-8, 26.23), but his reference to the renovation of the harbor of
Caesarea (Panagyr. Anast. 19) has found little confirmation in evidence from a
thorough survey. A modest harbor such as Aperlae seems to have continued to
operate, because it produced the lucrative murex or purple dye (Hohlfelder 2005).
The only known works that probably date from this period are the last phase of the
harbor of Anthedon in central Greece: two apparently hastily built moles and a
quay, with rubble fill and hydraulic mortar in compartments divided and faced by
rough courses of mainly reused blocks (Schlager et al. 1968). Because the courses
were irregular, shallow slots were cut in their upper surface and filled with mortar to
level off the top surface of the quay, methods very different from the sophisticated
harbor engineering of the early Empire. Although the Anthedon harbor structures
were probably a rapid construction by land engineers, intended to support the
urgent land defense of central Greece against invasions from the north in the early
sixth century, they have survived in good shape. A slipway, possibly a dry-dock, at
Marea on the shore of Lake Mareotis in Egypt is the only known example from late
Antiquity (Petruso and Gabel 1983).
MOLES AND QUAYS
Moles and breakwaters were built to provide protection against both heavy seas and
enemy attack. Their alignment, especially in relation to the prevailing winds and
currents, and type of construction were important in fulfilling their main purpose,
and also for dealing with the problem of siltation. The simplest breakwater was a
natural line of reef or rocks, and some early harbors in the Levant, such as Sidon,
made effective use of these, including wave traps to break the force of the waves. In
the Greek world, construction of long, submerged embankments of rubble may go
back at least to the seventh century (see above). At Cnidus in the fourth century,
breakwaters of huge, roughly-cut blocks were built in 30 m of water.
The outer faces of such breakwaters naturally were sloped, both to stabilize the
loose construction materials and to prevent undermining by wave action. The
permeability of such structures, along with their alignment-where possible at an
angle to the prevailing seas-helped to break the force of the waves. The height to
which breakwaters were built above sea level cannot now always be determined.
Some may have lost stones from their upper surface through storms or stone
robbing in later periods, but many still look very solid, possibly having grown
stronger with time through compaction of the rubble materials. It is now clear that
some breakwaters have subsided under their own weight, through lack of a solid
natural foundation or the destabilizing of the underlying levels by effects such as
liquefaction-the mobilization of sediments by the sudden infusion of water, often
as a result of earthquake. Evidence of subsidence can be seen in the very wide
breakwater at Cosa, the north mole at Cenchreae, and probably the north break-
water of the commercial harbor at Cnidus and the south breakwater at Phaselis. The
outer breakwater at Caesarea in Palestine seems to have sunk 5-6 m as a result of
either the shift of a geological fault or the liquefaction of the sand beneath the
harbor structures.
Some breakwaters did not support buildings and were intended to allow the
waves simply to break over them, creating a current within the harbor that would
help prevent silting. This seems to have been the original purpose of the Cosa
breakwater. Such a strategy would work only in virtually tideless seas such as the
Mediterranean, although tidal seas offered alternative methods of de-silting. Unless
a breakwater had this special purpose, it could also be used to add to the berthing
space available in the harbor. Thus the inner sides of moles were built up to serve as
quays, and fortification walls were built along their outer sides, to include the
harbors in the city fortifications (see below). There must often have been only a
narrow quay space inside such walls, or none at all; but merchant ships could at
least moor there in safety while waiting to unload at the shoreline quays or offload
into lighters without ever berthing at a quay. Excavations on the "Mole of the
Windmills" at Rhodes have shown that after Demetrius Poliorcetes landed on the
unfortified breakwater during his unsuccessful siege of 305 B.C., the Rhodians built
an impressive wall on it. By the second century B.C. its defensive role was less
important, and a quay for commercial purposes was added along its inner side). In
the period of the pax Romana, fewer moles around the Mediterranean bore forti-
fication walls, leaving more space for loading areas and for buildings. Some Roman
harbors such as Cenchreae and Leptis Magna had moles wide enough to carry
quays, warehouses, temples, statues, fish-tanks, signal-towers, and a lighthouse.
Horrea (grain warehouses) were a distinctive feature of Roman commercial harbors
such as Portus. The later addition of a fortification wall at Leptis was a sign of
renewed insecurity.
The new flexibility that hydraulic concrete gave to Roman engineers enabled
them not only to build solid structures of concrete, or concrete faced with ashlar
and brick, but also to experiment with arched structures, in particular detached
piers joined by arches supporting a paved surface. This design seems to have been
tried out experimentally for a short period during the first centuries B.C. and A.D.
in Latium and Campania. The arched mole at Puteoli was 372 m long and 16 m
wide, based on 15 piers of tufa blocks on concrete foundations. The arches were
built of brick and limestone blocks with a fill of rubble and pozzolana. The line of
piers is slightly curved, better to resist stormy seas from the south (Piromallo 2004:
273). The south mole of the naval harbor at nearby Misenum had two offset lines of
piers. Some contemporary works of art, mainly paintings from the same region (e.g.
Blackman 1982: fig. 5), illustrate similar structures, but we cannot be certain that the
superstructures were not built of timber. The purpose of the design may have been
to prevent silting up of the harbor basin by allowing currents of water to flow
through it. Perhaps the currents proved too strong and the waves unmanageable
during heavy seas and strong winds; but probably the main reason for the aban-
donment of the experiment was that the structures lacked longitudinal stability and
collapsed. The mole at Puteoli, for example, had to be rebuilt in A.D. 139 ( CIL
10.1640, 1641).
Another innovation, apparently of the Roman period, is the building of off-
shore breakwaters. Pliny the Younger (Ep. 6.31) describes being summoned by
Trajan to witness the building of a new harbor at Centumcellae (Civitavecchia), on
the harborless coast north of Rome: an artificial island was being built in front of the
harbor entrance, with enormous blocks brought out by boat and dropped, and
piers added to the pile to make an arcaded facing or superstructure. In the second
century A.D., a number of ports, such as Sabratha and Patras, were provided with
long offshore breakwaters as their only protection; the aim seems to have been to
prevent silting by allowing the coastal current to pass through unobstructed
(Blackman 1982: 198).
Though the inner side of a mole was often used as a quay, the main quays were
usually on the shore, backed by warehouses and all the facilities of a port. Quay
construction methods, we may assume, developed roughly in parallel with those of
moles: from rock-cut quays as at Sidon, to roughly faced and paved rubble, as in
early Delos, to well-dressed ashlar revetments, and then, in the Roman period, to
concrete structures, usually faced with ashlar or small stone blocks (opus re-
ticulatum) or brick, or sometimes with timber. Rock-cut quays (such as those at
Delos, Cnidus, and Mahdia) are very difficult to date except by general context.
A few solid Roman quays were faced with brick arches, and in some works of art
quays seem to be depicted as supported on freestanding arches. The only remains
that may belong to such a quay are those of the south quay at Puteoli, where two
offset lines of piers may have been joined by arches; but the wall may in fact simply
be the outer wall of a set of basins. Some quays and projecting jetties depicted in art
do appear to be timber structures (wall paintings from Stabiae and the Temple of
Isis at Pompeii, and the Palestrina mosaic from Praeneste: Blackman 1982: 83-85).
Such structures have long been known from north of the Alps, from Roman Britain
(e.g., the port of Roman London), Gaul, and the Rhineland (from river ports and
embankments); now some evidence has been found from the Mediterranean
coast-mainly from southern Gaul and northern Italy, but also in Alexandria from
about 400 B.C. Survival of such remains may be the result of favorable conditions
for the preservation of organic material, but it could also indicate local building
traditions. Timber structures are known from the small port of Lattes in southern
Gaul (Garcia et al. 2002: 67-70) and from the important regional port at Marseilles,
the ancient Greek colony of Massalia. Much timber was used in the slipways of pre-
Roman times (see below) and again in the quays of the first century A.D. One quay
was built of pozzolana concrete in an open form work of timber; the other involved
a more elaborate timber structure: a rubble fill and a clay and cobbled surface was
lined by a double row of pine piles, with horizontal planking between the piles
consisting of reused timbers from at least nine boats (Hesnard 2004). Hesnard
points out that Vitruvius may have been present at the siege of Massalia and
suggests that the techniques used to build these quays illustrate Vitruvius' guide-
lines. However that may be, timber facing would have reduced the risk of damage to
a quay from a ship coming alongside. Another timber structure at Massalia, a
landing-stage at right angles to the shore, also contained reused ship timbers; it ran
out 30 m into deeper water. The dramatic discovery of fifth-century A.D. ships in
the Roman port of Olbia in Sardinia provides a vivid impression of how the ships
were berthed-bow to shore in the typical "Mediterranean moor," probably be-
tween lines of timber piles (Kingsley 2004: 89-95).
Excavations in an ancient river port on the northwest edge of Pisa have revealed
both a large number of ships wrecked by floods between the second century B.C.
and late antiquity, and a number of timber structures along the river bank that
gradually moved northward as a result of alluvial deposits. This site illustrates
dramatically the problems faced by river ports. The earliest structure is a timber
bank revetment, apparently of the late sixth century B.C., possibly our earliest
evidence for Etruscan harborworks. A second timber revetment farther north,
continued by a stone quay, may date from ca 400 B.C. Still farther north, a piece of
timber first interpreted as part of a landing-stage may in fact be part of a gangway of
the earliest of the shipwrecks so far discovered (Camilli 2004). In the Bay of Naples,
a mole of concrete and timber has been discovered at Baiae. Some of these finds
represent the use of various timber form work techniques in quays and riverbanks;
others confirm the generally accepted hypothesis of the existence of freestanding
timber structures. A growing tendency toward regularity in the shape of harbor
basins can be traced back to classical times, most probably since a relationship had
to be established with the street plan of the city. Roman engineers were able to
impose regularity even when it was not provided by the natural features.
Projecting jetties greatly increased the amount of quay space available within a
harbor basin, which was otherwise restricted by natural contours until the great
man-made harbors of the Roman period. Hellenistic Delos had an unusual length
of quay for its period (1,700 m on a natural shoreline of 1,100 m), produced by
projections and indentations. There were several small jetties in the main harbor of
Piraeus. The north and south breakwaters at Sebastos, the Herodian harbor of
Caesarea in Palestine, were 300 and 800 m long respectively, and probably occupied
by quays along their entire interior length (figure 25.3; Josephus, BJ 1.411-13;).
Trajan's hexagonal harbor basin at Portus had sides nearly 358 m long, almost
entirely used as quay; the Claudian harbor basin had a combined quay length of
about 1,350 m; Leptis Magna and Terracina both 1,200 m. At Alexandria, the latest
estimate for the quay space in the first three inner basins, comprising only the
eastern half of the Great Harbor and much increased by projecting jetties from the
third century B.C. on, is about 3,000 m (figure 25.4; J. Cole, oral communication;
this estimate may well rise as excavation progresses).
A few quays, like that at Leptis Magna, had two levels, but most had only one.
The height of the quay surface above water level cannot have varied greatly from site
to site. The archaeological remains of harbor structures are now at various heights
with relation to sea level, but we have to allow for varied changes in relative sea level,
both eustatic and isostatic in character, along with the subsidence of quays and
moles, from the liquefaction of the sediments supporting them. As designed, quay
surfaces not in exposed positions need not have been much more than 1 m above sea
level, given the virtual lack of tidal change (Blackman 1982: 203). It has been argued
that a depth of 2-3 m was necessary to float a ship close to a quay, and at Ephesus the
harbor floor seems to have been 4 m below the waterline against a Roman quay; but
preliminary conclusions from the ROMACONS project indicate that quays may
have been built in much shallower water as well (Oleson et al. 2004: 221).
\.- ';.,._ ... ~
__.,.
_...... .. ·----· ---
- -·--.••-t• • ...-:::- ..... J_···;:· -·
.....~. ,.., ••~-- ,...,,;:,.-."':•.,.r:_;~~-:r,.T• ••"'• .. -$'"'-.,..· -

......... ........_~--~
Figure 25.3. Reconstruction of Sebastos, the harbor of Caesarea in Palestine, early first
century A.D. (Reconstruction by C. Brandon, by permission.)
Convenience suggests that a ship's deck should be level with the quay surface
for unloading, as depicted on the Torlonia Relief (figure 25.1), but some ancient
reliefs show the gangway at a steep angle (Basch 1987: fig. 1048; Casson 1996: fig.
174). Visual representations show that ancient merchant ships usually moored stem
or stern to shore, although not necessarily perpendicular to the quay. In river
harbors broadside mooring obviously would have been more common. Ships were
made fast to pierced mooring-stones or bollards. Study of these features deserves
more attention, for they may indicate ship sizes and mooring methods (Blackman
1988). A famous series are the pierced stone blocks set in the brick quay face of
Trajan's harbor at Portus, probably originally totaling more than 100. Such
mooring stones were built into the quays at the river port of Rome itself, as well as
the quays of Terracina and Leptis Magna and the river harbor of Aquileia (Brusin
1934: 16-26). Although less common, they are found in pre-Roman, Greek harbors.
At Teos they were wedge-shaped to take the lateral strain (Blackman 1973a: 115-17).
Stone bollards were usually set vertically in the quay. Appian (Pun. 96) refers to
cables in the commercial harbor at Carthage-presumably mooring cables that
mariners could pick up on entering harbor.
The operation of berthing and unloading in a busy harbor was complex. The
numbering of columns set back from the edge of the quays of the Trajanic harbor at
Portus must have had the purpose of defining individual berths; the main purpose
of the columns may have been to support a roof. Some ancient illustrations, almost
entirely of the Roman period, give us a glimpse of harbor operations, particularly
PHARO$ IN PAlACE AREA: MUSEUM
LIGHTHOUSE SEMA • ENCLOSUR
FOR TOMI
ANDll<Ef
,j
IN MIDOLE OF THE CITY: COURT OF J
GAOi/ES
PANEIUM
EUNOSTOS
HARBOUR
tv
''v'-=!? -~F=-ll --H= ~ ~~_;.. ------- --
I
1 PARALLEL
II {STREETS! I
---1i------- ---:
II
jj==[l== ji==Jl=- 91=~=~ ~-..:)1t_ =====v- ·--J'
I /
1r--G-= 9(--1F={l=--IJ,,,=:a{l,==I)--:
[l=--!l--- !i"'---11-- - --- --- .---, --"11"---"i
m-s~~/
NECROPOLIS
---y---1 ---r--i' ,,.-
____Ill... __ J 11j ,- --------'
/
/
/
/
.....__/ --- -- ,-
-------
-- ✓
~
LAKEMAREOTIS -
~
_r./""-..
~~~ )
0
-==- -=--= ==== ==-- ---,= ===- --___ ;5 KM
25.4. Alexandria: reconstr ucted plan of the port, with locations mention
ed by Strabo. (Courtesy of J. McK
unloading. Inscriptions, also mainly Roman in date, reflect the variety of the op-
erations and skilled trades involved: crane operators, stevedores, sand-ballastmen,
harbor-boatmen, lightermen, tugboatmen, divers, tally clerks, and so on (Casson
1996: 369-70). Elsewhere we have references to the operations of shipbuilding and
repair, but, as has been said, no reference to harbor engineers: they must have
counted simply as "engineers" (mechanikoi). All these were skilled trades that
continued to be practiced throughout antiquity, and one wonders whether the Arab
conquest made much difference to the composition of the skilled workforce in the
lands they conquered.
Cranes were used on quaysides already in the classical Greek period, although
the evidence is mainly epigraphic; various systems of pulley-blocks and tackle were
also in use, which is not surprising given the importance of such devices for ships'
rigging. For the Roman period there is somewhat more evidence for the use of
cranes on both quays and building sites (Blackman 1982: 211, n. 103; cf. chapter 13).
Vitruvius (De arch. 10.2.10) mentions a dockside crane on a rotating base that
allowed more convenient pickup and deposit of the loads.
The water supply system was always an important facility for those working in
the port and particularly for visiting mariners wishing to replenish supplies. At
Leptis Magna, there was a barrel-vaulted water-cistern right on the quay at the
southeast corner of the harbor; a watering point has been identified at Cosa; and
remains of an aqueduct have been found at Portus, as would be expected. At
Cenchreae and Forum Iulii (Frejus) a well has been found beside the harbor, and at
Antipolis an aqueduct and watering-point. At major fleet bases the need would
have been particularly great: at Misenum a vast underground reservoir fed by an
aqueduct was cut into the hill beside the harbor in the time of Augustus; it had a
capacity of 12,600 cubic meters of water.
Behind the quays of larger harbors lay the warehouses, workshops, buildings
for commercial transactions, bars, brothels, and hostels. At some Greek sites, such
as Piraeus, the emporion was a well-defined area with its own boundary stones,
aligned on the city grid, and the famous stoas. There and at Delos, Thasos, and
Ephesus, for example, it is possible to identify the nature of some of the buildings as
hostels (perhaps also brothels) or workshops. More evocative for Athens are our
literary sources, such as Aristophanes (Ach. 545-54), for the sounds and smells of a
port full of activity (Casson 1991: 97-112). Delos was an exceptional site, starting as a
major sanctuary and becoming an important entrepot and slave market, especially
after 167 B.C. when the Romans declared it a free port. Cenchreae, at least for the
Roman period, provides us with a vivid picture of harborside quays, warehouses,
and a sanctuary of Isis. Archaeological evidence is more plentiful for all these
features in the Roman period.
We obtain, mainly from epitaphs and other inscriptions, a clear impression of
the large numbers of foreigners resident in ports, and the frequent grants of special
legal status and the right to maintain their own cults. Ports were certainly the most
cosmopolitan of ancient sites, and the place where the greatest variety of cults was
found. For the Greeks the main focus was on divinities that ensured safe voyages,
but a variety of foreign cults are well attested. At Alexandria there was a "foreigners'
emporion," while Delos after 167 B.C., probably exceptionally, welcomed traders of
every nation. In Roman times the foreign community was a standard feature of
harbor towns. We know of many groups of foreign traders resident in places such as
Ostia and Puteoli, and they must have contributed greatly to the wealth and to the
rich maritime expertise of the harbor towns concerned. Dock laborers, however,
were an unruly element (Rickman 1988).
NAVAL HARBORS
From the early fifth century B.C. onward, one can see at most Greek coastal cities or
their out-ports the development of a separate military harbor, or a section for
warships within the harbor, provided with the greater security required for military
installations (Baika 2006). At Piraeus, the out-port of Athens from the early fifth
century, the two smaller harbors and part of the larger harbor were reserved for
warships. The dockyard was probably a restricted area, walled off from the land-
ward side, although the famous naval storehouse (Philon's Arsenal, built after 347/6
B.C.) could be visited by citizens to check public property. There is plenty of
evidence in Athenian literature of the fifth and fourth centuries that the dockyard
was a matter of great civic pride (e.g., Demosthenes 22.76, 23.207).
Other great ports such as Rhodes, Aegina, Thasos, Cnidus and, in the west,
Carthage, also had separate military harbors. The harbor of Cnidus has not been
investigated in detail, but the smaller, western harbor was known in antiquity as
"the trireme harbor." A limited survey at Aegina has confirmed the north harbor as
the military harbor, with remains of shipsheds. The separation of civil and military
harbors should not, however, be exaggerated. It is logical to assume that there was
one pool of skilled labor for specialist tasks in all port areas-rope- and sailmaking,
caulking and repair of ships, shipbuilding, etc.-and that most of the population of
the civil port would have also served as rowers in the warships of the Greek city
states.
In their heyday, major naval powers such as Athens and Rhodes maintained
naval bases at points on their own coasts, distant from the main naval base, to
provide defense in depth. Athens maintained bases even beyond her borders. In
many ways these are forerunners of the provincial fleet bases of the Roman Empire.
In the classical and Hellenistic Greek periods, coastal cities and the out-ports of
inland cities were normally fortified, and the actual harbor basin or basins was
usually enclosed within the fortifications (Blackman 1982: 193-95). The city walls
were extended along the harbor moles, to end in towers as at any normal city gate.
Entrances were kept narrow enough to be covered by artillery mounted on the
towers and for chains or booms to be used to close the entrance. Use of these devices
is attested by historians describing sieges, and by authors of technical handbooks on
siege techniques and countermeasures. Evidence for an entrance boom may be
preserved at ancient Halieis, although the fittings may be for a land gate rather than
a harbor entrance. Remains of towers and solid block buildings can still be iden-
tified at the entrance of many ancient harbors; in addition to housing artillery and
the apparatus for hauling the entrance chains, they may also have served as light-
houses. At Thasos and Phalasarna, the course of the harbor walls incorporated
towers. The improvements in artillery, which included ship-borne artillery, led to
more heavy fortifications in general in the fourth century and later. There are some
cases of harbor entrances being set back so that any enemy ship would be exposed to
flanking fire from the shore if it tried to attack: Cantharus and Zea harbors in
Piraeus and the "trireme harbor" at Cnidus are good examples. Controlled en-
trances were also a useful means of supervising commercial shipping; and research
between 1996 and 2006 has revealed the fortified entrance from the Nile to the
harbor basins at the sanctuary of Herakleion.
Ancient Greek geographers and coastal pilots refer to many cities as having a
"closable harbor" (kleistos limen), which must refer to such arrangements (Leh-
mann-Hartleben 1923: 65-74, 122-61). A harbor that was closable would presumably
normally have been enclosed within fortification walls as well. It is likely that
military harbors would also have been walled off on the landward side. The same
may have been true for major commercial harbors, where the authorities would
have wanted to control the movement of people, particularly foreigners, and goods
in and out of the city. In many cities the emporion was a clearly delimited area
(Lehmann-Hartleben 1923: 105-21).
Significant for the history of ancient town planning is the growing evidence for
the insertion of harbor installations into the city grid. We now know that the
dockyard at Sicilian Naxos was aligned on the street grid and clearly formed part
of the fifth-century B.C. city plan. At the new city of Rhodes (after 408 B.C.)
the rectangular military harbor Mandraki, lined by shipsheds, also conformed to
the city grid. It has even been suggested that the lines of the harbor influenced the
original layout of the grid-a sensible arrangement since harbor installations had to
take account of the line of the coast. In the fourth century B.C., Cnidus seems to
have had two street grids, aligned on the two harbors-the military and the
commercial. At Alexandria we know now from geophysical research that the
Heptastadion, the famous causeway linking the island of Pharos to the mainland,
was aligned on a main city street, and the main dockyard lay beside it (McKenzie
2003: 36-41; figure 25.4). The shipsheds in the Royal Harbor in the southeast corner
of the Great Harbor were, however, on a different alignment.
In military harbors the most distinctive feature was the shipsheds, which must
have taken up most of the available shoreline, leaving less space for quays than in
commercial harbors (see below). The two smaller and exclusively military harbors
of Piraeus were probably an extreme example. When the Athenian fleet went to sea,
Figure 25.5. Piraeus: Trireme harbor of Munichia in the fourth century B.C., recon-
struction by John Coates. (After Morrison et al. 2000: 229, fig. 71; by permission of
J. Coates.)
the triremes launched from the shipsheds of Munichia and Zea (figure 25.5) had to
sail round to the choma, a jetty in the largest harbor, Cantharus, for official in-
spection, and it was probably there that most of the crew embarked. In Rhodes the
military harbor may well have had shipsheds along most of its landward shoreline;
it is plausible that most quay space was provided along the inner side of the mole
that formed the eastern side of the harbor (Blackman et al. 1996). Archaeologists
studying the military harbor of Thasos assume a similar arrangement there.
Military harbors needed considerable space for equipment, although not di-
rectly at the waterside. Oars and spars seem normally to have been stored in the
shipsheds, and at Carthage there were storerooms for the gear above the shipsheds;
elsewhere probably above or alongside the installations. The Athenian naval lists
refer to a number of storage buildings in the dockyard area; the great naval
storehouse (for sails and rigging) built by Philon in Piraeus was rediscovered in
1988-89. The naval harbors of the Roman period are not well known, even those of
the Imperial period. In his magisterial survey, Redde (1986: 145-319) stresses how
little archaeological evidence we have for Roman naval installations, in striking
contrast to their commercial harbors.
ANCIENT SHIPSHEDS
By the early fifth century B.C., the constant struggle for naval superiority in the
Aegean had led to the refinement of one of the most extreme warships of antiquity,
the Greek trireme (Morrison et al. 2000). The successful employment of warships
depended on having them regularly dried out, and since they could not operate in
high waves, they were usually not taken out to sea after the end of the sailing
season. The extreme qualities of the warships also made their light wooden con-
struction vulnerable to the deteriorating forces of sun and rain. Furthermore, the
aggressive shipworm (Teredo navalis) of the Mediterranean, against which com-
mercial vessels could protect themselves by metal sheathing, inflicted serious
damage on the warships. Thus, the delicacy of the construction and the large
investment inherent in the ships made it necessary that they were carefully pro-
tected when not in use.
In order to protect warships effectively, it was necessary to haul them out of the
water and provide them with cover against sun and rain. For this reason, permanent
and purpose-built naval installations were constructed-neosoikoi, usually trans-
lated as "shipsheds." A single shipshed is basically a long and narrow roofed hall just
large enough to house a ship. It is open toward the sea and has a sloping floor
(slipway) with a gentle inclination from the water's edge up to the landward end.
The slipways were sometimes rock-cut, sometimes built up with ramps of earth,
plastered earth and rubble, or sand, most probably always clad with timber. The
ship was moved into position outside the shipshed and then hauled, stern first, up
the slipway until it was completely out of the water and under the roof. There it was
braced and fastened, and could be kept in store until it was launched the next time.
No remains of hauling-gear have yet been found at the top of slipways-a possible
bollard in Munichia could be just for making the ship fast-so we must assume that
the ships were manhandled up the slipways (as they were up the beach when away
on a mission). The exception may be some very steep slipways such as those at
Sounion.
Where a large number of ships had to be housed, several shipsheds were built
next to each other in a continuous row, presenting a common front toward the sea.
Good ventilation was necessary for the drying out and preservation of the ships,
and to facilitate work in the sheds; this prompted the use of columns, pillars, or
posts, rather than continuous walls, to support the roof. Only limited maintenance
work could be done on the ships in these confined spaces, but we now know that
this included painting, at least of the bow and stern where there was more space.
Red pigment has been found in the Naxos shipsheds, effective as an anti-fouling
paint, but dangerous for those applying it.
Some of the oldest shipsheds we know of were built on Samos around 530 B.C.
No definite archaeological remains have been found there as yet, and we have only
the testimony of Herodotus (3.39, 3.44-45) to confirm their existence. Herodotus
(2.159) also mentions having seen some kind of hauling installations (holkoi) built
on the Red Sea shore during the reign of the Egyptian pharaoh Necho II (c. 600
B.C.), but it is not clear whether these were true shipsheds or not. The extant
archaeological remains include shipsheds dated to the early fifth century B.C.
at Corcyra, Abdera, and Aegina, and point to the leading role of the Greek poleis
in the development of such facilities. It may be assumed that from this time
onward every major Greek city that boasted its own fleet also had numerous
shipsheds.
Athens had one of the largest fleets of the fifth and fourth centuries B.C., and
consequently needed a corresponding number of shipsheds. The naval inventory
lists prepared annually by the magistrates responsible for the dockyard in the
fourth century B.C. (JG 22 1604-1632) tell us of 372 shipsheds in the three harbors of
Piraeus at Munichia, Zea, and Cantharus (today the main harbor). The majority of
the sheds (196) lay in the military harbor of Zea. The series of shipsheds crowding
the waterfront of Zea and Munichia constituted monumental constructions,
outstanding among buildings in Athens for their sheer size (figure 25.5). Archae-
ological investigations, both on land and under water, were begun in 2001 on the
Zea sheds, first excavated in 1885; these have revealed large areas of the ancient
harbor installations, shipsheds, quays, fortifications, etc. Evidence indicating a
relative rise in sea level here of about 2 m since antiquity probably explains why so
much of the ancient harbor frontage has been preserved under water (http://
www.zeaharbourproject.dk).
Besides Athens, other major naval powers had large-scale military harbor
complexes with impressive numbers of shipsheds. Syracuse had 310 at the begin-
ning of the fourth century B.C. (Diodorus 14.42.5); the military port of Carthage
could accommodate 220 ships in the third/second century (Appian, Pun. 96); and
finally Strabo (12.8.11) reports more than 200 shipsheds in the harbor of Cyzicus.
Medium-sized naval powers such as Rhodes, Samos, Massalia or Chios had
naval bases with 50 to 100 shipsheds, while the norm seems to be bases for 10 to 50
warships, as in Aegina, Cos and Oeniadae, or Citium in Cyprus. Naxos in Sicily has
a surprisingly small dockyard (figure 25.6). Finally, we also find single shipsheds or
groups of two or three, usually rock-cut. They are mainly fortified and located on
remote promontories strategically situated for the surveillance of maritime sea-
lanes. Characteristic examples are the twin shipsheds at Sounion in Attica and the
single fortified shipshed of Trypeti near Seteia in Crete (Baika 2003). Larger is the
rock-cut complex of Oeniadae excavated in 1989 to 1995 (Kolonas 1989-1990).
Besides Piraeus, major excavations have been carried out in the shipsheds of
Carthage, Massalia, Citiium and Naxos (Blackman 2003; Hurst 1977, 1979, 1994;
Hesnard et al. 1999, 2001; Yon 2000; Blackman and Lentini 2003, forthcoming a, b).
The shipshed complexes in major dockyards such as Carthage and Piraeus reflect
more than mere utilitarian purposes. They attest a high level of technology and a
massive expenditure of resources and effort, which raise practical questions about
logistics and historical questions about political aspirations behind their creation.
Figure 25.6. Naxos: plan of dockyard, with the outline of the Olympias superimposed. (After
Blackman and Lentini 2003: 407, fig. 23; courtesy of D. J. Blackman and M. C. Lentini.)
Examples of Greek and Phoenician shipsheds have been found all over the
Mediterranean, from the Levantine coast (Raban 2003) to Massalia, and one of the
most spectacular shipshed complexes was built in Carthage. At Massalia there
appear to be remains of open slipways as well as roofed shipsheds. The use of
shipsheds is testified throughout the Greek world in the classical and Hellenistic
periods, and the need for them must have remained somewhat the same in the
Roman period. Although Roman shipsheds are attested in literature and icono-
graphic evidence, no definite archaeological evidence has yet been identified;
possible examples are now being studied (Ciampoltrini and Rendini 2004; Black-
man forthcoming; Rankov forthcoming; Cozza and Tucci 2006). Roman shipsheds
were not necessarily confined to naval bases, and small groups may have been
attached to major maritime villas.
Shipsheds are the only direct, concrete archaeological evidence that we possess
for calculating the dimensions of ancient warships. Intended to house various types
of oared vessels, a shipshed's dry length, clear width, and height (where it can be
calculated with a certain precision) offer us the approximate outside measurements
of the vessel that it once accommodated. The study of the hauling and launching
operations on the inclined slipways, and the evaluation of the basic functioning of
shipshed architecture, can help toward more precise understanding of the tech-
nology of the ships concerned, and of ancient maritime technology in general (cf.
Blackman and Rankov forthcoming; Blackman and Lentini forthcoming b).
HAULING-WAYS
Canal construction was a limited development in the classical Greek world (see
chapter 12). Periander of Corinth tried to cut a canal across the Isthmus of Corinth
in the sixth century B.C., to provide a shorter and safer route from the Aegean to the
west, but was not successful; instead he built a stone runway across it (a diolkos,
"haul-across"), over which warships and perhaps small merchant ships could be
hauled (figure 23.6). This structure remained in use into the Roman period, and
sections of it are still visible today. Thucydides (3-15.1) describes the Spartans
constructing holkoi to carry ships across the isthmus in 428 B.C.; this sounds like
hauling-gear or possibly wheeled cradles to transport the ships along the tracks still
visible in the surface of the runway. The cradle could have been slipped under the
hull at the submerged end of the runway, if it existed (but the ends are lost).
Thucydides' references to the operation (e.g., 8.7) speak of carrying the ships across
the isthmus, and do not use the word diolkos. There are possible remains of a crane
near the west end of the diolkos, but this could have been for lifting cargo, not ships.
It has been argued that the diolkos was mainly used for transporting cargo, and only
occasionally ships (MacDonald 1986). Thucydides also describes the hauling of
ships across the isthmus at Leucas (3.81.1, 4.8.2). A diolkos is mentioned at Alex-
andria, probably across the north end of the Heptastadion, but it has not been found
(Fraser 1961). A diolkos may have connected the military and west harbors at
Rhodes.
As mentioned above, Herodotus saw holkoi in Egypt, but it is not clear whether
these were hauling-gear or hauling-ways. The fact that they were still visible in his
time implies the latter. The holkoi at Delos seem to have been hauling-ways
(Duchene and Fraisse 2001: 158). Our evidence for horizontal hauling devices is
limited to references in literature. Windlasses, called onoi (donkeys) in Greek, were
used by the Persians to draw tight the cables of their bridge across the Hellespont
(Herodotus. 7.36), and by the Athenians in small boats to pull out the stakes of the
palisade protecting the dockyard in the Great Harbor of Syracuse (Thucydides
7.25.6). No evidence for hauling-gear has been found in shipsheds. The earliest
reference to ship-hauling gear is to the screw-windlass invented by Archimedes in
the third century B.C. to launch a superfreighter built for Hieron II of Syracuse
(reigned 269-215 B.C.; Athenaeus 5.207b; Casson 1996: 191-99; Baika 2002). It was
subsequently discovered that the freighter could not enter most existing harbors to
deliver its cargo of grain, so Hieron gave the ship to the king of Egypt (Ptolemy III
or IV). Perhaps the construction of enormous ships at Alexandria had forced the
harbor technicians to develop special facilities to accommodate them. The "Forty-
Banked" warship of Ptolemy IV (221-203 B.C.) was initially launched manhandled
on a cradle, but later a Phoenician engineer conceived of the dry-dock method for
docking it (Athenaeus 5.204c-d). In Latin literature we find references to machinae
for hauling ships (Horace, Carm. 1.4.2; Livy 25.11.18, 39.50.3).
LIGHTHOUSES
An important introduction of the early Hellenistic period was the lighthouse; the
first known example was the Pharos of Alexandria, named after the island on which
it was built. In antiquity lighthouses were used almost entirely as navigation marks
indicating the position of harbors, and only rarely to warn against potential hazards
to navigation (though harbor entrances could be dangerous, for example at
Alexandria (Redde 1979; Pensa 1999). Beacons, landmarks, and leading-marks were
used earlier (Morton 2001: 210-14). The Colossus of Rhodes fulfilled a similar
purpose: it almost certainly stood, not astride the entrance to the military harbor, as
has often been suggested, but on its eastern side. Both the Pharos and the Colossus
figured among the seven "Wonders of the Ancient World" (Brodersen 1992). The
lighthouse at Portus stood near the harbor entrance to the Claudian harbor, on the
west mole (but not on a separate island, as was once thought). It was commonly
depicted on mosaics, coins, reliefs, and lamps (figure 24.1; Redde 1979). The
lighthouse at Leptis Magna stood in a similar position, as did most others. A well-
preserved lighthouse dating to the third quarter of the first century A.D. was found
at Patara in 2003 (Yildirim and Gates 2007: 315). A few lighthouses, however, stood
on commanding heights: for example on the top of the East Island at Apollonia, the
port of Cyrene, and Roman lighthouses at Corunna in northwest Spain, and on two
hilltops above the port of Dover and across the Straits at Boulogne.
SHIPBUILDING YARDS
Shipsheds were not used for shipbuilding, since there was no space in them for
more than limited repair or maintenance. We have long assumed the existence of
impermanent structures, such as can be seen on beaches today, that have dis-
appeared without trace. At Marseilles, a shipbuilding yard of just this sort, dating
to the fifth and fourth centuries B.C., was found between 1992 and 1993, occupy-
ing at least 150 m of shoreline. Pieces of timber for shipbuilding lay on the beach,
and basins had been cut in the sand for soaking timbers (Hesnard et al. 1999: 34-37;
2001: 173-74). Partially worked timbers and parts of boats waiting for reuse, to-
gether with carpenter's tools, have been found on the shore of the Roman port of
Olbia in Sardinia (Kingsley 2004: 92). A navale for shipbuilders is attested at Ostia
in an inscription of the late second century A.D. (CIL 14.376; Rankov forthcoming).
SILTATION AND DREDGING
Siltation was a problem for harbor engineers in many parts of the Mediterranean,
particularly at or near river mouths, which were otherwise ideal places for the
harbors of cities with access to a deep hinterland (Blackman 1982: 199-202). Ac-
tivities on shore nearby sometimes added to the problem, for example, the dumping
of refuse from stoneworking activities into the harbor basin at Ephesus. One must
remember that the depths of water to be maintained were smaller because of the
lower draught of ships, and, in the Mediterranean, the small tide range, which was
not an effective method of flushing harbor basins.
Basic dredging methods were developed in antiquity. Excavations at Naples
have revealed dredging channels gouged across the sandy floor of the deepest levels
of the harbor (late fourth century B.C.), and at Marseilles there is clear evidence of
deep dredging in the first, third and probably fourth centuries A.D.; also the wrecks
of three Roman dredgers were found abandoned on the beach or reused in foun-
dations of a quay and landing-stage.
Much, however, depended on methods of preventing siltation, by the placing of
breakwaters to deflect silt-bearing currents or to allow controlled currents through
harbor basins (see above). This was easier to achieve if the harbor had two en-
trances. Flushing channels were already known in Phoenician Sidon, and under-
water ashlar-lined channels are clearly visible in the north harbor of Mytilene;
tunnels through moles have been reported elsewhere. At Paphos, there were two
broad channels through the eastern breakwater, protected by an outer breakwater.
Two long de-silting channels cut through the promontory at Cosa brought silt-free
water into the harbor, but their primary purpose was to provide brackish water and
a migration route for the adjacent fishery. A rock-cut de-silting channel has been
identified at Phalasarna. At Seleucia Pieria, port of Antioch near the mouth of
the Orontes, attempts were made to use the tide and then the river current, di-
verted through a canal, to flush the harbor basin. Roman engineers seem to have
learned that some breakwaters only made matters worse by blocking coastal cur-
rents, so they started to build offshore breakwaters not connected to the shore, as at
Centumcellae, and arched moles (see above).
SEA LEVEL CHANGE
Changes in sea level greatly affect the functioning of harbors, and there is evidence
that this possibility caused concern in antiquity. Actual relative sea level change at
any particular site could be caused by sea level change or by earth movements;
general appreciation of these factors grew during antiquity (Morton 2001: 13-29).
The central Mediterranean is a geologically unstable area, with a line of plate
tectonic boundaries. Evidence of human response in harbor rebuilding is so far
limited, and deserves further study. At Rhodes the slipway ramps were raised,
perhaps after the major earthquake of 227 B.C., and later uplift of the land may have
caused the abandonment of the dockyard; the factor of relative sea level change is
now taken into account in the study and interpretation of the shoreline features in
Zea harbor. In some cases, as at Cenchreae, the submergence was too catastrophic
for repair; the city of Helice was submerged by a tidal wave in the earthquake of 373
B.C. and disappeared. At other sites, such as Phalasarna in western Crete, the harbor
was put out of action by massive uplift of more than 6 m. See Blackman 1973b, 2005.
CONTEXT
Scholars have begun the complex task of setting ancient ports in their geographical,
historical, economic, and cultural context (e.g., Horden and Purcell 2000; Cunliffe
2001; Rickman 2005 and associated bibliography). The questions are numerous and
difficult to answer. How did ports relate with the city that they served, with their
hinterland, and with other ports? How separate were the ports, both military and
commercial, from the cities? How did some ports develop without apparent
hinterland? The outstanding example of this is medieval Amalfi, but there are many
ancient ports that seemed to have served rather for coastal traffic: cabotage needed
a regular series of stopping points. Cities near river-mouths may have connected
with river systems, canals, and lagoons-particularly in Spain, southern and At-
lantic France, and the west coast of Italy, including Pisa and Rome (Pasquinucci
and Weski 2004). The dramatic discovery of sunken boats in a silted-up river
channel at Pisa has drawn most attention to the actual boats, but one may ask
about the nature of the site where at least 20 boats were wrecked in antiquity.
SELECT BIBLIOGRAPHICAL GAZETTEER

OF SITES MENTIONED
Aegina: Knoblauch 1969, 1972.

Alexandria: Jondet 1916; Fraser 1961; Goddio 1998; McKenzie 2003; Goddio
and Bernand 2004; Goddio, Clauss et al. 2006; Empereur forthcoming;
McKenzie 2006.
Apollonia: Flemming 1972: 95-126; Laronde 1996; Blackman and Rankov
forthcoming; Sintes forthcoming.
Baiae: see Puteoli.
Caesarea in Palestine: Raban and Hoium 1996; Hohlfelder 1996.
Carthage: Hurst 1979, 1994.
Cenchreae: Scranton, Shaw, and Ibrahim 1978; Hohlfelder 1985.
Citium: Karageorghis and Michaelides 1995; Yon 2000.
Cosa: McCann et al. 1987; Ciampoltrini and Rendini 2004.
Delos: Duchene and Fraisse 2001.
Ephesus: Zabehlicky 1999.
Halieis: Jameson 1969, 1973.
Herakleion: Goddio forthcoming; Goddio et al. 2006.
Lecheum: Rothaus 1995.
Leptis Magna: Bartoccini 1958.
London: Milne 1985.
Massalia: Hesnard et al. 1999, 2001; Hesnard 2004.
Misenum: Maiuri 1958: 91-100; Doring 2003; Benini and Lanteri forthcoming.
Mytilene: Williams forthcoming.
Naples: Giampaola 2006; forthcoming.
Naxos: Blackman and Lentini 2003, forthcoming a; Lentini and Blackman
forthcoming.
Ostia: see Portus.
Phalasarna: Hadjidaki 1988; Frost and Hadjidaki 1990.
Phaselis: Schafer et al. 1981.
Piraeus: Garland 1987; Blackman and Rankov forthcoming.
Pisa: Bruni 2000, 2003; Camilli 2004.
Portus: Testaguzza 1970; Meiggs 1973; Keay et al. 2004; Rickman 2005; Keay
and Millett 2006, forthcoming.
Puteoli: Dubois 1907; Gianfrotta et al. 1993: 21-70; 2001 passim; 2002: 47-66;
Doring 2003; Piromallo 2004; Miniero forthcoming.
Rhodes: Bradford 1957: 277-86; Brodersen 1992; Blackman et al. 1996.
Samos: Simossi 1991.
Seleucia Pieria: Chapot 1907; Lehmann-Hartleben 1923: 214-16; Poidebard and
Lauffray 1951: 31-32.
Sidon: Poidebard and Lauffray 1951; Frost 1973.
Sounion: Goette 2000.
Thasos: Archontidou-Argyri et al. 1989; Blackman and Rankov forthcoming.
Tyre: Poidebard 1939; Frost 1971: 110-11.
REFERENCES
Archontidou-Argyri, A., A. Simossi, and J.-Y. Empereur 1989. "The underwater excavation
at the ancient port ofThasos, Greece," International Journal ofNautical Archaeology 18:
51-59.
Baik.a, K. 2002. "Dispositif du halage des hangars navals antiques: Etude ethno-
archeologique," Tropis 7: 43-83.
Baik.a, K. 2003. "Operating on shipsheds and slipways: Evidence of underwater configu-
ration of slipways from the neosoikos of 'Trypiti' (Crete)," in Beltrame 2003: 103-8.
Baik.a, K. 2006. "Early naval bases and military harbour infrastructure in the Mediterra-
nean," in The new view: Underwater archaeology and the historical picture. Antiqua 40.
Basel: Archaologie Schweiz, 176-92.
Bartoccini, R. 1958. Il porto romano di Leptis Magna. Bollettino del Centro di Studi per
la Storia dell'Architettura Suppl. 13. Rome: Centro di Studi per la Storia dell'Archi-
tettura.
Basch, L. 1987. Le musee imaginaire de la marine antique. Athens: Hellenic Institute for the
Preservation of the Nautical Tradition.
Beltrame, C. (ed.) 2003. Boats, ships and shipyards: Ninth International Symposium on Boat
and Ship Archaeology, Venice 2000, Proceedings. Oxford: Oxbow.
Benini, A., and L. Lanteri forthcoming. "11 porto romano di Misenum: Nuove acquisi-
zioni," in Blackman and Lentini forthcoming b.
Blackman, D. J. (ed.) 1973a. Marine archaeology. Colston Papers 23. London: Butterworths.
Blackman, D. J. 1973b. "Evidence of sea level change in ancient harbours and coastal
installations," in Blackman 1973a: 115-39.
Blackman, D. J. 1982. "Ancient harbours in the Mediterranean," International Journal of
Nautical Archaeology 11.2: 79-104; 11.3: 185-211.
Blackman, D. J. 1988. "Bollards and men," in I. Malkin and R. L. Hohlfelder (eds.),
Mediterranean cities: Historical perspectives: Mediterranean Historical Review 3.1.
London: Cass, 7-20.
Blackman, D. J. 1995. "Naval installations," in R. Gardiner (ed.), The age of the galley.
London: Conway Maritime Press, 224-33.
Blackman, D. J. 2003. "Progress in the study of ancient shipsheds: A review," in Beltrame

2003: 81-90.
Blackman, D. J. 2005. "Archaeological evidence for sea level changes," Zeitschrift fur
Geomorphologie, n.s. Suppl. 137: 61-70.
Blackman, D. J. forthcoming. "Roman shipsheds," in Hohlfelder, forthcoming.
Blackman, D. J., P. Knoblauch, and A. Yiannikouri 1996. "Die Schiffshauser am Man-
drakihafen in Rhodos," Archiiologischer Anzeiger 1996: 371-426.
Blackman, D. J., and M. C. Lentini 2003. "The shipsheds of Sicilian Naxos, researches 1998-
2001: A preliminary report," Annual of the British School at Athens 98: 389-436.
Blackman, D. J., and M. C. Lentini forthcoming a. "Further research on the dockyard of
Sicilian Naxos," Tropis 9.
Blackman, D. J., and M. C. Lentini forthcoming b. Ricoveri per navi militari nei porti del
Mediterraneo antico e medievale. Workshop Ravello, 4-5 nov. 2005. Bari: Edipuglia.
Blackman, D. J., and N. B. Rankov (eds.) forthcoming. Ship sheds ofthe ancient Mediterranean.
Bradford, J. 1957. Ancient landscapes. London: Bell.
Brandon, C. 1999. "Pozzolana, lime, and single-mission barges (Area K)," in K. Hoium, A.
Raban, and J. Patrich (eds.), Caesarea papers, 2. Journal of Roman Archaeology Suppl.
35. Portsmouth, RI: JRA, 169-78.
Brandon, C., R. L. Hohlfelder, J.P. Oleson, and C. Stern 2005. "The Roman Maritime
Concrete Study (RO MACONS): The harbour of Chersonisos in Crete and its Italian
connection," Mediterranee 1.2: 25-29.
Branton, G., and J. P. Oleson 1992. "The technology of King Herod's harbour," in R. L. Vann
(ed.), Caesarea papers: Straton's Tower, Herod's Harbour, and Roman and Byzantine
Caesarea. Journal of Roman Archaeology Suppl. 5. Ann Arbor, MI: JRA, 49-67.
Brodersen, K. 1992. Reisefuhrer zu den Sieben Weltwundern: Philon von Byzanz und andere
antike Texte. Frankfurt: Insel Verlag.
Bruni, S. 2000. Le navi antiche di Pisa: Ad un anno dall'inizio delle ricerche. Florence:
Polistampa.
Bruni, S. 2003. Il porto urbano di Pisa antica: La Jase etrusca, il contesto e il relitto ellenistico.
Milan: Silvana.
Brusin, G. 1934. Gli scavi di Aquileia. Udine: "La Panarie."
Camilli, A. 2004. "Le strutture 'portuali' dello scavo di Pisa-San Rossore," in Gallina Zevi
and Turchetti 2004: 67-86.
Casson, L. 1989. Periplus Maris Erythraei. Princeton: Princeton University Press.
Casson, L. 1991. The ancient mariners. 2nd ed. Princeton: Princeton University Press.
Casson. L. 1996. Ships and seamanship in the ancient world. Rev. ed. Princeton: Princeton
University Press.
Chapot, V. 1907. Seleucie de Pierie. Memoires de la Societe nationale des antiquaires de
France, 66: 1-78.
Ciampoltrini, G., and P. Rendini 2004. "11 sistema portuale dell'ager Cosanus e delle isole
del Giglio e di Giannutri," in Gallina Zevi and Turchetti 2004: 127-50.
Cozza, L., and P. L. Tucci 2006. "Navalia," Archeologia Classica 57: 175-201.
Cunliffe, B. 2001. Facing the ocean: The Atlantic and its peoples, 8000 BC-AD 1500. Oxford:
Delgado, J.P. 1998. Encyclopedia of underwater and maritime archaeology. New Haven: Yale
University Press.
Doring, M. 2003. "Romische Hafen- und Tunnelbauten der Phlegraschen Felder," in
C. P. J. Ohlig (ed.), Wasserhistorische Forschungen: Schwerpunkt Antike. Siegburg:
Deutsche W asserhistorische Gesellschaft, 35-53.
Dubois, C. 1902. "Observations sur un passage de Vitruve (V.12)," Melanges de l'Ecole

Dubois, C. 1907. Pouzzoles antique (Histoire et Topographie). Bibliotheque des ecoles
frarn;:aises d'Athenes et de Rome 98. Paris: A. Fontemoing.
Duchene, H., and P. Fraisse 2001. Le Paysage portuaire de la Delos antique: Recherches sur les
installations maritimes, commerciales et urbaines du littoral delien. Exploration ar-
cheologique de Delos 39. Athens: Ecole frarn;:aise d'Athenes.
Empereur, J.-Y. forthcoming. Pharos I, Etudes alexandrines, 9.
Felici, E. 1993. "Osservazioni sul Porto Neroniano di Anzio e sulla tecnica romana delle
costruzioni portuali in calcestruzzo," in Gianfrotta et al. 1993: 71-104.
Felici, E. 2002. "Scoperte epigraphiche e topographiche sulla costruzione del porto ner-
oniano di Antium," in Gianfrotta et al. 2002: 107-22.
Flemming, N. C. 1972. Cities in the sea. Garden City NY: Doubleday.
Fraser, P. M. 1961. "The diolkos of Alexandria," Journal of Egyptian Archaeology 47: 134-38.
Frost, H. 1971. "Recent observations on the submerged harbourworks at Tyre," Bulletin du
Musee de Beyrouth 24: 103-11.
Frost, H. 1973. "The offshore island harbour at Sidon and other Phoenician sites in the light
of new dating evidence," International Journal of Nautical Archaeology 2: 75-94.
Frost, H. 1995. "Harbours and proto-harbours, early Levantine engineering," in Kar-
ageorghis and Michaelides 1995: 1-21.
Frost F. J., and E. Hadjidaki 1990. "Excavations at the harbor of Phalasarna in Crete: The
1988 season." Hesperia 69: 513-27.
Gallina Zevi, A., and R. Turchetti 2004. Le strutture dei porti e degli approdi antichi. Atti del
II Seminario ANSER, Roma-Ostia Antica, 16-17.4. 2004. Soveria Mannelli: Rubbettino.
Garcia, D., L. Vallet et al. 2002. L'espace portuaire de Lattes antique. Lattara 15. Lattes:
Association pour le Developpement de l'Archeologie en Languedoc-Roussillon.
Garland, R. 1987. The Piraeus. London: Duckworth.
Giampaola, D ., et al. 2006. "La scoperta del porto di N eapolis: dalla ricostruzione topografica
allo scavo e al recupero dei relitti," Archaeologia Maritima Mediterranea 2: 47-92.
Giampaola, D. forthcoming. "11 porto di Napoli," in Blackman and Lentini forthcoming b.
Gianfrotta, P.A., P. Pelagatti, E. Felici, and P. G. Monti 1993. Archeologia Subacquea 1.
Universita di Tuscia, Viterbo. Rome: Istituto Poligrafico e Zecca dello Stato.
Gianfrotta, P.A., P. Pelagatti, E. Felici, and P. G. Monti 2002. Archeologia Subacquea 3.
Universita di Tuscia, Viterbo. Rome: Istituto Poligrafico e Zecca dello Stato.
Gianfrotta, P.A., and F. Maniscalco 2001. Forma Maris: Forum internazionale di Arche-
ologia Subacquea, Pozzuoli 22-24 sett. 1998. Naples: Massa.
Goddio, F. 1998. Alexandria: The submerged royal quarters. London: Periplus.
Goddio, F. forthcoming. Underwater excavations in the Canopic region: The excavation and
topography of Heracleion-Thonis.
Goddio, F., and A. Bernand 2004. Sunken Egypt: Alexandria. London: Periplus.
Goddio, F., M. Clauss et al. 2006 Egypt's sunken treasures. Munich/London: Prestel.
Goette, H. R. 2000. Ho axiologos demos Sounion: Landeskundliche Studien in Sudost-Attika.
Rahden/Westf.: Leidorf.
Gunther, R.T. 1903a. "The submerged Greek and Roman foreshore near Naples," Ar-
chaeologia 58.2: 499-560.
Gunther, R. T. 1903b. "Earth movements in the Bay of Naples," Geographical Journal 22:
121-49, 269-89.
Hadjidaki, E. 1988. "Preliminary report of excavation at the harbor of Phalasarna in West
Crete," American Journal of Archaeology 92: 463-79.
Hesnard, A. 2004. "Vitruve, De architectura, V, 12 et le port romain de Marseille," in

Gallina and Zevi 2004: 175-204.
Hesnard, A., M. Moliner., F. Conche, and M. Bouiron. 1999. Parcours de Villes. Marseille: 10
ans d'archeologie, 2600 ans d'histoire. Exposition du Musee d'Histoire de Marseille. Aix-
en-Provence: Edisud.
Hesnard, A., C. Maurel, and P. Bernardi 2001. "La topographie du port de Marseille, de
la fondation de la cite ala fin du Moyen Age," in Marseille: trames et passages urbains
de Gyptis au Roi Rene. Actes du colloque international d'archeologie, Marseille 3-5
nov. 1999. Publications du Centre Camille Julian 7. Aix-en-Provence: Edisud,
159-202.
Hohlfelder, R. L. 1985. "The building of the Roman harbour at Kenchreai: Old technology
in a new era," in A. Raban (ed.), Harbour Archaeology. British Archaeological Reports,
Intl. Series S257. Oxford: BAR, 165-72.
Hohlfelder, R. L. 1996. "Caesarea's master harbor builders: lessons learned, lessons ap-
plied?" in Raban and Holum 1996: 77-101.
Hohlfelder, R. L. 2005. "Swimming over time: Glimpses of the maritime life of Aperlae," in
Pollini 2005: 187-210.
Hohlfelder, R. L. forthcoming. The maritime world of ancient Rome. Ann Arbor: University
of Michigan Press.
Horden, P., and N. Purcell 2000. The corrupting sea: A study of Mediterranean history.
Oxford: Oxford University Press.
Hurst, H. R. 1977. "Excavations at Carthage: Third interim report." Antiquaries Journal
57: 232-61.
Hurst, H. R. 1979. "Excavations at Carthage: Fourth interim report." Antiquaries Journal
59: 19-49.
Hurst, H. R. 1994. Excavations of Carthage: The British Mission. Vol. II.I, The Circular
Harbour, North Side. London: The British Academy.
Jameson, M. H. 1969. "Excavations at Porto Cheli and vicinity, preliminary report, I:
Halieis 1962-68," Hesperia 38: 311-42.
Jameson, M. H. 1973. "Halieis at Porto Cheli," in Blackman 1973a: 219-31.
Jondet, G. 1916. Les ports submerges de l'ancienne Ile de Pharos. Memoires presentes a
l'Institut Egyptien 9. Cairo: Institut Egyptien.
Karageorghis, V., and D. Michaelides 1995. Cyprus and the sea. Proceedings of the Inter-
national Symposium, Nicosia. Nicosia: University of Cyprus/Cyprus Ports Authority.
Keay, S., M. Millett, and K. Strutt 2004. "Portus Romae: Recent survey work at the ports of
Claudius and Trajan," in Gallina Zevi and Turchetti 2004: 221-32.
Keay, S., and M. Millett 2006. Portus Romae. Rome: British School of Archaeology.
Keay, S., and M. Millett forthcoming. "Portus and Rome's maritime fayade," in Hohlfelder
forthcoming.
Kingsley, S. A. (ed.) 2004. Barbarian seas: Late Rome to Islam. London: Periplus.
Knoblauch, P. 1969. "Neuere Untersuchungen an den Hafen von A.gina," Bonner Jahrbu-
cher 169: 104-16.
Knoblauch, P. 1972. "Die Hafenanlagen der Stadt Agina," Archaiologikon Deltion 27A:
50-85.
Knoblauch, P. 1977. Die Hafenanlagen und die anschliessenden Seemauern von Side. Ankara:
Tiirk Tarih Kurumu Basimevi.
Kolonas, L. 1989-1990. "Anaskaphe Oiniadon," Archaiognosia 6: 153-58.
Laronde, A. 1996. "Apollonia de Cyrenai:que, archeologie et histoire," Journal des Savants
1996.1: 3-49.
Lehmann-Hartleben, K. 1923. Die antiken Hafenlagen des Mittelmeeres: Beitrage zur Ge-
schichte des Stadtebaus im Altertum. Klio Beiheft 14. Leipzig: Dieterich.
Lentini, M. C., and D. J. Blackman forthcoming. "Scavi recenti all'arsenale dell'antica
Naxos," in Blackman and Lentini forthcoming b.
MacDonald, B. R. 1986. "The Diolkos," Journal of Hellenic Studies 106: 191-95.
Maiuri, A. 1958. The Phlegraean Fields. 3rd ed. Rome: Instituto Poligrafico dello Stato.
Marriner, N., and C. Morhange 2007. "Geoscience of ancient Mediterranean harbours,"
Earth-Science Reviews So: 137-94.
McCann, A. M. 1977. "Underwater excavations at the Etruscan port of Populonia," Journal
of Field Archaeology 4: 275-96.
McCann, A. M., et al. 1987. The Roman port and fishery of Gosa. Princeton: Princeton
University Press.
Mccann, A. M., and J. P. Oleson 2004. Deep-water shipwrecks off Skerki Bank: The 1997
survey. Journal of Roman Archaeology Suppl. 58. Portsmouth, RI: JRA.
McKenzie, J. 2003. "Glimpsing Alexandria from archaeological evidence," Journal of Ro-
man Archaeology 16: 35-63.
McKenzie, J. 2006. The Architecture ofAlexandria and Egypt, 300 BC-AD 700. New Haven:
Yale University Press.
Meiggs, R. 1973. Roman Ostia. 2nd ed. Oxford: Oxford University Press.
Milne, G. 1985. The Port of Roman London. London: Batsford.
Miniero, P. forthcoming. "Baia sommersa e Portus Iulius: 11 rilievo con strumentazione
integrata multibeam," in Blackman and Lentini forthcoming b.
Morrison, J. S., J. F. Coates, and N. B. Rankov 2000. The Athenian trireme: The history and
reconstruction of an ancient Greek warship. 2nd ed. Cambridge: Cambridge Univer-
sity Press.
Morton, J. 2001. The role of the physical environment in ancient Greek seafaring. Mnemosyne
Suppl. 213. Leiden: Brill.
Oleson, J.P. 1977. "Underwater survey and excavation in the port of Pyrgi (Santa Severa),
1974," Journal of Field Archaeology 4: 297-308.
Oleson, J. P. 1988. "The technology of Roman harbours," International Journal of Nautical
"The ROMACONS Project: A contribution to the historical and engineering analysis
of hydraulic concrete in Roman maritime structures." International Journal of Nau-
tical Archaeology 33.2: 199-229.
"Reproduction of a Roman maritime structure with Vitruvian hydraulic concrete,"
Papageorgiou, S., M. Arnold, J. Laborel, and S. C. Stiros 1993. "Seismic uplift of the
harbour of ancient Aigeira, Central Greece," International Journal of Nautical Ar-
Pasquinucci, M., and T. Weski (eds.) 2004 Close encounters: Sea- and riverborne trade, ports
and hinterlands, ship construction and navigation in antiquity, the Middle Ages and in
modern time. British Archaeological Reports, Intl. Series S1283. Oxford: BAR.
Pensa, M. 1999. "Moli, fari e pescatori: La tradizione iconografica della citta portuale in eta
romana," Rivista di archeologia 23: 94-130.
Petruso, K., and C. Gabel 1983. "Marea: A Byzantine port on Egypt's northwestern fron-
tier," Archaeology 36.5: 62-63, 76-77.
Piromallo, M. 2004. "Puteoli, porto di Roma," in Gallina Zevi and Turchetti 2004: 267-78.
Poidebard, A. 1939. Un grand port disparu: Tyr. Recherches aeriennes et sous-marines, 1934-
36. Paris: Paul Geuthner.
Poidebard, A., and J. Lauffray. 1951. Sidon: Amenagements antiques du port de Saida. Etude
aerienne, au sol, et sous-marine, 1946-50. Beirut: Ministere des Travaux publics.
Pollini, J. 2005. Terra marique: Studies in art history and marine archaeology in honor of
Anna Marguerite Mccann. Oxford: Oxbow.
Raban, A. 1995. "The heritage of ancient harbour engineering in Cyprus and the Levant," in
Karageorghis and Michaelides 1995: 139-88.
Raban, A. 2003. "Ancient Slipways and Shipsheds on the Israeli Coast of the Mediterra-
nean," In Beltrame 2003: 91-102.
Raban, A., and K. G. Holum 1996. Caesarea Maritima: A retrospective after two millennia.
Leiden: Brill.
Rankov, N. B. forthcoming. "Roman ship sheds and Roman ships," in Hohlfelder,
forthcoming.
Redde, M. 1979. "La representation des phares a l'epoque romaine," Melanges de l'Ecole
Redde, M. 1986. Mare Nostrum. Bibliotheque des Ecoles franyaises d'Athenes et de Rome
260. Rome: Ecole franyaise de Rome.
Rickman, G. E. 1988. "The archaeology and history of Roman ports." International Journal
of Nautical Archaeology 17: 257-67.
Rickman, G. E. 2005. "Portus Romae?" in Pollini 2005: 232-37.
Rothaus, R. 1995. "Lechaion, western port of Corinth: A preliminary archaeology and
history," Oxford Journal of Archaeology 14: 293-306.
Schafer, J., D. J. Blackman, and H. Schlager 1981. Phaselis: Beitriige zur Topographie und
Geschichte der Stadt und ihrer Hafen. Istanbuler Mitteilungen Beiheft 24. Tiibingen:
Wasmuth.
Schlager, H., D. J. Blackman, and J. Schafer 1968. "Der Hafen von Anthedon." Arch-
iiologischer Anzeiger 1968: 21-98; 1969: 229-31.
Schmiedt, G. 1972. Il livello antico del mar Tirreno. Florence: Olschki.
Scranton, R. L., J. W. Shaw, and L. Ibrahim 1978. Kenchreai: Eastern port of Corinth, I:
Topography and architecture. Leiden: Brill.
Shaw, J. W. 1972. "Greek and Roman harbour works," in G. F. Bass (ed.), A history of
seafaring based on underwater archaeology. London: Thames and Hudson, 87-112.
Simossi, A. 1991. "Underwater excavation research in the ancient harbour of Samos:
September-October 1988," International Journal of Nautical Archaeology 20: 281-98.
Sintes, C. forthcoming. "Les neoria d'Apollonia," in Blackman and Lentini forthcoming b.
Testaguzza, 0. 1970. Portus. Rome: Julia.
Tusa, S. 2004. "11 sistema portuale di Mozia: il Kothon," in L. Nigro, Mozia X: rapporto
preliminare della XXII canpagna di scavi. Rome: Missione archeologica a Mozia, 445-64.
Walker, K. G. 2004. Archaic Eretria. London: Routledge.
Williams, H. forthcoming. "The harbours of ancient Lesbos," in P. Betancourt, M. C. Nelson,
and H. Williams (eds.), Krinoi kai limenai: Studies in honor ofJoseph and Maria Shaw.
Philadelphia: Institute for Aegean Prehistory.
Yildirim, B., and M.-H. Gates 2007. "Archaeology in Turkey, 2004-2005," American
Journal of Archaeology m: 275-356.
Yon, M. 2000. "Les hangars du port chypro-phenicien de Kition," Syria 77: 95-116.
Zabehlicky, H. 1999. "Die Grabungen im Hafen von Ephesos 1987-1989," in H. Friesinger
and F. Krinzinger (eds.), 100 Jahre osterreichische Forschungen in Ephesos: Akten des
Symposions, Wien 1995: 479-84. Vienna: Academy of Sciences.
PART VI
TECHNOLOGIES OF
DEATH
CHAPTER 26
GREEK WARFARE AND

FORTIFICATION
PHILIP DE SOUZA
SOURCES OF INFORMATION
The evidence for the study of ancient Greek warfare is not distributed evenly across
all periods. Following the collapse of the Mycenaean civilization in the twelfth
century B.C., there is a lengthy period in which artistic representations of armor
and weapons, relatively common during the Late Bronze Age, become extremely
scarce. The quantity and quality of archaeological finds also diminish rapidly after
1150 B.C., and at the same time the art of writing was lost to the Greeks. Written
documents only begin to emerge from this Dark Age around the middle of the
eighth century B.C. The descriptions of fighting in the Homeric poems, which were
probably composed in Ionia between 750 and 700 B.C., may feature some equip-
ment and practices recalled from earlier times, but for the most part they reflect a
contemporary martial culture (Morris 1986; Snodgrass 1974; van Wees 1992: 5-23).
Figurative art on painted pottery and other materials became widespread in the
latter part of the eighth century B.C., and increasing technical and artistic refine-
ments during the late seventh and early sixth centuries B.C. allowed artists and
sculptors to depict warfare in a more realistic manner than at any time since the
Late Bronze Age.
In the course of the eleventh, tenth, and ninth centuries B.C., iron was gradually
adopted as a metal for making tools, ornaments, and, above all, weapons. Around
this time the iron ore deposits of mainland Greece and the Aegean islands began to
be worked, supplementing supplies from Asia Minor, Cyprus, and sources further
674 TECHNOLOGIES OF DEA TH
away (Healy 1978: 62). It is sometimes assumed that iron was so far superior to
bronze that it was pointless to continue using bronze if iron were available, but
iron, unlike bronze, could not be cast by ancient metalworkers. It was also highly
brittle, and the technique for improving its malleability by a process known as
"carburization" involved a high level of wastage (chapter 4; Healy 1978: 225-36). So,
while iron was quite swiftly adopted as the metal for swords and spears, bronze
continued to be used in the Greek world for armor, especially helmets, and for
some weapons, particularly those that could be cast, such as arrowheads; lead was
also cast for slingshots (Snodgrass 1967: 36-40, 79-81, 84).
HOPLITE WARFARE
Archaeological finds dating from about 720 B.C. onward show that Greek warriors
were adopting new forms of defensive armor. These were the one-piece bronze
helmet, which covered most of the face and head, the bronze cuirass, made in two
or more pieces, which protected the torso, and bronze greaves, covering the legs
from the ankle to the knee at the front, but only as far as the calves at the back.
These accessories were used with a large round shield, nearly a meter in diameter,
made of wood and usually faced with a thin sheet of bronze, and an iron headed
spear, made of ash and typically 2.45 to 2.75 meters in length (Snodgrass 1967: 49-
58; Connolly 1981: 51-63; Sekunda 2000: 9-17). Modern scholars generally refer to
this combination of armor, shield, and spear as the "hoplite panoply." "Hoplite" is
a modern rendering of the Greek word hoplites, a heavily armed soldier, which is
derived from hoplon, meaning a tool or implement, but used in the plural hopla to
refer to the equipment of a soldier. Weapons are tools in the technology of war.
To judge from artistic depictions and widespread archaeological finds, the
most popular form of the helmet initially was the Corinthian, which enclosed the
entire head, apart from narrow openings at the front for the eyes combined with a
very narrow vertical breathing slit. Manufacturing such a helmet required great
skill, as it had to be beaten out from a single sheet of bronze. Many variant designs
were developed, however, which left more space for seeing and hearing, as well as
reducing the overall weight. The convex hoplite shield (aspis in Greek) could also
be extremely heavy. No complete shield has survived from antiquity, but reason-
ably accurate modern reconstructions have been prepared, based on partial rem-
nants and artistic representations. These indicate that a typical wooden shield, with
its bronze facing and bronze central arm-grip (porpax), could weigh as much as 7
kg. The convex shape and broad rim allowed the hop lite to bear some of the weight
on his shoulder, but the porpax was an essential aid. In combination with a smaller
handgrip (antilabe) attached to the inside of the rim, it enabled the hoplite to carry
and maneuver this weight for an extended period (Connolly 1981: 51-54).
GREEK WARFARE AND FORTIFICATION 675
Figure 26.1. Hoplite warriors fighting, Corinthian vase-painting, ca. 600 B.C., Louvre
Museum. (Photograph courtesy of Bridgeman Art Library.)
Modern scholarly accounts of how warriors actually fought using the hoplite
panoply are heavily dependent on the interpretation of various artistic representa-
tions of battle, dating from the mid seventh to the fifth centuries B.C., in combi-
nation with written sources from the seventh, sixth, fifth, and fourth centuries B.C.
(Connolly 1981: 29-63; Hanson 1991, 1995; van Wees 2000). Numerous vase paint-
ings show warriors in combat with the essential items of hop lite equipment, namely
the Corinthian helmet, the large, round shield with porpax and antilabe, and the
long thrusting spear (figure 26.1). Artistic conventions encouraged the portrayal of
warriors in "heroic nudity," so only one figure (on the ground) wears a cuirass, but
they all wear bronze greaves. The sideways stance and overhead thrusting action
with the spear seem to have constituted the usual combat style, judging from the
painted pottery and bronze or lead figurines. Most scholars argue that hoplites
marched and fought in a closely packed formation called a phalanx, which is
described by Thucydides and Xenophon (Hanson 1989, 2000). Descriptions that
are especially important because they were written by men with personal experi-
ence of similar battles are those of Thucydides on the Battle of Mantinea in 418 B.C.
(5.70-75) and Xenophon on the Battle of Coronea in 394 B.C. (Hell. 4.3-17-23) .
Hoplites are sometimes depicted in Greek art wearing or using a double-edged
sword with waisted sides (Greek xiphos), or a single-edged sword with a recurved
blade that broadens out toward the tip (Greek kopis or machaira) . Both were
designed to be most effective in slashing or chopping actions.
By the beginning of the fifth century B.C., hoplites fought in armies numbering
in the hundreds or thousands. They were grouped close together (how close is a
matter of considerable debate) in phalanx formations, typically from four to eight
men deep and hundreds of men long; a much greater depth of formation was
676 TECHNOLOGIES OF DEATH
Figure 26.2. Sword fight between Greek hoplite and Persian warrior, Attic red-figure
kylix by Triptolemos Painter, ca. 460 B.C. Edinburgh, National Museums of Scotland.
(Photograph: Bridgeman Art Library.)
sometimes adopted at the expense of breadth. Often confrontations between two

phalanxes, or groups of phalanxes if the armies comprised several allied contin-
gents, were resolved without a close engagement, because one side withdrew, either
before or soon after their opponents moved forward. If both sides were determined
enough to advance and come to blows, then combat at close quarters could last for
several hours, although there might be periods in which the front ranks drew apart
somewhat to rest, gathering their strength and resolve for further clashes. Hoplites
used their spears to stab at their opponents, relying mainly on their shields and
armor to protect them from such attacks. Swords were secondary weapons, to be
used if the spear was broken. This may be the situation depicted in the vase
painting shown in figure 26.2, on which a Greek hoplite and a Persian archer are
using swords of the kopis or machaira type. The hoplite's bronze helmet is Attic in
style, with hinged cheek pieces and no noseguard. He wears greaves and a cuirass of
stiffened linen. The Persian wears heavy clothing but no armor. The curved tips of
his bow, his principal weapon, are just visible. The most successful hoplite forces
were those that could keep up the pressure on their opponents without yielding. It
may have been their dogged determination not to give ground, as much as their
heavy armor and shields, that enabled Greek hoplites to defeat Persian armies
at Marathon in 490 B.C. and Plataea in 479 B.C. (de Souza 2003: 25-39, 68-79).
Eventually, one side would gain the upper hand, causing their opponents to waver,
retreat, and flee, pursued by the victors, who then expected to inflict more casu-
alties but suffer few themselves. Disintegration of the tight phalanx formation left
the individual soldiers exposed. The extent of the slaughter depended largely on the
duration of the pursuit, unless an enemy formation remained partially intact or
regrouped and was able to confront the pursuers. Casualties were surprisingly low,
perhaps because the majority of men in a phalanx did not actually come face to face
with the enemy. The best modern estimate is an average of 5 percent losses for
victors and 14 percent for the vanquished (Krentz 1985).
Scholars have long held that the introduction of the hoplite panoply neces-
sarily implies the adoption of phalanx tactics in battle. The consensus has been that
the new style of shield and armor were developed because in the late eighth century
Greek warriors fought in a relatively static manner, standing so close to each other
that they lacked the room to maneuver effectively, and they therefore needed greater
protection (Hanson 1991: 63-84; 2000). The painted images referred to above,
alongside references to massed ranks of fighters in the Iliad and Greek lyric poetry,
have been used as evidence to support that assumption. More recently, however,
a strong challenge has been mounted to this consensus (van Wees 2000; 2004: 45-
60 ), proposing that the formations described in the Iliad and lyric poetry, and de-
picted on painted pottery, are not the same as the phalanxes that fought at Plataea,
Mantinea, or Coronea. These early formations were looser, more fluid groupings
of men, some of them heavily armed and armored hop lites, but others more lightly
equipped with javelins and bows. Light-armed soldiers were an ever-present fea-
ture in Greek armies of the archaic, classical and Hellenistic periods, but ancient
Greek artists and writers preferred to concentrate on the heavily-armed hoplites in
their depictions of warfare. The pairs of spears that are regularly carried by hop lites
depicted on seventh and sixth century painted pottery may be a javelin and a lon-
ger, thrusting spear, indicating considerable overlap in the use of weapons between
heavy- and light-armed men. By the fifth century, however, hoplites seem only to
carry one, the thrusting spear. Javelins have become the weapon of the non-hoplite
soldiers, who are sometimes called peltasts, after the pelte, a lightweight, usually
crescent-shaped wicker and leather shield favored by javelin throwers.
During the Peloponnesian War (431-404 B.C.), when troops from many dif-
ferent parts of the Greek world came into conflict with one another, the predom-
inantly light-armed soldiers from northern, central and western Greece had con-
siderable success against hoplites from Athens and Sparta. In 425 B.C., a small force
of elite Spartan hoplites was trapped on the island of Sphacteria off the Pelopon-
nesian coast and driven to surrender by lightly armed Athenians and their allies (de
Souza 2002: 37-42). The historian Thucydides wrote that the other Greeks found
this Spartan capitulation "the most amazing thing that occurred in the whole of the
war" (4.40.1). This attitude of wonder reflects a major reason why less attention is
paid to the non-hoplite soldiers in the ancient sources: their social status and
military prestige were much lower than that of the wealthier, more heavily armed
and armored infantrymen. As a result, images of the latter abound in ancient Greek
art and are closely associated with the Homeric heroes (van Wees 2004: 61-85). In
some cases the distinction was between master and slave, most especially among
the Spartans, whose citizen hoplites were accompanied into battle by large num-
bers of helot attendants, as well as freeborn but subordinate allies from Laconia and
the surrounding regions. The Spartans at Pylos were initially disdainful of the
"spindles" shot and thrown at them on Sphacteria, but in the right circumstances
missile troops could be more effective than hoplites. The lightly equipped troops
are, however, far less visible archaeologically, since they wore little or no armor,
carried shields made of highly perishable materials, and only the killing tip of their
primary weapons was made of a durable material (Snodgrass 1967: 77-85).
The hoplite panoply seems to have been adopted in all parts of the Mediter-
ranean where the Greeks settled, but its relative popularity and importance varied
from place to place. There were far more hoplites among the Peloponnesians than
among the Thessalians, for example. One explanation offered for this variation is
that the hop lite style of warfare was ideally suited to the terrain of Southern Greece,
where the numerous small poleis (city-states) had relatively limited areas of flat
farmland to defend and therefore developed hoplite warfare as a means of settling
their differences (Hanson 1995, 2000). It is also clear that hoplite warfare does not
have to be associated with a particular form of government, whether it is democ-
racy, oligarchy, or tyranny (d'Agostino 1990).
As hoplite warfare became common, so the typical hoplite seems to have be-
come less of a bronze armored elite warrior and more of an ordinary soldier. Al-
though artistic images of hop lites frequently feature a considerable amount of ar-
mor, in practice it seems that many hoplites wore little more than a tunic and a
helmet. Most of the Greek hoplites who, with the aid of many light-armed men,
defeated the Persians at Plataea in 479 B.C. probably wore bronze body armor and
greaves, but cuirasses made of stiffened linen were already common in the fifth
century B.C. Even these were abandoned as hoplites were increasingly required to
fight against lightly armed troops in the late fifth and early fourth centuries, and
began to shed their weighty, cumbersome armor in order to improve their mobility
(van Wees 2004: 195-97). Cost was also a factor in determining how much of the
panoply an individual might use. It has been argued that bronze cuirasses, being
made of relatively thin metal, offered less protection than stiffened linen, but were
more impressive and suggestive of the wearer's social and economic status, de-
liberately modeled on Homer's descriptions of bronze and even gold armor on the
heroes who fought at Troy (van Wees 2004: 52-57). The citizen populations of the
Greek poleis in the classical period comprised many who could not afford such
status symbols, but nonetheless were needed to fight as hoplites in the numerous
wars of the fifth and fourth centuries B.C. As long as they had a spear, a shield, and
a helmet, they were well enough equipped. Conical felt caps (called piloi), worn
underneath bronze helmets to provide a more comfortable and stable fit, provided
the model for lighter, less restrictive helmets that became commonplace in the late
fifth and early fourth centuries (Sekunda 2000: 58-59). Thus it is clear that the
equipment and tactics of ancient Greek warfare were gradually evolving from
around 700 B.C. onward, although none of the particular technical and tactical
refinements can be considered very drastic until the military innovations that were
introduced in Macedon around the middle of the fourth century B.C.
THE MACEDONIAN WAR MACHINE
In the mid-fourth century B.C., after decades of dynastic conflict exacerbated by

outside interference, the kingdom of Macedon was stabilized by a strong, innova-
tive ruler, Philip II (359-338 B.C.). In order to achieve his imperialist ambitions he
created a professional army, at the core of which was a new form of phalanx. In its
fully developed form the Macedonian phalanx under Philip II and his son Alex-
ander the Great consisted oflarge groups of men (1,500 was a standard number in
the time of Alexander) each of whom was armed with a very long spear, or pike,
called a sarissa. The length of these weapons varied considerably but typically was
around 4 to 5 m. The shafts were made of hard wood; the Macedonians favored
cornel wood. The sarissa had an iron head, much narrower than that of the hop lite
spear, and a long pointed iron butt. It seems that sometimes the shaft came in two
sections joined by an iron sleeve, so that it could be taken apart for transportation.
The soldiers stood very close together (no more than about a meter apart), with the
spearheads of the front five ranks projecting ahead of the formation in a fearsome
"hedge" of sharp points. The men in the ten or more ranks behind held their
weapons poised above the heads of those in front, forming a loose but effective
screen to deflect missiles (Connolly 1981: 64-70; Sekunda 1984: 23-28).
Diodorus (16.3.2) claimed that the inspiration for this new tactical formation
came from the description of densely packed phalanxes of warriors in the Homeric
poems (Il. 13.128-35; 16.212-17; Lendon 2005: 121-24). Ancient sources also suggest
that the young Philip learnt to appreciate the possibilities of densely packed for-
mations of heavy infantry while an exile in Thebes (Justinus, Epit. 6.9.7; 7.5.1-3),
where similar ideas had recently been put into practice. Whether or not Philip II
was trying to imitate the fighting style of the Homeric infantry, it is clear that the
sarissa enabled Macedonians to engage their enemies at a much greater distance
than conventional hoplites. It was necessary to use both hands to wield the sarissa,
so the large, heavy shields used by hoplites would have been too cumbersome for
a Macedonian phalanx. Philip II's army may have used peltai, but the men in
Alexander's phalanx, and those of the Successors who carved kingdoms out of
Alexander's empire, carried small round shields with a diameter of about half a
meter, often faced with bronze, silver, or gold. They were supported by a neck strap
but also had the central armband and handgrip that are characteristic of the hop lite
shield, allowing the shield to be maneuvered more effectively in close-quarter com-
bat if the soldier lost or laid aside his sarissa.
It has already been noted that light-armed men with bows, slings, and espe-
cially javelins had been part of all ancient Greek armies throughout the archaic and
classical periods. In the new-style Macedonian army such troops were integral parts
of a carefully coordinated, combined-arms force. They provided crucial flanking
cover to the slow, unwieldy phalanx, often in concert with light cavalry. The greatest
innovation of all in the new-style Macedonian army was, however, the introduc-
tion of heavy cavalry as a striking force. The Greeks did not practice cavalry war-
fare before the classical period. Some of the wealthier hop lites had ridden to battle,
but they fought on foot. The Greeks began to encounter large Near Eastern cavalry
forces when the Persians conquered Ionia and invaded Greece. In the fifth century
B.C., the larger poleis organized small troops of cavalrymen, but they rarely played
a major role in battles. The stirrup was unknown to the ancient Greeks, they did
not shoe their horses, and their saddles were little more than cloths. Horse riding
was reserved for those who had both the time and the resources to develop eques-
trian skills, and it was only in plains of northern Greece, in regions like Thessaly
and Macedonia that horsemen were plentiful (Ducrey 1986: 94-103).
Philip II equipped the aristocratic Macedonian cavalry (known as Companions)
with a long spear called a xyston, shorter than the sarissa used by the Macedonian
phalanx, but still much longer than traditional Greek or Persian cavalry weapons.
It may have been modeled on the spears used by the nomadic peoples of the central
Asian and east European steppes. The Companions wore bronze, Boeotian-style
helmets with flared brims and cuirasses of small metal plates lined with linen and
leather. The overall effect was to make the Companion cavalry into shock troops,
whose charge could be effective against both cavalry and infantry formations
(Connolly 1981: 71-74; Sekunda 1984: 14-23). A further development of cavalry war-
fare in the Hellenistic period was the use of very heavily armored mounted troops,
usually called cataphracts (Greek katakphraktoi), by the Seleucid kings, descen-
dants of Alexander's general Seleucus Nicator (305-281 B.C.). The cataphracts were
usually Iranian horsemen who wore iron helmets and suits of armor made from
iron scales, as did their horses, and fought with long swords and lances the Greeks
called kontoi (Sekunda 1994: 21-22). With the growth of an independent Parthian
kingdom in the third and second centuries B.C., recruitment of these cataphracts
into the Seleucid armies declined, but the Parthian kings deployed them with great
success against the Romans in the first century B.C. (Mielczarek 1993).
In the early Iron Age, before the introduction of the hoplite panoply, some
Greek warriors may have ridden to a field of battle in chariots rather than on
horses. There is no clear evidence that they actually fought from their chariots, just
as the contemporary mounted warriors who rode to battle on horses do not seem
to have fought as cavalry. But, whereas cavalry gradually became standard element
of ancient Greek armies in the classical period, chariots did not, no doubt largely
due to the unsuitability of most of the terrain on which ancient Greeks actually
fought. It was only after the conquests of Alexander the Great opened up vast areas
of Egypt and the Near East that chariots were adopted by the Successors as part of
the huge armies with which they fought for control of the Near East. The favored
type of chariot was the scythed war chariot, often used by the Achaemenid Persian
kings in the fourth century B.C. This weapon was applied with great effect against
inexperienced infantry formations (Xenophon, Hell. 4.1.17-18), but it was of very
limited value against troops and commanders who were not terrified (Arrian, Anab.
3.13; Livy 37.40-41). The Seleucid monarchs frequently included four-horse or six-
horse chariots in their field armies (Sekunda 1994: 26).
The most exotic aspect of the armies of the Hellenistic period, however, was
the widespread use of elephants. The Greeks first encountered them when Alex-
ander led his men into northern India. Although Alexander did not attempt to
make tactical use of elephants he acquired as gifts, his successors deployed them by
the hundreds. Maintaining their numbers was always a problem. Although the
Seleucids attempted to breed them in Syria, they relied principally on eastern allies
to supply them. The Ptolemaic kings of Egypt made use of smaller African ele-
phants, as did the Carthaginians. The principal tactical application of elephants was
against enemy cavalry, but it quickly became apparent that they could be as much
of a liability as an asset. In spite of their training and the presence of mahouts, war
elephants could easily become so frightened and enraged that they would charge
into their own lines and cause havoc. As with scythed chariots, the tactical impact
of elephants was considerably lessened as armies became used to them (Scullard
1974: 64-190; Connolly 1981: 74-75).
Almost all of the varied troop types that contributed to the success of the
Macedonian and Hellenistic armies can be seen as adaptations of existing military
technology. Organizational and socioeconomic differences are of considerable im-
portance in our understanding of what made the Macedonian army so effective,
and what enabled the Hellenistic monarchs to wage war on a much greater scale
than had been the norm in the classical period. The underlying trend in the fourth
century onward can be characterized as increasing participation in warfare among
the lower socioeconomic classes, accompanied by what might be termed the
"professionalization" of Greek warfare. In the late classical and Hellenistic periods
professional, mercenary soldiers were used across the entire Greek world. Greek
mercenaries had sold their services since the seventh century B.C., but it was after
the Peloponnesian War that their numbers increased dramatically. Following the
death of Alexander, many emigrants from Crete or mainland Greece were encour-
aged to settle in the new Successor kingdoms, receiving land in return for their
military service (Cohen 1978; Ducrey 1986: 118-40).
State production and distribution of weapons and armor seems to have been
an important element in Philip H's reforms. A bronze spear-butt stamped with
MAK (for "Mac[edonian]") indicates that it was a state-issued item (Sekunda 1984:
28), and a late third century B.C. inscription from Amphipolis gives a schedule of
fines for Macedonian soldiers who do not have their full set of weapons and armor.
The fines are doubled for officers, whose equipment is expected to include a metal
cuirass, rather than just a linen one (Connolly 1981: 77-80; Sage 1996: 171). Officers
tended to be drawn from the wealthier, more aristocratic families, as had been the
case in the classical period, but they also took their role as military leaders more
seriously than their predecessors, who had prided themselves on their amateurism.
Manuals and pamphlets on how to command and train military forces began to
appear in the fourth century B.C., and the dedicated training of those engaged in
warfare, which had been neglected by most Greeks outside of Sparta, became a
standard polis activity (van Wees 2004: 87-95; Lendon 2005: 91-114).
The Greeks enjoyed a wide range of strenuous sports that helped prepare men
for the grueling physical and psychological demands of ancient warfare. Military
discipline, unit cohesion, and tactical coordination, however, were not skills that
the average Greek citizen trained for. Spartans, who prided themselves on being
trained soldiers, were the exception, along with some elite groups ofhoplites mod-
eled on the Spartans that were formed in Thebes and Argos. To man the Macedo-
nian phalanx, in which independent action was virtually impossible, regular train-
ing as a unit was vital. Ancient written sources indicate that a program of military
training known as the ephebeia was well established across the Greek world by the
third century B.C. The main areas of instruction were close combat, archery, javelin
throwing, and catapult firing. Trainees and mercenaries were often deployed on
guard duty on the city walls and frontier fortifications that had become prominent
features of the Greek landscape (Chaniotis 2005: 46-56).
FORTIFICATIONS AND SIEGE WARFARE
Fortified settlements and strongholds had existed in the Greek world since the
Bronze Age. Many of the Late Bronze Age Mycenaean citadels endured, their mas-
sive walls of polygonal masonry providing places like Athens, Corinth, and Miletus
with a fortified acropolis. There was very little development of such sites through
the Dark Age, except in a few cases where old Bronze Age walls were repaired or
reused. In the archaic period, new fortifications were constructed, but mostly in
different places and for different purposes than their Bronze Age predecessors. The
emerging polis communities wanted protection for their entire urban area, rather
than just a central place. Hence at Miletus the archaic-period fortifications en-
closed a substantial area of land suitable for houses, rather than just the acropolis,
as had been the case in the Bronze Age. By the end of the archaic period, the for-
tifications of the principal settlement on the northern Aegean island ofThasos had
been extended from the small acropolis to enclose most of city, right down to coast.
At Eretria on Euboea a similar extension of the acropolis defenses occurred. In the
numerous overseas settlements that were established, especially in the western Med-
iterranean, larger areas were also enclosed, to protect the settlers from hostile neigh-
bors or pirates. Walling off a peninsula was a preferred option, and was adopted at
Smyrna in Ionia, Taras in Southern Italy, and Naxos in Sicily (Winter 1971: 290-94;
Snodgrass 1986).
These defenses were little more than circuit walls five or six meters high, made
of rubble or brick on a stone socle, that presented a barrier to attacking soldiers.
Towers and bastions were rare, and gateways were often unprotected. Gradually,
however, in the seventh and sixth centuries B.C., as they came into contact with
Near Eastern peoples whose siege warfare techniques were well developed, the
Greeks of Ionia learned how to strengthen their defenses. Miletus, which had strong
walls and easy access to the sea, managed to resist both the Cimmerians and the
Lydian kings (Herodotus 1.6; 1.15-27), but most of the Greeks of Ionia lacked the
manpower to defend extensive walls over long periods, and many of their city areas
had in any case spread beyond the circuit of their walls (Winter 1971: 294-302).
When the Persians occupied Ionia in the second half of the sixth century B.C., they
found that few cities had substantial walls, and even those that did, like Miletus,
failed to keep Persians out. Nevertheless, the resistance that the Milesians put up
at the end of the Ionian Revolt in 494 B.C. suggests that they had worked hard
on improving their defenses; Herodotus (6.18) says the Persians needed to employ
both siege engines and mines.
In mainland Greece there were fewer incentives to improve defensive fortifi-
cations. It is clear that by the late sixth century B.C. Athens had a circuit of city
walls, because two passages refer to it in the immediate aftermath of the Persian
sack of 480 B.C. (Thucydides 1.89.3, 93.2). The city had expanded so much by 480
B.C., however, that defending the walls was not considered a worthwhile act
(Winter 1971: 62-64). In 478 B.C., following the expulsion of the Persians from the
Greek mainland, the Athenians not only rebuilt but also extended their walls
(Thucydides 1.93.2). According to Thucydides, Themistocles had a vision of Athens
"attached to the sea" and projecting her power by means of a navy. A large fleet
operating out of a strong harbor at Piraeus was essential to this vision, as it would
enable the Athenians to defy both the armies of their Greek neighbors and those of
the Persian king. The Athenians began to fortify the Piraeus peninsula in 493 B.C.,
at the instigation of Themistocles, but the project was unfinished in 480 B.C. The
rivalry between the major poleis of Greece proved a greater incentive than the
Persian threat for the construction of defensive fortifications (Winter 1971: 303-10 ).
The Athenians built the Long Walls to link their city directly to the Piraeus. They
were begun around 465 B.C. and completed in 446 B.C. The Northern Long Wall,
roughly 6 km in length, ran from the southwestern quadrant of the city walls to the
northeastern circuit of the Piraeus walls. A slightly shorter wall ran to the settle-
ment at Phaleron, effectively closing off landward access to the whole of the Bay of
Phaleron, but still leaving its shores exposed to potential invaders. In 446 B.C., a
Middle Wall was constructed, parallel to the North Wall and only a short distance
south of it, creating a narrow corridor between Piraeus and the city (Garland 1987:
14-26). The land within these walls was used for farming and residence during
periods when the countryside around Athens was vulnerable to enemy raids. The
Athenians provided similar fortifications for Megara, which also lay a short dis-
tance from its seaport of Nisaea around 460 B.C. These walls made it harder for
Corinth to control her neighbor Megara, and in turn encouraged the Corinthians
to build long walls linking their city to the harbor of Lechaeum, a few kilometers to
the north (Rothaus 1995). The proliferation in city walls forced the Greeks to be-
come more adept at sieges and assaults. In the fifth century B.C., the Athenians
achieved a reputation for capturing cities and were credited with introducing
the defensive tortoise and the battering ram to Greek warfare (Thucydides 1.102;
Diodorus 12.28). These devices were already centuries old in the Near East, and it is
clear that, in technical terms, defense seems to have had the upper hand over attack
in the Greek world until the latter part of the fourth century B.C. Although the
length of some ancient Greek defensive walls seems excessive, such as those encirc-
ling Messene in the Peloponnese, or the 6.5 km protecting the tiny city ofHeracleia
under Latmos in Anatolia, it is important to realize that they followed steep ridges
wherever possible, making them hard to approach and especially difficult for any
form of machines to be brought to bear on them (Winter 1971: 101-25).
The small city of Lato, situated on two slopes adjacent to the main ancient
route between eastern and central Crete, is a fine example of how to make the most
of both the natural setting and effective design in fortifications. Although Lato was
an independent polis in the archaic and classical periods, the defenses that survive
date from the mid-fourth to early second centuries B.C. A securely fortified urban
center was essential during the Hellenistic period, when Lato was involved in
the violent struggles for supremacy among the Cretan cities (Chaniotis 2005: 9-13.
Lato's main entrance was a narrow gateway less than 2 m wide (figure 26.3), which
led into a small, high-walled courtyard that could be closed off in the middle as well
as at both ends. If an attacking force managed to get through this gateway complex,
it had to turn sharply to the right and proceed up a steeply sloping main street,
whose 90 steps were partly built and partly cut out of the rock. Here the attackers
were faced by more strong fortifications. The walls of the houses on either side rise
in a series of terraces constructed of thick polygonal masonry that acts like a second
set of city walls, securing the houses and public buildings above and behind them.
These walls are reinforced by several small tower rooms that help to protect their
narrow entrances. On the right-hand (western) side, the rear walls are further
protected by a rampart wall of irregular limestone blocks. It is likely that the flat
roofs of compacted mud afforded good javelin, sling, or archery platforms. Because
Lato was built on hills that lack natural springs, its inhabitants relied on cisterns to
collect rainwater, which was not readily available to anyone attempting to invest
the city. These relatively simple devices were designed to make Lato difficult to at-
tack and easy to defend; there is no record of it ever having fallen to an assault.
Another significant feature of Lato's position is that it is some 8 km inland but
commands an excellent view over the Bay of Mirabello and the port of Lato pros
Kamara (modern Aghios Nikolaos). This coastal settlement became the main
urban center of the polis of Lato in the second century B.C. The physical move-
ment of the population center reflects a profound shift in local attitudes. In the
second century B.C., the people of Lato seem to have accorded greater importance
to the access to seaborne commerce that their excellent harbor could provide than
to the rugged defenses of their hilltop city (Picard 1992).
Figure 26.3. Remains of the fortified entrance courtyard at Lato, mid-fourth to early
second centuries B.C. (Photograph by D. de Souza.)
In response to improved assault techniques, more sophisticated defensive el-

ements became standard in fortifications. Gateways were strengthened by creating
corridor approaches between protective towers. Walls across open land were given
regular projecting towers and sally ports from which the defenders could coun-
terattack. Stone superseded mudbrick and rubble as the preferred building ma-
terial. Aesthetically pleasing polygonal and irregular masonry styles were largely
replaced by regular ashlar blocks that were more resistant to battering (Winter 1971:
77-91). From the middle of the fourth century B.C., however, the pace of change
quickened due to the emergence of professional armies under first the Macedonian
kings Philip and Alexander and later the rulers of the Hellenistic Successor king-
doms. These monarchs had at their disposal the resources necessary to maintain
and deploy effective siege trains, with experts who could construct and operate
numerous powerful, stone-throwing torsion catapults and mobile siege towers
with battering rams (Lawrence 1979: 39-66). They were skilled engineers as well,
capable of supervising complex operations to undermine, batter, or drill walls to
create breaches for assault. Ditches and moats were added in front of walls to make
mining more challenging and to keep catapults back from the walls, but they could
be quite easily overcome by filling and bridging (Winter 1982). At Syracuse in 213
B.C., Archimedes had the existing walls pierced at lower points to enable the mis-
siles to be fired at attackers who were very close to the walls (Polybius 8.5.6; Livy
24.34.9), but such responses were rare. For the majority of Greek cities it was be-
coming too costly in men and money to defend long stretches of wall against the
great concentrations of force available to the Hellenistic monarchs (McNicoll 1986).
The balance of power often lay with the attackers, provided they had sufficient
resources of men and materials to see the job through (de Souza 2007b: 447-60).
A further aspect of many ancient Greek fortifications was their mutual visibil-
ity, which allowed signals to be passed between strong-points and enabled them to
function as part of an integrated military system. The Athenians created such a de-
fensive system as a direct consequence of their experiences during the Pelopon-
nesian War, when the Spartans and their allies repeatedly invaded Attica and even-
tually occupied a fortified position of their own at Decelea, from which they raided
the countryside, severely disrupting agriculture and bringing silver production
in the Laurion region to a standstill. The Periclean strategy adopted at the outbreak
of the Peloponnesian War relied on the strength of Athens' city walls and the Long
Walls, but required the abandonment of much of the countryside of Attica. In the
fourth century the aim was to keep enemy forces out of Attica altogether (Ober
1985). The fortress at Eleutherae (modem Gyphtokastro) guarded a major moun-
tain pass into Attica from Boeotia (figure 26.4). By the late fourth century BC this
stronghold was a formidable barrier to the progress of invading forces, allowing
time to summon help from Athens via a series of watchtowers and relay stations
(Winter 1971: 43-45; Ober 1985: 191-207). Similar sophisticated and wide-ranging
defensive systems were developed elsewhere, like the chain of fortified settlements
and forts, linked by watch and signal towers and covering all the approaches to
Thebes south of Lake Copais (Fossey 1992).
NAVAL WARFARE
Warfare at sea in the Bronze Age Mediterranean was a maritime version of land
fighting, but the addition of a ram to the prows of oared ships, which occurred
around 900 B.C., turned the ships themselves into weapons (cf. chapter 24). Grad-
ually more oarsmen were added to make vessels faster (a major advantage in ram-
ming) and raised decks were incorporated to act as fighting platforms when ships
were in close proximity. The Phoenicians had added a second level of oarsmen by
700 B.C., and in the mid-sixth century B.C. warships incorporated a third level of
oarsmen, rowing through an oar box or outrigger, perhaps again as the result of
Phoenician innovation (figure 24.4). Three levels of oarsmen proved the maximum
that could be accommodated by ancient oared warships, although an increase in
the number of men pulling an individual oar permitted the construction of larger
and more robust vessels (Casson 1996: 49-65, 77-123). The most common Greek
warship type in the classical period was the "three," with one man per oar, usually
in the form of the trireme, the standard vessel in the celebrated fleets of fifth-
century Athens. The Rhodians preferred a sleeker version of the "three" called a
trihemiolia, with two levels of oars operating through the outrigger (Morrison
Figure 26.4. Walls of the Athenian fortress at Eleutherae, ca. 450-late fourth century B.C.
(Photograph by J. P. Oleson.)
1996: 319-21). The "four" (quadrireme) was similar in size to the trireme but had
oars on just two levels, each pulled by two men. The "five" (quinquereme) usually
had oars on three levels, with two men pulling the top two. This arrangement was
favored by the Seleucid and Ptolemaic kingdoms, which could command large
numbers of oarsmen and marines. A few very big warships were built in the late
fourth and early third centuries B.C. They were powered by hundreds or even
thousands of oarsmen, but these "polyremes" were of only limited use in sea
battles, as they were slow and cumbersome. It may well be that they were designed
to act as artillery platforms and assault ships in attacks on coastal cities and har-
bors, although their principal achievement was to enhance the prestige of individ-
ual monarchs (de Souza 2007a).
Naval warfare in the classical and Hellenistic periods usually involved ram-
ming an enemy ship with the intention of holing it, or of grappling and then
boarding. Many rams were broken off in combat, or taken from captured ships and
displayed as trophies. So far, however, archaeologists have recovered only a single
large bronze ram, at Athlit, on the coast of Israel. It weighs 462 kg, measures just
over two meters long, and is blunt-ended with fins that would encourage splitting
of the timbers of a ship's hull. It is clearly intended to penetrate to a short distance,
presumably to reduce the risk of entanglement with its victim (Casson and Steffy
1991). The use of ramming without grappling or boarding as a fleet's primary tac-
tic would have required relatively fast ships, which in turn implies well trained
crews. The best mode of attack was from the rear or at an acute angle, which nor-
mally meant outmaneuvering an enemy ship in open water. The ship's captain and
helmsman needed skill and judgment to time the attack. The ramming ship aimed
to withdraw and attack another ship while its first victim flooded and sank, al-
though the lack of heavy ballast in warships usually meant that when holed they
would often just heel over, or continue to float with their decks awash. A variation
on actual ramming was to use the stem post, the ram, and the projecting beams
(Greek epotides) of the outrigger at a ship's prow to smash an enemy ship's oars
and outrigger. Warship prows were designed to withstand the impact of ramming,
and the mortise and tenon construction, along with heavy wales, allowed much of
the shock to be transmitted along the length of the vessel. Less speed and maneu-
vering skill were needed to ram a vessel head on, or square on, but to do either
risked serious damage to the ramming ship. Fleets whose captains were not con-
fident in the speed and agility of their ships would prefer to ram prow to prow or
obliquely, and then, as the opposing ships came together, the marines they carried
would assault the enemy with arrows and javelins, often fired or thrown from a
crouching position for better stability and then try to board their ship, a daunting
challenge even in calm seas. One way to make it easier was to use grappling hooks
to secure the enemy ship at close quarters. These devices, called cheires siderai (iron
hands) by the Greeks, were routinely used when ships were operating in confined
spaces such as straits or harbors (Casson 1996: 120-22).
Other types of missile weapons were also used to clear enemy soldiers and
sailors from the decks of a ship. The earliest attested use of torsion catapults, hurl-
ing either bolts or stones, comes in the description of the naval battle at Salamis
in 306 B.C., between the forces of Ptolemy I Soter and Demetrius Poliorcetes.
Diodorus Siculus (20.49.4) says that the latter put stone-throwing catapults on his
ships, along with arrow-shooting catapults on their prows. Since he had recently
been assaulting the Cypriot city of Salamis, it is quite likely that these artillery
weapons came from his siege train. They would have been most effective against the
fighting personnel of an enemy ship and any artillery it may have had, when the
warships were some distance from each other (de Souza, 2007b: 441-43). In a naval
battle in 184 B.C., the exiled Carthaginian general Hannibal, commanding of the fleet
of king Prusias of Bithynian fleet, used his catapults to hurl pots filled with poison-
ous snakes onto the decks of the warships of Eumenes II of Pergamon (Frontinus,
Strat. 4.7.10-11).
The tactical, technical, and engineering achievements of the ancient Greeks in the
sphere of warfare were impressive, but not outstanding. They did develop some
distinctive and highly effective weapons, or weapons systems, particularly the hop-
lite and the Macedonian phalanxes, but neither of these was a radical departure
from what had gone before, nor was either so innovative that it completely changed
the nature of warfare. It could be argued that there was very little technological
change in ancient Greek warfare from the eighth to the first century B.C. Armies
still consisted of men who fought with spear, sword, bow, and sling, some on horse-
back but the majority on foot, as they had done in the preceding centuries, and as
they would continue to do for many centuries to come. Greek achievements in
fortifications and siege warfare were similarly modest. In the fourth and third cen-
turies B.C., some very elaborate walls and defensive systems were constructed, but
the introduction of siege towers and torsion catapults cannot be compared to the
adoption of gunpowder-fired weapons in the fourteenth century, which forced
wholesale changes in the engineering and architecture of defensive fortifications
(Black 2002: 69-96). Even such a celebrated naval achievement as the trireme
(possibly not a Greek invention anyway) was still just another rowed ship with a
fighting a deck and a ram. Innovations like the mounting of torsion catapults and
the increase in the size of the ships did not fundamentally change the way that naval
warfare was conducted, as was the case when cannons were deployed on warships in
the sixteenth century A.D. (de Souza 2001: 20-23, 97-118). While there can be no
denying the lasting fame achieved by military leaders like Themistocles and Alex-
ander, ancient Greek achievements in the techniques and tools of war between the
eighth to the first century B.C. cannot be ranked as highly on a global scale as those
in literature, architecture, philosophy, or science.
REFERENCES
Black, J. 2002. European warfare, 1494-1660. London: Routledge.

Casson, L. 1996. Ships and seamanship in the ancient world. Rev. ed. Princeton: Princeton
University Press.
Casson, L., and R. J. Steffy (eds.) 1991. The Athlit Ram. Texas A & M University Press.
Chaniotis, A. 2005. War in the Hellenistic world: A social and cultural history. Oxford:
Blackwell.
Connolly, P. 1981. Greece and Rome at war. London: Macmillan Phoebus.
Cohen, G. M. 1978. The Seleucid colonies: Studies in founding, administration and organi-
sation. Historia Einzelschriften 30. Wiesbaden: Steiner.
d'Agostino, B. 1990. "Military organisation and social structure in archaic Etruria," in
0. Murray and S. Price (eds.), The Greek city from Homer to Alexander. Oxford:
Oxford University Press, 59-84.
de Souza, P. 2001. Seafaring and civilization: Maritime perspectives on world history. Lon-
don: Profile Books.
de Souza, P. 2002. The Peloponnesian War, 431-404 B.C. Oxford: Osprey Publishing.
de Souza, P. 2003. The Greek and Persian Wars, 499-386 B.C. Oxford: Osprey Publishing.
de Souza, P. 2007a. "Naval forces," in P. Sabin, H. van Wees, and M. Whitby (eds.), The
Cambridge history of Greek and Roman warfare, vol. 1. Cambridge: Cambridge Uni-
versity Press, 357-67
de Souza, P. 2007b. "Naval battles and siege warfare," in P. Sabin, H. van Wees, and M.
Whitby (eds.), The Cambridge history of Greek and Roman warfare, vol. 1. Cambridge:
Cambridge University Press, 434-60.
Ducrey, P. 1986. Warfare in ancient Greece. New York: Schocken Books.
Fossey, J.M. 1992. "The development of some defensive networks in Eastern Central
Greece during the classical period," in S. van de Maele and J.M. Fossey (eds.),
Fortificationes antiquae. Amsterdam: Gieben.
Garland, R. 1987. The Piraeus. London: Duckwortli.
Hanson, V. D. 1989. The Western way of war: Infantry battle in classical Greece. Oxford:
Hanson, V. D. (ed.) 1991. Hoplites: The classical Greek battle experience. London: Routledge.
Hanson, V. D. 1995. The other Greeks: The family farm and the agrarian roots of Western
civilization. New York: Free Press.
Hanson, V. D. 2000. "Hoplite battle as ancient Greek warfare: When, where and why?" in
H. van Wees (ed.), War and violence in ancient Greece. London: Duckworth, 201-32.
Healy, J. F. 1978. Mining and metallurgy in the Greek and Roman world. London: Thames
and Hudson.
Krentz, P. 1985. "Casualties in hoplite battles." Greek, Roman and Byzantine Studies 26:
13-20.
Lawrence, A. W. 1979. Greek aims in fortification. Oxford: Oxford University Press.
Lendon, J.E. 2005. Soldiers and ghosts: A history of battle in classical antiquity. New Haven:
Yale University Press.
McNicoll, A. 1986. "Developments in techniques of siegecraft and fortification in tlie Greek
World ca. 400-100 B. C.," in P. Leriche and H. Treziny (eds.), La fortification dans
l'histoire du monde grec. Paris: CNRS, 305-13.
Mielczarek, M. 1993. Cataphracti and Clibanarii: Studies in the heavy armoured cavalry of
the ancient world. Lodz: Oficyna Naukowa.
Morris, I. 1986. "The use and abuse of Homer," Classical Antiquity 5: 81-138.
Morisson, J. S. 1996. Greek and Roman oared warships. Oxford: Oxbow Books.
Ober, J. 1985. Fortress Attica: Defense of the Athenian land frontier, 404-322 B.C. Mnemosyne
Suppl. 84. Leiden: Brill.
Picard, 0. 1992. "Lato," in J. W. Myers, E. E. Myers, and G. Cadogan (eds.), The aerial atlas
of ancient Crete. University of California Press, 154-59.
Rotliaus, R. 1995. "Lechaion, western port of Corinth: A preliminary archaeology and
history." Oxford Journal of Archaeology 14: 293-306.
Sage, M. M. 1996. Warfare in ancient Greece. A sourcebook. London: Routledge.
Scullard, H. H. 1974. The elephant in the Greek and Roman world. London: Thames and
Hudson.
Sekunda, N. 1984. The army of Alexander the Great. Oxford: Osprey Publishing.
Sekunda, N. 1994. Seleucid and Ptolemaic reformed armies, 168-145 B.C. Vol. 1: The Seleucid
army under Antiochus N Epiphanes. Stockport: Montvert Publications.
Sekunda, N. 2000. Greek Hoplite, 480-323 B.C. Oxford: Osprey Publishing.
Snodgrass, A. M. 1967. Arms and armour of the Greeks. London: Thames and Hudson.
Snodgrass, A. M. 1974. "An historical Homeric Society?" Journal of Hellenic Studies 94:
114-25.
Snodgrass, A. M. 1986. "The historical significance of fortification in archaic Greece," in
P. Leriche and H. Treziny (eds.), La fortification dans l'histoire du monde grec. Paris:
CNRS, 125-31.
van Wees, H. 1992. Status warriors: War, violence and society in Homer and history. Am-
sterdam: Gieben.
van Wees, H. (ed.) 2000. War and violence in ancient Greece. London: Duckwortli.
van Wees, H. 2004. Greek warfare: Myths and realities. London: Duckworth.
Winter, F. E. 1971. Greek fortifications. Toronto: University of Toronto Press.
Winter, F. E. 1982. "A summary of recent work on Greek fortifications in Greece and Asia
Minor," in P. Leriche and H. Treziny (eds.), La Fortification dans l'histoire du monde
grec. Paris: CNRS, 23-29.
CHAPTER 27
ROMAN WARFARE AND

FORTIFICATION
GWYN DAVIES
WARFARE AND THE ROMANS
A message relayed to the Roman people by Romulus after his translation to the
heavens, stands as an unambiguous endorsement of Roman military prowess. "Tell
the Romans that it is the gods' will that my Rome shall be the capital of the world;
therefore let them cultivate the arts of war and let them know and teach their chil-
dren that no human force can resist Roman arms" (Livy 1.16.7). For Livy, the equa-
tion of Roman greatness with martial accomplishment was a truism that stood
as a simple and unassailable explanation for the rise of the city to its pinnacle of
imperial dominion. As the gods themselves had mandated, Roman success, the
well-being of the state and its continuing expansion, were inextricably linked with
the maintenance of military supremacy.
This glorification of war and its elevation as the primary causal factor in the
construction of Roman hegemony was not just the view of a leading propagandist
of the new Augustan age. Instead, a belief in the efficacy of warfare as both a tool of
state policy and a worthy undertaking in its own right permeated Roman society.
For the elite, political reputations were consolidated (or established in the first
place) by demonstrable military achievements, and service in the army formed an
important component of the successful cursus honorum. Further down the social
scale, the status of ordinary soldiers was reflected in their receipt of a respectable
income enhanced by occasional donatives and (for the imperial period at least) the
prospect of significant "retirement" benefits at discharge. Additionally, the steady
transformation of the army into a professional standing force (substantially com-

pleted by the start of the first century A.D.) meant that the state enjoyed access to
trained personnel sufficiently versatile to assist the process of civil governance over
a broad spectrum. In such a climate, the virtues of training, discipline, and com-
petence were increasingly viewed as essential prerequisites for the army, both on
campaign and in the performance of its domestic duties. As a result, it is not sur-
prising to note that for the late Roman historian Vegetius "we see no other expla-
nation of the conquest of the world by the Roman people than their drill-at-arms,
camp discipline, and military expertise" (Mil. 1.1).
A considerable body of scholarship has been devoted to explaining the basis of
Roman military power and to elucidating the technical aspects of its development
and utilization, and it is not necessary to recapitulate these discussions here. Rather,
this chapter will focus on the role of technology (and the complementary dimen-
sion of engineering competence) in improving the operational capabilities of the
Roman state, allowing for the projection of power on a systematic and sustained
basis. To this end, we will trace the organizational and weapon system develop-
ments that enabled Roman armies to engage their enemies with confidence in the
field, alongside the evolution of fortification schemes that enabled economies of
force essential to imperial security.
THE ARMY IN THE FIELD:

ORGANIZATION, WEAPONS,
AND TACTICS
It is not unrealistic to claim that the Roman army was the most proficient military
force of the ancient world. The successes that it enjoyed over several centuries of
campaigning against a diversity of enemies in often difficult terrain and harsh en-
vironmental conditions oblige us to acknowledge the resilience of both individual
soldiers and the system that sustained them in the field. This is particularly true
when we recall that the general trajectory of battlefield superiority was not seriously
checked even by such calamitous reverses as Cannae, Carrhae, Kalkriese (Teuto-
berg Forest), and Edessa that involved the complete loss of entire armies and even
the capture (in the last instance) of the emperor himself (Valerian). This resilience
stemmed in part from the inherent flexibility of the Roman army and its capacity
to adapt to new challenges, whether in terms of the adoption of revised formations
and weapon systems or the wholesale reorientation of its strategic posture to con-
front changing threats along the frontiers. As tempting as it might seem to ascribe a
monolithic character to the Roman army and its institutions-a habit fostered by
ROMAN WARFARE AND FORTIFICATION 693
unhelpful modern references to the existence of a "Roman military machine"-it

is increasingly apparent that an essential factor in the long-lived success of the
Roman state was the capacity of its armed forces to evolve in the face of new
threats.
ARMY STRUCTURE AND ORGANIZATION
The centrality of the citizen army to the early Republic is made clear by the mea-
sures that formed part of the reforms generally attributed to Servius Tullius but
whose introduction probably spanned a much wider period of time. A key com-
ponent here was the division of society into different voting blocs based on gra-
dations of wealth, and both Livy (1.43) and Dionysius of Halicarnassus (Ant. Rom.
4.19) make an explicit connection between the citizen's capacity to arm himself for
battle and his place in the new social framework. Although considerable doubt has
been cast on this particular equation of wealth, fighting potential, and political
representation (Cornell 1995: 186-90), it is clear that the original impetus behind
some of these constitutional measures lay in the more efficient mobilization of the
state's military resources.
The expansion of the Roman state allowed for the steady increase in the size of
the military forces at its disposal, and by the time of the First Punic War it was not
unusual to have armies up to 40,000 strong operating at the end of lengthy lines
of communication. This capacity to campaign at long range was instrumental for
the projection of Roman power. During the Second Punic War, large-scale en-
terprises were sustained simultaneously in Spain, Italy, and Greece, a remarkable
multi-front effort that testifies to Roman logistical and organizational capabilities.
These demonstrations of Rome's developing strategic reach were also reinforced
on the tactical level by the posture adopted for pitched battles. The Roman legion
as described by Polybius (6.19-24) was composed of 30 maniples of paired cen-
turies with a mix of heavy and light infantry. Each of the two front line formations
(in the standard three lines of battle) had ten maniples each made up of about 120
heavy infantry and 40 skirmishers, while the rearmost line of battle had ten smaller
maniples of 60 heavy infantry and 40 skirmishers. These lines were drawn up on
the battlefield in open order, that is, each maniple was separated from its neighbors
by a gap equivalent in width to the frontage of another maniple. The gaps in the
line of battle were covered by the troops of the next line to the rear. The difference
in manipular strength is explicable by the different tactical roles that each of these
formations was intended to fulfill. The two front line formations were intended
to engage the enemy forces directly, and if all went well the third line was not called
into action. However, if matters proved less than successful for the troops of
the first two lines, then the open order enabled these units to disengage and to fall
back through the gaps and seek cover to reorganize behind the solid screen of the
third line. Although the heavy infantry component of the maniples of the third line
amounted to only half the strength of those of the other formations, these troops
were exclusively comprised of veteran soldiers who might be relied upon to keep
steady even in the face of a pressing enemy attack. The advantage of this system lay
in its flexibility. A Roman army could in this way maintain a strong front line,
with the luxury of a secure backstop provision at a reasonable economy of force.
Furthermore, the provision of gaps in the formation allowed for the forward (and
backward) movement of units without disrupting the entire battle line. Of course,
crucial to the success of such maneuvers was the exercise of competent command
and control and a high degree of unit discipline and training, allowing these re-
deployments to be effected swiftly and efficiently. As a result of the reforms carried
out by Marius at the end of the second century B.C., the professionalism of the
army was dramatically enhanced as the legionary ranks were opened to the landless
poor whose only qualification was the possession of Roman citizenship (Keppie
1984: 61-63). This measure, linked to the introduction of much better pay, meant
that the army for the first time became an attractive career choice, and more citi-
zens served longer terms in the ranks, allowing for continuity in useful military
skills.
By the late Republic, this manipular system had undergone further revision,
and the cohort now became the key tactical unit on the battlefield. The cohort com-
prised three of the old maniples (one being drawn from each of the three lines of
battle) but with the crucial difference that the skirmishers formerly attached to
each maniple were now transformed into heavy infantry as well. The cohort was
divided into six centuries of 80 men each, giving it a field strength on paper of 480
men (the standard establishment by the early imperial period). Each legion com-
prised ten of these cohorts. Sometime in the first century A.D. (possibly as early as
the reign of the practical soldier-emperor Tiberius), the first cohort of the legion
was nearly doubled in size to improve overall striking power, giving each legion a
combat strength of about 5,200 heavy infantry, supported by a small body of cav-
alry (Hyginus, de Mun. Cast. 3; Goldsworthy 1996: 14-15; Keppie 1984: 174-76).
This absorption of the skirmishers converted the legion into an exclusively
heavy infantry formation and raised the pressing need to recruit additional forces
that could fulfill other specialist combat roles. Although allied states had tradition-
ally supplied a disproportionate element of cavalry strength to the Roman army,
it was now necessary to expand (and regularize) this contribution. Accordingly,
during the Principate of Augustus, the old ad hoc methods of securing allied troops
for short-term campaigns were replaced by the systematic recruitment of these
forces into permanent units of auxilia. These auxiliary forces, organized into cohorts
of infantry or mixed infantry and cavalry, as well as pure alae (wings) of horsemen,
played an essential role in the garrisoning of the provinces and added a useful
dimension to Roman battlefield capability.
WEAPONS AND EQUIPMENT
The Roman army was not shy in changing its tactical dispositions to better con-
front its enemies, and it was equally prepared to adapt its armaments to the same
end. This was particularly true when the Romans encountered equipage that had
proved successful in the hands of their opponents, and several weapon types were
acquired through this process of borrowing. Beyond this talent for improvisation,
the Roman state also developed an impressive capacity for the large-scale manu-
facture of military equipment, complemented by a complex network of suppliers
and producers who furnished the troops with most of their material requirements,
from pack animals to pottery (Bishop and Coulston 2006: 233-52). A brief discus-
sion of the basic panoply of the Roman soldier (figure 27.1) will introduce the tech-
nology of death employed on the battlefield.
Given the emphasis that the Romans placed on the aggressive use of force and
the importance of seeking decisive battle at the earliest opportunity, it is appropriate
to begin with a description of offensive rather than defensive equipment. As a gen-
eral point, it is worth noting from the outset that our evidence derives from a com-
bination of literary, representational, and archaeological sources, and that assump-
tions made as to the developmental sequence of any given type of equipment-or
even its universal adoption-remain speculative.
Spear and Sword

The personal weapon most closely associated with the Roman legionary soldier is
the pilum or throwing spear. This javelin, with its characteristic long iron shank
terminating in a pyramidal head, was a close-range shock weapon that was thrown
at the enemy immediately before the physical collision of the confronting ranks.
The intention was that a barrage of pila would serve to disrupt an opponent's front
line. This result would have been achieved by the specific design of the weapon that
allowed the 'bodkin' head of the javelin to punch through an enemy's shield, while
the narrow iron shank up to 55 cm long (Bishop and Coulston 2006: 52), offered
the possibility of striking the soldier sheltering behind. Even if the opposing soldier
was not disabled by a direct hit on his body, the effect of the pilum crashing
through the shield would have reduced much of its protective value by making the
wielding of the shield much more cumbersome. As a result, the enemy may have
dropped his shield altogether, making him much more vulnerable to sword thrusts
when the lines of battle eventually collided. In effect, the use of the pilum made up
for the relative paucity of missile troops in the Roman army, giving its heavy
infantry a capability of engaging the enemy beyond the thrusting range of oppos-
ing spearmen or the stabbing/slashing arc of swordsmen. Training in the discharge
of these pila was an important element in the daily weapons drill of the Roman
l D
~
E
'I I
I
I
1, !
B I ''
V
0 20 40cm
Figure 27.1. Elements of the Roman panoply: a: an "imperial" helmet of the first cen-
tury A.D.; b: a first-century A.D. legionary shield; c: a late Republican pilum; d: a second-
century A.D. spatha; e: a first-century A.D. gladius Hispaniensis. (Drawn by R. Carmenate;
copyright G. Davies)
legionary, and Vegetius informs us (Mil. 1.14) that overweight javelins were em-
ployed in practice in order that the standard weapons could be hurled with greater
effect in combat.
The antecedents of the pilum remain disputed, and although the earliest ar-
chaeologically recovered examples date from the last years of the third century
B.C., it is uncertain whether the weapon was an adaptation of some foreign pro-
totype or was an original Roman innovation. It seems, however, that the other
main offensive weapon of the Roman infantryman, the sword, was introduced as a
result of direct contact with foreign influences. The Roman identification of the
weapon as the gladius Hispaniensis suggests that its point of origin lay in Spain and
that the Romans became acquainted with the type as a result of their nearly con-
tinuous military involvement with the Celtiberians from the Second Punic War
through the second century B.C. The essential characteristic of this sword was its
dual purpose as a cut and thrust weapon, equally adept in a stabbing or slashing
action (although Vegetius was adamant that military training emphasized the for-
mer over the latter as a more efficient killing stroke; Mil. 1.12). The earliest example
of a recognizable sword of this type comes from Smihel in Slovenia, from a con-
text that has been dated to around 175 B.C. These weapons start off as slightly
waisted blades approximately 65 cm in length, with long tapering points (Connolly

1998: 130), although by the early imperial period the edges of the blade had be-
come straighter and the point shorter, giving a reduced sword length of 50 cm. The
parallel-edged, still shorter "Pompeian" variant, seemingly introduced during the
mid first century A.D., represents the final evolution of this form of multipurpose
short sword, the use of which seems to have continued (from iconographic evi-
dence) down to the end of the second century (Bishop and Coulston 2006: 154-57).
Although some short swords of a notably different type appear to have persisted in
use during the third century, the majority of troops at that time seem to have been
equipped with the spatha, a longer bladed and narrower weapon originally issued
to the cavalry (where a longer reach was essential for down-swinging strokes di-
rected against infantry from horseback). The spathae that have survived in the ar-
chaeological record are less uniform in character than the short swords, with blade
lengths varying considerably from 62 cm to 91 cm, although most group around
approximately 70 cm (Bishop and Coulston 2006: 82, 130; Connolly 1998: 260).
Presumably this degree of variation reflected the wider range of users that these
blades were meant to accommodate. The change to a longer sword type suggests
that the dynamics of dose-order combat had also undergone revision. The empha-
sis on a short, upward stabbing action aimed at gutting an opponent immediately
after the attacker had collided with him shield first, must have been replaced by
a different style of contest. By stressing the continuing importance of using the
point and not the edge of the sword, Vegetius may be implying that the spatha was
employed in stabbing down over the shield top or sideways around the shield, the
aim still being to inflict the enemy with a deep puncture wound rather than a
slashing cut. It is equally possible, however, and perhaps more probable, that the
adoption of the spatha is evidence for the dramatic transition to an open melee
type of combat in which longer edged blades provided a more versatile repertoire
prioritizing the cut over the thrust. Whichever explanation is preferred, it is worth
noting that the spatha was not a finely balanced weapon, as its center of gravity was
located low down the blade and the light pommel offered little in the way of
counterweight (James 2004: 141). This characteristic made precision handling more
difficult, possibly indicating the redundancy of the finer techniques of weapons
drill and thus the abandonment of the regimented dose-order fighting techniques
of the past.
Missiles
Although the pilum and the sword may have formed the primary offensive weap-
ons of the Roman heavy infantry, the troops had recourse to other types of equip-
ment from time to time. Prior to the reforms of the late Republic, the veteran
soldiers who made up the third line of battle employed a heavy thrusting spear in
lieu of a javelin, a weapon well suited to the task of repelling the onslaught of a
successful foe who had broken through the first two Roman lines. In the later
Empire, we can note the appearance of a wide range of new types of thrown mis-
siles, as replacements for the pilum (the rarity of which Vegetius notes with regret,
Mil.1.20). These comprised a diverse set of light javelins and weighted darts or
flechettes (plumbata), five of the latter being slotted into the back of the shield, to be
"thrown at the first assault" according to Vegetius (Mil. 2.15). Reconstructions
derived from examples of plumbata found at Wroxeter proved effective up to a
range of 60 m (Eagle 1989).
Furthermore, legionaries as well as specialist units of auxilia also made use of
other forms of missile weaponry. Since slings, bows, and catapults not only offered
the potential to strike at an enemy from a distance but also delivered their pro-
jectiles with much greater kinetic force, it is not surprising that troops regularly
underwent training in the use of such weapons. The large number of ovoid lead
slingshot recovered from Perugia, many examples of which are impressed with
propaganda messages attributable to both sets of combatants in the siege of 41-40
B.C. (Keppie 1984: 123-25), demonstrates how rapidly a significant quantity of these
projectiles could be manufactured in the field. Archery equipment, arrows, and
related materials are also excavated on a regular basis, suggesting the widespread
use of composite bows in the Roman army. These bows were made of wood, lam-
inated with horn on the inside and glued tendon on the outside to resist com-
pressive pressure and add tensile strength whenever the weapons were drawn and
released (Coulston 1985: 245-59). Unfortunately, no complete example of a Roman
bow has survived, and our knowledge of the exact performance of these weapons
must remain speculative.
Artillery
One very important branch of missile forces was artillery, first manufactured on a
significant scale in the Syracusan workshops ofDionysius I (ca. 399 B.C.; Diodorus
Siculus 14.41.3-42.1), although the Ptolemies sponsored research as well (Philo of
Byzantium, Bel. 50.3). These early weapons, essentially large bow-armed bolt-firers,
were much improved by the invention of torsion-powered engines, seemingly by
the corps of engineers and craftsmen employed by Philip II of Macedon (Marsden
1969). Such machines allowed an assailant to deliver heavier missiles with greater
accuracy at a much longer range than could be offered by conventional bows or
slings. Although Livy (6.9) makes anachronistic references to the use of artillery in
Roman contexts as early as the first years of the fourth century, it is implausible
that the Roman state had access to the appropriate technology until much before
the First Punic War (264-241 B.C.). Thereafter, however, the Romans enthusi-
astically adopted these engines, although mainly employing them in static roles,
either as emplaced weapons (on a city circuit or onboard ships) or as siege bat-
teries. In doing so, they were following established tradition, as, notwithstanding
their rare appearances on the ancient battlefield, artillery pieces remained too un-
wieldy with too slow a rate of fire to be anything more than fixed-base weapons. By
the time of the First Dacian War, however, Roman technical proficiency had not
only given rise to the "most powerful arrow-shooting engines ever produced in the
ancient world" (Marsden 1969: 190 ), but had also endowed their artillery with a
new and lethal mobile dimension.
Making its first appearance on the sculpted reliefs of Trajan's Column (scenes
XL, XLIX, LXVI) is the cart-mounted carroballista. The sinew coils of the spring
mechanism that provided this weapon's torsion power were protected by iron
casings set further apart than in earlier catapultae. This not only provided a better
degree of weatherproofing in the field, but also allowed the artillerymen to sight
their weapon more efficiently. Further important advantages also accrued from the
new all-iron frame: the catapult arm could be retracted further, allowing the bolt to
be discharged with approximately 25 percent more power, and the compact design,
with its standardized metal components, was not only easier to fabricate, but was
also much lighter than its wooden-framed and metal-plated predecessors (Mars-
den 1971: 227-32). It is not surprising, therefore, that the relative mobility and
efficacy of the carroballista made it a useful adjunct to field operations, lending
plausibility to Vegetius' claim (Mil. 2.25) that 55 of these machines comprised the
traditional complement of the legion. Of course, bolt-firers were not the only ar-
tillery employed by the army, as stone projectors of various calibers were also
deployed in some numbers. The projectiles delivered by these latter engines varied
from 0.5 kg grapefruit-sized shot to the large 25 kg missiles delivered by the siege
artillery before Jerusalem in A.D. 70 (Josephus, BJ 5.270). These weapons could
either be variants on the arrow-firing types or single-armed hurling devices (com-
monly identified as onagri during the later empire), ten of which were allocated to
each legion according to Vegetius (Mil. 2.25). Large stone projectors of this nature
were unsuitable for open battlefield conditions and could only be used effectively
from prepared positions.
Shields
The material discussed so far has been largely offensive in nature, but, of course,
the Roman army also relied upon a wide range of defensive equipment to lend
protection to its men. One element of the panoply of the Roman soldier forms a
bridge linking both the means of attack and defense, namely the infantry shield.
Although the curved, rectangular shield (often described in modem literature as a
scutum, notwithstanding that this was the generic term employed for most Roman
shields) may be familiar from its frequent depiction in contemporary reliefs, it is
clear that the range of shield forms actually deployed was quite extensive. One
important point to stress regarding the large infantry shield (whether it was of the
rectangular, oval or even hexagonal type), was that this piece of equipment formed
more than a passive shelter for the individual soldier. The aggressive use of the
heavy infantry shield was an integral part of Roman weapons drill, with the sol-
dier being encouraged to thrust his shield against his opponent's body as the battle
lines collided, making use of the momentum of the charge to gain maximum
advantage.
We are fortunate that a small number of surviving shields (particularly one
from the Egyptian Fayum and others from Dura-Europos in Syria) allows us to
reconstruct the manner in which these items were assembled. The strength of the
Roman infantry shield derived from its laminated character, with three layers of
light wooden strips being glued together with the central layer of the laminate laid
crosswise between the horizontal strips of the front and back layers. The wooden
strips were thicker at the center than at the edges (Bishop and Coulston 2006: 61),
giving added strength at the central boss and greater flexibility at the sides, making
the shield less liable to fracture at the moment of impact. A leather or felt cover
encased the wooden core, held together by rawhide or (later in the first century)
bronze stitching, and the face of the shield could be painted to display the device of
the unit in question. A reevaluation of the Dura material suggests that there was no
supplementary linen cover and that the paint or gesso for the shield decoration was
applied directly to the leather casing (James 2004: 162-63). The back of the shield
had a light wooden frame glued on over the leather which incorporated a central
rib that provided a handgrip at the aperture left for this purpose in the middle of
the board. A protruding metal boss both served to protect the hand of the wielder
and to provide extra force during the clash of arms.
Helmets and Body Armor

More exclusively protective in nature than the shield are two other classes of equip-
ment: the helmet and body armor. Roman helmets owed much of their original
inspiration to Celtic designs. The classic "Montefortino" helmet of the Republican
period, an elongated bronze bowl with riveted or hinged cheekpieces, a distinctive
crest fitting, and a shallow disc neck guard, was derived from north Italian types
introduced by the Gallic invaders of the peninsula (Bishop and Coulston 2006: 65).
The mainstream replacement of this helmet, the late Republican "Coolus" or "joc-
key cap" type, a more ergonomic, flattened (usually) iron bowl with larger cheek-
pieces and a ribbed, downward flaring neck guard, was also adapted from indige-
nous models encountered during the expansion into Celtic Europe beyond the Alps
(Connolly 1998: 230). Although the Montefortino style was retained by the Prae-
torians long after its abandonment by frontline forces, its Coolus replacement was
also phased out during the first century in favor of the "Imperial" helmet. This latest
variant had better neck protection (eventually extending coverage to include the
upper shoulders), built-in brow and ear guards and transversal reinforcing strips
forming a crosspiece across the bowl top. Internal padding or a separate lining cap (as
identified from the Dura assemblage, James 2004: 109) provided further shock ab-
sorption and, perhaps more importantly, improved comfort for the user. By the
fourth century A.D., the one-piece helmet and neck guard seems to have been re-
placed by a simpler casque, often including an integrated nasal guard, comprised of
(at least) two parts, with the join along the bowl crest being protected by a raised
ridge. The cheekpieces and neck guards were now detachable or were replaced al-
together by a camail suspended from the base of the helmet, prefiguring the form of a
medieval basinet.
The earliest body armor employed by the Roman soldier was the mail shirt
worn over an undergarment that combined a fair degree of protection with flex-
ibility of movement. Mail shirts, however, were both expensive to manufacture (al-
though relatively easy to repair) and heavy to wear, a drawback sometimes rem-
edied by the substitution of a skirt of cured leather strips worn in conjunction with
a short mail tunic rather than the full, hip-length shirt. Officers of sufficient means
might have preferred a Greek-style bronze cuirass (often of the muscled variety)
instead of mail, while the poorer citizen recruits might have had to make do with
little more than a simple bronze pectoral plate of doubtful value. Although the date
for its introduction remains uncertain, another possible option in lieu of mail
was the use of scale armor, comprising thin bronze platelets wired together in an
overlapping pattern and sewn onto a fabric undergarment. This so-called lorica
squamata was easier to manufacture than pure mail and was less heavy to wear, but
these advantages may have been discounted by the reduction in movement allowed
by the more rigid sewn strips and a probable diminution in the protection offered,
particularly against the impact of missiles (Bishop and Coulston 2006: 64).
The introduction of segmental plate armor (usually termed lorica segmentata)
during the early first century provided heavy infantry with a new and effective form
of body protection. Two complete sets of this armor, recovered from a second-
century A.D. context at Corbridge, have allowed us to reconstruct the manner in
which these articulating plates were intended to operate. The soldier's torso and
back were protected by a set of overlapping horizontal iron strips or half-hoops
linked to leather strapping by a series of hooks and buckles. The shoulders and
upper arms were encased within a further set of vertically arranged iron plates that
were hinged together and hooked onto the girdle plates of the chest and back. The
whole ensemble combined the advantages of relative freedom of movement with
the capacity to dissipate impact shock over a wider area than mail shirts could pro-
vide. The complex series of fasteners, straps, and buckles that gave the armor its
articulating character, however, also proved its point of weakness, at least insofar as
durability was concerned. The copper alloy elements of these fittings were not only
relatively fragile, but were also subject to corrosion, particularly when in direct
contact with the iron plates (Bishop and Coulston 2006: 98). The evidence we have
for broken fastenings and for the replacement of hinges with rivets on some of the
plates, suggests the degree to which unit fabricae were kept busy with the repair and
maintenance of this equipment.
Although it was formerly believed that the use of segmental armor came to an
end sometime in the third century, it now seems clear that the use of articulating
plate in heavy infantry contexts continued into the fifth century. It is also worth
pointing out that the adoption of heavy cavalry, from modest beginnings in the
second century to more widespread deployment from the third century onward,
also meant that some mounted units were provided with significant armored
protection. The use of such cataphractarii or clibanarii, a concept originally bor-
rowed from the Persian enemy, was more common in the East than the West, and
we have our best evidence for the equipage of such troops from Dura-Europos. The
site not only preserves graffiti depicting armored cavalry, but also produced at least
four sets of equine "barding" covers, or armored coats, two of which were recov-
ered almost intact. These examples of barding demonstrate that the Romans em-
ployed two large sheets of double-thickness linen which acted as backing for
sewn rows of scale armor (both copper alloy and iron scales were employed) that
were attached to one another by a wide leather strip that ran along the horse's spine
(James 2004: 113-14). Presumably these covers were also supplemented by some
form of neck guards and chamfrons (protective headpieces), as have been iden-
tified at other late Roman sites. The cavalryman was also encased in scale armor,
possibly with some form of supplementary thigh guard, and his primary weapon
appears to have been the long, heavy two-handed lance known as the contus (Dixon
and Southern 1992: 49-50 ), a shock weapon that would have served equally well
against other heavy horsemen and heavy infantry.
SIEGEWORK CONSTRUCTION
The preceding discussion has allowed us to explore how Roman tactics and
equipment may have been adapted to retain a technical edge on the battlefield, and
it is now necessary to turn to the related issue of the engineering competence that
also set the Roman army apart from most of its contemporaries. A convenient
demonstration of these skills can be obtained by looking at Roman approaches to
siege warfare.
As siege engines and related technical aids are considered elsewhere in this book
(chapter 13), attention will be focused here on fixed structures built by the army
to assist in the task of reducing stubbornly defended targets. These include works
of circumvallation, assault ramps, siege mounds, mines, and other miscellaneous
engineering projects that were intended to facilitate operational progress (Davies
2006). Diodorus Siculus (13.2.1) informs us that the Romans learned their siegecraft
from the Greeks, but by being "pupils who always outstripped their masters ...
[they] had then forced the cities of their teachers to do their bidding." Certainly,
the Hellenistic world had proved a rich testing ground for the emerging science of
poliorcetics, but the determination, skill, and imagination with which Roman com-
manders waged festungskrieg make them the primary exponents of the art.
A regular part of the Roman training regimen concerned the rapid raising of
earth-and-timber fortifications. Although the primary purpose of such exercises
was to enable rapid preparation of temporary camps for protection in hostile

territory (Welfare and Swan 1995), the process also built up familiarity with field
engineering techniques in general. This was to serve the Roman army well on those
occasions when siege operations were commenced in the course of an ongoing
campaign when the troops at hand were required to undertake the requisite field-
works without recourse to any specialist corps. Indeed, the Roman armies that
produced the most sophisticated and extensive siege works were usually those that
had been in the field for several years and whose personnel had developed a high
degree of proficiency by repeated demands to reduce strongholds in their paths.
Caesar's Gallic campaign and the suppression of the First Jewish Revolt by Ves-
pasian and Titus are cases in point.
For sheer scale of endeavor it is impossible to ignore the works Caesar built
before Alesia in 52 B.C., where the double encirclement, stretching for a combined
distance of 35 km, was designed to hem in Vercingetorix and his 80,000 men while
simultaneously defending the besieging force against the huge relief army massing
beyond their lines. Although Caesar was hard-pressed at times, the elaborate mea-
sures that he put in hand for the protection of his army behind their lines of con-
travallation (facing the besieged) and circumvallation (facing the external threat),
ensured that he had sufficient time to transfer reserves to threatened sectors before
they were overwhelmed. The key to Caesar's success was the provision of im-
pressive obstacle fields that fronted the most threatened sectors, backed up with a
wide turf rampart amply provided with towers for the accommodation of light
artillery. Modern excavation has revealed that these works may not have had the
monolithic character that Caesar's own account imputes to them (BG 7.74), but
they remain a testament to his ingenuity and the industry of his troops. In the most
vulnerable section on the plain to the west of Alesia, the contravallation comprised
a rampart nearly 6 m wide with towers set at 15 m intervals fronted by a ditch up to
3.2 m wide and 1.4 m deep. This ditch, however, was only the last element of the
obstacle field. In front of it was a glacis up to 16 m wide with evidence for two
bedding trenches for cippi (the ancient equivalent of barbed wire, comprising
interleaved sharpened branches) and five rows of stimuli (logs set with iron spikes
buried in prepared pits). Two ditches lay to the east of the glacis, the inner, a V-
profiled feature of a consistent 2.7 m width and up to 1.5 m depth and the outer, up
to 6.5m wide and 1.5m deep, which for some of its length was flooded with water
diverted from a nearby stream. In other sectors the defensive configuration may
not have been so elaborate, but nonetheless could include such features as /ilia
(man-traps with sharpened stakes at the base) or ground liberally studded with
caltrops in lieu of some of these other obstacles.
Schemes of circumvallation were only one component of the siege repertoire
deployed by Roman field engineers. More direct approaches might demand the
investment of time, energy and expertise in the raising of either assault ramps or
siege mounds. The former structures were designed to parallel the height of de-
fensive walls in order to facilitate the passage of storming parties and to mount
engines capable of effecting a breach, while the latter were intended to parallel or
overtop the defenses so that emplaced artillery might sweep the ramparts clear of
defenders and allow escalades or sapping operations to be mounted. The most
sophisticated ramps built by the Romans (that can be termed "Avaricum-style"
structures after the Gallic stronghold assaulted by Caesar) served to combine both
functions. The most famous (and substantially intact) assault ramp is that at
Masada (figure 27.2), a funnel-shaped structure 225 m long, between 50 and 200 m
wide at the base, with a vertical rise of about 75 m. This latter dimension is a little
misleading in the sense that the Roman engineers took advantage of an existing
natural outcrop so that the actual built feature is only 25-30 m thick. Although this
remains to be proven by excavation, it seems likely that the work crews of the
Tenth Legion first cut a series of horizontal steps at regular intervals along each
flank of the bedrock spur. Then a timber box-revetment (using local desert species,
as the surviving elements demonstrate) was installed on each of these platforms
and filled with rammed earth and rubble. The ramp core could then have been
deposited between the stacked boxes of the two retaining revetments.
The ramp has a more or less uniform gradient of 17 degrees but an original
slope of 20 degrees is likely, allowing for 1,900 years of weathering. This relatively
steep profile did not prevent the attackers from deploying a heavy iron-dad tower
and large ram at the summit, engines that presumably were winched up the ramp
in short increments along a prepared track, on each occasion having chocks
rammed in place to prevent slippage. Although much has been made of the use of
Jewish corvee labor at the siege, there is no basis to suppose that they had anything
to do with the process of raising this assault ramp. The Tenth Legion, long ha-
bituated to such endeavors after three years of campaigning, was more than ca-
pable of constructing such a work without assistance and in relatively short order.
Indeed, there is no reason to extend any estimate for the time taken to complete
this difficult technical operation (including the time taken to construct an unnec-
essarily comprehensive circumvallation) beyond five months, if it took that long
(Roth 1995).
Although the use of circumvallation and assault ramps became a rarity in the
later empire, there are still ample demonstrations of the technical proficiency of
Roman soldier-engineers. During the late-third-century A.D. siege of Cremna, for
example, the Roman commander built two lines of containment, well-sited artillery
platforms, and an enormous siege mound to allow his assault parties to approach
the enemy wall (see Mitchell et al. 1995; Davies 2000). At Boulogne, Constantius
Chlorus' men denied the use of the harbor to the defenders by sinking piles into the
seabed, despite the problems caused by the ebb and flow of the tide and the pres-
ence of a hostile flotilla (Pan. Lat. 6.6.2). Finally, in the course of his Persian cam-
paign, the emperor Julian threw up a siege mound at Maiozamalcha while simul-
taneously extending mines beneath the enemy circuit, one of which at least, was
intended as a mine of attack, to allow his troops surreptitious entry into the city
rather than simply undermining the walls (Zosimus 3.22.4)
Figure 27.2. Roman assault ramp at Masada, A.D. 73. (Photograph by G. Davies.)
ROMAN FORTIFICATIONS
Although the Roman army was often on the attack and made use of complex siege
technology, it was also highly skilled in the preparation of defensive fortifications.
The standard Roman "playing-card" fort of the first and second centuries A.D., an
elongated rectangle with rounded corners and a limited number of internally
projecting angle and interval towers, was laid out with only limited provision for
point defense (Gregory 1986, 1996b). It is often claimed that this procedure con-
formed to the Roman preference for meeting an enemy in the field rather than
awaiting an attack from the security of a prepared position, but it may also reflect
the realities of the early empire, when Rome's enemies generally lacked the capacity
to mount a serious threat to even modestly fortified positions. This second ex-
planation is supported by the fact that certain forts sited in more remote and more
vulnerable locations were often provided with a reinforced set of defenses, usually
comprising multiple ditches that extended the obstacle field on their most vul-
nerable flanks. Thus, instead of the conventional two ditch configuration, Whitley
Castle in Northumberland has seven ditches covering the approaches to its weakest
point on the southwest. Birrens in southwest Scotland has six ditches on its north
side, and Ardoch, in the wild country beyond the Antonine Wall, has five to its east.
The intention in all cases was to disrupt the momentum of any attack and to
subject the assailant to more effective missile fire as men bunched together while
they sought to negotiate the obstacle field.
Greater care was lavished on overtly defensive measures during the third
century, with new constructions often combining an irregular plan (to take max-
imum tactical advantage of the terrain) with the provision of many more towers
either partially or fully projecting beyond the wall circuit (Johnson 1983; Lander
1984; Gregory 1996a, 1996b ). The impetus for this latter measure was clearly driven
by the need to provide enfilading fire along the line of the fort wall (a recognition
that an enemy might now be in a position to threaten the same much more di-
rectly), and to provide firing platforms for artillery that would enable defenders to
engage attackers at a considerable range. Such towers were usually provided with
wide embrasures to allow catapults a greater traversing arc or were constructed
with a solid concrete-rubble core that would have absorbed the recoil shock gen-
erated by heavy onagri (notwithstanding Ammianus' caution against emplacing
these engines on stone-built structures, 32,4.5). The so-called Saxon Shore forts of
southern Britain demonstrate many of these innovations: Pevensey has large, solid
U-shaped towers boldly projecting outward, and Cardiff has five-sided solid bas-
tions as interval towers fully bonded into the fort wall, hollow towers at the gate-
ways, and an earth rampart terreplein reinforcing the inner face of the fort wall.
Elsewhere in the empire we see a similar pattern of evolution, with increasing
emphasis being given to defensive capability over time. For example, on the eastern
frontier we can see a process whereby the relatively small number of forts suggested
as having second-century origins, such as Umm el-Quttein and the putative first
phase at Qasr el-Azraq (Kennedy 2004), mostly of rectangular plan with no towers,
are replaced by a much more extensive network of castella usually of fourth-
century date, with enhanced defensive capabilities (Gregory 1996b; Parker 2000).
Atypically, the Trajanic fort at Humayma (ancient Havarra) had a rectangular plan
with projecting square towers (Oleson et al. 2003). The later castella range from the
small garrison posts styled as quadriburgia to full-scale legionary fortresses-even
if the late legion was a much different creature than its early empire predecessor
(Tomlin 2000). Although there are variations in form from site to site, the quad-
riburgium at Yotvata (dated by inscription to A.D. 296/299) can be seen as typical.
This is a broadly square enclosure, measuring about 40 m x 40 m, with four square
angle towers that project about 3 m beyond the 2.5 m wide circuit wall that is built
of stone to a height of nearly 2 m with mud-bricks above (up to seven courses still
survive in situ). A simple entrance was built in the center of the east wall, and a
narrow postern gate is set directly adjacent to the southwest angle tower. The care
taken over these features suggests the new importance ascribed to providing even
a small garrison post like this with a formidable defensive capability. This is even
more apparent at the late legionary sites of el-Lejjun and Udruh in southern Jordan.
These 4.7-hectare fortresses, with numerous U-shaped interval towers projecting
about 11 m beyond the thick stone-faced and rubble mortared walls, and four mas-
sive semicircular fan-shaped corner towers with internal spiral staircases (see Ken-
nedy 2004; Parker 2006), were as defensible as any crusader castle.
We can see a similar pattern of increasing concern with practical protective
devices in the construction of urban fortifications. Wailed cities were nothing new
in Roman contexts: the defenses provided for veteran colonies such as Aosta or
Turin, or for the new Augustan showpieces such as Autun or Frejus, were often
imposing and extensive, incorporating high circuit walls with ample interval
towers (figure 27.3) and particularly fine multiple gateways (Johnson 1983: 17-18).
We should not, however, forget that in the ancient world an impressive enceinte
was considered to lend dignitas to a city and the construction of elaborate urban
fortifications often had more to do with issues of status display rather than any
pressing concern with security (Gros 1992). A dramatic change in attitude can be
seen for the western provinces of the empire in the second half of the third century
when, in the wake of the breakdown of central authority, most open cities in Gaul,
Spain, and the Germanies underwent a rapid program of fortification as the pop-
ulation lost confidence in the illusive claim of an impermeable frontier. These new
urban defenses were often limited in terms of the area enclosed by the walls (fre-
quently following advantageous ground as at Angers or Senlis, or incorporating
existing structures to provide an ersatz fortress, as with the amphitheater at Amiens
or the forum at Bavai; see Johnson 1983: chap. 5) and the impression gained from
the amount of spolia included within the fabric of these walls is of hasty demoli-
tions put in hand to furnish the requisite building materials without delay (figure
27.4). But despite the speed with which many of these urban fortifications were
raised, there can be little doubt that they were provided with sophisticated de-
fensive attributes, particularly in terms of powerful towers with multiple embra-
sures suitable for mounting several light catapults with wide fields of fire. In many
ways, Aurelian's new defenses of Rome itself provided the role model (Todd 1978).
This decade-long building scheme (ca. A.D. 272-282) resulted in the creation of an
18 km circuit for the city, employing favorable terrain wherever possible, with
eighteen major gates and several additional posterns. The wall, about 4 m thick and
surviving up to 20 m in height, must have involved extensive clearance and de-
molition, although several existing structures along its line were pressed into
service where feasible. Thus, the southern arcades of the Amphitheatrum Castrense
were bricked up and its curving profile became the new wall line and the 3 m tall
pyramid-tomb of Gaius Cestius was too monumental a structure to be removed, so
the wall was simply butted up against it.
Aurelian's great achievement at Rome is overshadowed only by the tremen-
dous complex that we know as the Theodosian Land Walls of Constantinople,
which formed the most powerful urban fortifications of antiquity. Here, a 20 m
wide moat formed the first line of defense, the water being retained by a series of
dams. A low outwork and patrol track provided direct oversight of the moat and
was backed by a proteichisma studded with towers. The final element was the inner
wall, a stone and coursed-brick construction, 5 m wide and 12 m high, that was
strengthened by 96 towers 20 m in height, offset from the towers on the second wall
in order that the fire of their many engines might not be masked. This remarkable
ensemble of works served the New Rome well: even during the final siege of 1453, it
was human error and an open postern that brought about the fall of the city, rather
than any failure of the thousand-year-old defenses.
There can be little doubt that the centrality of the army to social and political life
serves to distinguish the Romans from their contemporary neighbors. With the
Figure 27.3. Urban defenses of Augustan Autun. (Photograph by G. Davies.)
possible exception of Macedonia in the reign of Philip II, no other state of the
classical world laid such an emphasis on the role of the army either in the per-
formance of its traditional security mission or in the management of the fabric of
empire. Army units transcended their direct combat functions by assisting civilian
administrators in a plethora of duties, ranging from the provision of specialist
engineering and construction assistance, to the assignment of soldiers to mundane
clerical and support positions on the staffs of provincial governors. The Vindo-
landa tablets suggest something of the scale of this commitment, with one strength
report (Tab. Vindol. 1.154) revealing that over 60 percent of the fort garrison was
out-posted on detached duty even though the north British frontier must still have
been regarded as an active operational theater (Bowman 1994).
To a degree, this participation in the civilian realm served to offset the burden
of maintaining the army in its provincial deployments, but nonetheless, the cost to
the state remained substantial. Even though the early imperial field strength of some
30 legions (supplemented by auxiliaries in equal numbers) represented a remark-
able economy of force, supplying the dispersed network of garrisons in the frontier
zones with the necessary resources involved sustained investment. The distribution
of wages to the troops caused a significant outflow of bullion from the core to the
periphery, only partially offset by the exploitation of new precious metal sources.
In fact, it is likely that many coin issues were struck for the purpose of paying troops
(see chapter 30, part 2). This transfer of resources had the side effect of stimulat-
ing local spending in the areas where the soldiers were based. Even retired soldiers
continued to be useful to the state. Although veteran settlements continued to serve
a limited security role and to act as nodes of acculturation in under-urbanized
ROMAN WARFARE AND FOR TI FICA TION 709
Figure 27.4. City wall at Side incorporating spolia, possibly third century A.D. (Photo-
graph by G. Davies.)
areas, less formal arrangements could also be beneficial. When discharged soldiers
reentered civilian life, perhaps moving into the vici that had grown up outside their
old bases or returning to their home villages, a transfer of valuable skills and fresh
capital also followed. This gave the former soldier enhanced status among his
neighbors and also made future recruitment easier, as the benefits of service were
made apparent to all.
In order to fulfill their heavenly mandate to be the rulers of the world, the
Romans diligently applied themselves to the arts of war. Their successful mastery of
battlefield techniques and their adoption, where appropriate, of equipment and
technologies first introduced by their opponents allowed Roman armies to sustain
the state over several hundreds of years of challenge and change. For reasons of
length, this chapter has focused on the experiences and perspectives of the heavy
infantry citizen soldiers, the definitive image of Roman power. As a result, little
could be said about the vital role of auxilia and their organization and equipage,
Roman naval capabilities, the various complex frontier systems, and the central-
ized production of state-organized fabricae manufacturing equipment during the
later Empire. Nevertheless, enough has been said to confirm the complexity and
sophistication of this technology and its role in establishing and maintaining the
Roman imperial system.
REFERENCES
Bishop, M. C., and J. C. N. Coulston 2006. Roman military equipment. 2nd ed. Oxford:
Oxbow.
Bowman, A. K. 1994. Life and letters on the Roman frontier: Vindolanda and its people.
London: Routledge.
Brewer, R. J. (ed.) 2000. Roman fortresses and their legions. London: The Society of Anti-
quaries.
Connolly, P. 1998. Greece and Rome at war. 2nd ed. London: Greenhill.
Cornell, T. J. 1995. The beginnings of Rome. London: Routledge.
Coulston, J.C. N. 1985. "Roman archery equipment," in M. C. Bishop (ed.), The production
and distribution of Roman military equipment: Proceedings of the Second Roman
Military Equipment Research Seminar. British Archaeological Reports, Intl Series S275.
Oxford: BAR, 220-336.
Davies, G. 2000. "Cremna in Pisidia: A re-appraisal of tlie siege works," Anatolian Studies
50: 151-58.
Davies, G. 2006. Roman siege works. Stroud: Tempus.
Dixon, K. R., and P. Soutliern 1992. The Roman cavalry. London: Batsford.
Eagle, J. 1989. "Testing plumbatae," in C. van Driel-Murray (ed.), Proceedings of the Fifth
Military Equipment Conference. British Archaeological Reports, Intl. Series S476.
Oxford: BAR, 247-53.
Goldswortliy, A.K. 1996. The Roman army at war, 100 BC-AD 200. Oxford: Clarendon.
Gregory, S. 1986. "Not 'Why not Playing Cards?' but 'Why Playing Cards in tlie First
Place," in P. Freeman and D. Kennedy, The Defence of the Roman and Byzantine East. 2
vols. British Archaeological Reports, Intl. Series S297. Oxford: BAR, 1: 169-75.
Gregory, S. 1996a. Roman military architecture on the eastern frontier, A.D. 200-600. 3 vols.
Amsterdam: Hakkert.
Gregory, S. 1996b. "Was There an Eastern Origin for tlie Design of Late Roman fortifi-
cations? Some Problems for Research on Forts of Rome's Eastern Frontier," in
D. Kennedy (ed.), The Roman Army in the East. Journal of Roman Archaeology, Suppl.
18. Ann Arbor: JRA, 169-210.
Gros, P. 1992. "Moenia: Aspects defensifs et aspects representatifs des fortifications," in
S. van de Maele and J.M. Fossey (eds.), Fortificationes antiquae. McGill University
Monographs in Classical Archaeology and History 12. Amsterdam: Gieben, 211-25.
James, S. 2004. The excavations at Dura-Europos, 1928-1937. Final Report VII: The arms and
armour and other military equipment. London: The British Museum Press.
Johnson, S. 1983. Late Roman fortifications. London: Batsford.
Kennedy, D. 2004. The Roman army in Jordan. 2nd ed. London: CBRL.
Keppie, L. 1984. The making of the Roman army. London: Batsford.
Lander, J. 1984. Roman stone fortifications: Variations and change from the first century
AD to the fourth. British Archaeological Reports, Intl. Series S206. Oxford: BAR.
Marsden, E.W. 1969. Greek and Roman artillery: Historical development. Oxford:
Clarendon Press.
Marsden, E. W. 1971. Greek and Roman artillery: Technical treatises. Oxford: Clarendon
Press.
Mitchell, S., S. Cormack, R. Fursdon, E. Owens, and J. Oztiirk 1995. Cremna in Pisidia: An
ancient city in peace and war. London: Duckworth and The Classical Press of Wales.
Oleson, J.P., G. S. Baker, E. De Bruijn, R. M. Foote, J. Logan, M. B. Reeves, and A. N.
Sherwood 2003. "Preliminary report of the al-Humayma Excavation Project, 2000,
2002," Annual of the Department of Antiquities of Jordan 47: 37-64.
Parker, S. T. 2000. "Roman legionary fortresses in the East," in Brewer 2000: 121-38.
Parker, S. T. The Roman Frontier in Central Jordan: Final Report on the Limes
2006.
Arabicus Project, 1980-1989. Princeton: Princeton University Press.
Roth, J. 1995. "The length of the siege of Masada," Scripta Classica Israelica 14: 87-110.
Todd, M. 1978. The walls of Rome. Totowa, NJ: Rowman and Littlefield.
Tomlin, R. S. 0. 2000. "The legions in the late Empire," in Brewer 2000: 159-78.
Welfare, H., and V. Swan 1995. Roman camps in England: The field archaeology. London:
HMSO.
PART VII
TECHNOLOGIES
OF THE MIND
CHAPTER 28
INFORMATION
TECHNOLOGIES:
WRITING, BOOK
PRODUCTION,
AND THE ROLE
OF LITERACY
WILLY CLARYSSE
KA TELIJN V ANDORPE
THE ARRIVAL OF WRITING
Today, virtually all humans are affected directly daily by technologies of writing, so
it is difficult for us to imagine a completely oral society. "We live in a world of
visible words" (Small 1997: 3). The invention of writing was a huge step in the
development of civilization, because it extended human memory from the brain of
single individuals to an external storage medium. Writing provided an artificial
memory that could be consulted at any time; seeing replaced hearing as a means of
communication and as the means of storing communication (Havelock 1986: 100).
Instead of experts who could remember the things of the past, the new experts were
able to "decode" the written symbols. In an oral society, versified language played a
716 TECHNOLOGIES OF THE MIND
100%
90%
80%
70%
60%
■ Prose
50%
■ Poetry
40%
30%
20%
10%
0%
BC BC BC AD AD AD AD AD AD AD
III II I I II III IV V VI VII
Figure 28.1. Percentage relationship between Greek and Latin texts in prose and poetry
found in Egypt, third century B.C. to seventh century A.D.
much larger role because memorization largely depended on the use of rhythm, as
Plato realized (Tht. 267a) .
The earliest Greek literature was written in verse, and this is surely a remnant
of the oral society. Identification of the Muses as the "daughters of memory" is
linked with the functional use of poetry and music as "storage of cultural infor-
mation for re-use, the instrument for the establishment of a cultural tradition"
(Havelock 1986: 55-56). Greek prose literature starts only in the later fifth century
B.C. and takes over from verse in such fields as philosophy, history, and epideictic
oratory. The ongoing shift from verse to prose continues even in the later period, as
is shown by papyrus documentation from Egypt (figure 28.1). Notwithstanding the
continuing popularity of Homer in the school curriculum, prose texts climb from
30 percent of literary papyri in Hellenistic Egypt, to nearly 50 percent in early
Byzantine Egypt (for actual totals rather than percentages, see LDAB). In Christian
literature, church hymns and psalms take over from lyric and tragic poetry.
Sumerian cuneiform (Green 1981: 345-72) and Egyptian hieroglyphic writing
(Dreyer 1998, with corrections by Kahl 2003) both developed at the end of the
fourth millennium B.C., no doubt as devices to meet the administrative require-
ments of an increasingly complex society. It is no accident that the oldest texts are
written for the purpose of bookkeeping, both in Sumer and in Egypt: writing came
about through economic necessity, not for religious or cultural purposes. Cunei-
form writing was developed to record transactions at the temple complex in Uruk,
INFORMATION TECHNOLOGIES 717
a city with perhaps 50,000 inhabitants, enormous for that period. It was the only
way to collect and preserve information on such a large scale. Tokens and seals,
which can be considered precursors of writing, were no longer sufficient.
The long-term evolution of writing tended toward an ever closer adaptation of
script to sound. Some symbols did remain (most prominently digits, which are in
fact multilingual symbols even today), but every new script rendered more closely
the spoken word. Consonantal writing perhaps started as early as the Middle King-
dom in the so-called proto-Sinaitic scripts (Daniels and Bright 1996), which wrote
Semitic languages with signs derived from Egyptian hieroglyphs. In Ugarit, cune-
iform writing was used as the basis for two alphabets around 1200 B.C.; one of
them already has the order aleph, bet, gimel (A, B, C). Writing systems based on a
consonantal skeleton marking only the long vowels were adapted for several West-
Semitic languages (Aramaic, Phoenician); such a system still functions quite well in
modern Arabic.
The Greek alphabet was adopted and adapted from the North-Semitic (Phoe-
nician) script. Both the time (first part of the eighth century B.C.?) (Jeffery 1990:
12-21; Powell 1991: 18-20; Harris 1996: 58-59) and place (the southern part of the
Aegean, perhaps Rhodes, Cyprus, Crete, or Euboea, or a Greek trade post on the
coast of North Syria, such as el-Mina; Jeffery 1990: 5-12; Powell 1991: 12-18;
Woodard 1997) are still under discussion. There can be no doubt that Greek liter-
acy started under the impulse of the Near East. Herodotus' well-known statement
(5.58) that the Phoenicians brought the alphabet, is confirmed by the names of the
letters (acrophonic a-leph, b-et, and so forth. have a meaning in Semitic, but not in
Greek), the order of the alphabet, the shapes of the letters, and the term phoinikeia
(grammata) used in early Greek inscriptions. From the start, the Greek alphabet
differed from its model in its use of five signs to denote the vowels. Vowel notation
may first have been used within Phoenician for foreign names and then have
become generalized in Greek.
The adoption of the Greek alphabet in Italy in the late eighth century B.C. is
better documented. The alphabet of the Greek colony Cumae was taken over by the
neighboring Etruscans and then quickly by the other peoples of the peninsula.
There are about 120 Etruscan texts surviving from seventh-century B.C. Italy, com-
pared with only 10 Latin texts (Cornell 1991: 8). Etruria was not far behind Greece,
and the rest of Italy was not more than a generation behind them. Since the
Romans came into contact with a local alphabet of the Greeks who lived in the
West, the Latin alphabet differs from the classical Greek alphabet in a few details,
such as the value of the symbols X (ks instead of kh) and H (h instead of the vowel
eta) and the absence of a symbol for the sound ps at the end of the alphabet. In
addition, Latin abandoned some letters that it did not need (such as theta and phi)
and at a later stage created new ones (Lassere 2005: 34-35).
The earliest texts written in Greek date from the end of the eighth century B.C.
They are written on pottery (as on the Dipylon vase) or stone; other materials (such
as wax tablets) have not been preserved. The direction of writing was from right
to left, as in their Semitic models. The oldest Latin texts, such as the heavily
debated Fibula of Praeneste and the Duenos vases, were also written from right to
left (Lassere 2005: 34, 81). In seventh-century B.C. Greece, the second line was often
written from left to right and the next lines alternated. This boustrophedon script,
named after the turning of the ox (Greek bous) in plowing, was popular in the ar-
chaic period and is also attested in some early Latin inscriptions. By the early fifth
century B.C. our left to right direction-better suited to the right-handed majority,
since the written letters remain visible and ink is not smeared by the hand-was
already firmly established, although exceptions are found until the end of the cen-
tury (Jeffery 1990).
Many of the earliest texts are "speaking objects" (Bodel 2001: 19), often met-
rical: "I am the cup of Nestor," "[So-and-so] dedicated me." The preserved texts
written before 700 B.C. include a price inscription (Dipylon vase), an erotic spell
(Nestor's cup), and a dedication (Mantiklos bronze), all three in hexameter form.
There are also marks of ownership (personal names) and a maker's name. From the
first half of the seventh century B.C., abecedaria (sample alphabets for instruction),
funerary inscriptions, and curses are also found. The surveys found in Powell (1991:
119-81) and Thomas (1992: 52-61) are based on the standard work ofJeffery (1990).
The city-states start to make use of written laws and lists of officials after 650 B.C.
This marks the beginning of public inscriptions, which will become one of the
hallmarks of Greco-Roman society. They were first written on wooden boards
(Stroud 1979), later on slabs of marble or stones especially prepared for writing,
and often set up in a temple precinct. The coloring of the letters in alternating red
and blue and the stoichedon style, in which the letters are aligned both horizon-
tally and vertically as in a crossword puzzle, a style popular in sixth- and fifth-
century B.C. Greece, contributed to their monumentality and visibility rather than
to their legibility.
Written law, however, did not supersede oral law: it added monumental weight
and perhaps religious authority to some special laws that were meant to be read
aloud by officials. Oral law continued, and sometimes laws were put in verse so that
they could be learned by heart (Camassa 1988). The Cretan scribe Spensithius
(called poinikastas, referring to the Phoenician origin of the script) was at the same
time a mnemon (remembrancer; SEG 27.631 = Jeffery and Morpurgo-Davies 1970:
118-54; Detienne 1988: 67-70; Ruze 1988: 83-85). Written law and oral tradition are
not exclusive, but interact: archaic writing was in the service of the spoken word.
The diversity of local alphabets in the archaic and classical periods suggests
that writing spread quickly, before all the problems of adaptation to Greek had
been confronted (Harris 1996: 61). After Athens abandoned its own alphabet in
403 B.C., this diversity made place for a growing unity. From the Hellenistic period
onward, Greek writing is remarkably uniform across the circum-Mediterranean
world from Afghanistan to southern Italy. The Latin script became diversified
again in late antiquity, when monastic scriptoria developed different styles in the
provinces.
WRITING MATERIALS
The variety of writing materials in antiquity may be illustrated by the last dispo-
sitions of Augustus: these were written on two (wooden) notebooks (codices) and
copied onto three papyrus rolls (volumina) by the Vestal Virgins. The part record-
ing his personal accomplishments (res gestae) was then displayed on bronze tablets
in front of Augustus' mausoleum in Rome and on stone monuments in the prov-
inces of the Greek East (Suetonius, Aug. 101.1, 4; Ridley 2003). The choice of writing
materials was dictated by their availability and by the purpose they were meant to
serve: standard materials, usually more perishable, for archival purposes and books,
durable surfaces for public display.
Several perishable materials were available (cf. Bouquiaux-Simon et al. 2004).
In ancient Egypt, the papyrus plant (Lewis 1974) played the same role as bamboo in
the Far East. It was used for ropes, baskets, sandals, roof coverings, small boats, and
furniture. The stem could even be eaten, giving the Egyptians the epithet "papyrus
eaters" (papyrophagoi; Scholion Aeschylus, Supp. 761). But above all papyrus was
used as a writing material. In Egypt it was cheap (Skeat 1995: 87-90) and remained
the most common writing material for all kinds of texts until the arrival of paper
around A.D. 900. Elsewhere it had to be imported and was therefore relatively ex-
pensive, which made other writing materials more competitive. Nevertheless, pa-
pyrus was, in both Greece and Rome, the most common writing surface for longer
texts, especially literary works.
Animal skin is attested as a writing material from an early period in Greece and
in the Near East (diphtherai; Herodotus 5.58). Finds from Dura-Europos at the
Euphrates and Afghanistan (ancient Bactria) show that it was in common use in
the East in the Hellenistic period. Parchment, a form of untanned leather, is tra-
ditionally linked with the royal court of the Attalids at Pergamon in Asia Minor,
whence it gets its name (through Latin charta Pergamena and French parchemin).
Parchment is more durable than papyrus but far more expensive to produce. In
Egypt it is not found before the early second century A.D., and the oldest example
is a Latin text (LDAB 4472). From the fourth century A.D. onward, parchment
takes over from papyrus for luxury books in the Mediterranean, but not for ordi-
nary documents. This only happens after the Arabs conquered Egypt (A.D. 650)
and overseas import of papyrus became more expensive and difficult.
For everyday documents and archival purposes, thick wooden tablets (Greek
deltoi; Latin tabulae, tabellae), often coated with wax and written on with a stilus
(figure 28.2), were also commonly used in the Greco-Roman world. The example
illustrated is one of five wax tablets forming the schoolbook of Aurelius Antonius, a
young boy in Egypt who copied several times a Greek saying of the Athenian orator
Isocrates: "For some people honor their friends only when they are present, other
people also love them when they are far away." The beech tablet is deepened on
both sides and coated with black wax (beeswax mixed with carbon). The Romans
in particular preferred to write legal documents of a private and public nature on
Figure 28.2. Wax tablet from schoolbook of Aurelius Antonius, Egypt, fourth century
A.O. (Papyrological Institute Leiden inv. V.20 = LDAB 2530.) (Photograph by permission
of Papyrological Institute Leiden.)
tablets (Meyer 2004; Marichal 1992a; Lassere 2005: 456-57). Such tablets, usually
bundled into booklets of two, three or multiple tablets (diptychs, triptychs, or
polyptychs), were, for instance, filed in Rome's state archives and in the Campa-
nian business archives (see below). Roman Britain produced another type of writ-
ing material: thin leaf tablets of wood, written on with ink and sometimes bundled
into multileaved notebooks, have been excavated at the fort of Vindolanda im-
mediately south of Hadrian's wall (Bowman and Thomas 1983: 32-45; 2003). It is
easy to underestimate the importance of wooden tablets in Roman daily life, since
only a small fraction of them have survived (Eck 1998).
Ostraka (potsherds) were a cheap alternative as a writing material, giving their
name to the Athenian exile procedure called ostracism (Forsdyke 2005) . Thousands
of ostraka, mainly private letters, accounts, and tax receipts, have been discovered
in Greco-Roman Egypt (e.g., Cuvigny 2000 ), but other regions, such as Bu Njem in
modern Libya, have produced numerous examples as well (Marichal 1992b; Lassere
2005: 457-58) . Lead was used for magical purposes, but also for letters and even for
a record of horses of the Athenian cavalry (Kroll 1977).
More durable surfaces were preferred for public notices. Wooden boards,
often whitened (Greek leukomata, Latin alba), were used extensively for temporary
public notices in the Greek and Roman world, while stone and bronze (the latter
especially in the Roman world) were meant for long term publicity and as symbolic
memorials of important decisions. Only exceptionally were wooden boards used
for permanent publication, as, for instance, in archaic Greece for the laws of Draco
and Solon (Stroud 1979). Writing was not only used for the dissemination of public
documents, but legends or other inscriptions could also accompany and clarify all
kinds of objects and monuments, such as tombs, statues, public buildings, mile-
stones, altars, vases (Immerwahr 1990), and even paintings, reliefs, and mosaics
(Corbier 1995).
In Greece, "local variations in natural resources and political systems produced
very different epigraphic profiles in different cities and regions" (Bodel 2001: 10-
15). The importance of inscriptions in classical Athens, for instance, was closely
linked to its democratic institutions and was enhanced by the ready supply of
marble. An explosion of epigraphic activity in the Roman world is attested from
the reign of Augustus onward, when strong, white marble from the Apuan Alps
north of Pisa was shipped to Rome; "whether or not Augustus purposefully set out
to reshape the epigraphic landscape, the example he set at the capital for acceptable
forms of public display established a pattern and a set of standards that quickly
spread throughout Italy and the western provinces" (Bodel 2001: 6-10). The Ro-
man epigraphic habit during the first and second centuries seems to be typical of
urbanized and militarized areas where individuals wanted to assert their status in a
complex and changing society (Woolf 1998).
ROLL AND CODEX
Bookrolls were made of papyrus, rarely of parchment (Bouquiaux-Simon et al.

2004). For the production of papyrus sheets, the triangular stem of Cyperus papyrus
L. was peeled and the stem was cut in thin slices, which were put side by side on a
wooden board; a second row was placed over the first at 90 degrees. Some 20 or 40
such double thickness sheets (kollemata) formed a roll (see Pliny, HN 13.68-72, 74,
77, 81, 89). The dimensions of the sheets varied, but30 cm was a normal height for a
roll (figure 28.3). On the inside surface (recto) the papyrus fibers ran horizontally,
along with the length of the roll, and the sheet-joins were vertical; on the outside
(verso) both the papyrus fibers and the sheet-joins ran vertically. The sheet-joins
(kolleseis) are usually 15-20 cm apart. In order to protect the roll, an extra leaf called
protokollon (literally "first sheet") was provided. The difference between recto and
verso is visible not only through the sheet-joins, however; the recto is also more
polished, because writing always starts on that side. The verso, originally on the
outside of the roll, was only used after the original text on the inside surface was no
longer needed.
Since the Old Kingdom, Egyptian papyrus rolls (tomoi in Greek, volumina in
Latin) were written, or rather painted with a brush, in successive columns (selides
in Greek, paginae in Latin) side by side from top to bottom of a scroll oriented
sideways. Since the Greeks wrote from left to right they just turned the rolls 180
degrees so that the reed pen (kalamos) still went "downhill" over the sheet joins,
but they continued to write in columns (figure 28.4). Only in the Byzantine period
kollesls kollesis
r, r-.
Protokolfon
sheet joins
Figure 28.3. Diagram of the construction of a papyrus roll.
are some rolls written transversa charta, in a single column from top to bottom
oriented the long way, as in medieval rolls ( rotuli) in the West.
A standard roll of 20 sheets, measuring 320-360 cm (Skeat 1982), could contain
only a limited amount of text, for instance, one play of Euripides. Multiple lengths
were frequent, especially doubles, with some reaching up to 15 and 20 m (Johnson
2004: 143-52) . Since very long rolls were cumbersome to manipulate, the scholars
of the Alexandrian Museum cut up longer texts, such as Homer, into sections of
approximately the same length, which we still call "books." Later authors spon-
taneously divided their works into similar "books," adapting themselves to the
constraints of the medium. Bookrolls are attested from about 3000 B.C., and they
0 1 2 l 4 5 G 7 8 9 10 ~1 12 1:\ 1~ tS 16 17 IA ,fl :7n
--- - ~ - 11111 - --------=------ - -. j
Figure 28.4. Late-third-century B.C. papyrus from Egypt containing fragment of Me-
nander, Sikyonios, copied by a professional scribe. It was later used as secondhand paper
in mummy cartonnage. Papyrus Sorbonne inv. 2272b = LDAB 2738. (Photograph by J.
Gascou, by permission.)
■ Codex Christian
D Codex classical
■ Roll Christian
■ Roll classical
Figure 28.5. Percentage relationship of codex and roll in Greek and Latin books found in
Egypt, third century B.C. to the seventh century A.D.
were still used in the sixth century A.D. for administrative purposes as well as for
literature. In order to facilitate the unrolling, a stick called a "navel" (Greek om-
phalos, Latin umbilicus) could be attached to the roll.
From the second century A.D. onward a new means of uniting longer texts
became common: the book consisting of pages that are turned over rather than
unrolled. The origin of the modern book, the codex, is the wooden notebook al-
ready known to the Assyrians. In Rome, a new type of notebook made from parch-
ment (membranae) but modeled on the codex design, spread to literary texts. The
new format had distinct advantages: both sides of the pages were now used from
the start, saving space, and the book was easier to handle, transport, and consult. It
took several centuries before the codex definitely superseded the roll. Clearly the
habits of the Christians played a decisive part in this change: whereas in the second
and third centuries most classical authors were still copied on rolls, Christian texts,
especially the New Testament, were codices from the start. The Bible may have
been considered a form of a handbook rather than a literary work, or the use of the
codex may reflect some opposition to the Torah (still written on rolls today), or
perhaps the conservative reflex just did not affect this new kind of book (figure
28.5). The scaled (percentage) graph illustrates the disappearance of the roll in the
fourth century for Greek and Latin texts found in Egypt, and the absence of rolls
for Christian literature.
The codex also allowed for easier and more precise referencing within a work.
In the early fourth century A.D., Eusebius of Caesarea divided the gospels into
short numbered paragraphs, which he used for concordances between parallel pas-
sages. Numbered columns are already found in the Herculaneum rolls, but pagi-
nation is much more common in codices. In any case, before the printing press was
invented each book remained an individual copy, and pagination could not be
used for reference.
It took some time before the new book form completely distanced itself from
the scroll: the pages of the oldest papyrus codices were simply cut from rolls and
were not yet divided into quires. Single-quire codices of more than 100 pages
(25 folded sheets), however, are quite common in the third and fourth centuries
(Turner 1977: 58). When the book is open, narrow columns, two and even three to
the page, give the same panoramic view as a roll. The titles of a work in a bookroll
came at the end, since it was up to the next reader to roll the book back to the be-
ginning, and this habit continued for some time in the codex format as well.
BOOK PRODUCTION AND BOOK TRADE
To us, ancient books seem notoriously unfriendly to the reader. Words were not
separated, there were no capital letters and hardly any punctuation to mark a new
beginning, no paragraphs or indents. Each column was an unbroken block of let-
ters. Until the third century B.C., even poetry was written in continuous lines, as in
the papyrus of the Persae of Timotheus (LDAB 4123). This scriptio continua is in
fact a step backward from the Semitic alphabets, which distinguished words with
spaces or dots, as do some of early Greek inscriptions and many Latin inscriptions.
Changes of speaker in a dialogue were marked with a paragraphos-stroke at the
beginning of the line, but the reader had to guess at the names of the speakers.
Aesthetic motives apparently took priority over practical considerations, as the
literary script imitated inscriptions. Accents and punctuation were gradually ad-
ded, but sparingly, and word division was only introduced with minuscule writing
in the ninth century A.D. Although reading Greek or Latin was, of course, far easier
for native speakers than it is for us, the deficient layout certainly slowed down the
reading oflong or complicated texts. It also made word searches almost impossible.
In fact, many texts were written down primarily to be read aloud by a reader or
as mnemonic aids for a well-known text, whether it was the Iliad or the Bible. In
documentary papyri, retrieving texts and adding figures is often facilitated by the
layout.
The beginning of the book trade goes back to the late fifth century B.C.
(Bouquiaux-Simon et al. 2004). Socrates tells us how the works of Anaxagoras could
be bought at the marketplace for one drachma (Xenophon, Ap. 26d), Xenophon
gives a description of ships full of books (bibloi gegrammenai) run aground in the
Black Sea, wrecked and plundered by the Thracians (An. 7.5.14), and Aristophanes
mocks the new habit of "reading books for the sake of instruction" (Ran. 1114).
Since no copyright existed, anybody could copy and sell any work, nor was there
any quality control outside the private copies made directly from the original (for
instance in the philosophical schools).
This situation changed thanks partly to the founding of the famous library of
Alexandria and the philological work in the Museum. Whereas in the third century
B.C. copies of Homer are widely divergent, with verses added and omitted at
random, the later texts become standardized; even the uniform verse numbering is
the same as in modern editions. The new science of philology, based on careful
copying of authoritative models, was taken over by the Christians, whose bibles
present very few deviations from the norm. Of course, not all editions were perfect,
and there was a great distinction between school books and the deluxe copies of the
rich. In Rome, commercial reproduction by well-organized publishing firms could
put a new book on the market in hundreds of copies in a few days time. Purchasers
sometimes checked several copies in order to obtain a correct text. Martial (1.117)
gives us an entertaining description of a bookshop in Rome: "Next to the Forum is
a bookshop, where both doors are plastered with advertisements. These display the
titles of the books in stock, and you need only cast a glance at the list. Go in and ask
for my book. The owner-his name is Atrectus-will be extremely pleased to get a
fine copy of Martial out of his first or second shelf and let you have it for five
denarii."
LIBRARIES
The oldest collections of books are attributed to the tyrants Polycrates of Samas
and Pisistratus of Athens-the latter is even said to have opened his library to the
public. Euripides also owned a private library. Aristotle's library, partly captured
by Sulla, was the first to be systematically arranged (Casson 2001: 17-30). It was a
model for the great libraries of Alexandria and Rome and had an enormous influ-
ence on the development of scholarship. Unlike Plato, who preferred oral discus-
sion to the bookish accumulation of knowledge, Aristotle's encyclopedic type of
scholarship was only possible with the help of an extensive collection of books. In
the Hellenistic period large, semipublic libraries were sponsored by the eastern
Mediterranean kings, most famously that in Alexandria, which for the first time
attempted at being complete and formed the basis of Alexandrian scholarship
(Casson 2001: 31-47; Bagnall 2006, with a warning against overestimation of its
size). Here the best manuscripts were compared and the first commentaries and
critical editions of texts were made. Roman nobles owned extensive libraries, of
which only that of L. Calpurnius Pisa, the father-in-law of Caesar, has partially
survived in Herculaneum (1,800 charred scrolls). It was housed in a small room
with shelves around the walls and a reading table in the middle. But in Rome and in
the provinces many libraries were available to the public (28 in Rome alone at the
time of Constantine; Curiosum, Appendix). Most libraries were designed for read-
ing on the premises, but in some borrowing was possible (Casson 2001: 107-8, 161).
Literary culture had indeed become one of the hallmarks of Greco-Roman civili-
zation; its retreat within the walls of palaces and monasteries contributed to the
new civilization of late antiquity.
With the barbarian invasions in the fifth century, literacy retreated in the west-
ern empire, while at the same time the wealth of inscriptions dried up in the cities
of the east, and even in Egypt the number of papyri temporarily dwindled. But
simultaneously literacy was on the rise, since natives started writing their own
languages in a Greek alphabet (Coptic and Gothic) or with renewed vigor (Syriac).
In religious matters, books suddenly became of prime importance. Whereas the
pagan rites and myths were handed on by word of mouth and only very partially
written down, the new religions, Christianity, Manichaeism, and later Islam, were
"religions of the book," in which God communicated with this world through
sacred scriptures. Christ himself, although he never seems to have written anything
down, is represented in the apse of the churches with a codex Bible in his hands.
Bible texts were copied for magic, read aloud in the church, learned by heart, and
endlessly commented on by theologians. Believers split along religious lines (gnos-
tics, Arians, orthodox, monophysites) according to the books they read or the
interpretations they gave to them. This religious divide became one of the main
problems of the early Byzantine Empire.
RECORD-KEEPING
A discussion of record-keeping in antiquity may focus on various points of interest

(compare Brosius 2003 and Nicolet 1994). Public and temple records are usually
kept in or near buildings of officials or institutions in charge of them; even public
records may be kept under the custody of gods. The writing material determined
the arrangement of the records (wooden tablets were stored and labeled in a dif-
ferent way from papyrus rolls) and the construction of the archive rooms (Meso-
potamian clay tablets had to be stored in rooms with the proper humidity to avoid
crumbling). The classification systems used by the Greeks and Romans were
gradually refined as the public archives increased in number, became more com-
plex, and had to cope with falsifications and fraud. Roman Egypt developed an
efficient system of archiving public and even private records. The availability of
public records is an elementary part of modern archiving, but it is not always
attested for ancient archives. Records on stone or bronze were not only meant for
publishing or publicizing, but were also authoritative. For us they often form the
only traces of the original records.
Our knowledge of techniques of record-making and record-keeping in the
Greco-Roman world is scanty. Whereas the remains of several archival depositories
have survived from the ancient Near East and the Minoan and Mycenaean world,
the practice of archiving in Greece and Rome is only known when recorded on
stone or bronze, or when referred to in literary works. Occasional building remains
(such as the Athenian Metroon and the Roman Tabularium) or seals attached to
archival records and baked by fire (e.g., in the Selinunte temple, Zoppi 1996, and in
a Delian house, Boussac 1996) witness Greco-Roman record-keeping. Apart from
some find spots such as Vindolanda in Roman Britain, Greco-Roman Egypt is the
main exception. Here large groups of state and temple records have been uncov-
ered in ancient dumps or through the antiquities market (Vandorpe forthcoming),
but in general even Egypt has produced only a few archival establishments with
their holdings. Some carbonized papyrus rolls were discovered in the ruins of the
Roman record-office (bibliotheke) in Thmouis, the capital of the Mendesian nome,
set on fire by furious taxpayers (Kambitsis 1985).
Notwithstanding the scarcity of archival records, the importance of Greek
record-keeping may be clearly seen from the terms arche (magistracy) and archeion
(office of the chief magistrates where public records could be kept), which have
become internationally accepted archival terms in the Western world. This should
not surprise us, since the polis "gave rise to a novel type of political life and gov-
ernment that could not fail to have a decisive bearing on archival organization and
service" (Posner 1972: 91). Whereas the earlier palace civilizations mainly filed
simple administrative records, the Greek democratic institutions and offices gave
rise to numerous decrees, laws, and so forth, which were to be classified in the
magistrates' or state archives. Record-keeping became even more complex when
Greek cities made registration (anagraphe) and validation of private transactions
obligatory, so that not only public but also private records had to be kept.
Very soon the Greeks realized the importance of written records for democ-
racy: Aristotle regarded official archives as indispensable for his model state (Po-
litica 6.1321B.34) and Euripides (Supplices 430-38) contrasts tyranny and democracy
by comparing the availability of written laws: "Where the tyrant is, there are no
common laws, but one man rules, in whose keeping and in his alone the law
resides; so is equality no more. But when the laws are written down, the weak and
the wealthy have equal justice and it is open to the weaker to use the same language
to the prosperous when he is reviled by him, and the weaker prevails over the
stronger if he has justice on his side."
Heavily debated is the Metroon or state archive of Athens (Posner 1972;
Pritchett 1996: 14-36). When the enlarged Boule or Council became the supreme
board by the reforms of Clisthenes in 509/8 B.C., its archive became the most im-
portant one, and its records were probably housed in a new building on the agora
known as the Old Bouleuterion. This building was at the same time the meeting
place for the newly organized board and the record office of documents needed by
the board (Shear 1995; for a radical revision of the archaeological evidence, see
Miller 1995). The alternative suggestion that the fifth-century B.C. Council relied
on memory and oral communication, and that stone inscriptions served in some
way as an official archive (Thomas 1989 and 1992) has been called into doubt
(Sickinger 1994a; Pritchett 1996; Shear 1995: 186). Near the end of the fifth century a
New Bouleuterion was constructed nearby. The old Council chamber was replaced
by the Metroon, the sanctuary of the Mother of the Gods, which sheltered the
growing number of records under the goddess's custody (Munn 2006: 330-32). The
Metroon, often referred to as the "public archive" (demosion), played an important
role in the storage of public documents (Georgoudi 1988). On a more optimistic
view, it was a kind of central archive, where the Athenians gradually entered
ancient public records and systematically filed new ones (Shear 1995). In the period
when the Metroon was erected, a committee of "establishers of the laws" (no-
mothetai) was elected, who could undoubtedly find their working material in the
new building (Munn 2006: 331).
The state archive holdings in the Metroon could be consulted not only by
Council members, but also by litigants and other interested parties, for instance for
research purposes (Posner 1972: 113; Sickinger 2004). The layout of the original
building is unclear, as it was overbuilt by a Hellenistic Metroon in the second cen-
tury B.C. (Miller 1995: 142; Shear 1995: 158). Here the archives probably occupied
two of the four rooms.
The original records in Greek cities were written on perishable materials:
wooden tablets, papyrus rolls (Sickinger 1994a: 295; Shear 1995: 186, 188), or even
parchment in cities of Asia Minor when a security copy was needed. Wooden
tablets were probably useful for records that had to be consulted regularly or had to
be supplemented. It is unclear how the records were arranged. As suggested by
Posner (1972: 112), wooden tablets may have been stored in jars. The enormous jar
of the Metroon in which Diogenes the Cynic took up residence may have been such
a storage jar. Records on papyrus may have been kept in niches comparable to
those in contemporary library buildings. Undoubtedly, the date and the category
to which a document belonged played a crucial role in the classification; catalogues
are nowhere mentioned, but public slaves probably served as finding aids (Sick-
inger 2004: 103-4).
Alongside oral dissemination of decrees and laws "essential to democracy"
(Thomas 1989: 61-68; 1992), state records could be published, although this did not
happen systematically. Proposals and decisions might be temporarily displayed
in public on wooden boards, for instance fixed on the base of the statues of the
eponymous heroes in the agora (Thomas 1989: 62). Possibly these wooden boards
were afterward entered in the archival depository so that they functioned twice:
"exposees pour quelques jours ala vue de tous, afin de s'acquitter de leur mission
informative, elles gagneraient ensuite leur depot" (Georgoudi 1988: 236). From the
end of the sixth century B.C. onward, state documents that were considered im-
portant might be published on stone, serving a more permanent purpose (Sickinger
1994a: 2004). Some types of documents were even entrusted to bronze (Sickinger
2004: 100). The inscriptions were set up in the vicinity of places where the inscribed
documents had particular relevance. Some inscriptions contained less detailed
information than the original records (Sickinger 1994a). Nevertheless, the stone
copies were for the Greeks as valid and authoritative as the originals; they could
be consulted or quoted instead of the archival records. The Greeks did not know
the "modern concept of a single, authoritative version of a particular document"
(Sickinger 1994b: 277), although they clearly distinguished between the original
decrees (psephismata auta) and their copy (antigraphon) (Shear 1995: 187). The sym-
bolic value often attributed to inscriptions does not rule out their practical value
(Sickinger 2002: 154-55).
In the Hellenistic period, a selection of inscribed public documents could con-
stitute an archive on its own, as the texts were "picked out by the community (or
responsible authority) to create and broadcast a particular theme and message." In
an "inscriptional archive," like that displayed on the anta of the temple of Athena
Polias, the people of Priene decided, after a period of hostilities, to bring together
and publicize those records that established its entitlement to rights obtained in the
past (Sherwin-White 1985: 74-80).
The Greek temples were also important centers of record-keeping. Like
state institutions, they used wooden tablets or papyrus for their records and
published part of them on stone. Their inscribed inventories, as for instance at the
temple of Apollo at Delos, informed the city council and the gods about precious
temple objects, but were at the same time symbols showing the wealth of a temple
(Linders 1988; Thomas 1992: 86-87). The mobile archives of the naopoioi, re-
sponsible for the rebuilding of Apollo's temple at Delphi in the fourth century B.C.,
deserve special mention; their records were stored in wooden cases, which were
carried by paid porters, from the seat of the officials to the workshop and vice
versa, on the days when contracts and accounts were to be verified (Georgoudi
1988: 235-36).
The Athenian Metroon had counterparts in Rome in the Aerarium and Tabu-
larium. The Aerarium, housed in the temple of Saturn at the foot of the Capitoline
Hill, was foremost the repository of the state treasury. It was established in the early
Republic and was for a long time run by the young and inexperienced quaestors,
the consuls' financial assistants, until eventually, under Nero, more experienced
prefects were put in charge (Millar 1964; Corbier 1974). Understandably, financial
records pertaining to the treasury were kept at the Aerarium as well. At a later
stage, nonfinancial documents such as copies of laws (leges) and senatorial de-
crees (senatus consulta) were also deposited in there. In 78 B.C., after a fire on the
Capitoline Hill, a new archives building, the Tabularium, was constructed as an
annex to Saturn's temple, presumably to store the increasing number of records. It
was an impressive monument with a protected staircase leading traffic from the
Forum to the Capitoline Hill. In what remains, no rooms really suitable for record-
keeping were found; perhaps these were located on the second floor, which has not
survived (Culham 1989: 101-2).
The Aerarium and annex-Tabularium cannot be considered a central archive

in Republican times (compare the radical view of Culham 1989). The record-
keeping of senatorial decrees (senatus consulta) may clarify this point (Talbert 1984
and Coudry 1994, who also discuss the acta senatus, records of the proceedings in
the senate). Senatorial decrees were copied for a first time in the records (Greek:
deltoi) of the consuls, undoubtedly kept by these magistrates as part of their private
journals (commentarii) and taken home with them after their term of office. These
could be consulted as long as the consuls lived and are considered to have played a
key role in keeping a record of senatorial decrees, "un systeme complexe ou les
categories du public et du prive se distinguent mal" (Coudry 1994).
But senatorial decrees became valid only after deposition (delatio) in the Aera-
rium, "une formalite juridique ... une operation marginale" (Coudry 1994: 94).
The reason for deposition may be the belief that a document became effective only
after it had been filed under the guardianship of a god, in this case Saturn (Culham
1989). Whatever the motive, a second copy was kept in the quaestorial records of
the Aerarium (Greek: deltoi demosioi tamieutikoi), in the form of diptychs (two
tablets hinged together, producing four pages). Both consular and quaestorial
records were classified by year (named after the consuls or quaestors in charge) and
within a year, by month; the records of each month were subsequently numbered.
The diptychs were referred to by this number and by the page of the little book.
Undoubtedly, these tablets took up much more space than papyrus documents.
The approximately 12,500 characters of the senatorial decree recording the judicial
proceedings against Piso under Tiberius discovered on bronzes (see below), were
copied by the emperor's quaestor on fourteen tablets (tabulae), probably written
on both sides and undoubtedly bundled into one booklet (Eck et al. 1996).
Probably no public access was provided to these essential state records of the
Aerarium; only persons holding a magistracy could consult them. In addition, the
record-keeping in the late Republic Aerarium stood in bad repute: records were not
easily accessible and some were even falsified or destroyed (Culham 1989; Coudry
1994: 67, 71; Moreau 1994). Mark Antony, for instance, apparently deposited sen-
atorial decrees that had never been voted (Cicero, Phil. 5.12). As they served for only
a short term, the inexperienced quaestors had to rely on their subordinates, public
servants (apparitores) who became the "real magistrates" (Plutarch, Cat. Min. 16.3).
Cicero (Leg. 2.46) complained about the poor service: "We have no custody of the
laws and therefore they are whatever the apparitores want them to be; we have to
ask the copies from them, but we have no memory confided to the public records."
As a quaestor, Cato the Younger tried to improve the situation.
The public archive of the Aerarium was slow in developing, partly because
there were good alternatives for consulting, for instance, the Senate decrees (Posner
1972; Culham 1989; Coudry 1994). Apart from the consular records mentioned
above, the plebeians started early in the Republic to deposit in the temple of Ceres
both their own laws (plebiscita) and their own copies of Senate decrees which "were
wont to be suppressed or falsified" (Livy 3.55.13). This archival depository was
abolished by Augustus in 11 B.C. Finally, private record-keeping was well <level-
oped: Romans entered their transactions in their daybooks (adversaria) and at the
end of the month made a summary in the register of receipts and expenditures
(tabulae or codex accepti et expensi) (Thilo 1980). Rich households stored their re-
cords in a separate room called a tablinum; among these, privately purchased cop-
ies of public documents could be kept (Williamson 1987: 164).
Even though the Republican Aerarium did not function as a central archive,
there was "un archivage tres complet des documents issus du Senat ... , consta-
tation qui va a l' encontre de l'idee courante selon laquelle la Republique aurait
souffert d'une so rte de sous-administration" (Coudry 1994), but the role of the men
of status in this system should not be underrated, as they were present as witnesses
at all stages of the process of record-making (Williamson 1995) and had easy access
to the public records.
Augustus tried to make the public archive of the Aerarium more efficient by
redefining the tasks of the quaestors; probably in view of these reforms, the archive
in the temple of Ceres was abolished (Coudry 1994: 67, 74). Subsequently, the
public records at the Aerarium were no longer neglected. The records (tabulae),
books (libri), and accounts (rationes) probably had to be written down (scribere)
and classified (ordinare), and the responsible magistrates had to swear that there
was no falsification or fraud (Coudry 1994: 75). As the Senate's powers diminished,
an imperial archive developed composed of a private archive (sacrarium, secre-
tarium), an accounts section of the imperial household (tabularium castrense), an
imperial tabularium, and the chanceries called ab epistulis, a libellis, and a memoria.
How were public records given publicity in Rome? As the rate of literacy was
relatively low, proclamation was important: heralds read laws and important de-
crees aloud in public meetings. In addition, some information, such as the prae-
tor's annual edict and lists of prices, was temporarily posted in a visible place on
whitened wooden boards (alba) (Eck 1998). Finally, important statutes, interna-
tional treaties, and decrees could be published in bronze in Rome or even in provin-
cial cities. After Cn. Calpurnius Piso, accused of the murder by poison of Germa-
nicus and of other serious political offences, was condemned, the Senate ordered its
resolution to be displayed on bronze in Rome in whatever place seemed best to
Tiberius and likewise "in the most frequented city of each province and in the most
frequented place in that city," and "in the winter quarters of each legion where the
standards are kept." The Spanish province Baetica produced at least six bronze copies
of this senatorial resolution (Eck et al. 1996; Griffin 1997; Damon and Takacs 1999).
Like the Greek copies on stone, these bronze copies were authoritative, and
legal copies could be made from them, like military diplomas that had been ver-
ified from the bronzes in Rome. The choice of the "eternal" bronze for permanent
display was due to the absence of stone suitable for engraving in Rome's early days.
Even when marble became common during the Empire, Romans stuck to their old
traditions (Williamson 1987: 179); they had always linked bronze tablets to the
notion of inviolability. Bronze tablets are part of a divine context: these "objects for
or belonging to and therefore protected by the gods" are sacrosanct (Williamson
1987). In the western provinces, the habit of posting decisions on bronze was taken
over, whereas the Greek cities in the East remained faithful to the medium with
which they were familiar and displayed them on stone.
The most famous bronze collection was displayed on the Capitoline Hill,
Rome's ritual center in front of Jupiter's temple. The bronze versions are some-
times considered an archive constituted cumulatively and accessible to every Ro-
man citizen. State documents, however, were not systematically entrusted to bronze.
In addition, the bronze copies were hard to read, even for literate individuals, be-
cause of both the juridical jargon and the cramped lettering. Therefore, the bronze
tablets should not be considered as an archival collection meant to be consulted.
They rather serve another purpose: "they ceremonially reinforce Roman great-
ness," already embodied on the Capitoline Hill (Williamson 1987). In contrast,
display on stone is attested, for instance, for priestly records: the yearly acta (min-
utes of the proceedings) of the Arval Brethren, were, over a period of more than
250 years, inscribed on marble tablets at their headquarters in the grove of the Dea
Dia shrine near Rome (Scheid 1990, 1994; Bodel 2001: 20).
For Roman provincial record-keeping we may turn to Roman Egypt (Cockle
1984; Burkhalter 1990), where the Romans built on an existing system of public
record-keeping. This Roman province had "an extremely detailed system of ar-
chiving public and private documents, which seems to have been used to the full"
(Cockle 1984), organized from the lowest administrative levels up to the central
services in Alexandria and able to cope with falsification and fraud.
The public papyrus records, such as tax returns and census lists generated by
the village scribes, were transferred, through the intermediary level of the toparchy,
to the nome capital, where the officials in charge (the strategos and royal scribe)
deposited them into the central "record office of public documents" (bibliotheke
demosion logon). The archive keepers, often former gymnasiarchs, were responsible
for their conservation and classification; they were not allowed to alter, add or de-
lete anything in the texts. They gathered related documents, ordered them chro-
nologically or alphabetically (for first-letter alphabetical order, see Verhoogt 1998:
215) and pasted them together into rolls (tomoi sunkollesimoi; the pasting of pa-
pyrus documents is already attested for official texts in Hellenistic Egypt, Clarysse
2003). These rolls were provided with a general title referring to the document type,
such as "declaration by house." These rolls were subsequently classified into sec-
tions referring to the nome strategos in charge, the document type, the year, and the
village. The records could be consulted on a regular basis by, for instance, the
assistants of the provincial governor or of the nome strategos, or by litigants. Record
offices in the nome capitals kept copies, as the original documents were sent to the
Patrika record office in Alexandria.
A second bibliotheke in the nome metropolis, which split off from the public
record office just mentioned in A.D. 72, was the "record office oflanded property"
(bibliotheke ton enkteseon). Here private persons registered their holdings in order
to have their rights as owners or creditors guaranteed officially. Only declarations
of property addressed to this record office, no title deeds, were kept in this bib-
liotheke; on the basis of these, the archive keepers drew up survey sheets (dia-
stromata) recording the owners with their properties. The holders of property
within the nome were entered, in alphabetical order, by their place of origin; each
owner took up one or more papyrus sheets (kollema[ta]) of a roll on which their
holdings were classified by village and by category. The rolls as well as the papyrus
sheets were referred to by numbers.
Further record offices are attested in Alexandria under the supervision of the
archidikastes, the chief justice or "master of rolls." The Patrika, named after an
Alexandrian quarter, housed the original public documents from all over Egypt. In
addition, private documents were centrally registered in the capital: after entering
the register of the katalogeion, they were deposited in the bibliotheke of the Nanaion,
the temple of Isis Nanaia. After the creation of the temple of the divine Hadrian
(Hadrianeion), the original documents were kept and protected in this new deposit
registry, and duplicates could be consulted in the Nanaion.
Record-keeping techniques are also attested for some private archives, which
are, as a rule, less complex than their public counterparts. Business archives of
third-century B.C. Egypt have a registration system for letters showing that the
documents were systematically arranged according to date and contents (Van-
dorpe forthcoming). Egypt was not alone in developing record keeping. A house
on Delos that sheltered a large number of sealed papyrus records when it was des-
troyed by fire in 69 B.C., produced thousands of baked seals that allow the basic
classification system to be reconstructed. The businessman entered documents by
category and by client (Boussac 1996). Some Campanian archives of sealed waxed
tablets furnish evidence for Italy (Andreau 1999: 71-79; Marichal 1992a; Meyer
2004). The first-century A.D. archive of the Sulpicii, a firm of financiers from
Puteoli, for instance, consists of waxed tablets bound together in booklets (dip-
tychs or triptychs). A summary is sometimes added either on one of the unwaxed
outside tablets or on the spines of the booklets, or even on both (Camodeca 1999:
31-36).
LITERACY
A person is illiterate who "cannot with understanding both read and write a short
simple statement on his everyday life" (UNESCO 1977: 12). In antiquity, as now,
elementary forms of reading (Greek anagignoskein = to recognize the letters) may
well have been more widespread than the art of writing (Greek graphein = to
scratch), and many writers were able just to scratch their name and not much more
(for such a "signature literate," see figure 28.6). Statistics of ancient literacy rates
can therefore be misleading.
Because of the difficulty of the Bronze Age cuneiform and hieroglyphic writing
systems, each with hundreds of signs, partly phonetic and partly symbolic (idea-
rr,,1~
If.Ill<
Figure 28.6. Libellus of the Decian persecution (A.D. 250 ), illustrating the difference
between the hand of a professional scribe and the clumsy signature of one of the mem-
bers of the commission. John Rylands Library inv. 112a and b = V. Martin et. al. (eds.),
Catalogue of the Greek and Latin Papyri in the John Rylands Library, Manchester, Vol 2,
Documents of the Ptolemaic and Roman Periods (Manchester: Manchester University
Press, 1915), nos. 112 b and c. (Reproduced courtesy of the University Librarian and
Director, the John Rylands University Library, The University of Manchester.)
grams and determinatives), writing and reading was mastered only by a small pro-
fessional class of scribes, a typical feature of the cultures of this period. The intro-
duction of the alphabet made it far easier to master the arts of reading and writing
and no doubt resulted in a far higher literacy rate, although it is hardly possible to
work out percentages. According to Harris (1989) only a minority of adult males (less
than 10% in classical Greece, maximum 30% in imperial Rome) could read and
write; mass literacy was thus never achieved in classical times. This does not,
however, preclude the possibility that writing could penetrate deeply into the lives of
ordinary people: Egyptian peasants who could neither read nor write carefully kept
the potsherds with Greek tax receipts as proof that their salt tax or dyke tax was paid.
Even if only 10 percent of the male population could write in the Greco-Roman
world, this was ten times more than in the Ancient Near East, where literacy was in
fact limited to a small class of specialists. In Greece, writing was used from the start
for private notices, graffiti, and everyday messages between individuals. When the
community took hold of the new technology, it was applied for codifying laws and
became a powerful political instrument. In Athens, the political implications were
already used to the full at the end of the sixth century, when leading politicians
could be banned for ten years by a negative naming or vote made on potsherds,
known as ostracism (Forsdyke 2005). Thousands of such ostraka have been found
(Martin 1989), but they constitute only a small percentage of what was written.
Even if not every person was literate, this system implies a high degree of basic
literacy among the male citizen population. Illiteracy carried no stigma, but being
able to read and write was both practical in dealings with the government and a
sign of prestige (Hopkins 1991). Although literates were still a minority, the Roman
Empire later was bound together by writing, from an administrative, economic,
and cultural point of view.
The spread of literacy became possible when education was no longer private
and oriented toward physical development, music, and poetry, but became col-
lective and based also on the teaching of letters. The "teacher of letters" (gram-
matistes) gradually became a "teacher" (didaskalos). Herodotus (6.27.2) refers to a
school in Chios in 494 B.C. in which 119 children were killed by a collapsing roof. In
Hellenistic Miletus and Teos, schools for freeborn boys and even girls were spon-
sored by a wealthy citizen, and clearly the children of citizens were expected to at-
tend daily courses (SIG 2.577-78; cf. SEG 30 [1980] no. 1535.24-28, for Xanthus in
Lyda) (Marrou 1965: 78-93, 218-42). The Ptolemies subsidized Greek education
by exempting schoolteachers from the salt tax, and the ratio of teachers to male
children of the whole population (including Egyptians) is as high as 1:165. This
statistic confirms Marrou' s optimistic view of widespread diffusion of Greek edu-
cation in the Egyptian countryside, against Harris' more pessimistic view (Clarysse
and Thompson 2006: 125-33). In Hellenistic and Roman Egypt, girls of the upper
strata of society also received some degree of education, and even women teach-
ers are not unknown (Cribiore 2001: 74-101). Literacy did not diminish in the early
Byzantine period, as is shown by the numerous school texts (Cribiore 1996). Al-
though the monastic movement prided itself on simplicity and even illiteracy,
Pachomius' monastic rules explicitly stipulate that illiterate persons should be
taught to read, in order to learn the sacred scriptures by heart.
Orality and literacy are indeed not opposites, but orality continued to hold a strong
position throughout antiquity. Thus Plato's dialogues, masterpieces of written lit-
erature, explicitly proclaim the primacy of living discussion over the static written
word. The dialogue genre is not only used for philosophical discussions, but even for
biography and history, as in Thucydides' dialogue of the Melians. At school, Homer
and Demosthenes were not only copied, but also learned by heart by generations of
pupils; similarly the monks used the books in the monastic library to learn by heart
the gospels and psalms: the written word is in the service of the spoken word. Ancient
poets, orators, and even historians would give declamations of their works to a
friendly audience before publishing it in written form, and oratory in general, like
drama, could only be fully appreciated through performance.
REFERENCES
Andreau, J. 1999. Banking and business in the Roman world. Cambridge: Cambridge Uni-
versity Press.
Bagnall, R. S. 2006. Hellenistic and Roman Egypt: Sources and approaches. Aldershot-
Burlington: Ashgate.
Bodel, J. 2001. Epigraphic evidence: Ancient history from inscriptions. London: Routledge.
Bouquiaux-Simon, 0., M.-H. Marganne, W. Clarysse, K. Vandorpe, and J.-C. Didderen
2004. Les livres dans le monde greco-romain. Liege: Editions de l'Universite de Liege.
a
Boussac, M.-F., 1996. "Etude statistique d'un depot d'archives Delos," in M.-F. Boussac
and A. Invernizzi (eds.), Archives et sceaux du monde hellenistique. Bulletin de corre-
spondance hellenique Suppl. 29. Atlienes: Ecole franc;aise d' Atlienes, 511-23.
Bowman, A. K., and J. D. Thomas 1983. Vindolanda: The Latin writing-tablets. London:
Society for tlie Promotion of Roman Studies.
Bowman, A. K., and J. D. Thomas 2003. The Vindolanda writing-tablets: Tabulae Vindo-
landenses II. London: British Museum Press.
Brosius, M. (ed.) 2003. Ancient archives and archival traditions: Concepts of record-keeping
in the ancient world. Oxford: Oxford University Press.
Burkhalter, F. 1990. "Archives locales et archives centrales en Egypte romaine," Chiron 20:
191-215.
Camassa, G. 1988. "Aux origines de la codification ecrite des lois en Grece," in Detienne
1988: 130-55.
Camodeca, G. 1999. Tabulae Pompeianae Sulpiciorum (TPSulp.): Edizione critica del-
l'archivio puteolano dei Sulpicii. 2 vols. Rome: Quasar.
Casson, L. 2001. Libraries in the ancient world. New Haven: Yale University Press.
Clarysse, W. 2003. "Tomoi Synkollesimoi," in Brosius 2003: 344-39.
Clarysse, W., and D. J. Thompson 2006. Counting the people in Hellenistic Egypt. 2 vols.
Cockle, W. E. H. 1984. "State archives in Graeco-Roman Egypt from 30 B.C. to tlie reign of
Septimius Severus," Journal of Egyptian Archaeology 70: 106-22.
Corbier, M. 1974. L'aerarium Saturni et l'aerarium militare: Administration et prosopo-
graphie senatoriale. Collection de l'ecole franyaise de Rome 24. Rome: Ecole franyaise
de Rome.
Corbier, M. 1995. "L'ecriture dans l'image," in Solin and Salomies 1995: 113-61.
Cornell, T. J. 1991. "The tyranny of evidence: A discussion of tlie possible uses of liter-
acy in Etruria and Latium in tlie archaic age," in M. Beard, A. K. Bowman, and
M. Corbier (eds.), Literacy in the Roman world. Journal of Roman Archaeology Suppl. 3.
Ann Arbor: University of Michigan Press: 133-58.
Coudry, M. 1994. "Senatus-consultes et acta senatus: redaction, conservation et archivage
a
des documents emanant du senat, de l'epoque de Cesar celle des Severes," in De-
mougin 1994: 65-102.
Cribiore, R. 1996. Writing, teachers, and students in Graeco-Roman Egypt. Atlanta, GA:
Scholars Press.
Cribiore, R. 2001. Gymnastics of the mind. Greek education in Hellenistic and Roman Egypt.
Princeton: Princeton University Press.
Culham, P. 1989. "Archives and alternatives in Republican Rome," Classical Philology 84:
100-15.
Cuvigny, H. 2000. Mons Claudianus: Ostraca graeca et latina. III. Les rerus pour avances ala
"familia." Cairo: Institut franc;ais d'archeologie orientale du Caire.
Damon, C., and A. S. Takacs (eds.). 1999. "The Senatus Consultum de Cn. Pisone patre:
Text, translation, discussion," American Journal of Philology 120: 1-162.
Daniels, P. T., and W. Bright 1996. The world's writing systems. Oxford: Oxford University
Press.
Demougin, S. (ed.) 1994, La memoire perdue: A la recherche des archives oubliees, publiques
et privees, de la Rome antique. Serie Histoire ancienne et medievale 30. Paris: Pub-
lications de la Sorbonne.
Detienne, M. 1988. Les savoirs de l'ecriture en Grece ancienne. Lille: Presses Universitaires.
Dreyer, G. 1998. Umm el-Qaab I: Das predynastische Konigsgrab U-j und seine fruhen
Schriftzeugnisse. Mainz: von Zabern.
Eck, W. 1998. "Inschriften auf Holz. Ein unterschatztes Phanomen der epigraphischen
Kultur Roms," in P. Kneissl and V. Losemann (eds.). Imperium Romanum: Studien zu
Geschichte und Rezeption. Stuttgart: Steiner, 203-17.
Eck, W., A. Caballos, and F. Fernandez 1996. Das Senatus Consultum de Cn. Pisone patre.
Munich: Beck.
Forsdyke, S. 2005. Exile, ostracism, and democracy: The politics of expulsion in ancient
Greece. Princeton: Princeton University Press.
Georgoudi, S. 1988. "Manieres d'archivage et archives de cites," in Detienne 1998: 221-47.
Green, M. W. 1981. "The construction and implementation of the cuneiform writing
system," Visible Language 15: 345-72.
Griffin, M. 1997. "The Senate's story," Journal of Roman Studies 87: 249-63.
Harris, W. V. 1989. Ancient literacy. Cambridge, MA: Harvard University Press.
Harris, W. V. 1996. "Writing and literacy in the archaic Greek city," in Strubbe, Thybout,
and Versnel 1996: 416-23.
Havelock, E. A. 1986. The muse learns to write: Reflections on orality and literacy from
antiquity to the present. New Haven: Yale University Press.
Hopkins, K. 1991. "Conquest by book," in M. Beard, A. K. Bowman, and M. Corbier (eds.),
Literacy in the Roman world. Journal of Roman Archaeology Suppl. 3. Ann Arbor, MI:
University of Michigan Press, 133-58.
Jeffery, L. H. 1990. The local scripts of archaic Greece. Rev. by A. W. Johnston. Oxford:
Clarendon Press.
Jeffery, L. H., and A. Morpurgo-Davies 1970. "Poinikastas and poinikazein," Kadmos 9:
118-54.
Johnson, W. A. 2004. Bookrolls and scribes in Oxyrhynchus. Toronto: University of Toronto
Press.
Immerwahr, H. R. 1990. Attic script: A survey. Oxford: Clarendon.
Kahl, J. 2003. "Die friihen Schriftzeugnisse aus dem Grab U-j in Umm el-Qaab," Chronique
d'Egypte 78: 112-35.
Kambitsis, S. 1985. Le Papyrus Thmouis 1, colonnes 68-160. Paris: Publications de la
Sorbonne.
Kroll, J. H. 1977. "An archive of the Athenian cavalry," Hesperia 46: 83-140.
Lassere, J.-M. 2005. Manuel d'epigraphie romaine. Paris: Picard.
LDAB = Leuven Database of Ancient Books. Available on line: http://ldab.arts.kuleuven.be.
Lewis, N. 1974. Papyrus in classical antiquity. Oxford: Clarendon Press.

Linders, T. 1988. "The purpose of inventories," in D. Knoepfler (ed.), Comptes et in-
ventaires dans la cite grecque: Actes du colloque international d'epigraphie tenu a
Neuchatel du 23 au 26 septembre 1986 en l'honneur de Jacques Treheux. Neuchatel:
Universite de Neuchatel. Faculte des lettres, 37-47.
a
Marichal, R. 1992a. "Les tablettes ecrire dans le monde romain," in E. Lalou, Les tablettes a
ecrire de l'Antiquite a l'epoque moderne. Turnhout: Brepols.
Marichal, R. 1992b. Les ostraca de Bu Njem. Tripoli: Jamahira Arabe libyenne populaire
socialiste.
Marrou, H.-1. 1965. Histoire de l'education dans l'Antiquite, rev. ed. Paris: Editions du Seuil.
Martin, A. 1989. "L'ostracisme athenien," Revue des Etudes Grecques 102: 124-45.
Meyer, E. A. 2004. Legitimacy and law in the Roman world: Tabulae in Roman belief and
practice. Cambridge: Cambridge University Press.
Millar, F. 1964. "The Aerarium and its officials under the Empire," Journal of Roman
Studies 54: 33-40.
Miller, S. G. 1995. "Old Metroon and old Bouleuterion in the classical agora of Athens," in
M. H. Hansen, K. A. Raaflaub (eds.), Studies in the ancient Greek polis. Stuttgart:
Steiner, 133-56.
Moreau, P. 1994. "La memoire fragile: Falsification et destruction des documents publics
au Ier s. av. J.-C," in Demougin 1994: 121-47.
Munn, M. H. 2006. The Mother of the Gods, Athens, and the tyranny of Asia: A study of
sovereignty in ancient religion. Berkeley: University of California press.
a
Nicolet, C. 1994. "A la recherche des archives oubliees: Une contribution l'histoire de la
bureaucratie romaine," in Demougin 1994: v-xvii.
Posner, E. 1972. Archives in the ancient world. Cambridge, MA: Harvard University Press.
Powell, B. B. 1991. Homer and the origin of the Greek alphabet. Cambridge: Cambridge
University Press.
Pritchett, W. K. 1996. Greek archives, cults and topography. Amsterdam: Gieben.
Ridley, R. T. 2003. The emperor's retrospect: Augustus' Res gestae in epigraphy, historiography
and commentary. Leuven: Peeters.
Ruze, F. 1988. "Aux debuts de l'ecriture politique: Le pouvoir de l'ecrit dans la cite," in
Detienne 1988: 82-94.
Scheid, J. 1990. Romulus et ses freres: Le college des freres arvales, modele du culte public dans
la Rome des empereurs. Bibliotheque des ecoles frarn;:aises d'Athenes et de Rome 275.
Rome: Ecole frarn;:aise de Rome.
Scheid, J. 1994. "Les archives de la piete: Reflexions sur les livres sacerdotaux," in
Demougin 1994: 173-85.
Shear, T. L. 1995. "Bouleuterion, Metroon, and the archives at Athens," in M. Hansen and
K. Raaflaub (eds.), Studies in the ancient Greek polis. Stuttgart: Steiner: 157-90.
Sherwin-White, S. M. 1985. "Ancient archives: The Edict of Alexander to Priene, A re-
appraisal," Journal of Hellenic Studies 105: 69-89.
Sickinger, J. P. 1994a. "Inscriptions and archives in classical Athens," Historia 43:
286-96.
Sickinger, J.P. 1994b. "Review of R. Thomas, Literacy and orality in ancient Greece.
Cambridge: Cambridge University Press, 1992," Classical Philology 89: 273-78.
Sickinger, J. P. 2002. "Literacy, orality, and legislative procedure in classical Athens," in
I. Worthington and J.M. Foley (eds.), Epea and grammata: Oral and written com-
munication in ancient Greece. Mnemosyne Suppl. 230. Leiden: Brill, 147-69.
Sickinger, J.P. 2004. "The Laws of Athens: Publication, preservation, consultation," in

E. M. Harris, L. Rubinstein (eds.), The law and the courts in ancient Greece. London:
Duckworth, 93-109.
Skeat, T. C. 1982. "The length of the standard papyrus roll and the cost-advantage of the
codex," Zeitschrift fur Papyrologie und Epigraphik 45: 169-76.
Skeat, T. C. 1995. "Was papyrus regarded as 'cheap' or 'expensive' in the ancient world?"
Aegyptus 75: 75-93.
Small, J. P. 1997. Wax tablets of the mind: Cognitive studies of memory and literacy in classical
antiquity. London: Routledge.
Solin, H., and 0. Salomies (eds.) 1995. Acta colloquii epigraphici latini. Societas scientiarum
Fennica. Commentationes humanarum litterarum 104. Helsinki: Academia scien-
tiarum Fennica.
Stroud, R. S. 1979. The Axones and Kurbeis of Drakon and Solon. Berkeley: University of
California Press.
Strubbe, J. H. M., R. A. Tybout, and H. S. Versnel (eds.) 1996. Energeia: Studies on ancient
history and epigraphy presented to H. W. Pleket. Amsterdam: Gieben.
Talbert, R. J. A. 1984. The Senate of imperial Rome. Princeton: Princeton University Press.
Thilo, R. M. 1980. Der Codex accepti et expensi im romischen Recht: Bin Beitrag zur Lehre von
der Litteralobligation. Gi:ittingen: Muster-Schmidt.
Thomas, R. 1989. Oral tradition and written record in classical Athens. Cambridge: Cam-
bridge University Press.
Thomas, R. 1992. Literacy and orality in ancient Greece. Cambridge: Cambridge University
Press.
Turner, E. 1977. The typology of the early codex. Philadelphia: University of Pennsylvania
Press.
UNESCO 1977. Statistics of educational attainment and illiteracy, 1945-74. Statistical Re-
ports and Studies 22. Paris: UNESCO.
Vandorpe, K. forthcoming. "Archives and dossiers," in Oxford handbook of papyrology.
Verhoogt, A. M. F. W. 1998. Menches, Komogrammateus of Kerkeosiris: The doings and
dealings of a village scribe in the late Ptolemaic period (120-110 B.C.). Leiden: Brill.
Williamson, C. 1987. "Monuments of bronze: Roman legal documents on bronze tablets,"
Classical Antiquity 6: 160-83.
Williamson, C. 1995. "The display of law and archival practice in Rome," in Solin and
Salomies 1995, 239-51.
Woodard, R. D. 1997. Greek writing from Knossos to Homer: A linguistic interpretation of
the origin of the Greek alphabet and the continuity of ancient Greek literacy. Oxford:
Woolf, G. 1998. "Monumental writing and the expansion of Roman society in the early
Empire," Journal of Roman Studies 86: 22-39.
Zoppi, C. 1996. "Le cretule di Selinunte," in M.-F. Boussac and A. Invernizzi (eds.),
Archives et sceaux du monde hellenistique. Bulletin de correspondance hellenique Suppl.
29. Athenes: Ecole franc;aise d' Athenes, 327-40.
CHAPTER 29
TIMEKEEPING
ROBERT HANNAH
TOWARD A SOCIOLOGY OF TIME
For most of the thirty years that followed the publication of Sharon Gibbs' still
fundamental Greek and Roman Sundials (1976), work on the early technology of
timekeeping remained predominantly descriptive. Indeed, Turner commented in
1993 that the history of this technology could only just then start to rise above the
level of description to questions of synthesis and interpretation (Turner 1993: ix-
xi); this process has been slow. In related domains, such as the organization and
division of time, or the representation of time, a shift had already been made from
essential, but basic, data collection to studies of the philosophy, psychology, and
sociology of time. This progress grew from a century of work on a social anthro-
pology of time, starting with the research of scholars like Durkheim (1915) and
Evans-Pritchard (1940). Nilsson (1920) to a degree represented a classicist's con-
tribution to this new development, although his study tended still to be more
descriptive than conceptually analytical. The period between the 1950s and the 1970s
then witnessed a particularly rich maturation of this initial growth in what may be
termed the study of "social time." Social anthropologists turned sociologists, such
as Goody (1968) and Bourdieu (1990), increasingly problematized and theorized
this notion of social time, in the process extending the scope of research from so-
called primitive societies to their own western cultures. By the end of the twentieth
century, Gell (1992) could successfully synthesize almost a hundred years of eth-
nographic and sociological work on time among modern societies.
But it was only from the mid-198os that we have witnessed the proliferation of
studies on the sociology of Greek and Roman time. Zerubavel (1985), with his
investigation of the week in antiquity, was a pioneer. Since then, increasing
TIMEKEEPING 741
numbers of researchers have displayed an interest in the social constructs of time as

evinced by the literature, documentary texts, art and philosophy of the Greeks and
Romans (e.g., Hannah 1986a, 1986b, 1989, 1993a, 1993b, 1997; Bettini 1991: Rupke
1995; Dohrn-van Rossum 1996; Csapo and Miller 1998; Turetzky 1998; Darbo-
Peschanski 2000; Rossiter and Suksi 2003). Yet the perceptions of time encapsulated
in the mechanisms created to measure it generally escaped scholarly interest un-
til early in the twenty-first century. Indicative of a shift in attitude has been the
treatment given to the discovery of the most significant addition to Gibbs' corpus,
the physical remains of the horologium Augusti in Rome. The excavator's detailed, if
controversial, discussion of the scientific aspects of this sundial (Buchner 1982,
1993-1994; cf. Schutz 1990) was very soon accompanied by cultural readings of the
monument (Wallace-Hadrill 1987; Beck 1994). Turner (1990) and Lippincott (1999)
also showed the way, but because both are exhibition catalogues, they still neces-
sarily depend on a taxonomy of time instruments and thereby maintain an artificial
separation of these devices from their broader social contexts. The instruments
of timekeeping are becoming less ends in themselves and more gateways into the
ancient mentalite about time. This trend is worth pursuing, as it is the instruments
themselves that are most likely to tell us what ordinary people thought on a daily
basis about time.
NATURAL CYCLES OF TIME
The needs of the human body-for food, relief, sleep-provide the simplest
marker of the passage of time. This is as true today as it was in the ancient world,
however much we allow technology to order our days and so to control these basic
demands. The story of how the Greeks and Romans learned to measure time can
seem to be a history of the gradual distancing of people from nature, of their
"denaturization" (Turner 1990: 20). The development of sundials and similar
means of dividing the day and year into artificial segments can be paralleled by the
development of calendars that achieve the same purpose and effects. Yet in both
instances the abstraction from natural measures of time never creates a complete
divorce. Underneath the complexities of built timepieces such as the Tower of the
Winds in Hellenistic Athens or of temporal constructs such as the Julian calendar
in late Republican Rome lie still visible vestiges of the natural cycles and phe-
nomena from which they ultimately derived.
Underlying most of the ancient means of keeping time is the use of the celestial
bodies-the sun, the moon, and the stars-either directly through observation or
indirectly through calculated schemata based originally on observation, to enable
people to track their lives and activities from day to day, month to month, year to
year. Societies privileged the dawn and evening observations of celestial bodies, no
doubt because these were the pivotal periods when people shifted from daytime
activities to those of nighttime or vice versa (Goody 1968: 32). Even the political
world could be subservient to this very simple, natural mechanism. The Assembly
in classical Athens, for instance, began its regular meetings at sunrise (Aris-
tophanes, Ach. 19-20), and the Roman Senate would usually meet after sunrise and
cease business by sunset, to judge from the fact that decrees passed outside those
times were deemed invalid (Aulus Gellius, NA 14.7.8).
The stars observed at sunrise and sunset provided a useful sequence of first and
last visible risings and settings. Our earliest certain indications of the Greeks using
this timekeeping method occur around 750-700 B.C., when Homer and Hesiod
embedded such star lore in poems of quite different types, but with the same
purpose: to indicate times of the year significant for agriculture or seafaring. Other
seasonal indicators could also be used to triangulate the proper time for agricultural
activity, such the migration of birds (Hannah 2005: 20-22). Hesiod mentions just
ten observations of the risings, settings, or culmination of five stars or star groups.
While this may seem a very small number of observations over a year, it has rightly
been pointed out that Hesiod's economical set of data still provides a functional
safety net of observations over the crucial parts of the agricultural year, from sowing
to reaping (Reiche 1989). It has been argued that astronomical observations were
used even earlier, in the Late Bronze Age, particularly in Crete, where it has been
proposed that some buildings were oriented toward solar or stellar phenomena on
the horizon (Henriksson and Blomberg 1996, 1997-1998; Blomberg and Henriksson
2003). But while the orientations seem convincing, in the absence of decipherable,
contemporary documentary evidence it is difficult to see what purpose they served.
PARAPEGMATA
Hesiod's data may have derived from a dedicated astronomical poem-an Astro-
nomica was attributed to him-and it would be interesting to know how this
compared with Egyptian and Babylonian texts, which are earlier, more extensive,
and more systematized than what Hesiod gives us (Belmonte 2003; Parker 1974;
Hunger and Pingree 1989). In the Near East, we have a sense of real star "calendars"
into which events could be slotted. We have nothing of this sort in Greece until
the fifth century B.C. This type of calendar-perhaps better called an almanac-
appears in the form of an instrument called a parapegma. Its invention is connected
with two Athenians, Meton and Euctemon, and some of the leading astronomers of
the classical period were involved in its further refinement, notably Eudoxus and
Callippus. Excerpts from these parapegmata were included in later literary com-
pilations. In addition, fragmentary examples have been discovered in public urban
contexts across the Mediterranean, most dating to the Hellenistic period (Lehoux
TIMEKEEPING 743
2000, 2007). Technologically, they are very simple instruments; surviving examples
consist of sets of inscribed stone tablets, on which someone (it is odd that we do not
know who) had to move a peg manually one day after another throughout the year,
through a series of 365 holes. Alongside some of the holes were chiseled the star
observations for the day (Hannah 2001b: 76-79, 2005: 59-61; Taub 2003: 20-27).
How the observations were made, and with what instrumentation beyond the
naked eye, we do not know. Meton's teacher, a metic called Phaeinus, observed the
solstices from Mount Lycabettus in Athens ( [Theophrastus], Sign. 4), while Meton
himself (fl. 432 B.C.) set up stelai and recorded the solstices (Aelian, VH 10.7).
Meton also erected an instrument called a heliotropion on the Pnyx in Athens
(scholiast on Aristophanes, Av. 997). The very name of this instrument suggests that
it had something to do with the solar tropics of the solstices. Since astronomical
observations were regularly oriented to the horizon at this time, the heliotropion is
likely to have been a device aligned to an horizon rising or setting point, rather than
a sundial casting the shadow of a noon sun, and perhaps it pinpointed the place of
the solstices rather than assisted in discovering their time, which may already have
been a "given" from Babylonian astronomy (Hannah 2005: 52-54; Bowen and
Goldstein 1988). It is remarkable that the rising of the summer solstice sun, when
seen from the Pnyx, occurs near the peak of Lycabettus, so natural topographical
features might have assisted in the alignment of the heliotropion.
There is a good deal of evidence for the parapegma ofEuctemon (fl. 432 B.C.), as
it was one of the most popular star almanacs used in later times. The most extensive
quotation appears in the collection of parapegmata attached to the first-century
B.C. Eisagoge of Geminus. There, the observations from Euctemon's and others'
parapegmata are arranged under the artificial signs of the zodiac (i.e., under solar
months), recording that on the nth day of the sun's passage through a given zodiacal
sign, this star or that rose or set at dawn or dusk, and occasionally coincided with
certain weather conditions. This organization probably postdates Euctemon by a
century or more, since artificial zodiacal signs of 30 degrees are not attested in Greek
texts before the third century B.C., while he himself arguably arranged the obser-
vations simply according to the number of days between them (Bowen and Gold-
stein 1991; Hannah 2002). We can extract from this compilation the observations
associated only with Euctemon, and so gain a "text" which reflects part, if not all, of
the original. Overall, this text gives a greatly increased number of observations over
Hesiod's-42 observations of 15 stars or star groups-together with notices of the
solstices and equinoxes. Even the points in between, which are nowadays termed
"quarter days," can be found, serving as markers for significant farming activity
across the end of one season and the start of another (Hannah 2005: 59-70).
The accumulation of so many star observations was not dependant upon lit-
eracy, since oral societies also engaged in such activity. Nevertheless, writing
probably enabled the Greeks to decontextualize these data and to place them in
other quite abstract contexts, such as parapegmata, which in turn enabled them to
discover more precise temporal relationships between these phenomena and the
cycle of the sun, and indeed to understand that cycle much better (Hannah 2001a).
Others further up the chain of interpretation, so to speak, were helped by the

parapegmata to organize their activities in time with the seasonal year. Columella
(Rust. 9.14.12) refers explicitly to the star calendars of Meton and other astronomers
being adapted to public (religious) festivals. A papyrus from Hibeh in Egypt (P Hib.
27), dating to about 300 B.C., preserves a festival calendar for the temple ofNeith at
Sais, which lists various religious festivals alongside astronomical and meteoro-
logical observations. The astronomical observations in this case were derived from
the parapegma of Eudoxus, who worked a generation after Euctemon, and were
structured within a scheme of 12 zodiacal months and then further worked into the
native Egyptian calendar.
THE ANTIKYTHERA MECHANISM
Although the development of the lunisolar calendars of the Greeks and of the
subsequent Julian calendar of the Romans is extremely important in the history of
timekeeping as a whole, it demonstrates little further technological dependency,
except perhaps in the matter of correcting the leap-year error that was embedded
with the initial adoption of the Julian calendar in Rome (Buxton and Hannah
2005). Further calendaric development is therefore not the subject of this study in
any detail (see further Hannah 2005; Blackburn and Holford-Strevens 1999: 669-76,
712-16). Nonetheless, it is worth noting the invention in the fifth century B.C., in
both Greece and Persia, of the 19-year "Metonic" cycle to help synchronize the
otherwise incommensurate lunar and solar cycles, because this overarching cycle is
incorporated into one of the most intriguing timekeeping instruments to have
survived antiquity. This is the so-called Antikythera Mechanism, a highly sophis-
ticated geared instrument recovered from a shipwreck off the coast of the Aegean
island of Antikythera in 1900 and dating at the latest to the mid-first century B.C.
(Price 1974; Zeeman 1986; Bromley 1986). This unique artifact has been the object of
renewed critical study in the past twenty years or so, culminating in a physical
reconstruction based on a mixture of autopsy of the badly corroded mechanism
and educated conjecture (figure 29.1), and in a better understanding of the complex
gearing (without the need, it may be noted, for a differential geared train, which had
been previously assumed and yet was otherwise unique from antiquity). The in-
strument managed to correlate, in an ingenious system of geared wheels, the mo-
tions certainly of the sun and the moon, and arguably of the five planets known to
antiquity, in epicyclic motion through the zodiac, all timed against the Egyptian
calendar and a parapegma (Bromley 1990; Wright 2002, 2003a, 2003b, 2004, 2005a,
2005b, 2005c, 2005d, 2006a, 2006b). New reflectance imaging techniques and high-
resolution X-ray tomography have begun to provide new information, indicating
that the instrument could also be used to compute eclipses (Freeth et al. 2006).
TIMEKEEPING 745
Figure 29.1. Reconstruction of the Antikythera Mechanism by Michael Wright. (Photo-

graph by M. Wright.)
There is evidence within the mechanism that it was a composite, cannibalizing

parts of older devices, and thereby suggesting the existence of other similarly
elaborate astronomical instruments (Wright 2006b) . This device affords us a
glimpse of the class of sophisticated instruments that replicated the motions of the
celestial bodies and were so admired in antiquity, such as Archimedes' sphere
(Cicero, Rep. 1.14.21-22).
We do not know for what purpose the Antikythera Mechanism was devised. It
was originally identified as a "calendar computer" (Price 1974), and it has often
subsequently been misinterpreted as a navigational aid, simply because it was re-
covered from a shipwreck. It is probably better termed a planetarium, which offered
the facility to measure time in terms of both the Egyptian calendar on the front of
the dial and of the Callippic lunisolar cycle of 76 years (four Metonic cycles less a
day) on the back, a combined system used later by Ptolemy in the Almagest. Indeed,
correlations with local civil, lunar calendars of the Greek world may also have been
possible, particularly if those calendars were still aligned with the moon, as now
seems proven (Wright 2005b; Buxton and Hannah 2005: 302). The parapegma-if
that indeed is the appropriate term for the attached star-observation list-seems to
be one of a kind, corresponding in its details to no single example known. Eclipse
recording or forecasting are also likely functions, but to what specific end is not yet
clear. Banal though it may now seem for so complex an instrument, horoscopal
astrology has been suggested as one of its possible functions, since it would permit
the rapid calculation of the positions of all the major planetary bodies essential to
ancient astrology, positions which are recorded with a remarkable degree of ac-
curacy in surviving tables from the imperial Roman period (Neugebauer 1941-1943:
209-50; cf. Neugebauer and Parker 1969: 225-35).
Certainly parapegmata were associated with forecasting, although of weather

rather than other events. Astrometeorology had a long history throughout antiq-
uity (Taub 2003: 15-69), and indeed it has been proposed that before the first
century B.C. the Greeks were interested not so much in measuring time per se as in
observing the orderly sequences of "omen events" such as star-rise and star-set,
equinoxes and solstices, on which the sequences of agriculture and religion relied
(Price 1975). This argument is attractive, but some caution is warranted: "omen-
events" need careful definition, if we are not to do an injustice to the parapegmata,
some of which (particularly the early Greek ones) expend much more space on pure
fixed-star phase prediction than they do on the meteorological forecasts that might
be read as a causal result of those star phases. The cultural context of astro-
meteorology in both the Near East and Greek worlds is now being given serious
attention (Rochberg 2004; Taub 2001, 2003; Lehoux 2000, 2007).
SUNDIALS
The risings and settings of stars have a very long history as a timekeeping method,
continuous from the Babylonian period to the Middle Ages. By the latter period,
Christian monks were using the stars to signal the hours for prayer at night
(McCluskey1998). It is apparent from the remains of the horologiumAugusti, where
notice is preserved of the Etesian Winds, that the astrometeorological lore of the
parapegmata not only persisted into the period of the Julian calendar, but also could
be incorporated into sundial technology.
The shadow-casting gnomon may have found its way into Greece just before 500
B.C., when it is associated with the philosopher Anaximander (Diogenes Laertius
2.1). Herodotus (2.109) seems to refer to it later in the fifth century, saying that the
Greeks gained knowledge of the polos and gnomon as well as the 12-part division of
the day from the Babylonians (Lloyd 1988: 34-36). While there is nothing to suggest
that it was used by ordinary people at that stage (Gibbs 1976: 6-7), nevertheless the
use made by Greek comic playwrights from the late fifth century B.C. of the idea of
reckoning the time for a meal by gauging the length of a person's shadow (Aris-
tophanes, Eccl. 651-52; Athenaeus, 1.8b-c, 6.243a) demonstrates a familiarity with
the principle underlying the gnomon. How sophisticated this human sundial was in
reality is unclear. Although fixed shadow lengths throughout the year would lead to
a variable time for the same meal, this may well have suited people's biological
clocks, since the length of the day varied throughout the year, and meals would
literally have been movable feasts. The moment it was decided that a fixed hour
should serve for mealtimes, or for any other activity, a tension was created between
people's physical needs and mechanical demands. This is the point behind the
TIMEKEEPING 747
complaints against sundials as slave drivers retailed by Aulus Gellius (NA 3.3.4-5)
and Alciphron (3,1) in the second century A.D., both of whom rely on a comic topos
of at least three hundred years' vintage and so demonstrate the longevity of the
issue.
At some stage, a type of sundial was developed that showed the passage of the
shadow of the gnomon into each of the 12 zodiacal signs in the course of a year. A
liberal reading of Vitruvius' attribution to Eudoxus of a type of sundial called the
arachne (spider's web) (De arch. 9.8.1) has been seen as a reference to the zodiacal
sundial and so has placed its invention in the fourth century B.C. (Ardaillon 1900:
257; Rehm 1913: 2418-19), but this interpretation is misguided. Although the arachne
may well have derived its name from a weblike network of hour and season lines
across its face (Gibbs 1976: 60-61), it is quite possible that the first dials to show the
sun moving through the zodiac were much simpler, similar to a dial found on Chios
dating perhaps to the second century B.C. (Hunt 1940-1945: 41-42). This is just a
small, flat north-south meridian line, with the summer and winter solstices marked,
together with the divisions between the zodiacal signs. If, on the other hand, the
term arachne is correctly understood, and its association with Eudoxus reasonably
founded, then it demonstrates the increased refinement of the divisions of the
month and day that pertained by the fourth century B.C.
It is easy to imagine how helpful the basic meridian type of sundial would have
been in marking out the passage of the seasons through the year, and even of
dividing the year into smaller manageable chunks, the 12 zodiacal months. But how
useful would it be on a day to day basis? Unfortunately, at the latitude of Athens in
Greece (38° N), the shadow measured by even a human-sized gnomon changes only
a little from day to day, too little to help in the distinction of one day from the next,
the more so if the gnomon is even smaller in scale (the Chian dial may have had one
only 72 cm in height). Such a sundial might be useful only over longer periods of
several days or a month, if the events being timed by the dial are spaced well apart in
the year. Still, the movement of the sun across the sky in the course of the day offered
a better opportunity to organize events within the day, and we can see other
sundials taking much more advantage of this feature.
The earliest surviving Greek sundial, dating to the third century B.C., comes
from the small site of Delos, where 25 dials have been excavated. Another 35 have
been unearthed in the town of Pompeii. Between them, the two sites demonstrate
the popularity of the instrument from the Hellenistic period into the Roman
Empire. The principal kinds of sundial available to the Greeks and Romans were the
hemispherical, the cylindrical, the conical, and the planar. The hemispherical type
of sundial was the most labor-intensive and difficult to construct, as it entailed
carving out an even hemisphere of stone. But it was also the simplest to mark out,
because it captured the celestial dome on a matching concave surface. Its gnomon
hung out over the hollow hemisphere (figure 29.2). The conical type, with the
related but much less common cylindrical form, was a variation on the hemi-
spherical, representing a simpler, partial hollowing out of the stone block. At least
Figure 29.2. Spherical sundial, from Aphrodisias, Sel~uk Archaeological Museum no. 375.
(Photograph by R. Hannah.)
in the surviving archaeological record, it is easily the most popular type. This may
occasion some surprise, given the apparent complexity of its theory and the greater
difficulties in marking out the requisite interior lines, as these are now projected on
to an awkwardly shaped curved surface, but it is likely that the conical was much
easier to construct out of stone (the preferred material) than the hemispherical
type, while its theoretical underpinning was kept to a minimum-and indeed
probably not understood by many of the makers, to judge by their inaccuracies
(Gibbs 1976: 17, 74-75). The ubiquity of such sundials in both the Greek and Roman
worlds is well captured in a quip, attributed to the Roman emperor Trajan early in
the second century A.O., that makes a real dial of the human face (Anthologia
Palatina 11.418): "If you put your nose facing the sun and open your mouth wide,
you'll show all the passersby the time of day."
The flat-plane type of sundial, which occurs in both horizontal and vertical
forms (figure 29.3), is the easiest to construct but the most difficult to mark out.
The difficulty arises from the projection of the hemispherical dome of the sky on
to a completely flat surface. A shadow tracking the movement of the sun through
the year is cast by a gnomon which is usually stuck perpendicularly into the
flat surface of the sundial. Vitruvius (De arch. 9.7) provides an analemma for the
dial's construction, which is now well understood despite the awkwardness of
his text, and extendable to a variety of dial types (Drecker 1925: 3-4; Evans 1999:
247-51).
TIMEKEEPING 749
.,' I .,'
~
.. .
:~ ·'='
-~~
·•
Figure 29.3. Tower of the Winds, Athens, elevation. (After J. Stuart and N. Revett, The
Antiquities of Athens [London: Priestley and Weale, 1825], vol. 1, pl. 14, fig. 1.)
HOURS
Sundials helped inculcate into society the concept of the seasonal, or unequal,
hour. For most purposes in antiquity, such hours were the norm. From Egypt came
the notion that each day or night could be divided into 12 hours from sunrise to
sunset, and another 12 from sunset to sunrise (Parker 1974: 53; Quirke 2001: 42) .
Since daytime and nighttime change in length with the seasons, the length of each
hour therefore changed also according to the season. Only at the spring and
autumn equinoxes were the hours equal through the whole day. This gives us what
is called the "equinoctial" hour, which we are used to because mechanical clocks
and hourglasses have required them since their invention in the Middle Ages. But
only astronomers tended to use them in antiquity, as can be seen in the Hibeh
Papyrus (P Hib. 27) of about 300 B.C., where they are subdivided down to ½, 5th of
an hour.
The shorter winter hour crops up in a joke in Plautus (Pseud. 1304), indicating
that the concept of the hour had already made its way to Rome by the early second
century B.C., perhaps in the train of actual Greek sundials. The rarer "half-hour" first
appears in a fourth century B.C. comedy by Menander (fr. 1015), and is inscribed on
just a couple of surviving Greek sundials (Gibbs 1976: 226 no. 3007, 239 no. 3020;
1979: 45, fig. 4). In his instructions for making sundials, Cetius Faventinus in the third
century A.D. remarked (ironically to us moderns) that people were "in too much of a
hurry to want to know more than what hour it is" (Pattenden 1979: 204, 207-8).
Nonetheless, the concept of the half-hour was already sufficiently well known in the
first century A.D. to appear in Christian apocalyptic literature (Rev. 8:1).
The accuracy of a sundial can be checked for its particular locality, because the
solstice and equinox lines should correspond to a specific latitude. The lines on the
dials from Delos show that most were made to be used there, whereas the dials from
Pompeii show far less accuracy and do not suit its latitude well. This inaccuracy
means that the days of the solstices and equinoxes in Pompeii were wrongly
marked. On the other hand, the daily hour lines on Pompeian dials do match reality
reasonably well, so it may be that among the Pompeians there was more emphasis
on the time of day and on business within each day than on the time of year (Gibbs
1976: 90-92). Pliny's story (HN 7.214) of earlier Romans living for 99 years in
ignorance of the inaccuracy of a sundial transported from Catania to Rome in 263
B.C. depends for its force on the dial's engraved lines not agreeing with the hours
(horas) appropriate for Rome. Yet, ironically, the four-degree difference in latitude
between the two cities affects the hours of the day far less than it does the days of the
year: in the summer there would be a negligible error of 0.07 hours in the day,
whereas the gnomon's shadow in Rome would never have fallen on the dial's
summer solstice line, and would have fallen on the winter solstice line twice (Gibbs
1976: 96 n. 25). We do not know how the Romans discovered that the dial was
inaccurate for them; the first water clock was not set up there until 159 B.C., so they
do not seem to have relied on another type of instrument to tell them the correct
time of day. On the other hand, perhaps it really was the time of year that was
recognized as being inaccurately measured by the sundial, and the traditional
translation of Pliny's horas as "hours," should be changed to the less common but
alternate meaning "seasons" (as in Pliny, HN 9.107, 12.15, 17.132). In that case, he
would be correctly representing the problem of the Sicilian sundial in Rome. If
Roman comic playwrights in the early second century B.C. could raise a laugh at the
notion of a town being "stuffed with sundials" that controlled ordinary life (Aulus
Gellius, NA 3.3.5), then, even if the joke was a transplant from Greek New Comedy,
it is possible that the presence in Rome itself of more and more sundials provided
the observant with evidence that the Catania dial was unreliable.
Certainly we gain a distinct impression of a careful parceling out of the hours of
the day among the Romans. In Rome, the accensus had the job of announcing when
it was the third, sixth, and ninth hours of the day (Varro, Ling. 6.89; Pliny, HN
7.212), and a few sundials survive with these hours specifically marked out (Gibbs
1976: 300, no. 3080 for a list). In addition, the hour lines on dials were sometimes
TIMEKEEPING 751
numbered. If this was done in Greek, the letters of the alphabet were used for
numbers (see figure 29.2). A Greek epigram (Anthologia Palatina 10.43) puns neatly:
"Six hours are sufficient for work. But the rest, when set out in letters, say 'Live!' to
mortals." The first six hours of the day, to noon, were devoted to work, but the next
four to leisure. This is a simple play on the letters for the hours 7, 8, 9, and 10, which
were Z, H, 0, and I read together, they formed the word "Live." The satirist Martial
(4.8) provides sharper definition of the day: the first two hours were occupied by
the salutatio between patron and clients; at the third hour the law courts opened;
work throughout the city lasted till the end of the fifth hour, followed by a rest at the
sixth, and a complete end to work at the seventh; the eighth hour was spent at the
gymnasium; dinner came at the ninth; the tenth, the poet hopes, is when the em-
peror will read this latest book of epigrams! These hours may represent the norm,
but variations certainly occurred. Retirement, not surprisingly, brought a more re-
laxed timetable for the rich, as Pliny the Younger (Ep. 3,1) demonstrates in his
description of the typical day of the elderly but very fit Spurinna. Allowance could
also be made for the variability of the seasons, with the time for bathing being
shifted from the ninth hour in winter back to the eighth in summer.
PORTABLE DIALS
As we have seen, already by the early second century B.C. the Romans found humor
in the idea of a town being full of sundials, and indeed they were everywhere-in
public squares, temples, town houses, and country villas. They were also used up
and down the length and breadth of the Roman world, from Spain to Greece, and
from Africa to Germany. Under the Empire they were even in people's hands,
miniaturized and portable, the ancient equivalent of the modern pocket watch. The
earliest portable sundial that we have is from Herculaneum. It is known as the
"Ham Dial," because its distorted bronze plate looks just like a small leg of ham. A
spike on one side threw a shadow onto a series of crisscrossing lines on the plate,
from which one could read the hour of the day. Other portable dials are regularly
circular in shape, or cylindrical, and some come with extra plates to suit different
latitudes (Price 1969; Arnaldi and Schaldach 1997; Wright 2000). One small dial of
perhaps mid-third century A.D. date consists of just two plates and a gnomon and
yet permits the reading of the time of day anywhere between latitudes 30° and 60°
N; 30 locations are specifically listed (Oxford, Museum for the History of Science,
inv. no. 51358; Price 1969: 253-56; Schaldach 1998: 45-47; cf. figure 31.3).
These dials foreshadow the astrolabe, the portable timepiece par excellence of
the Middle Ages. While no ancient astrolabe survives, the instrument was most
likely a late Greek invention. It utilized stereographic projection to represent the
celestial hemisphere two-dimensionally, a method of representation described by
Ptolemy in his Planisphere in the second century A.D. and perhaps known to
Hipparchus in the second century B.C. The earliest treatise on the astrolabe, by
Philoponus in the sixth century A.D., reflects an earlier one by Theon in the fourth
century A.D. (Price 1957: 603-9; Neugebauer 1975: 868-79; Evans 1999: 253-54; Taub
2001: 920-21).
WATER CLOCKS
The major timekeeping mechanism in the classical world, the water clock, initially
did not rely on emulating the motions of celestial bodies; but even it eventually was
adapted to do so. In origin this was not a clock in the strict sense, but a simple egg-
timer type of mechanism, measuring the set times for certain activities by a fall or
rise in water level. The earliest surviving water clock comes from Egypt and dates to
about 1400 B.C., while the earliest mention of such an instrument there is from the
late 16th century B.C. The Egyptians may well have borrowed the mechanism from
the Babylonians (Turner 1990: 58, no. 49; Cotterell et al. 1986; Spalinger 1996). In the
Greek world the term klepsydra (water thief) was borrowed from a device which
worked like a large pipette (e.g., Aristotle, Gael. 294b.14-30; [Aristotle], Pr. 914b,
Lewis 2000: 344, fig. 1), and it was applied to other containers, such as a bucketlike
vase with a hole near the base for the outflow, which was used from the fifth century
B.C. in Athens to time the speeches in the law courts ([Aristotle], Ath. Pol. 67;
Aristophanes, Ach. 694, Vesp. 93). Indeed, the klepsydra became synonymous with
the courts (Allen 1996: 158). A preserved example holds the equivalent of six
minutes' worth of water (Young 1939). Something of the same aversion to being
controlled by a time machine that we encountered with sundials is found also with
the water clock (Plato, Tht. 172c-e), but while the necessity for brevity imposed by
the klepsydra might have led to fallible judgments (Allen 1996: 159), it also ensured
equality through standardization, a fundamental characteristic of the democracy.
Other utensils could be used as simple timers also, such as the ubiquitous clay
oil lamp. This is occasionally described in Greek papyri from Egypt as containing
specific amounts of oil. The contexts suggest that fixed periods of time are intended
for the lamplight to burn during various magical rituals, notably in conjuring up
and holding on to spirits (Betz 1986: 172-82, 336). Pliny (HN 33.96-97) also notes
that work shifts for miners in Spain could be calculated by the exhaustion of their
lamp oil.
The standardization of time through the klepsydra extended also to the defi-
nition of the legal day, which was divided into a certain quantity of water, which in
turn was then further subdivided into various volumetric measures. This "day"
itself was equated with the shortest days of the year, those of the Athenian month
Poseideon in midwinter (Young 1939: 281). Klepsydrai of some kind were also used
TIMEKEEPING 753
to measure out the length of night watches in the military world, but there a
measure based on subdivisions of a standardized, winter day could not work for the
longer days of the campaigning season, so attempts were made to ensure that the
watches were equalized by adjusting the volume of water that the clocks held
through coating the inner surface of the klepsydra with varying layers of wax
(Aeneas Tacticus 22.24-25; Whitehead 2002: 158-60 ). The Romans continued to use
klepsydrai in political, legal, and military contexts (e.g., Caesar, B Gall. 5.13; Pliny,
Ep. 2.11.14).
Perhaps under the influence of these improvements in taking account of the
variability of the day and night through the year, in the mid-fourth century an
outflow water clock was built near the Heliaia in the agora in Athens, the social and
political center of the city, with a capacity of 1,000 liters, large enough to operate
uninterruptedly over the whole of a long summer's day (Camp 1986: 157-59). A very
similar example has been found at the Amphiaraeum at Oropus (Turner 1990: 62-
63, no. 65), where it may have timed rituals or performances at the nearby theater
(cf. Aristotle, Poet. 1451a8). Others have been found around the Greek world
(Dohrn-van Rossum 2003: 463).
From the Hellenistic period on, the inflow water clock was developed as an
alternative to the outflow type (Lewis 2000: 363-66). A float, whether in the outflow
tank or in the inflow, could be connected via gear wheels and even automated figures
to some means of displaying the passage of time as water entered or exited the tank.
This time display could be a simple cylinder marked with a scale of hour-lines,
although the representation of the uneven seasonal hours presented difficulties.
Alternatively, the passage of time could be tied explicitly to the motions of the
celestial bodies through the incorporation of images of the zodiac or stereographic
projections of the celestial sphere. Vitruvius describes sophisticated examples de-
vised by Ctesibius in the third century B.C., including the anaphoric clock, which
told the time via an automated representation of the sequential risings of the stars
(De arch. 9.8.4-15; Lewis 2000: 366; Rowland and Howe 1999: 117, 290-91; Dohrn-
van Rossum 1996: 26-27; Price 1957: 601-3). Physical remains of such complex
machinery are rare. Fragments have been found of the star-plate of two such water
clocks, both dating perhaps to the second century A.D., one from Salzburg, which
was monumental in scale at a diameter of over one meter, and another much
smaller one from Grand in Lorraine (Evans 1999: 251-53; Lewis 2000: 366-67).
In the Tower of the Winds in Athens (figure 29.3), built by Andronicus from
Cyrrha (Muller 2001: 43-44), there are traces on the floor of channels for piping to
some type of water clock that is no longer preserved. This 13 m high octagonal
building, well known in antiquity to judge from its mention by both Varro (Rust.
3.5.17) and Vitruvius (De arch. 1.6.4-7), was a tour de force of timekeeping instru-
ments. It incorporated not only a water clock inside but also nine sundials on its
exterior walls and an annex. New research suggests an earlier date in the second
century B.C. rather than the mid-first century B.C. On architectural grounds, eight
of the external sundials, one on each outside wall, are accepted as an original ele-
ment of the design, despite the absence of any mention of them by Vitruvius. This
omission caused Delambre, their great elucidator, some concern and led to the
suspicion that they may have been added afterward (Delambre 1817: 487-503). The
influential reconstruction by Noble and Price of the interior water clock is now
disputed (Noble and Price 1968; von Freeden 1983; Kienast 1993, 1997, 2005, and
personal communication).
The fact that there is still something to be learned and understood in a mon-
ument that has been so familiar to western Europeans since its depiction in the
eighteenth century is testament to the patchy state of our knowledge generally
about ancient developments in the technology of timekeeping. So while future
studies will undoubtedly continue the present trend to discuss the ancient artifacts
from a sociological perspective, there is room still for the basic scientific, analytical
description of those very same objects. Without that fundamental understanding,
any further interpretation is simply speculation.
REFERENCES
Allen, D. 1996. "A schedule of boundaries: An exploration, launched from the water-dock,
of Athenian time," Greece and Rome 43: 157-68.
Ardaillon, E. 1900. "Horologium," in DarSag 3: 256-64.
Arnaldi, M., and K. Schaldach 1997. "A Roman cylinder dial: Witness to a forgotten
tradition," Journal for the History of Astronomy 28: 107-117.
Beck, R. 1994. "Cosmic models: Some uses of Hellenistic science in Roman religion," in
T. D. Barnes (ed.), The sciences in Greco-Roman society. Apeiron 27. Edmonton:
Academic Printing and Publication, 99-117.
Betz, H. D. (ed.) 1986. The Greek magical papyri in translation, including the demotic spells.
Chicago: University of Chicago Press.
Belmonte, J. A. 2003. "The Ramesside star clocks and the ancient Egyptian constellations,"
in M. P. Blomberg, E. Blomberg, and G. Henriksson, (eds.), Calendars, symbols, and
orientations: Legacies of astronomy in culture. Uppsala: Uppsala Astronomical Ob-
servatory, 57-65.
Bettini, M. 1991. Anthropology and Roman culture: Kinship, time, images of the soul. Balti-
more: Johns Hopkins University Press.
Blackburn, B., and L. Holford-Strevens 1999. The Oxford companion to the year: An ex-
ploration of calendar customs and time-reckoning. Oxford: Oxford University Press.
Blomberg, M., and G. Henriksson 2003. "The Minoan peak sanctuary on Pyrgos and its
context," in Blomberg, Blomberg, and Henriksson 2003, 127-34.
Bourdieu, P. 1990. The logic of practice. Trans. R. Nice. Cambridge: Polity Press.
Bowen, A. C., and B. R. Goldstein 1988. "Meton of Athens and astronomy in the late fifth
century B.C," in E. Leichty, M. de J. Ellis, and P. Gerardi (eds.), A scientific humanist:
Studies in memory of Abraham Sachs. Philadelphia: The University Museum, 39-81.
Bowen, A. C., and B. R. Goldstein 1991. "Hipparchus' treatment of early Greek astronomy:
The case of Eudoxus and the length of daytime," Proceedings of the American Philo-
sophical Society 135: 233-54.
Bromley, A. G. 1986. "Notes on the Antikythera Mechanism," Centaurus 29: 5-27.
TIMEKEEPING 755
Bromley, A. G. 1990. "Observations of the Antikythera Mechanism," Antiquarian Horology

18: 641-52.
Buchner, E. 1982. Die Sonnenuhr des Augustus. Mainz: von Zabern.
Buchner, E. 1993-1994. "Neues zur Sonnenuhr des Augustus," Nurnberger Blatter zur
Archiiologie 10: 77-84.
Buxton, B., and R. Hannah 2005. "OGIS 458, the Augustan calendar, and the succession,"
in C. Deroux (ed.), Studies in Latin literature and Roman history, vol. XII. Collection
Latomus 287. Brussels: Latomus, 290-306.
Camp, J. M. 1986. The Athenian agora: Excavations in the heart of classical Athens. London:
Thames and Hudson.
Cotterell, B., F. P. Dickson, and J. Kamminga 1986. "Ancient Egyptian water-clocks: A
reappraisal," Journal of Archaeological Science 13: 31-50.
Csapo, E., and M. Miller 1998. "Democracy, empire, and art: Toward a politics of time and
narrative," in D. Boedeker and K. A. Raaflaub (eds.), Democracy, empire, and the arts
in fifth-century Athens. Cambridge, MA: Harvard University Press, 87-125.
Darbo-Peschanski, C. (ed.) 2000. Constructions du temps dans le monde grec ancien. Paris:
CNRS Editions.
Delambre, J. B. J. 1817. Histoire de l'astronomie ancienne. vol. 2. Paris: V. Courcier.
Dohrn-van Rossum, G. 1996. History of the hour: Clocks and modern temporal orders. Trans.
T. Dunlap. Chicago: University of Chicago Press.
Dohrn-van Rossum, G. 2003. "Clocks," in H. Cancik and H. Schneider (eds.), Brill's New
Pauly. Leiden: Brill, 3:458-64.
Drecker, J. 1925. Die Theorie der Sonnenuhren. Berlin: Vereinigung Wissenschaftlicher
Verleger, Walter de Gruyter.
Durkheim, E. 1915. The elementary forms of the religious life. Trans. J. W. Swain. London:
Allen and Unwin.
Evans, J. 1999. "The material culture of Greek astronomy," Journal for the History of
Astronomy 30: 237-307.
Evans-Pritchard, E. 1940. The Nuer: A description of the modes of livelihood and political
institutions of a Nilotic people. Oxford: Clarendon Press.
Freeth, T., Y. Bitsakis, X. Moussas, J. H. Seiradakis, A. Tselikas, E. Magkou, M. Zafeir-
opoulou, R. Hadland, D. Bate, A. Ramsey, M. Allen, A. Crawley, P. Hockley,
T. Malzbender, D. Gelb, W. Ambrisco, and M. G. Edmunds 2006. "Decoding the
ancient Greek astronomical calculator known as the Antikythera Mechanism," Nature
444 (30 November 2006): 587-91.
Gell, A. 1992. The anthropology of time. Oxford: Berg.
Gibbs, S. L. 1976. Greek and Roman sundials. New Haven: Yale University Press.
Gibbs, S. L. 1979. "The first scientific instruments," in K. Brecher and M. Feirtag (eds.),
Astronomy of the ancients. Cambridge, MA: MIT Press, 39-59.
Goody, J. 1968. "Time: Social Organization," in D. L. Sills and R. K. Merton (eds.),
International encyclopedia of the social sciences. New York: Macmillan, vol.I6: 30-42.
Hannah, R. 1986a. "The Emperor's stars: The Conservatori portrait of Commodus,"
Hannah, R. 1986b. "Et in Arcadia ego?-The finding ofTelephos," Antichthon 20: 86-105.
Hannah, R. 1989. " ... praevolante nescio qua ingenti humana specie ... : A reassessment of
the Winged Genius on the base of the Antonine Column," Papers of the British School
in Rome 57: 90-105.
Hannah, R. 1993a. "Alcumena's long night: Plautus, Amphitruo 273-76," Latomus 52:
65-74.
Hannah, R. 1993b. "The stars of Iopas and Palinurus," American Journal of Philology 114:
123-35.
Hannah, R. 1997. "The Temple of Mars Ultor and 12 May," Romische Mitteilungen 104:
374-86.
Hannah, R. 2001a. "From orality to literacy? The case of the Parapegma," in J. Watson
(ed.), Speaking volumes: Orality and literacy in the Greek and Roman world. Leiden:
Brill, 139-59.
Hannah, R. 2001b. "The moon, the sun and the stars: Counting the days and the years," in
S. Mccready ( ed.), The discovery of time. London: MQ Publications, 56-99.
Hannah, R. 2002. "Euctemon's parapegma," in C. J. Tuplin and T. E. Rihll (eds.), Science
and mathematics in ancient Greek culture. Oxford: Oxford University Press, 112-32.
Hannah, R. 2005. Greek and Roman calendars: Constructions of time in the classical world.
London: Duckworth.
Henriksson, G., and M. Blomberg 1996. "Evidence for Minoan astronomical observa-
tions from the peak sanctuaries on Petsophas and Traostalos," Opuscula Atheniensia
21: 99-114.
Henriksson, G., and M. Blomberg 1997-1998. "Petsophas and the summer solstice,"
Opuscula Atheniensia 22-23: 147-51.
Hunger, H., and D. Pingree 1989. MUL.APIN: An astronomical compendium in cuneiform.
Horn: Verlag Ferdinand Berger und Sohne.
Hunt, D. W. S. 1940-1945. "An archaeological survey of the classical antiquities of the
island of Chios carried out between the months of March and July, 1938," Annual of
the British School in Athens 41: 29-52.
Kienast, H. J. 1993. "Untersuchungen am Turm der Winde," Archiiologischer Anzeiger 1993:
271-75.
Kienast, H. J. 1997. "The Tower of the Winds in Athens: Hellenistic or Roman?" in
M. C. Hoff and S. I. Rotroff (eds.), The Romanization of Athens: Proceedings of an
international conference held at Lincoln, Nebraska (April 1996). Oxford: Oxbow, 53-65.
Kienast, H.J. 2005. "La Torre dei Venti di Atene," in E. Lo Sardo (ed.), Eureka! Il Genia
degli Antichi. Naples: Electa, 245-51.
Lehoux, D. R. 2000. "Parapegmata; or, Astrology, weather, and calendars in the ancient
world." PhD thesis, University of Toronto.
Lehoux, D.R. 2007. Astronomy, weather, and calendars in the ancient world: Parapegmata
and related texts in classical and Near-Eastern societies. Cambridge: Cambridge Uni-
versity Press.
Lewis, M. 2000. "Theoretical hydraulics, automata, and water clocks," in 6. Wikander
(ed.), Handbook of ancient water technology. Leiden: Brill, 343-69.
Lippincott, K. 1999. The story of time. London: Merrell Holberton and National Maritime
Museum.
Lloyd, A. B. 1988. Herodotus Book II. Vol. 3. Leiden: Brill.
McCluskey, S. 1998. Astronomies and cultures in early medieval Europe. Cambridge:
Millier, W. 2001. "Andronikos (1)," in R. Vollkommer (ed.), Kunstlerlexikon der Antike.
Munich and Leipzig: K. G. Saur, 1.43-44.
Neugebauer, 0. 1941-1943. "Egyptian planetary texts," Transactions of the American Phi-
losophical Society 32: 209-50.
Neugebauer, 0. 1975. A history of ancient mathematical astronomy. Berlin: Springer-Verlag.
Neugebauer, 0., and R. A. Parker 1969. Egyptian astronomical Texts, vol. 3. Providence, Rl:
Brown University Press.
TIMEKEEPING 757
Nilsson, M. 1920. Primitive time-reckoning: A study in the origins and first development of the
art of counting time among primitive and early culture peoples. Lund: Gleerup.
Noble, J. V., and D. J. de Solla Price 1968. "The water clock in the Tower of the Winds,"
Parker, R. A. 1974. "Ancient Egyptian astronomy," in F. R. Hodson (ed.), The place of
astronomy in the ancient world. London: The British Academy and Oxford University
Press, 51-65.
Pattenden, P. 1979. "Sundials in Cetius Faventinus," Classical Quarterly 29: 203-12.
Price, D. J. de Solla 1957. "Precision instruments: To 1500," in C. Singer, E. J. Holmyard, A.
R. Hall, and T. Williams (eds.),A history of technology, vol. 3. Oxford: Clarendon Press,
582-619.
Price, D. J. de Solla. 1969. "Portable sundials in antiquity, including an account of a new
example from Aphrodisias," Centaurus 14: 242-66.
Price, D. J. de Solla. 1974. Gears from the Greeks. Transactions of the American Philoso-
phical Society 64.7. Philadelphia: American Philosophical Society.
Price, D. J. de Solla. 1975. "Clockwork before the clock and timekeepers before time-
keeping," in J. T. Fraser and N. Lawrence (eds.), The study of time, vol. 2. Berlin:
Springer-Verlag, 367-80.
Quirke, S. 2001. The cult ofRa: Sun-worship in ancient Egypt. London: Thames and Hudson.
Rehm, A. 1913. "Horologium," in RE 8, 2416-33.
Reiche, H. A. T. 1989. "Fail-safe stellar dating: Forgotten phases," Transactions of the
American Philological Association 119: 37-53.
Rochberg, F. 2004. The heavenly writing: Divination, horoscopy, and astronomy in Meso-
potamian culture. Cambridge: Cambridge University Press.
Rossiter, J. J., and A. Suksi 2003. "The seasons: Greek and Roman perspectives." Special
issue, Mouseion 47.3.
Rowland, I. D., and T. N. Howe (eds.) 1999. Vitruvius: Ten Books on Architecture. Cam-
bridge: Cambridge University Press.
Rupke, J. 1995. Kalender und Offentlichkeit: Die Geschichte der Repriisentation und religiosen
Qualifikation von Zeit in Rom. Berlin: Walter de Gruyter.
Schaldach, K. 1998. Romische Sonnenuhren: Eine Einfuhrung in die antike Gnomonik.
Frankfurt am Main: Verlag Harri Deutsch.
Schutz, M. 1990. "Zur Sonnenuhr des Augustus auf dem Marsfeld," Gymnasium 97: 432-57.
Spalinger, A. 1996. "Some times," Revue d'Egyptologie 47: 67-77.
Taub, L. 2001. "Destini della scienza greca: eredita e longevita degli strumenti scientifici,"
in S. Settis (ed.), I Greci: Storia, Cultura, Arte, Societa. 4 vols. Turin: Einaudi, 3:
889-930.
Taub, L. 2003. Ancient meteorology. London: Routledge.
Turetzky, P. 1998. Time. London: Routledge.
Turner, A. J. 1990. Time. The Hague: Tijd voor Tijd Foundation.
Turner, A. J. 1993. Of time and measurement: Studies in the history of horology and fine
technology. Aldershot: Variorum.
von Freeden, J. 1983. OIKIA KYPPHI:TOY: Studien zum sogenannten Turm der Winde in
Athen. Rome: Giorgio Bretschneider.
Wallace-Hadrill, A. 1987. "Time for Augustus: Ovid, Augustus and the Fasti," in M. Whitby
and P. Hardie (eds.), Homo viator: Classical essays for John Bramble. Bristol: Bristol
Classical Press, 221-30.
Whitehead, D. 2002. Aineias the Tactician: How to Survive Under Siege. 2nd ed. London:
Bristol Classical Press.
Wright, M. T. 2000. "Greek and Roman portable sundials: An ancient essay in approxi-
mation," Archive for History of Exact Sciences 55=177-87.
Wright, M. T. 2002. "A planetarium display for the Antikythera Mechanism," The Hor-
ological Journal 144: 169-73, 193.
Wright, M. T. 2003a. "Epicyclic gearing and the Antikythera Mechanism, Part 1," Anti-
quarian Horology 27: 270-79.
Wright, M. T. 2003b. "In the steps of the master mechanic," in He archaia Hellada kai ho
Sunchronos Kosmos. Patras: University of Patras, 86-97.
Wright, M. T. 2004. "The scholar, the mechanic and the Antikythera Mechanism: Com-
plementary approaches to the study of an instrument," Bulletin of the Scientific In-
strument Society 80: 4-11.
Wright, M. T. 2005a. "The Antikythera Mechanism: A new gearing scheme," Bulletin of the
Scientific Instrument Society 85: 2-7.
Wright, M. T. 2005b. "Counting months and years: The upper back dial of the Antikythera
Mechanism," Bulletin of the Scientific Instrument Society 87: 8-13.
Wright, M. T. 2005c. "Epicyclic gearing and the Antikythera Mechanism, Part 2," Anti-
quarian Horology 29: 51-63.
Wright, M. T. 2005d. "11 meccanismo di Anticitera: L'antica tradizione dei meccanismi
ad ingranaggio," in E. Lo Sardo (ed.), Eureka! Il Genio deglio Antichi. Naples: Electa,
240-44.
Wright, M. T. 2006a. "The Antikythera Mechanism and the early history of the moon-
phase display," Antiquarian Horology 29: 319-29.
Wright, M. T. 2006b. "Understanding the Antikythera Mechanism," in 2nd International
Conference on Ancient Greek Technology. Athens: Technical Chamber of Greece, 49-60.
Young, S. 1939. "An Athenian clepsydra," Hesperia 8: 272-84.
Zeeman, E. C. 1986. "Gears from the Greeks," Proceedings of the Royal Institution of Great
Britain 58: 137-56.
Zerubavel, E. 1985. The seven day circle: The history and meaning of the week. New York: Free
Press.
CHAPTER 30
TECHNOLOGIES OF
CALCULATION
PART 1: WEIGHTS AND MEASURES

(CHARLOTTE WIKANDER)
The various systems of units used for weights and for measures both of volume and
of space in the classical world are bewilderingly diverse. The only phenomenon that
seems comparable is the measurement of time, particularly the many ways of
organizing solar and lunar calendars (chapter 29). Obviously, all humans needed to
measure weight, volume, space, and time from a very early period of social and
technological development, explaining the variety of approaches-which survived
in large part until the appearance of the metric system in the late eighteenth
century. As with monthly and yearly calendars, weights and measures for the archaic
and classical periods in Greece also present strictly local systems, some of which,
however, through growing trade, could be used over larger areas as a complement
to local measurements. This spread of influence happened mainly with the Aegi-
netan and, above all, Athenian standards.
The establishment of the new Hellenistic monarchies on the ruins of the Persian
empire during the last decades of the fourth century B.C. does not seem to have
changed the situation to any great degree. Both the Seleucid and Ptolemaic kingdoms
to a large extent kept to traditional measuring standards, a mixture oflocal traditions
and Persian customs, instead of trying to impose any consistent systems, even the
widely accepted Athenian system. For example, in spite of the strongly centralized
economic administration, the Persian dry-measure artabe, about 55 liters, was used in
Ptolemaic Egypt, but with widely different volumes ranging from 36 to 42 liters
depending on time, place, and type of goods (Rostovtseff 1941: 1296-1300).
The situation gradually changed after Rome's subjugation of the Eastern
Mediterranean: with the Mediterranean as an arena for trade within the Roman
Empire, local systems were gradually adapted in order to accommodate exchange
between regions, as well as customs regulations. These adaptations, however, were
slow and for the most part took place only in the first century A.D. A pertinent
example is the stone tablet of Leptis Magna in Libya, which, still in the second
century A.D., gives standard comparative measures of Punic, Roman, and Ptole-
maic length (Ioppolo 1967: 87-98).
In view of this diversity, the presentation below will be divided into various
categories of measurements (volumes: dry or wet, etc.) over a wide geographical
and chronological span, illustrated by some tables with short comments-this,
hopefully, to avoid total confusion to the reader. The tables will reflect the figures
normally given in modern works, but may occasionally deviate from, for example,
archaeological finds from different locations.
Metrology
The collective category "metrology" was established as a field of investigation in the
nineteenth century, as part of the German academic trend of enforcing positivistic
order on the earlier chaos of research concerning classical antiquity. The first
attempt at synthesis was published in 1838 (Bockh 1838) but was superseded in 1864
by Hultsch's Griechische und riimische Metrologie (revised and enlarged edition,
1882). By then, Hultsch had also established a base for further research with his
Metrologicorum Scriptorum Reliquiae (1864-1866), a collection of all known ancient
texts dealing with metrology. Hultsch has in many instances remained a standard
reference, the first serious attempt to systematize our knowledge of ancient mea-
surement systems. Viedebantt provided the next major contribution, Forschungen
zur Metrologie des Altertums (1917), then followed up and subdivided the field with
his Antike Gewichtsnormen und Munzfusse (1923); he also wrote the RE article
"Metrologie."
Once these fundamental bases for further research were established, then
augmented by such works as Berriman 1953 and Richardson 2003, the metrological
field tended to divide into different subsections, although with some cross-refer-
encing. The most important of these subfields were weights (frequently coupled
with numismatics), volume measurements (frequently coupled with amphora
studies), and the study of Roman land measurement. The most obvious example of
crossover into another field is the separation of the study of weight systems, which
became strongly connected to numismatics, since the ancient coinage systems were
based on metal weight.
Great help has been brought to the entire field in the twentieth century by a few
important publications of archaeological finds, mainly standard tokens of weight
TECHNOLOGIES OF CALCULATION 761
and standard measures of volumes. Most notable is the work of Lang (1964) con-
cerning the weights, measures, and tokens found in the Athenian agora, more
recently complemented by Hitzl (1996) on the weights from Olympia. This volume
also contains a report on the state of research concerning weights and coin values.
Apart from amphora studies, the only significant contextual evidence for weights
and measures from excavations other than those at Athens and Olympia is the
material from Olynthus (Robinson 1941).
Since the ancient volumes and measures presented were normally based on a
duodecimal system, they may puzzle the modern reader when translated into
modern decimal, metric terminology. The modern situation, however, retains a
certain amount of chaos in the measurement of distance, weight, and volume. Great
Britain and the United States, for example, still make use of miles, yards, and inches,
along with pounds and ounces, while the international oil market continues to
measure crude oil in "barrels" (159 liters) whether or not the producing or con-
suming country uses the metric system for other purposes. Even in Europe length
measurements based on parts of the human body (e.g., the cubit) were widely used
through the mid-nineteenth century. There is thus nothing particularly abnormal
in the ancient diversity.
Dry Measures
In the various Greek city-states, measures for dry and liquid contents were strictly
separated (see Table 30.1). The medimnos was the largest measurement, used gen-
erally although with different volumes in the different cities. The Attic standard,
however, was widely used, for example in Syracuse. The medimnos was a large
measurement, normally above 50 liters (although at Aegina and Sparta it was even
more), and thus for normal use had to be subdivided into four smaller units: the
hekteis, hemihekton, choinix, and kotyle. This smallest measurement, the kotyle
(varying from 0.273 to 0.379 l.), was also used as a normal drinking cup.
The medimnos has attracted a great deal of interest from scholars of ancient
history, since (according to Plutarch, Sol. 23.3) this was the basis of Solon's division
of the political rights of the Athenian population during his archonship of 594/3
B.C. The highest class was defined by production, the pentakosiomedimnoi, those
who had fields producing a minimum of 500 medimnoi of grain (ca. 25,000 liters,
which for modern conditions would be a very low yield).
Excavations in Athens have yielded several examples from the fifth and fourth
centuries B.C. of what has been plausibly identified as examples of standard mea-
surements for dry goods, enabling traders and customers to check the correctness of
volumes bought and sold. In the Athenian agora, many fragments of such cylin-
drical ceramic vessels, in some cases given the painted designation DEM (demosion,
"of the people") were found in the immediate vicinity of the Tholos, the building
where the five official magistrates overseeing the agora trade, the agoranomoi, had
their office (Lang 1964). The conscientious official control of weights and measures
Table 30.1. Dry measures

Greek (Athenian)
1 kotyle 0.274 l
2 kotylai dikotylon!xestes 0.547 l
3 kotylai trikotylon 0.821 l
4 kotylai choinix 1.094 l
16 kotylai hemihekton 4,378 1
32 kotylai hekteus 8.768 l
192 kotylai medimnos 52.536 1
Roman
1 cochlear 0.0111
4 cochlearia cyathus 0.046 l
6 cochlearia acetabulum 0.068 l
12 cochlearia quartarius 0.136 1
24 cochlearia hemina 0.273 l
48 cochlearia sextarius 0.546 l
8 sextarii semodius 4.366 l
16 sextarii modius 8.732 l
48 sextarii quadrantal!amphora 26.196 l
is further illustrated by the fact that the agoranomoi had the assistance of a college of
five Metronomoi, specialized in the overseeing of official measurements. Similar
magistrates were also employed in the Piraeus harbor and at Eleusis.
The basic Roman dry measure was the modius (shovel), originally and pri-
marily used for grain. It held about 8.7 liters, and could be divided into various
smaller measures (sextarii, heminae, quartarii, acetabula). Several physical examples
of measuring containers for a modius have been preserved in the archaeological
record, mainly from Roman Britain. They were open containers on low feet, made
of wood or metal. The modius was the common measure of grain (wheat as well as
barley) for all purposes: sowing, distribution of grain to the urban plebs of Rome,
measuring rations for agricultural slaves (cf. Cato, Ag. 56; Columella, Rust. 2.12).
Liquid Measures
The largest volume measure of fluids (primarily used for oil and wine) in the Greek
world was the metretes, also named amphoreus, after the most common ceramic
container (see Table 30.2). The Attic metretes measured 39.4 liters, and was divided
into four smaller units (chous, oxybathon, kotyle, and kyathos). Both kotyle and
kyathos were also names for drinking vessels, as the chous was for pouring. As usual,
Table 30.2. Liquid measures

Greek (Athenian)
1/6 kotyle kyathos 0.045 I
1/4 kotyle oxybaphon 0.07 I
1 kotyle 0.27 I
12 kotylai chous 3-281
144 kotylai metretes/amphoreus 39.4 I
Roman
1 cochlear 0.011 I
4 cochlearia cyathus 0.046 I
6 cochlearia acetabulum 0.068 I
12 cochlearia quartarius 0.136 I
24 cochlearia hemina 0.273 I
48 cochlearia sextarius 0.546 I
6 sextarii congius 3.275 I
24 sextarii urna 13.09 I
48 sextarii quadrantal!amphora 26.196 I
960 sextarii culleus 524.00 I
Aeginetan and Spartan measurements had their own standards, the Aeginetan
substantially larger than the Athenian, 54.6 liters, while the Spartan equivalent was
smaller, about 29 liters.
The Romans also used the vase shape amphora, a large container, as a measure.
Its volume was based on the weight of wine, 80 librae (libra of ca. 327.5 g) giving a
standard theoretical capacity of 26.2 liters. The amphora had seven subdivisions,
the urna, congius, sextarius, hemina, quartarius, acetabulum, cyathus, and cochlear.
Until the advanced second century A.D., however, the local volumes contained in
actual amphoras could vary significantly (Peacock and Williams 1986: 51-53). With
growing international trade, measures tended toward greater standardization from
the reign of Nero (A.D. 54-68) on. At both Pompeii and Minturnae, stone tables
with hollowed-out standards for measuring of fluids have been found (Mau 1908:
88-89).
Weights
The weights used at the local markets or for long-distance trade in the Greek world
(see Table 30.3) are very tangible archaeological objects, often of distinctive di-
mensions. They were fashioned of lead, bronze, or stone. For special purposes (in
medicine or for precious metals) smaller weights were used, but they have left no
Table 30.3. Weight standards

Greek Aeginetan Attic
obolos 1.04g 0.73 g
6 oboloi = 1 drachme 6.24 g 4.37 g
100 drachmai = 1 mina!mna 623.7 g 436.6 g
60 minai = 1 talent 37.42 kg 26.2 kg
Roman
1/144 uncia siliqua 0.19 g

1/48 uncia obolus 0.57 g
1/24 uncia scripulum 1.14 g
1/12 uncia dimidia sextula 2.27 g
1/8 uncia drachma 3.41 g
1/6 uncia sextula 4.55 g
1/4 uncia sicilicus 6.82 g
1/3 uncia binae sextulae 9.1 g
1/2 uncia semuncia 13.64 g
1 uncia 27.29 g
1.5 unciae sescuncia 40.93 g
2 unciae sextans 54.58 g
3 unciae quadrans 81.86 g
4 unciae triens 109.15 g
5 unciae quincunx 136.44 g
6 unciae semis 163.73 g
7 unciae septunx 191.02 g
8 unciae bes 218.3 g
9 unciae dodrans 245.59 g
10 unciae dextans 272.88 g
n unciae deunx 300.16 g
12 unciae libra!pondus 327.45 g
archaeological record. Stone weights were sometimes marked with a pair of female
breasts, while the mold-made metal weights could be square, rectangular, or tri-
angular (Kisch 1965, O'Brien 1981-1984).
The close connection between the coinage system and daily weighing activities
is reflected in the names for denominations of coins. The smallest weight was the
obol-probably derived from the word obeloi, thin metal spits in bunches used by
the early seventh century B.C. as a pre-stage of the weight and coinage systems
combined (Orion, Etymologikon s.v. obolos), which was also the name of the small
bronze coins in everyday use for shopping in classical Athens. Six obols more or less
made a drachme, giving an Attic drachma of 4.37 g in weight. The mina or mna
could vary from 70 to 150 drachmas. A common unit was the stater, with the weight
of two minai. The talent, weighing 60 minai, was so large that it was rarely used
outside state or official business. The standardization of weights and measures in
Athens is regularly credited to Solon in the early sixth century B.C., following the
influential passage of Aristotle (Ath. Pol. 10). Several complications are caused by
the fact that "market weights" and standard coin weights usually were not quite the
same, despite the use of the same names-the market weights being slightly heavier.
The two sites so far important as find-spots of weights are the agora in Athens
(Lang 1964: 2-38) and the sanctuary at Olympia (Hitzl 1996). Olympia in particular
has yielded sufficient archaeological material to allow precise analysis of the
original weights, thus eliminating many earlier points of speculation. The Olympia
excavations have clearly established the existence of an Aeginetan weight/coin
system different from the Athenian, a supposition earlier based mainly on passages
in Aristotle (Ath. Pol. 10) and Pollux ( Onomastikon 9.76, 86-87; late second century
A.D.), and subsequently hotly disputed.
The tolerance for aberration of precise weight was apparently quite wide: both
Lang and Hitzl record a range of variation of approximately 20 percent in weight.
This is surely not surprising, given the difficulties of fashioning cut-stone and cast-
metal weights to a precise weight. The scales on which the weights were used were
beam-balances, the oldest with equal-length arms, known from literature of the
fifth century. The arms held bowls, into which the weights and the substance to
be weighed could be placed. At least from the Hellenistic period, scales with
unequal arms were in use, allowing the measurement of greater weights without
the need for equivalent counterweights but requiring carefully calibrated yards
(Lazzarini 1948).
The Aeginetan and the Athenian drachmas were exchanged at the rate of 70 to
100, but they were both used at the same time in the classical period at Olympia
(Hitzl 1996: 53-54). One mina weighed 110 Attic drachmas, but only 77 Aeginetan
drachmas. Hitzl's latest group, however, shows a combination: a standard mina of
110 drachmas (480.26 gr.). This standard was imposed on the Athenian subjects of
the Delian league in the late decades of the fifth century, and seems to have pre-
vailed also at Olympia.
While the growing Roman empire gradually imposed a standardized system of
weights from the second century B.C. onward, local standards survived for a sur-
prisingly long period in both Italy and the eastern Mediterranean. The original
Roman/Italic basis for weight standard was the libra or pondus, the pound, which
provided the unit both for early coinage (aes signatum, aes grave) and for market
weight (as). While the coin and metal weights coincided, the actual standard weight
varied substantially during the Republic, just as in Greece. There are, however,
indications that the Roman pound was related and adjusted to the Greek system:
thus, the Roman pound of 327.45 g was exactly the equivalent of three quarters of a
Greek market mina.
The libra was the standard unit; there was no nomenclature for heavier weights,
although archaeological evidence shows stone and metal weights of 2, 3, 4, and 5
pounds to be fairly usual; the 5-pound measure seems to have been particularly
common. The libra itself had several subdivisions, notably the semis, uncia, triens,
quadrans, and sextans.
Measures of Length and Area

As in most premodern societies, in Greece and Rome length measurements were
based on the human body (see Table 30.4). The basic measurements were taken
from the arm (from the elbow to the tips of the fingers-pechys, cubitus, ell; ca. 46
cm), the foot (pous, pes, two-thirds of an ell; -ca. 30 cm), and the distance across
the outstretched arms (orgyia, fathom; ca. 6 feet- or 1.8 m). For measuring longer
distances, the foot was multiplied by five (bema diplooun, passus; 1.48 m), or by
5,000 to yield the mile (milia passuum; 1.48 km), the normal measurement for road
distances in the Roman empire. One should, however, beware of regarding these
measurements as very precise, since they vary significantly between localities and
across time (Hecht 1979; Wikander 1993). In Greece, particularly in athletic con-
texts, we also encounter the stadion (600 feet, ca. 177-85 m, depending on local
standards). The stadion for races at Olympia was exceptionally long, 192.3 m.
From an early stage in the sixth century B.C., we encounter in Greece the system
of defining field surfaces as units of "plough-land," that is, the area which it was
possible to plough during a working day, from sunup to sundown. Clearly this is an
imprecise unit, since topography, type of soil, and the condition of the draft animals
can affect the result. Nevertheless, the same method for measuring land is reflected
in the Roman measurement iugerum (from iugum, yoke). The iugerum, however,
which was the normal private and state unit for agricultural properties during
the Republic, was well defined: a rectangle measuring 120 x 240 Roman feet (35.52 x
71.04 m, 2,523 sq m). It was subdivided according to a duodecimal system into actus,
unciae, and scripula, suitable for smaller cultivation, such as kitchen gardens.
Roman engineers used collapsible brass measuring rulers, which gave the
measure of a foot of 29.6 cm, subdivided into sixteen digiti (fingers), or twelve
unciae. In order to measure long distances, particularly roads, a wheeled hodometer
was used (Lewis 2001: 134-39). The engineer and architect Vitruvius described this
machine around 25 B.C. (see chapter 31). It consisted of two wheels connected to a
series of cogwheels, one vertical and one horizontal, which, counting the revolving
of the wheels, dropped a small ball into a container for each mile passed. This
corresponded to the system of the milestones (miliaria), spread along the roads of
the Empire, giving the distances from Rome on the main military roads, and from
the nearest town or city to the next urban center. Such milestones, giving also
details of repairs, have been found in great numbers throughout the empire, their
standard interval (where that can be determined) being 1,481.5 meters, around.
1,000 passus or 5,000 pedes.
For measuring areas and distances the Roman mensores (measurers) used an
instrument called the groma, which made it possible to plot straight lines and
Table 30.4. Measures of length and area

Greek length Roman length
1/16 foot daktylos 1.7-2.2 cm 1/16 foot digitus 1.85 cm
1/8 foot kondylos 3,4-4.4 cm 1/12 foot uncia 2.47 cm
1/4 foot palaiste 6.8-8.8 cm
1 foot pus 27-35 cm 1 foot pes 29.6 cm
(Athenian
32.5 cm)
1.5 feet pechys 40-53 cm 1.5 feet cubitus 44.4 cm
5 feet bema 1.35-1.75 m 5 feet passus 1.48 m
diplooun
6 feet orgyia 1.62-2.10 m 2 passus pertica 2.96m
10 feet akaina 2.70-3.50 m 24 passus actus 35.56 m
100 feet plethron 27-35 m 125 passus stadium 185.2 m
600 feet stadion 177-213 m 1000 Roman 1,481.5 m
passus mile
Greek area Roman area
1 sq foot 0.73-1.23 sq m 1 sq foot pes 0.088 sq m

quadratus
l0XlO akaina 7.3-12.3 sq m 100 sq feet scripulum 8.76 sq m
feet (Ptolemaic
Egypt)
l00XlO0 plethron ca. 700-1,200 2,400 sq uncia 210.28 sq m
feet sqm feet
medimnon ca. 2,700 sq m 6 unciae = 120X120 1,262.00 sq m
( Cyrenaica) actus feet
quadratus
l00XlO0 aroura 27.56 sq m 12 unciae = 24ox120 2,523.34 sq m
ells (Ptolemaic iugerum feet
Egypt)
2 iugera heredium 5,046.68 sq m
200 iugera centuria 50.47 ha
800 iugera saltus 201.87 ha
90-degree angles off established lines (Lewis 2001: 120-33). The use of this device is
described by Heron of Alexandria (Dioptra 33), and it is illustrated on funeral
reliefs; one example survived at Pompeii (Schi0ler 1994: Brodersen 1995: 202). It was
based on a vertical rod of man-height, with an offset crown to which a plumb line
was attached to mark the survey point on the ground. A cross with four equal arms
perpendicular to each other was mounted on the crown in such a way that it could
turn. A plumb line was attached to the end of each arm, serving as crosshairs for
sighting. Once the central plumb bob was in position, the horizontal cross arms
could be adjusted to allow sighting on physical landmarks or the compass points,

and the marking of straight lines or perpendiculars by men with range rods.
Sighting from these lines across not too large distances, milestones or boundary
markers for fields could be erected; the use of back-sightings to confirm the lines
allowed remarkable accuracy with a simple device (Lewis 2001).
The systems of weighing and measuring in the classical world may seem bewildering
in their diversity; nevertheless, they are usually quite logical and useful within their
time and place, and they fostered significant accomplishments in agriculture, en-
gineering, architecture, and commerce. The carefully calculated weight standards
behind Athenian coinage in the fifth century B.C. and behind the coinage of
Alexander in the fourth helped promote trade and prosperity in both periods (see
part 2 of this chapter). The production of stone columns and architectural members
in a graduated series of standard dimensions was an important contribution to the
extraordinary productivity of the Roman imperial building program (chapter 10).
The roads and colonial land allocations measured off by Roman surveyors opened
up enormous areas of Europe and North Africa to settlement and commerce. Like
literacy, numeracy, and timekeeping (chapters 28 and 29), the measurement of
weight, length, and area constituted an important subsidiary technology facilitating
the many other technical accomplishments of the classical world.
REFERENCES
Berriman, A. E. 1953. Historical metrology: A new analysis of the archaeological and the
historical evidence relating to weights and measures. London: Dutton.
Bockh A. 1838. Metrologische Untersuchungen uber Gewichte, Munzfiisse und Masse der
Alterthums in ihrem Zusammenhange, Berlin: Veit.
Brodersen, K. 1995. Terra cognita: Studien zur romischen Raumerfassung. Spudasmata 59.
Hildesheim: Olms.
Hecht, K. 1979. "Zurn romischen Fuss," Abhandlungen der Braunschweigischen Wis-
senschaftlichen Gesellschaft 30:1-34.
Hitzl, K. 1996. Die Gewichte griechischer Zeit aus Olympia. Olympische Forschungen 25.
Berlin: de Gruyter.
Hultsch, F. 1864-1866. Metrologicorum Scriptorum Reliquiae. 2 vols. Leipzig: Teubner.
Hultsch, F. 1882. Griechische und romische Metrologie. 2nd ed. Berlin: Weidmann.
Ioppolo, G. 1967. "La tavola delle unita di misura nel mercato augusteo di Leptis Magna,"
Quaderni di Archeologia Libia 5: 89-98.
Kisch, B. 1965. Scales and weights: A historical outline. New Haven: Yale University Press.
Lang, M. 1964. Excavations in the Athenian agora. Vol. 10, Weights, measures and tokens.
Princeton: The American School of Classical Studies at Athens.
Lazzarini, M. 1948. "Le bilance romane del Museo N azionale e dell' Antiquarium Comunale
di Roma," Rendiconti della Classe di Scienze Morali, Storiche e Filologiche dell' Acca-
demia dei Lincei Ser. 8.3: 221-54.

University Press.
Mau, A. 1908. Pompeji in Leben und Kunst. Leipzig: Wilhelm Engelmann.
O'Brien, D. 1981-1984. Theories of weight in the ancient world. 2 vols. Leiden: Brill.
Peacock, D. P. S., and D. F. Williams 1986. Amphorae and the Roman economy: An in-
troductory guide. London: Longman.
Richardson, W. F. 2003. Numbering and measuring in the classical world. 2nd ed. Bristol:
Bristol Classical Press.
Robinson, M. 1941. Metal and minor miscellaneous finds: An original contribution to Greek
life. Olynthus 10. Baltimore: Johns Hopkins University Press.
Rostovtzeff, M. 1941. The social and economic history of the Hellenistic world. 3 vols. Oxford:
Schi0ler, T. 1994. "The Pompeji-Groma in new light," Analecta Romana 22: 45-60.
Viedebantt, 0. 1917. Forschungen zur Metrologie des Altertums. Leipzig: Teubner.
Viedebantt, 0. 1923. Antike Gewichtsnormen und Munzfusse. Berlin: Weidmann.
Wikander, 6. 1993. "Terracotta modules, Oscan feet and tile standards," in E. Rystedt, C.
Wikander, and 6. Wikander (eds.), Deliciae fictiles: Proceedings of the First Interna-
tional Conference on Central Italic Architectural Terracottas at the Swedish Institute in
Rome, 10-12 December, 1990. Acta Instituti Romani Regni Sueciae 50. Stockholm:
Astroms Forlag, 67-70.
PART 2: COINAGE
(ANDREW MEADOWS)
The Technology of Coinage

Coinage is a technology in its own right. It is also the result of specific processes
that represented, at least in the early stages, major technological advances. Within
the context of this handbook, it is appropriate to stress the role of coinage as a
facilitating technology. Without it, for example, many of the major engineering
activities described in the other chapters of this book could never have left the
realm of the theoretical. It was also responsible for major paradigmatic shifts in
other areas of ancient behavior and thought. These are three bold claims; each
must be addressed in turn.
In some form or another, money (by which we mean a substance that can be
used as a store of wealth, an expression of value, or a medium of exchange) has
existed almost as far back as the human documentary record extends, at least as far
as the third millennium B.C. Coinage is a specific subset of money that might be
defined as a piece of metal used as money which conforms to standards and bears a
design (Howgego 1995: 1). As recently as the 1970s, the beginning of coinage ap-
peared as a radical technological jump in economic history, but more recent
scholarship has tended to situate coinage within a continuum of the development
of monetary instruments. On the one hand, there has been a movement to play up
the extent to which the Greek world had, before the arrival of coinage, adopted the
Near Eastern practice of using both silver and gold bullion as a monetary medium
(see especially Kroll 1998, 2001). At the same time, analysis of the Near Eastern
archaeological record has suggested two ways in which coinage was foreshadowed
by the use of silver ingots. There is fairly widespread evidence for the production of
so-called chocolate-bar ingots, which were cast in a preportioned form, apparently
to facilitate their subdivision for use in specific transactions (Thompson 2003).
There is also growing evidence from as early as the twelfth to eleventh centuries
B.C., both from literary and documentary texts and in finds of silver and bullae, for
the practice of sealing determined amounts of silver, probably in linen bags, to
provide easily usable quantities of guaranteed weight and fineness (Thompson
2003). Thus, in all three ways in which it is a distinctive monetary form-metallic
composition, production to a fixed weight, and stamping with a design-it appears
that coinage was a development from an existing Near Eastern tradition rather
than a fundamentally new technology.
Nonetheless, coinage did bring with it technological advance. Research has
suggested, for example in the case of the early electrum coinage, that the regular-
ization of the production of gold stimulated the systematic use of gold refining. "It
would seem that here was a technology awaiting a problem, a problem which was
finally posed by the introduction of coinage, that is pieces of precious metal where
the weight and purity were guaranteed" (Ramage and Craddock 2000: 212). The
authors of this passage were puzzled that while the technological ability to refine
gold by removing from it naturally occurring silver had long existed, there is no
evidence for its systematic exploitation until the invention of electrum coinage.
They push this hypothesis further to suggest that "everywhere, except possibly in
India, the introduction of a gold coinage seems to have provided a stimulus for
gold-refining" (Ramage and Craddock 2000: 12). This observation provides us with
an important example of the way in which coinage as a concept drove the full
development or application of a new technology from infancy to maturity.
Gold coinage, however, remained a marginal part of the economy for much of
the ancient Greek and Roman world. Silver was the mainstay of the monetary
market. Metal analysis that has been carried out on pre-coin monetary silver (Stos-
Gale 2001) suggests that the metal was in general of high quality (97% pure or
better). The earliest coinages matched these levels. Archaic Greek coinage, at many
of the mints at which it was produced, generally achieved silver fineness of 98
percent or better (Gale et al. 1980). Nonetheless, debasement was always an option,
and it was deployed at times as a financial stratagem but was rarely used in a
"deceptive" sense. An interesting example is provided by a series of Roman coins of
the early second century B.C. known to modern scholars as quadrigati (from their
distinctive reverse types showing a four-horse chariot). These coins appear to have
been produced routinely at about 72 to 93 percent fineness (Walker 1980: 58-61). At
the same time, however, the Roman mint was producing the clearly separate de-
Figure 30.1. Brass sestertius of the emperor Titus (A.D. 79-81), with a design commem-
orating the completion of the Flavian amphitheater in A.D. 81. British Museum, BMCRE
Titus 190. (By permission of the Trustees of the British Museum.)
narius coinage at full fineness. It seems that the easily recognizable debased
quadrigati had a function separate from that of the denarii, and they were not
intended to pass for full-fineness coins. When debasement of the coinage did began
to occur more systematically, notably under the Roman Empire, it seems that this
occurred as part of a considered and closely monitored strategy. Analysis of the
imperial silver coinage of the Severan dynasty, for example, has suggested that a
silver fineness of 46 percent was established empire-wide and observed not just at
the mint of Rome, but also at the provincial mints in Syria and Tyre (Gitler and
Ponting 2003). Such debasement continued into the third century and became a
central element in the Roman state's attempts to increase the money supply when
faced with increasing demands for expenditure.
The advent of coinage thus led to an important monetary development in the
form of fiduciary financial instruments, although there were, in fact, much earlier
precedents. Late in the fifth century B.C., coinage began to be produced in copper-
based alloys. The value of these coins was, in general, notionally linked to their
metal content. Bigger and heavier coins were worth more, but this higher value was
not directly proportional to the extra amounts of metal in most cases. Manipula-
tion of the alloys used to produce such fiduciary, lower-value coins, was at its most
sophisticated following the reform of the coinage by Augustus (37 B.C.-A.D. 14).
Under Augustus, two new high-value base coins were introduced, the sestertius
(previously a silver denomination) and the dupondius (e.g., figure 30.1). Unlike the
lower value as, which was a bronze coin, these higher denominations were pro-
duced in brass. The yellowish color of this alloy (orichalcum) seems to have had
golden associations (see chapter 4; figure 4.5), which were played upon by the
minting authorities to create confidence in higher-value base metal coin (Burnett
1987: 54) . Later in the Roman imperial period, there is evidence for similar ma-
nipulation of expectation in the phenomenon of deliberate surface enrichment of
low-fineness silver coins in an attempt to make them look more silver than they
really were; Butcher and Ponting (1998: 310) discuss the "devious nature of ancient
moneyers."
Coinage as a New Technology

Coinage was hand produced, and as such it represented no great technological
advance. The engraving of the dies used to strike the coins was a skill inherited from
and shared with the producers of sealstones, which had a long history before the
arrival of coinage and which continues even now. The application of the engraved
dies to the coin blank-the obverse die below, embedded in the "anvil," the reverse
die above-and the striking of the upper die with a hammer was not a sophisticated
process either. Striking the coin represents nothing more than an adaptation of the
application of seals to clay. Again the process is foreshadowed and paralleled by the
stamping of such items as documents, bricks, or amphora handles. Nor is there
anything even in the scale of use of coin dies, at least in the first three centuries of
coin production, that represents a decisive shift from previous practice. Although
the evidence may not be plentiful, we must imagine that the seals of important
officials within the great empires of the Near East were a common sight on official
documents, and widely recognized.
The technological advances that coinage brought through its mass-produced
nature were rather advances of application, and we may divide them essentially into
two categories: those intended by the original issuers, and those that resulted
spontaneously from use.
Although, as has been noted above, several important facets of coinage as a
monetary instrument were present in preexisting monetary traditions and objects,
two key elements were not. First, coinage was, in the form that it was produced,
ready-stamped with a seal that guaranteed it and gave it value within the context of
the society that recognized the device. Second, it was produced at a series of fixed
weights in a denominational system that provided for its use in making fixed
payments at certain, different levels of value without any further treatment, such as
cutting. These two elements combined to create a monetary form that was ideally
suited to the making oflarge-scale identical payments in a way that had never been
possible before, other than in kind. Indeed, it has been argued that the need to make
such payments was always the motivating factor for the production of coinage in
the ancient world (Crawford 1970). The technology embodied in coinage in this
sense became the facilitator of a whole series of activities that had been unthinkable
or impossible without coinage and forms the subject of this chapter's final section.
Such developments, however, were not necessarily foreseen, desired, or intended by
the issuers of coin.
There is another sense in which the design on the coinage presented new
possibilities. For although personal seals and sealings had undoubtedly been and
remained a common occurrence within administrative circles, the designs that
.. .
- ~ , ~,-
~ --
-~---.
~ \ ~~
'\' -.
..~·· . -----~ ,-,
~"£, -' " .\, ........
• ' -· ' '\
~·1],, . <"\""I -
'./F::.;:;..,,. 1 i 1'1 -,
;; ~ -1 ·•· 1- ••··· \ > :
.. :_· ~~'" \~;:;.\ i / '
•!_,_ -~,1%,~ -~... ~-\ -
j_:i,.;.Jt.,_
~-,..l,A,,. l1~~1?.•{#:
.. --~
'-~ --......;.~.· .·•.....·-,:;;;?". .
•~
Figure 30.2. Silver tetradrachm of Alexander the Great (336-323 B.C.). British Museum,
BMC 5c. (By permission of the Trustees of the British Museum.)
appeared on coins were different in two fundamental respects. First, by the fifth
century B.C. and onward through the Hellenistic period down to the appearance of
Roman coinage in late third and second centuries B.C., the designs on the coins
had transcended the personal to become symbolic of the state that guaranteed the
coinage. Even when personal badges did appear as coin designs, this was a function
of the individual's role within the state apparatus, whether as servant of the state or
embodiment of it. Second, after coinage had entered the mainstream of economic
behavior, and once lower denominations suitable for small-scale transactions had
become common, in the Hellenistic period and under the Roman Empire, the
designs on coins penetrated to levels of society on a far wider basis than did the seal
designs of the elites. In a world with no other form of mass medium, there was a
golden opportunity to be seized here. Alexander the Great was the first to do so.
The mechanism seems childishly simple to a modern eye. As Alexander moved his
ways eastward with his army and began to establish himself at the head of what had
been the Persian Empire, he established mints to produce the coinage that the
maintenance of his prodigious conquest required. Unlike his Persian predecessors,
who had adopted no consistent monetary form, Alexander ensured that each of his
mints produced coinage in exactly the same form in pure gold and silver (figure
30.2). By the time of his death, almost thirty different production centers were
producing virtually identical coins, from Macedonia in the west to Babylon in the
east (Price 1991: 72). To many of those inhabiting the vast realm that had become,
within a very short space of time, Alexander's kingdom, coinage, and the messages
it conveyed about Alexander's divine origins, royal status, and military prowess,
would have been the single obvious sign of the new regime. Within a few years
there could have been virtually no one within this kingdom who had not been
exposed to the ideological ensemble of the coin designs. This was the birth of mass
communication.
A similar approach was taken by some of the Hellenistic monarchies, but the
phenomenon would not occur again on such a scale until the reforms of the third
century A.O. that unified coin design across the Roman Empire. An interesting
Figure 30.3. Electrum stater signed by Phanes, perhaps minted at Ephesus, ca. 600 BC.
British Museum, BMC 1. (By permission of the Trustees of the British Museum.)
early development of this potential occurred in Republican Rome during the later
second and first centuries B.C., when the coin designs chosen by the monetary
magistrates became for a while the vehicle for the development of family histories
and the memorializing of ancestral tradition, as part of the political interplay of the
major Roman gentes (Meadows and Williams 2001).
Coinage as Facilitator of Technologies

As has already been noted, coinage was well adapted from the outset to perform a
particular function, and it is often inferred from this suitability that that this func-
tion was one of the prime drivers behind coinage's invention, its uptake in the Greek
world, and its eventual success as a monetary medium in the Greco-Roman world
and beyond. That function is the making of broadly acceptable mass payments.
For the earliest coinages, the nature of the issuing and controlling body is
unclear. The oldest Greek coin inscription informs us that the design on the coin is
the badge of an individual named Phanes (figure 30.3). It is thus not impossible
that early coins were produced by wealthy individuals. Throughout the classical
Greek and Roman periods, however, virtually all coinage may reasonably be as-
sumed to have been guaranteed and issued by the state (although the source of the
bullion may at times have been private). Coinage will have entered circulation by
being paid out by the state in return for services, labor, and produce. As such,
coinage becomes indicative of an approach to public activity on a fundamentally
different model than that of the royal and temple states of the Near East. Coinage
did not emerge as a major part of the economy in Egypt, Mesopotamia, and Iran
before the fourth century B.C. It was not produced in these areas much before the
arrival of Alexander the Great. As a state payment device, coinage is the evidence
for a state that functions by purchasing what it needs rather than commanding
it. The precise relationship between the development of the Greek city and the
concept of citizenship on the one hand, and the appearance of coinage on the
other, remains unclear. There may be an element of cause and effect operating in
both directions between the two; but it is certainly the case that coinage spread
Figure 30.4. Silver denarius issued by Mark Antony as triumvir, 32-31 B.C. British Mu-
seum, BMCRR East 189. (By permission of the Trustees of the British Museum.)
swiftly throughout the Greek world in the latter half of the sixth century B.C., the
point at which many Greek states seem to have been coalescing as political units,
drawing up law-codes, and wrestling with constitutional questions such as the
nature of tyranny.
It is certainly true to say that key elements of the Greek and Roman states that
we now consider definitive are scarcely imaginable without coinage. A prime case is
that of the radical democracy at Athens. By the late fifth century B.C., a number of
the integral institutions, such as the law-courts, the rotating council, and the
assembly of all citizens, offered cash payments to facilitate participation on the part
of the populace at large. Coinage was the oil that greased the wheels of this political
apparatus. It was also at Athens and at this period that it became possible to
maintain a paid, standing military force. The real implications of this development
become most evident under the Roman Empire when, by the early third century
A.D., there may been as many as a third of a million men earning more than 500
denarii per year (Alston 1994). At times the military nature of coin issues is explicit
in their design, as in the case of the massive silver issues of Mark Antony in the 30s
B.C. These coins are often described as "legionary denarii," since they were marked
with the numbers of the legions for which they were destined to serve as pay (figure
30-4). The impact of such a development goes far beyond the military, of course.
The soldiers themselves were engaged in areas that we would today regard as civil
engineering, such as the building of roads and bridges. But other workers, too,
from the fifth century B.C. onward, were paid for their services: masons, architects,
doctors, lawyers, prostitutes, artists, actors, and musicians (Loomis 1998) (see
figure 30.1). Coinage made it possible not only for the state to patronize these
activities on a broad scale, but also for private individuals to pay for services.
The existence of a large body of citizens effectively divorced from the land by
their full-time military, technical or artistic employment, and thus outside the
basic subsistence farming structure of ancient society, created a cash-based market
for the basic goods required to survive. The role of coinage in this new cash-based
economy is clearly stated by ancient political philosophers such as Plato and
Aristotle (Pol. 1.9):
It came to be agreed between men that for exchange purposes they would give and
receive some sort of commodity, which was easy to handle and also useful for life,
such as iron and silver and any other such substance. In the beginning value was
decided by size and weight. Subsequently, stamps were placed upon it so as to
remove the need for measurement: for the stamp was a mark of value. Once
money had come into being to serve the needs of necessary exchange, the other
form of moneymaking quickly came into being: trade. In the beginning this was
perhaps fairly simple, but then became with practice more complex, as men
learned where and how exchange might yield the greatest profit. This is why it
appears that moneymaking principally involves coinage, and that the chief re-
quirement for such moneymaking is the ability to see where the most money will
come from, for that is where wealth and money are made. Indeed, it is often held
that wealth is a large quantity of coin, because that is what moneymaking and
trade is all about.
Coinage for Aristotle was clearly a technology that allowed exchange to function
more smoothly within an advanced society, and also facilitated the accumulation
of wealth. So far had it become embedded within everyday experience in the major
cities of the late fifth century B.C., that it was impossible even for Plato, who
gravely mistrusted the accumulation of wealth that it facilitated, to envisage a city
without it (e.g., Leg. 5.742a-b). This is all the more remarkable when we consider
that even in Plato and Aristotle's day, coinage was still being produced predom-
inantly in denominations unsuitable for the fulfillment of small-scale transactions.
Most Greek city states in the archaic and classical periods produced only silver
coinage, and only at denominations of the value of about a week's pay. It was only
in the Hellenistic and Roman periods that smaller silver and bronze denominations
became the norm.
But by this point no one could stop coinage from performing precisely the
function that Plato feared. By allowing individuals to take up paid employment
and free themselves from the land, coinage created not just individuals able to
devote themselves to professional activity and thus advance the technologies of
their societies, but also a class with the leisure to participate in the enjoyment of the
fruits of their labor. The Romans and Greeks who enjoyed the lavish bath com-
plexes that adorned their cities, congregated around the splendid nymphaea, and
traveled the fine new roads could do so because they had money. Coinage had
given them that freedom and, for this reason above all others, was the most potent
of all ancient technological advances.
REFERENCES
Alston, R. 1994. "Roman military pay from Caesar to Diocletian," Journal of Roman Studies
84: 113-23.
Balmuth, M. (ed.) 2001. Hacksilber to coinage: New insights into the monetary history of the
Near East and Greece. New York: The American Numismatic Society.
Burnett, A. M. 1987. Coinage in the Roman world. London: Seaby.

Butcher, K., and M. Ponting 1998. "Atomic absorption spectrometry and Roman silver
coins," in W. A. Oddy and M. Cowell (eds.), Metallurgy in numismatics, N. Royal
Numismatic Society Special Publication 30. London: Royal Numismatic Society,
308-34.
Crawford, M. H. 1970. "Money and exchange in the Roman world," Journal of Roman
Studies 60: 40-48.
Gale, N. H., W. Gentner, and G. A. Wagner 1980. "Mineralogical and geographical silver
sources of archaic Greek coinage," in Metcalf and Oddy 1980, 3-49.
Gitler, H., and M. Ponting 2003. The silver coinage of Septimius Severus and his family
(193-211 AD: A study of the chemical composition of the Roman and Eastern issues.
Milan: Ennere.
Howgego, C. J. 1995. Ancient history from coins. London: Routledge.
Kroll, J. 1998. "Silver in Solon's laws," in R. Ashton and S. Hurter (eds.), Studies in
Greek numismatics in memory of Martin Jessop Price. London: Spink, 225-32.
Kroll, J. 2001. "Observations on monetary instruments in pre-coinage Greece," in Balmuth
2001, 77-91.
Loomis, W. T. 1998. Wages, welfare costs and inflation in classical Athens. Ann Arbor:
University of Michigan Press.
Meadows, A. R., and J. H. C. Williams 2001. "Moneta and the monuments: Coinage
and politics in republican Rome," Journal of Roman Studies 91: 27-49.
Metcalf, D. M., and W. A. Oddy (eds.) 1980. Metallurgy in numismatics, I. Royal Numis-
matic Society Special Publication 13. London: Royal Numismatic Society.
Price, M. J. 1991. Coinage in the name of Alexander the Great and Philip Arrhidaeus.
Ramage, A., and P. Craddock 2000. King Croesus' gold: Excavations at Sardis and the history
of gold refining. London: The British Museum Press.
Stos-Gale, Z. A. 2001. "The impact of the natural sciences on studies of Hacksilber and
early silver coinage," in Balmuth 2001, 53-76.
Thompson, C.M. 2003. "Sealed silver in Iron Age Cisjordan and the 'invention' of
coinage," Oxford Journal of Archaeology 22: 67-107.
Walker, D.R., 1980. "The silver contents of the Roman Republican coinage," in Metcalf
and Oddy 1980, 55-72
PART 3: PRACTICAL MATHEMATICS

(KARIN TYBJERG)
Practical mathematics is related to technology in several ways. It is both a tech-

nology in its own right for administrating and facilitating daily life, economic
transactions, and administration, and a tool for other technologies, such as land
measuring, construction, warfare, and hydraulics. Practical math is not simply a
matter of applying mathematics to practical ends; the relationship between
mathematics and technology is two-directional, and practical methods are reflected
back into theoretical mathematics. The evidence shows that mathematics played a
central role in political and administrative control, that professional practitioners
used mathematics to solve problems as well as to create an identity for themselves,
and that the relationship between practical and theoretical mathematics is complex
and dynamic. There is, however, one caveat: the source material on practical and
particularly everyday mathematics is limited, varied, and frequently indirect. Often
the methods and techniques for practical calculations and geometrical construc-
tions have been lost, and only the results are preserved: accounts, measurements,
loans, and records of inheritance and tax. Little is known about the relationship
between what was taught and what was used, and the information transfer was
predominantly oral, leaving little conclusive material.
Mathematics as a Technology
In Aristophanes' Wasps (656; quoted from Rihll 1999: 45), a quick calculation is
introduced with the remark "Reckon up, not with counters, but just on your hands,
the amount of tribute." This comment illustrates two points about the use of
mathematics in classical Athens: that finger reckoning was second best to a more
sophisticated method of counters, and that mathematical calculations were per-
formed and checked in public. It also indicates that, since numeracy played an
important role in public life, a history of numeracy should be taken as seriously as
the much-researched problem of literacy in the Greek world (see Netz 2002 on the
role of numeracy and counters in classical Greek culture).
The earliest formal notation for numbers is known as the acrophonic system,
and its symbols derive from the first letter in the word for each number (5 is IT for
penta, 10 is d for deka and so on); the unit is simply I. This system developed in close
conjunction with finance and trade and includes symbols for monetary units such
as obols. While there are no literary sources for this notation, it is common on
official inscriptions from Athens and other Greek communities. It is also found on
stones probably used as counting boards and abaci, such as the famous Salamis
tablet (of disputed date, but possibly late fourth century B.C.), whose markings can
be interpreted as obol and drachma fractions (JG 112 2777; see Pullan 1970: 23-25;
Heath 1921: 49-51) (figure 30.5). There is no evidence on how these boards were
used, but it may be surmised that pebbles were arranged and manipulated on lines
representing different numbers of units. This could be the method of counters
alluded to by Aristophanes.
The importance of mathematical techniques in public life is clear from nu-
merous inscriptions of records, lists, and accounts. They show, for instance, the
tributes paid by other cities to Athens with amounts, subtotals, and totals, lists of
loans to the state or temples, and inventories and accounts of construction work.
The mathematical practices of accounting and recording thus had both a practical
and a symbolic role. Not only did they allow officials to administrate resources; the
public display of accounts also served the symbolic purpose of displaying power
,_
X
u
:..,..
[.
<I
(!_
..,.. :i:
n e-.
-I X
X
Figure 30.5. Greek calculation tablet, probably for use with pebbles. Salamis, uncertain
date. (After E. Guillaume in DarSag 1.1, 1877: 2.)
and organization as well as transparency and objectivity. Practical mathematical

techniques were used to produce "accountable" knowledge and acted as a guar-
antee for objectivity in social and political transactions as well as in commerce and
finance (Cuomo 2001: 13-16) .
From the third century B.C. on, alphabetic notation took over as standard.
Numbers were named consecutively after the letters of the alphabet, in which the
first nine represent 1-9, the next nine 10-90, and the last nine 100-900. In conse-
quence, this notation was integrated with written culture and was used in high level
Greek geometry, as well as for inscriptions and administrative documents.
Mathematical techniques played a central role in the complex bureaucracy of
Ptolemaic and Roman Egypt, which monitored land holdings and land use. A large
number of papyri and tablets deal with land surveys and provide detailed infor-
mation on owners, location, dimensions, and crops. Typically, for Egyptian and
later Greek mathematics, the numbers are listed in so-called unit-fractions-that is,
the reciprocal values half, third, fourth, fifth, and so on. Values are then expressed
as whole numbers plus decreasing fractions. The calculations performed in these
surveys were approximations and often contained mistakes. Areas of quadrilateral
fields were, for example, usually estimated by multiplying the averages of the two
pairs of opposite sides. This gives advantage to the tax-man as the area is always
overestimated, except in the case of a perfectly rectangular field.
Very similar arithmetical and geometrical techniques are found in Babylonian,
Islamic, and medieval European sources, indicating a long-lived tradition of pro-
fessional mathematics. Throughout this tradition we find not only simple approx-
imation techniques, but also solutions to complex problems, which are totally
unrealistic despite being set in practical contexts such as land measurement or horse-
trading (see H0yrup 1990 on "sub-scientific" mathematical traditions). These prob-
lems demonstrate a theoretical outgrowth from practical techniques and solutions.
Clear traces of this development are found in more theoretical authors, such as
Diophantus, who presented the same arithmetical problems but without the horse-
trading, and in Hero of Alexandria, who placed problems ofland measurement in an
Archimedean context. The boundaries between practical and theoretical mathe-
matics were not watertight.
In general, the Roman world featured a grand administrative system of
mathematically trained officials in both army and government. Roman magistrates
were obliged to submit accounts after their period in office to a central archive, and
precise accounts were also kept at lower levels. Some of the most precise and
thorough accounts known from the Roman world stem from the pay-sheets of a
couple of first-century A.D. soldiers (Ste. Croix 1956: 39-40). Training was sys-
tematized, and there is evidence from the Roman world for basic mathematical
education that supported recording, accounting, and measuring, along with tables
for conversion of measures and units (Cuomo 2001; Fowler 1999: 268-76).
The Roman numeral system was-like the Greek acrophonic system-closely
linked to trade and calculation on the abacus, as the counters on the abacus are
easily noted down, with each counter matching a symbol (Taisbak 1965). The
abacus works by representing numbers by counters on lines with values 1, 5, 10, and
50; the number 15 was thus represented on the abacus by one counter on 10 and one
on 5, and in Roman numerals by X for 10 and V for 5 ( on the abacus, see Heath 1921,
Pullan 1970). Figure 30.6 illustrates a reproduction of an abacus found at Pompeii;
the counters in the upper rank are worth 5, while those below are worth 1. From left
to right the columns represent 1,000,000, 100,000, 10,000, 1,000, 100, 10, 1, 1½
(uncia, ounce), and fractions of an uncia.
Both measurement and finance in the Roman world operated with a system of
twelfths-the so-called uncia fractions (¾ 2 , '7,: 2 , etc.). Surveyors, for example, divided
their unit, the iugerum, into twelve parts, and the foot was divided into twelve
unciae pedis; incidentally, this use of the word uncia gives rise to the English inch. In
Ars Poetica (325-330 ), Horace describes schoolboys at work: "Tell us Albinus' son: if
you subtract uncia (¾ 2 ) from quincunx (¾2 ), what is left? Come on, out with it!"
"Triens (~)." "Good; you'll be able to look after your property" (quoted from
Richardson 2004: 21-22).
A striking example of how mathematics could be used to demonstrate orderly
government is found in Frontinus' account of the water-supply system for Rome,
De aquaeductu urbis Romae. Fron tinus estimates the amount of water entering the
Figure 30.6. Modern reproduction of Roman abacus with sliding counters. Science
Museum, London. (Courtesy of Science Museum/Science & Society Picture Library.)
water system for the city and, to account for these resources, he asks whether the
intake matches the discharge to official supply structures. He found a considerable
loss and sought to stop fraudulent diversion by standardizing pipe sizes. In this way,
mathematics was used to secure a fair distribution of resources and-not unlike
Horace's schoolboy-Frontinus demonstrates his competence at looking after the
property of the state through a display of calculations with uncia fractions (cf.
chapters 2 and 11).
Practical mathematics was thus not only essential in daily interactions of trade
and commerce, but also highly significant in the administration of complex soci-
eties. Mathematics was an instrument for recording and control, quantification of
power, and collective accountability.
Mathematics in Technology
It is often claimed that mathematics and technology were unconnected in antiquity
and that this separation barred progress. On closer scrutiny, however, many con-
nections become apparent, and it is clear that mathematics was used in a variety of
technologies, including land measurement, construction, hydraulics, town plan-
ning, astrology, and military engineering. Techniques for measuring and estimating
areas and volumes were related to geometrical theory. The preserved treatises that
employ geometry in technological problems provide, however, mainly a theoretical
superstructure although they deal with ostensibly practical problems. The real ex-
tent of practical use of geometry is therefore uncertain.
The origin of geometry was said-for example, by Herodotus (2.109), Plato
(Phdr. 274c), and Strabo (17.3)-to be in measuring land, and the term geometria
had the dual significations of both land measurement and geometry. At the more
theoretical end of the spectrum stands Hero of Alexandria with his Metrica, on
calculating and dividing areas and volumes. This treatise offers a systematic
treatment of geometrical objects and contains both methods of calculation and
geometrical proofs. Although the work goes far beyond practical use with its
consideration of polygons and sections of complex curves, it maintains a practical
perspective. The section on dividing areas is explicitly related to land measure-
ment, and three-dimensional objects are described as architectural elements.
That practical and theoretical work cannot easily be distinguished is also made
clear in the Corpus Agrimensorum Romanorum, a large collection of Latin works on
land measurement that contains a mixture of extracts from Euclid, treatises on
techniques for land measurement, and laws on land administration (Blume et al.
1848-1852; Lewis 2001). In the Corpus, Balbus, who worked as a surveyor in the
Roman Army under Domitian or Trajan, writes about his practical tasks, but also
produces a taxonomy of geometrical objects and concepts. Practical experience is
thus inserted in geometrical discourse when the curved, irregular lines of the
landscape are listed with the straight and circular lines of geometry. In this way
Balbus offers fellow surveyors a geometrical framework for their practical tech-
niques and thus elevates the status of the mathematical expertise of surveyors.
Columella takes almost the opposite approach in his treatise on farm management
for gentlemen, De re rustica (A.D. 60-65), since he includes some practical geo-
metrical techniques to demonstrate that he knows how it is done but specifically
states that he leaves this sort of calculation to the surveyor (mensor; 5.1.3). He offers
simple rules for calculating the areas of fields shaped like various geometrical
objects-triangles, rectangles, and circles-among them the simple rule for quad-
rilaterals used in Roman Egypt (Rust. 5.1.4-2.10).
A similar range of techniques and geometrical superstructure is found in many
other technical disciplines. Little is known of architectural geometry, but an
Egyptian papyrus of Euclid refers to geometrical objects as architectural elements,
such as stones or columns. The text thus appears to have been used in the training
or study of architects, although the Euclidean geometry probably constituted no
more than an educated background. The Roman architect Vitruvius points out in
his handbook De architectura (ca. 25 B.C.) that geometry and arithmetic are needed
by the architect to calculate cost, make measurements, lay out buildings, and solve
problems of symmetry (De arch. 1.1.4). Vitruvius, however, lists mathematics
among many other disciplines and thus demonstrates the learning of architects
rather than explaining what happens on the building site; he uses mathematics to
upgrade his technical expertise. At the same time, mathematics does appear to have
been used for large building projects. In a fifth-century B.C. tunnel construction for
water-supply-the Eupalinus tunnel-numbers are found engraved along the sides
of the tunnel wall, probably measuring the distance into the mountain (Kienast
1995; cf. chapter 12). A more theoretical account of tunnel construction is found in
Hero's instructions for triangulating across a mountain, so that the teams digging a
tunnel from either end will meet in the middle. An inscription from the Roman
colony of Saldae shows that this was a real concern. Two teams of workers had failed
to meet in the middle because each veered from the straight line, that is, from proper
geometrical control (CIL 8.2728; chapter 12).
Military machines were also subjected to mathematical treatment. In the third
century B.C., Philo of Byzantium wrote on catapult construction and offered advice
based not only on experience but also on geometrical proof. He relates the problem
of scaling well-functioning catapults up or down to the famous geometrical prob-
lem of doubling the volume of a unit cube. He solves the problem with a moving
ruler and thus combines mechanics and geometry in two ways: a geometrical
problem is set in a practical context and proved by an instrument. In his Poliorcetica
(Siegecraft), Philo also applies geometry to the problem of constructing fortifica-
tions. He recommends building walls in a zigzag shape of regular triangles, leaving
the fortification less vulnerable to sieges by providing a more comprehensive range
of fire from the walls. Although these models are more complex than strictly
practicable, archaeological ruins of fortifications suggest that the principle was
indeed applied. The importance of mathematics and in particular measurement to
the military engineer is further illustrated by the oft-repeated admonition to
measure the height of the walls of a fortification to ensure that the siege machines
are sufficiently large (Polybius 9.19; Hero, Dioptra 190.14-19).
Geometrical techniques and proofs for measuring heights, areas, and volumes
are thus employed in a number of technical disciplines. The role of the geometrical
techniques is not just practical; they also serve to emphasize the special expertise of
practitioners and raise the status of their disciplines. Although practical mathe-
matics was often looked down upon-most notably by Plato (e.g. Resp. 7.525b-
526b )-practical efficacy held status in its own right. Hero of Alexandria shows that
practical methods provide a natural extension of geometrical methods for dealing
with irregular areas and volumes (Tybjerg 2004) and Balbus inscribes irregular lines
in a Euclidean context, but both proudly point to the practical prowess of their
methods. The combination of geometry and technology may have been behind the
complex tunnel buildings in antiquity, but certainly bolstered professional standing
and status of many technical experts.
REFERENCES
Blume, F., K. Lachmann, and A. Rudorff 1848-1852. Die Schriften der romischen Feldmesser.
2 vols. Berlin: G. Reimer.
Cuomo, S. 2001. Ancient mathematics. London: Routledge.
Fowler, D. 1999. The mathematics ofPlato's Academy: A new reconstruction. 2nd ed. Oxford:
Heath, T. L. 1921. A history of Greek mathematics. 2 vols. Oxford: Clarendon Press.
H0yrup, J. 1990. "Sub-scientific mathematics: Observations on a pre-modern phenome-
non," History of Science 28: 63-86.
Kienast, H.J. 1995. Die Wasserleitung des Eupalinos auf Samos. Bonn: Habelt.
University Press.
Netz, R. 2002. "Counter culture: Towards a history of Greek numeracy," History of Science
40: 321-51.
Pullan, J. M. 1970. The history of the abacus. 2nd ed. London: Hutchinson.
Richardson, W. F. 2004. Numbering and measuring in the classical world. 2nd ed. Bristol:
Bristol Phoenix Press.
Rihll, T. E. 1999. Greek science. Greece and Rome New Surveys in the Classics 29. Oxford:
Ste. Croix, G. E. M. de 1956. "Greek and Roman accounting," in A. Littleton and B. Yamey
(eds.), Studies in the history of accounting. London: Sweet and Maxwell.
Taisbak, C. M. 1965. "Roman numerals and the abacus," Classica et medievalia 26: 147-60.
Tybjerg, K. 2004. "Hero of Alexandria's geometry of mechanics," Apeiron 37: 29-56.
CHAPTER 31
GADGETS AND
SCIENTIFIC
INSTRUMENTS
ORJAN WIKANDER
THE category of "gadgets" includes a large number of devices (automata) and

machines described by Greek inventors and engineers. These more or less com-
plex devices, which constitute an important source of information on the theo-
retical as well as practical mechanical standard of the classical world, have been the
subject of remarkably little detailed study, They have often been-and occasionally
still are-dismissed as worthless toys or "marvels," intended to evoke religious awe
among the superstitious public. For some time, however, a more serious judg-
ment of the automata has prevailed, describing them as object lessons in me-
chanical and pneumatic principles, rather than as tricks intended to inspire wonder
(e.g., Price 1962; Drachmann 1963; Bedini 1964). In addition to these ancient au-
tomata, this chapter will treat some practical applications of the physical princi-
ples they illustrated, for example water clocks, astronomical instruments, and
hodometers.
BACKGROUND
The so-called five simple machines were known and discussed by Greek engineers
and scientists from at least the late third century B.C. onward, perhaps beginning
with Philo of Byzantium. The early-fourth-century mathematical commentator
Pappus describes them (Collectio 8.52):
There are five powers [machines] by the use of which a given weight is moved
by a given force .... But it has been shown by Hero and Philo that the above-
mentioned powers [machines] are reducible to a single principle, even though
they differ considerably from another in form. Their names are the follow-
ing: wheel and axle, lever, compound pulley, wedge, and, moreover, the so-
called endless screw.
The "single principle" referred to is, of course, that what you lose in distance, you
gain in power. The five simple machines are fundamental for most mechanical
devices. The wheel and axle, the lever, and the wedge were known by the Early
Bronze Age, the pulley by the eighth or seventh century B.C., while the screw is not
attested with certainty until the mid-third century B.C. Although four of the simple
machines were known before the age of Alexander the Great, in general the knowl-
edge of mechanical or pneumatic devices remained almost nonexistent before that
time. The development of the screw and gearwheel, allowing the precise application
of force in a circular motion, were essential to this development.
In consequence, it is not surprising that the phenomenon of automata is so
strongly associated with the Hellenistic period, when their necessary prerequisites
were finally at hand and the Museum at Alexandria was a hive of inventive activity.
Given the lack of a mechanical and cultural context, there is good reason to feel
dubious about the few gadgets alleged to be earlier than the third century B.C.
Athenaeus (4.174c), for example, ascribes a complex clock to Plato: "But it is said
that Plato provided a small notion of its [the water-organ's] construction by
having made a clock for use at night that was similar to a water-organ, although it
was a very large water-dock." There are numerous technical objections to the
flying bird of Archytas, as described by Aulus Gellius (NA 10.12.9-10):
By a certain theoretical knowledge and mechanical education, Archytas made a

wooden model of a dove that was able to fly. It was obviously balanced by weights
and driven forward by a current of air enclosed and concealed in it. Concern-
ing such an absurd story, I would like to quote Favorinus' own words: "Archytas
of Tarentum who, among other things, was an engineer, made a wooden, flying
dove. When it alighted, it did not take flight again."
These two stories show interesting parallels. Both gadgets are attributed to very
well known philosophers active in the first half of the fourth century B.C., and both
descriptions come from sources about five centuries later: Gellius and his infor-
mant Favorinus wrote in the second century A.D., Athenaeus about A.D. 200.
Gellius seems to doubt his own story, and Athenaeus-otherwise a notorious
GADGETS AND SCIENTIFIC INSTRUMENTS 787
name-dropper-cannot refer to anything more specific than "it is said" (legetai).

Diels (1915) attempted a reconstruction of Plato's alarm clock, but both the clock
and the bird may well be as mythical as the robots allegedly constructed by He-
phaestus (Homer, Il. 18.369-79) and Daedalus (Aristotle, De an. 1.3.406b).
THE TECHNOLOGICAL BOOM OF THE

THIRD CENTURY B.C.
The basis for Hellenistic science was laid by Aristotle and his pupils in the Lyceum in
the period from about 336 to 323 B.C., then further elaborated by his followers at the
Lyceum and the scientists at the Museum of Alexandria. A fair picture of the state of
mechanical research immediately before the explosion of technological accom-
plishments during the reigns of Ptolemy II Philadelphus and Ptolemy III Euergetes
(283-221 B.C.) can be found in the Mechanika (Mechanical Problems) traditionally
attributed to Aristotle, but presumably written by one of his pupils in the 280s B.C.
The treatise is devoted mostly to the lever, wheel, wedge, and pulley, and the
practical applications discussed show clearly the low level of mechanical knowledge
at the time. But a dynamic development of both theoretical and practical mechanics
was soon to begin. One particularly important discovery was the possibilities of-
fered by air, water pressure, and steam resulting from studies concerning vacuum
executed by Strato of Lampsacus, leader of the Lyceum from 287 to 269 B.C. This
knowledge is partly preserved in the introduction to Heron's Pneumatica.
No less important was the invention of various devices for changing the speed or
direction of a linear or rotating motion: the gear, the cam, and the screw. Only
relatively recently has the fundamental development of ancient gearwheels been
understood (see, particularly, Foley et al. 1982; Lewis 1997: 37-57). A number of
uncertain points remain but, for the time being, our knowledge of the earliest history
of the cogwheel or gearwheel may be summarized as follows. About 300 B.C., power
was transmitted by the use of chains passing over pentagonal prisms, in the repeating
catapult designed by Dionysius of Alexandria (Philo of Byzantium, Bel. 73-76). If the
prisms had been provided with notches on each side in order to engage the chain
better, the construction would not have been very dissimilar from a true cogwheel. In
the 270s B.C., Ctesibius' water clock included a rack and pinion and perhaps gear-
wheels (Vitruvius, De arch. 9.8.5). By about 260 B.C. there is indirect but credible
evidence for the worm gear, and shortly afterward for ordinary cogwheels engaging
in parallel or at right angles, in descriptions of the hodometer and the planetarium of
Archimedes (Vitruvius, De arch. 10.9.1-4; Cicero, Rep. 1.14.21-22).
As in the case of the water-mill, much scholarly energy has been devoted to
questioning, not the existence, but the actual importance of the cam in antiquity.
Figure 31.1. Reconstruction of gears and cam on Nysa's vehicle. (Lewis 1997: fig. 25;
courtesy of Dr. Michael J. T. Lewis, University of Hull.)
Today, however, there is little doubt that cams on the shaft of water-powered wheels
played an important part in some ancient Roman industries (see chapters 6 and 13),
and this device has been traced in the ancient descriptions of at least four Hellenistic
automata (Lewis 1997: 84-88, figs. 25-28). The earliest certain example is the arm of a
puppet carpenter moving a hammer up and down in the automatic theater of Philo
of Byzantium, from the 230s B.C. (preserved in Heron, Automatopoiika 24; Schmidt
1899: 422-27, fig. 103). Lewis argues persuasively for two even earlier examples: the
flute player of Apollonius of Perge around 240 B.C. (preserved only in a reworked
Arabic version by Banu Musa) and the 4 m high, seated statue of Nysa carried in a
procession of Ptolemy II Philadelphus in the 270s B.C. (figure 31.1). Athenaeus
(5.198f) quotes an almost contemporary account by Callixeinus of Rhodes: "This
stood up automatically without anyone putting his hands to it, and after pouring a
libation of milk from a golden bowl sat down again."
According to ancient sources, the history of the screw started with Archimedes,
who allegedly invented the water-screw during his stay as a youth in Egypt, perhaps in
the 260s (Diodorus Siculus 1.34.2, 5.37.3; Strabo 3.2.9; Athenaeus 5.208f; Drachmann
1958). Through a misunderstanding of Diodorus' text (Kiechle 1967: 14, notes 1-2),
some scholars maintain that Archimedes found the screw already in use in the Nile at
his arrival. Nor is the attempt of Dalley convincing (in Dalley and Oleson 2003: 7-17)
to envisage water-screws in Sennacherib's palace gardens in Nineveh four hundred
years earlier. Obviously the water-screw was from the very beginning intended for
practical use, but so were most applications of the screw, particularly the screw-
presses discussed by Heron and Pliny (Heron, Mechanika 3-15, 19-21; Pliny, HN 18.317;
Drachmann 1932: 50-85). Small metal bolts intended for joining pieces together
remain scarce in antiquity; they are mostly to be found in small luxury objects, parade
helmets, and medical instruments (Mutz 1969; Klumbach 1973; Gaitzsch 1983; Dep-
pert-Lippitz 1995). The endless screw played an important part in the worm gear used
by automata. The tapering wood-screw did not exist in the classical world.
The crank, finally, remains subject to controversy. No certain literary or ar-
chaeological evidence proves its existence in antiquity, even though some scholars
argue for the invention of the crank by Archimedes (Drachmann 1973; Sleeswyk
1981a). Certainly it seems remarkable that a civilization familiar with the rotating
hand-mill would not have applied the same principle to the crank, whose handle
rotates in more or less the same way. In any case, none of the known ancient
automata or other gadgets requires the crank to function.
AUTOMATA
The adjective automatos, meaning "self-acting" or "spontaneous," can be found as

early as Homer's description of the wheeled tripod robots of Hephaestus (Il. 18.376).
Today, the word is the technical term for various mechanical gadgets, in particular
those described by Philo of Byzantium and Heron of Alexandria. The power of the
automata derives from forces such as hanging weights, revolving wheels, or the
pressure of heated air, and in exceptional cases even the force of wind or steam
pressure (Heron, Pneumatika 1.43, 2.6, 2.11; see chapter 6). Again, it is doubtful
whether these gadgets could correctly be described as "toys," and it is a myth that
they could to any appreciable extent have functioned as marvels stimulating reli-
gious awe. Landels (2000: 202-3) points out that, of 75 devices in Heron's Pneu-
matika, only five can justifiably be labeled "temple miracles": for example, burning
incense on an altar is extinguished automatically by a libation (Heron, Pneumatika
1.12); holy water runs from a vending machine when a five-drachma coin is inserted
(1.21); temple doors open when a fire is lit on the altar outside (1.38).
There is reason to doubt that even these rather few "miracles" were, in fact,
realized in order to awake awe among the public. Five drachmas is an extremely
high price for a dose of holy water, and the mechanism for opening the temple
doors in the last example is so complex that it is very questionable whether it would
function in practice. Nonreligious "miracles" are far more common, such as fig-
urines of various animals drinking water (e.g., Heron, Pneumatika 1.29-31, 2.36-37)
or of singing birds (e.g., 1.15-16, 2.4-5, 2.32). Even many of these may, in fact, be
mere armchair inventions. Few, if any, of the gadgets could have been of any
practical use. Possible exceptions are automatically replenished oil lamps (1.22-24)
and lamps whose wick was pushed forward as the oil burned, by means of a rack
engaging a cogwheel coupled to a float (1.34).
Many automata were certainly intended to cause astonishment, but, as has
been pointed out by various scholars, the ultimate intent of these miracles was
r
Figure 31.2. Magic drinking horn designed by Heron of Alexandria, Pneumatika 1.18. (6.
Wikander, after Schmidt 1899: fig. 19.)
doubtless to illustrate physical and mechanical principles. There are many exam-
ples showing the function of communicating vessels and siphons, the powers of
air-pressure and vacuum, the use of floats and gears, and so on. If the goal was
educational, why was there so much emphasis on the manifestation of marvels? We
are not in a position to say, but part of the answer may be connected with the social
situation of the Alexandrian inventors. They were almost entirely dependent on the
good will of the Ptolemaic rulers who sponsored the Museum and every scientist
associated with it. The royal court (rather than the general public) probably was the
most appreciate and lucrative audience for the entertainment and amusement
offered by these miracula, long afterward collected in the works of Heron.
Several gadgets concerned with the pouring of wine or other beverages would
have provided suitable entertainment at royal banquets. For example, one jug
delivers the same amount of liquid every time one pours from it (Heron, Pneu-
matika 2.1). Another one pours various liquids separately from the same mouth
(1.22). Yet another magic device (1.18) pours water, wine, or both (figure 31.2):
Some drinking-horns (rhyta) are so designed that, when first wine and then water
is poured into them, pure water flows out from them at one occasion, a mixture
[of wine and water] at another, and pure wine at yet another. This is arranged in
this way. A drinking-horn a. ~ y has two partitions 8 i:: and~ l). Through both
of them is led a pipe 0 K soldered together with the partitions and pierced with the
little hole 'A, which is placed slightly above the partition ~ l). But below the par-
tition 8 i::, there should be a ventµ in the hollow of the drinking-horn. Under these
conditions, if one intercepts the outlet y and pours in the wine, it will proceed
through the hole 'A to the space 8 i:: ~ l). For the air in that will depart through the
ventµ . When we cover the ventµ with the finger, it will keep the wine in the part
os~ TJ. Now, when we pour water into the part ix~ so of the drinking-horn, while
covering the vent µ, the water will flow pure. But if we release the vent µ while the
water is still above, a mixture will flow. When the water has flowed out, then
the pure wine will flow. It is also possible, by [covering and] releasing the vent
µ several times, to vary the outflows. But it is better first to pour water into the
space o s ~ TJ and then pour wine, when the vent is covered. For it will follow at
one occasion that pure wine will flow out, but, when the vent is released again, a
mixture, and, when the vent is covered again, pure wine will flow. And this will
happen as many times as we want.
WATER-CLOCKS AND ASTRONOMICAL

INSTRUMENTS
The inflow water-dock attributed to Ctesibius (ca. 280 B.C.) is the first device
certainly including a true cogwheel: "The water flowing in at a regular volume
through that opening raises an inverted bowl, which is called the 'cork' or 'drum' by
the craftsmen. A bar and revolving drum are attached to this apparatus and both are
fitted with regularly spaced teeth which, when meshing with one another, make
measured rotations and movements" (Vitruvius, De arch. 9.8.5; Drachmann 1976).
Archimedes further elaborated Ctesibius' water-dock, substituting a right-angle gear
for Ctesibius' rack and cogwheel (Hill 1976; Sleeswyk 1990: 28-34; Lewis 1997: 37-41).
Together, they became the point of departure for the construction of water-clocks
during the Greco-Roman period and well into the medieval period (chapter 29).
From the very beginning, the construction of inflow water-clocks was closely
related to that of astronomic instruments such as astrolabes, which later enjoyed
popularity in the Islamic world and, in the fourteenth century, became the in-
spiration for the mechanical astronomical clocks of western Europe (Price 1974:
54-55; King and Millburn 1978). It is consequently quite appropriate that Archi-
medes, apart from his clock, was famous for his exquisite planetarium. In fact, this
complex device, which was brought to Rome after the capture of Syracuse in 212
B.C., was described with admiration by four authors writing in Latin over four
centuries: Cicero (ca. 54-51 B.C.; Rep. 1.14.21-22; Tusc. 1.63), Ovid (A.D. 2-18; Fast.
6.263-83), Lactantius (early fourth century A.D.; Div. inst. 2.5.18), and Claudian
(ca. A.D. 400; Carmina minora 51 [68]). The most detailed description is given in
Cicero's De re publica (whose dramatic date is 129 B.C.). The speaker is L. Furius
Philus, consul in 136 B.C. and one of the leading statesmen of his time.
I remember that C. Sulpicius Gallus-as you know, a most learned man-

when he happened to visit M. Marcellus, who had been consul together with him,
asked him to bring forward the celestial globe which Marcellus' grandfather
had removed from the capture of the rich and beautiful Syracuse, even though he
had taken nothing else to his home from the immense booty captured. Although
I had heard this globe mentioned frequently because of Archimedes' fame, I was
not particularly impressed by its appearance; for the other globe that the same
Archimedes had made and the same Marcellus had placed in the temple ofVirtus
was more beautiful and more widely known. But when Gallus had begun to
explain the function of this device in an authoritative way, I realized that there had
been greater genius in that famous Sicilian than it seemed possible for a human
being to possess. For Gallus told us that the other solid and massive globe was an
old invention, first constructed by Thales of Miletus ....
But this kind of globe, he said, which showed the motions of the sun and the
moon and of those five stars which are called "wandering" and, so to speak,
"strolling about" [i.e., the planets], contained more than could be accommodated
in the solid globe. The invention of Archimedes was admirable inasmuch as he
had contrived that one single turning device could represent uneven and varying
movements with totally different rates of speed. When Gallus set this globe in
motion, it so happened that the moon remained as many revolutions behind the
sun in that bronze device as the number of days it was behind in the sky; ac-
cordingly, the same eclipse of the sun happened on the globe [as in the sun], and
the moon came to the limit constituted by the shadow of the earth, when the sun
from the region ... [lacuna]
Although neither this description nor any of the others presents a clear picture
of the function of this device, it clearly was capable of illustrating the interrela-
tionships of numerous heavenly bodies. Even though there is no explicit mention of
gears in Archimedes' "globe," we must presume that cogwheels transmitted the
motion to its various different parts. In Hellenistic gadgets motion was more
often transmitted by pulleys or plain friction wheels (the entire "automatic theater"
of Hero worked in this way [Schmidt 1899: 334-453] ), but with these methods it
would have been impossible to reach the precision necessary for the functions
described.
It is tempting to envisage Archimedes' "globe" as something reminiscent of a
nineteenth-century orrery or planetarium, that is, a three-dimensional model of a pre-
Copernican solar system in which the turning of a crank gave a clear illustration of the
relative movements of the planets, moon, and sun around the earth. In all probability,
however, it looked more like the well-known Antikythera Mechanism, discovered in a
Roman shipwreck south of the Peloponnesus in 1900, but scientifically published only
in 1974 (Price 1974: 5-51; with significant corrections by Bromley 1986, Wright 2002,
2006; see also chapter 29). This instrument was originally enclosed in a rectangular
wood-and-metal box, 30 cm long, 15 cm wide and 7.5 cm high (figure 29.1). Although
the details of the mechanism and its precise function are still under investigation, it
may have included as many as 31 parallel cogwheels of bronze with between 15 and
about 225 cogs each, by means of which complex astronomical calculations could be
carried out. The results could be read on three dial plates on the front and back faces of
the instrument. On the dials, solar years could be compared with lunar years and
synodic months, and it was possible to predict solar and lunar eclipses, to read the
rising and setting of various constellations, and perhaps other functions.
Although the myth survives in modern scholarship, Price and the scholars
following him emphasize that this device had nothing whatsoever to do with the
navigation of the ship on which it was found. It is an instrument whose scientific
and technical complexity far surpasses what would seem reasonable for practical
purposes. It is rather a precious luxury article, comparable with the Greek bronze
statues conveyed, presumably to Rome, on the same ship. From astronomical
arguments, Price suggested that the instrument was constructed around 87 B.C., a
date that accords well with the probable date of the wreck, approximately 80 B.C.
This is also a period when members of the Roman elite showed great interest in
technical devices of this kind. Cicero also mentions (Nat. D. 2.34.88) a similar
orrery constructed by the philosopher Posidonius, "which at each revolution
presents the same result in the sun and the moon and in the five wandering stars
[the planets] that takes place in heaven every day and night." Cicero was well
acquainted with the device, as he knew Posidonius personally from his studies in
Rhodes in 78 B.C. It is very unlikely, however, that the mechanism discovered at
Antikythera is identical with Posidonius' planetarium.
Price saw a direct connection between devices like the Antikythera Mechanism
and the Islamic astrolabes known from about AD 1000 (al-Biruni) onward, a
hypothesis confirmed only nine years after the publication of his book by the
discovery of four fragments of a Byzantine, brass sundial-calendar (Field and
Wright 1985a and 1985b). It has been dated to the early sixth century A.D. and may
be seen as the missing link between the devices of the classical world and those of
early Islamic science (figure 31.3). The device made use of internal gears to provide
a display of the god for each day of the week and the phases of the moon. The face
plate carries coordinates for 16 towns and provinces around the Mediterranean;
when the gnomon was set to the appropriate location and the indicator on the
suspension set to the corresponding declination, the shadow would fall on the
appropriate hour. As the sundial-calendar is in many ways almost identical with al-
Biruni's astrolabe, it shows beyond doubt that that the astrolabe was not an Islamic
innovation but, in fact, a late development from Hellenistic precursors.
In antiquity, the word astrolabium was used for two different devices, the
astrolabium planisphaerium (related to the devices already described) and the "ring-
astrolabium" invented by Hipparchus in the second century B.C. and described in
detail by Ptolemy (Alm. 5-1; Kauffmann 1896). It consisted of several connected
metal rings, one corresponding the ecliptic, another the colure of the solstices. The
rings were divided into degrees and marked with astronomical symbols; by turning
them, various astronomical calculations could be made. We get a clear picture of
this device from an astronomical instrument discovered at Philippi in 1965
(Gounares 1978). It consists of two rotating rings connected with a third, larger one,
with a diameter of 7.25 cm. Inscriptions make it possible to adjust the instrument
according to months and geographic position (Vienne, Rome, Rhodes, Alexan-
dria). The Philippi device, which has been dated to the fourth century A.D., could
be seen as something between a true astrolabe and a portable sundial, about 15 of
which have been found throughout the Roman Empire (Price 1969; Buchner 1971).
Figure 31.3. Reconstruction of a Byzantine portable sundial-calendar with settings for 16

locations, ca. A.D. 400-600. Science Museum, London, inv. 1985-222. (Courtesy Science &
Society Picture Library.)
HODO METERS
Hodometers, mechanical instruments for measuring travel distances both on land

and at sea, are described in detail by both Vitruvius (De arch. 10.9.1-4, 5-7) and
Heron (Dioptra 34 and 38) and may have served a more important role than is
often assumed. Many attempts have been made to reconstruct these complex
machines (e.g., Drachmann 1963: 157-68). This discussion will concentrate on the
land hodometer of Vitruvius.
Let the wheels of the carriage have a diameter of four feet, so that, if a wheel
has a marked point and starts moving forward from this point, rotating on the
surface of the road, it will have completed a distance of precisely 12½ feet, when it
arrives at the point from which it began its revolution. Then let a drum (tym-
panum) with one single tooth projecting from the circumference be fastened to
the inner side of the wheel-hub. Moreover, in the wagon-body a box should be
securely fastened with a drum rotating vertically and fixed to an axle. On the
outside of this drum four hundred small teeth (cogs) shall be placed at equal
distances, which mesh with the teeth of the lower drum. Moreover, at the side of
the upper drum shall be fixed another little tooth projecting beyond the (other)
teeth. But above, there should be placed a horizontal drum, toothed in the same
way and enclosed in another box, whose teeth mesh with the little tooth, which
had been fixed to the side of the second drum. In this (third) drum, as many holes
should be made as the number of miles that can be covered with the carriage in a
day's journey.... In all these holes small round stones should be placed, and in the
case of this drum ... there shall be one opening with a small channel through
which the stones placed in this drum, when they have arrived at this place, can fall
one by one into the wagon-body and a bronze vessel placed below. Thus, when the
wheel moves forward, it carries with it the lowest drum, and as the little tooth on
this drum at each revolution drives the teeth in the upper [drum] forward, the
result will be that, when the lowest [drum] has revolved 400 times, the upper
drum revolves once; and the little tooth that is fixed to its side moves one tooth of
the horizontal drum. Thus, as the upper [drum] will revolve once compared to
400 revolutions of the lowest drum, it will record a distance of five thousand feet,
that is, one thousand paces. Consequently, the number of stones that fall will
announce by their sound the passing of a single mile. But the number of stones
collected from the bottom will show, by their sum, the number of miles for the
day's journey.
Vitruvius' description is detailed, but not entirely clear. It is no surprise that

Leonardo da Vinci was fascinated by it and tried to recreate the machine at his
drawing-board (Sleeswyk 1981b: 159-62). Two fundamentally different solutions to
the problem are preserved by his hand, but both differ decisively from Vitruvius'
text, and later scholars have resigned themselves to the insolubility of the problem.
Even Drachmann (1963: 156-59, fig. 61) considered the hodometer an armchair
design impossible to construct in practice. In the late 1970s, however, Sleeswyk
contrived a solution based in its entirety on Vitruvius' description, although he
resorts to less conventional solutions to some details for which Vitruvius supplies
no information. He has demonstrated the viability of his reconstruction by
building a quarter-size scale model. In his drawn reconstructions (figure 31.4), the
number of cogs has been drastically reduced for the sake of clarity, but the prin-
ciple remains the same (Sleeswyk 1979, 1990: 23-28; Lendle 1990).
Sleeswyk's solution seems convincing, but his historical analysis is even more
exciting, when he asks to whom Vitruvius referred when attributing the machine to
"our predecessors" ( rationem ... a maioribus traditam). His argument starts from a
small but revealing detail: the use of falling balls or stones-not uncommon in the
Islamic world, but known in the classical world from one more case only, the
water-dock attributed to Archimedes. The attribution is not absolutely certain, but
it can be supported by a variety of arguments-for example, the very motive for the
invention of the hodometer. What could be the purpose of providing a carriage
with so complex (and certainly expensive) machine, when the main Roman roads
were supplied with milestones at intervals of one Roman mile (1.48 km)? Sleeswyk
argues convincingly that in fact it was the Romans who needed the hodometer,
when they began to put out milestones in the 250s B.C. Perhaps some Roman
official solicited a solution to the need for laying out roads over long distances and
calculating quickly the requirements for materials and labor (Lewis 2001: 134-39).
Figure 31.4. Reconstruction ofVitruvius' hodometer. ( 6. Wikander, after Sleeswyk 1990:

fig. 5.)
At this time, Archimedes was a man in his thirties, already respected for his
scientific gifts and active at the court of Hieron II of Syracuse, one of Rome's close
allies during the First Punic War. Sleeswyk's theory remains hypothetical, but
plausible. Among Archimedes' earliest works is Dimension of the Circle, in which he
fixed the value of pi between 3 10/70 and 3 10/71 ( ca. 3,1418, within about 0.2 percent
of the real figure), a study in accordance with approximately simultaneous work on
the hodometer, a practical application whose function is dependent on a tolerably
accurate value of pi.
GADGETS IN THE ROMAN EMPIRE
The technology of the Roman Empire is mainly associated with large, low-tech-
nology machines based on Hellenistic inventions, for example the saqiya gear
drive, the water-mill, and olive-presses. But various types of evidence prove that
the knowledge of automata and other gadgets was passed on, too. Even though the
majority of Heron's automata presumably belong to the third century B.C., it is

worthy of note that he found it meaningful to write about them as late as the
second half of the first century A.D. Astrolabes and planetaria survived. The series
of literary testimonia concerning the planetarium of Archimedes over the four
centuries from Cicero to Claudian shows the importance attached to the device.
The same is true of portable sundials, the astrolabe from Philippi, and water-
clocks, which indicate the technical skills that survived throughout the Roman
Empire. We even have examples of the use of gadgets in Roman ceremonies and
processions. The Nysa statue shown in Ptolemy Philadelphus' procession has its
counterpart in an automated image of Julius Caesar's corpse, made of wax, dis-
played at his funeral: "The image was turned round by a mechanism in all di-
rections, and all over the body and the face the twenty-three wounds could be seen
that had so brutally been inflicted on him" (Appian, B.Civ. 2.20.147). When
Herodes Atticus arranged the Panathenaic festival at Athens (ca. A.D. 139), the ship
carrying the new peplos of Athena in the procession "was not propelled by draft
animals, but slid forward by machinery hidden below" (Philostratus, VS 550).
The most natural place to look for imitations of the court miracles of the early
Ptolemies is, of course, the Roman imperial court. It is hardly mere coincidence
that some of the Roman gadgets mentioned in historical sources concern the
emperor Nero and the contemporary satirical figure Trimalchio. Petronius (Sat.
54.4) speaks about walls in the triclinium of Trimalchio, which might split open to
let in "some automata," perhaps referring to a practical joke actually realized at one
of Nero's parties. In the Golden House of Nero, there was, among other technical
refinements, a central "circular banquet hall, which revolved incessantly, day and
night, like the heavens" (Suet., Ner. 31.2). Excavations in a central octagonal hall
probably identical with the banquet hall reveal that this mysterious rotation may
have been effected by water power (Priickner and Stortz 1974: 338-39). The ref-
erence to the revolution of the heavens may indicate that it was, in fact, not the
banquet hall that turned, but rather a shell inside the vaulted roof. A similar
construction was described almost a century earlier in an elaborate dining room at
Varro's villa in Casinum (Rust. 3.5,17). A later passage in Pappus (Collectio 8.2)
proves the survival of this type of construction into the fourth century; it mentions
"sphere-makers, who construct models of the heavens based on the even and
circular motion of water" (Oleson 1984: 65).
The pieces of evidence preserved are few, but considering the impressive, high-
technology development demonstrated by the Antikythera mechanism, which has
survived by the purest of accidents, it would be unwise to underestimate the
accomplishments of imperial Rome in this area. Nevertheless, the direct technol-
ogy transfer that took place from the Roman Empire to medieval western Europe
as far as water-mills and similar machines are concerned, had apparently no
counterparts in high technology. Here, the very limited knowledge actually
transferred to the European Middle Ages apparently took the way through By-
zantium and the Islamic world.
REFERENCES
Bedini, S. 1964. "The role of automata in the history of technology," Technology and
Culture 5: 24-41.
Bromley, A. G. 1986. "Notes on the Antikythera Mechanism," Centaurus 29: 5-27.
Buchner, E. 1971. "Antike Reiseuhren," Chiron 1: 457-82.
Dalley, S., and J.P. Oleson 2003. "Sennacherib, Archimedes, and the water screw,"
Technology and Culture 44: 1-26.
Deppert-Lippitz, B. 1995. Die Schraube zwischen Macht und Pracht: Das Gewinde in der
Antike. Sigmaringen: Jan Thorbecke.
Diels, H. 1915. "Uber Platons Nachtuhr," Sitzungsberichte der Preussischen Akademie der
Wissenschaften 2: 824-30.
Drachmann, A.G. 1932. Ancient oil mills and presses. Det Kgl. Danske Videnskabernes
Selskab. Archaeologisk-kunsthistoriske Meddelelser 1.1. Copenhagen: Levin and
Munksgaard.
Drachmann, A. G. 1958. "The screw of Archimedes," in Actes du VIIIe Congres international
d'Histoire des Sciences, Vol. 3. Milan, 940-43.
Drachmann, A.G. 1963. The mechanical technology of Greek and Roman antiquity: A study of
the literary sources. Copenhagen: Munksgaard.
Drachmann, A.G. 1973. "The crank in Graeco-Roman Antiquity," in T. Mikulas and R.
Young (eds.), Changing perspectives in the history of science. London: Heinemann,
33-51.
Drachmann, A. G. 1976. "Ktesibios's waterclock and Heron's adjustable siphon," Cen-
taurus 20: 1-10.
Field, J. V., and M. T. Wright 1985a. Early gearing: Geared mechanisms in the ancient and
mediaeval world. London: Science Museum.
Field, J. V., and M. T. Wright 1985b. "Gears from the Byzantines," Annals of Science 42:
87-138.
Foley, V., W. Soedel, J. Turner, and B. Wilhoite 1982. "The origin of gearing," History of
Technology 7: 101-29.
Gaitzsch, W. 1983. "Die 'romische' Schraube aus dem Kastell von Niederbieber," Bonner
Jahrbucher 183: 595-602.
Gounares, G. 1978. "Chalkino phoreto heliako horologio apo tous Philippous," Arch-
aiologike Ephemeris 1978: 181-91.
Hill, D.R. 1976. On the construction of water-clocks: Kitab Arshimidas ft' amal al-binkamat.
London: Turner and Devereaux.
Kauffmann, G. 1896. "Astrolabium," RE 2, cols. 1798-1802.
14-22.
King, H. C., with J. R. Millburn 1978. Geared to the stars: The evolution of planetariums,
orreries, and astronomical clocks. Toronto: University of Toronto Press.
Klumbach, H. (ed.) 1973. Spiitromische Gardehelme. Miinchner Beitrage zur Vor- und
Friihgeschichte 15. Munich: Beck.
Landels, J. G. 2000. Engineering in the ancient world. Rev. ed. London: Constable.
Lendle, 0. 1990. "Vitruvs Meilenzahler (De Arch. 10.9.1-4)," in W. Goerler and S. Koster
(eds.), Pratum Saraviense: Festgabe fur Peter Steinmetz. Stuttgart: F. Steiner, 75-88.
Hull Press.

University Press.
Mutz, A. 1969. "Romische Bronzegewinde," Technikgeschichte 36:161-67.
Price, D. J. de Solla. 1962. "Automata and the origins of mechanism and mechanistic
philosophy," Technology and Culture 5: 9-23.
Price, D. J. de Solla. 1969. "Portable sundials in antiquity, including an account of a new
example from Aphrodesias," Centaurus 14: 242-66.
Price, D. J. de Solla. 1974. Gears from the Greeks: The Antikythera Mechanism-A calendar
computer from ca. Bo B.C. Transactions of the American Philosophical Society 64.7.
Philadelphia: American Philosophical Society.
Priickner, H., and S. Stortz 1974. "Beobachtungen im Oktagon des Domus Aurea," Ro-
mische Mitteilungen 81: 323-39.
Schmidt, W. 1899. Heronis Alexandrini Opera quae supersunt omnia. Vol. 1, Pneumatica et
Automata. Leipzig: Teubner.
Sleeswyk, A. W. 1979. "Vitruvius' waywiser," Archives Internationales d'Histoire des Sciences
29:11-22.
Sleeswyk, A. W. 1981a. "Hand-cranking in Egyptian antiquity," History of Technology 6:
23-37.
Sleeswyk, A. W. 1981b. "Vitruvius' odometer," Scientific American 245.4: 158-71.
Sleeswyk, A. W. 1990. "Archimedes' odometer and waterclock," in Ancient technology,
Finnish Institute at Athens, symposium held 30.3-4.4. 1987. Tekniikan museon julk-
aisuja 5. Helsinki: Tekniikan museon. 23-37.
Wright, M. T. 2002. "A planetarium display for the Antikythera Mechanism," Horological
Journal 144: 169-73.
Wright, M. T. 2006. "The Antikythera Mechanism and the early history of the moon-phase
display," Antiquarian Horology 29.3: 319-29.
CHAPTER 32
INVENTORS,
INVENTION, AND
ATTITUDES TOW ARD
TECHNOLOGY AND
INNOVATION
KEVIN GREENE
THIS chapter has three parts. The first section looks at a selection of comments
about invention in ancient sources; the second considers the nature of some an-
cient inventors; and the third surveys examples of Greek and Roman technology
that reveal stability or change.
INVENTORS, INVENTION, AND ATTITUDES TOWARD INNOVATION 801
ATTITUDES TO INVENTION AND

INNOVATION IN GREEK AND
ROMAN WRITERS
Optimism, Pessimism, and Human Ingenuity

What marvelous, stupendous accomplishments human effort has achieved in the
fields of construction and textile production! How far it has progressed in agri-
culture and navigation! What accomplishments of imagination and application
in the production of all kinds of vessels, various types of statues and paintings ....
With what sharp minds humans grasp the paths and arrangement of the heav-
enly bodies! How great the knowledge of this world humans have filled them-
selves with! Who could describe it? (St. Augustine, de Civ. D. 22.24)
In the context of the western Empire in the early fifth century A.D., St. Augustine's
assessment of the accumulated worldly achievements of Greek and Roman civili-
zation is probably tempered by consciousness of its fragility, but credit for both
useful and entertaining inventions is firmly attributed to the human intellect. This
theme runs through many classical works, despite expressions of regret about de-
cline from a better past or criticisms of utilitarian knowledge and labor. Hesiod's
gloomy perception of a "fifth race" anticipated further destruction: "For now is the
iron race, when humans never will cease from labor and sorrow by day and from
suffering at night, since the gods will give only grievous concerns.... And Zeus will
destroy this race of mortals too" (Hesiod, Op. 107-78). Plato associated a lack of iron
with a lost life of simplicity when the gods provided humans with all they needed for
material comfort-clothing, houses, and pottery (Plato, Leg. 3.677a-679b).
While many writers made obvious points about necessity prompting invention,
a more interesting theme is that of humans gaining inspiration from nature. Lu-
cretius credited humans with the ability to experiment: "Nature herself was the
mother who first brought forth the model of sowing and the beginning of grafting:
berries and acorns, having fallen from the trees, sent out swarms of sprouts on the
ground beneath in proper season.... Then they tried one method after another to
cultivate their dear little plot ... " (Lucretius 5.1361-69).
Vitruvius' knowledge of Hellenistic technical treatises gave him a perception of
the development of things that "were more easily done with machines [machinae]
and their revolutions, some others with instruments [organa]" and how their au-
thors had gradually improved "by their learning all those things that they believed
useful for research, for the arts, and for established traditions" (De arch. 10.1.4-6).
Nevertheless, nature remained the inspiration for human ingenuity: "Now all
machinery has been modeled on Nature and set up with the revolution of the
universe as instructor and teacher.... So when our ancestors noticed that this was
the case, they took their models from Nature and imitating her, led on by divine
help, they developed the amenities of life" (Vitruvius, De arch. 10.1.1-4).
Contemporary Perceptions of Machines

The special character of devices powered by water rather than by human or animal
labor, such as mills or water-wheels, impressed Greek and Roman writers over
many centuries, from Antipater to Procopius (Humphrey et al. 1998: 31-34). A
noria (water-lifting wheel powered by paddles) caught the imagination of one
Latin poet: "It pours out and scoops up water, discharging on high the stream it
carries, and it drinks up a river only to disgorge it. A marvelous achievement! It
carries water and is carried along by water. In this fashion a stream is pushed up by
a stream, and a new device scoops up the old-fashioned fluid" (Anth. Lat. 284).
A careful observer could deduce how a watermill or noria worked, whereas many
Hellenistic automata produced magical effects from invisible mechanisms. Just as in
eighteenth-century Europe, automata expressed the ambiguity of mechanical de-
vices that mimicked nature (Dugan and Dugan 2000: 40-42) and led to speculation
about robots and flying model birds (Humphrey et al. 1998: 61-62): "Craftsmen
prepare a device whose first principle they conceal, so that only the wonder of the
machine is apparent, the cause hidden" ([Aristotle], Pr. Preface, 848a.20-38).
Pappus reiterated a fundamental classification of machines in the fourth century
A.D.: "There are five machines by the use of which a given weight is moved by a given
force, and we will undertake to give the forms, the applications, and the names of
these machines.... The names are as follows: wheel and axle, lever, system of pulleys,
wedge, and, finally, the so-called endless screw" (Pappus, Coll. 8.52). Pappus also
repeated Hero's specifications for the education of mechanicians, and defined their
areas of operation: "conjurers" (manganarioi) or mechanicians (mechanikoi) who
move weights; constructors of siege machines (organopoioi); machine-builders
(mechanopoioi), especially for water-lifting; gadget-designers (thaumasiourgoi) who
make automata, clocks, and so forth; and sphere-makers, who construct models of
the heavens. It is interesting to note the persistence of an association between con-
juring and mechanical ability (Pappus, Coll. 8.1-2; Humphrey et al. 1998: 47-48).
There was widespread awareness of the strategic importance of fostering the-
oretical and practical knowledge of technology in the Hellenistic world. Philo of
Byzantium observed that royal patronage in Alexandria encompassed both forms of
knowledge, for craftsmen were "heavily subsidized because they had ambitious
kings who fostered craftsmanship. Not everything can be accomplished by the
theoretical methods of pure mechanics, but much is to be found by experiment"
(Bel. 50.3). Long before the famous engagement of Archimedes in the defense of
Syracuse in 212 B.C., Dionysius I (ca. 430-367 B.C.) had assembled a "think tank" to
generate military hardware-although Traina (1994: 68-69) considers modification
or development of the catapult to have been more likely than actual invention. The
focused work groups, high salaries, and performance bonuses all sound surprisingly
modem. Diodorus Siculus (14.41.3-4, 42.1) recounts that
Dionysius, therefore, immediately assembled technicians, commanding them to

come from the cities he ruled, and luring them from Italy and Greece-and
even from Carthaginian territory-with high wages. For he intended to manu-
facture weapons in great numbers and projectiles of every sort, and in addition
tetraremes and quinqueremes-no quinqueremes yet having been constructed

at that time. After assembling a great number of technicians, he divided them
into work-groups according to each one's own talents .... In fact, the catapult was
invented in Syracuse on this occasion, since the most able technicians were
gathered together from all over into one place. The high wages stimulated
their enthusiasm, along with the numerous prizes offered to those judged the
best.
Ingenuity and the Status of Work

One of the best-known traditions in Greek and Roman literature is the association
of leisure with an honorable life-Finley's "primitive" representation of the an-
cient economy (1973) relies heavily on generalizing the view of the literary elite to
the rest of society. Seneca (Ep. 90.10-13) distinguished between wisdom and in-
genuity, and attributed invention to the latter:
I differ from Posidonius, when he concludes that tools for the crafts were invented
by wise men .... It was human ingenuity, not human wisdom, that invented
those things. In this also I differ from him, that it was the wise who devised mines
for iron and copper when the earth, scorched by forest fires, poured out metal
from liquefied surface veins of ore. The sort of men who discovered these mines
are the same sort who work them today. Nor does the question of whether the
hammer or the tongs came first seem as subtle to me as it did to Posidonius. Both
were invented by someone with a mind that was nimble and sharp, but not
great or elevated-along with anything else that must be sought with a bent body
and a mind focused on the ground.
Seneca's specific example echoes the general statement by Aristotle that "we term
'banausic' those crafts that make the condition of the body worse, and the
workshops where wages are earned, for they leave the mind preoccupied and
debased" (Pol. 8.2.1 [1337b]). This attitude was foreshadowed by Xenophon ( Oec.
4.2-3). However, approval of physical activity was also expressed in Greek and
Roman literature (Hesiod, Op. 303-11, St. Paul, Thes. 3: 6-10). Furthermore, the
living associates of deceased individuals saw no shame in commemorating them
with gravestones that named their crafts or trades, and even pictured some of them
at work in sculpted scenes (Zimmer 1982). Agriculture was normally considered an
honorable exception among forms of human work, and, in addition, agricultural
labor enhanced physical fitness for war ([Aristotle], Oec. 1.2.2-3). Such exceptions
smoothed the way for machines associated with irrigation and the processing of
crops, and all devices connected with warfare.
Inventions
Antibanausic judgments remind us that we are not dealing with the utilitarian
values of nineteenth-century Europe: "As more and more arts were discovered,
some pertaining to necessities and some to pastimes, the inventors of the latter
were always considered wiser than the inventors of the former, because their
knowledge was not oriented toward utility" (Aristotle, Metaph. 1.1.11-17 [981a-
982a]). Finley (1965: 33-34) contrasted Pliny the Eider's passion for naming in-
ventors of nonutilitarian arts and crafts with his inability to identify the inventors
of new machines such as the screw-press. He also cited two stories (Humphrey et al.
1998: 595) in which inventors were penalized or rejected, rather than rewarded; he
argued that these "technological fables"-one about unbreakable glass (involving
Tiberius), the other about a device for lifting columns (involving Vespasian)-
demonstrated a real antagonism toward innovation that underpinned a culture of
stagnation. Like Plutarch's comments about Archimedes, however, such stories
were told to illustrate the character of an individual rather than to express an-
tipathy to inventions (Greene 2000: 46-50).
In the second century B.C., Polybius noted that Romans copied Greek shields:
"For this too is one of their virtues, to adopt new customs and emulate what is bet-
ter" (6.25.10-11). In the following century, Cicero offered unambiguous praise for the
qualities of Romans, who "have constantly shown themselves more clever at in-
vention than the Greeks, or have improved on what they received from the Greeks-
at least in those fields they judged worthy of their effort" (Tusc. 1.1.2). However, by
the first century A.D. Pliny the Elder was complaining that Roman achievements had
been accompanied by ignorance of Greek knowledge (HN 14.2-4):
The research of the men of long ago was so much more productive or their
industry so much more fortunate when, a thousand years ago at the very begin-
nings of literacy, Hesiod began to publish his instructions to farmers and nu-
merous others followed his line of research. For this reason our task is greater,
since now we have to investigate not only what was found out later, but also the
discoveries made by the pioneers.
Pliny's near-contemporary Frontinus appears to confirm fears of compla-
cency, for in a discussion of engineering works and catapults he observed, "I see no
further scope for the applied arts" (Str. 3, Preface). This view, however, is belied by
the anonymous fourth-century A.D. writer who set out designs for many devices to
assist the late Roman army, including an ox-powered paddle-wheel warship using
"animal power, enhanced by the resources of the human intellect"-specifically,
90-degree gears familiar from mechanical mills, and wheels with projecting vanes
of the kind used for lifting water from rivers (De rebus bellicis 17.1-3).
GREEK AND ROMAN INVENTORS
But as for those benefits to humans that lay hidden in the earth-bronze, iron,
silver, and gold-who else before me could claim to have found them
first? ... Learn the whole matter in a brief phrase: all arts possessed by mortals
come from Prometheus. (Aeschylus, PV 442-506)
Identifying Inventors in the Ancient World

Many fundamental crafts or inventions were associated with gods-for example,
weaving (Athena), agriculture (Demeter), fire (Prometheus/Hephaestus), or wine
(Dionysus). Only in recent centuries have inventions been ascribed to named
individuals with any regularity, for the good reason that no clear concept of in-
tellectual property rights backed by law existed before the fifteenth century A.D.
(May 2002). Most modem classicists and archaeologists, when challenged to name
Greek or Roman technical inventors or inventions, rarely get further than Archi-
medes and his water-screw or Heron and his steam-driven rotating sphere. Rather
than indicating that invention was rare, this may indicate that we are looking for the
wrong thing. In a preindustrial society, we should perhaps focus on processes by
which technological innovations might be diffused and implemented, whether or
not they were inventions sensu strictu.
How many of the names associated with technical inventions in ancient sources
may be relied upon? Archimedes' water-screw raises most of the difficulties asso-
ciated with this question (Dalley and Oleson 2003). The plentiful archaeological
evidence and pictorial representations of water-screws in use dates from the first
century A.D. at the earliest. Literary evidence begins in the second half of the third
century B.C., however, and Vitruvius provided practical instructions for building
one in the first century B.C. (without naming Archimedes). Diodorus Siculus (mid-
first century B.C.) refers to the "so-called Egyptian screw, which Archimedes the
Syracusan invented, called the screw on account of its design," but the Greek verb
eurisko could mean "found" or "observed" as well as "discovered." The matter is
complicated by Dalley's claim that this device had already been invented in Me-
sopotamia in the seventh century B.C., was introduced to Egypt when it fell under
Assyrian control, and was observed there by Archimedes in the third century B.C.
Thus, we have ambiguous vocabulary and contested origins for one of the most
frequently cited technological inventions of the ancient world.
Pliny the Elder compiled comprehensive lists of origins in his Natural History
(7.191-209), considering discoveries to be part of the subject of human nature. A
short extract (HN 7.195-97) illustrates both his all-embracing approach and his
inability to resolve differences between authorities:
Tiles were invented by Cinyra, son of Agriopa, as well as mining for copper, both
in the island of Cyprus, and also the tongs, hammer, crowbar and anvil; wells
by Danaus who came from Egypt to Greece to the region that used to be called Dry
Argos; stone quarrying by Cadmus at Thebes, or according to Theophrastus,
in Phoenicia; walls were introduced by Thrason, towers by the Cyclopes accord-
ing to Aristotle but according to Theophrastus by the Tirynthians; woven fab-
rics by the Egyptians, dyeing woolen stuffs by the Lydians at Sardis, the use of the
spindle in the manufacture of woolen by Closter son of Arachne, linen and nets
by Arachne, the fuller's craft by Nicias of Megara, the shoemaker's by Tychius
of Boeotia.
On the other hand, people with technical skills were brought together with the
purpose of producing useful devices, notably at the Museum at Alexandria in Egypt
and at Syracuse in Sicily, where royal houses were "interested as much in broad-
casting the prestige and imagery of technological success, the tangible evidence of
human triumph over the forces of nature, as in the practical results" (Dalley and
Oleson 2003: 25). Individuals such as Ctesibius, Philon, and Archimedes, whose
names became associated with specific inventions, were at the very least responsible
for memorable innovations, whether or not the devices that they brought into
practical use were their own inventions.
Archimedes (287-212/11 B.C.)

Archimedes' reputation is largely based not on technical inventions but on math-
ematical skills and a scientific discovery. His popular renown relies upon the flash
of brilliance (Eureka!) involved in relating the level of bathwater to the measure-
ment of specific gravity and on an antiheroic (but culturally symbolic) death while
attempting to complete some calculations during the fall of Syracuse to the Ro-
mans. Toomer (1996) saw Archimedes as "the greatest mathematician of antiquity"
and credited him with the invention of hydrostatics, but he made much less of his
role as technical inventor: "Popular history knew him as the inventor of marvelous
machines used against the Romans at the siege of Syracuse, and of devices such as
the screw for raising water." The idea that science is somehow debased by practical
application was attributed to Archimedes several centuries after his death but is
often held to encapsulate the attitude of the entire ancient world:
He possessed so high a spirit, so profound a soul, and such treasures of scien-
tific knowledge that, though these inventions had obtained for him the re-
nown of more than human sagacity, he yet would not deign to leave behind
him any written work on such subjects, but, regarding as ignoble and sordid
the business of mechanics and every sort of art which is directed to use and
profit, he placed his whole ambition in those speculations in whose beauty
and subtlety there is no admixture of the common needs of life. (Plutarch, Marc.
17.3-4)
Wilson has argued that Plutarch's point was to show how unusual Archimedes
was in not putting technological knowledge to practical uses (2002: 4). Astron-
omy and mathematics were entirely respectable pursuits, while defending a city or
raising royal revenues clearly placed mechanics in a worthy military and official
context.
The most familiar device to which Archimedes' name is attached today is the
water-screw, a cylinder with a close-fitting internal spiral that lifted water when the
cylinder was rotated, but the association with Archimedes is not beyond all doubt
(above). The water-screw is a good example of the kind of device associated with
the Museum at Alexandria that involved scientific or mathematical principles but
could be applied to the productivity of irrigated farmland on which the prosper-
ity of Egypt depended. Furthermore, the relationship between Archimedes and
Hieron II of Syracuse-who is also famous for creating a ship of unparalleled size
(Athenaeus, Deip. 5.206e-209b )-inevitably draws attention to the relationship

between hydrostatics and shipbuilding (Pomey and Tchernia 2005). The renown of
Archimedes' astronomical spheres (Cicero, Rep. 1.14.21-22) implies command of
precision engineering as well as large machines. Lewis credits Archimedes with the
invention of the hodometer, another very practical application of science to the
needs of the state, for it could be used in contexts such as the marking of distances
on the new Roman highways that were being extended through Italy (2000b: 632).
Thus, if the superficial impression gained from Plutarch is set aside, Archimedes
appears similar to other Hellenistic inventors.
Ctesibius (fl. 270 B.C.)

The reputation of Ctesibius (Drachmann 1948: 1-41) is not based on surviving
writings, but on authorship of treatises on machines that were quoted extensively
by Philon, Vitruvius, and Heron. Such works give a clear impression that "the art
of the ancient mechanikos apparently combined enquiries into the theory and
design of devices with meticulous concern over details of their manufacture and
operation" (Knorr 1996). As with Archimedes and the water-screw, it will never be
possible to claim definitive priority for ancient inventors and the devices that are
attributed to them, but Ctesibius is generally credited with inventing the force-
pump, a keyboard-operated organ in which a constant flow of air was assured by
water pressure, and with adding accuracy to the water-dock by ensuring a regular
flow of water. His writings about artillery (known largely from Philon) probably
put existing developments into a systematic form, since the innovation of using
torsion springs on catapults has been attributed to unnamed engineers working for
Philip II around 340 B.C.
Along with the Archimedean water-screw, the force-pump-known to Vi-
truvius (10.6-7) as machina ctesibica-is one of the very few ancient inventions to
be named after an individual in antiquity. The water-dock and organ were not basic
inventions, but refinements to existing devices involving the creation of an even
flow of water or air. The twin-chambered pump, in which valves and pistons
opened and closed alternately, also generated a continuous flow of water. The
pneumatic devices associated with Ctesibius fit well into the range of activities that
would have interested a Hellenistic ruler and the concerns of the Museum of
Alexandria: water lifting, time keeping, and warfare. Vitruvius commented that in
addition to the force-pump, Ctesibian devices included "many others of varied
design driven by water pressure. Pneumatic pressure can be seen to provide effects
which imitate nature, such as singing of blackbirds and mechanical acrobats, and
little figures which drink and move, and other things that delight the senses by
pleasing the eye and catching the ear (De arch. 10.7.1-4). The symposium and the
entertainment of guests was an important state activity, while automata associated
with temple rituals had another clearly utilitarian function in managing belief
(Schiirmann 1991).
Philon of Byzantium (fl. 200 B.C.)

The name of Philon of Byzantium is not linked to any inventions, and much of his
surviving work is concerned with artillery and siege works (Drachmann 1948: 41-
74). His use of Ctesibius (and the subsequent use of both by Heron) contributed to
the preservation and transmission of a body of knowledge that allowed Hill to speak
of "a Greek school of mechanical engineering, stemming from Ctesibius (300-270
B.C.) and continuing through Philo and Hero, probably into Byzantium, whence it
passed into Sasanid Persia" (1979: 21).
Vitruvius (fl. 35/25 B.C.)

Knowledge of the achievements of the "Greek school of mechanical engineering" was
made available in Latin in book 10 of Vitruvius' De architectura just as the Roman
Empire began to be created under Augustus. Vitruvius is most closely associated with
architecture (Howe and Rowland 1999), but his book on mechanics is particularly
interesting to technological historians. Indeed, many of the devices attributed to
Greek inventors-such as the hydraulic organ-proliferated in Roman times (Fleury
1993: 198-204). Vitruvius is frequently described in rather patronizing terms, as if
Greek science had already given way to Roman pragmatism, despite his prolific
citation of Greek philosophers, scientists and mathematicians. Vitruvius and Heron
are particularly important because their works survive so much better than those of
Ctesibius or Philon, while containing much material inherited from them. The
cultural context is also particularly interesting, since Vitruvius is the first "Roman"
writer on engineering, whereas the other major figures from Archimedes to Heron
wrote in Greek and were connected, more or less directly, to the theoretical and
applied mathematics and mechanics of the Museum of Alexandria. Few general
writers associate Vitruvius with any topic other than architecture, and most fail to
comment on his significance as a historian of science and philosophy who "appre-
ciated that in its general and most humane form, architecture included everything
which touches on the physical and intellectual life of man and his surroundings"
(Tomlinson and Vallance 1996).
Like Archimedes, Vitruvius throws doubt on the supposed conflict between
theory and practice in the ancient world, "often employing the theories of the most
antibanausic Greek thinkers to elucidate his very practical subject.... For a man
with interests practical and theoretical in equal measure, understanding the nature
of nature was central to all" (Tomlinson and Vallance 1996). Perhaps because of the
widespread belief that techne was contrary to nature, Vitruvius related the move-
ment of machinery to that of the universe in a rather pious manner that contrasts
with his practical description of the optimization of human power by means of
machines such as catapults or oil presses (De arch. 10. 1. 1-4).
Heron of Alexandria (fl. A.D. 62)

In addition to maintaining extensive knowledge of mathematics and science as a
"Greek" writer operating in the early Roman Empire, Heron of Alexandria wrote
about the same technical subjects as Ctesibius and Philon: artillery, surveying,
mechanics, pneumatics, and so forth. Drachmann (1948: 74-99) considered his

technical work to be working notes rather than a finished treatise (1948: 159). His
popular reputation as an "inventor" rests largely on items that appear to anticipate
important medieval and later developments. His description of a Ctesibian water-
organ includes a pump powered by a windmill and trip-hammers, but most at-
tention has been given to the aeolipile, a primitive reaction turbine persistently and
incorrectly described as "an early form of steam engine" (e.g., Wilkinson 2000: 147).
The aeolipile is one of many devices that use hydraulic or pneumatic power assisted
by heating, and it was designed to demonstrate a scientific principle, not to serve as
a power source (Keyser 1992). Despite wise words from Drachmann (1963: 260), it
has carried a great historiographical burden as a symbol of the ancient world's
inability to exploit technology:
You can hear people contend, on the strength of such play-things, that the an-
cient Greeks could have invented the steam engine, if only they did not have the
slaves, which made such an invention superfluous. But slave labor was not cheap,
and the presence of slaves did not prevent the invention of the watermill, which
could be constructed by the means in hand. The construction of the steam en-
gine had to wait until it was possible to make iron pipes and put them together
with screws.
STABILITY, CONTINUOUS
DEVELOPMENT, AND CHANGE IN
GREEK AND ROMAN TECHNOLOGY
Stability
It is important to stress that many areas of Greek and Roman technology remained
remarkably static. No fundamental changes took place in techniques of forging or
casting metal objects, whether made from individual metals or alloys, although
brass came into regular use by the Roman period (Craddock 1990; see also chapter
4). Most changes were increases in scale: producing a life-size bronze statue for a
Greek temple clearly demanded more people and better workshop organization
than casting prehistoric bronze axes (Mattusch 1996; see also chapter 16). Likewise,
the scale of demand for precious and base metals in the Roman Empire exceeded
anything known previously, and the extraction of metals from mines provided an
opportunity for the diffusion of many Hellenistic inventions for water-lifting and
for using reciprocating machinery for crushing ores (Lewis 1997; Wilson 2002).
Thus, a static manufacturing technology can coexist with innovations in extraction
and processing.
Methods used in forming ceramics by hand or on a potter's wheel all existed

long before archaic Greek times and persisted through the Roman Empire; finished
vessels could be fired in simple bonfires or permanent kilns (Peacock 1982; see also
chapter 19). Surface treatments included vitreous glazes, glossy slips, and painted
decoration, but changes were the result of aesthetic or economic factors rather than
technology. Potters making the relief-molded bowls that became popular in Hel-
lenistic times were able to adopt the kinds of molds employed in casting their metal
prototypes; they were already accustomed to molding other ceramic products,
notably figurines, and often worked closely with metalworkers (Merker 2003). Most
changes in the production, distribution, and consumption of pottery result from
cultural or economic factors rather than technology. The diffusion of Greek and
Roman cuisine demanded new serving and preparation vessels (Bats 1996), while
classical architecture required the dissemination of knowledge about making spe-
cialized ceramic bricks, roof tiles and water pipes.
Continuous Development
Continuous development occurred in one fundamentally important aspect of
Greek and Roman civilization: warfare (chapters 26 and 27). Logistic and organi-
zational innovations were demanded by changing scales of operation-conflict
between city-states, war with Persia, Alexander's eastern conquests-culminating
in the mature Roman Empire, which required permanent garrisons, frontier de-
fenses and a professional standing army. Technical changes in equipment included
the "invention" of mechanical catapults by Dionysius I's "think tank" in 399 B.C.
Torsion springs improved Hellenistic artillery devices a little later, and remained
the essential form of propulsion throughout the Roman imperial period. Archae-
ological study of surviving remains has revealed continual modification and im-
provement of this mechanism, as well as proliferation of its use (Marsden 1969;
Baatz 1994). A heavier device for throwing large stones-the onager-was brought
into extensive use by the third century A.D., presumably because devices suited to
defensive rather than offensive warfare were required in the face of growing threats
from "barbarians" beyond the frontiers.
Siege warfare became particularly important in later classical and Hellenistic
Greece, and vivid accounts reveal valuable details. In terms of attacking a defended
city, Demetrius I of Macedonia (336-283 B.C.), known as Poliorketes-"Besieger of
Cities"-used large constructions and technical devices in a prolonged but ulti-
mately unsuccessful siege of Rhodes. Defensive strategies are associated with Ar-
chimedes (287-212/11 B.C.), who turned scientific knowledge into technological
inventions to counter the siege of Syracuse by Roman attackers. With the exception
of episodes of warfare with their eastern neighbors, the Roman imperial army rarely
had occasion either to besiege walled cities, or to defend their own cities from
attackers equipped with machinery; the principal threats came from invading
Germanic and other peoples who lacked technical resources. Physical barriers such
as Hadrian's Wall (defending northern Britain, second century A.D.) or the Ana-
stasian Wall (protecting Constantinople's western approaches, fifth century A.D.)
gained defense in depth as city walls became increasingly elaborate, reinforced by
multiple ditches and projecting towers to maximize the effectiveness of defense by
long-range mechanical artillery and shorter-range archery. Similar fortifications
and mechanical artillery devices continued in later Roman and Byzantine times;
indeed, a nameless fourth-century writer proposed a number of technical and
organizational innovations that were of sufficient interest in subsequent centuries
for copies of the text and illustrations of De rebus bellicis to survive (Thompson
1952; Hassall 1979).
The personal equipment of soldiers underwent considerable change and de-
velopment from classical Greek to Roman times, notably in the quantity of pro-
tective armor or chain mail available to ordinary soldiers. Plentiful weaponry and
armor required extensive iron production and workshops for the manufacture and
repair of equipment; clearly, a professional standing army provided the best context
for infrastructure of this kind. Nevertheless, many items that became standard
features of Roman armor and equipment had originally been adopted from ene-
mies (Bishop and Coulston 1993: 194-95).
At sea, Greek naval power relied on the trireme galley with a projecting prow
for ramming other ships; it was probably adopted from Carthage or Egypt.
Competition between navies led to larger galleys with more oars and rowers-such
as quinqueremes-better suited to boarding than to high-speed ramming. Trir-
emes remained the staple of Roman naval power, as there were no longer any
competing navies in the Mediterranean to stimulate greater size or further devel-
opment. Away from the Mediterranean, Roman imperial military ships had to cope
with large rivers such as the Rhine, and seas characterized by tides and large waves.
Archaeological finds have revealed a mixture of Mediterranean-style galleys and
more robust local vessels built in a different tradition using overlapping planks and
nails, possibly related to the Venetic ships that had given Caesar so much trouble in
the first century B.C. (Greene 1986:17-34).
Roman imperial buildings, such as the Coliseum (late first century A.D.), the
Baths of Caracalla (ca. 200 A.D.) or the Basilica ofMaxentius (ca. 300 A.D.), bear
little resemblance to anything Greek predating the first century B.C. Their forms are
related to the concerns and activities of the Roman state and its relationship with its
citizens. Every component: bricks, arches, vaulting, concrete, surface decoration
with exotic stone-can be traced back in one way or another to Greek architectural
and ornamental features ultimately adopted from Mesopotamia and Egypt. What is
new is a kind of building where internal space is supremely important (Delaine
2002; chapter 10 )-but is a vaulted bathing establishment or an imperial reception
hall an invention? Hydraulic concrete is frequently cited as a Roman invention, but
the term "discovery" seems more appropriate. Rather than being developed in-
tentionally, it resulted from observing the properties of mortar in areas of Italy
where volcanic sand was readily available. The chemical reaction between volcanic
sand and mortar that allowed it to set underwater led to export of this sand for
creating harbors or waterproofing aqueducts; this is a clear example of innovation

leading to technology-in-use (Oleson and Branton 1992; Oleson et al. 2005, 2006).
Some classical Greek and Hellenistic inventions were directly linked to devel-
opments in science, notably instruments associated with astronomy and medicine
(see chapters 29 and 31). Scientific analysis of harmony may have had an impact on a
further class of objects for which little physical or documentary evidence survives:
musical instruments (Landels 1998). There are direct links between practical ap-
plication and abstract scientific knowledge in the water-organ associated with
Ctesibius. Representations of lyres indicate changes in size, number of strings, and
mechanisms for tuning that may have had a scientific basis.
A particularly interesting example of continuous development is coinage
(Chapter 30, part 2). These small discs of metal were invented around 600 B.C., and,
following the example of Lydia, provided Greek city-states with a convenient
method of discharging obligations in pieces of bullion of guaranteed purity (Car-
radice and Price 1988). Their exchangeability, however, allowed a diversification of
economic activities that broke tight social links with known and trusted partners,
and provided new means of wealth storage and transfer over long distances. Smaller
denominations were gradually introduced in base metals, and systematically re-
organized in the early Roman Empire into something resembling modern ideas of
token currency (Greene 1986: 45-52). Once coins had become a regular part of
commercial exchanges, continual adjustments were necessary to maintain the cor-
rect weight ratios between gold, silver, and base denominations, requiring careful
production of consistent alloys in different mints. In terms of technologies that
change the way that people engage with the everyday world through material
objects, coins were a major Greek invention, while the formalization of denomi-
nations by Augustus was a significant innovation that diffused to become tech-
nology-in-use from Britain to Egypt.
Stepwise Change
Glass is the best example of significant technological change in an industrial product
in classical times, and it illustrates the care with which terms need to be used to
distinguish between invention and innovation (cf. chapter 21). The raw material
had been discovered or invented in Mesopotamia and/or Egypt by 3000 B.C.,
probably as a by-product of heating minerals and chemicals for other purposes, and
innovated as a means of adding a decorative coating to clay or faience. Hollow
vessels made entirely from glass were an invention, but their production required
no innovation since the raw material already existed, and early vessels were formed
on a clay core, which-in contrast to that of glass-coated objects-was then re-
moved. However, a dramatic change occurred in the first century B.C. when the
discovery that a drop of molten glass on the end of a hollow rod could be expanded
into a bubble led to the invention of hollow blown glass vessels (Israeli 1991). Since
an extensive glass industry already existed, putting a new technique into use and
creating a new range of hollow vessels was an innovation, which led to further
technical innovations such as decorating vessels in relief by blowing them into a
mold. In addition there was a considerable proliferation of decorative effects (for
example that the inclusion of gold leaf), as well as diversification of production into
mirrors, window panes, and mosaic cubes for wall decorations.
Our understanding of the adoption of blowing and the subsequent prolifera-
tion of forms relies on the excavation of workshops and studies of vessels in
museum collections (Weinberg 1988; Whitehouse 1997, 2001). Perhaps the most
interesting effect of changes in production technology was that blowing trans-
formed glass vessels from luxuries to mass-produced everyday items (Fleming
1999 ). After the brief period of invention and innovation between about 50 B.C. and
A.D. 50, it provides an example of Edgerton's concept of technology-in-use (1999:
112) and draws its significance from the fact that glass vessels came to be used by a
wider social range of people. Glass was produced in increasing quantities and in a
larger number of forms in the following centuries as production diffused to many
parts of the Roman Empire.
The codex, in which separate leaves are bound along one edge between stiff
covers, rarely appears in popular lists of Roman inventions, but it had displaced the
scroll by the late Roman period and has remained the standard form of book ever
since (Roberts and Skeat 1983; chapter 28). The earliest examples were small note-
books of parchment leaves modeled on collections of wooden writing tablets bound
together in sets. A codex could accommodate much longer texts than a scroll, and it
was better suited to parchment, which (although expensive) was more durable than
Egyptian papyrus. It is not hard to see the functional convenience of the codex in a
Roman imperial context, for the codex facilitated the use of reference works, law
codes, and biblical texts. The codex would have assisted one of the most significant
innovations of the Roman Empire: the diffusion of literacy from the Mediterra-
nean basin to Europe and North Africa, which allowed technical knowledge to
be transferred in written form in the wake of military engineers and provincial
administrators.
There are less dramatic examples of technologies where a stepwise change
occurred. In everyday commerce, the adjustable steelyard was an improvement on
conventional balances because a single small counterweight could be moved to
balance much larger objects. It could have been invented anywhere that an under-
standing of levers existed, and the principal was explained in the anonymous early
mechanical treatise attributed to Aristotle (Pr. 20.853b-854a, 1.849b--85oa; Hum-
phrey et al. 1998: 50-51). It became common in Roman times, and many have been
found at Pompeii, in its suggested region of invention (di Pasquale 1999). In terms
of instrumentum domesticum, individual property could be protected more effec-
tively when doors could be secured by locks with unique keys; the modern form of
lock with a rotating key, however, is not found before the Roman period. This
mechanism was sufficiently different from the Greek slotted key, which lifted a peg
securing a bolt, to be classed as an invention-as can the idea of a padlock, inde-
pendent of a door, which was in common use in Pompeii by A.D. 79 (Manning
1986: 90-97). The water-dock (clepsydra), which had been in use for many cen-
turies, was refined (reportedly by Ctesibius) to ensure an even flow, and the hy-
draulic technology involved allowed entertaining "side effects" such as moving
figures to be added (chapters 29 and 31). Likewise, mathematical understanding of
angles allowed the sundial to be made more accurate from the third century B.C.
Numerous examples were in everyday use in Pompeii (Gibbs 1976), and it could be
made in a portable form (Field 1990) or scaled up to the Augustan Solarium in
Rome ("die gro~te Uhr aller Zeiten"; Buchner 1982: 7).
It is a commonplace that farming was the most significant industry of the
ancient world, and many technological developments are closely related to the
processing of its products (chapter 14). Research has shown the innovative char-
acter of animal husbandry as well (chapter 8). Weaving was particularly extensive,
and Wild has observed several developments in loom technology within the Roman
imperial period that appear to be associated with the production of increasingly
complex patterns (chapter 18). In particular, there was a change from a vertical to a
horizontal loom (Wild 1987). Processing crops involved mills and presses for ex-
tracting juice from grapes or extracting oil from olives. The fitting of a counter-
weight to the lever of a press by means of a screw (rather than pulling it down with
ropes) may be connected with Heron of Alexandria's description of a thread-
cutting device (Humphrey et al. 1998: 55). The invention of direct screw presses,
which did not require a lever, might have resulted from observing screwed coun-
terweights, or could have been scaled up from traction instruments used by doctors
setting broken bones. The water-powered mill used for grinding corn certainly was
an invention, whether driven by a horizontal wheel that turns a millstone either
directly, or indirectly by a vertical wheel whose drive is transmitted to the stone
through 90 degrees by gears that allow the stone to rotate at a different speed to the
waterwheel (chapter 6). All of the components described by Vitruvius in the late
first century A.D. existed previously: rotary millstones, 90-degree angle gears, and
large waterwheels (Lewis 1997). The water-mill could not have been created until all
of the individual elements had been developed; this process conforms to Usher's
(1955) "cumulative synthesis" model of invention involving perception of a
problem (substituting water power for human or animal labor), insight (awareness
of a range of water wheels and gears), and critical revision (getting the materials,
components, and operating speeds right). Compare also Schiffer's definition of
"invention cascades" associated with "complex technological systems" (chapter
33). Specialist scientists/engineers associated with centers such as Alexandria would
be well placed to make inventions of this kind, since they knew about the needs of
their sponsors, understood mechanical principles, and explored the working of
hydraulic machinery in devices such as water-clocks with displays and side effects-
including wheels with gears, turned by water, as described by Philon of Byzantium
(Lewis 2000a: 354). The expansion of the Roman Empire, and its enormous con-
sumption of wine, oil, and grain, provided a context for the wide diffusion of
pressing and milling technologies (Amouretti and Brun 1993; Mattingly 1996).
This discussion has demonstrated that each technological item may have gone
through a different trajectory of discovery, invention, innovation, diffusion, and
use. This process emphasizes the need for considerable care with vocabulary, and it
complicates any attempt to examine processes of invention and innovation, for we
only ever see the results in technical treatises or archaeological discoveries. There is a
certain consistency about the best known inventors. Archimedes was primarily
interested in mathematics, but many areas of his research had practical implica-
tions. Understanding of levers, weights, and hydrostatics, combined with the
practical skills and concepts necessary for building astronomical spheres, provided
ample knowledge for the construction of siege equipment or water-lifting devices.
Hellenistic monarchs had every reason to support abstract philosophical and sci-
entific pursuits, since intellectual knowledge could be turned to useful purposes:
government, taxation, courtly entertainment, temple rituals, warfare, architecture,
bulk transport, water supply, and irrigation. Whatever misgivings about banausic
activities may have existed, there was clearly a need for knowledge about inventions
to be transmitted in written form, and for new compilations to be assembled as
political and economic circumstances changed. Descriptions and illustrations of
the working of machines commonly dismissed as toys allowed basic principles to be
grasped by engineers who lacked the scientific background of Archimedes or Heron
(Lewis 2000a). Thus, while neither Vitruvius nor Heron can be classed as "in-
ventors," they demonstrate the existence of an audience for treatises on technical
subjects, which persisted into the late imperial and Byzantine period (Pappus,
Pseudo-Heron). Without systems for creating and sharing such information, in-
ventions could not have become innovations to be diffused into technology-in-use.
Vitruvius, Ptolemy, Galen, and Heron all consolidated centuries of scientific
knowledge into systematic treatises at around the time that Pliny the Elder at-
tempted to gather together as much knowledge as possible into his encyclopedic
Natural History. This process did not end in the Roman period, of course; the
survival of much of what we possess results from similar demands for systematic
knowledge in the contexts of Byzantium, Persia, Islam, and Renaissance Italy, where
rulers were faced with the same problems of government, defense, transport, food
supply, ritual, or entertainment. There are no grounds in the technical writers
whose work we know for making essentialist generalizations that contrast a Greek
scientific spirit with a Roman talent for practical application.
Nothing survives of subliterary texts about such craft skills as making pumps
out of blocks of wood rather than bronze, or the practical advantages of glass
windows, of wooden barrels, or of codices. The only developments in the provinces
to reach Pliny's ears were related to agriculture; the vallus and the wheeled plough
presumably impressed him as an estate owner. In contrast, he had little to say about
concrete vaulting or the expansion of glass production; only a garbled tale about
"unbreakable" glass suggests awareness ofrecent innovations (Greene 2000: 46-47).
Archaeological finds provide a physical image of a Greek and Roman world in
which technology was disseminated and implemented according to the scale of
political units, with an apogee in the Roman Empire (Greene 1992, 1994). Multiple
estate ownership by rich families would have encouraged such processes, and the
same social class provided army officers and government officials who gained
detailed knowledge of a number of provinces during their careers. Regional
"schools" of mosaic designers, or standardized services of tablewares made in
multiple production centers, reflect the dissemination of knowledge between craft
workers, whether through molds, pattern books or wooden templates. Many Ro-
man industries incorporated name-stamps into their products (e.g., tiles, pottery,
bronze and glass vessels) which demonstrate that branch workshops could be
opened and workers transferred between them (Harris 1993). The most interesting
questions about Greek and Roman invention and innovation are not about priority
or originality, but about the contexts in which such processes took place.
REFERENCES
Amouretti, M.-C., and J.-P. Brun (eds.) 1993. La production du vin et de l'huile en
Mediterranee. Bulletin de Correspondance Hellenique Suppl. 26. Athens: Ecole Franc;aise.
Baatz, D. 1994. Bauten und Katapulte des romischen Heeres. Stuttgart: Franz Steiner.
Bats, M. (ed.) 1996. Les ceramiques communes de Campanie et de Narbonnaise (le av. J-C-Ile
ap. J-C): La vaisselle de cuisine et de table. Naples: Centre Jean Berard.
Bishop, M. C., and J. Coulston 1993. Roman military equipment from the Punic wars to the
fall of Rome. London: Batsford.
Buchner, E. 1982. Die Sonnenuhr des Augustus: Nachdruck aus RM 1976 und 1980 und
Nachwort uber die Ausgrabung 1980/81. Mainz: von Zabern.
Carradice, I., and M. J. Price 1988. Coinage in the Greek world. London: Seaby.
Craddock, P. T. (ed.) 1990. 2000 years of zinc and brass. British Museum Occasional Paper
50. London: British Museum
context of invention in the ancient world," Technology and Culture: 1-26.
Delaine, J. 2002. "The Temple of Hadrian at Cyzicus and Roman attitudes to exceptional
construction," Papers of the British School at Rome 70: 205-30.
di Pasquale, G. 1999. "Weighing instruments," in A. Ciarallo and E. de Carolis (eds.),
Pompeii: Life in a Roman town. Milan: Electa, 283-85.
Drachmann, A.G. 1948. Ktesibios, Philon and Heron: A study in andent pneumatics. Co-
penhagen: Munksgaard.
Drachmann, A.G. 1963. The mechanical technology of Greek and Roman antiquity: A study of
the literary sources. Copenhagen: Munksgaard.
Dugan, S., and D. Dugan 2000. The day the world took off: The roots of the Industrial
Revolution. London: Channel 4 Books.
Edgerton, D. 1999. "From innovation to use: Ten eclectic theses on the historiography of
technology," History and Technology 16.2: m-36.
Field, J. V. 1990. "Some Roman and Byzantine portable sundials and the London sundial-
calendar," History of Technology 12: 103-35.
Finley, M. I. 1973. The ancient economy. London: Chatto and Windus.

Fleming, S. J. 1999. Roman glass: Reflections on cultural change. Philadelphia: University of
Pennsylvania Museum.
Fleury, P. 1993. La mecanique de Vitruve. Caen: Centre d'Etudes et de Recherche sur
l'Antiquite, Universite de Caen.
Gibbs, S. L. 1976. Greek and Roman sundials. New Haven: Yale University Press.
Greene, K. 1986. The archaeology of the Roman economy. Berkeley: University of California
Press.
Greene, K. 1992. "How was technology transferred in the Roman empire?" in M. Wood and
F. Queiroga (eds.), Current research on the Romanization of the western provinces.
British Archaeological Reports, Intl. Series S575. Oxford: BAR, 101-5.
Greene, K. 1994. "Technology and innovation in context: the Roman background to
mediaeval and later developments," Journal of Roman Archaeology 7: 22-33.
Greene, K. 2000. "Technological innovation and economic progress in tlie ancient world:
Empire in the light of instrumentum domesticum. Journal of Roman Archaeology Suppl.
Hassall, M. (ed.) 1979. Aspects of the De rebus bellicis: Papers presented to E. A. Thompson.
British Archaeological Reports, Intl. Series S63. Oxford: BAR.
Hill, D. R. 1979. The book of ingenious devices. Dordrecht: Reidel.
Howe, T., and I. D. Rowland (eds.) 1999. Vitruvius: Ten books on architecture. A new English
translation with commentary and illustrations. Cambridge: Cambridge University
Press.
Humphrey, J. W., J.P. Oleson, and A. N. Sherwood (eds.) 1998. Greek and Roman tech-
nology: A sourcebook. London: Routledge.
Israeli, Y 1991. "The invention of blowing," in M. Newby and K. Painter (eds.), Roman
glass: Two centuries of art and invention. London: Society of Antiquaries: 46-55.
Keyser, P. 1992. "A new look at Heron's 'steam engine.'" Archive for the History of Exact
Sciences 44: 107-24.
Knorr, W. R. 1996. "Mechanics," OCD, 943-44.
Landels, J. G. 1998. Music in ancient Greece and Rome. London: Routledge.
versity Press.
Lewis, M. J. T. 2000a. "Theoretical hydraulics, automata, and water clocks," in 0. Wikander
(ed.), Handbook of ancient water technology. Leiden: Brill: 343-69.
Lewis, M. J. T. 2000b. "The Hellenistic period," in 0. Wikander (ed.), Handbook of ancient
water technology. Leiden: Brill: 631-48.
Manning, W. H. 1986. Catalogue of the Romano-British iron tools, fittings and weapons in the
British Museum. London: British Museum.
Marsden, E.W. 1969. Greek and Roman artillery: Historical development. Oxford: Clar-
endon.
Mattingly, D. J. 1996. "Olive presses in Roman Africa: Technical evolution or stagnation?"
in M. Khanoussi, P. Ruggeri, and C. Vismara (eds.), Africa romana: Atti dell'XI con-
vegno di studio Cartagine, 1994. Sassari: Universita degli Studi, 577-95.
Mattusch, C. C. 1996. Classical bronzes: The art and craft of Greek and Roman statuary.
Itliaca, NY: Cornell University Press.
May, C. 2002. "Antecedents to intellectual property: The European pre-history of the
'ownership' of knowledge," History of Technology 24: 1-20.
Merker, G. S. 2003. "Corinthian terracotta figurines: The development of an industry," in

C. K. Williams and N. Bookidis (eds.), Corinth: The centenary, 1896-1996. Princeton:
American School of Classical Studies at Athens, 233-45.
Oleson, J.P. 2004. "Well-pumps for dummies: Was there a Roman tradition of popular sub-
literary engineering manuals?" in F. Minonzio (ed.), Problemi di macchinismo in
ambito romano. Archaeologia dell'Italia Settentrionale 8. Como: Comune di Como,
65-86.
Oleson, J.P., and G. Branton 1992. "The technology of King Herod's harbour," in
R. L. Vann (ed.), Caesarea papers: Straton's Tower, Herod's Harbour, and Roman and
Byzantine Caesarea. Journal of Roman Archaeology Suppl. 5. Ann Arbor, MI: JRA,
49-67.
Oleson, J.P., C. Brandon, S. M. Cramer, R. Cucitore, E. Gotti, R. L. Hohlfelder 2005. "The
ROMACONS Project: A contribution to the historical and engineering analysis of
hydraulic concrete in Roman maritime structures," International Journal of Nautical
Oleson, J.P., L. Bottalico, C. Brandon, R. Cucitore, E. Gotti, R. L. Hohlfelder 2006.
"Reproducing a Roman maritime structure with Vitruvian pozzolanic concrete,"
London: Longman.
Pomey, P., and A. Tchernia 2005. "Archimede e la Syrakosia," in E. lo Sardo (ed.), Eurika!
Il genio degli antichi. Naples: Electra, 228-32.
Roberts, C.H., and T. C. Skeat 1983. The birth of the codex. London: British Academy/
Schiirmann, A. 1991. Griechische Mechanik und antike Gesellschaft: Studien zur staatlichen
Forderung einer technischen Wissenschaft. Stuttgart: Franz Steiner.
Thompson, E. A. 1952. A Roman inventor and reformer: Being a new text of the treatise De
rebus bellicis. Oxford: Clarendon Press.
Tomlinson, R., and J. Vallance 1996. "Vitruvius (Pol(l)io)," OCD, 1609-10.
Toomer, G. J. 1996. "Archimedes," OCD, 146-47.
Traina, G. 1994. La tecnica in Grecia ea Roma. Roma: Laterza.
Usher, A. P. 1929. A history of mechanical inventions. Cambridge, MA: Harvard University
Press.
Usher, A. P. 1955. "Technical change and capital formation," in Universities National
Bureau Committee for Economic Growth, Capital formation and economic growth.
Princeton: Princeton University Press, 423-550.
Weinberg, G.D. 1988. Excavations at Jalame: Site of a glass factory in late Roman Palestine.
Columbia: University of Missouri Press.
Whitehouse, D. 1997. Roman glass in the Corning Museum of Glass, I. Corning: Corning
Museum of Glass.
Whitehouse, D. 2001. Roman glass in the Corning Museum of Glass, II. Corning: Corning
Museum of Glass.
Wild, J.-P. 1987. "The Roman horizontal loom," American Journal of Archaeology 91:
459-72.
Wilkinson, P. 2000. What the Romans did for us. London: Boxtree.
Wilson, A. 2002. "Machines, power and the ancient economy," Journal of Roman Studies
92: 1-32.
Zimmer, G. 1982. Antike Werkstattbilder. Berlin: Gebr. Mann.
PART VIII
ANCIENT
TECHNOLOGIES
IN THE MODERN
WORLD
CHAPTER 33
EXPANDING
ETHNOARCHAEOLOGY:
HISTORICAL EVIDENCE
AND MODEL-BUILDING
IN THE STUDY OF
TECHNOLOGICAL
CHANGE
MICHAEL B. SCHIFFER
EXPANDING ETHNOARCHAEOLOGY
Although ethnographic information has long been incorporated into inferences

about past human behavior, not until the last third of the twentieth century did
archaeologists undertake their own sustained and systematic observations in on-
going societies (e.g., Binford 1978; Donnan and Clewlow 1974; Gould 1978, 1980;
Longacre and Skibo 1994; Skibo 1992; Yellen 1977). This research activity is some-
times called "living archaeology," although "ethnoarchaeology" now seems to be the
preferred designation. This chapter suggests that an expanded ethnoarchaeology
that exploits evidence from the historical records of both ancient and modem
societies can become an important research strategy to obtain, refine, and evaluate
822 ANCIENT TECHNOLOGIES IN THE MODERN WORLD
general models and heuristics for investigating technological change. This proposal is
grounded in the tenets of behavioral archaeology, a research program that privileges
the study of people-artifact relationships in all times and all places (Reid et al. 1975;
LaMotta and Schiffer 2001; Schiffer 1992, 1995a; Skibo et al. 1995; Zedefio 2000).
Ethnoarchaeology is usually regarded as the study, by archaeologists, of "tradi-
tional," non-Western societies, such as Australian Aborigines or Mayan Indians (for a
synthesis of ethnoarchaeology, see David and Kramer 2001). Yet since the late 1970s
ethnoarchaeological investigations have also taken place in W estem industrial so-
cieties (e.g., Buchli and Lucas 2001; Gould and Schiffer 1981; Graves-Brown 2000;
Rathje 1979; Rathje and Murphy 1992; Schiffer 1976, 1978). In the latter studies,
ethnoarchaeology has come to resemble, in the artifacts examined, both historical
archaeology and modem material-culture studies. Indeed, societies today are so
interconnected by global commerce and travel that it is almost impossible to dis-
tinguish between ethnoarchaeology, historical archaeology, and modem material-
culture studies on the basis of where and when the objects under investigation were
made.
I suggest that ethnoarchaeology is set apart from other disciplines that study
similar artifacts because its practitioners often ask general questions about the ma-
teriality of human behavior, including phenomena such as refuse-disposal practices,
social boundary maintenance, and ritual technologies. Indeed, ethnoarchaeological
research frequently leads to the provision of general principles-models, law-
like statements, and theories-useful in constructing inferences and for explaining
diversity and change in human behavior.
Ethnoarchaeologists are in an advantageous position to formulate general
principles because researchers study the artifacts themselves and also draw on strong
evidence describing the participation of those very same artifacts in activities. The
latter evidence usually comes from the investigator's own observations, as in the
study of a living society-any living society-or from accounts of people who once
took part in activities no longer being practiced (i.e., oral history). If ethnoarch-
aeology is the study of general relationships between activities and artifacts when
strong evidence is available on both, then in an expanded ethnoarchaeology re-
searchers can make use of evidence from the historical record as well. After all, his-
torical accounts may contain useful descriptions of both artifacts and the activities
in which they took part.
Thus, I propose that ethnoarchaeology be regarded as a wide-ranging research
strategy that includes the establishment of general principles through the study of
strong evidence on both activities and artifacts. It matters not whether the evidence
derives from the investigator's own fieldwork in a living society or from digging in
the documents of dead ones. Thus, classical archaeologists, who rely on historical
and archaeological materials, can-among other pursuits-conduct ethnoarchae-
ological research that aims to furnish and evaluate general principles. In the re-
mainder of this chapter, I illustrate the vision of an expanded ethnoarchaeology by
presenting models and heuristics derived from research on electrical technologies
of the eighteenth and nineteenth centuries.
EXPANDING ETHNOARCHAEOLOGY 823
EARLY ELECTRICAL TECHNOLOGIES

AS A MODEL FOR EXPANDED
ETHNOARCHAEOLOGY
Source material for studying early electrical technologies is copious. Accounts of

the constituent artifacts, along with descriptions of activities such as manufacture
and use, can be found in textbooks, monographs, journal articles, letters, in-
struction manuals, diaries, newspapers, magazines, and so on. Moreover, many of
the artifacts themselves survive in private collections and in museums of technol-
ogy throughout the West. Of course the historical record exhibits frustrating gaps,
but the sheer amount of information available, even for the eighteenth century, is at
times overwhelming. In the course of asking historical questions while seeking to
write narratives about how our modern electrical world came to be, I have some-
times also asked general questions about technological change and exploited the
rich historical record seeking answers. The result has been the provision of models
and heuristics that might be useful in studying technological change in many social
contexts.
BEHAVIORAL FOUNDATIONS
The key framework that informs many behavioral models of technological change is
the life history of artifacts and technologies. Long implicit in studies of chipped-
stone technology (Bleed 2001), the life-history framework was made explicit and
generalized during the 1970s in studies on inference and on the formation processes
of the archaeological record (Schiffer 1972, 1975, 1976; Sullivan 1978). More recently,
it has served in research on diverse topics including subsistence (Gumerman 1997)
and territory formation (Zedefi.o 1997). This framework reminds us that every ar-
tifact, structure, and technological system had a life history that may have included
processes such as procurement of raw materials, manufacture, use, reuse, and de-
position. Life histories can be modeled in countless ways, depending on the in-
vestigator's research interests and the technology under investigation.
The most fine-grained life-history models are termed "behavioral chains"; they
involve the sequence of specific activities in which an artifact participated during its
entire existence. By focusing on the constituent activities of a behavioral chain, the
investigator can isolate specific person-person, person-artifact, and artifact-arti-
fact interactions. Behavioral chains are essential, for example, in building models to
explain the design of artifacts (e.g., Schiffer and Skibo 1997; Skibo and Schiffer
2001).
Researchers can also employ a life-history model that describes three basic
processes in the development of a type of artifact or technology: invention, com-
mercialization (manufacture or replication), and adoption (LaMotta and Schiffer
2001; Schiffer 1995a). Keeping such processes distinct at the research-design stage is
important because, for example, one cannot explain the occurrence of inventive
activities by using models that account for the adoption of new technologies, and
vice versa. Clearly, explanations of technological change require a plethora of
process-specific models and theories (Schiffer 1995a; Schiffer et al. 2001).
Because behavioral theories and models attend assiduously to the materiality of
human life, they often make reference to the behavioral capabilities of artifacts that
facilitate specific interactions in specific activities. These behavioral capabilities
are termed performance characteristics and enable symbolic and utilitarian func-
tions (on performance characteristics, see Schiffer and Miller 1999: 16-20 ). Thus, in
functioning as a cooking pot, a ceramic container has to possess, among other
performance characteristics, adequate resistance to thermal shock. Likewise, a cru-
cifix must have a certain form in order to be recognizable as a symbol of Christian
beliefs.
The crucifix example suggests that performance characteristics also include
behavioral capabilities dependent on human senses: tactile, olfactory, gustatory,
auditory, and visual. Sensory performance characteristics, which contribute im-
portantly to symbolic and aesthetic interactions, include a silk shirt's ability to feel
"silken" when touching someone's skin, a clarinet's capacity to emit clarinetlike
sounds when played, and a halibut's ability to smell like "fresh fish." During the
conduct of most activities, diverse performance characteristics of the constituent
artifacts-sensory and nonsensory-come into play and facilitate the activity's
forward motion. The section below "Ddifferential adoption further expands the
fundamental notion of performance characteristic.
I now tum to three examples of an "expanded ethnoarchaeology": techno-
logical differentiation, differential adoption, and the cascade model of invention
processes. Readers interested in the details of these studies and the written mate-
rials employed as evidence can tum to the original works cited in these discussions.
For a survey of archaeological studies of technology, see Millar 2007.
TECHNOLOGICAL DIFFERENTIATION
New technologies such as glassworking, aqueduct construction, and ceramic pro-

duction often begin their life histories as a small number of forms definable on the
basis of utilitarian and/or symbolic functions. Thus, the first pottery in any area
often consisted of just a few jars with limited food-preparation functions; centuries
later, vessels were manufactured in dozens of shapes and sizes that took part in
diverse act1VIt1es, from marriage ceremonies to brewing alcoholic beverages.

Technologies that endure and proliferate possess an adaptability that permits
people, in a changing social context, to create new forms to serve new utilitarian
and symbolic functions. The increase in the varieties of a technology is a recurrent
pattern discernible in countless archaeological sequences; the process responsible
for this pattern is termed "technological differentiation."
In a study of twentieth-century home electronics, I called attention to tech-
nological differentiation, noting that the process is usually accompanied by func-
tional specialization in the new forms (Schiffer 1992: 107-8). However, because the
aim of that study was to deconstruct the notion of "technological revolution," no
models of technological differentiation were built. Rather, the task of model
building was taken up in a later study of eighteenth-century electrical technology
(Schiffer 2002; Schiffer et al. 2003).
Electrical technology became distinguishable early in the eighteenth century
with the appearance of a small number of esoteric devices-including Leyden jars
(the first capacitors) and frictional generators-that natural philosophers em-
ployed to explore surprising phenomena, such as electricity's ability to produce
light and sparks. From this modest beginning, people over the course of decades
invented hundreds of specialized electrical artifacts that carried out utilitarian and
symbolic functions in activities as varied as healing, giving public lectures, and
experimenting with gases. Dozens of these inventions were replicated by philo-
sophical instrument-makers and adopted by consumers.
To gain a behavioral understanding of technological differentiation, I con-
structed a framework based on the transfer of technologies among "techno-
communities." A techno-community (or, simply, "community") is a group of
people whose members-often drawn from different social, political, and ethnic
groups-carry out particular activities employing particular technologies. Thus,
electrophysicists used certain electrical devices for conducting cutting-edge physics
experiments, whereas electrotherapists in their healing activities employed, in ad-
dition to generators and Leyden jars, some rather different sorts of electrical things.
As a technology is transferred from one community to another, the recipient
community's members invent new functional types whose performance character-
istics are more suitable for participating in their own activities. For illustrative
purposes, the following discussion makes reference to the technology transfer from
electrophysicists to electrotherapists.
It is convenient to model technology transfer as a six-phase process, which
begins with information transfer. In this phase people who are potential nuclei of
new recipient communities learn about the technology in various ways, such as by
word of mouth, through print media, and by examining the hardware itself. Thus,
magazine articles and lectures about electrophysicists' experiments reached a wide
public, including people who envisioned the use of electricity for ameliorating
various afflictions. Obviously, inferring from the archaeological and historical re-
cords the precise modes of information transfer can be difficult, if not impossible.
This is not, however, a fatal limitation because we know that, often, only a fraction
of the individuals who learn about a new technology actually go on to modify or

acquire it. The latter, who make up the recipient community, are the focus of the
model.
The next phase, experimentation, usually begins when a few people try out the
new technology, learning how well it performs in their activities. These early efforts
often reveal that the technology shows some promise but its weighting of perfor-
mance characteristics is unsuitable for the new activity. For example, it was found
that the cumbersome instruments of early physics were not convenient for use in
electromedicine.
In the redesign phase, people invent specialized types of the technology, which
embody a more appropriate weighting of performance characteristics. Thus, elec-
trotherapists designed compact and portable generators along with specialized
conductors that could be used to apply electricity to afflicted areas of the human
body, including feet, teeth, and eyes.
Replication involves manufacture of the redesigned technology and its distri-
bution to purchasers and users. Thus, instrument makers manufactured and sold
many devices that had been redesigned specifically for electromedical activities.
During acquisition (also called consumption or adoption), people obtain exam-
ples of the new technology and thereby become members of the recipient techno-
community. Apothecaries, surgeons, and physicians, among many others, com-
prised the community of electrotherapists. During the final phase, use, the acquired
technology's use-related performance characteristics came into play as it interacted
with healers, assistants, and patients.
The technology-transfer framework was formulated in tandem with my efforts
to explain, in historical terms, the differentiation of eighteenth-century electrical
technology. The rich historical record made it possible to fashion the six-phase
model, whose application, in turn, enabled the crafting of a narrative that accounted
for the differentiation of that very same technology in the hands of more than half
a dozen techno-communities. Whether the technology-transfer model can be ap-
plied with profit to other cases of technological differentiation, such as the pro-
liferation of worked-bone tools and ornaments during the Upper Paleolithic, glass
in Rome, and lasers in the twentieth century, remains to be learned through future
research.
At the very least, this framework invites the archaeologist to ask a great many
questions about a differentiated technology. Research to answer these questions can
lead to diverse inferences about, for example, the activities in which the technol-
ogy's new varieties participated, contextual factors (social and environmental) that
contributed to changes in activities, the functions of the technology in those ac-
tivities, the performance characteristics of the new varieties that promoted their
adoption over alternative technologies that might have served the same functions,
and the social composition of the associated techno-communities. By synthesizing
these inferences, the archaeologist can fashion a narrative to explain technological
differentiation, one that contextualizes the new technologies in relation to changing
social and environmental factors.
DIFFERENTIAL ADOPTION
A common problem in the study of technological change is that of explaining

patterns of adoption of a new technology that becomes available in the marketplace.
Usually, new technologies are not adopted uniformly; that is, some but not all
individuals, social units, or communities acquire and use the new technology. Thus,
an understanding of adoption processes requires one to frame the problem as
differential adoption, and invites us to seek explanations for both adoption and
nonadoption decisions among potential adopters (i.e., those familiar with the
technology). In the course of investigating the differential adoption of electrically-
lit lighthouses during the late nineteenth century (Schiffer 2005a), I was able to
refine the performance matrix, a heuristic tool that had been employed in previous
studies of adoption processes (e.g., Schiffer and Skibo 1987; Schiffer 1995b, 2000).
The performance matrix is particularly applicable to cases in which a new tech-
nology is competing with existing technologies for application in ongoing activities.
If a new technology lacks competition, the study of adoption requires other models
and heuristics, such as Hayden's (1998) "aggrandizer" model.
A behavioral analysis of adoption is founded on the premise that potential
adopters consider the known and anticipated performance characteristics of the
competing technologies. In a deliberative process that may not be explicit, potential
adopters assign varying weights to specific performance characteristics in response
to contextual factors-political, social, religious, demographic, and economic, for
example. These weightings determine whether the technology is adopted. Because
these diverse contextual factors differ among potential adopters, their weightings
of performance characteristics also differ; the result is differential adoption. As I
will show, patterns in the performance matrix enable the archaeologist to ascer-
tain which performance characteristics were weighted in given adoption decisions.
The researcher can then offer hypotheses about the specific contextual factors
that affected the adoption decisions of individuals or groups. An investigation of
electrically-lit lighthouses of the late nineteenth century will illustrate this research
process.
Generator-powered electric lights, employing carbon arcs, were first installed
in a lighthouse in 1859, in England. During the following decades, when many
hundreds of new lighthouses were being built worldwide, the electric light was
adopted only infrequently. Indeed, at the end of the century, the use of electric-arc
illumination in lighthouses peaked at around 30. The adoption decisions are in-
triguing: France and England together installed nearly 20 electric lights, most other
nations acquired none, and a few-including the United States-adopted just one
or two. Electric arc lights had to compete against lamps that burned some kind of
hydrocarbon (e.g., lard oil, rapeseed oil, or kerosene).
Adoption decisions were made by lighthouse boards, which were governmental
or quasi-governmental organizations. They built new lighthouses, maintained old
ones, and in a few nations, especially England, France, and the United States,
supported experimental programs that assessed new illuminants and other light-
house technologies. These experiments were often documented in published re-
ports, which furnished information on the performance characteristics of oil and
electric illuminants to potential adopters-and to this researcher. Moreover, con-
textual information was available from abundant primary and secondary sources.
Thus, this case provided an ideal opportunity to refine the performance matrix as a
tool for exploring differentiation adoption.
The performance matrix is a table for comparing, side by side, the performance
characteristics of two or more competing technologies in relevant life-history ac-
tivities. For each performance characteristic, the investigator indicates with plus
and minus signs which technology did (+) or did not ( - ) achieve minimally
adequate performance. Performance matrices can also be built with numerical
values or presence/absence notations. To reveal any patterns in the performance
matrix, the investigator shuffles the rows to create clusters of plusses and minuses
within an activity category.
Because decisions of the boards affected most life-history activities of light-
house illuminants, their members presumably took into account a wide range of
performance characteristics. Thus, in table 33.1, performance characteristics are
aggregated by major life-history activities: (1) acquisition and installation of com-
ponents, (2) functions during use, and (3) operation, regular maintenance, and
repair. In constructing this matrix it was necessary to employ an expansive con-
ception of performance characteristics so as to implicate the entire range of con-
textual factors that might have influenced adoption decisions. Indeed, the entries in
table 33.1 run the gamut of behavioral capabilities, such as costs of acquiring
components, ease of administration, and the ability to convey particular meanings.
This broad conception of performance characteristics, along with plus-and-minus
entries, allows the archaeologist to incorporate into the analysis seemingly incom-
mensurable factors, qualitative and quantitative. As a result, one can deal explicitly
with the diverse contextual factors that likely influenced the weightings of per-
formance characteristics and affected adoption decisions.
As long as the researcher includes all potentially relevant performance char-
acteristics (from the standpoint of the individual or social unit making the adop-
tion decision), the performance matrix serves as a causally neutral tool of analysis.
Naturally, the reconstruction of performance characteristics in the context of the
classical cultures will rely more on archaeological and ethnographic evidence than
on the rich contemporary documentation available for nineteenth-century light-
houses. In principle, patterns appearing as clusters of plusses and minuses should
indicate which performance characteristics were heavily weighted by past decision-
makers and which were not.
The performance matrix of lighthouse illuminants exhibits several patterns.
The dominant pattern is that the electric light competed poorly in acquisition,
installation, operation, maintenance, and repair activities. This pattern suggests
that the vast majority of maritime nations, those that adopted no electric lights, had
assigned heavy weight to utilitarian and financial factors. By not adopting electric
Table 33.1. A performance matrix for lighthouse illumination, ca. 1860-99

(adapted from Schiffer 2005a, Table 1).
Acquisition of the Components, and
Installation of the System Electric Oil
Ability to acquire system components + +
commercially
System can be installed in lighthouses in +
any location
System can be easily installed in existing +
lighthouse structures
Affordability of a system's "first costs" +
Existing expertise adequate for designing +
and installing the system
Functions During Use Electric Oil
Yields the whitest, brightest, most +

penetrating light
Can produce sufficiently steady light + +
Long outages are avoidable + +
Does not cast confusing shadows +
Can avoid blinding mariners +
Ability to symbolize special concern for +
the safety of ships and sailors
Can symbolize a nation's wealth and +
political power
Can symbolize modernity +
Able to symbolize scientific/technological +
prowess
Operation, Regular Maintenance, and

Repairs Electric Oil
Operable with traditional staff of keepers +

Operable without complete backup +
systems
Ease of repairing breakdowns +
Affordability of operating expenses +
Ease of administration +
lights, these nations avoided readily predicted expenses and hassles. On the other
hand, a lesser pattern in use-related functions indicates that the electric light was an
adequate illuminant, especially in haze and light fog, and it especially excelled in
symbolic capabilities because its light was distinctive in color and brightness. This
lesser pattern helps one to formulate explanations as to why two nations, England
and France, adopted a handful of electric lights and why several others acquired one
or two. It would appear that lighthouse boards in adopting nations had assigned,
for political purposes, a heavy weight to symbolic capabilities, which outweighed
the electric light's many utilitarian and financial performance deficiencies. To make
a long story short, I concluded that adoptions were promoted by technological
rivalries between long-time adversaries England and France as well as by the in-
terests of a few other nations in advertising to all mariners a commitment to
modernity through their mastery of a cutting-edge electrical technology.
The lighthouse case suggests that an archaeologist, using patterns in a per-
formance matrix as a foundation, can construct a contextualized narrative that
accounts, with reference to local factors, for instances of adoption and nonadoption
of a new technology. Although the performance matrix had been employed pre-
viously in studies of technological change, even in a prehistoric case (Schiffer and
Skibo 1987), the lighthouse project led to its significant refinement as a tool of
analysis. Indeed, the lighthouse study sharply defined the role of the performance
matrix as a causally neutral tool suitable for investigating differential adoption.
INVENTION CASCADES
Many technologies, like lighthouse illuminants, canal-aqueduct systems, and do-

mestic cooking utensils, consist of a set of artifacts that must interact appropriately
with people and themselves, especially during activities of manufacture and use.
How these relatively harmonious relationships come to be is the subject of the
"cascade model" of invention processes (Schiffer 2005b). The cascade model assists
in understanding the spurts of inventive activity that accompany development of a
"complex technological system" (CTS); it does not, however, explain why the de-
velopment of a CTS is initiated or pursued. The archaeologist enjoys wide latitude
in defining a CTS, for it need consist only of a set of interacting artifacts. As such,
CTSs can be expected to develop in virtually every human society, including tech-
nologies that range from the bow and arrow to the nuclear-powered submarine.
The cascade model was inspired by historian Thomas Hughes' (1983) model of
"reverse salients." According to Hughes, during the development of a complex
socio-technical system, such as an electric power network, certain components lag
and present critical problems-generators of insufficient capacity to meet demand
and power poles vulnerable to lightning strikes, for example. If the system is to
become functional, then such problems must be solved-at first through invention.
This key insight was the starting point for building the cascade model, which is a
behavioral elaboration and generalization of Hughes' model.
The cascade model took shape as I confronted the bewildering array of inven-
tions needed to create a functional electromagnetic telegraph system during the mid-
nineteenth century (Schiffer 2005b). Especially intriguing were the numerous and
diverse inventions offered to solve particular performance problems, many of which
failed to be widely replicated and adopted. In a nutshell, the cascade model posits
that, during a CTS's development, emergent performance problems-recognized by
people as shortcomings in that technology's constituent interactions-stimulate
sequential spurts of inventive activity. As adopted inventions solve one problem,
people encounter new and often unanticipated performance problems, which stim-
ulate more inventive spurts, and so on. The result is a series of "invention cascades."
The distinctive feature of this model, which promotes its generality, is the premise
that processes in a CTS's life history are the immediate contexts in which perfor-
mance problems emerge and provoke invention cascades. Life-history processes,
then, are the appropriate analytic units for investigating invention processes in CTSs.
Any CTS has a life history consisting of a minimal set of processes: creating a
prototype, replication or manufacture, use, and maintenance. These processes do
not comprise a unilinear sequence, for some may occur in parallel and others can
recur. And, depending on the specific CTS, it may be necessary to delineate many
more processes. Thus, to accommodate the telegraph's many invention cascades, I
specified as analytic units additional processes that may apply only to CTSs in
capitalist-industrial societies, including technological display, demonstrating
practicality to potential financiers, and marketing and sales. Although life-history
processes can be elaborated ad infinitum, the key premise remains invariant: life
history processes, however enumerated, are the proximate contexts of invention
cascades.
If the development of a CTS is to proceed, people must judge that its com-
ponent artifacts have reached acceptable values of important performance char-
acteristics. The failure of a component to perform adequately its major symbolic
and/or utilitarian functions usually calls forth a spurt of inventive activity. Al-
though most of the resultant inventions do not solve the performance problem, one
or a few might, and these may be replicated and adopted. As development con-
tinues, other performance shortcomings are often encountered, which stimulate
additional inventive spurts. The result is an invention cascade. The development of
the electromagnetic telegraph, for example, involved cascades of inventions for
telegraph senders, receiver-printers, poles, insulators, lightning protectors, un-
derwater cables, and so forth.
The cascade model invites archaeologists to rewrite the story of any CTS by
seeking evidence that its development proceeded in fits and starts, marked by
invention cascades and the likely proliferation of unsuccessful components. In this
way, the archaeologist may be able to account for previously ignored variability,
such as unusual or unique artifacts that do not fit into established classifications.
The search for traces of invention cascades can be guided by an appreciation for the
performance requirements of the developing CTS, which dictates that the archae-
ologist understand how the system would have worked. To acquire such knowl-
edge, one can exploit modem engineering literature and expertise, conduct exper-
iments, and draw on ethnographic, archaeological, and historical information.
Using life-history processes as analytical units, the archaeologist can then seek
evidence for the successful and unsuccessful components that arose during the
CTS's development.
As Arnold (2007) demonstrates, the cascade model is also a potent antidote to
diffusionist explanations of technological change. In fact, acquiring the "idea" for a
specific CTS is usually the easiest part of the process. As one early experimenter
pointed out, the idea of the electromagnetic telegraph was "obvious" (Barlow 1825:
105). Indeed, many people familiar with electrical technology in many nations tried
to construct electromagnetic telegraphs. They soon learned that it was far from
obvious how to design an entire suite of components that had to function together
effectively in a telegraph system. Thus, each development team-for example, Morse
and Vail in the United States, or Wheatstone and Cook in England-had to con-
trive specific components, usually by trial and error, with failures far outnumbering
successes. In view of this process, in which the inventors' expertise and skill grew
along with the proliferation of components, it would be facile to attribute multiple
developments of the telegraph in several western nations to the diffusion of the
telegraph idea. At best, the idea of a particular CTS is a starting point for a some-
times lengthy and costly development process that often makes use oflocal human
and material resources (Arnold 2007).
Lacking familiarity with classical archaeology, I cannot presume to prescribe how a

researcher might conduct ethnoarchaeological studies that exploit historical (and
archaeological) evidence, such as those I carried out on electrical technologies of the
eighteenth and nineteenth centuries. However, I would be dumbfounded if the
surviving evidence of technologies in classical societies did not offer archaeologists
a significant opportunity to generate and refine models of technological change. In
the classical world, complex technological systems were developed, some tech-
nologies became highly differentiated, and technologies were no doubt differentially
adopted. At the very least, classical archaeologists can assess whether the general
models, concepts, and heuristics presented here have any utility for understanding
specific cases of technological change.
It might appear that the expanded ethnoarchaeology advocated here is merely
historical research. In a superficial sense that is true: the ethnoarchaeologist, like the
historian, obtains evidence from the written record. By the same sort of argument
one might also claim that traditional ethnoarchaeology is ethnography, since its
practitioners also obtain information through studies of living people. I suggest,
however, that the narrow focus on data gathering and sources of evidence is mis-
placed. What distinguishes traditional ethnoarchaeology from ethnography is the
distinctive set of general questions that archaeologists bring to the field: a con-
cern with people-artifact relationships of every kind, from processes of stone-tool
manufacture to disposal practices for ritual technologies (e.g., Walker 1995). Like-
wise, what distinguishes an expanded ethnoarchaeology from history is the asking
of general questions and an unapologetic concern with artifacts and processes of
technological change. Working with historical materials as a historian, the re-

searcher fashions explanatory narratives; working as an ethnoarchaeologist, that
same researcher can build and evaluate models, theories, and heuristics of poten-
tially widespread applicability.
REFERENCES
Arnold, J. E. 2007. "Credit where credit is due: The Chumash ocean-going plank canoe,"
American Antiquity 72: 196-209. Barlow, P. 1825. "On the laws of electro-magnetic
action, as depending on the length and dimensions of the conducting wire, and on the
question, whether electrical phenomena are due to the transmission of a simple or
compound fluid," The Edinburgh Philosophical Journal 12: 105-14.
Binford, L. R. 1978. Nunamiut ethnoarchaeology. New York: Academic Press.
Bleed, P. 2001. "Trees or chains, links or branches: Conceptual alternatives for consider-
ation of stone tool production and other sequential activities," Journal of Archae-
ological Method and Theory 8: 101-27.
Buchli, V., and G. Lucas (eds.) 2001. Archaeologies of the contemporary past. London:
Routledge.
David, N., and C. Kramer 2001. Ethnoarchaeology in action. Cambridge: Cambridge Uni-
versity Press.
Donnan, C. B., and C. W. Clewlow, Jr. (eds.) 1974. Ethnoarchaeology. Los Angeles: UCLA
Institute of Archaeology.
Gould, R. A. (ed.) 1978. Explorations in ethnoarchaeology. Albuquerque: University of New
Mexico Press.
Gould, R. A. 1980. Living archaeology. Cambridge: Cambridge University Press.
Gould, R. A., and M. B. Schiffer (eds.) 1981. Modern material culture studies: The archae-
ology of us. New York: Academic Press.
Graves-Brown, P. (ed.) 2000. Matter, materiality and modern culture. London: Routledge.
Gumerman, G. 1997. "Food and complex societies," Journal of Archaeological Method and
Theory 2: 105-39.
Hayden, B. 1998. "Practical and prestige technologies: The evolution of material systems,"
Journal of Archaeological Method and Theory 5: 1-55.
Hughes, T. P. 1983. Networks of power: Electrification in Western sodeties, 1880-1930. Bal-
timore: Johns Hopkins University Press.
LaMotta, V. M., and M. B. Schiffer 2001. "Behavioral archaeology: Towards a new syn-
thesis," in I. Hodder (ed.), Archaeological theory today. Cambridge: Polity Press,
14-64.
Longacre, W. A., and J.M. Skibo (eds.) 1994. Kalinga ethnoarchaeology: Expanding ar-
chaeological method and theory. Washington, D.C.: Smithsonian Institution Press.
Millar, H. M. -L. 2007. Archaeological approaches to technology. London: Academic Press.
Rathje, W. L. 1979. "Modern material culture studies," Advances in Archaeological Method
and Theory 2: 1-37.
Rathje, W. L., and C. Murphy 1992. Rubbish! The archaeology ofgarbage. New York: Harper
Collins.
Reid, J. J., M. B. Schiffer, and W. L. Rathje 1975. "Behavioral archaeology: Four strategies,"
American Anthropologist 77: 864-69.
Schiffer, M. B. 1972. "Archaeological context and systemic context," American Antiquity 37:
156-65.
Schiffer, M. B. 1975. "Behavioral chain analysis: Activities, organization, and the use of
space," in "Chapters in the prehistory of eastern Arizona, IV," special issue, Fieldiana:
Anthropology 65: 103-19.
Schiffer, M. B. 1976. Behavioral archeology. New York: Academic Press.
Schiffer, M. B. 1978. "Methodological issues in ethnoarchaeology," in R. A. Gould (ed.),
Explorations in ethnoarchaeology. Albuquerque: University of New Mexico Press,
229-47.
Schiffer, M. B. 1992. Technological perspectives on behavioral change. Tucson: University of
Arizona Press.
Schiffer, M. B. (ed.) 1995a. Behavioral archaeology: First prindples. Salt Lake City: University
of Utah Press.
Schiffer, M. B. 1995b. "Social theory and history in behavioral archaeology," in J.M. Skibo,
W. H. Walker, and A. E. Nielsen (eds.), Expanding archaeology. Salt Lake City: Uni-
versity of Utah Press, 22-35.
Schiffer, M. B. 2000. "Indigenous theories, scientific theories and product histories," in
P. Graves-Brown (ed.), Matter, materiality and modern culture. London: Routledge,
72-96.
Schiffer, M. B. 2002. "Studying technological differentiation: The case of 18th-century
electrical technology," American Anthropologist 104: 1148-61.
Schiffer, M. B. 2005a. "The electric lighthouse in the nineteenth century: Aid to navigation
and political technology," Technology and Culture 45: 275-305.
Schiffer, M. B. 2005b. "The devil is in the details: The cascade model of invention pro-
cesses," American Antiquity 70: 485-502.
Schiffer, M. B., K. L. Hollenback, and C. L. Bell 2003. Draw the LIGHTNING Down:
Benjamin Franklin and electrical technology in the Age of Enlightenment. Berkeley:
University of California Press.
Schiffer, M. B., and A. R. Miller 1999. The material life of human beings: Artifacts, behavior,
and communication. London: Routledge.
Schiffer, M. B., and J.M. Skibo 1987. "Theory and experiment in the study of technological
change," Current Anthropology 28: 595-622.
Schiffer, M. B., and J.M. Skibo 1997. "The explanation of artifact variability," American
Antiquity 62: 27-50.
Schiffer, M. B., J.M. Skibo, J. Griffitts, K. Hollenback, and W. A. Longacre 2001. "Beha-
vioral archaeology and the study of technology," American Antiquity 66: 729-38.
Skibo, J.M. 1992. Pottery function: A use-alteration perspective. New York: Plenum.
Skibo, J.M., and M. B. Schiffer 2001. "Understanding artifact variability and change: A
behavioral framework," in M. B. Schiffer (ed.), Anthropological perspectives on tech-
nology. Albuquerque: University of New Mexico Press, 139-49.
Skibo, J.M., W. H. Walker, and A. E. Nielsen (eds.) 1995. Expanding archaeology. Salt Lake
City: University of Utah Press.
Sullivan, A. P. 1978. "The Structure of archaeological inference: A critical examination of
logic and procedure," Advances in Archaeological Method and Theory 1: 183-222.
Walker, W. H. 1995. "Ceremonial trash?" in J.M. Skibo, W. H. Walker, and A. E. Nielsen
(eds.), Expanding archaeology. Salt Lake City: University of U tab Press, 67-79.
Yellen, J. 1977. Archaeological approaches to the present: models for reconstructing the past.
New York: Academic Press.
Zedefi.o, M. N. 1997. "Landscapes, land use, and the history of territory formation: An
example from the Puebloan Southwest," Journal of Archaeological Method and Theory
4: 67-103.
Zedefi.o, M. N. 2000 "On what people make of places: A behavioral cartography," In
Michael B. Schiffer (ed,) Social theory in archaeology. Salt Lake City: University of Utah
Press, 97-111.
INDEX
Note: page numbers followed by f and t indicate figures and tables.
abacus, 780, 781f Alexander the Great, 679, 680, 685, 773, 773f
accensus, 750 Alexandria
Achaemenid vessels, 528 Great Harbor, 650, 652f
acquisition phase, 826 Library of, 725
acrophonic system, 778 Museum at, 338, 787, 805-806
Acropolis, 250, 433 Pharos of, 661
Adam, J.-P., 64 "Alexandrian" glass, 529
adits, 97 alkali in glass, 521
Admiralty Charts of the Mediterranean, 639 Allan, J., 501
adobe walls and plaster coating, 236 alphabetic notation of numbers, 779
adoption, differential, 827-830 alphabets, 717
adzes, 445 altar dedications by guilds, 443
Aemilius Lepidus, M., 558 alum, 485
aeolipiles, 154, 809 Amasis Painter, 52f
Aerarium, 729, 731 Ammianus Marcellinus, 348-350
Aeschylus, 19, 804 Amphictyonic League, 229
Agesistratus, 348 amphitheaters, 361, 455-456
agonistic ideal, 17 Amphitheatrum Castrense, 707
Agora II mill (Athens), 147 amphoras, 259, 260!, 267f, 277-278,
agoranomoi, 761-762 508-509, 763
Agricola, 66, 113 amphoreus, 762-763
Agricola (Tacitus), 26-27 Amyntas III (King of the Macedonians), 250
agriculture anaphoric water clock, 340
overview, 158-159 Anatolia, 95-96
arable cultivation, 165-169 Anaxagoras, 3,724
arboriculture, 162-165 Anaximander, 746
fodder crops, 182 Ancient Society (Morgan), 62-63
livestock farming integrated with, 181-182 Andreas, bench of, 345
meadows, gardens, and moldboard plows, Andronicus from Cyrrha, 753
169-171 anemouria ("windwhirls"), 153
Mediterranean ecosystem adaptation, animal by-products, 384-386, 719
159-161, 16of animal fibers, 466-468, 468f
terracing as passive irrigation, 309 animal husbandry
water management and irrigation, 161-162 overview, 175-177
Agrigento temple of Zeus (Sicily), 241, 242f, agriculture and, 161
244-245 breeding, 177-180
Agrippina, 327 fowling, 188, 192, 193t-203t, 204
air movement in mine workings, 97 game farming, 186-188, 189t-191t
alchemists, 527 health and veterinary care, 183-185
Alciphron, 747 housing, 183, 184!
Alesia, 703 nutrition, 180-183
838 INDEX
animal husbandry (continued) arable cultivation, 165-169

principal domestic species, 178t arachne sundial, 747
relief of Roman livestock, 177f arboriculture, 162-165
reproductive rate, 181 arcades, 299, 301!, 573
transhumance, 161, 183 Archaemenid vessels, 532
water supply and, 167 archaeological evidence. See sources and
animal power evidence
overview, 140 "Archaeology: The Loss of Innocence"
domestication and, 136 (Clarke), 76
human energy vs., 137 archeion (office of the chief magistrates), 727
Pompeian mill, 376, 377f arches
water-lifting wheels and, 162 overview, 257-259
See also transport on land by harnessed in bridges, 570, 572-573, 572f
animals buttressing, 269, 269f
animals, hoisting of, 361 in harbors, 648, 649
Anio Vetus, 297 reinforcement of, 269-273, 271f
Annius, T., 560 relieving, 259-260, 26of
ANSER (Porti antichi del mar Tirreno) Archimedean water screw
project, 641 agriculture and, 162
Anthedon harbor, 646-647 depictions of, 42
Anthemius of Tralles, 39 force pump and, 353
Antigone (Sophocles), 19 mining and, 97
Antike Technik (Diels), 67 reliability of attribution, 805
Antikythera Mechanism, 341, 744-746, 745f, theoretical and practical considerations,
792-793 806
antimony, 527 Archimedes
Antioch, 298, 357 continuous development and, 810
Antipater of Thessalonica, 144, 148 Dimension of the Circle, 796
Apollodorus of Damascus, 348, 456 Equilibrium of Planes, 338, 339
Apollonius (Egyptian finance minister), fire started with mirrors, 138
466,467 as innovator, 806-807, 815
Apollonius of Perge, 338, 340-341, 360, 788 planetarium, 791-792
Apollonius the Athenian, 436 on practical and theoretical advances, 584
Appian Way (Via Appia), 553-558, 554/, 560,563 screw-windlass, 661
Appius Claudius Caecus, 554, 555, 556 sources on, 23
appropriate technology movement, 75-76 trispastos and polyspastos, 344-345
Aqua Appia, 296-297 warfare and, 685
Aqua Claudia, 302 on water clock, 791
Aqua Marcia, 297, 302 water-screw (See Archimedean water screw)
Aqua Tepula, 297 worm drive, 340, 344
Aqua Traiana, 300!, 302 Archimedes, Syracusan, 39
aqueducts Archi of S. Lidano, 569
De aquaeductu urbis Romae (On the architectural construction. See construction
Aqueducts of the City of Rome) and engineering
(Frontinus), 25-26, 64, 780-781 architectural geometry, 782
Greek, 293-296 architecture, business of, 228-229
Near Eastern, 290-293, 292f architraves, 270
qanats (underground tunnels), 162, 291-293, archives. See record-keeping
292f, 322-324 Archytas of Tarentum, 584, 786
Roman, 296-302 arcuated lintels, 270
Saldae aqueduct tunnel, 329-333 ards, 166, 167f
water power and, 143 area measures, 766-768
See also tunnels Ariccia viaduct, 557, 569
INDEX 839
Ariston, 21 metalworking shops, 432-433

Aristophanes, 528, 653, 725, 778 ostracism, 735
Aristotle shipsheds, 658
on animal husbandry, 181 technical texts in, 17-20
on cheese, 385 Tower of the Winds, 749!, 753
on coinage, 775-776 written sources from, 17-20
on construction, 251 Augustine, 30-31, 395, 801
on division of labor, 5 Augustus, 24, 127, 719, 731
on glass, 521 Aurelian, 707
on ichthyology, 205 Aurelius Antonius, 719, 720/
library of, 725 Ausher, A. P., 67
on tools, 3 AutoCAD and AutoLISP, 227
on written records, 727 automata. See gadgets and scientific
armchair voussoirs, 275, 276/ instruments
arms and armor. See warfare and fortification Avenches mill (Switzerland), 142, 144
Arnold, B., 626 axes, 445
Arnold, J.E., 830
Arretine production, 397 Baalbek, Temple of Jupiter at, 258, 259
Ars Poetica (Horace), 780 Babylonian Talmud, 27
Artemidorus, 6, 489 back-sights, 554-555
artillery Baebelo mine, 95, 97
overview, 346-350, 349/ bag-bellows, 103
matliematics and, 783 baking
in naval harbors, 654-655 overview, 378-379
Roman, 698-699 division of labor in, 406-408, 407/
See also warfare and fortification ovens, 377/
arubah mills, 145, 356 representations of, 38, 408
Arzeron of tlie Queen, 568 Balbus, 26, 782, 783
asbestos, 469 Balkans, 95
ash used in glass, 525 ballistics. See artillery
assault ramps, 704, 705/ Bammer, A., 231-232
asses. See transport on land by harnessed banausic arts, 5, 803, 815
animals Barbegal mills (France), 149-150, 149/
Assos sarcophagi, 403 barding covers, 702
Assurnasipal II of Assyria, 290 barley, 165
Assyrian water supply systems, 290 Baroulkos (Hero), 39-40, 338-339,
astrolabes, 751-752, 791, 793 344-345
astrology, 745 barrel vaults, 272-273, 274
astrometeorology, 746 Basalla, George, 74
Astronomica (Hesiod), 742 bases, naval, 654
astronomical observation, 741-745 Basilica of Maxentius, Rome, 259
asymmetrical plowshares, 170 Bassae temple, 245!, 247-248
Athenaeus baths
on Archimedes' helix, 344 barrel vaulting, 272-273
on Corinthian bronze, 434 of Caracalla (Rome), 147, 402
on Nysa statue, 788 at Chester (Britain), 275
on Plato's clock, 786-787 of Diocletian (Rome), 269/
on salted fish, 386 heating of buildings, 273
on wine, 164 heating of Byzantium bath, 139
Athena Promachus bronze, 433 Morgantina, Sicily, 275-277
Athens reservoir cisterns for, 305
defensive signal system, 686 Roman power and, 27
fortifications, 683-684 water system and, 307-308
840 INDEX
beams, 248-249, 454 bows, 346, 698

Beckmann,Johann, 67,72 box joint, 453
beehive ovens, 377f, 378-379 boxwood, 450
beehive-shaped glassworking furnaces, 526, Bradford, J., 640
526f Bradley, M., 73, 80
behavioral chains, 823 braiding, 475
behavioral theories, 823-824 brass, 110-113, 771/
Belevi quarry, 124, 124/ bread production. See baking
Belisarius, 357 bread wheat, 170
bellows, 102-103 breaking machine (fractaria machina), 357
belly shooter (gastraphetes), 346 breakwaters, 642, 647-649
Belopoeica (Hero), 24 breeding of livestock, 177-180
Belopoeica (Philo of Byzantium), 21-23 brick
bench of Andreas, 345 ceramics and, 506-507
bench of Hippocrates, 345 construction with, 263-265, 265/
bench of Neileus, 345 industry, 272
bending moment design, 246-247 mass production of, 401-402
bending of wood, 460 pitched brick vaulting, 274-275
benefication of ores, 101, 101!, 102/ ribbing, 270-272, 271/
Benignus, Quintus Candidus, 28 bridges
Berlin Foundry Cup, 36, 54, 419, 420!, Augustan, 571-572
421!, 422/ in the Babylonian Talmud, 27
berthing, 651 depictions of, 44
Beytriigen zur Geschichte der Erfindungen engineering outside Italy, 573-574
(Beckmann), 67 in Italy, 569-573
bibliotheke ton enkteseon (record office of Julius Caesar's bridge over the Rhine,
landed property), 732 345, 441
Biographical Objects (Hoskins), 73-74 Narni, 571, 572/
biological roots of technology, 3-4 Pons Aelia, 572
bipedales, 265, 265f, 267, 267f, 272 Pons Aemilius, 569
birds as game species, 188, 192, 193t-203t, Pons Sublicius (Pile Bridge), 454
204 Pont du Gard (Nimes River, France), 299
bitumen, 139 Ponte Alto, 557
bleaching, 476 Ponte Antico, 557, 572
Bled Zehna aqueduct, 302/ Ponte Cecco, 571
blind dowels, 241 Ponte del Diavolo, 570, 572
Bloch, M., 71, 149 Ponte della Catena, 570
bloomery process, 107 Ponte Diavolo, 566
blown glass. See glass Ponte di Pele, 557
blowpipes, 536-537 Ponte di S. Cono, 570
Bliimner, Hugo, 35-36, 67 Ponte Maggiore, 557
Boardman, J., 499 Ponte Manlio, 570
boat mills, 357 Ponte Milvio, 570
Boeotia, 311 Ponte Nona, 569
Bokonyi, S., 181 Ponte Rotto, 558
bone ash in glass, 525 Ponte Salario, 570-571
book rolls, 719, 721-723, 722/ Ponte Sambuco, 566
books. See writing and book production Ponte Solesta, 571
bottle cisterns, 287 Trajan's bridge over the Danube,
Boule (Council), 727-728 455f, 456
Bourdieu, P., 740 on Via Appia, 557-558
boustrophedon script, 718 Volturnus bridge, 567
bow and drill, 41f, 43, 445 wooden, 455-456
INDEX 841
Britain Canachus, 427

Chester baths, 275 Canal de l'Est (France), 334-335
coal in, 139 canals, 206, 333-335, 660
fortifications in, 705-706 See also aqueducts
mills, 357 capstans, 340, 343-344
pottery in, 503, 505, 514 Caputi Hydria, 42, 44!
ships in, 623 Caracalla baths (Rome), 147, 402
water-lifting devices, 352f caravel/as, 627
wetland reclamation, 312 Car Dyke, 312
bronze carpentry, depictions of, 42f
alloys and color, 431-432 carpet knots, 473
armor, 424-425 carroballistae, 348, 350, 699
Berlin Foundry Cup and, 420-421, 42of, Carthage
421f, 422f aqueduct shafts, 310
brass vs., 110, 113 cisterns in, 288
cauldrons and other vessels, 425-426 dockyard, 658
as construction material, 241-242 Zaghouan-Carthage aqueduct, 301f, 303
figurines, 422-423 carts and wagons
large-scale casting, introduction of, 424 catapults mounted on, 348
market for bronze statuary, 436-437 diversity of, 596-598
number of statues, 434-435 hamaxai (two-wheeled cart), 592, 598
pumps, 353 for quarried rock, 125
screws, 341-342 terakukloi (four-wheeled wagon), 592
suitability of, 425 See also transport on land by harnessed
tablets of, 731-732 animals
weapons and armor of, 674 cash-based market, 775-776
brooches from Roman Britain, 113 See also coinage
Buchanan, R. A., 71, 74 cashmere, 467
buckets, 353 castella, 706
Buildings (Procopius), 31 castella divisoria (distribution tanks), 302-303
building technology. See construction and casting
engineering bronze armor and vessels, 424-426
business of architects, 228-229 bronze statuary, 427-431
Butcher, K., 772 glass, 530, 531
butchery, 384-385 cast iron, 107
butter, 385 catapults, 346-347, 688, 699, 783, 787
butt-joint, 452f Cato, 380, 487
buttressing arches, 269, 269f Cato the Younger, 730
cattle. See animal husbandry
Caesar, Julius cauldrons, 425-426
building campaign, 127 cavalry, 680, 701-702
force movement and, 552 cedar, Lebanese, 450
funeral of, 797 Celer, Petronius, 331-332
Rhine bridge, 345, 441 cella (sekos), 235
siegework, 703 Celsus, 26
on Veneti ships, 625-626 Celtic technology
Caesarea harbor, 646f, 647, 650, 651f boats and ships, 625-627, 627f
calendars, 742-745 helmets, 700
Caley, E. R., 112 transport, 597
Callixeinus of Rhodes, 788 cement, 236
Calpurnius Piso, L., 726, 731 cementation, 111
camelina sativa (Gold-of-Pleasure), 485 Centumcellae (Civitavecchia) harbor, 648-649
cams, 787-788, 788f ceramic pipes, 303
842 INDEX
cerd, 140 citrus, 450

cereals City of God (De Civitate Dei) (Augustine),
arable cultivation, 165-169 30-31
baking, 378-379 city walls, 683-684, 706-707
grain milling factories, 412 clamp and dowel assemblies, 241
milling, 356, 374-378 (See also mills; Clarke, David, 76
water-mills) classification, 74, 726
mortar and pestle, 373-374 Claudius, 327
processing of, 371f clay, 238-239, 503
threshing, 373 Clemens, Varius, 331, 332
water-powered pestles, 151 Cleon, 20-21
Cetius Faventinus, 750 Cleon of Atliens, 490
Chaerephanes, 312 Clistlienes, 727
chain mail, 108 clocks. See timekeeping
chain pump, 355 closable harbor (kleistos limen), 655
Chalcidian League, 250 clotliing, depictions of, 56-57, 58f, 59f
charcoal, 138-139 coal, 139, 432
chariots, 680-681 Coates, John, 620
cheese, 385 codices, 719, 723-724, 813
cheires siderai (grappling hooks), 688 cogwheels, 787, 791
cheiroballistra (hand-ballista), 348, See also gears
349[, 350 cohorts, 694
cheirolabe, 345 coinage
chemical energy, 138-139 continuous development and, 812
Chemtou marble workshops, 409, 410f copper and brass, no, 111f
Chemtou mill (Tunisia), 356 as facilitator of technologies, 774-776,
Cheops ship, 607, 608/, 614 774!, 775f
chest of Galen, 345 mass production, 401
chest of Nymphodorus, 345 metallurgy and, 93
chickens, 180, 183, 204 as new technology, 772-774, 773f
See also animal husbandry technology of, 769-772, 771f
Childe, V. Gordon, 71, 72 weights and, 764
China, 362 collegia, 142, 443, 600-601
chisels, 125, 447 colonnades, interior, 244
chocolate-bar ingots, 770 colored glasses, 526-528
Christianity Colosseum, 259
books and, 725, 726 Colossus of Rhodes, 661
bread wheat and, 170 Columeau, P., 186
codices and, 723-724 Columella
technology, Christianization of, 30-31 on animal breeding, 179
churches, 31 De re rustica, 782
Cicero on foods, 385
Greek sculptures and, 435 on mules, 590
on planetaria, 791-792, 793 on olives, 164
on public records, 730 on star calendars, 744
Romans, praise for, 804 column drums, 130-131, 233f
on vulgar occupations, 5 columns, 130-131, 235, 258-259
circuit walls, 683-684, 706-707 combustion, 138
circular saw, alleged, 341 Companions, 680
circumvallation, 703 competition among technicians, 32
circuses, 360-361 complex technological systems (CTSs),
cisterns, 287-290 830-832
cisterns, reservoir, 304-305, 306f compound pulley, 344
INDEX 843
concrete solar power and, 137-138

development of, 260-261 See also bridges; roads
facing, 261-265, 263f Construction of Catapults (Belopoeica) (Philo
hydraulic concrete for harbors, 644-645, of Byzantium), 21-23
811-812 consumption, conspicuous, 308
vaults and domes, 266-269 contractors (redemptores), 257
wood and, 454 contravallation, 703
Conophagos, C. E., 102 contus (lance), 702
consonantal writing, 717 conversion of energy, 153-154
Constantine, 29 cooling with ice or snow, 383-384
Constantinople, 305, 707 Coolus helmet, 700
Constantius Chlorus, 704 copies of statues, 435-436
construction and engineering copper, 106-107, 527, 771
architectural geometry, 782 core-formed vessels, 530
ceramics and, 506-507 Corinth, Isthmus of, 334, 592-594, 660
continuous development and, 811-812 Corinth bronze industry, 433-434
Greek Corpus Agrimensorum Romanorum, 16,
overview, 225-227 26,782
archaeological evidence, 227 Corpus Hippocraticum, 16, 17, 20
design, 228-229 Cosa/Heba road, 567
design bending moment, 246-247 cothon (artificial harbor basin), 642-643
fire protection engineering, 251-252 cotton, 469
geotechnical, 229-234 coulters, 170
as interdisciplinary, 251-252 Coulton, J. J., 64, 226, 228, 241, 251, 344
lime-based building materials, 235-238 counter-excavation technique, 323, 326
metal building materials, 240-244 cranes, 37, 39, 342-345, 343f, 653
Parthenon, inscription describing cranks, 354-355
construction of, 18 Crates, 311
peristyle and cella as independent units, Crete, glasswork in, 534
235 criminals, 100
roof slope, 240 Crispinus, L. Novius, 331, 332
specifications and models, 40 Crocodilion River mill, 147, 148, 148f
stone building materials, 244-246 cross-bar wheel, 597
stone roof tiles, 247-248 cross beams, 248
terracotta roof tiles, 238-239 cross vaults, 268-269
wood building materials, 248-251 crowbars, 339
Roman crucibles, lidded, 111, 112f
arch, vault, and trabeated architecture, crucible steel, 108, 109
257-260 crucifix, 824
architectural plans, 40 Crummy, N., 74
built environment, 256 crushed glass, 540
concrete and its facings, 260-266 crushing of grapes, 379-380
depictions of, 47 crushing of olives, 380-381
heating systems, 273 Ctesibius, 340, 346, 360, 753, 807
labor supply and contracts, 257 cuirasses, 701
marble trade and, 126-128 cumulative synthesis approach, 77
provincial techniques, 274-278, 276f cuneiform writing, 716-717
structural reinforcement of vaults and cuniculi (Etruscan drainage galleries), 312,
arches, 269-273 325-326
technological innovation, conditions of, Cuomo di Caprio, N., 503
278-279 cupids (erotes), 38, 44-45, 54-55
vaulted structures, large-scale, curdling, 385
266-269 curing, 485-487
844 INDEX
currency. See coinage Description and Analysis of All Figures

cursus publicus (imperial message service), (Balbus), 26
600 design stage in architecture, 228-229
cut-loop pile, 473 Destree, A., 71
cut stone floors (opus sectile), 131 deus ex machina, 360
Cyclopes, 4, 440 devices. See gadgets and scientific instruments
Cynegetica (Xenophon), 186 Diels, H., 67, 787
Cyprus, 95-96, 106-107 dies, engraved, 772
Cyropaedia (Xenophon), 394-395 differential adoption, 827-830
differentiation, technological, 824-826
Dacia, 26 diffusion theory, 79-80
Daedalus, 423,427, 442, 447f digging, 164, 165, 286
Dalley, S., 805 Dimension of the Circle (Archimedes), 796
damask twill, 473 Dinsmoor, W. B., 241
dams, 309-310 Diocletian's Edict on Maximum Prices,
Daumas, M., 73 132-133, 143, 388, 601
Davies, 0., 68 Diodorus, 423, 679, 788
Day the World Took Off, The (TV), 65 Diodorus Siculus, 346, 688, 702, 802-803, 805
De aquaeductu urbis Romae (On the Aqueducts diolkos (portage road across Isthmus of
of the City of Rome) (Frontinus), 25-26, Corinth), 592-594, 594!, 660
64, 780-781 Dionysius I, 698, 802
De architectura (Vitruvius), 23-24, 225-226, Dionysius of Alexandria, 346, 787
338, 808 Dionysius of Halicamassus, 552
de Camp, L. Sprague, 68 Dionysius of Syracuse, 346
De causis plantarum (Theophrastus), 163 Diophantus, 780
De Civitate Dei (City of God) (Augustine), Dioptra (Hero), 338-339, 341-342
30-31 direct screw presses, 359-360, 814
deer, 188 discovery, theory of, 77
defensive fortifications. See fortifications "Discus-Thrower" (Myron), 436
deflections, in construction, 228 disease in animals, 185
Defoe, Daniel, 66 distaff, 470
deforestation, 105, 139, 153 distillation, 138, 383
Deipnosophistae (Athenaeus), 164, 386 division of labor, 5, 394-395, 405-408, 407f
Delaine, J., 65, 257 Divus Claudius, Temple of, 306
Delambre, J.B. J., 754 Docimium quarry, 123!, 130!, 132
Del Chicca, F., 64 docks. See harbors
Delian mill, 375 documents. See record-keeping; writing and
Delos quay and breakwater, 642, 650 book production
Delphi, 27-28, 232 dogs, 187t
Del Rosso, R., 205 Dolaucothi mines (Wales), 94-95, 114
Demetrias of Ephesus, 229 dolia, 508
Demetrius I of Macedonia, 810 domes, 266-267, 267f
Demetrius Poliorcetes, 688 Domitian, 566-567
democracy, written records and, 727 Domitius Ahenobarbus, Cn., 574
denarii, 770-771, 775f donkey-driven mills, 376, 377f
depictions. See representations of technical door pivots, 241
processes Doric columns, dynamics of seismic forces on,
De rebus bellicis (anonymous), 29-30, 66, 235
358-359 "the Doryphorus" (Polyclitus), 435-436
De re metallica (Agricola), 66, 113 double mold-pressing, 533
De re publica (Cicero), 791-792 double-pitched roofs, 457
De re rustica (Columella), 782 dough-kneading machines, 152, 358, 378
Derry, T. K., 72 dousing, 98
INDEX 845
doves, 192 Eisagoge of Geminus, 743

dovetail joint, 452/ elements, concept of, 106, 521
dowels, 241 elephants, 681
Drachmann, A. G., 68, 344, 358, 795, 809 Eleutherae, fortress at, 686, 687f
drachmas, 765 Emissario Romano, 211
drachme, 764 Empedocles, 106
drainage, 234, 311-312 emporia, 653, 654, 655
dredging, 662-663 endless screw, 789
drilling energy and power
in Homer, 440 overview, 136-137
images of, 40, 41f, 42f, 43/ chemical, 138-139
woodworking drills, 445 economic decline of Roman Empire and,
drinking horn, magic, 790-791 153-154
drop-tower mills, 145, 356 human, 137, 140
Drover Berg Tunnel (Germany), 328-329, potential, 584-588
328f solar radiation, direct, 137-138
Drusus Canal (Germany), 335 wind, 152-153
dry distillation (pyrolysis), 138 See also animal power; water power
drying by sunlight, 138 Engels, F., 69, 71
dry measures, 761-762, 762f engraved dies, 772
Dubois, C., 639 Enneakrounos, 294
ducks, 192 environmental impacts of mining, 105
dupondius, 771 ephebeia (military training), 682
Durkheim, E., 740 Ephesus, 231, 650
dyeing, 476-477 epigraphs, 27-29, 721
epistatai, 18
"eared" plows, 166, 168f epitaphs as sources, 28-29
earthquakes and construction, 259 Equilibrium of Planes (Archimedes), 338, 339
Ebro valley (Spain), 309, 310 erotes (cupids), 38, 44-45, 54-55
Eckardt, H., 74 Esarhaddon, 290
economic histories, 68-70 ethnoarchaeology
Edgerton, D., 82-83 overview, 821-822
edge-to-edge brick barrel vaulting, 274 behavioral theories and life-history
Edict on Maximum Prices of Diocletian, framework, 823-824
132-133, 143, 388, 601 differential adoption and performance
education, 735 matrices, 827-830, 829t
eggs, 188, 204 electrical technologies as model for, 823
Egnatius, Cn., 575 invention cascades and complex
Egypt technological systems (CTSs), 830-832
engineers and technical traditions in, 20-21 technological differentiation and technology
glassworking in, 527 transfer, 824-826
irrigation in, 309 Etruscans, 312, 325-326, 717
mathematics in, 779 Euctemon, 743
metalworking and, 423-424 Eumenes II, 296
mining in, 96 Eupalinus of Megara, 294, 324
papyrus rolls, 721-722, 722f Eupalinus Tunnel, Samos, 294, 324-325, 325[,
pitched brick vaulting in, 274 782-783
quarrying in, 114-115, 129 Euripides, 727
Roman provincial record-keeping in, Eurysaces tomb, 38, 51-53
732-733 Eusebius of Caesarea, 724
shipbuilding, 614-617 euthytonoi (straight spring), 347
water-lifting devices, 162, 350, 352, 355 Eutropos sarcophagus, 43f
wool in, 466-467 Evans-Pritchard, E., 740
846 INDEX
evolution, historiography of, 74-75, 79 fortifications

experimentation phase, 826 continuous development and, 811
Greek, 682-686, 685f
fabrics. See textile production mathematics and, 783
facing, concrete, 261-265, 263f Roman, 705-707, 708!, 709f
factories. See large-scale production See also siege warfare
fanciful figures, 54-55 forums, 27
farming. See agriculture; animal husbandry Fossa Mariana, 335
Feddersen Wierde site (Germany), 170 foundation embankments, 568
Fellmeth, U., 581 foundations, 230-234, 232f, 259-260, 26of
fermentation of grapes, 383 fountain houses, public, 294
fibers for textiles, 466-469, 468f fountains, public, 306-307
figurines, 50, 507 fowling, 188, 192, 193t-203t, 204
Finley, Moses, 5-6, 63-64, 68, 70, 75, 77, fractaria machina (breaking machine), 357
372-373, 803, 804 free blown glass vessels, 538
fire, 4, 138 freedmen, 257
fire protection engineering, 251 free shaping of glass, 534
fish consumption, 385-386 Frontinus, Sextus Julius, 25-26, 64, 303,
fish farming, 206, 211-213, 212f 780-781, 804
fishing, 204-206, 207t-21ot fruits, 384
"five simple machines," 786, 802 Fucine Lake, 327, 343
Flaminius, Gaius, 558 Fiille, G., 397
flax, 468 fullers, 475-476, 476f
flechettes (plumbata), 698 fumes from smelting, 105
Fleury, P., 64 Furius Philus, L., 791-792
floating mills, 143-144, 357 furnaces
floors, 131, 237-238, 456 glassworking, 526, 526f, 536, 537f, 541
flute-player, mechanical, 338, 360, 788 smelting, 102, 103f
"flying needle" technique, 472-473 furniture, 459-460
fodder crops, 182 furs, 484
fodder management, 181
foggara. See qanats (underground tunnels) gabled roofs, 457
food processing and preparation gadgets and scientific instruments
overview, 369-373 overview, 785-787
animal by-products, 384-386, 387f automata, 24, 54, 789-791, 802
baking scene, 38 hodometers, 358,766, 794-796
cereal processing, 371f, 373-379 in Roman Empire, 796-797
dough-kneading machines, 152, 358 technological boom of 3rd century B.C.
game meat, 188 and, 787-789
grape and olive processing, 372f, water clocks and astronomical instruments,
379-384 340,752-754, 791-793, 814
innovation and invention, See also water-lifting devices
386-389 Gaitzsch, W., 446
large-scale production, 409-412 Gaius Cestius, 707
oil and wine presses, 359-360 Galen, chest of, 345
vegetables, fruits, and nuts, 384 Gallatin Painter, 42f
water-powered grain pestles, 151 gamebirds, 188,192, 193t-203~ 204
See also agriculture game farming, 186-188, 189t-191t
footwear, 487, 491-492 gardens, ornamental, 169-170
Forbes, H. A., 164 Gasparin, Cte. de, 588
Forbes, Robert, 68 gastraphetes (belly shooter), 346
force pumps, 97, 353-354, 354!, 807 gears, 340-341, 787, 788f
Forma Urbis, 40 geison (crown moulding), 248
INDEX 847
Gell, A., 740 grappling hooks, 688

Gellius, Aulus, 747, 786 gravel roads, 567-568
gemstones, counterfeit, 527, 540 great wheel, 42
geometry, 26, 782, 783 Greek and Roman Technology (White), 64
Georgics (Vergil), 166 Gregory of Nyssa, 274
geotechnical engineering, 229-234 Gregory of Tours, 147
Germann, K. von, 246 griffins' heads, 425
Gibbs, Sharon, 740, 741 groma, 766-768
gilding, 431 ground water drainage, 234
Gille, B., 585 grouting, 235
Gimpel, Jean, 71 Guardapasso viaduct, 564-565
gladius Hispaniensis, 696-697, 696f gynecology, 342
glass
in classical Greece, 529-531 Hadrian, 553
colored glass, 526-528 Hagia Sophia (Constantinople), 31
colorless glass, 528-529 half-hours, 750
glassblowing, Roman, 535-539, 537/ half-lap joint, 452[, 453
glassmaking, 521-522 Halieis, 655
glass pottery, 531-534, 535/ Haltwhistle Bum mill (Britain), 143
mass production, 400-401 hamaxai (two-wheeled cart), 592, 598
other decorative techniques, 540 Ham Dial, 751
output of glassblowers, 541-542 hammers, water-powered, 151-152
stepwise change and, 812-813 hammonitrum ("sandsoda"), 528
working properties of glass, 522-526, 523[, Handbook of Ancient Water Technology
524!, 525!, 526/ (Wikander), 64
workshops, primary and secondary, hand mills, rotary, 375-376
520-521 Hannibal, 688
glazed pottery, 513 harbors
Glotz, G., 69 ancient technical literature on, 643
glues, 235-236, 453 context of, 663-664
gnomon (sundial), 746-748 development, historical overview of,
goat hair, 467 641-643
Goblot, H., 292 hauling-ways, 660-661
gold, 104, 106, 770 history of research on, 638-641
See also mining and metallurgy lighthouses, 661
gold leaf, 531 moles, breakwaters, and quays,
Gold-of-Pleasure (camelina sativa), 485 647-653
gold thread, 470 naval, 654-656
Goody, J., 740 port activity, 653-654
Gordon, R. B., 76 sea level change and, 663
gouges, 447 shipbuilding yards, 662
grain processing. See cereals shipsheds (neosoikoi), 657-660
granaries, 169, 171/ siltation and dredging, 662-663
granite, 127-128, 130-131, 246 technological evolution, 643-647
grape cultivation, 163-164 wood used in construction of, 454
grape must (mustum), 383 harmony, 812
grape processing and winemaking harnessed animals. See transport on land by
overview, 372/, 379 harnessed animals
ceramic containers and, 508 Harpalos, 28-29
crushing and treading, 379-380 harvesting equipment, 168-169
fermentation, 383-384 Haterii tomb, 37, 39
large-scale, 412 Hayden, B., 827
presses, 359-360, 381-383 hay meadows, 170
848 INDEX
heating Higginbotham, J., 211

chemical, 138 Hill, D. R., 64
in construction, 273 hinges, 241
solar, 137-138 Hippocrates, bench of, 345
water, 307-308 Hippocratic texts, 16, 17, 20
See also energy and power hippodromes, 360-361
Hegeso, 441 Histoire generale des techniques (Daumas), 73
Heidegger, M., 74 Historia plantarum (Theophrastus), 163, 249
Hieron II of Syracuse, 661 historiography of technology
hekatompeda ("hundred-foot" temples), alternative technology, 75-76
231-232, 244 beginnings, 62-68
heliotropion, 743 economic histories, 68-70
helix, 344 ethnography and material culture, 73-74
Hellmann, M.-C., 64, 248 evolution and progress, 74-75
helmets, 674, 676, 696/, 700-701 relevance and continuity, 72-73
hemp, 468-469 revolutions and determinism, 70-72
Hephaestus, 54, 419-420, 421, 433 See also theoretical approaches
Hephaisteion (Athens), 233[, 237 History of Civilisation, 69
Heracles, 54 History of Many Memorable Things Lost, Which
Herculaneum, 303 Were in Use among the Ancients, The
Hermes of the Agora, 435 (Pancirolli), 66
Hermodorus of Salamis, 127 History of Mechanical Inventions (Usher), 67
Herod, 298 History of Technology (Singer et al.), 67-68, 72
Herodes Atticus, 797 Hitzl, K., 761, 765
Herodotus hobnailed shoes, 487, 491-492
on cauldrons, 425 Hodge, A. T., 64, 248
on Egypt, 614 Hodkinson, S., 160
on harbors, 642, 657-658 hodometers, 358, 766, 794-796, 796f
on mechana, 251 holkoi, 660
on navigation, 628 hollow voussoirs, 275, 276f
on roads, 552-553 Homer
on tunnels, 325 on Cyclopes, 4
on Xerxes canal, 334 navigation described by, 628
Hero (Heron) of Alexandria ships in poems by, 617-618, 621
automata, 24-25, 154 standardization of texts of, 725
Baroulkos, 39-40, 338-339, 344-345 on strap drill, 440
on glass, 529 Honeycomb Bowls, 539
on hodometers, 358 Hopfen, H.J., 588
magic drinking horn, 790-791, 79of Hoplite warfare, 674-679, 675f, 676f
mathematics and, 780, 782, 783 hopper mills, 374-375
on mechanics, 337, 338-339 Horace, 780
on olive-presses, 359 horizontal-wheeled mills, 144-145, 146/, 356
Pneumatika, 789 horologium Augusti, 746
reputation of, 808-809 horses, 180, 680, 701-702
on screws, 341-342 See also animal husbandry; transport on land
systematic treatises and, 815 by harnessed animals
wind-powered organ, 152-153 Hortensius, Q., 186
Hesiod, 163, 742, 801 horticulture. See agriculture
Hesnard, A., 649 Hoskins, Janet, 73-74
Hezekiah Tunnel, Jerusalem, 324 hours, 749-751
hieroglyphic writing, 716 See also timekeeping
Hieron II of Syracuse, 796, 806-807 House of Simon the Shoemaker, 487
Hiero of Syracuse, 23 housing of livestock, 183
INDEX 849
Hughes, Thomas, 830 invention cascades, 814, 830-832

Hultsch, F., 760 inverted siphon, 295-296, 295f, 297,
Humayma, Trajanic fort at, 706 299-300, 302/
Humphrey, J. H., 361 Ionia, 683
hunting, 185-186 iron
hunting dogs, 187t blowpipes of, 536-537
hydraulic concrete, 644-645, 811-812 as building material, 240-242, 242f, 243/
hydraulic engineering. See water management emergence of, 93
and hydraulic engineering in harvesting equipment, 168-169, 170
hydraulics and building construction, 234 introduction of, 419-420
hydraulis (water organ), 360 smelting of, 107-109
hysplex, 361 tunnel-building and, 322
weapons of, 673-674
ice, 383-384 irrigation, 161-162, 355
ichthyology, 205 Isthmus of Corinth, 334, 592-594, 660
Ictinus, 251 iugerum, 766, 780
Ihde, D., 74 Iulius Alcimus, C., 491
Iliad (Homer), 677 Iulius Helius, C., 491
illustrations. See representations of technical
processes Janiculum mills (Rome), 149-150
imperial message service (cursus publicus), 600 jars, tablets stored in, 728
Impey, 0., 501 javelins, 677, 695-696, 697-698
impluvia, 289 Jerash (Gerasa) Jordan sawmill,
inclined planes, 339 150-151, 151/
industry. See large-scale production Jerusalem glass workshop, 527, 535
information technology. See writing and book jetties, 650
production Johnson, J. Seward, 432
information transfer, 825-826 joinery, 450-453, 452/
Die Ingenieurtechnik im Alterthum jointing in foundations, 234
(Merckel), 67 Jones, Wilson, 258-259
innovation. See invention, inventors, and Josephus, 348
innovation Julian, 704
Inquiry into Plants (Theophrastus), 226-227 Julian calendar, 744
inscriptional archives, 729 Justinian, 31
inscriptions, 27-29, 718, 721
See also Nonius Datus inscription Al-Karaji, 322-323
insets in statuary, 431 Kelly, J.M., 231
instruments. See gadgets and scientific keys, 813
instruments Kibbutz Ma'ayan Barukh, 148
invention, inventors, and innovation kilns
in animal husbandry, 175, 177 construction of, 277
attitudes toward, 801-804 depictions of, 45-46, 46f, 47[
conditions influencing Roman innovation, firing in, 503-504
278-279 for terra sigillata, 505!, 506/
continuous development, 810-812 King's Highway, 552
food technology, 386-389 kleistos limen (closable harbor), 655
inventors, Greek and Roman, Klemm, F., 68
804-809 klepsydrai ("water thiefs"), 752-753
in mining and metallurgy, 93-94 kneading machines, 152, 358, 378
in quarrying, 129-132 Kondratieff, N. D., 69
stability and, 809-810 Ki:irber-Grohne, U., 170
stepwise change, 812-816 Kyrenia ship reconstruction, 612, 622-623,
theory, 5-6,77,78-79 622/
850 INDEX
labor force. See workforce Les Avrillages mill at St. Doulard, France,
La Graufesenque potteries, 398-399 142,144
Lambeau, F., 64 lever-and-drum presses, 382
lamps, 508 lever and screw presses, 359
Lancaster, L., 64-65 lever-and-weight presses, 381-382, 382f
lances, 702 lever presses, 359
Landels, J. G., 65, 226, 581, 789 levers, 339-340
land transport. See transport on land levigation, 502
by harnessed animals Lewis, M. J. T., 64, 81, 357, 788, 807
Lang, M., 761, 765 lex Oppia, 600
larchwood, 450 liberal vs. illiberal arts, 5
large-scale production Libon of Elis, 229
overview, 393-395 librae, 763, 765-766
Chemtou marble workshops, 409 libraries, 725-726
of coins, 773 Liebhart, R. F., 248
division of labor, 405-408 Lierke, R., 532, 533, 534
in food sector, 409-412 life-history framework, 823-824
marble trade, 402-405 lifting technology in Roman
mass-produced artifacts, 396-402 construction, 258
production sites, 395-396 See also water-lifting devices
lateral thrusts, countering, 268-270, 269f lighthouses, 661, 827-830, 829t
lathes lime-based building materials, 235-238
depictions of, 41-42 lime in glass, 521, 522
furniture making and, 459-460 Limenopoii"ka (Harbor Construction)
innovation of, 132 (Philon), 643
woodworking and, 445-446 limestone, 246, 570
Lato, 684, 685f linen, 468, 476
lattice ribbing, 267f Ling, W., 358-359, 362
Laurence, R., 581 lintels, arcuated, 270
Laurion mines, 100, 101 Lippincott, K., 741
Law of Constant Composition, 106 liquid measures, 762-763
laws and decrees, written, 718, 730-731 literacy, 733-735, 734.f
lead, 241, 419, 527 Livy
lead pipes, 300, 303, 304.f on Archimedes' catapults, 23
leather. See tanning and leather on Roman warfare, 691, 693, 698
Lebanese cedar, 450 on Via Appia, 555, 563
Lechaeum harbor, 643 Zizza paths, 124-125
Le Corbusier, 251-252 locatio conductio contract, 257, 397
Lefebvre des Noettes, Ct., 70-71, 581, 588, locked mortise and tenon fastenings, 6nf, 618,
590,600 621-622, 627-628
legal codices, 29 locks, 813
legion, Roman, 693-694 logs, 449
legionary denarii, 775 See also wood and timber
Leguilloux, M., 488 Longidienus, Publius, 51, 441, 442f
Lehmann-Hartleben, K., 639-640 Long Walls, 683
Lemonnier, P., 81 looms, 471-473, 472f, 814
length measures, 766-768 lorica segmentata (segmental plate armor), 701
Leningrad Painter, 42, 44.f lorica squamata (scale armor), 701
lenses, glass, 528 lost-wax method, 422, 427-431
Leonardo da Vinci, 795 Louis, P., 69-70
Leonidas of Tarentum, 205 Lucan, 628
Lepidus, M. Aemilius, 127 Lucian, 435
Le Rozier workshop, 399, 400 Lucretius, 4, 801
INDEX 851
Lucullan marble, 127, 131/ Marsden, E. W., 346, 348

Lucullus, L. Licinius, 127, 186 Martial, 725, 751
Lyceum, 787 Martin, R., 226
Lysippus of Sicyon, 430 Marx, Karl, 69, 71
Lysistratus, 430 Masada assault ramp, 704, 705/
masonry, 235-236
Ma'agan Mikhael shipwrecks, 619 Massalia (Marseilles) port, 640, 649,
Macedonian warfare, 679-682 659,662
machines mass production, 396-402
overview, 337-339 See also large-scale production
artillery, 346-350,349/ mathematics
cranes and traction, 342-345 geometry, 26, 782, 783
entertainment and, 360-361 Greek calculation tablet, 779/
extent of use of, 362 practical, 777-778
Hero of Alexandria on, 24 techniques, mathematical, 778-781
oil and wine presses, 359-360 technology, mathematics in,
simple machines or mechanical powers, 781-783
339-342 Mattingly, D., 581
social and economic development and, Mausolus of Caria, 132
361-362 Mazarron I shipwreck, 618, 621
surgical traction, 345-346 McCann, A. M., 206
transport and paddlewheels, 358-359 McNeil, I., 72-73
water-power diversification, 357-358 meadows, hay, 170
See also gadgets and scientific instruments; measurement. See weights and measures
water-lifting devices; water-mills mechana, 251
Madra Dagaqueduct, 296 Mechanica (attrib. Aristotle), 251, 787
magistrates, 727, 761-762, 780 mechanical advantage (MA), 339-340
mail shirts, 701 Mechanical Problems, 338, 340
Malone, P. M., 76 Mechanice syntaxis (Philo of Byzantium),
manganese in glass, 527 21-22, 338
manhole technique, 323 Mechanics, 344
manufactories, 396, 400 Mechanics (Hero), 344-345, 359
manufacturing. See large-scale production mechanics of forces, 584-588
marble mechanistic view, 77
Chemtou marble workshops, 409, 410/ medicine, 185, 342,345-346
epigraphs on, 721 medimnoi, 761
grades of, 246 Mediterranean ecosystems, adaptation to,
plaster, marble-compound, 236, 237 159-161
polishing and weathering of Mediterranean Triad, 370
statuary, 237 Megara, 683
quarrying of, 124 Megarian bowls, 510
roof tiles of, 247 Meiggs, R., 248
standardization and mass trade of, Meissner, B., 64
402-405 melting point of glass, 523-524, 524!, 525/
trade in, 126-132 mensores (measurers), 766-768
types of, 127-128 mercenaries, 681
veneers, 131-132 Merckel, C., 67, 71-72
wagon transport of, 599/ Merkel, J., 103
See also quarrying and stonework Merritt, F. S., 232, 244, 249
marble bureau (ratio marmorum), 127 Mesopotamia, 274, 521-522
Marius, 694 message service, imperial (cursus
Marius Canal (France), 335 publicus), 600
Mark Antony, 730, 775, 775/ metallurgy. See mining and metallurgy
852 INDEX
metalworking survival and development after fall of

overview, 418-419 Roman Empire, 113-114
alloys and color, 431-432 technology of mining, 96-99
bronze techniques and vessels, 424-426 tradition and innovation, 93-94
building materials, metal, 240-244, water-powered trip-hammers, 151-152
242f, 243f mirrors, 138, 528
copies, role of, 435-436 missile weaponry, Roman, 697-698
early bronze work and figurines, 422-423 mitre joint, 451, 452f, 453
Egyptian influence, 423-424 modius (shovel), 762
harvesting equipment, 168-169 Mokyr, Joel, 74
Hephaestus and ancient metalworkers, mold-blowing, 538-539
419-420 moldboard plows, 166, 169f, 170
House ofVettii scene of, 38, 44-45 molding on ceramics, 510
market for bronze statuary, 436-437 moldings, wooden, 459
number of bronzes, 434-435 mold-pressing, 531-533, 533f
statuary techniques, 426-431 Mole of the Windmills (Rhodes), 648
in vase paintings, 420-421, 420!, 421f, 422f moles and breakwaters, 647-649
workshops, evidence from, 432-434 moles necessariae ("indispensable structures"),
meteorology, 746 551
Meton, 743 Mommsen, T., 322, 329-330, 333
metretes, 762 money. See coinage
metrology, 760-761 monolithic columns, 130-131, 258-259
Metroon, 727-728 Mons Porphyrites, 127, 129
Midfendic, 312 Montefortino helmet, 700
milestones, 556, 766 Montel Salviano tunnel, 327
Miletus, 682, 683 Morgan, Lewis, 62-63, 71
military. See warfare and fortification Morgantina baths (Sicily), 275-277
milk, 385 mortar and pestle, 373-374
millponds, 143 mortared rubble construction, 260-261,
mills 264-265, 265f, 288-289
floating, 143-144, 357 mortaria, 508
grain milling factories, 412 mortars, 235, 261, 288
hopper, 374-375 mortise-and-tenon joint, 451-453, 452f
Pompeian, 376, 377f mortising chisels, 447
rotary hand mills, 375-376 mules. See transport on land by harnessed
sawmills, 150-151, 357 animals
See also water-mills Mumford, Lewis, 71
millstones, 147-148, 356 Mummius, 435
mina, 764-765 Muro del Peccato, 569
Minerva altar relief (Rome), 443, 444! Museum at Alexandria, 338, 787, 805-806
Minerva Medica temple, Rome, 267f music, 152-153, 360, 812
mining and metallurgy mustum (grape must), 383
beneficiation of ores, 101-102 Mycenaean roads, 553
brass, 110-113 Myron, 428-429, 436
depictions of, 37, 57f mythology, 4, 53-55
early Greek metallurgy, 422-423 See also Homer
environmental impacts, 105
iron and steel, 107-109 naalebinding (type of needlework), 474
organization and administration of mining, nailed footwear, 487, 491-492
99---100 nails, 241, 401, 451
refining, 105-107 name-stamps, 397, 509, 815
smelting, 102-104 naopoioi, 729
sources of metal ores, 94-96 Narcissus, 327
INDEX 853
Narni bridge, 571, 572f ceramic containers and, 508

nautae, 600-601 crushing, 380-381, 381f
naval harbors, 654-656 large-scale, 412
naval warfare, 686-688 presses, 359-360, 381-383
navigation, 628-632 separation, 383
Naxos dockyard (Sicily), 658, 659f Olympia, 229, 244, 432, 531
Needham, J., 358-359, 362 Olympias reconstruction, 612, 613f, 620-621
Negreiron-Negres (France), 183, 184! Olynthus mills, 374-375
Neileus, bench of, 345 onagri, 348-350, 699,706
Nelson, R. R., 74 On the Fluteplayer (Apollonius of Perge),
Neolithic period and animal power, 136 340-341
neosoikoi (shipsheds), 657-660 opaion roof tiles, 251
Nero, 456,797 open cast pits, 98
Neuberger, A., 67 Optics (Ptolemy), 529
Neuburger, Albert, 35-36 optimism, 801
New Bouleuterion, 728 opus caementicium (mortared rubble
Nile River, 614 construction), 260-261
Nilotic scenes, 54 opus craticium, 455
Nilsson, M., 740 opus incertum, 262, 263f
Nimes aqueduct (France), 301 opus mixtum, 262, 263, 263f, 264-265
Nineveh glassmaking, 522, 524, 527 opus reticulatum, 262-263, 263f, 279
Nin/Zaton boats of Croatia, 619-620 opus sectile (cut stone floors), 131
Noble, J. V., 754 opus signinum, 261, 289
nome capital and nome metropolis, 732 opus testaceum, 263-264, 263f
Nonius Datus, 28,326 orality, 715-716, 735-736
Nonius Datus inscription, 322, 326, Orata, 213
329-333, 33of organs, 152-153, 360
noria device, 351-353 Oribasios, 344, 345
Noricum iron, 108 Orlandos, A., 226
North, Douglass, 62-63, 74 orreries, 793
notices, public, 720 Ostia bakeries, 406, 407f, 408
nucleated workshops, 396 ostracism, 735
nudity, 56 ostraka (potsherds), 720
number notations, 778, 779 ovens,377f, 378-379
nutrition, veterinary, 180-183 See also furnaces
nuts, 384 overshot mills, 144, 146-147
Nymphadorus, chest of, 345 oxen. See transport on land by harnessed
nymphaea, 306-307 animals
Nysa statue, 788, 788f, 797 Oxrhynchus, 491
oxygen, 211-212
oak, 450 oyster cultivation, 212-213
obolos, 764
obsidian, 166 packsaddles, 589-590, 589t
Odyssey (Homer), 440, 617-618, 621, 628 See also transport on land by harnessed
Oeconomicus (On Household Management) animals
(Xenophon), 5, 163-164 paddlewheels, 358-359
oil presses, 359-360 pagination, 724
Old Bouleuterion, 727 Palestine, 96
Old Oligarch, 249 palintonoi (back-spring devices), 346, 348, 350
Oleson, J. P., 63, 68, 81, 643 Palladius, 380
olive cultivation, 160!, 164-165 Pancirolli, Guido, 66, 497
olive processing panoply, hoplite, 674, 675f, 677, 678
overview, 372f, 379 panoply, Roman, 696f
854 INDEX
Pantheon, 130, 258-259, 268 Mechanice syntaxis, 21-22, 338

Pappus,339, 344,786,802 Poliorcetica (Siegecraft), 783
papyrus and papyrus rolls, 719, 721 purposes of, 30
paradeigmata (models of buildings), 40 reputation of, 808
parapegmata, 742-744, 745-746 theoretical vs. practical knowledge
Parasceuastica (Philo), 352 and, 802
parchment, 719 on water-lifting devices, 351-352
Paros quarries, 124 Philoponus, 752
Parthenon, 18, 250-251 philosophical study, 24-25
pasturing, 182 Phoenicians
patination, 431-432 alphabet and, 717
patrons, 20, 55-56 harborwork, 641-642, 643-644
pattern welding, 109 navigation, 629
Pausanias, 423, 424, 427, 433 shipbuilding, 615-616
Pausias, 528-529 warships, 686
paving, 555-556 phosphoric iron, 107
Peacock, D. P. S., 396, 502 picks, 122-123, 129
Peloponnesian War, 677, 686 pigs, 181, 183
pentakosiomedimnoi, 761 See also animal husbandry
Pentelicon, Mt., 124-125, 127 pikes, 679
Penteskouphia, 37f, 45, 46/, 47/ Pile Bridge (Pons Sublicius), 454
peppercorns, 169 pile drivers, 345
Perennius, C., 397 piles, wooden, 259, 260/, 454
performance characteristics, 824 pilings, 454
performance matrices, 827-830, 829t piloi (conical felt caps), 678
Pergamom, 295-296, 361 pilum (javelin), 695-696, 696/
Periander, 340, 643, 660 Pinna nobilis, 469
Pericles, 18, 229 pipes for water, 300, 302-304
periploi (written sailing instructions), Piraeus, 654, 655-656, 656/, 658, 683
629-630 Pisa harbor, 650
Periplus Maris Erythraei, 629-630 Pisistratids, 229
peristyle, 235 Pisistratus, 294, 725
Perseus (Cellini), 428 pitched brick vaulting, 274-275
Persian warriors, 676, 676/, 683 pitched roofs, 457
Peschlow-Bindokat, A., 246 Place Jules-Verne shipwreck, 619
pessimism, 801 planes, 339, 446
pestles, 373-374 planetaria, 791-792, 793, 797
Petrakos, B., 232 plaster, 236-238
petrochemicals, 139 plate armor, 701
Petronius, 433-434, 797 platforms, revolving, 361
Phaeinus, 743 Plato
phalanx formation, 675-677, 679, 682 clock of, 786
Phanes, 774, 774/ on coinage, 775-776
Phari potter's workshop, 504 on division oflabor, 5
Pharos of Alexandria, 661 on glass, 521
Phidias' workshop at Olympia, 531 on necessity, 801
Philip II of Macedon, 679-682, 680, 681, 685 on Prometheus, 4
philology, 725 Plautus, 750
Philon, 643 plays, technical knowledge in, 19
Philo of Byzantium Pliny the Elder
on artillery, 346-347, 783 on bronze, 432, 433
Belopoeica (Construction of Catapults), on construction, 258
21-23, 339 on curing, 485
INDEX 855
on fractaria machina (breaking Pons Aemilius, 569

machine), 357 Pons Antico, 557, 572
on fuel shortages, 139 Pons Sublicius (Pile Bridge), 454
on glass, 522, 527, 528, 535, 538 Pont du Gard (Nimes River, France), 299
on helpless infants, 4 Ponte Alto, 557
on lathes, 132 Ponte Cecco, 571
list of origins, 805 Ponte del Diavolo, 570, 572
on mining and smelting, 97--98, 100, 104, Ponte della Catena, 570
106-107 Ponte Diavolo, 566
on Monte Salviano tunnel, 327 Ponte di Pele, 557
on plows, 166 Ponte di S. Cono, 570
on pottery, 504-505 Ponte Ladrone viaduct, 564-565
on Roman achievements, 804 Ponte Maggiore, 557
on Samian craftsmen, 424 Ponte Manlio, 570
on Theodorus, 231 Ponte Milvio, 570
on timber, 448-450 Ponte Nona, 569
on time, 750, 752 Ponte Rotto, 558
on wood bending, 460 Ponte Salario, 570-571
Pliny the Younger, 335, 648-649, 751 Ponte Sambuco, 566
plows, 166, 167f, 168!, 169f, 170 Ponte Solesta, 571
plowshares, asymmetrical, 170 Pontine marsh canal, 335
plumbing. See water management and Ponting, M., 772
hydraulic engineering Popilius, P., 558, 560
plumb lines, 767-768 porcelain, 497-498
Plutarch, 23, 429, 431, 432, 806 portable sundials, 751-752
pneumatic catapults, 346-347 Port Berteau 2 shipwreck, 627
Pneumatics (Philo), 351-352 porters, 589
Pneumatika (Heron), 789 ports. See harbors
poaching, 185-186 Portus harbor, 638, 639f, 640, 651, 661
poetry vs. prose, 716, 716f Portus workshop, 130, 131f
Poidebard, Antoine, 640 Poseidon, Penteskouphia sanctuary of, 45
pointilli technique, 123 Posidonius, 793, 803
Polanyi, Karl, 70 post and lintel systems, 269-270
Poliorcetica (Philo), 783 post-holes, 454
Politics (Aristotle), 5 Postumius, Sp., 558
Polybius, 23, 292, 804 pot-bellows, 103
polychrome glass objects, 540 potential energy, 584-588
polychrome marble, 127 pottery and ceramic production
Polyclitus, 429, 435-436 clay preparation and pottery forming,
Polycrates, 21, 294, 642, 725 501-503
polygonal foundation stonework, 232f, 234 consumption practices, diffusion of,
polyremes, 687 513-514
polyspastos, 344, 345 depictions of, 45, 47
Pompeian mill, 376, 377f firing technology, 503-504
Pompeii glazed, 513
carpenters procession fresco, 446, 447f historical, geographical, and cultural
street fountains, 306, 307 perspectives, 496-498
sundials, 750 mass production, 396-400
Temple of Venus, 132 perceptions of, 498-501
water system, 297, 302, 303, 304/ range of products, 504-509
See also Vettii House, Pompeii stability of technology, 810
pondus (pound), 765 surface finishes and decoration, 509-513
Pons Aelia, 572 workshops vs. manufactories, 396
856 INDEX
power. See energy and power quarrying and stonework

pozzolana, 261, 648 overview, 121
practical mathematics. See under mathematics construction with stone, 244-246
prefabrication, 131 Diocletian's price edict and, 132-133
preservation of meat and fish, 385-386, 387/ documents on stone, 728-729
presses for oil and wine, 359-360, 380-383, 382/ foundations, 230-234, 232/
pressing of glass, 531-534, 533/ Greek, 121-125
Price, D. J. de Solla, 754, 793 marble trade, Roman, 126-132
Price Edict of Diocletian, 132-133, 143, 388, 601 millstones, 147-148
primitivist view, 581 paving, 555-556, 562-568, 56ef, 565/
prisms, 787 preferential selection of stone, 244-246
proclamations, 731 Roman quarrying, 125-126
Procopius, 31, 357, 646 stone roof tiles, 247-248
production, large-scale. See large-scale volcanic, 267-268
production See also marble
production-line operation, 408 quays, 649-653, 651/
progress, historiography of, 74-75 querns, 374
Prometheus, 4
Prometheus Bound (Aeschylus), 19 rabbit hair, 469
prop-and-lintel system, 457, 457f, 458 rack-and-pinion gearing, 340, 341
Propylaea at Athens, 241-242, 242!, radial brick barrel vaulting, 272-273, 274, 275
244-245, 247 Raepsaet, G., 64
prose vs. poetry, 716, 716/ rafters, 248-249
proto-Sinaitic scripts, 717 rainwater-collection cisterns, 287-290
pruning, 165 ramps, 704, 705/
Psammetichus I, 423 rams, warships with, 620, 687-688
psychai, 38f, 44-45, 54 ranching. See animal husbandry
Ptolemies, the, 735 rangeland management, 182-183
Ptolemy, Claudius, 529 Raschpetzer-Tunnel (Luxemburg), 328
Ptolemy Soter, 431, 688 ratio marmorum (marble bureau), 127
Pugsley, P., 445 reading, 733-735
pulleys,340,342-344, 343/ See also writing and book production
pumps reconstructions of Mediterranean vessels,
chain pump, 355 611-613, 613/
force pump, 97, 353-354, 35ef, 807 record-keeping
punctuation, 724 overview, 726-727
Punic hydraulic linings, 288 Greek, 727-729
Punic salteries, 386 in Roman Egypt, 732-733
Punic Wars, 693 in Rome, 729-732
puppetry, 788 recycling of glass, 536
purification of water, 290, 301 Redde, M., 656
purity standard for metals, 106 redemptores (contractors), 257
purlins, 248-249 redesign phase, 826
puteal (wellhead), 286 red-slip wares, 396-397, 500, 502,
Puteoli (Pozzuoli), 557, 566 509, 510
pygmies, 54 refinements, in construction, 228
pyrolysis (dry distillation), 138 refining, 105-107
relief carving, 460
qanats (underground water channels), 162, relief decoration on pottery, 509-510
291-293, 292!, 322-324 relieving arches, 259-260, 260/
Qasr Ibrim, 491-492 religion and books, 725, 726
quadriburgia, 706 repeater catapults, 346, 787
quadrigati, 770-771 replication phase, 826
INDEX 857
representations of technical processes Via Latina, 555, 560

overview, 35-36 Via Nomentana, 563
artistic intent, 36-46 Via Ostiense, 564-565
baking, 38f, 408 Via Popilia, 558, 560, 567, 568
categories of depictions, 48-50 Via Postumia, 558, 567, 568, 573
clothing and personal accessories, 56-57, Via Praenestina, 564f, 569
58f, 59f Via Salaria, 561-562, 565-566, 570
harbors, 638 Via Sarsinate, 568
mythological or fantastic settings, 53-55 Via Tiburtina, 563, 564
range of depictions, 46-48 Via Valeria, 563
smelting, 102 Rodgers, R. H., 64
textiles, 465 Roeder, Josef, 126
woodcraft, 441-443 Rogers, D. M., 79
workforce, 55-56 ROMACONS project, 650
reproductive rate in animal husbandry, 181 Romano-Celtic boats and ships, 625-627, 627f
reservoir cisterns, 304-305, 306f Roman Empire, economic decline of, 153-154
revolving platforms, 361 Roman legion, 693-694
Rhamnus, temple of Nemesis at, 232 Roman numeral system, 780
Rhodes, 648, 656, 661 Rome, J. J., 267
Rhodes, R. F., 247 Rommelaere, C., 64
Rhoecus, 423-424 Romulus, 691
Riace Bronzes, 428, 429f, 43of, 431-432 roofing
ribs, vaulted, 270-272, 271f ceramic, 507
Rickets, J. T., 249 Corinthian style tiled roof, 239f
ridge beams, 248, 458-459 opaion roof tiles, 251
Rieti, 566 pitch, 240, 457
Rio Tinto mines, 95, 97, 98, 100, 101, 102f stone roof tiles, 247-248
roads terracotta roof tiles, 238-239
Arzeron of the Queen, 568 wooden construction, 456-459, 457f, 458f
construction and paving, 562-568, 564f, 565f rope-making, 470
design and layout of Roman roads, Rostovtzeff, M., 70, 75
560-562 rotary hand mills, 375-376
diolkos (portage road across Isthmus of rotary mold-pressing, 531-532
Corinth), 592-594, 594f, 660 rotary olive crushers, 380-381, 381f
engineering outside Italy, 573-576 rotary pressing, 533-534
gravel, 567-568 rotation of crops, 160-161
Italy, major Roman roads in, 558-560, 559f roughed-out statues, 404
pre-Roman roads, 552-553 Royal Road, 552
study of, 551-552 rubble, mortared, 260-261, 264-265, 265f,
transport systems and, 601 288-289
Via Aemilia, 558, 567-568 rubble breakwaters, 642
Via Amerina, 559 Ruttan, Vernon, 77, 79
Via Annia, 567
Via Appia, 553-558, 554f, 560, 563 saccarii (porters), 589
Via Appia Traiana, 567, 572 saddle joint, 452f, 453
Via Aurelia, 559, 560, 569-570 saddle querns, 374
Via Cassia, 559-560, 563 sagging technique, 531, 534
Via Clodia, 559 Sahrhage, Dietrich, 205
Via Domitia, 574 sails, 616-617, 620
Via Domitiana, 566-567 Salamis, 688
Via Egnatia, 574-575 Saldae aqueduct tunnel, 329-333
Via Flaminia, 558, 561, 563-564, 565f, 567, 569, Saliari quarries, 403
570, 571, 572f Salmon, J., 581
858 INDEX
salt preservation, 385-386, 387/ sewn-plank fastenings, 610/, 614-615, 618-620

Samian ware, 396-397, 398 shadufs, 162, 309, 350-351
Samos, 294, 423-424, 657-658 shaft furnace, 103/
Sanctuary of Hercules Victor at Tivoli, 271 shafts, 310, 323
Sanctuary of Zeus at Olympia, 432 shafts-and-gallery technique, 293, 311-312
Sanderac tree, 450 shears, 537-538
"sandsoda" (hammonitrum), 528 sheep, 179,180, 466-468
saqiya device, 351-353, 355 See also animal husbandry
sarcophagus epitaphs, 28 sheepfolds, 183, 184/
sarcophagus trade, 402-403 shellfish, 212-213
sarissa (pike), 679 shields, 674, 696/, 699-700
sawmills, 150-151, 357 ships and navigation
saws continuous development and, 811
circular saw, alleged, 341 Egyptian and Phoenician shipbuilding,
for stone veneers, 132 614-617
types of, 446 maritime environments, 613-614
water-powered, 150-151 Mediterranean-style vessels, 625
Saxon Shore forts, 706 nautical and navigational terms (glossary),
scale armor, 701 630-632
scales, 765 naval warfare, 686-688
scarf joint, 451, 453 navigation before A.D. 500, 628-630
Schmiedt, G., 640 reconstructions, 611-613, 613/
Schneider, H., 65, 67, 68 river transport as water power, 141
Schumacher, E. F., 76 Romano-Celtic boats and ships, 625-627
Schumpeter, J. A., 69, 70 shipbuilding 800 B.C. to A.D. 500, 617-621
Schiirmann, A., 80 shipbuilding yards, 662
scythed war chariots, 681 sources of evidence, 606-614
Scipio Africanus, 126 technological achievements of classical
Scoppieto workshop, 399-400 world, 627-628
scoria, Vesuvian, 268 technological changes 800 to 300 B.C.,
scratch plows, 166 621-623
screw presses, 359-360, 383, 814 technological changes after 300 B.C.,
screws, 341-342, 451, 788-789 623-624
screw-windlass, 661 See also transport
scriptio continua, 724 shipsheds (neosoikoi), 657-660
sculpture. See figurines; statuary Shirley, E., 65
scythes, 170 shoes, 487, 491-492
sea level changes, 641, 663 Side city wall, 709/
seasonal indicators, 742 sidewalks, 563
sea transport. See harbors; ships and navigation Sidon glasswork, 535, 538
Sebastos harbor, 650, 651/ siege warfare, 682-686, 702-704, 705!, 810
segmental plate armor, 701 sieving, 378
sekos (cella), 235 signals, defensive, 686
Semitic languages, 717 silica, 521, 522
senatorial decrees, 730 silk, 469
Seneca, 529, 803 siltation, 662-663
Sennacherib, 291 silver, 104, 770-771
separation of olive oil, 383 See also coinage; mining and metallurgy
Serino aqueduct, 298 Singer, Charles, 67-68, 72
Servius Tullius, 693 single-yoke harnesses, 595-596, 596/, 597/
sestertius, 771 siphon, inverted, 295-296, 295!, 297,
Severus, Septimius, 139 299-300, 302/
sewers, 311-312 skins as writing materials, 719
INDEX 859
slavery specialization, 405

architects, slaves as, 229 See also division of labor
mining and, 100 specula, 342
opus reticulatum and, 262 Spensithius, 718
pottery and, 398 sphyrelaton, 427
sleeper beams, 454 spindles, 470
Sleeswyk, A. W., 615, 795-796 spinning, 469-470
slingshot, 698 Spintharus of Corinth, 229
slips, in ceramic production, 510-511 spiral reticella bowls, 532, 534
slipways, 657 Spruytte, J., 590
Small is Beautiful (Schumacher), 76 stability of technology, 809-810
smelting, 102-104 stadia, 360-361
Smith, Adam, 394 stadion, 766
snow used for cooling, 383-384 stamp marks, 397, 509, 815
Social Construction of Technology (SCOT), standardization
80-81 Homeric texts, 725
social evolution theories, 4 marble trade, 402-405
Socrates, 724 statue-types, 428
soda, 521 weights and measures, 765, 768
sod-breaking tools, 166 star navigation, 628, 63of
softening point of glass, 523-524, 524!, 525f stars and timekeeping, 742-744
soil condition, 230 stater, 765, 77,if
solar power, 137-138 state symbolism on coins, 773
Solon, 761 Statius, 566, 567
sophists, 17 statuary
Sophocles, 4, 19 alloys, color, and patination, 431-432
sources and evidence bronzes, number of, 434-435
agriculture, 159 bronze techniques, 426-431
classical Athens, 17-20 copies, role of, 435-436
construction, 227 insets, 431
fishing, 205, 207t-210t market for bronze statuary, 436-437
fowling, 192 polishing and weathering of, 237
future work to be done on, 9-10 poses, 424
gamebird species, 193t-203t roughed-out, 404
game species, 189t-191t standardization of marble and, 404-405
general features of, 31-33 Staudenmaier, J. M., 80
Hellenistic kingdoms, 20-23 steel, 108
hunting dogs, 187t steelyard, adjustable, 813
land transport, 582 Steffy, J. R., 623
late antiquity, 29-31 steining, 286
mechanics, 338 step-rings, 267, 267f
metalworking, 418-419 stoichedon letters, 718
Roman Empire, 23-29 stone weights, 764
shipbuilding and navigation, 606-614 stonework. See quarrying and stonework
textile production, 465 stopcocks, 303, 30,if
types of, 15-17 Strabo
warfare, 673-674 on construction, 256
woodcraft, 440-444 on deforestation, 139
Spain, Islamic, 309, 310 on glass, 528
Spartans, 677-678, 682 on mining, 105
spatha, 696f, 697 on Pontine marsh canal, 335
"speaking objects," 718 on water devices, 162
spears, 677, 680, 695-696 strap drill, 43f, 440
860 INDEX
Strato of Lampsacus, 787 Technology in Western Civilisation (Kranzberg

street grid, 655 and Pursell), 73
strongholds, 682-684 technology shelf concept, 75
structural design, 228 technology transfer, 75-76, 79-80, 825-826
See also construction and engineering tegulae mammatae, 273
stucco, 236-237 Telecles, 423-424
Studies in Ancient Technology (Forbes), 68 tempered steel, 108
sulphur dioxide, 105 temple architecture. See construction and
sun. See solar power engineering
Sundbo, J., 78 Temple of Hephaestus, 433
sundial-calendar, 793, 794! Temple of Venus, Pompeii, 132
sundials and hours, 746-752, 748f, 749f, 814 temples and record keeping, 729
Suovetaurilia sacrifice, 177f terakukloi (four-wheeled wagon), 592
surgery, veterinary, 185 terminal reservoir cisterns, 305
surgical traction machines, 345-346 terracing, agricultural, 309
surveying, 554-555, 561 terracotta
swords, 675, 676, 696-697 figurines, 507
symbolism of the state on coins, 773 heating and, 273
syngraphai (verbal specifications), 40 hollow voussoirs, 275, 276f
Syria, 96, 575 pipes, 300
Syro-Palestinian glassblowing, 536, 539 roof tiles of, 238-239
in Tunisia, 277
tablets, 719-720, 720!, 731-732 terra sigillata, 396-397, 502, 504, 506, 510
tablet-weaving, 473-474 Testour mill (Tunisia), 356
tablewares. See pottery and ceramic production testudo, 307-308
Tabula Contrebiensis, 310 textile production
Tabularium, 729-730 overview, 465-466
Tacitus, 26-27 dyeing, 476-477
Tagliata canal at Cosa, 206 fibers, 466-469, 468f
talent, 765 finishing and fulling, 475-476
Tameanko, Marvin, 45 interlaced and non-woven fabrics, 474-475
tanning and leather looms and weave structure, 470-474
overview, 483-484, 484t spinning, 469-470
curing, 485-487 stepwise change and, 814
leatherworkers and tanners, 490-491 Theagenes, 293
procedures associated with tanning, 486t theaters, 289, 360, 455-456
scale of production, 490 Theatrum Pompeii (Theatre of Pompey,
tanneries, 488-489 Rome), 127
technological change, 491-492 Themistocles, 294, 683
vegetable tanning, 487 Theodorus, 21, 423-424
taps, 303-304 Theodorus of Samos, 225, 231
tap slag, 104 Theodosian Code, 600
Tarquinia/Gravisca road, 567 Theodosian Land Walls of Constantinople, 707
Tarracina (Terracina), 555 Theodosius, obelisk of (Constantinople), 258
Die Technik des Altertums (Neuberger), 67 Theophrastus
Technological Arts and Sciences of the Ancients on bleaching, 476
(Neuberger), 67 on charcoal, 138-139
Technological Choices (Lemonnier), 81 on coal, 432
technological differentiation, 824-826 on construction, 226, 241
Technologie und Terminologie der Gewerbe und on glues, 235-236
Kunst bei Griechen und Riimern Inquiry into Plants, 226-227
(Bliimner), 67 on lathes, 132
technology-in-use, 82-84, 813 on mirrors, 528
INDEX 861
on timber, 448-450 representations of, 41-42, 42f, 43f, 48-51

on viticulture, 163 sod-breaking, 166, 167f, 168-169, 168f
on wax, 435 woodworking, 443-447
on wood mechanics, 249-250 Toomer, G. J., 806
theoretical approaches Torlonia Relief, 638, 639f, 651
overview, 76-77 torno terere ("tooling on a potter's wheel"), 534
behavioral theories and life-history Torre Astura piscina, 212f
framework, 823-824 torsion artillery, 346-350, 349f, 698-699
diffusion and technology transfer, 79-80 torture implements, 345
discovery and innovation, 77 Towards a New Architecture (Le Corbusier),
innovation, 78-79 251-252
social construction theory, 80-81 Tower of the Winds, Athens, 749f, 753
technology, 78 trabeated architecture, 257-260
technology-in-use theory, 82-84, 813 traction, surgical, 345-346
See also ethnoarchaeology; historiography trade and marble standardization, 402-405
of technology trade and transport. See transport
theriotropheion (game farm), 186 Trajan, 335, 455f, 456, 557-558, 748
Theseus, 553 Trajan's Column, 44, 53, 258, 442-443, 455/
Thucydides, 241, 660, 675, 677 transcendentalist concept, 77
Tiberius, 538 transfer of technology, 75-76, 79-80, 825-826
tide-mills, 143-144 transhumance, 161, 183, 466-468
tie bar construction, 269-270, 269f transport
tie-beam truss roofs, 457-458, 458f, 459 by canal, 333-334
tiles of energy, 153
Corinthian style tiled roof, 239f of foods, 169
opaion roof tiles, 251 hodometers, 358, 766, 794-796, 796f
stone roof tiles, 247-248 paddlewheels, 358-359
terracotta roof tiles, 238-239 of quarried stone, 124-125
timber. See wood and timber by river, 141
timber truss roofs, 457-458, 458f, 459 by sea (See harbors; ships and navigation)
timekeeping transport on land by harnessed animals
overview, 740-741 choice of vehicle, 592
anaphoric water clock, 340 cost of, 601
Antikythera Mechanism, 341, 744-746, 745/ customs, context, and cost, 598-601
hours, 749-751 diolkos (portage road across Isthmus of
natural cycles of time, 741-742 Corinth), 592-594, 594!
parapegmata, 742-744 diversity of cartage and wheels, 596-598
portable dials, 751-752 harness and yoke, 590-592, 591f, 592f
sundial-calendar, 793, 794! historiography and methodology,
sundials, 746-748, 748f, 749/ 580--584
water clocks, 340, 752-754 mechanics of forces and potential energy,
Timgad, 341 584-588
Timosthenes of Rhodes, 643 packsaddle and portage, 589-590, 589t
Titus Tunnel (Turkey), 326, 327f single-yoke harnessing, 595-596, 596f, 597f
tombs, 28-29 universality of, 588-590
tongue and groove joint, 452f trapetum (rotary olive crusher), 380-381, 381f
tooling of glass, 534, 535/ travertine, 271
tools treading of grapes, 379-380
glassblowing, 536-538 treatises, systematic, 815
harvesting, 170 trenching technique (quarrying), 122-124, 123f,
iron and steel, introduction of, 98 124!
for quarrying and stonework, 122-123, Tres Minas mines, 114
125, 129 trihemiolia, 686
862 INDEX
Trimalchio, 797 Velitrae (Velletri), 555

Trimble, J., 405 velocity ratio, 339
triple pulley, 340, 344 veneers, stone, 131-132
tripod cauldrons, 425-426 veneers, wood, 460
Trireme Project, 612 Veneti, ships of, 623-624
trispastos, 344, 346 Veneto school of bridge architecture,
trolleys, 340 572-573
tubi fittili (vaulting tubes), 275-277, 276f Vercingetorix, 703
tubuli, 273 Vergil, Polydor, 65
Tuck, A., 473 Vergil, 166
Tunisia, 277 vertical-wheeled mills, 141-144, 145f, 146, 148f,
tunnel cisterns, 288 355-356
tunnels Vesuvian scoria, 268
early tunnels and the qanat, 162, 291-293, veterinary care, 183-185
292f, 322-324 Vettii House, Pompeii
engineering of, 319-322 frieze of erotes and psychai engaged in
important tunnels, 32ot-321t metalworking, 38, 44-45
mathematics and, 782-783 metalworking scene, 43
medieval, 333 plumbing, 304!
Nonius Datus and the Saldae aqueduct wall painting of Dadaelus and Icarus,
tunnel, 329-333 442, 443/
special achievements in ancient Vetustinus, Porcius, 331, 332
construction, 324-329 Via Aemilia, 558, 567-568
See also aqueducts; canals Via Amerina, 559
turbines, 154 Via Annia, 567
Turner, A. J., 740, 741 Via Appia, 553-558, 55ef, 560, 563
two-beam vertical loom, 471-472 Via Appia Traiana, 567, 572
Tylecote, R. F., 109 Via Aurelia, 559, 560, 569-570
Tyrrhenian coast, 211 Via Cassia, 559-560, 563
Via Clodia, 559
Uluburun shipwreck, 615, 616, 618, 621 Via dei Mulini bakery, 406-408
Umm el-Quttein, 706 Via Domitia, 574
uncia fractions, 780 Via Domitiana, 566-567
unit-fractions, 779 Via Egnatia, 574-575
Usher, A. P., 77 Via Flaminia, 558, 561, 563-564, 565f, 567, 569,
570, 571, 572f
Valle d'Aosta, 562, 571 Via Latina, 555, 560
vallicoltura, 206 Via Nomentana, 563
Varro, 192, 591, 599 Via Ostiense, 564-565
Varus, L. Betilienus, 297 Via Popilia, 558, 560, 567, 568
vase paintings. See Berlin Foundry Cup Via Postumia, 558, 567, 568, 573
vaults Via Praenestina, 564/
in cisterns, 288-289 Via Prenestina, 569
development of, 266-269 Via Salaria, 561-562, 565-566, 570
provincial construction techniques, Via Sarsinate, 568
274-278, 276f Via Tiburtina, 563, 564
reinforcement of, 269-273, 269f, 271f Via Valeria, 563
vegetable fibers, 468 Vickers, M., 499, 501
vegetables, 384 Viedebantt, 0., 760
vegetable tanning, 487 Vindolanda tannery, 488
See also tanning and leather viscosities of glass, 523-524, 523f, 524!
Vegetius, 692, 696, 697 Vita S. Romani abbatis, 143
Vegneron, P ., 179 viticulture, 163-164
INDEX 863
Vitruvius iron used in armor, 108, 109

on aqueducts, 298-299 resilience and battlefield superiority,
on balancing of weights, 586 692-693
on construction, 225-226, 228, 257, 262 siegework construction, 702-704, 705f
De architectura, 23-24, 225-226, 338, 808 significance in Roman society,
on force-pump, 807 707-710
on harbors, 643, 649 weapons and equipment, 695-702
hodometer, 358, 766, 794-796, 796f transport services, 599-600
on human capabilities, 3-4 warp-weighted loom, 471, 472f
on invention, 801 warships with rams, 620
lost works considered by, 39 washing floors (mining), 101, 102f
on mathematics, 782 Wasps (Aristophanes), 778
on mechanics, 338 water clocks, 340, 752-754, 791-792, 814
on mills, 147, 148 Water Clocks (Hero), 339
on plaster, 236 water-lifting devices
on practical and theoretical advances, 584 overview, 350-355
reputation of, 808 agriculture and, 162, 355
on solar heating, 138 animal power and, 140
systematic treatises and, 815 force pumps and, 353-355
on time, 753 mining and, 97
on woodcraft, 441, 454-455, 456 saqiya and, 351-353
Vitudurum tannery (Switzerland), 488, shadufs, 162, 309, 350-351
489f water screw and. See Archimedean
volcanic stone, 267-268 water screw
Volturnus bridge, 567 wheels, 42, 351-352, 351f
voussoirs, 275, 276f, 570, 571 water management and hydraulic
engineering
Wadi Amran (Israel) mines, 99 overview, 285
wagons. See carts and wagons; transport on agriculture and, 161-162
land by harnessed animals baths, 307-308
walled cities, 683-684, 706-707 cisterns, 287-290
walls, 236-238, 244, 273 diversion from public systems, 294
Walters, H.B., 498-499 fountains and nymphaea, 306-307
Ward, C. A., 615 Greek systems, 293-296
Ward-Perkins, John, 126, 129, 131 irrigation systems, 309-311
war engines, 346-350 Near Eastern aqueducts and systems,
warfare and fortification 290-293
artillery, 346-350, 349f, 698-699 purification, 290, 301
casting of armor, 424-425 quay water supply systems, 653
continuous development and, 810-811 reservoir cisterns, 304-305, 306f
Greek Roman aqueducts, 296-302, 308-309
fortifications and siege warfare, 682-686 sewers and drainage, 311-312
Hoplite warfare, 674-679, 675f, 676f urban distribution network, 302-304
Macedonian war machine, 679-682 water-lifting devices, 350-355
naval warfare, 686-689 wells, 285-287
sources of information, 673-674 water-mills
mass production of arms and armor, 401 cams and, 787-788
mathematics and, 783 cereal processing and, 376-378, 412
naval harbors, 654-656 invention of, 141-143
Roman mechanics, 143-150, 145f, 146f, 148f,
army structure and organization, 693-694 355-357
culture of, 691-692 stepwise change and, 814
fortifications, 705-707, 708!, 709f water organ (hydraulis), 360
864 INDEX
water power Wilson, R. J. A., 277

dough-kneading machines, 152, 358 Wilson Jones, M., 131
fractaria machina (breaking machine), winches, 340
357 Winder, N., 76
grain-processing pestles, 151 winding of glass, 534
interest in, 802 windlasses, 340, 345, 661
mining and, 113-114 windmills, 153
sawmills, 150-151, 357 wind power, 152-153
saws, water-wheel driven, 132 wine making. See grape processing and wine
transport by, 141 making
waterproofing, 237, 288 Winter, S., 74
water screw. See Archimedean water screw women, depictions of, 55
"water snail." See Archimedean water Wood, John T., 231
screw wood and timber
wax. See lost-wax method as building material, 248-251
wax tablets, 719, 72of centering structures in domes and vaults,
weapons. See warfare and fortification 266-267
weather forecasting, 746 as commodity, 251
weathering, 236, 237, 528 harbor framework, 645, 649, 650
weaving, 52f, 470-474, 814 mechanics and, 249
See also textile production pumps, 353-354
wedge catapults, 346 sawmills, 150-151, 357
wedge holes in quarrying, 123 tablets of, 719-720
wedge presses, 360, 388 trees favored, 448-450, 449t
weeding, 165 water transport of, 141
weft-faced, compound weave, 473, 474! woodworking and wood technology
weights and measures overview, 439-440
overview, 759-760 evidence and representations,
balancing of weights, 586 440-444
dry measures, 761-762, 762f furniture, 459-460
length and area, measures of, joinery, 450-453, 452f
766-768 in large building projects, 453-459, 457f, 458f
liquid measures, 762-763, 763f raw materials, Theophrastus and Pliny on,
metrology, 760-761 448-450
weights, 763-766, 764! tools, 443-447
weirs, 143 wooden artifacts, 444
wells, 285-287 wool, 466-468, 468f
What the Romans Did for Us (BBC), 65 workforce, 55-56, 100, 257
wheat, 166, 170 Works and Days (Hesiod), 163
wheel-made glass vessels, 532 workshops, 396, 399, 432-434
wheels worm drive, 340, 344
cross-bar, 597 See also Archimedean water screw
quality of, 597-598 worm gear, 787, 789
water-lifting, 42, 97, 140, 162, 351-352, 351f writing and book production
in water-mills, 145f, 146-147 alphabets, 717
Whitbread, I. K., 503 book production and trade,
White, K. D., 64, 70, 226 724-725
White, Lynn, Jr., 71 earliest texts, 717-718
whorls, 470 emergence of written tradition, 17
Wikander, 6., 63, 64, 71, 143, 150 invention of writing, 715-717
Williams, T. I., 72, 74-75 libraries, 725-726
Wilson, A., 64, 71, 83, 806 literacy, 733-736
INDEX 865
poetry vs. prose, 716, 716/ Xerxes, 334

record-keeping, 726-733 xyston (spear), 680
rolls and codices, 721-724
writing materials, 719-721 Yassi Ada shipwreck, 623
written sources. See sources and Yegill, F. K., 265
evidence yokes. See transport on land by harnessed
wrought iron, 108 animals
Xenophon Zaghouan-Carthage aqueduct, 301!, 303

on banausic arts, 5 zarte Rippenschalen, 540
on books, 724-725 Zerubavel, E., 740
on division oflabor, 394-395 zilu loom, 473
on fumes, 105 zinc, 110
on grape cultivation, 163-164 zinc oxide, 111-112
on hunting, 186 zodiacal signs, 747
on warfare, 675 Zosimos, 109

Oxford Handbook of Engineering and Technology in The Classical World - 2008

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Oxford Handbook of Engineering and Technology in The Classical World - 2008

Uploaded by

Copyright:

Available Formats

THE OXFORD HANDBOOK OF

JOHN PETER OLESON

Oxford University Press, Inc., publishes works that further

Oxford New York

CopYTight © 2008 by Oxford University Press, Inc.

Published by Oxford University Press, Inc.

Library of Congress Cataloging-in-Publication Data

Printed in the United States of America

Abbreviations and Spelling Norms, xvii

1. Ancient Written Sources for Engineering and Technology, 15

2. Representations of Technical Processes, 35

3. Historiography and Theoretical Approaches, 62

PART II PRIMARY, EXTRACTIVE TECHNOLOGIES

4. Mining and Metallurgy, 93

5. Quarrying and Stoneworking, 121

6. Sources of Energy and Exploitation of Power, 136

7. Greek and Roman Agriculture, 158

8. Animal Husbandry, Hunting, Fishing, and Fish Production, 175

PART III ENGINEERING AND COMPLEX MACHINES

9. Greek Engineering and Construction, 225

10. Roman Engineering and Construction, 256

11. Hydraulic Engineering and Water Supply, 285

12. Tunnels and Canals, 319

13. Machines in Greek and Roman Technology, 337

PART IV SECONDARY PROCESSES AND MANUFACTURING

14. Food Processing and Preparation, 369

15. Large-Scale Manufacturing, Standardization, and Trade, 393

16. Metalworking and Tools, 418

17. Woodworking, 439

18. Textile Production, 465

19. Tanning and Leather, 483

20. Ceramic Production, 496

21. Glass Production, 520

PART V TECHNOLOGIES OF MOVEMENT AND TRANSPORT

22. Land Transport, Part 1: Roads and Bridges, 551

23. Land Transport, Part 2: Riding, Harnesses, and Vehicles, 580

24. Sea Transport, Part 1: Ships and Navigation, 606

25. Sea Transport, Part 2: Harbors, 638

PART VI TECHNOLOGIES OF DEATH

26. Greek Warfare and Fortification, 673

27. Roman Warfare and Fortification, 691

PART VII TECHNOLOGIES OF THE MIND

28. Information Technologies: Writing, Book Production,

29. Timekeeping, 740

30. Technologies of Calculation, 759

31. Gadgets and Scientific Instruments, 785

32. Inventors, Invention, and Attitudes toward Innovation, 800

PART VIII ANCIENT TECHNOLOGIES IN THE MODERN WORLD

33. Expanding Ethnoarchaeology: Historical Evidence

Prof. David Blackman Dr. Gwyn Davies

Prof. Willy Clarysse Dr. Philip de Souza

Dr. Mark Jackson Prof. John Peter Oleson

Prof. Martin Kenneth Jones Prof. Lorenzo Quilici

Dr. Geoffrey Kron Prof. Georges Raepsaet

Prof. Michael Brian Schiffer

Dr. Andrew Meadows Dr. Katelijn Vandorp

Prof. Charlotte Wikander Prof. John Peter Wild

JOHN PETER OLESON

THE GREEK AND ROMAN VIEW

In addition to recognizing the age-old biological-perhaps even evolutionary-

discredited interpretation, and there is an extensive literature on both sides of the

ORGANIZATION OF THE MATERIAL

This book is not intended to be a compendium of all the technological procedures,