This Content Downloaded From 193.255.211.64 On Tue, 19 Dec 2023 10:29:44 +00:00

Roman Glassblowing in a Cultural Context
Author(s): E. Marianne Stern

Source: American Journal of Archaeology , Jul., 1999, Vol. 103, No. 3 (Jul., 1999), pp.
441-484
Published by: Archaeological Institute of America
Stable URL: https://www.jstor.org/stable/506970
REFERENCES
Linked references are available on JSTOR for this article:
https://www.jstor.org/stable/506970?seq=1&cid=pdf-
reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms
Archaeological Institute of America is collaborating with JSTOR to digitize, preserve and

extend access to American Journal of Archaeology
This content downloaded from

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
All use subject to https://about.jstor.org/terms
Roman Glassblowing in a Cultural Context
E. MARIANNE STERN
Abstract between glassblowers are explored. Diocletian's Pr

Edict (PE) provides important clues to start-up busin
Commercial glassblowing dates from the beginning
expenses.
of Augustus' rule. This paper focuses on the impact of Dominated by the division into two branch
glass commerce and trade were brisk, both within and
this novel technique on Roman society: the develop-
ment of the technique, the artisans who madeyond
the the borders of the empire. Glass vessels played a
nificant role in the daily life of all segments of socie
glass, the merchants who marketed it, and the custom-
The forms and function of glass vessels in the West and
ers who bought and used glass vessels. The perfection
of glassblowing is characterized by improvementsthe in
East are discussed separately.*
tools and equipment and the discovery that molten
WONDROUS GLASS
glass can be blown, a discovery that was closely con-
nected with recycling. The division into two separate
Petronius relates the following story abo
branches-one for making raw glass from primary in-
"There
gredients, the other for working the material and cre- was a craftsman once who made a g
ating glass objects-determined the structure that
of thedidn't break. So he got an audience
industry. Gender, names, and business relationships
emperor, taking his present with him. The
* I wish to thank the anonymous readers of AJA for sav- Study in the Design, Construction, and
ing me from an embarrassing oversight as well as for many Economics of Large-Scale Building Projects
useful comments and suggestions. J. Reynolds (Cambridge)in Imperial Rome (JRA Suppl. 25,
commented on my new suggestions for PE 16. 7-9. My re- Portsmouth 1997).
search benefited greatly from discussions with Heimo
Giacchero Do-
M. Giacchero, Edictum Diocletiani et Colle-
lenz (Magdalensberg), Andrea Rottloff (Augsburg), Lu- garum de pretiis rerum venalium (Genoa
cia Sagui (Rome), Mara Sternini (Rome), and Luigi and 1974).
Luisa Taborelli (Torino). T. Gagos and P. Goitein Heilporn (AnnA Mediterranean Society 1:
S.D. Goitein,
Arbor) assisted with locating papyrological sources. S.E. Economic Foundations (Berkeley 1967).
Knudsen (Toledo), E. Roffia (Milan), and A. Rottloff Isings C. Isings, Roman Glass from Dated Finds
provided information on vessels in their care and do- (Archaeologica Traiectina 9, Gronin-
nated photos for use. Additional thanks for photographs gen 1957).
goes to Lee Mooney (Toledo); Soprintendenza Archeo- van Lith and S.M.E. van Lith and K. Randsborg,
logica (Milan); and the Ernesto Wolf Collection (Stutt- Randsborg "Roman Glass in the West: A Social
gart and Paris). I am very grateful to many unnamed Study," Berichten van de Rijksdienst
colleagues who provided me with publications of their voor het Oudheidkundig Bodemonderzoek
own and others. Finally, my thanks for many years of 35 (1985) 413-32.
friendship go to Gladys D. Weinberg, to whom this arti- PE Diocletian's PE, quoted after Giacchero
cle is dedicated. unless otherwise noted.
All dates are A.D. unless otherwise noted. All references
Rfitti B. Riitti, Die rdmischen Gldser aus Augst
to pounds refer to Fustat or Roman pounds. Literature ci- und Kaiseraugst (Forschungen in Augst
tations in notes are arranged in chronological order of 13, Augst 1991).
publication. Stern 1994 E.M. Stern and B. Schlick-Nolte, Early
The following abbreviations are used: Glass of the Ancient World 1600 B.C. -
AnnAIHV Vols. 1-9 published in Liege: vols. 1- A.D. 50: Ernesto Wolf Collection (Ostfil-
3: Annales du ler/2l/31 congres des dern-Ruit 1994).
Journees internationales du Verre; vol.Stern 1995 E.M. Stern, Roman Mold-blown Glass: The
4: Annales du 4P Congres Interna- Toledo Museum ofArt (Rome 1995).
tional d 'Etude Historique du Verre, vol.Stern (in prep.) E.M. Stern, Roman, Byzantine and Early
5 ff: Annales du ... Congres de l'Associ- Medieval Glass: Ernesto Wolf Collection
ation Internationale pour l'Histoire du (in prep.).
Verre; vols. 10-12 publ. in Amster-Two Centuries M. Newby and K. Painter eds., Roman
dam; vols. 13- publ. in Lochem, Glass: Two Centuries of Art and Inven-
Netherlands.
tion (The Society of Antiquaries of
Cool and Price H.E.M. Cool and J. Price, Roman Vessel London, Occasional Paper 13, Lon-
Glass from Excavations in Colchester don 1991).
1971-85 (Colchester ArchaeologicalWeinberg G.D. Weinberg ed., Excavations atJalame:
Report 8, Colchester 1995). Site of a Glass Factory in Late Roman
DeLaine J. DeLaine, The Baths of Caracalla. A Palestine (Columbia 1988).
441
A merican Journal of Archaeology 103 (1999) 441-484

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
442 E. MARIANNE STERN [AJA 103
Caesar hand it back to

tinues him
to intrigue chemistsand dropped
and physicists.4 It is not
floor. The emperor couldn't have been more difficult to imagine how an eyewitness account of a
shaken. The man picked the bowl off the ground- workshop incident, as described above, could have
it had been dented like a bronze dish-took a ham- been embellished with physically impossible details
mer from his pocket and easily got the bowl as such goodas the glassblower pulling a little hammer out
as new."'1 of his pocket and hammering out the dent in front
Vitrum2 was the most versatile material known to of the emperor. After all, Pliny himself was of the
the Romans. Pliny (HN 36.195) calls it vitrumflexile opinion that glass could be hammered into relief
"flexible/malleable glass." Like modern plastics, like silver: argenti modo caelare, "(some glass is) chased
glass mimicked other materials in shape, color, and like silver" (HN36.193). This misconception reappears
design: neque est alia nunc sequacior materia "there is in Martial (Epigr. 14.94): audacis plebeia toreumata
no other material nowadays that is more pliable" vitri "plebeian chased cups of dreadnought glass" (au-
(Plin. HN 36.198). When glass is soft, it can be thor's emphasis). In reality, glass vessels decorated
stretched and expanded or made to compress. The with relief were produced by the technique known as
almost miraculous fluidity of the material when it is mold-blowing.5
hot, and the unlimited possibilities of transforming its
ROMAN GLASSBLOWING: THE PERFECTION
shape on the blowpipe from tube to sphere to square
OF A CRAFT
and back again to cylindrical, as well as stretching,
widening, and pinching the opening of the mouth The discovery that glass can be expanded
once the glass is separated from the pipe, are a never shaped by human breath revolutionized glass
ending source of fascination for glassblower and ing to such an extent that today "glassblowing
spectator alike. During blowing, glass seems to defy become the generic term for all glassworking, w
the laws of nature. Even if a piece falls off the blow- the glass is blown or formed by other techn
pipe or punty, it will not shatter! If the blower reat- The invention of the blowpipe meant that holl
taches it quickly and reheats, he or she can restore jects and vessels that previously required labor
the shape and complete the piece as planned. sive operations6 could now be made much m
Such a workshop accident might be at the root of quickly and that less glass per object was neces
Petronius' story about an unbreakable glass cup that Moreover, blowing permitted the production o
was dropped to the floor, got dented like a metal classes of items.7
vase, and was hammered back into shape. The story The beginning of commercial glassblowing coin-
is told by several ancient authors, sometimes with the cides roughly with the creation of the Roman empire.
comment that the glass had been tempered (tempera- Within half a century the art of glassblowing was
mentum) to make it malleable or flexible (Plin. HN transformed from a local Syro-Palestinian craft to an
36.195; Origenes 16.16.6).3 In spite of Pliny's dis- empire-wide enterprise. Around the mid-first century
missal of the story as a fantasy, the suggestion that glassblowing facilities began to spring up through-
the Romans actually invented unbreakable glass con- out the empire and beyond. Important first-century
I Fuit tamen faber qui fecit phialam vitream quae non frange- (University of Illinois Studies in Language and Literature
batur. Admissus ergo Caesarem est cum suo munere, deinde fecit 13, 1930) 110-12.
reporrigere Caesarem et illam in pavimentum proiecit. Caesar non 4 R.C.A. Rottlinder, "Naturwissenschaftliche Untersu-
pote validius quam expavit. At ille sustulit phialam de terra. Col- chungen zum r6mischen Glas in K61ln," K'lnJb 23 (1990)
lisa erat tamquam vasum aeneum. Deinde martiolum de sinu 563-82; G. Eggert,"Vitrum flexile als Rheinischer Boden-
protulit et phialam otio belle correxit (Sat. 51; all translations of fund," KolnJb 24 (1991) 287-96.
this work are those by J.P. Sullivan, The Satyricon, and the 5 Stern 1995, 68.
Fragments (New York 1965). All other translations, unless 6 On alternative techniques for shaping glass vessels:
otherwise noted, are taken from Loeb Classical editions.
Stern 1994 (with lit.); R. Lierke, Antike Glast6pferei (Sonder-
2The etymology of vitrum has given rise to many conjec- heft AntW, Mainz am Rhein 1999).
tures, but linguists agree that the origin of the word is not 7 Useful general introductions to Roman glass are given
Latin. E.R. Knauer, "III Glass and Pigment," MMAJ 28 by D.B. Harden, "Ancient Glass, II: Roman," AJ 126 (1969)
(1993) 28-34 suggests that vitrum is of Celtic derivation, 44-77;J. Price, "Glass," in D. Strong and D. Brown, Roman
perhaps from a root uei "bend, twist" (cf. English wire) as Crafts (London 1976) 111-25; Price, "Glass," in M. Henig
preserved in the Celtic word viriolae (Celtiberic viriae). ed., A Handbook of Roman Art (Oxford 1983) 205-19.
' M.L. Trowbridge, Philological Studies in Ancient Glass

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
1999] ROMAN GLASSBLOWING IN A CULTURAL CONTEXT 443
glassblowing sites include Avenches, corners Lyon, and

of the empire, and especially from the eastern
Saintes, to mention just a few.8 Several workshops
Mediterranean. Most ofap-
the areas where glassblowers
settled-Rome,
pear to have produced vessel glass and windowCampania,
panes, and the northern Adri-
atic coast-already
e.g., at Sentinum (Italy), Aix-en-Provence, had longstanding commercial
Bet She'an,
and-perhaps-at Sardis,9 but at other contactssites where
with Greece and the eastern Mediterra-
glass vessels were made there is no indication
nean.12 Finally,of the
glassblowing itself probably did not
production of flat glass.10 Glass beads
require and
a huge jewelry
investment in expensive new tools be-
were almost always made in workshops cause the
thatearliest vessels could all have been blown
special-
ized in this particular aspect of glass production.
with inexpensive blowpipes fashioned by the glass-
blowers
The earliest glassblowers plied their themselves.
craft on the
Syro-Palestinian coast and in Italy. North
Whereas Italy, Dal-
the initial discovery that glass can be in-
matia, and the Ticino Valley appear flated
to have been at
took place somewhere along the Syro-Palestinian
the forefront of glassblowing, but Campania
coast,13 was
where glassworking and glassmaking boasted
probably very active as well.11 According to Pliny
a centuries-old tradition, glassblowing was perfected
(HN 36.193), glassblowing (flatu figurare,
in Italy. The "shaping
range in quality and quantity of early
by breath") was formerly a specialty of glass
blown Sidon (mod-in Italy and western Europe
excavated
ern Saida, in southern Lebanon). far The unprece-
surpasses that from the eastern Mediterranean
dented speed with which the new with technique spreadin shapes, decorative tech-
regard to variation
throughout the empire was due to a niques, rangeand offunction.
factors: In Egypt, where the tradition
political, economical, and technical. Augustus'began
of glassworking rule in pharaonic times, artisans
ended a century of civil strife in Italy and created
were notoriously a adopt the new technique of
slow to
vast network of pacified provinces. blowing. SpeedyAlexandria's
commu-glass industry, renowned in
nication became possible from onethe end of the
Hellenistic em-appears to have suffered a
period,
pire to the other. Italy experienced an economic
marked decline in the first century. Pliny does not
boom that attracted artisans and merchants from all mention Alexandria when he discusses the glass cen-
8J. Morel et al., "Un atelier de verrier du milieu du 1erbridge, 1980) 91, 92. No workshop has been identified at
siecle apr. J.-C. ?i Avenches," ArchSchw 15 (1992) 2-17; H.Sardis, although cullet and wasters show that glassworkers
Amrein (forthcoming); M.-D. Nenna et al., "L'atelier de were active in the city. The similarity in fabric of vessels and
verrier de Lyon, du 1er siecle apres J.-C., et l'origine deswindow panes suggested to von Saldern that both were
verres 'romains'," Revue d'Archeometrie 21 (1997) 81-87. made in the same workshops, but it is also possible that dif-
Saintes (two sites): A. Hochuli-Gysel, "R6misches Glas ausferent workshops used raw glass made in one factory.
dem Siidwesten von Frankreich," AnnAIHV 12, 1991 (Am- 10 Morel et al. (supra n. 8) (Avenches); Weinberg
sterdam 1992) 79-88; B. Velde and A. Hochuli-Gysel, (Jalame).
"Correlations between Antimony, Manganese and Iron 1" Archaeological evidence for glassblowing: Campania:
Content in Gallo-Roman Glass," AnnAIHV 13, 1995 (Lochem E.M. Stern, "A Small Glass Bottle on Three Pinched Feet,"
1996) 185-91. Glassblowing in Britain began in the early in Festoen Opgedragen aan A.N. Zadoks-Josephus fitta bij haar
Flavian period: Cool and Price 226. On early glassblowingzeventigste verjaardag (Scripta Archaeologica Groningana 6,
in Spain: J. Price, "Glass Production in Southern Iberia inGroningen n.d., ca. 1976) 527-38; L.A. Scatozza H6richt,
the First and Second Centuries A.D.: A Survey of the Evi-"Syrian Elements among the Glass from Pompei and Her-
dence," JGS 29 (1987) 30-39. For a survey of early Imperialculaneum," in Two Centuries 76-85. Ticino: S. Biaggio
glassblowing facilities, see also D. Foy and G. Sennequier Simona, I vetri romani provenienti dalle terre dell' attuale Can-
eds., Ateliers de Verriers de l'antiquit da la peri ode pre-industrielle,tone Ticino (Locarno 1991); HelvArch 22 (1991) 78-143.
Association Franfaise pour l'Archeologie du Verre, Actes des 4e North Italy: Vetro e vetri, exhibition catalogue, Milan, Mu-
Rencontres, Rouen 24-25 Novembre 1989 (Rouen 1991);seo Archeologico, 1 November 1998-18 April 1999 (Milan
Stern 1995, 22. 1998) 13-146 (glass from recent excavations in Milano
9 L.Taborelli, "Elementi per l'individuazione di una offi-and vicinity); M. Calvi, I vetri romani del Museo di Aquileia
cina vetraria e della sua produzione a Sentinum," ArchCl (Aquileia 1968); Calvi, "Arte vetraria Ticinese e arte vetraria
32 (1980) 138-66 (workshop dated mid-first century); L. Aquileiese: raffronti e analogie," in HelvArch 22 (1991)
Rivet, "Un quartier artisanal d'epoque romaine A Aix-en-133-43. Dalmatia: Trasparenze imperiali Vetri romani dalla
Provence," RANarb 25 (1992) 325-96 (workshop datedCroazia, exhibition catalogue, Rome, Palazzo Barberini,
mid-second to early third century); Y. Gorin-Rosen, "Glass1998 (Milano 1997).
Workshop," in G. Mazor and R. Bar-Nathan, "The Bet 12 L. Taborelli, "Un antico forno vetrario ad Ancona,"
She'an Excavation Project 1992-1994," in Excavations and Picus 18 (1998) 219-24, esp. 224.
Surveys in Israel 17 (1998) 27-29, esp. 29; A. von Saldern, '1 Y. Israeli, "The Invention of Blowing," in Two Centuries
Ancient and Byzantine Glass from Sardis (SardisMon 6, Cam-46-55.

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
.c C
Fig. 1. Modern technique: Gathering hot glass from a pot

furnace. (Drawing D.F. Giberson)
ters of his day; nor is the city mentioned by the Flavian

poets for whom crystalla "crystal glasses" and calices
Fig. 2. Modern technique: Transferring a piece from the
blowpipe to the pontil; the blowpipe rests on the arms of
vitrei "glass cups" from the Nile were a literary topos.
the glassblower's bench. (Photo L. Dorfrnan)
Thebes appears to have been the main exporting
glass center of the first century.14 Political reality may
have reinforced Egypt's traditional tendency to look
Roman Improvements in Tools and Equipment
inward for cultural and artistic stimuli. Most descriptions of ancient glassblowing are
based on the assumption that the craft and its tools
Attracted by the magnet of good commissions avail-
able in Rome, many Sidonian glassblowers migratedare so simple that there was no development in the
to Italy and set up shop in Rome, Campania, andtechnique. As recently as 1987, D.B. Harden wrote:
"a glassblower blows and finishes a vessel using pro-
Aquileia. Numerous glass skyphos-handles stamped
with the names of Sidonians bear witness to their cesses that have never altered, at least in principle,
presence in Rome and other sites in the western since glassblowing originated."'16 His detailed de-
Mediterranean. Glassblowers named Ariston, Artas, scription of the technique is based exclusively on
Philippos, Neikoon, Eirenaios, and their colleagues 20th-century European practice: a gob of molten
may be credited with introducing the art of glass- glass is gathered on an iron blowpipe about 3-5 ft.
blowing to Rome and the West. Epigraphical and ar- long (fig. 1) and expanded by blowing; a solid iron
chaeological evidence indicates that a vicus vetrarius rod (also known as a punty or pontil) about 2.5-
"glassworkers quarter" existed in the vicinity of the 3.5 ft. long is affixed to the bottom of the vessel
Porta Capena at Rome.15 In Italy the Sidonians came with a wad of glass (fig. 2); the vessel is separated
into close contact with the strong utilitarian charac- from the blowpipe and held on the punty while the
ter of Roman technology. Referring to a period be- mouth of the vessel is finished (fig. 3). He con-
fore 7 B.C., Strabo (16.758) noted already that many cludes with the remark that the work is usually
improvements in the glass industry were taking place done with a team of four "with a master-blower in
in Rome "both with respect to the coloring of glass charge, who performs the main blowing and fash-
and to facilitate production techniques, for example ioning" seated on a wooden chair (also known as
for making colorless glass resembling rock crystal" glassblower's bench) with projecting arms on which
(krystallophanes). The interaction and exchange of he balances the blowpipe and pontil (fig. 4). Harden
ideas between Sidonians and Romans furnished the offers no historical or archaeological evidence for
any
impetus for the innovations and improvements of this description.
that
created the great Roman glass industry. No serious discussion of the early development of
1' E.M. Stern, "Hellenistic Glass from Kush," AnnAIHV Studio e la Conservazione del Mosaico, Rome, 5-7 December
8, 1979 (Liege 1981) 35-59, esp. 49. 1994 (Bordighera 1995) 447-66, esp. 455 n. 41.
15 Stern 1995, 68-69 (Sidonians in Rome). On the vicus 1' D.B. Harden, in Harden et al., Glass of the Caesars, ex-
vetrarius: M. Bacchelli et al., "Nuove scoperte sulla prove-hibition catalogue, The Corning Museum of Glass, British
nienza dei panelli in opus sectile vitreo della collezione Museum, R6misch-Germanisches Museum, Cologne (Milan
1987)
Gorga," in Atti del 2 Convegno dell' Associazione Italiana per lo 87, since then followed by many scholars.

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
Fig. 3. Modern technique: Widening the mouth of the ves

the pontil rests on the arms of the bench. (Photo L. Dorfma
lier cultures.'7 It is my contention,

dence from ancient vessels, historical research of
glassblowing tools, primitive glass furnaces, and my
own experience as a glassblower, that neither the
tools Harden mentions nor the manufacturing and
organizational practices he describes existed when
commercial glassblowing began. For example, the
glassblower's bench that features prominently in his
account-and in most modern reconstruction at-
tempts of ancient techniques-was not invented un-

til the 17th century.'8 Assuming its existence in an-
tiquity thoroughly confuses our understanding of
ancient glassworking processes. Another important
tool, now indispensable but apparently unknown to
Roman glassblowers, is the cross-cutting "scissor-type"
iron shears. Many idiosyncrasies of Roman glass ves
sels, such as the way handles are drawn out thin and
folded back and forth at the point of attachmen
rather than being cut cleanly, may be due to the lac
of this tool.19
My research leads to the conclusion that more
than 100 years of experiments, discoveries, inven-
tions, and improvements separate the first trial infl
Fig. 4. Modern glassblower's bench. The glassblower rolls

the blowpipe back and forth on the arms of the bench
while she makes a crease, also known
17 as ajack
So also mark,
T. Gain, at the in Glass. Present an
"Experiments
point where the glass will be separated from
Future," AnnAIHV 12,the
1991 pipe.
(Amsterdam 1993) 261-70,
(Photo L. Dorfman [1992]) esp. 262-63 "Analogies."
'8y. Ohira, "Lo scanno e la bardella a Murano e in altr
localita Europee," JGS 29 (1987) 72-80. For a historic
Roman glassblowing has been undertaken
evaluation in the
of some of the tools and equipment used by Ro
man glassblowers: Stern
past. The problem is that the archaeologist needs 1995,to19-29; more complete
Stern (in prep.).
know enough about glass to ask the right questions,
whereas the glassblower must be W,, Weinberg
aware 35,the
of 66; Stern (in prep.). Stern 1995, fig. 3
archae-
shows the modern scissor-type shears used for trimmin
the rim of a vessel.
ological reality and the technical limitations of ear-

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
Fig. 5. Core-forming furnace with vertical

Fig. 6. Reheating in ahea
clo
(Design and drawing D.F.
heatGiberson)
chamber. (Drawing
tion of
heat-softened glass20
instead from
ofjust full fle
heating
man glassblowing
inair the second half
surrounds the glass of
century. Most of the tools and for
important techniques blowing
for granted as integral
of heat to the allows craft the wergla
nace with a introduction
during this period. The horizontal heat chamber was a Romanof a n
invention.with
of glassworking furnace To judge from Roman clay oil lamps d
a horizontal h
ber, the constructionpicting
of this the piece iron
of equipment (fig. 7),22 it wa
blowpipe,
molten hot glass, andfirmlythein placepontil technique
in the third quarter of the first cen
finishing the rim of tury,a when
vessel,
the lamps werewere the
made. Their mo
findspots i
Asseria (Dalmatia) and Ferrara
tant steps in the development of are consistent with a
glassblow
(north?) Italian origin.
if not all of these techniques were perfecte
While it is impossible to date
If glassblowing these
began with impro
inexpensive clay blow-
pipes, such
precisely, the essentials canas appear
beto deduced
be depicted on the two
fro
sources. Before lamps, this could explain
the invention ofthe rapid spread of th
glassblowi
glassworking operations were
technique, because probably
the glassblowers p
themselves could
above a vertically rising
easily makeflame,
the blowpipes.23aTheset-up
hypothesis ofththe
ceramic blowpipe
the glassworker to work the(fig. 8) is attractive;
glass while iron tubes
he ar
softening (heating) not it (fig.
present 5).21 The
in the archaeological record of mod
the Au-
worker's furnace hasgustan
a closed, horizontal
period. A sturdy hea
iron tube was difficult t
ber, that is, a heat chamber into which one
make with ancient technology. Apart from the ex-
penses makes
pipe horizontally. This involved, the need to custom design and
it impossible
ulate the glass during reheats.
commission The
an iron blowpipe from advanta
a blacksmith un
20 Glass tubes pinched closed at the lower end and di officina

sub- vetraria: Alcune considerazioni sulla lavorazi-
one del vetro soffiato nell' antichita," JGS 29 (1987) 22-29.

sequently inflated through the other end are the earliest
evidence for the discovery that heat-softened glass can 23 be
The hypothesis of clay blowpipes, based on archaeo-
expanded by air. On these tubes, excavated in the wastelogical
ofconsiderations and modern workshop experience,
washalf
a glassworker's shop in Jerusalem, assigned to the first tested in practice and first suggested by E.M. Stern,
of the first century B.C.: Israeli (supra n. 13) 46-55."Art and Archaeology at the Toledo Museum of Art Glass
21 On glassworking facilities before the invention Crafts
of the Building," The Glass Art Society 1993Journal (Seattle
closed glassblower's furnace: Stern 1994, 24-25; Stern, 1993) 70-77, esp. 74-77. See also Stern 1994, 81-85, figs.
"Interaction between Glassworkers and Ceramists," in P. 156-71; Stern 1995, 39-43, fig. 20 left, figs. 28-32. My
McCray and W.D. Kingery eds., The Prehistory and History of thanks are due to Kathleen McCarthy for demonstrating
Glassmaking Technology (Ceramics and Civilization 8, Wester- techniques (figs. 2-4, 8, 18, 20, 21). Neither ceramic nor
ville 1998) 183-204, esp. 188, 203 (with lit.); D.F. Giberson, metal blowpipes have been identified in excavations of
A Glassblower's Companion (Warner 1998) 19, 47-50. sites predating the mid-first century.
22 D. Baldoni, "Una lucerna romana con raffigurazione

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
........
! .
Fig. 8. Blowing with a

of Art glass studio, 19
Fig. 7. Clay oil lamp ation

from of earlier
Asseria, luxu
depictin
blower at work before excavated in richca
a closed furnace, t
tesy Split Archaeological
weigh Museum,
495 g andno. 590
Fc
Such heavy vessels a
familiar with its construction
ern Mediterraneanmight bl
as a deterrent forfirst
many glassworkers
century. The
the new technique. heavy-weight vessels
By the year 70, however,
origin of iron blowp
the iron b
use in many if not The
all workshops, altho
earliest physica
apparently not usedpipefor purposes
comes in theoth
fo
blowing. The evidence is indirect.
on the interior of Thew
vessels are small bottles weighing
(mid-first 14-
century)
up to 166 g. In the second half
preserving of sha
the the
sizes,
large bottles, plates, andancinerary
indication t
urn
handles became common.
pipes differed. An urn in
Rema
1066 g without itsdiameter
lid, of 1.3 cm
a and inner diameter of 0.5 cm
cylindrical bot
Collectionweighs 572 at g
were excavated (figs.
a mid-second 9, 10
to mid-third-century
were to too
be heavy
glassworking site at Aix-en-Provence; latea
blown with cla
Roman
production required a pipe with suf
blowpipes were excavated in Spain and perhaps in
strength to carrythe
the
southern weight
Ukraine.5" Nothing ofis known the
about theg
fore length of the Roman iron blowpipe. It may have
the middle of the first century fo
tion of the iron blowpipe is consisten
been relatively short (3 ft. or even less) like the pipes
du V
24 Vetro e Vetri (supra n. 11) 34, no. 1, and figs. 6, 7, pl. 1I, s. aprasJ.-C.) et at Saintes (Mediolanum) (fin du ler
(jug from Dello); 64, 66, no. 3, and figs. 17, 18, pl.s. XVI apras J.-C.," AnnAIHV 14, 1998 (forthcoming); Velde
(jug from Valeggio Lomellina). See also 30, no. 2, andand figs.Hochuli-Gysel (supra n. 8) 186, fig. 2. Small fragments
2, 4, pl. II: flecked amphora (Isings Form 15) from of iron tubes were excavated at the site of a glassworking
Carpenedolo (Bs), tomb 1, inv. St. 78987, H 27 cm,furnace pre- at Aix-en-Provence: Rivet (supra n. 9) 356. On late
served weight 335 g. A 26.6 cm tall jug published by B.
Roman iron tubes, probably blowpipes:J. Lang and.J. Price,
Czurda-Ruth, Die rimischen Glser vom Madasberg (Kirnt- "Iron Tubes from a Late Roman Glassmaking Site at M6r-
ner Museumnsschriften 65, Klagenfurt 1979) 131, no. ida 1013, (Badajoz), in Spain,"JAS 2 (1975) 289-96; Stern 1995,
color pl. 15, weighs over 910 g. 41-42, ns. 26-29; M. Sternini, La fenice di sabbia. Storia e
25 H. Amrein and A. Hochuli-Gysel, "Le soufflage tecnoloqia
libre del vetro antico (Bari 1995) 83-85.
du verre dans les ateliers i Avenches (Aventicum) (milieu

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
Fig. 10. Cylindrical bottle. H. 14.7 cm; wt. 572 g. No pon-

til scar. Last quarter of first to first quarter of second cen-
tury. Probably made in Asia Minor. (Courtesy Ernesto
Fig. 9. Lidded cinerary Wolf Collection)
urn. H. 26 cm; wt. 1,066.0
out lid). No pontil scar. Late first to early second
Western Europe. (Courtesy Toledo Museum of
also known
1977.14. Purchased with fundsas pontil mark, on the
from thebottom of the
Libbey
ment, gift of Edward vessel. Shape, size, andLibbey.)
Drummond depth of the scar vary; it
can take the form of gashes or stand up as a ridge
still used in Hebron and
and make other primitive
the vessel wobbly. Today we usually re- fac
the eastern Mediterranean. move the scar by grinding and polishing. Fortu-
nately ancient glassblowers rarely removed the
The vessel's rim was always rough and irregular af-
ter separating the glass from the pipe, and finishing
scars. They provide important clues to the develop-
the rim presented a challenge. In the earliest yearsment of the technique, which did not become
of glassblowing, some mold-blown vessels appear to widely used until the last decennia of the first cen-
have been held with a clamp to fire-finish the rim; tury. Even then many artisans shied away from re-
the necks are distorted and/or show crimp marks.26 attaching the vessel, as evidenced by many large
Another possibility would have been to fire-finish thevessels with intricate rim folding but without a pon-
til scar (see figs. 9, 10).
vessel after it was annealed. The glass was so thin that
local reheating above a heat source can not be ex- The ring-shaped annular scar, the most common
cluded; this technique could explain why many early scar in East and West throughout antiquity (fig. 17),
closed vessels show no distortion or tool marks. The does not prove the use of a pontil rod (punty). The
rims of open vessels were usually left unworked if annular scar can be produced by a punty (fig. 18)
they were very thin (fig. 13), or else ground and pol-but it may also result from reattaching the vessel to
ished by cold working after the vessel was annealedthe collar of glass that remains on the blowpipe after
(fig. 14). Plates and wide bowls were provided with a the vessel is "cracked off." Because the modern glass-
blown foot (fig. 15).27 The artisan used a secondblower sits on a bench at a distance from the furnace
paraison (glass bubble), blown against the underside (figs. 2 and 4),28 reattaching the vessel is now usually
of the vessel, to hold it in the heat, flare it open, anddone with the help of an assistant, but the ancient
finish the rim (fig. 16). glassblower could easily perform this operation by
The pontil technique solved the problem of heat- him or herself. Seated on a stool directly in front of
finishing the rim. Instead of a second paraison the the furnace (fig. 19), the artisan cracked off the ves-
glassworker reattached the vessel with a wad of hotsel onto a flat working shelf and then (re)attached
glass that acted as a "glue." This glass leaves a scar, the tool to the bottom of the vessel, steadying it
26 Stern 1995, 20, 21 s.v. Clamp. JGS 19 (1977) 20, figs. 3 and 4.
27 Good examples in D.F. Grose, "Early Blown Glass," 28 Stern 1995, figs. 37-39.

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
Fig. 11. Jug from Dello (Bs), tomb 3; H. 25 cm; wt. 495 g. No pontil
Probably made in north Italy. (Courtesy Soprintendenza Archeol
Fig. 12. Jug from Valeggio Lomellina, Cascina Tessera, tomb 54bis;
pontil scar. Mid-first century. Probably made in north Italy. (Co
logica, Milan inv. St. 59234. Gift of Edward Drummond Libbey.)

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
against the furnace wall.29 The shelf, als

marver, is shown protruding from the
front of the glassblower on the Roman la
The earliest scars known to me are annular scars on
vessels excavated at Magdalensberg. They are very rare

but predate the year 45.3() A rich mining and industrial
center founded by Roman traders in the southern
Alps, Magdalensberg had close ties to north Italy, in
particular with large merchant families from Aqui-
leia,: which is also where most of the glass came
from. The Aquileian connection suggests a north Ital-
ian origin for the concept of reattaching the vessel.
The solid scar is positive proof for the use of a
punty. Depending on how the glassblower shapes the
glass the punty wad will be solid or hollow.2 A small
solid scar of 1 cm diameter, noted in a Flavian con- Fig. 14. Hofheim cup. H. 6.4 cm; Dm. rim 7.5 cm; wt.
144.8 g. No pontil scar. Mid-first century. Western Europe.
text at Augsburg, is proof that the technique of
(Courtesy Toledo Museum of Art, no. 1951.376. Purchased
transferring the vessel to a punty was developed be- with funds from the Libbey Endowment, gift of Edward
fore the end of the first century.'" The punty wad Drummond Libbey.)
can be made with reheated chunks, of glass or by
coating the tip of the rod with molten glass. Any or
The invention of the iron blowpipe and the con-
all of the techniques described above could have
struction of the closed furnace were necessary for
been used in antiquity.
gathering molten glass. The blowpipe was necessary
because iron was the only material that could resist
the temperature of molten glass (ca. 1050-11500C).
The closed furnace was the only way to achieve and
hold that temperature.
Molten Glass and Recycling

The discovery that molten glass could be blown
was nothing less than revolutionary. It was closely re-
lated to the equally momentous discovery that bro-
ken glass artifacts could be totally remelted, a break-
through that kindled a literary response in the
Fig. 13. Ribbed bowl (zarte Rippenschale). H. 5.5 cm; Dm. Flavian period (69-96) equal only to the excitement
rim 7.0 cm; wt. 72.4 g. No pontil scar. First half of first cen-
tury. Probably made in north Italy. (Courtesy Toledo Mu-
of Augustan poets about glassblowing. The concept
seum of Art, no. 1923.426. Gift of Edward Drummond of recycling, with all its social and economic manifes-
Libbey. ) tations, reverberated widely in Roman literature and
29 This technique, still practised in Hebron, is docu- curn," JGS 25 (1983) 79-86; G. Piccottini and H. Vetters,
mented in a video made at the Haaretz Museum, Ramat Fiihrer durch die Ausgrabungen auf dem Magdalensberg (Klagen-
Aviv, and shown in conjunction with the exhibit "Ancient furt 1990); G. Piccottini, "Gold und Kristall am Magdalens-
Glass from the Holy Land," Detroit Institute of Arts, 21 berg," Germania 72 (1994) 467-75; H. Dolenz, Eisenfunde
November 1998-31 January 1999. aus der Stadt auf dem Magdalensberg (KIirntner Museum-
:' Glass specialists tend to doubt the early date for aban- schriften 75, Klagenfurt 1998).
donment of Magdalensberg, but a terminus ad quem of 45 32 Glassblowers now make a hollow punty for a variety of
has been independently confirmed for all other categories reasons: to diminish the size of the scar, to facilitate sepa-
of objects from the site (personal communication H. Do- ration from the punty, to attach a punty to a piece with a
lenz, whom I would also like to thank for permission to pointed bottom, etc.
study the glass). I noted annular scars on Czurda-Ruth (su- 3 Personal communication A. Rottloff (1997). On the
pra n. 24) 65, no. 515, pl. 3: diameter of scar 1.8 cm, glass from Augsburg: A. Rottloff, "Zwei bedeutende Fund-
gashes; 87, no. 641, pl. 4, diam. of scar 2.0 cm, gashes. No. komplexe r6mischer Gliser aus Augusta Vindelicum-Augs-
780, pl. 5, has a solid, comma-shaped gash (pontil scar?) burg," AnnAIHV 13, 1995 (Lochem 1996) 163-74. Com-
with a diam. of ca. 1.5 cm. In addition, two unpublished pare eastern Mediterranean mold-blown bottles: Stern
base fragments excavated in the 1990s have annular scars. 1995, nos. 120-128, solid pontil scar illustrated on p. 191.
31 M.R. DeMaine, "Ancient Glass Distribution in Illyri-

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
Fig. 15. Plate with blown foot from Cosa, Atrium Publicum, no. C
cm. Before A.D. 40/45. Italian. (After D.F. Grose, "Early Blown Gla
dence,"JGS 19 [1977] 20, fig. 3c)
ken
caught the imagination of all classes of glass excavated
society. Recy- at Pompei may
cling became a poetical topos for Flavian
as fragments
poets such
collected for remelting
as Martial (Epigr. 1.41.3-5; 10.3.3-4), Statius
may date (Silv.
before 79.38 The realization
1.6.73-74), andJuvenal (Sat. 5.47-48).
be totally remelted led to the delib
Recycling had already been common before
of broken the and recycling b
vessels,
invention of blowing, but it was on a small
mous with scale and
remelting. At this time s
glass was
did not involve remelting. Glassworkers andalso going out of fashion;
artisans
in related fields reused fragmentsglass ofof precious
the last col- quarter of the first c
ored glass vessels and sandwich following
gold-glass. Bits of was either natur
centuries
broken glass, including bicoloredor colorless.
pieces, This could
were be remelted
in- without the risk
of becoming
cluded in early architectural mosaics.34 an indistinct
Curved glassmuddy color as would
vessel fragments formed the eyes inlaid
have been in bronze
the result of remelting mixed fragments
statues and mummies.35 Scraps and chips
of colored from mo-
glass.
saic canes were used as backing to mosaic glass
To blow molten tiles"6
glass it must be held at a constant
high temperature
and tesserae made from broken sandwich (ca. 1050-1150TC) for the dura-
gold-glass
vessels decorated early mosaic glass
tiondishes.37
of the work. Such a high temperature can be
Literary evidence suggests that achieved
the discovery that
only with sophisticated pyrotechnology.
broken glass can be totally remelted took
The furnace place
design in
is complicated by the fact that
the early Flavian period. Pliny the
does not
working portseem toModern furnaces have a
emits heat.
have been aware of this propertyshield
of glass.
or door He
whichwrote
the glassblower can open and
(HN 36.199): fragmenta teporata adglutinantur tan- and reheating. Perhaps
close quickly for gathering
tum, rursus tota fundi non queuntRoman furnacesfragments
"broken also had a door, but we do not
can only be made to stick to each other,
know. they of
The sophistication can
the Roman furnace can
not be totally remelted." Thus webest
can probably
be appreciated date
by comparing the quality of an-
the discovery to some time between
cient glassca. 70,
vessels when
to those made in primitive fur-
naces in theHistory,
Pliny had finished most of his Natural eastern Mediterranean.
and Like the Roman
Martial's Epigrams in A.D. 86. If aglassblower's
basket full offurnaces
furnace, bro- in Herat (Afghani-
3' F.B. Sear, Roman Wall and Vault Mosaics (Heidelberg 79 [Naples 1979] 256). Dio Cassius's statement (60.17.6)
1977) 40. that the Roman emperor Claudius made citizenship so
3?5 Personal observation, April 1983, Graeco-Roman Mu- widely available that one could obtain it "for a piece of
seum, Alexandria, inv. 14475/20818 and 20847. One of broken glass" cannot be used to date the beginning of re-
the inlaid glass eyes preserves cut grooves on the reverse. cycling to his rule (37-54). Dio wrote in the late second to
6 Stern 1994, 63. early third century when recycling was so common that
37 Stern 1994, 109-10, 112. "broken glass" had become an idiomatic expression for in-
: A. Pasqui, "La villa pompeiana della Pisanella presso dicating cheapness.
Boscoreale," Monumenti Lincei 7 (1897) 518 (quoted after 39 The reasons for the change to colorless glass were
J.-P. Morel, "La ceramica e il vetro," in F. Zevi ed., Pompeiprobably unrelated to remelting: Stern 1995, 186.

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
Fig. 16. Using a second bubble to flare open th

dish. (Drawing A. and L. Marty)
stan), Damascus, Hebron, and Cairo still fu

ing in the 1960s and 1970s remelted broke
but the quality of their output was poor. T
bubbly and full of striae and other impuriti
because they did not achieve the high
Fig. 18. Hollow punty
tures required for complete wad. (Photo L. Dorfman)
fusion. The tem
achieved by modern furnaces at Cairo and H
around 8000C. the beginning of glassworking in the second millen-
Not all Roman glassblowers worked with molten nium B.C. Individual chunks of glass may be heated
glass. Many glassworkers continued to pick up pre-even without a furnace, above a simple fire-pot. The
heated chunks of glass (fig. 20) and soften them attemperature needed for softening glass to a work-
the tip of the tool (fig. 21), as they had done since able state is not more than ca. 900-9500C, possibly
even lower.41 Very little glass is wasted in the process
because there is no crucible to which the glass can
adhere and less glass remains on the pipe. Molten
glass was not commonly used in bead making: bead
makers needed so many different colors in small
amounts that heating individual chunks of raw glass
or fragments of broken colored glass was always
more economical than melting glass in a crucible.42
The glassblower starting with a chunk of glass took
care not to get it too hot, because then the glass
would drip off the pipe and be wasted. The discovery
that drippy hot glass could actually be manipulated
and blown must have come as a shock-probably
equal in intensity to the disbelief of glass historians
when they were first confronted with the concept of
blowing a chunk of glass.43 Working with molten
glass required a total rethinking of techniques. The
Fig. 17. Ring-shaped pontil scar, gashes. Fragment of plate
of Isings Form 46a, from Augsburg. Augsburg, Glasschicht, glassblower gathering drippy hot glass has to cool
no. 3080 L. Flavian. (Photo A. Rottloff) the outer skin of the glass on his pipe before it can
40 Chafic Imam, "L'artisanat du verre en Syrie," AnnAIHV scribed by 6. Kifiiikerman, Glass Beads. Anatolian Glass Bead
3 (Liege 1964) 184-90; D. Charlesworth, "A Primitive Glass Making (Istanbul 1988).
Furnace in Cairo," JGS 9 (1967) 129-32; G. Lehrer, Hebron 41 The temperature depends on the composition of the
City of Glassmaking (Museum Haaretz, Tel-Aviv n.d., ca. glass. On the temperatures needed to soften ancient glass:
1970); M. Reut, "Le verre souffle d'Herat," StIr 2.1 (1973)Stern 1994, 21-23 and Stern 1995, 34-36. Giberson (su-
94-111; N.H. Henein, Le verre souffl&e en Egypte (Cairopra n. 21) 47 emphasizes the fact that "the use of hot glass
1974); L. Taborelli, "Un modo arcaico di produzione as an application does not in itself prove the use of a pot of
vetraria: Viaggio nel tempo al seguito di una fonte contem-hot glass."
poranea," in A. Avanzini ed., Profumi d'Arabia, Atti del con- 42 On ancient bead making techniques: Gam (supra n.
vegno Pisa 1995 (Saggi di Storia Antica 11, Rome 1997) 17).
149-66. A primitive furnace in Turkey, used exclusively to 43 E.M. Stern, "Glass Working before Glass Blowing,"
produce nonblown objects, such as beads and bangles, is de-AnnAIIHV12, 1991 (Amsterdam 1993) 21-31, esp. 22-23.

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
Fig. 19. Marver(ing surface) adjacent to furnace. Hebron glass

at work. (After G. Lehrer, Hebron City of Glassmaking [Tel Aviv n
Fig. 20. Picking up a preheated chunk of glass at the tip of a

blowpipe. The Toledo Museum of Art glass studio, 1992. (P
Dorfman)
Fig. 21. Heating a chunk of glass at the tip of a ceramic blowpipe

prior to blowing. The Toledo Museum of Art glass studio, 1992.
(Photo L. Dorfman)

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
be blown. The advantage

THE STRUCTURE of
OF THE working
GLASS INDUSTRY wit
glass is that cooling the skin of a hot gather
In antiquity, glassmaking and glasswork
time than heating a chunk; it is also easie
two separate crafts. This had been the
large vessels, because one
the very can dip
beginning in thethe
second pip
mille
times into the molten glass to gather new g
and remained customary throughout ant
the glass already on the pipe.
into the Middle Ages.48 The division in
Archaeological and literary evidence sugges
workshops for making the material an
in antiquity, working with
workshops molten
for working glass
and shaping the g
common in Italy and western Europe tha
portant consequences for the structure
eastern Mediterranean, where
man glass glassblowers
industry. It is generally accept
the fourth century still
primary workshopsup
picked individua
existed in the centur
known in Greek as harpazein
the invention of glassblowing. "chun
bolon A recent
ing."44 This may have been
Roman due
glass from to a areas
different problem
and dif
ing fuel in areas with little
turies or
concludes no
that thewood. Eve
chemical compo
cient furnaces stoked as diverse an assortment of uniform that the same source of sand must have
materials as traditional kilns in south Italy and Sicily
been used to make the glass.49 If this conclusion is
in the 1950s and 1960s-branches, roots, trimmings, correct, primary glassmaking was probably still con-
as well as straw, olive husks, shavings, and sawdust centrated within a very few areas.
(whatever happened to be available)45-it would Pliny (HN 36.190) mentions as primary glassmak-
still have been a daunting task to assemble suffi-
ing areas the Syro-Palestinian coast (confirmed for
cient fuel to work with molten glass. The calorificthe Byzantine period by excavations of huge glass
value per gram of these materials is constant,tanks in Israel),50 as well as Campania, Spain, and
about 4.5 Kcal for dry wood, less if the wood Gaul is (HN 36.194). Strabo (16.758) mentions Egypt
green, but the volume of fuel and the energy spent(confirmed by archaeological remains, although the
on gathering it depends on the species of tree anddating is less certain).51 The hypothesis that glass was
type of material (brush or heartwood).46 The amount
made only in a small number of primary workshops
of fuel needed for a primitive glassworker's furnace
in the Roman Imperial period is consistent with the
in modern Cairo is enlightening: it stoked one-third
fact that numerous remains of glass furnaces exca-
of a ton of wood per day, using mostly old railroad
vated in western Europe and Britain have all been
ties and wood from disassembled ships,47 that identified
is, as secondary workshops (where glass was
compact materials. shaped into objects).
Ancient artisans were slow to discover that glass Glassblowing, unlike other industries associated
can be totally remelted because they did not make with fire (pottery, bronze, and metalworking), did
their own glass. They bought it as solid ingots notor develop into a large-scale enterprise in antiq-
chunks and had little understanding of how the ma- uity and the early Middle Ages, in spite of a formi-
terial was made. The reason for this lay in the struc-
dable output. This was probably due entirely to
ture of the glass industry. physical restraints. Ancient depictions show that the
44 P Oxy. 50, no. 3536, line 3. On the literary and archae-

sels in Roman Cilicia," KolnJb 22 (1989) 121-28, esp. 121-
ological evidence for this technique: E.M. Stern, "A Fourth 23; Stern 1994, 19-27.
Century Factory for Gathering and Blowing Chunks of 49 Nenna et al. (supra n. 8) 81-87; D. Foy and M. Picon,
Glass?" JRA 5 (1992) 490-94; Stern 1994, 28; Stern 1995,
"Lingots de verre en Mediterranee occidentale (3e siecle
36-37; Stern, "Glassblowers in Greek Poetry," AJA 101 avantJ.-C.-5e siecle apresJ.-C.): Approvisionnement et mise
(1997) 342-43 (abstract). en oeuvre," AnnAIHV 14, 1998 (forthcoming). On glass-
45 R. Hampe and A. Winter, Bei Tdpfern und Zieglern in making: Stern 1995, 23-24. New insights can be expected
Siiditalien Sizilien und Griechenland (Mainz 1965) 196; Gib-
from M.-D. Nenna ed., Ateliers de verriers. Decouvertes recentes
erson (supra n. 21) 50 mentions straw, with documentary (Travaux de la Maison d'Orient, Lyon 1999, in press), es-
evidence for this practice. On stoking a mixture of locally
pecially the contributions by M.-D. Nenna, Y. Gorin-Rosen,
available materials including palm fronds, see also C.M. M. Picon, and M. Vichy (brought to my attention by M.-D.
Jackson et al., "Glassmaking at Tell el-Amarna: An Inte- Nenna).
grated Approach," JGS 40 (1998) 11-23, esp. 19-21. 50 Infra n. 261.
46 DeLaine 113. 51 Nenna et al. (supra n. 8) 85-86; M.-D. Nenna, "Ate-
47 Henein (supra n. 40) 10.
liers de production et sites de consommation en Egypte:
48 On the separation of the glass industry into two sepa-
Bilan de cinq annees de recherches 1993-1998," Ann-
rate branches: E.M. Stern, "The Production of Glass Ves- AIHV 14, 1998 (forthcoming).

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
Roman glassworking furnace was small52 (figs. 7,

22). The interior circumference was approximately
45-65 cm, a measurement confirmed by excavated
remains at Avenches and elsewhere.5 The furnace
had one working port. Because each glassblower
needed his or her own working port to reheat the
glass on the pipe while he or she was actually mak-
ing an object, the number of working ports dictated
how many glassblowers could work simultaneously
at one furnace.
Although excavations often find more than one

furnace in close proximity, it is not clear to what ex-
tent they operated simultaneously. For example, at
Avenches, where four furnaces were in use between
40 and 70, two overlap, suggesting rebuilding, a task
that must have often been necessary (annually?) Fig.
be- 22. Fragment of clay oil lamp from Carthage. Private
cause of heavy wear on the walls of the furnace.54 Collection. End of fourth to early fifth century.
Multiple furnaces may have been necessary in order
to blow on a daily basis. A primitive glassblowing the
fur-previous day; otherwise it can take up to three
nace at Herat could operate only every other day hours.
be- An additional two hours are reserved for
cause the furnace needed 24 hours to cool down af- melting the glass (i.e., the broken glass vessels).57 To
ter a day's work. When demand suddenly increased achieve the relatively pure glass seen in many Ro
the owner built a second furnace to be able to blow man objects, the time allotted to melting woul
every day.55 probably have been longer.
Other impediments to expanding into large scaleIn view of the particular exigencies of the metier,
it is unrealistic to assume that ancient glassblowing
operations included the size of the crucible contain-
was a large-scale industry comparable to the potter
ing the molten glass and the amount of space avail-
industry-with hundreds of employees or even
able for annealing, the slow cooling process neces-
slaves laboring in one establishment.58 Available py
sary to avoid creating stress in the glass that might
cause it to crack. Depending on the thickness of rotechnology
its made it impossible to enlarge the an-
walls, a glass vessel anneals in about 18-20 hours. cient
It glass furnace so that it could accommodat
cannot be safely removed from the space reserved more than one working port. In theory it would hav
for annealing until the glass reaches room tempera-been possible to create a series of small furnaces like
ture.56 Adjacent to one of the furnaces at Avenches the metal workshops at Magdalensberg, but eve
was a rectangular structure identified as an anneal-there inscriptions indicate that individual shops be
ing area. Perhaps the annealing chamber was tacked longed to individual owners and/or managers,59 an
onto the main structure and heated by the same fire no comparable agglomeration is known for Roman
as the furnace itself, as it is commonly organized glass
in furnaces. Large-scale glass industry began i
primitive furnaces. Rekindling and bringing the fur-15th-century Europe, when huge furnaces with mul
nace up to temperature in these primitive furnaces tiple working ports enabled numbers of glassblower
takes about two hours if the furnace has been used to work simultaneously. Earlier medieval illustration
52B. Caron and C. Lavoie, "Un fragment de lampe 55 Reut (supra n. 40) 107.
representant un four verrier," JGS 39 (1997) 197-98. 56 Annealing appears to have been a major problem in
53 Four first-century furnaces at Avenches, innerRoman diame-workshops. AtJalame, many fragments testify to a
ter 50-65 cm: Morel et al. (supra n. 8) 5-6, figs. 3-7 cidents
(with during the annealing process: Weinberg 35.
57 Reut (supra n. 40) 104.
refs. to similar size furnaces at Martigny and Kaiseraugst);
furnace at Aix-en-Provence, postdating 150, inner 58 F.K Kiechle, "Die Struktur der gewerblichen Glaser
diam.
45 cm, outer diam. ca. 75 cm: Rivet 1992 (supra n.zeugung 9) 349. in der frfihen Kaiserzeit," AnnAIHV 6, 1973 (Lieg
Although a fourth-century glassworking furnace excavated1974) 53-64. An AJA reviewer notes that large-scale enter
at Jalame was rectangular and covered a larger area, prisesca.are also uncharacteristic of the pot industry. Muc
2.40 X 3.60 m, there is no indication that it had more than and metalworking was done in small-scale units.
pottery
one working port: Weinberg 28-33. 59 Piccottini and Vetters (supra n. 31) 60-63; Dolen
54 Morel et al. (supra n. 8) 5-6, figs. 3, 4, 6, and 7.(supra n. 31) 15-37.

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
of glassblowing still show a solitary master at

nace.60 Literary evidence attests that this was
norm in the 12th century. Glassblowers took
blowing but did not blow simultaneously (se
The Glassblowers).
To gain a sense of the ancient glassblower
put, it is useful to look at the output of glass
working at primitive furnaces. The glassb
Herat produced about 100 vessels per day.61 T
nace in Cairo accommodated three pots of
glass and three working ports. Depending
size of the vessels and the complexity of thei
(handles, spouts), one day's production utiliz
two larger pots could be either 100 large
complicated vessels or 100 medium-sized ves
working port sufficed for the production
small vessels.62
It is not known how many days per year the an-

cient glassblower worked. Recent calculation of the
Roman working year suggests 220 days for a seven
month season, 290 days for a nine month season.63 If
the glassblower had to let the furnace cool down en-
tirely between blowing cycles (see above), he or she Fig. 23. Rectangular bottle from Linz, grave 99a. H. 28.5
could blow only every other day. For calculating the cm. Grave dated first half of second century. Made in Aqui-
leia. (Courtesy Ober6sterreichisches Landesmuseum)
ancient glassblower's minimum output, it is here as-
sumed that blowing was seasonal and took place on
110 days for a 12-month year of 220 days. Based on with each other, to external transactions such as buy-
the examples of the glassblowers in Herat and Cairo, ing raw glass and the marketing and selling of the
the output averaged 100 vessels per day or 11,000 per finished product.
year.64 This works out to 330,000 vessels in 30 years!
Although glassblowing is not only dangerous but also The Glassblowers
unhealthy because of poisonous fumes, such a long It has been suggested that the economy of the Ro-
period of activity would not have been impossible. man empire can be compared to the western Euro-
The tombstone of the opifex artis vitriae "glass artist"Ju- pean economy between 1400 and 1800.66 However,
lius Alexsander records his death in Lyon at the ven- this may not apply to glassmaking and glassworking,
erable age of 75 after 48 years of happy marriage.65 since Roman workshop practises differed consider-
The size and design of the ancient glassworking ably from those common in late medieval and early
furnace imposed certain physical restrictions. This industrial Europe.67 The limited number of people
affected many aspects of organization of the indus- who could blow at one furnace has been noted
try, from the number of people who could work si- above. Another difference regards the gender of th
multaneously in one shop and their relationships glassblowers. Whereas until very recently glassblow
60 R.J. Charleston, "Glass Furnaces through the Ages,"

imum output of 36,000 vessels per year for one furnace
JGS 20 (1978) 9-33, esp. 11, fig. 1 The earliest with
furnaces
two master blowers and one assistant blowing on al-
with multiple workports appear in illustrationsternate
dated to
days.
the late 15th century (Charleston 13, figs. 2, 3). 65 D. Foy and G. Sennequier, A travers le verre du moyen dge
61 Taborelli (supra n. 40) 159. d la renaissance, exhibition catalogue, Musee des Antiquites
62 Henein (supra n. 40) 38. It seems strange thatde Seine-Maritime
there ' Rouen (Rouen 1989), 61, 62, no. 8.
would have been no difference between the number of 66 W.V. Harris, "Between Archaic and Modern: Some
large and medium-size vessels. Production includedCurrent
some Problems in the History of the Roman Economy,"
in Harris
56 shapes with diameters varying between 55 and 4 cm; the (infra n. 84) 11-29, esp. 15.
height of bottles varied between 23 and 6 cm. The glass-
67 The main concern of this paper is vessel glass. Com-
blowers blew five days per week, from 10 a.m. to 10parative
p.m. evidence regarding architectural glass is noted
breaking only for meals. only where it is relevant to the topic. Later documents that
63 DeLaine 105-106; the figures of 220 and 290 includemight elucidate practices in Roman glassblowing are
one day off in eight. brought into the discussion where evidence for the Roman
64 Taborelli (supra n. 12) 223 with n. 6 calculates a period
max- is inconclusive.

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
. .... ..... .
?X.
?40
ov,
.. ........
x
.... ......
?gi
-X-R
-M
'MUM
? . 'a gg
........... . ......
......
........... ......
.......
... .. p
....
X,
. ............
... .........
.........
...
... ...........................
.
.... ... ..... ..

...................
..... ........ .. . .... . .. .... .....
.............
................ .... .?.'
...
... .....
.
.....
...
..................
..
.....
......
M ...
&.....
.... ...
... .........
.. ...
.. .. ....... ..... . . . . . 'D iv

..... .. .. .... . ..........
. .. ......... ...
... ... ....... ...
......
X&: ...... ..... ...... ....
g ..........
...................................
.......... ..................
...........
. ......... 'X'
....... ... ............ .... .. ......
. . . . . ........ . ........ .......
:-x,
....................... . . .
x
.. ..... ...
....... . ... ..... .....
..... ..... ... ....... ...... ... ........ ... ........

........... ..............
. . . . .... .....
...... ...... .. ... .........
....... .. .. ... .. ...

4. " .1Q, '.- V '? ..... .. . ....
.. .........
?'gq
-W.
...... ..... .
. ..... ....
X.
..........
YX. ..... ..........

;:OX ... ...... ....
sm
iZ . ..........
E
XI:
Ik'
W 'xg%. ............
. xj
.. ............... . . .....
X..
PAPA
*x
-A
??. 4g?-
4,-
..................
... .............. 7.:
.. . ..... ............ .. . ...
.... ...................
?:X
Fig. 24. Detail of underside of fig. 23, 25 X 13.4 cm. Signed SENTIASE/CUNDAFA/CITAQ[uileia]
VITR[earia]. (Courtesy Ober6sterreichisches Landesmuseum)
A number of glassblowers of the Roman period

are known by name, but little is known about their fi-
CITAQ V1T
CVNDAFA
nancial and social positions. The names of two glass-
blowers inscribed on a Roman pottery lamp (fig. 7)
ENTIA
are of interest because they suggest the glassblowers
were liberti, freedmen. Next to the glassblower blow-
ing a tall-necked bottle appears the name TRELLUS;
Fig. 25. Detail of fig. 23. Drawing of signature, beginning
his assistant's name is ATHENIO69 suggesting he (or
in bottom line, left.
his ancestor) hailed from Athens. Since the lamps
were made in Italy, we may assume the scene repre-
ing in Europe was an exclusively male occupation, sents a workshop in Italy.
the names of several women glassblowers active in In the eastern Mediterranean, one glassblower
the first century are known: Sentia Secunda had a stands out above all others: Ennion. His name has
shop in Aquileia (figs. 23-25); Neikais worked in the been identified as a Hellenized Semitic name.70
Greek-speaking eastern Mediterranean.68 Ennion specialized in mold-blown tablewares.
68 P. Karnitsch, "Der r6mische Urnenfriedhof," Jahrbuch mische Lampe mit Darstellung des Glasblasens," BJ 159
Stadt Linz 1952, 385-489, esp. 437-46, a discussion of two (1959) 149-51, who first published the lamp and studied
rectangular bottles, each with a different base molding, the original object. Like the lines representing flames at
from women's graves 99a and 99c. I would like to thank A. the top of the furnace the inscriptions were added before
Rottloff for sending a copy of the relevant pages. See also firing, when the clay was leather hard.
E.M. Stern, "Women Glassblowers in the Roman Empire," 70 G. Lehrer, Ennion, a First-century Glassmaker, exhibition
AJA 97 (1993) 338 (abstract); Stern 1995, 100-101, no. 5; catalogue, Haaretz Museum (Ramat Aviv 1979) 14; it is
Stern, "Neikais-A Woman Glassblower of the First Cen- also known from a third-century builder's inscription at
tury A.D.?" in G. Erath, M. Lehner, and G. Schwarz eds., Damascus: SEG II, 829. On Ennion see also Harden et al.
Komos Festschriftfiir Thuri Lorenz zum 65. Geburtstag (Vienna (supra n. 16) 164-66, nos. 86, 87; Y. Israeli, "Ennion in
1997) 129-32, pls. 27-28. Jerusalem," JGS 25 (1983) 65-69; Stern 1995, 69-73; D.P.
69 Foy and Sennequier (supra n. 65) 109-10, no. 44 Barag, "Phoenicia and Mould-blowing in the Early Roman
state that the names were scratched into the clay after fir- Period," AnnAIHV 13, 1995 (Amsterdam 1996) 77-92. En-
ing, but that is not what is said by M. Abramic, "Eine r6- nion's floruit was in the first half to mid-first century.

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
Glass vessels showing

may havehis been boundmold-blown
to their place of residence andsi
Greek, Ennion epoiese, "Ennion
trade by a hereditary tie: "We decree that made
the workers (
been found throughout the
in the crafts included in the Mediterran
list appended below, abid-
Israel to Spain as well
ing in theas on
individual the should
communities, north be exemptedco
Black Sea. The notably
from all kinds wideof compulsory distributio
services, as leisure
products is a measure
should be ofemployed his success,
for learning no
their crafts thor-
glassblower but alsooughlyas a that
in order businessman
they wish both to become more w
ther familiar withskilled
allthemselves
the and to train their sons."76
intricacies of
Before the Constantinian
tance trade or else knew how to edict, find
wares originat-
the
ner(s) for this venture.
ing in different parts of the empire were made us-
It is not easy to assess
ing specialized the
techniqueseconomic
such as snake-thread re
glassworkers from Roman tax
glass, sandwich gold-glass, and laws
flasks withinof
flasks. the
fourth centuries. If Such techniques are so complicated
a statement that they pre-
ascribed to
ius in the Historia Augusta can
suppose contacts between bethemselves.77
glassblowers trusted
Sev. 24.5), The hypothesis(vitriarii)
glassworkers that in the Roman empirewere
glass-
enough in the
early blowers
third century
moved about freely, setting upto be inc
shop where
other craftsmen who were
there was taxed
a market for in
their products, order
is supported
the emperor
Alexander
by epigraphical Severus'
evidence. The glassworkerbuildin
Julius Al-
Vitriarii exander, who
probably heredied at Lyon ca.
refers to 200, those
hailed from wh
chitectural glass (windows, mosaics)-in
Carthage.78 The fourth-century restriction of glass-
in demand for large public buildings-ra
blowers' movements is consistent with a new phe-
vessel glass. In the small Egyptian
nomenon: "international" town
fashions in glass are char- of
chus, for
example,acterized
6000 pounds
primarily of glass
by the imitation of elements that
total of 1320 talents, went into the "warm
a glassblower can duplicate just from seeing an ob-
thermon) of the city's
ject made public
elsewhere. bath.7' The d
of the walls and vaults of the baths of Caracalla inThe organization of the glass industry was proba-
Rome included 16,900 m2 of glass mosaic.72 Aurelian
bly not uniform throughout the Roman empire. In
(270-275) taxed glass and other commodities im-
the northwest provinces, for example, workshops
ported from Egypt into Rome (Vopiscus Vit. Aurel.were characteristically situated on the edge of town,
45.1), 73perhaps to protect local craftsmen. Italy and
an arrangement that was destined to become domi-
Rome were just beginning to recover from a century- nant in medieval Europe.79 To judge from the dis-
and-a-half long malaise that had plagued all seg-
covery of a Byzantine workshop in the center of town
ments of private life.74 at Bet She'an, glassblowing in the eastern Mediterra-
In August 337 diatretarii, perhaps "engravers and/nean was not always relegated to the outskirts.80
or cutters," and vitrarii, along with other groups of In the temperate climate of Europe, glassblowers
skilled laborers and artisans, professionals and semi-
would have had no difficulty working year-round. In
professionals, were exempted from personal taxes.
the eastern Mediterranean, where summers were
The law, which was probably issued by one of emperorharsh and hot, blowing may have been a seasonal oc-
Constantine's sons,75 remained in force into the sixth
cupation reserved for the winter. The summer
century. Its phrasing suggests that the glassworkers
months could have been occupied with marketing
71 P Oxy., vol. 45, no. 3265 (infra n. 130). civitates morantes ab universis muneribus vacare praecipimus, si
72 DeLaine 180-81.
quidem ediscendis artibus otium sit adcommodandum; quo magis
7" On authorship and credibility of events cited in the cupiant et ipsi peritiores fieri et suos filios erudire. Cod. Theod.
Hist. Aug.: KlPauly (Munich 1979) 2.1191-93. 13.4.2; Cod. lust. 10.66.1; author's emphasis. On the edict:
74 C. Panella, "Le merci: Produzioni, itinerari, destini," Trowbridge (supra n. 3) 119 with n. 34; E.M. Stern, Ancient
in A. Giardina ed., Societi romana ed impero tardo antico 3: Le Glass at the Fondation Custodia (Collection Frits Lugt) Paris (Ar-
merci. Gli insediamenti (Bari 1986) 431-59; infra n. 164. chaeologica Traiectina 12, Groningen 1977) 156-58.
75 The edict was issued by Constantine II, according to 77 Stern (in prep.). On migrating Syrian glassworkers,
O. Seeck, Regesten der Kaiser und Pipstefiir dieJahre 311 bis infra n. 263 (snake-thread), n. 264 (flask within flask).
476 n.Chr. (Stuttgart 1919) 185. On the receiver, Valerius 78 Supra n. 65.
Maximus, Praetorian Prefect of Dalmatius Caesar (?), see 79 Stern (supra n. 76) 152-55. For recent publications
A.H.M. Jones, The Prosopography of the Later Roman Empire 1: of Roman workshops on the outskirts of towns: Riutti 150-
260-395 (Cambridge 1971) 590-91. 52 (Augst); Rottloff (supra n. 33) 170 (Augsburg).
76 Artifices artium brevi subdito comprehensarum per singulas so0 Gorin-Rosen 1998 (supra n. 9) 27-29.

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
blown
and selling the glass, ordering new signatures and
supplies, on certain
re- glass vessels appear to
indicate that
building the furnace, or by activity individualother
in some glassworkers entered freely
field. Such a seasonal division of work has been into business partnerships. The group of lason,
noted for Cypriot potters. Meges, and Neikais, all three of whom specialized in
mold-blown bulbous beakers of one specific type
Business Relationships and blown in molds of exactly the same technical con-
Business relationships appear to have differedstruction,
as a is the most obvious example. The Sidonians
who migrated to Italy and made skyphoi with stamped
result of the very different conditions in the eastern
Mediterranean, Africa, and Europe during the for-handles are another group which may have formed
mation of the empire. In the west, where the Ro-
partnerships.85 In third-century Egypt glassworkers
organized locally in guilds.86
mans generally had a higher level of organizational,
technical, and business skills than the populationsOnein recent hypothesis is the possibility of the ex-
the areas they annexed and converted into change prov- of molds for mold blowing, implying "a se-
inces, powerful senatorial families with freedmen ries of local workshops, perhaps trading actual molds
and slaves as business managers and agents domi- among themselves."87 Originally suggested to replace
nated production lines in several industries. A good
Harden's hypothesis that the glassblower Ennion re-
example is the metal industry at Magdalensberg.81 It in midcareer from the Syro-Palestinian coast
located
is conceivable that the production of architectural
to north Italy,88 the concept of mold exchange may
bethe
glass was organized along lines similar to those in comparable to the production of signed clay
metal and clay industries (bricks and tiles), but lamps
ar- (Firmalampen) in the western part of the Ro-
chaeological evidence is lacking. In this connection,
man empire.89 There is, however, no evidence that
it is unfortunate that we do not know whether the branch workshops played an active role in the east-
base moldings on the underside of prismatic glass
ern Mediterranean, not even in the pottery indus-
bottles refer to the makers of the bottles or to those
try.90 Elsewhere I have argued that the distribution
who produced their contents.82 If the moldings referpattern of Ennion's products and other early eastern
to the glass workshop, the distribution pattern of Mediterranean mold-blown wares is indicative of
bottles carrying the name of C. Salvius Gratus mightlong distance trade.91 Ennion's enormous output-
be consistent with the hypothesis of branch work- over 30 vessels preserving his mold-blown signature
are known-may very reasonably be the production
shops, one in north Italy, the other in Augsburg, both
active in the late second to early third century.83 of one artist (see above).
Most research on business practices has focused Glassblowing and pot making are similar in that
both industries produced household containers and
on conditions in Italy and the western part of the Ro-
man empire.84 The following observations and re-
tablewares. Yet it is not clear to what extent business
marks concentrate on the eastern Mediterranean. practices of Roman glassblowers compare with those
In the eastern provinces glassworking already of Roman potters. Several contracts from Roman
boasted an established tradition of business prac- Egypt provide details regarding the lease of facilities
tices long before the Romans arrived. The mold-and equipment to potters. The exact juridical inter-
81 Supra ns. 31, 59. domesticum (JRA Suppl. 6, Ann Arbor 1993);J.-J. Aubert,
82 On the problems of names, infra pp. 467- 69. Business Managers in Ancient Rome (Leiden 1994).
8 A survey of north Italian findspots casts doubt on the 85 Stern 1995, 68-69 and 73-74.
hypothesis that the bottles marked by C. Salvius Gratus 86 P Oxy., vol. 45, no. 3265 and vol. 54, no. 3742, both
were made in Aquileia: E. Roffia, "Osservazioni su alcune quoted in full infra pp. 464, 465. In Rome some of the colle-
bottiglie in vetro con marchio di C. Salvius Gratus," Rivista gia may have acted as guilds for the benefit of their mem-
Archeologica dell'Antica Provincia eDiocesi di Como 163 (1981) bers: DeLaine 204.
115-29, pls. I-V; G.M. Facchini, "La circolazione dei vetri 87 M. McClellan, "Recent Finds from Greece of First-
romani nella Cisalpina: il ruolo di Calvatone-Bedriacum," Century A.D. Mold-Blown Glass," JGS 25 (1983) 71-78;
Quaderni del Giornale Economico Suppl. 5/96 (1996) 53-58. Cool and Price 43, 227.
No workshop has been located in north Italy. On the possi- 88 D.B. Harden, "Romano-Syrian Glasses with Mould-
bility of a manufacturing center at Augsburg, documented blown Inscriptions,"JRS 25 (1935) 163-86, esp. 164-65.
by deformed fragments of Salvius Gratus' bottles and 89 W.V. Harris, "Roman Terracotta Lamps: The Organi-
waste: Rottloff (supra n. 33) 170-72. More on Salvius Gra- zation of an Industry,"JRS 70 (1980) 126-45.
tus, infra n. 152.
"o Aubert (supra n. 84) 302 cites only one stamp, from
84 W.V. Harris ed., The Inscribed Economy. Production and Asia Minor.
Distribution in the Roman Empire in the Light ofinstrumentum 91 Stern 1995, 69-72.

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
pretation of these made on Fridays belonged

contracts is to not
the Jews alwa
and those
they do show an interesting made on Saturdays accruedvarietyto the Moslims." in
concerning the lease From theseof workshop
agreements and
it is clear that the glass-
space, equipment,blowers and even raw
were owner-entrepreneurs, mate
even if they could
clay and firewood. In not provide
one cashcontract,
capital. The contract featuring
dated Abu
lessee named Aurelius Paesis undertook to deliver a Sa'd is the only one to mention the hiring of a skilled
set number of wine amphoras at a set time for which craftsman. The use of hired labor may have been
the landlord would pay a fixed price in money and more common in primary glassmaking. A fragmen-
kind. The duration of the contract was for two years.
tary contract dated to the spring of 1057 describes
The potter brought his own assistants and had total an agreement between two partners and a laborer in
control of the leased premises. Apparently, he wasCairo,
an "all three being indiscriminately termed zajjaj,
independent craftsman, managing his own business glassmakers.... The laborer undertook to work on
and entering the contract in his own right.92 No thesuchmelting furnace for the duration of a year. ..
contracts for glassworking have been identified,His so remuneration would consist in 5 dirhems and
we do not know if they existed or if they resembled lunch worth 1 dirhem on any day he worked. He
potters' contracts.93 would not work for anyone else during the period of
In the absence of evidence for the Roman period,
the contract."'7 The laborer's wages were the common
a hoard of documents detailing economic and social
wage of the time, paid at the end of each day. The
conditions in the late 10th to early 13th centuries
addition of a meal appears to be rooted in Roman
sheds light on the types of contracts and agreements
practice. Goitein suggests that the stipulation that
glassblowers made among themselves. Knownthe as man was not to work for anyone else indicated a
the Geniza documents, they were recovered from tight
the labor market. Another explanation might be
Geniza (religious archive) of a synagogue in Cairo.94
the owners' fear of disclosure of glassmaking recipes.
In one document two glassblowers agree to blow
The agreements described in the Geniza docu-
glass together for a period of six-and-a-half months:
ments show that great differences existed between
a glassblower named Abu Sa'd provides 20 dinars, the income (and presumably the social status) of in-
while his partner does not contribute any capitaldividual
but glassblowers. The glassblower Abu Sa'd men-
receives a personal loan (from Abu Sa'd) of 10 tioned
di- in the first of the agreements cited above was
nars, on condition that Abu Sa'd will work only two
very successful. Two years later he contributed 400
"turns" a week, and his partner the rest.95 In a sec-
dinars in cash to a partnership for making wine. At
ond contract two other glassblowers agree to workthe other extreme stands his partner who needed a
together for the duration of one year: one partner
personal loan of 10 dinars to pay for his own contri-
bution. He offered the title deeds to his house as
provides 199 dinars, the other only 6, contributing
also a small quantity (10 qintars, i.e., 1000 Fustat
security, but he entered the contract as an owner-
pounds or ca. 450 kg) of raw material.96 They invested
partner. The enterprise itself must have been mod-
the 199 dinars in raw glass: 105 qintars of locally made
est, since they had no more than 30 dinars to begin
glass and 108 qintars of imported red glass. A thirdwith. The two glassblowers mentioned in the second
contract mentions "a partnership in the manufactur- contract began with 205 dinars (199 + 6) and ca. 450
ing of glass vessels, which was done in a store of kg
cop-
(10 qintars) of raw glass.
perware" (!) (in Damascus). A fourth document de-
scribes a partnership between Jewish and Muslim
Diocletian's Price Edict
craftsmen (silversmithing or glasswork): their tools
The single most important document recording
were "common property, on condition that profitsthe prices of glass in the Roman empire is Dio-
92 Aubert (supra n. 84) 253-55. 96 "The basic weight was the dirhem (not to be confused
93 The Egyptian papyri are currently being analyzed for with the coin bearing the same name) weighing 3.125 g.
references to the glass industry; work in progress by Tra- The common pound of Fustat consisted of 144 dirhems
ianos Gagos and myself. (or 12 ounces of 12 dirhems), approximately 450 g, com-
94 Goitein 87-88; the contracts cited pp. 363-65, nos. 9, parable to the present day U.S.A. pound... One hundred
19, 8, and 17 respectively. pounds made a qintar" (Goitein 360). The word was de-
95 The Muslim gold coin dinarweighed 4.233 g. Two di- rived from kentenarion "one hundred pounds," which was
nars "were regarded as monthly income sufficient for a the basic weight for glass in Roman Egypt. See infra 464-66
lower middle class family ... The dirhem was a coin of low with discussion P Oxy. vol. 45, no. 3265 and P Oxy. vol. 54,
silver content." Approximately 36-40 dirhems had the no. 3742.
value of one dinar (Goitein 359-60). 97 Goitein 94. On the value of the dirhem, supra n. 95.

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
cletian's Price Edict (hereafter PE). Issued in No- called "natural colored glass" in modern glass litera-
vember/December 301, presumably at Alexandria ture. This is also indicated by the word vir<i>dis,
but almost certainly prepared while the emperor was "greenish."'05 The more expensive Alexandrian glass
residing in Antioch, the PE lists prices for six types was colorless, i.e., intentionally decolorized. Barag
of glass.98 shows that the geographic designations date from a
The declared purpose of the PE was to check in- period long before the PE, presumably from the first
flation. The prices mentioned in the PE were maxi- century, but certainly in the case of Judaean glass,
mum prices, not fixed prices. The preamble specifi- before the year 135; in that year Hadrian officially
cally encourages lower prices in places where goods abolished the province of Judaea. As a penalty for
were abundant. Originally composed to aid soldiers, the Bar Kochba revolt of 132-135 he renamed the
the PE aimed to benefit the entire population of the province Syria Palestina. The use of a geographical
Roman empire. The prices would have been particu- designation that had ceased to be meaningful at the
larly beneficial to groups living on a fixed income. time of the PE is known also from other goods.106
State purchases for army and imperial court supplies The prices for glass vary according to the stages of
seem to have been made at the listed prices.99 production. One may compare the prices for pairs
Because the PE was prepared in Antioch, the of wooden wagon parts qualified by the terms fabrica-
choice of items to be included in the list is thought tum and infabricatum.107 Raw glass and vessel glass
to reflect, to some extent, Antiochene conditions. were both sold by the pound, i.e., the Roman pound
The Antiochene connection is of interest for glass of 327.45 g.108 The PE uses two words to indicate the
studies because the city is not actually mentioned as weight: libra (Gk. litra), referring to raw glass, and
a glass center in the Roman period,100 although the pondo to vessel glass. The choice of the word proba-
general region was renowned for its glass, and glass bly reflects the reality of transactions: raw glass was
was certainly being worked (or made?) in Antioch in sold in multiples of one pound (libra), whereas the
the 12th century.101 A new interpretation of the merchant needed to use a balance with a weight
prices for architectural glass (see below) is consistent (pondus) to calculate the price of a glass vessel. The
with the observation that the PE offers no positive pricing of glass vessels by weight rather than per piece
proof "that western or even non-Antiochene condi- may well have been common practice in the late
tions were taken into account."'02 The PE's maxi- Roman empire, especially if the addition of levis
mum prices for glass are outlined in Table 1.103 "smooth," in lines 3 and 4 indicates that the vessels
Barag'04 argues that Alexandrian and Judaean dowere not decorated by engraving or otherwise.109
not refer to the origin of the glass but to generic
Selling glass vessels by weight can be compared to
types (qualities) of glass. He convincingly identifies
basing the price of a pottery container on its capac-
Judaean glass with common bluish green glass, often ity."10 Both types of prices are objective: they reflect
98 Giacchero. The Latin version has been reedited by

tant Epigraphic Discoveries Related to the History of
Glassmaking in the Roman Period," AnnAIHV 10, 1985
J.M. Reynolds in C. Rouech6, Aphrodisias in Late Antiquity
(JRS Monograph 5, 1989) 265-318. For a recent discus- (Amsterdam 1987) 109-16, esp. 113-16. The Greek text is
sion of date, place of issue, intended audience, choicefrom
of Giacchero 171.
items to be included, prices, and general background of the104 Barag (supra n. 103) 113-14.
PE: S. Corcoran, The Empire of the Tetrarchs. Imperial Pro-105 Erim and Reynolds (supra n. 103) published the text
nouncements and Government A.D. 284-324 (Oxford 1996) as S<ub>VIR<i>DIS and commented "probably a mis-
205-33, reviewed by W. Turpin,JRA 11 (1998) 652-56. take for viridis or subviridis" (103).
99 Turpin (supra n. 98) 655 with n. 11. 106 Corcoran (supra n. 98) 222 quotes as an example
100 On the glassblower Paulinos who identified himself
"dalmatics."
as an Antiochean, infra n. 155. 107 PE 15.1-29; see also Corcoran (supra n. 98) 225.
101 Benjamin of Tudela: "Ten Jews dwell here, engaged10s Giacchero 117. D.K. Charlesworth, in Erim and Reyn-
in glassmaking .. .," cited by C.J. Lamm, Mittelalterliche olds (supra n. 103) 109 saw "no obvious rationale" in the
Gliiser und Steinarbeiten aus dem Nahen Osten 1: Text (Berlin
use of different words to indicate the same weight of one
1930) 491, no. 46. pound.
102 Corcoran (supra n. 98) 220-23; the citation is from109 So already Charlesworth (supra n. 108) 109. One
p. 223. may compare the use of Greek leia for smooth-walled
103 PE 16.1-6; first published by K.T. Erim and J. Rey- metal and glass vessels: E.M. Stern, "Glass in Athenian
nolds, "The Aphrodisias Copy of Diocletian's PE on Maxi- Temple Treasures," JGS 41 (1999).
mum Prices," JRS 63 (1973) 99-110, esp. 103, 108, 109, 110 PE 15.101; Erim and Reynolds (supra n. 103) 108,
column III, lines 35-39. The Latin text here follows Rey- commentary to line 33 cetera vascula pro ratione [capacita-
nolds (supra n. 98) 281. See also D. Barag, "Recent Impor- tis?] "other clay vessels according to their capacity."

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
Table 1. Diocletian's Price Edict (PE) 16.1-
16.1 DE VITRO 16, 1 GREEK TEXT
la Vitri Alexandrini libra una [X] viginti quattuor lcc
2 [Vitri I]udaicis virdis libra una [X t]redecim 2
3 [Vitri Ale]xandrini in calicibus et 3
vasis levibus in pondo uno X triginta
4 Vitri Iudaici in calicibus et vasis 4
levibus in po(ndo) unum X viginti [i
5 Speclaris optimi libra una X octo 5 [- - - - - --- - k(TQl)] ' X I
6 Secundi libra una X sex 6 [- -- - - -- ] k(ca) a' . 9 g'
7 [DE----?---]BUS 7 [FIEPI------]
7a [-- - - lib(ra)] una X qua[drag
8 [ .. .c.10... coloris? li] b (ra) una X tr[
9 [ ... c.10 ... coloris? 1] ib (ra) una X vi
FOR GLASS
la Alexandrian glass one pound denarii 24

2 Judaean greenish glass one pound denarii 13
3 Alexandrian glass cups and
smooth vessels one pound denarii 30
4 Judaean glass cups and smooth
vessels one pound denarii 20
5 Window glass, best (quality) one
pound denarii 8
6 [Window glass] second (quality)
one pound denarii 6
See p. 466 for translation and discussion of lines 7-9
the amount of raw material and labor(fig.

pound in making the
26); late Roman east
object rather than its degree of aesthetic
spherical perfection.
bottles of comparable c
The rationale is utilitarian: functionality
proximately is the
one deter-
to two Roman p
mining factor. The same rationaleuralstill prevails in Af-
bluish-green glass, the bot
ghanistan.'11 It is not clear howonegenerally glass ves-
pound would have cost app
sels were sold by weight in antiquity"112
narii, the or when this
two-pounders over 40
practice began. Luxury glass 20
and decorated
times vessels
as much as a comparable
were always sold per piece."l1 The PE (7.1-23) lists maximum
There are two ways to evaluate whether the
occupations. Theprice
minimum daily
of vessel glass was expensive labor
or inexpensive. One is
was 25 denarii plus meals
by comparing the prices of different
narii, a types of 7,700
total of goods;denarii ove
the other is by trying to relateor
the
10,150 denariicost
price to the overof
a year of 2
living and prevailing wages. laborers earned 50-60 denarii pl
In the PE, the price of a pottery container
age weight of with a Roman glas
a late
capacity of two sextarii (1.094 mately
It) was two denarii.114
150-350 g. At the PE's p
Although a capacity of two sextarii was per
denarii small for a depending
pound, clay on
vessel, it was relatively large for
wasa made
glass vessel. Barrel-
of Alexandrian colorles
shaped "Frontinus" bottles with a capacity
green, of two or equivalent
25 denarii-the
one-and-a-half sextarii had an of
average weight laborer-would
an unskilled of one hav
"t Reut (supra n. 40) 107; Taborelli quier, Verrerie1997 (supra

d'epoque romaine. n. 40)
Musie des Antiquitis de Rouen,
160.
Collections des musees departementaux de Seine Mari-
12 Barag (supra n. 103) 116 quotes several examples time (Rouen 1985) 169 notes that Frontinus bottles were
dating from the 12th to the 20th century of glass sold by made with standard sizes of 3 sextarii (1.62 It); 2 sextarii
weight.
(1.078 It) (here fig. 26); 1.5 sextarius (0.80 It); 0.5 sextarius
113 Infra p. 61.
(or 6 cyathi, 0.27 It); and 1.5 cyathus (0.068 It).
1"4 PE 15.98. The sextarius was 0.547 liter, corresponding 115 On the relative value of a worker's meal: DeLaine
to one sixteenth of a modius: Giacchero 117. G. Senne- 210; see supra p. 460 with n. 97 for medieval Cairo.

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
approximately 38.168 modii or 19.0

or, in the PE's prices, the equivalen
denarii per year (1908.4 denarii, to
It is interesting to note that th
glass vessels are much lower than t
eval Cairo. A Geniza document da
mentions 18 empty glass vessels wi
two dinars, a price considered t
equivalent of a lower middle-class
income.119 The price is even highe
into account that the 18 vessels w
items for sale in a legume shop, a c
they were utilitarian items made o
green glass. As will be shown below
were probably more realistic.
Whereas the prices of finished ves
tant for the glassblower/retailer an
public, the glassblower's primary c
the prices of raw glass, because he
certain amount of it to set up bu
quirements included work and stor
ment (furnace), fuel, and tools. An
items, including the raw material, m
leased, as several potters' contra
Egypt suggest (see above). Most gla
probably have preferred to work
tools, because blowpipes and hand-
adapt their shape to fit into the
mouth), and also because one's own
Fig. 26. Frontinus bottle. H. 22.4 cm;
with Dm.
the most care. base 8.5 cm;
Weight 377 g. Inscribed FROTINIANA.The contractsAnnular pontil
mentioned in the Geniza docu-
scar. Late first to early second century. Made in northwest
ments
Europe. (Courtesy Toledo Museum ofindicate
Art, that the amount of raw glass involve
no 1948.220.
in setting
Purchased with funds from the Libbey up shop could
Endowment, varyof
gift considerably. The bar
Edward Drummond Libbey.) minimum would probably have been near 10 qintar
(ca. 450 kg), the amount contributed by a glass
one or two average size vesselsblower
of mentioned in one of the
Judaean Geniza documents.120
glass or
one average size vessel of Alexandrian glass.
At the PE's prices, this would translate into an invest-
Another way to relate prices ment
to the cost
of 33,720 denariiofforliving
1405 pounds of Alexan-
uses the price of wheat or its equivalent.
drian glass, or 18,265 The
denariiPE
for set
the same quantity
the maximum price of one kastrensis
of Judaean glass.
modius
In either scenario,
"army the investment
modius" (hereafter KM) of wheatin rawatglass100
was a denarii.116
huge expense. Taking into account
The minimum net consumption the per
inevitable waste of glass
person perduring
year blowing (ca. 40-
45%),121 450
has been estimated at the equivalent ofkg250raw glass
kgsufficed
wheat. to blow approxi-
mately 1080
At a weight of 6.55 kg per modius1"7 vessels
this with an average
works out to weight of 250 g,
116 PE 1.1a. Wheat was measured by volume. One income.

modius was 8.754 It; the KM was twice as much: 17.51 It: 120 Goitein 365, no. 19.
Giacchero 117. 121 Henein (supra n. 40) 20 states that 1250 g of broken
117 K. Hopkins, "Taxes and Trade in the Roman Em-glass yields 1000 g of molten glass, a loss of 20%; during
pire,"JRS70 (1980) 101-25, esp. 118-19. blowing, waste from material remaining on pipe and
118 DeLaine 220 quotes a normal monthly rate of five punty and sticking to crucible accounts for an additional
modii per recipient. loss of ca. 20-25%.
119 Goitein 151: price of 18 vessels; supra n. 95: monthly

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
approximately admitting
10-11 dayslight, especially
worth in the caldaria.
ofThe use
blow
luck some of the of glass
first glassfor this vessels
purpose is mentioned
might by several
be
time first-century
raw glass ran low, Roman authors.125
allowing the glassbl
new supplies or take Two out a loan.
early fourth-century papyri from Oxyrhyn-
The PE's maximum prices
chus suggest that windowwould have
glass was also common in
Egypt. The glassblowers
very difficult for most specificity of these documents requires
to ear a
short discussion
Lactantius's statement (De mort. in spite of this pers.
paper's focus 7.6)
on ves-
"drove goods off the sel glass.market (evidently
In a declaration of prices dated 26 Novem-
price-ceiling was tooberlow to ofallow
317, a member any
the glassworkers' prof
guild at Oxy-
well have been true rhynchus
for cites utilitarian vessel
a price of four talents per hundred
glassblower would have had to work below
pounds of glass:126
of production. The total weight of the 10
To Valerius Ammonianus alias Gerontius, curator of
that could be blown from 450 kg raw g
the Oxyrhynchite, from the guild of the glassworkers
have been
kg, or
of the 825ca.
illustrious 270
Roman
and most pound
illustrious city of the Oxy-
allows a maximum sales price of
rhynchites, through me Aurelius Areion, son 30of... d
pound for AlexandrianIn accordance
glass with orders, at my own risk whic
vessels, I declare
the pricethe
into 24,750 denarii for entered below for the goods which I han-
lot-significant
dle, and I swear the divine oath that I have been de-
the 33,720 denarii necessary to buy 450
ceitful in nothing. As follows:
glass. The maximum Glass,
sales by weight price100 pounds for the sa
talents four.
of vessels made of Judaean glass was
In the consulship of Ovinius Gallicanus and Caeso- 16,50
again less than the cost of
nius Bassus, viri raw glass.
clarissimi. Hathyr Under
30. I Aurelius
Areion, have presented this, making my
cumstances every glassblower's top declaration
prio
as aforesaid.-I Aurelius Pathermouthis, wrote on
have been to cut back on waste
his behalf as he is illiterate.
and recycl
as possible.
The last type of glass mentioned
Previous publications in
associate the price of four t
spec<u>laris, probably "window
talents with glass,"
the PE's price of 24 denarii per pound of cer
chitectural glass (16.5-6). Its
Alexandrian raw glass, low
the most expensiveprice
raw glass is
tion that this was a for
lowvessels. The declaration does not state the pur-
quality glass. It ha
gested that the inclusion
pose of the glass,of
but the window
price appears rather lowglass
for
was more relevant for the western half of the Ro-
vessel glass. Four talents per hundred pounds of
man empire than for the eastern Mediterranean, glass translates into 6,000 denarii,127 or 60 denarii
because in the third century glazed windows wereper pound. If this was the price of Alexandrian raw
not widely used and especially not in the East.123
glass in 317, it suggests an average annual com-
Most documentation of ancient glazing has been pound inflation of 5.89% in the 16 years following
concentrated on Italy and the West. However, therethe PE.128 Inflation rates fluctuated and varied ac-
is increasing evidence for extensive use of window cording to commodity. A low inflation rate has been
glass in the eastern Mediterranean long before the noted for a few commodities, but 5.89% is very low
Byzantine period. The windows of the South Baths in comparison to the average inflation of 13.91% be-
at Bosra are coeval with the original construction tween 301 and 359, and 18.97% between 310/11 and
of the building in the second century.124 Window 359.129 If the annual inflation percentage of glass
glass was an important item in all Roman Bathwas the same as the average annual inflation, the
buildings: it was necessary to keep in the heat while
declaration of 317 should refer to window glass,
122 R. Duncan-Jones, The Economy of the Roman Empire. 126 P Oxy., vol. 54, no. 3742.
Quantitative Studies (Cambridge 1974) 367. 127 After Diocletian's reform, the talent in Egypt
123 Barag (supra n. 103) 116, following D.K. Charles- equaled 1,500 denarii: R.S. Bagnall, Currency and Inflation
worth's hypothesis in Erim and Reynolds (supra n. 103) in Fourth Century Egypt (Bulletin of the American Society of
109.
Papyrologists Suppl. 5, 1985) 16-17.
124 H. Broise, "Vitrages et volets des fen tres thermales A 128 P Oxy. vol. 54, p. 238.
I'epoque imperiale," in Les Thermes Romains, Actes de la table 129 P Oxy. vol. 54, pp. 233; Corcoran (supra n. 98) 225-
ronde organiz&e par 1'Ecole Franfaise de Rome, Nov. 1988 (Col- 26 rounds off these figures at 14% and 19% respectively. I
lection de l'Ecole FranCaise de Rome 42, 1991) 61-78, thank D. Black, University of Toledo Department of Eco-
esp. 68-74. nomics, for calculating inflation rates.
125 Broise (supra n. 124) 61.

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
which the PE priced at 8 and 6 denarii higher raterespectively

of ca. 19% does not (see above). Taking
for first and second quality. The into price account
of 60 thatdenarii
the two glass declarations belong
per pound in 317 is consistent with to the
anfirst quarterinfla-
annual of the century, and noting that
tion rate of 13.42% for first quality the average
window increase in the
glass andvalue of the gold solidus
a rate of 15.48% for second quality was window
16.33% in the years 301-340, I use a hypotheti-
glass.
A second declaration by the same glassworkers'
cal inflation rate of 16% for the years 301-326. Even
guild at Oxyrhynchus, datedJune/July if it is not
326,accurate, it gives an idea of the price
specifies
the use of 6000 pounds of glass in range.
theWith an inflation rate of 16%, second quality
construction
of the public bath:130 window glass priced at 6 denarii per pound in 301
would have cost about 245.25 denarii per pound in
In the consulship of our masters Constantinus Au-
gustus for the seventh time and Constantius the most
326. This translates into 16.35 talents per 100 pounds
illustrious Caesar for the first time. To Flavius Leuca- of glass. If this is an acceptable hypothesis, the re-
dius logistes of the Oxyrhynchite nome from the maining 5.65 talents per 100 pound, about one quar-
guild of glassworkers of the glorious and most glori-
ter of the price, represented labor and other costs.
ous city of the Oxyrhynchites through me, Aurelius
The total invoice of the glassworkers' guild was
Zoilus .... In response to your demand for an ac-
1320 talents. Of this sum 339 talents, or 508,500 de-
count of all the matters affecting our profession re-
narii, were left to cover the cost of wages, scaffolding,
lating to the service of fitting out the warm baths in
the public bath of the city, I have perforce drawn transport
it to the building site, and other construc-
up and submit it in order that your grace may be
tion costs. Six thousand pounds of glass would have
able to know. It is: for the work needed on the warm
sufficed for approximately 242.37 m2 of window
baths, x hundred pounds; for the work needed on
the gymnasium, x hundred pounds; at a rate of panes.133
22 No data are available for calculating the
talents per hundred pounds. Total 6000 pounds, to- cost of glazing, but there are indications for the cost
of placing mosaic cubes. Window panes were much
tal 1320 talents. Which we accordingly report. In the
aforementioned consulship, Epeiph. .... I, Aureliuslarger than mosaic cubes, but-unlike the cubes-
Zoilus, have presented this as set out above.
each individual pane must be framed in wood, plas-
The glass used in the public baths in 326 canter, or metal. The construction of the frame itself
hardly be anything other than architectural glass.
might be complicated by provisions for opening and
Previous publications have combined the prices men- closing the window for ventilation.134 In the absence
tioned in the two papyri and interpreted them as evi- of data on glazing, I therefore tentatively substitute
the wages that would have been needed to fit the
dence for an increase of 450% in the price of glass.131
This sounds very high for a period of just eight-and- baths at Oxyrhynchus with 6000 pounds of decora-
a-half years, but it translates into an annual inflation
tive glass mosaics instead of with windows.
of 20.22% for the years 317-326. Although the rate Mosaic cubes were made from flat glass "cakes"
may thus seem acceptable,1'32 it cannot be used. The having the thickness of the cubes (ca. 0.7 cm). The
glassworkers' declaration of 326 states explicitly that
cake was scored and broken up into tesserae.135 Col-
the 22 talents per pound include the cost of "fitting"
ored glass could be imported in the form of cakes, but
the glass. I propose to establish the price of the glass
glassworkers may have also prepared the cakes them-
itself by basing it on the average annual inflation selves. The process was simple and involved little la-
rate for the period 301-326. bor apart from gathering and transporting fuel. The
Two average inflation rates for the first half of the
natural surface tension of glass causes any chunk of
fourth century have been calculated; the lower rate raw glass to flatten out into a puddle when melted
of ca. 14% includes the years 301-310/11, the that, upon cooling, stiffens into a cake or disk of 0.7
130 P Oxy., vol. 45, no. 3265. bridgeshire 1980-85 (London 1996) 397-409, esp. 396-97.
131 P Oxy. vol. 54, no. 3742, commentary to line 13. Bag- Blown window panes, which became common in Britain in
nall (supra n. 127) 69 does not calculate the rate of infla- the late third and fourth centuries, would probably have
tion but lists the prices for glass at 4 and 22 talents as been slightly thinner and covered a larger surface. On the
though they refer to the same item. possibility of primary glassmaking (for window glass?) in
132 The commentary to P Oxy. vol. 54, no. 3742 (supra n. northern Britain: C.M.Jackson et al., "The Manufacture of
131) states that the average annual inflation was 22.2%, Glass in Roman York," JGS 40 (1998) 55-61.
but it does not specify the time period. 134 Broise (supra n. 124). One window usually consisted
133 Calculation based on the fact that 8390 cm2 of cast of several panes.
glass window panes weighed 6.8 kg. For this and the meth- 135 S.M. Goldstein, "Glass Fragments from Tell Hesban,"
ods of producing window panes, see J. Price: "Glass," in Andrews University Seminary Studies 14.1 (1976) 127-32,
R.P.J. Jackson and T.W. Potter, Excavations at Stonea, Cam- esp. 129.

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
to 0.8 cm
thickness.136 Numerous
green Judaean broken
glass. Unless glass mosaic cubes were c
glass in various colors
as inexpensive
(several as window glass,
tints
which seemsofimprob-
blui
gray, black, and red) were
able, the found
price of 22 talents, at
which included Gera
fitting,
total weight of 45-50 kg.
cannot apply toRoughly circular
mosaics. I suggest that the two decla-
some measured up to rations
40 of the
cm Oxyrhynchus glassworkers' guild in th
in diameter;
thickness was 3-4 mm.137
317 and 326 both refer to window glass.
Six thousand pounds of glass would have su
PE 16.7-9
to cover ca. 100-109 m2 of wall and vaults in
lic bath at An entry for
Oxyrhynchus, glass mosaic cubescounting
without is missing in the
between the cubes (ca. 20%).
PE's section At
on glass. I would likeca. 15,000
to suggest that it was
the square meter, DeLaine
contained in PE chaptercalculated
16, lines 7-9, given in tha
Table
took ca. 2.8 days per
1. Thesquare meter.138
main reason for suggesting that this sectionTh
re-
that the glassworkersfersat Oxyrhynchus
to glass mosaic cubes (or the cakes forneedemaking
280 them) is the occurrence
man-days to complete the of the word
job. At "colored"
an in lines
av
nual increase of 16%, the
8 and 9. The wordPE's daily
itself is not preserved inwage
the Latin
narii for one unskilled
version but laborer
appears in the Greek(plus
edition offood
the PE:
10 denarii in kind) would have
chromatos "(of) color." Glassrisen
is one of theto som
few materials
between 1,216.32 and 1,420.69
where color denarii
is a significant factor (f
in determining the
price. Today's
worth 30 and 35 denarii prices vary between U.S. $18.99
respectively). Two and
eighty days of labor$70.85
cost per kg,the
dependingguild 240,570-
on the coloring agent.140
denarii. This seems Thepossible
number of letterswith 508,50
missing in line 7, the sec-
available for wages, plus
tion title,the cost
is unknown. of originally
J. Reynolds scaffold inter-
supervision, and transport.
preted the last letters of the last word in the title as
The above calculations show
ARIS, which is also that window
the text given by Giacchero. g
If
the specifications ofthat
both
reading is papyri. An
correct, the missing word altern
was perhaps
sibility is that the glass
related toused
MUSEARIUS, in PE the constru
chapter 7, line 6, with
the baths was not for window
reference to the wage of glass
a musearius, but wa
a fitter of glass
Glass mosaic cubes were a common form of decora- mosaics. In her 1989 reedition of the text Reynolds
tion in public buildings. J. DeLaine calculated that proposes to read BUS at the end of the heading,
16,900 m2 of glass mosaic decorated the walls and which she tentatively interprets as coloribus "col-
vaults of Caracalla's baths at Rome.'39 The above cal- ors."141 The Greek copy of the PE preserves in line
culations have already shown that 339 talents would 7a a masculine genitive ending in ou, which is consis-
have sufficed to fit glass mosaic cubes. tent with a lost word chrysou "of gold (leaf)" glass.'42
For the Baths of Caracalla in Rome, which were If my hypothesis is correct, that this section is about
decorated with figural mosaic designs, DeLaine glass for mosaics, the three maximum prices of 40,
based her calculations on the price of Alexandrian 30, and 20 denarii respectively could refer to gold
raw glass because she reasoned that colored glass leaf, colored, and natural bluish-green cubes (or the
would have been more expensive than natural bluish- cakes for making them).
36 Stern (supra n. 43) 25-29; Stern 1994, 66-67. Von have another look at the stone in the summer of 1999. In
Saldern (supra n. 9) 97, no. 729, pl. 17 illustrates a frag-
the Aphrodisias copy, the section on pens and ink (four
ment of a blue cake for making cubes. On Roman imperiallines) are cut immediately after the section on glass; the
trade in cakes of colored glass, infra n. 213. lines numbered 15.7-9 by Giacchero appear at the top of
137 P.V.C. Baur, "The Glassware," in C.H. Kraeling the
ed.,next column. The coverage of the PE was not com-
Gerasa, City of the Decapolis (New Haven 1938) 505-46, plete,
esp. but "glass cubes for wall mosaics are a very attractive
517-18.
idea, and while logic does not seem to have been an obvi-
8 DeLaine 180-82. ously guiding factor in the organization of the edict, there
139. DeLaine 180; large numbers of glass mosaic cubes of could be some logic in the progression Glass to Glass Mo-
the fifth to sixth century were excavated at Sardis: von Sal-
saic Cubes to Ivory and Tortoise Shell" (J. Reynolds, per-
dern (supra n. 9) 92-94, pl. 17. sonal communication January 1999).
140 On current prices for raw glass: E.M. Stern, "Glass 142 On sandwiched gold leaf tesserae see von Saldern
and Rock Crystal. A Multifaceted Relationship," JRA 10 (supra n. 9) 93; a late Roman or early Byzantine cake of
(1997) 193 with n. 5.
sandwiched gold leaf for making mosaic cubes was exca-
141 This section is part of the Aphrodisias copy: Erim vated at Heshbon in Israel: Goldstein (supra n. 135) 129
and Reynolds (supra n. 103). Reynolds 1989 (supra n. 98) and pl. XI:B, bottom row.
was brought to my attention by the author who plans to

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
COMMERCE AND TRADE Flavian poets mention peddlers hawking sulphur
for broken
The chief interest of the glass industry for glass.
eco-Martial (Epigr. 1.41) disparagingly
compares someone to a transtiberinus ambulator/qui
nomic historians lies in the fact that its products be-
pallentia sulpurata fractis/permutat vitreis "tramping
long to several categories-raw materials, house-
hawker from beyond the Tiber who exchanges pale
hold and utility ware, and luxury items including
sulphur matches for broken glass."'144 Itinerant mer-
fine tableware-each of which can reflect the pros-
chants and peddlers probably included some glass
perity of different groups of the population. The aim
among the wares they brought to outlying villages
of this section is to discuss some of the evidence for
and they may have brought back fragments of bro-
commerce and trade as it relates specifically to the
ken glass vessels.
glass industry, not to present an economic distribu-
Distribution maps provide an interesting basis
tion model for Roman glass. The emphasis is on
for reconstructing patterns of ancient trade. Table-
distribution patterns and means of transportation.
ware and most other glass vessels were sold empty,
Vessel glass, for example, was made in all parts of
so their trade patterns reflect commerce in glass,
the Roman empire, but there is literary and archae-
but certain types of glass bottles appear to have
ological evidence that it was also traded by water
been sold filled with specialized contents, implying
and by land, both within and beyond the borders of
cooperation between the manufacturer of the con-
the empire.
tent and the glassblower. For example, unguentaria
Just as the glass industry was divided into two sepa-
shaped like birds and spheres were filled with cos-
rate branches, one concerned with making the raw
metic powders and fire-closed at the tip.145 The dis-
material, the other with working it into objects, so also
tribution of such vessels reflects the commerce of
commerce in glass was twofold. Workshops needed
their contents.
raw glass, and finished vessels needed to reach cus-
tomers. There was probably also a limited trade in
Inscribed Bottles: Evidence for Glass Trade?
blanks and half products for engraving, cutting, andSeveral classes of larger storage and/or transport
painting. Recycling, a fourth commercial outlet, would
glass bottles carry mold-blown inscriptions. Mercury
engage merchants from the smallest peddler to large-
bottles, for example (fig. 27, Isings Form 84), were
scale enterprises involving shipments of tons of cul-
named after the base molding that often includes
let (broken glass vessels). a representation of Mercury. Mold-blown barrel-
Whereas long distance trade in fine tablewares, shaped bottles with one or two handles were also
raw glass, and cullet could be economically profit-known as Frontinus bottles (Isings Forms 89 and
able (see below), most unguentaria, ordinary house- 128, fig. 26) because of the frequency with which
hold containers, tableware for daily use, and funer-
that name appears in the base molding. Many pris-
ary urns were destined for local and regional matic bottles (usually square, Isings Form 50), have
markets. Variations in the finishing of individual base moldings with inscriptions and/or geometric
objects and the prevalence of specific shapes in designs on the bottom. The greatest interpretive di-
one region are evidence for increasing regionaliza- lemma associated with these bottles is whether the
tion of production beginning in the second half of base moldings refer to the glassblower or to the pro-
the first century.143 ducer of the contents.146
143 Price (supra n. 8) 30-39; Cool and Price 225-27. 108, n. 9, notes that an intact glass bird in Turin containe
144 Compare Juv. 5.48 and Stat. Silv. 1.6.74. See C. Isings, a rose-scented liquid. On birds, most recently: G.M. Fac
"Exchanged for Sulphur," in Festoen Opgedragen aan A.N. chini in Vetro e vetri (supra n. 11) 131-36.
Zadoks-Josephus Jitta bij haar zeventigste verjaardag (Scripta 146 M. Sternini, "I vetri," in Harris (supra n. 84) 431-5
Archaeologica Groningana 6, Groningen n.d., ca. 1976) gives a useful survey of all classes of inscribed glass vesse
353-56.
On Mercury bottles: Stern (supra n. 76) 64-72, no. 18
G. M. Facchini et al., "Studio di una forma vitrea di etai ro
145 Isings Forms 10 and 11. For recent analyses confirm-
ing that the content was colored powder: L.A. Scatozzamana: La Merkurflasche," Postumia 6.6 (1995) 150-73; M
H6richt et al., "Prime osservazioni ed analisi sul contenuto
Sternini et al., "Unguentari in vetro con bollo nelle coll
di alcuni recipienti in vetro rinvenuti nell' area archeolog-
zioni del Museo Nazionale Romano," Annali della Facoltd
ica di Pompei," in L. Franchi dell' Orto ed., Ercolano 1738- e Filosofia 17 (Universiti di Siena 1997) 55-10
Lettere
1988. 250 anni di ricerca archeologica, Atti del convegnoG.M.
inter-
Facchini in Vetro e vetri (supra n. 11) 139-46. O
nazionale Ravello-Ercolano-Napoli-Pompei 1988 (Rome 1993)
Frontinus bottles: Sennequier (supra n. 114) 169-82;
557; J. Perez-Arantegui et al., "Analysis of the Products
Cool and Price 204-206. On prismatic bottles: Sternin
Contained in Two Roman Glass Unguentaria from the Col-
1993, 88-93: group III; Cool and Price 183-84 (square
ony of Celsa (Spain)," JAS 23 (1996) 649-55. C. Macca-
bottles 179-99).
bruni, I vetri romani dei Musei Civici di Pavia (Pavia 1983)

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
one glassblower or one workshop to produce bottles

with different base moldings for several commission-
ing parties, although there is no proof for this practice.
The base moldings of Sentia Secunda include an
abbreviation VITR[a (or vitrearia)] for "glass" (or
"glassworker"), and the word FECIT "made it" (figs.
23-25).148 The occurrence of the word fecit makes it
obvious that these moldings refer to the producer of
the bottles; their findspots reflect overland trade in
glass. Sentia Secunda's workshop was in Aquileia; the
bottles were excavated in Linz, Austria. However,
most moldings on square bottles do not include the
word fecit.
The square bottles carrying the name of Salvius
Gratus-without the addition of fecit-may serve to
illustrate the problem of inscribed glass bottles. Bot-
tles featuring his name in the base molding are com-
mon in north Italy and southern Germany. A ship
that sank six miles from Grado (near Aquileia) with
a mixed cargo of amphoras, glass (mostly cullet),
and other goods included several fragments of bot-
tles marked C Salvius Gratus, which may or may not
have been cullet at the time of their sea passage.'49
If Salvius Gratus was a glassblower, the findspots are
evidence of overland trade and short distance trade
Fig. 27. Mercury bottle. H. 20.5 cm; wt. 243.8 g.
GFHI. No pontil scar. by
Second century.
ship (either with the bottles or with Made
their shards).in
northwest Europe. (Courtesy Toledo
The findspots might Museum
even indicate the existence of aof
1987.216. Gift of Rabi R. Soleimani.)
branch glass workshop at Augst.150 On the other
hand, if the base moldings refer to the contents, new
Most prismatic bottles
bottles were with base
not representative moldin
of trade in glass but
from findspots in trade western
the in whatever filled Salvius Gratus's bottles.
half of the Ro
pire. Their production technique
Archaeological appears
evidence suggests that the absence t
characteristic of the of a maker's identification
West: they does not of itself
were mean
blown in
walled molds, which that
madethe name itseen in
possible
a base molding refers
to tospee
the
manufacturing process for
contents. creating
In a shop at Herculaneum, ancontaine
order of glass
standardized capacity.
vessels This
still enclosed
method
in packing materials
for included
im
(economizing) the production
an empty square bottle ofwith purely
the name of P. Gessius
utili
vessels was very rare Ampliatus
in the in theeastern
base molding, a find circumstance
Mediterra
The bottom section of the
that implies thismold, carrying
square bottle was sold empty; in other th
molding with the name, had
words, this signature four
without slots,
the addition of fecit re- o
each edge, for inserting panels
ferred to the glass that
shop, not to the contents.151made
walls. The curious construction of the mold with in-
In other respects, the problems regarding the in-
terpretation of the names and the form in which
terchangeable parts would, in theory, have enabled
147 Stern (in prep.). On the technique: V. Seitter, "Be-149 A.J. Parker, Ancient Shipwrecks of the Mediterranean &
merkungen zur Herstellung von formgeblasenen r6misch- the Roman Provinces (BAR International Series 580, Oxford
en Gldisern mit Bodenmarken," ArchKorrBl 21 (1991) 527- 1992) 197, no. 464, dated ca. 200; on the overland routes
33. Molds for square bottles have been excavated at Augst
for bottles marked by Salvius Gratus: Roffia 1981 (supra
and Saintes: Riltti 163-64, fig. 103, pl. 218: 05 and n.06;83).
Hochuli-Gysel 1991 (supra n. 8), 85-87, figs. 5-7: six mar-150 Rottloff (supra n. 33) 170-72. On the possibility of
ble panels blackened through use, the panels themselves branch workshops producing bottles marked by Salvius
were reused revetment slabs. The mold published by F.
Gratus: supra p. 459 with n. 83.
Fremersdorf, "Die Anfainge der r6mischen Glashfitten 151 A. de Franciscis, "Vetri antichi scoperati ad Erco-
K6lns," KoilnJb 8 (1965/66) 24-43, esp. 29 and fig. 2:9lano,"
is JGS 5 (1963) 137-39; on the gens Gessia and its
now thought to be "from the base of a pottery version connections
of with Campania: Scatozza H6richt 1991 (supra
a square bottle": Cool and Price 180. n. 11) 76-79. More on this shop: infra p. 471 with n. 174.
148 On Sentia Secunda (supra n. 68).

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
they appear are similar to those stamps.156 Most of these bottles

encountered have been found in
in pot-
tery stamps. No agreement appears to exist
Italy and the western part of theon
Roman empire; they
whether the latter refer to the owner's social status have not yet been reported from the eastern Medi-
and/or should be interpreted as documents of busi- terranean. It is therefore of interest that the coin
ness relationships.152 The names on the glass bottles used to create the stamp of one bottle found in
north Italy has been identified as minted by the Koi-
are usually Latin, but some inscriptions are in Greek.153
The grammatical form of the name varies, appearing non Bithynia (128/129).157 The wide-bodied unguent
either in the nominative or in the genitive, possibly bottle (also known as candlestick unguentarium)
implying a different relationship with regard to pro- that became fashionable in the second half of the
duction, such as the nominative (Sentia Secunda) first century might actually have been designed spe-
for a master glassblower/owner and the genitive for a cifically to create space for this type of administrative
workshop product "of so-and-so." Other differences inscription.158 However, its exact purpose remains
regard the convention of Latin names that can take unclear. The fact that relatively few bottles are in-
the form of a freedman's name or of a Roman citi- scribed suggests that the inscription was not meant
zen. Two tria nomina stamps, each consisting of threeto guarantee the quality of the contents at the retail
initials that were sometimes combined on one bot- level. It is still unclear at which point in production,
tle, may indicate persons playing the "composite rolepackaging, or distribution the inscription played a
role and for whom it was destined, especially since
of producer-refiner-dealer of the valuable contents."154
In addition, there are some names that add a top- wooden labels might be attached to the bottles to
onymic like "the Antiocheian" or "of/from Chios."155 identify groups or "batches" of vessels.159
Whereas Sentia Secunda used the old-fashioned loc-
ative case to indicate that her workshop was at Retail

Aqui-
leia, a toponymic can refer either to the location of
Thea excavations of the cities buried by Vesuvius's
workshop or to the origin of an artist or artisan work-
eruption provide fascinating opportunities to com-
ing far from home. pare the number of silver, glass, ceramic, and bronze
Goods (scents?) packaged and sold in glass vessels
con- in use at one moment in time. In these finds
tainers with imperial administrative inscriptions in vessels outnumbered thin-walled pottery by as

glass
relief impressed on the underside were a special many as two or three times, a proportion strongly
case; all aspects of production, packaging, and distri-
suggesting that glass had largely replaced thin-walled
bution were in the hands of the emperor. The in-
pottery as common tableware.160 The glass vessels
scriptions often include words such as VECTIGAL,
available for sale came from an astonishing range of
PATRIMONIUM etc.; others include monetary locations in the West and East, suggesting intensive
152 On the nomenclature in pottery stamps: Aubert (su- 156 Sternini 1993 (supra n. 146) 85-88, Group II (with
pra n. 84) 284-95. In the case of C. Salvius Gratus, the
lit.). See also A. Frova, "Vetri romani con marchi," JGS 13
cognomen Gratus was equally used by citizens, freedmen (1971) 36-44;J. Price, "Roman Unguent Bottles from Rio
and slaves: Roffia (supra n. 83) 123. Tinto (Huelva) in Spain," JGS 19 (1977) 30-39; L.
153G. Lehrer-Jacobson, "Greek Names on Prismatic Taborelli, "Vasi di vetro con bollo monetale," Opus Rivista
Jugs,"JGS34 (1992) 35-43; Trowbridge (supra n. 3) 120-28
Internazionale per la Storia Economica e Sociale dell' Antichitd
provides a list of Greek and Latin names and letter combi-
1.2 (1982) 315-40; Taborelli, "Nuovi esemplari di bolli gia
nations found on glass vessels but it is obviously outdated noti
by su contenitori vitrei dell' area centro-italica," Picus 3
the numerous finds that have come to light since the publi- (1983) 23-69; Taborelli, "A proposito della genesi del
cation of her study in 1930; some of these are mentionedbolloin sui contenitori vitrei," Athenaeum 63 (1985) 198-217;
EAA, Supplement 1970, s.v. vitrarius (M.C. Calvi). Sternini et al. 1997 (supra n. 146) 77-90.
154 L. Taborelli, "Contenitori di vetro con bollo: Un caso 157 H. Bilsing, "Der Miinzabdruck im Boden einer Glas-
esemplare della loro problematica," Rivista archeologica
flasche von Ficarolo (I)," AntW22 (1991) 21.
dell'antica provincia e diocesi di Como 177 (1995) 71-87;158 Sternini 1993 (supra n. 146) 91.
Taborelli, "Riflessioni sul caso di un bollo vitreo con tria 159 L. Taborelli, "Sulle ampulle vitreae. Spunte per l'appro-
nomina forse ridotta a sigla," in Athenaeum Studi di Letter- fondimento della loro problematica nell'ottica del rapporto
atura e Storia dell'Antichita (Universiti di Pavia) 86.1 (1998) tra contenitore e contenuto," ArchC144 (1992) 309-28, fig. 1.
287-89, pls. I, II. On names of glassblowers indicating 160 Morel (supra n. 38) 258-61; L.A. Scatozza H6richt, I
freedmen: supra p. 457. vetri romani di Ercolano (Rome 1986) 22 gives the overall
155E.g. Paulinos Antiocheus (nominative, in Greek, percentages of glass and thin-walled pottery in the site mu-
"Paulinos of Antiocheia"): Barag (supra n. 103) 109-11, seum at Herculaneum as 260 glass vessels (71.04%) and
figs. 1-3; and "Tiberinou Chio[u]" (genitive, in Greek, "of 106 thin-walled pottery (28.96%). At Cosa, thin-walled pot-
Tiberinos of Chio[s]"), from Tharros, Sardinia: G. Pesce, tery went out of use in the Claudian-Neronian period (41-
in StSard 14/15 (1955/57) 356 and fig. 104 (interpreted as 68): M.T. Marabini Moevs, Roman Thin Walled Pottery from
"chio[n] (wine measure) of Tiberinos"). On the problem of Cosa (1948-1954) (MAAR 32, Rome 1973) 45.
interpreting toponymics: Stern 1995, 72.

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
commercial contacts able in the second
ranging century. Glass
from north usage wasItaly
domi-
Asia Minor, and nated by plain but good
Syro-Palestine. quality glass plates, storage
Significantly, th
merce was geared more toward
bottles, trade
simple household in
unguentaria, and single
glass cin-
and small quantities or series
erary urns, most ofthan
which were large-scale
probably produced re- b
ports.161 This is also borne out
gionally. Finds by the
from controlled glass
excavations in north-ves
covered from shipwrecks west Italy(see below).
suggest that in the third century relatively
Beginning in the early
simple second
glass vessels werecentury, vesse
indicators of wealth.165 The
sale tend to come from fewer sources and from a designation of Milan as one of the capitals of the em-
more restricted area than previously. An evalua-pire (end of third century) signaled the beginning
tion of the glass excavated at Colchester concludesof economic revival in north Italy. Costly engraved
that by the early second century, the city "wouldglass tablewares, occasionally decorated with unpro-
have been supplied with glass produced by local tected gold foil on the exterior, are imports from the
centres, by centres further afield in the provinceRhineland (Cologne) and from Rome.166 There is
(i.e., Britain), and almost certainly by imports from no evidence for the production of luxury glass in
the Rhineland, Belgium and northern France."'162north Italy in the fourth century. Glass cups, beakers,
Increased regional production destined for re- and bottles produced in the region (at Aquileia?)
gional markets appears to have been typical for were available for domestic use.
many trade goods.163 In general, the manufacture and the sale of locally-
Whereas the influx of technology and industry led made household wares were probably not widely sepa-
to economic growth in Gaul and the northwest prov-rated. In Italy the production of fine glass tableware
inces of the empire, Italy at the end of the secondpeaked in the first century, a period when artifacts
century began to experience a depopulation and anwere "commonly produced within the household or in
economic crisis that affected all areas of life in thesmall workshops appended to stores, where indepen-
peninsula. In the archaeological record of north Italydent craftsmen sold their finished products on a local
the crisis is tangible even earlier, beginning in the scale."'67 There is archaeological evidence for a glass-
second half of the second century.164 Problems for blower/retailer from the Byzantine period (see below).
the glass industry may have begun earlier still. The When the glassblower doubled as a shopkeeper,
decrease of vessel glass postdating the first century isthis involved a whole new set of opportunities for dif-
now becoming increasingly obvious. Areas that were ferent types of cooperative agreements and hiring
at the forefront of luxury production for conspicu-practises. Cash flow would have been a never-ending
ous consumption in the mid-first century (see figs. problem for those manufacturers who were also
11, 12), producing not only for local markets butshopkeepers. In second-century Jewish circles,
also for export, appear to have dropped out of busi- where Rabbi Akiba's saying, "The shopkeeper ex-
ness toward the end of the century. A rough count oftends credit," was held in honor,168 buying on credit,
the glasses in museums in northeast Italy (Veneto)whether wholesale or at the retail level, appears to
suggests that luxury tablewares were not readily avail-have been the rule. The owner of a commercial
161 Morel (supra n. 38) 250-51. Museo Archeologico Nazionale di Adria (CCAVV 2, Murano
162 Cool and Price 227.
1996); A. Larese and E. Zerbinati, Vetri antichi di raccolte con-
163 Panella (supra n. 74) 431-37; one may also compare cordiesi e polesane (CCAVV 4, Murano 1998); G. Zampieri,
the glass finds from the villa at Settefinestre: G. DeTom- Vetri antichi del Museo Civico Archeologico di Padova (CCAVV 3,
maso, "Vetro," in Settefinestre: una villa schiavistica nell'Etruria Murano 1998). A similar decline, beginning slightly later,
romana 2: La villa e i suoi reperti (Modena 1985) 173-211. has been noted in the Ticino area: Maccabruni (supra n.
164 L. Brecciaroli Taborelli, "I1 vasellame da mensa in 145); S. Biaggio Simona, I vethi romani provenienti dalle lerre
etA tardoantica," Archeologia in Piemonte 2: L'eti romana dell'attuale Cantone Ticino (Locarno 1991) 27-29; HelvArch
(Torino 1998) 271-89. On the crisis in Italy, its effects and 22 (1991) 78-143.
its probable causes: Panella (supra n. 74). See also Panella, 166 F. Paolucci, I vetri incisi dall' Italia settentrionale e dalla
"Merci e scambi nel Mediterraneo tardoantico," Storia di Rezia nel periodo medio e tardo imperiale (Firenze 1997) 196-
Roma 3.II (Torino 1993) 613-97. 97; Brecciaroli Taborelli (supra n. 164) 275-77. On local
'65 Brecciaroli Taborelli (supra n. 164) 273-75; Vetro e production for regional use: M. Buora, "Una produzione
vetri (supra n. 11) 77-128. Compare the glass excavated in artigianale di un vetraio a Sevegliano (agro di Aquileia,
cemeteries at Asti, Alessandria, Susa, Brescia, and Vog- Italia settentrionale) nel IV sec. d.C.,"JGS 39 (1997) 23-31.
henza. Glass in north Italian museums: Calvi (supra n. 11); 167 Aubert (supra n. 84) 201.
G.L. Ravagnan, Vetri antichi del Museo Vetrario di Murano
168 Mishna Aboth 3:16, cited after Goitein 151 and 438,
(Corpus delle collezioni archeologici del vetro nel Veneto n. 8.
[CCAVV] 1, Murano 1994); S. Bonomi, Vetri antichi del

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
varietygreeted
flower garden at Pompei, however, of fine tablewares.
everyone The preliminary publica-
who crossed the threshold into his house with the tion documents at least 31 different vessel shapes,
words CRAS CREDO "I will give credit tomorrow" in-
which means that most shapes for sale were available
laid in the mosaic doormat at the entrance.169 in very small quantities.'73
Some glassblowers may have diversified their At the time of its destruction in 79, a shop near
stock and sold products made by colleagues/part-the forum of Herculaneum held a contingent of
ners (glassblowers and other artisans). Similarly,
glass vessels packed in straw and other materials and
pharmacies and drugstores sold glass vessels that
divided into separate packages according to vessel
shape. The glass vessels included:174 10 monochrome
they filled with herbs and scents. Many fragments of
small glass unguentaria and a few almost complete shallow ribbed bowls (Erc. 2a), two small bowls with
bottles were excavated in a commercial flower gar- tubular rims and base-rings (Isings Form 44a, Erc.
den at Pompei.170 Epigraphical evidence indicates 8), six large bowls (Isings 44b, Erc. 8), another large
that in Pompei glassworking and the sale of frank-bowl (Isings 42, Erc. 9), four undecorated ladles and
incense were concentrated in the same part of the four with spiral thread (Erc. 17), one tall, straight-
walled mold-blown beaker (Isings 31, Erc. 19), two
city: regio clivi vitrari sive vici turari "the quarter of
the glassworkers also known as the quarter of the indented beakers (Isings 32, Erc. 21), three cylindri-
frankincense dealers."171 cal beakers (Isings 30, Erc. 23); one square bottle
Imported glassware appears to have been sold in marked P. GESSI AMPLIATI (Erc. 25), one cylindri-
combination with imported pottery. A mid-first-century cal bottle (Isings 51b, Erc. 25), one bulbous jug with
store at Colchester stocked a selection of glass vessels upturned spout (Erc. 29), one mold-blown cup
in addition to Samian ware and various other types shaped like the head of a black (Erc. 33), two arybal-
of fine pottery and clay lamps. The glass vessels had loi with dolphin handles (Erc. 40), one small spheri-
been stacked on shelves above the pottery. When fire cal bottle (Erc. 41), one small spherical bottle (Erc.
destroyed the store (ca. 50-55) much of the glass 46), two tubular unguentaria (one Erc. 47a, the
melted and dripped down on the pottery. Neverthe- other Erc. 47d), three piriform unguentaria (Erc.
less, several glass vessel shapes have been identified:
49), one carinated bottle (Erc. 50), perhaps one urn
shallow sagged bowls, plates, natural bluish-green with M-shaped handles (Erc. 57), and one lid (for an
ribbed bowls and the more luxurious monochrome urn) (Erc. 59). Apparently, the glass tablewares were
blue and polychrome mosaic ribbed bowls. Blown sold in sets, like metal and pottery.
vessels included small cylindrical cups of the typeThe buying of glassware in sets is also documented
known as Hofheim cup (see fig. 14), a yellow sky-by the presence of sets in houses in Herculaneum as
phos, and a cylindrical bottle.172 well as in first-century tombs throughout the Roman
An even more mixed assortment of pottery, glass,
empire. Glass sets were found in tombs at Vervoz, Bel-
and lamps made up a merchant's stock at Cosa, gium
de- (dated 60-75), at Saintes, southwest France
stroyed in 40-45 when one of the walls of the forum-(40-60), in Dalmatia (first century), and at Vize, East-
basilica collapsed: Arretine pottery, amphorae, lamps,
ern Thrace (mid-first century, possibly before 44).175
thin-walled tablewares, coarse pottery, and 76 glass Literary evidence attests the use of glass sets in
vessels. The glass included mold-formed, ribbed, mold-
Egypt. In a letter ascribed to the early second cen-
blown, and free-blown vessels, including an amazing
tury a certain Claudius Terentianus lists among the
169 W.F. Jashemski, "The Garden of Hercules at Pompeii 174 On the shop on the Decumanus Maximus: de Fran-
(II.viii.6): The Discovery of a Commercial Flower Garden,"
ciscis (supra n. 151); Scatozza H6richt (supra n. 11). The
AJA 83 (1979) 403-11, esp. 410. following compilation is made from Scatozza Horicht
170Jashemski (supra n. 169) 407. (supra n. 162); the numbers preceded by "Erc." refer to
171 ILS 1224b, quoted by Isings 5, with n. 3. On connec-
her forms.
tions between medicinal preparations and glass contain- 175 M.-C. Gueury and M. Vanderhoeven, "La tombe
ers: L. Taborelli, "I contenitori per medicamenti nelle pre-
gallo-romaine de Vervoz aux Musees Royaux d'Art et
scrizioni di Scribonio Largo e la diffusione del vetro
d'Histoire," BMusBrux 60 (1989) 107-24; H. Chew, "La
soffiato," Latomus 55 (1996) 148-56. tombe gallo-romaine de Saintes. Nouvel examen du mate-
172H.E.M. Cool, "The Boudiccan Uprising and theriel," Antiquitis Nationales 20 (1988) 35-61; M.R. DeMaine,
Glass Vessels from Colchester," Expedition 38.2 (1996) "The
52- Northern Necropolis at Emona: Banquet burials with
62, esp. 57-58 and fig. 9. ladles," AnnAIHV 11, 1988 (Amsterdam 1990) 129-44;
173 F. Grose, "Roman Glass of the First Century AD.A.M.A Mansel, "Les fouilles de Thrace," Belleten 4 (1940)
Dated Deposit of Glassware from Cosa, Italy," AnnAIHV 6,
115-39, esp. 133 with figs. 47-49.
1973 (Liege 1974) 31-52.

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
objects he is sendingconical
fromnecked bottles
Alexandria
(tapering both up and down),
to K
et accipias caveam one bottle with a funnel neck"
gallinaria(m) in andquaglass rods.182
ha[bes]
vitriae et phialas A Byzantine glass shop
quinarias (sixth-seventh
p[ar century)
u]nu<m
paria sex .... was recently
"Receive also excavated
a in the center of Bet She'an,
chicken coop
you have sets near the bazaar.183
of glassware, It consisted
two of two connecting
bowls (lit.
rooms and a courtyard.
of quinarius size, a dozen goblets From the (lit.
description"six
it ap- p
A second-century papyrus
pears that this shop
from
was run by the
Oxyrhyn
glassblower. The
cates that glass was wall of one of the two rooms showed
bought even traces of shelves
in half
count of articles at for order of
storing glass vessels, Eugenetor
but most of the glass objects
sack: (. .. .) 2 procheiria
appear to have been (handboxes)
found in the other room that con
opened eastwardof
hemisyntheseis (half-sets) onto the street and also
glass, 4 contained
... pote
ing cups) and 1 . . the.,furnace-a
4 batellai
setup that suggests the(plates),
furnace (and
(bowls), 1 oxybaphon (saucer)."177
the glassblower) benefited from the draftIn fourt
entering
through the
Karanis, excavated sets of door.184
glassThe front room was divided
were almo
composed of dishes, into
bowls,jars, flasks,
two workspaces, where "numerous and
glass vessels-
to-do residents stored
many of them intact-found along the walls and in tog
glass tablewares
red-polished pottery several concentrations,in
dishes had been stored on shelves, and
baskets
in pithoi that eitherracks
stood
of shelves or on the
in baskets." floor
The shapes were com- or
into it.178 A late fourth-century
mon utilitarian shapes for daily use: two shop
types of a
stocked glass and pottery vessels
lamps for use in of spherical
a polycandelon, goblets, the or
sam
suggesting the shopkeeper had
piriform bottles, small jugs ordered
with large loop handles,
cially to sell sets in
different
bowls, and window panes materials.
(round and rectangular).
In the row of Byzantine shops abutting t
gogue at Sardis, two adjoining
Long-Distance Trade within the Romandouble-st
Empire
probably Jewish owned or managed,
Most long-distance conta
trade in glass took place within
amounts of brokentheglass vessels
borders of the and win
Roman empire. Archaeological
dated to the fifth-sixth centuries. There is evidence
evidence suggests this trade was concerned primarily
for a dye shop on the lower floor.180 The reports dowith raw glass and fine tableware, though not exclu-
not specify whether the glass fragments were fromsively, as can be seen by the occasional recovery of
new or from used vessels. If the fragments were notprismatic bottles from shipwrecks.185 Glass was pref-
from broken glass assembled for recycling, theirerably shipped by sea since it was much more eco-
large number suggests they might represent one or nomical to transport goods over water than over
more deposits destined for wholesale (to local shops? land. The PE provides evidence for the cost of
for regional export?). Apparently, the two shops transport. A cart load of 1200 pounds (388 kg) cost
were located not far from the manufacturing facili-20 denarii per mile; a donkey load, probably of 200
ties.181 The glass vessels from shop E12 included pounds (65 kg), cost 4 denarii per mile (PE 17.3-5
"globular vessels, bottles, numerous glass lamps, gob-combined with PE 14.8-11). Based on these prices,
lets, concave vessel bases, and 350 window panes." A DeLaine has calculated the following average costs
"closet" in the stairway was filled with fragments ofof transport: 0.52 KM per ton per Roman mile for
glassware. Shop E13 yielded a total of about 4000 glass ox-carts, 0.12 KM per ton per Roman mile up-
fragments: 90% vessels, 10% window panes. The ves- stream, 0.059 KM per ton per Roman mile down-
sels included "two lamps, over 350 goblets, a salver, stream, and 0.012 KM per ton per Roman mile by
base-rings, a cup or bowl, cylindrical bottles, many sea.186 With a specific gravity of ca. 2.60, ancient
176 H.C. Youtie and J.G. Winter, Papyri and Ostraca from 181 Von Saldern (supra n. 9) 95.
Karanis (Michigan Papyri 8, 2nd ser., Ann Arbor 1951) no. 182 Crawford (supra n. 180) 78-79.
468, lines 15-18. 183 Gorin-Rosen 1998 (supra n. 9) 27-29.
177 P Oxy. vol. 4, no. 741.
184 The primitive furnaces in Egypt and Afghanistan
178 D.B. Harden, Roman Glass from Karanis (Ann Arbor were also oriented in such a way as to benefit from the
1936) 34-38. draft: Reut (supra n. 40), Henein (supra n. 40).
179 C.K. Williams and O.H. Zervos, "Corinth, 1982: East 185 Parker (supra n. 149) 197, no. 464, sunk ca. 200 near
of the Theater," Hesperia 52 (1983) 24-25, pl. 10: 64, 65. Grado carrying, inter alia, square bottles marked by C.
180sJ.S. Crawford, The Byzantine Shops at Sardis (SardisMon Salvius Gratus; see also supra ns. 33 and 83.
9, Cambridge 1990) 78-86; von Saldern (supra n. 9). The 186 DeLaine 210-11.
dye shop mentioned by Crawford p. 79.

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
pottery
soda-lime-silica glass weighed ca. 2,600 kg tableware,
per cubic lamps, and glass, these commodi-
meter.187 Weight would have beenties a primary con-
were never a major item of cargo on shipwrecks
discovered
cern in deciding which type of transport toinchoose
the Mediterranean.190 Nine wrecks of
for shipping raw glass if a choiceRoman
was ships
available.
carrying glass vessels in their cargo
With shipments of vessel glass, volume have been
and identified:
the fra- of these five date from the first
gility of the merchandise would have made
century,191 trans-
three from the second and third centu-
portation by boat attractive. ries,192 and two are of uncertain date.'19 All of these
In addition to their own cargo, most shipsshippers
sank in the also
western Mediterranean. The dates
are consistent
carried merchants with their wares, charging themwithpas-
an overall pattern of sea trade in-
sage money and for freight. The distinction
dicating that between
it was most intense in the late Republi-
shipowner (navicularius/naukleros), captain (magister/
can and early Roman Imperial periods.
kubernetes or pronaukleros), and merchant (mercator,
Most ships did not carry large amounts of vessel
glassagent
negotiator/emporos, pragmateutes) or his (on raw(pistikos)
glass see below), nor was glass the sole
was fluid: one man might fill all theseorroles at once.188
main cargo. The trade in glass vessels was proba-
Guilds, religious communities, and blyresident
handled by fellow
general merchants who took on indi-
countrymen, organized in stationes, assisted
vidual mem-
consignments of glass.'94 It is sometimes diffi-
bers travelling abroad. Shipping companies from aall
cult to distinguish merchant's goods'195 from his
over the Mediterranean had offices in Ostia;
personal logo-
property,'96 since in most cases the quan-
types announcing their names andtity specialties can is small and varied. Thus,
of glass they carried
still be seen in the mosaics of the Piazzale delle Cor- the status of the few and varied glass vessels exca-
porazioni. Many eastern Mediterranean cities were vated at Port Vendres and Mellieha Bay'97 is not
represented in the Roman forum; the Tyrians and clear. A basket filled with nine glass unguentaria,
Beirutians had offices in Pozzuoli. Whereas Syrians some still holding the remains of a cosmetic, may
played an important role in the early Roman empire, be evidence for an individual consignment on
Jews became increasingly visible in the fourth cen- board a ship that sank in the harbor at Fos-sur-Mer
tury.189 The active role played by eastern Mediterra- in the mid-second century.198
nean merchants in the long distance sea trade may Individual consignments were probably common
to some extent account for similarities between glass also in overland trade. At Kempten (southern Ger-
shapes made in the eastern and western part of the many) a small concentration of ca. 12 square glass
Roman empire. jugs among 89,565 kg of Rheinzabern Samian pot-
Although findspots indicate extensive trade in tery, ironwork, and bronze suggests a wholesale rather
187 I thank Fred E. Schaefer, Toledo, for this calculation, lot de verres du 1er siecle provenant du port de Narbonne
which is based on the fact that ancient glass was denser (Aude)," RANarb 25 (1992) 177-206.
(contained proportionately less silica) than modern soda- 192 Parker (supra n. 149) no. 464 (Grado); no. 691
lime-silica glass which has a specific gravity of 2.50. For (Mellieha); and no. 906 (Procchio).
comparison, the specific gravity of quartz (rock crystal) is 193 Parker (supra n. 149) no. 530 (La Jaumegarde A)
2.65, marble 2.72, and oakwood 0.75. and no. 614 (Maddalena).
188 J. Rouge, Recherches sur l'organisation du commerce mari- 194 On mixed cargoes and numbers of merchants sailing
time en M&diterran&e sous l'empire romain (Paris 1966) 214- on Roman ships: G.W. Houston, "Ports in Perspective:
94; A.H.M. Jones, The Later Roman Empire 284-602 (Ox- Some Comparative Materials on Roman Merchant Ships
ford 1964, repr. 1973) 866-71. and Ports," AJA 92 (1988) 553-64, esp. 558.
189 Rouge (supra n. 188) 302-19. On trade with Britain: 195 Ship wrecks containing glass vessels thought to be
J. du Plat Taylor and H. Cleere eds., Roman shipping and commercial consignments rather than personal properties
trade: Britain and the Rhine provinces (Council for British Ar- of travellers were found, inter alia, at Antikythera (late Hel-
chaeology Research Report 24, 1978). lenistic): G.D. Weinberg, Glass Vessels in Ancient Greece (Ath-
190 Parker (supra n. 149) 16-17. ens 1992) 28-33; la Tradeliere, Narbonne, and Cavallo I
191 Parker (supra n. 149) no. 283 (Cavallo A); no. 584 (ca. 50-60): supra n. 191; Serge Limani (11th c.): G.F.
(Lavezzi A); no. 875 (Port Vendres B; more on the glass in Bass, "The Nature of the Serge Limani Glass," JGS 26
A.J. Parker and J. Price, "Spanish Exports of the Claudian (1984) 64-69.
Period: The Significance of the Port Vendres II Wreck Re- 196 Parker (supra n. 149) lists 21 wrecks of the Roman
considered," IJNA 10.3 [1981] 221-28, esp. 224-27); no. and early Byzantine periods containing glass thought to
1174 (La Tradeliere; on the glass from this wreck, see also have been used on board.
M. Feugere and F. Leyge, "La cargaison de verrerie au- 197 Supra ns. 191, 192.
gustecnne de l'6pave de la Tradeliere [Iles de Lerins]," in 198 Parker (supra n. 149) 373-74, no. 1002 (Saint Ger-
M. Feugere ed., Le verre preromain en Europe occidentale vais C).
[Montagnac 1989] 169-76). Narbonne: M. Feugere, "Un

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
were packed
than a retail depot.199 The in a hardwood
remainsbox.203 Basketry ofand pa- wh
to pyrus were contingent
have been a wholesale the preferred packaging materials
of glas in
excavated at Augsburg. Dated
Egypt, while straw was more tocommon the Flavia
in the West.204
the finds include over
In some cases,600 individual
the glass vessels themselves contained g
representing some 40 the merchandise,
shapes for example,
and the unguentaria
subtypes from
sels are thought to For-sur-Mer
comeand from
perhaps some one
prismatic bottles.
large It s
glass from north Italy (perhaps
has been suggested that some large lateAquilei
Roman and
early Byzantine
damaged before it could enterglass containers
the served to transport ci
retail
wine and fish
The practice of taking on sauce.205
individual cons
is similar to business Thepractices
direction of the trade inin Roman medieva
glass vessels
ranean trade and commerce. Numerous Geniza is an intriguing question. While it is generally as-
documents and business letters record the workings
sumed that most trade in the first century went from
of a trade based on individual consignments.the
Timing
Syro-Palestinian coast to the North Pontic cities
was of the utmost importance. Business letters oftenthe possibility of trade in the opposite di-
and Italy,
ended with lists of that day's prices for a wide range
rection, from Italy to the Syro-Palestinian coast, is also
of goods. Vendors aimed to be first on the market.
very likely. The dominance of Italy and the West in
Countless letters advise the addressee to send a com- early glassblowing suggests glasses made in the West
modity "with the very first ship sailing" or to sell im-were exported to the East,206 not only to the coasts of
mediately upon arrival. Business relationships in-the Black Sea where the Romans had a foothold207
cluded "friendship" and other forms of informalbut even to Syria and Palestine. Anomalous finds
cooperation that might last for a lifetime, partner- such as several modioli and a large, one-handled
ships and commenda (in principle for short termsquat cylindrical bottle208 were probably personal be-
duration and limited to specific undertakings), fam- longings that do not represent regular trade. For
ily partnerships, commissions and agencies.201 other shapes the question can only be addressed
Adequate packing was important for glass vessels. when more data is available and reliable distribution
Egyptian papyri from the Roman period mentionmaps have been made. Western-made glass vessels
glass vessels packed in a cavea gallinaria "chickenhave come to light in the East in surprising numbers,
coop," a procheirion "handbox," and a panarion "bread such as the many Hofheim cups (see fig. 14).209
basket."202 Archaeological evidence confirms the prac- In many cities, the presence of a high quality local
tice. The nine glass unguentaria from Fos-sur-Merglassblowing industry appears to have created de-
were transported in a basket; the second-to-thirdmand for imported fine wares. Excavations at sites
century glass vessels from the wreck in Mellieha Bayknown to have had a thriving glass industry frequently
199 M. Rhodes, "Roman Pottery Lost en route from the 206J. Price, "Glass Tablewares in Use in Mytilene (Les-
Kiln Site to the User," Journal of Pottery Studies 2 (1989) 44- bos) in the 1st c. A.D.," AnnAIHV 14, 1998 (forthcoming)
58, esp. 45, 53, 54; on the bottles and other glass finds notes that many groups are similar to the glass excavated
from Kempten: P. Fasold, "Die friih- und mittelromischen at Frejus.
Glaser von Kempten-Cambodunum," in Forschungen zur 207 N. Sorokina, "Das antike Glas der Nordschwarz-
provinzialr6mischen Archdologie in Bayerisch Schwaben (Schwia- meerkfiste," AnnAIHV 4 (Liege 1967) 67-79, esp. 77; So-
bische Geschichtsquellen und Forschungen 14, Augsburg rokina, "Facettenschliffglfiser des 2. und 3. Jhts. u. Z. aus
1985) 197-230, square bottles, pp. 200, 206-208, 218-22, dem Schwarzmeergebiet," AnnAIHV 7, 1977 (1978) 111-
no. 43, figs. 14, 15. 22, esp. 122; Sorokina, "Glass Aryballoi (First-third Cen-
2() For a preliminary discussion of the finds: Rottloff turies A.D.) from the Northern Black Sea Region," JGS 29
(supra n. 33) 166-70. (1987) 40-46, esp. 43.
201 Goitein 164-86.
208 Modioli and squat cylindrical bottle from Syro-Pales-
202 Youtie and Winter (supra n. 176) no. 468, lines 15- tinian findspots: Stern (in prep.). Five small square bottles
18; P Oxy., vol. 4, no. 741, lines 14-15, and P Oxy., vol. 10, and two small cylindrical bottles from tombs at Castra look
no. 1294, line 6 respectively. very similar to those made in the West: Y. Gorin-Rosen, An-
203 Parker (supra n. 149) nos. 1002 and 691 respectively. cient Glass from the Holy Land, exhibition catalogue, Fine Arts
204 E.g.: C.C. Edgar, Greco-Egyptian Glass, Catalogue genral Museum of San Francisco/Israel Antiquities Authority
des antiquitis Egyptiennes du Mus&e du Caire (1905, repr. (1998) 20; Y. Gorin-Rosen, presentation at 14th congress
Osnabrfick 1974) nos. 32.655; 32.656; 32.661; F.J. Hassel, of the Association Internationale pour l'Histoire du Verre,
"Glasamphore im Deckelkorb," JRGZM 33 (1986) 908-909, Venice/Milan 1998.
fig. 94; de Franciscis (supra n. 151). 209 Isings Form 12. On Hofheim cups in general: Cool
205 L. Taborelli, "Indagine preliminare sui contenitori and Price 64-68; on eastern Mediterranean finds: Stern
in vetro per trasporto e la conservazione del vino e del (in prep.).
garum," Opus 12-13 (1993-1994) 1-23.

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
run.212
yield more imported glass vessels than sitesGlass
wherefindsno
from later shipwrecks are evi-
glass was blown. Pompei and Aquileia arethat
dence good exam-ingots remained the norm for
preformed
ples of cities in Italy where local glassworking
trade in colored glass andduring the Roman empire.213
Natural bluish-green glass became common in the
imported glass vessels are attested in quantity.
Unlike the finished glass object whose potential
late Hellenistic period, although it was a trade good
buyers were numerous and locatedasthroughout
early as the thirdthe century B.C.214 The bulk density
empire and beyond, raw glass was destined
of glass is higherfor a
than that of sand or clay, which was
small, select number of clients whose workshops
commonly used asspe-
ballast. Sold in the form of amor-
cialized in the production of luxury phousor utilitarian
chunks, the raw glass could be transported by
wares. It has recently been suggested sea at almostthe
that no cost because it doubled as ballast. A
mer-
first-century
chants who provisioned the workshops with raw shipwreck
glassexcavated off the Croatian
might have done so in return for a coast
specified part of
at Mljet yielded more than 100 kg of natural
the production as was to become customary much
bluish-green raw glass, dispersed throughout the
later in medieval France.210 In a situation described area of the wreck.215 Chunks of transparent blue-
in the Babylonian Talmud (third century), Rabbi green glass were discovered off the Israeli coast at
Huna explains the different Sabbath rules for when Apollonia/Arsuf and Carmel Beach. Whereas the
a merchant needs to unload finished vessels and glass off Carmel Beach is thought to be from a third-
chunks of raw glass from his donkey's pack.21' Thecentury merchant ship "sailing between the shores
Talmud citation shows that one merchant might of Lebanon and Alexandria," glass was also made lo-
bring raw glass and/or cullet to the glassblower cally
and at Apollonia/Arsuf.216
leave with finished vessels. There is, however, no evi-
Once recycling became common, cullet could be-
dence of a formal business arrangement for this type
come part of the ballast, preferably mixed with raw
of transaction either in the Roman period or in glass
the because cullet has a lower bulk density than
Geniza documents.
raw glass. An Islamic ship that sank at Serge Limani
Archaeological evidence for long distance carried
sea 3 metric tons (3,000 kg) of glass "both in the
trade in raw glass reflects the important changes that
form of chunks of raw glass and broken vessels" (cul-
took place in glassmaking and glassworking at differ-
let) as ballast in its hold.217 For comparison, a Geniza
ent periods in history. Up to the late Hellenistic
document dated 1011 mentions 37 bales of glass (at
period, almost all glass objects were made of inten- about 227 kg each) sent by three Jewish firms from
tionally colored or decolorized glass. Expensive Tyre,
and that is, a total of more than eight metric tons,
produced in small quantities, colored glass was sold (8,399 kg, to be precise).218
in the form of preformed ingots. Long distance trade It is not clear how the transportation of this much
in ingots is illustrated by the carefully packed cakes of as a ship's ballast was organized, and whether
glass
colored glass excavated in a Bronze Age ship that
the initiative was with the glassmaker, the buyer, or
sank off the southwest coast of Turkey, at Ulu Bu-
the merchant/ship's owner. Perhaps the latter sold
210 Nenna et al. (supra n. 8) 86. chunks of glass and both dated third century B.C.
211 Weinberg 25, n. 2. 215 I. Radic and M. Jurisic, "Das antike Schiffswrack von
212 G.F. Bass, "A Bronze Age Shipwreck at Ulu Burun Mljet, Kroatien," Germania 71 (1993) 113-38, esp. 122-23.
(Kas): 1984 Campaign," AJA 90 (1986) 269-96; Bass, On ingots and chunks of raw glass from land and under-
"Splendors of the Bronze Age," National Geographic 172 water excavations in the western Mediterranean, see also
(1987) 693-733; Parker (supra n. 149) 439-40, no. 1193. Picon (supra n. 49) and D. Foy, "Archeologie: Une epave
See also R.J. Charleston, "Glass 'Cakes' as Raw Material chargee de lingots et de vaisselle de verre," Verre 3.3 (1997)
and Articles of Commerce," JGS 5 (1963) 54-68; D. Barag, 65-70.
Catalogue of Western Asiatic Glass in the British Museum (Lon- 216 E. Galili et al., "Underwater Surveys and Rescue Ex-
don 1985) 107-110; Stern 1999 (supra n. 109). cavations along the Israeli coast," IJNA 22.1 (Febr. 1993)
213 Ingots of raw glass: Parker (supra n. 149) 221, no. 61-77, esp. 65 (Apollonia/Arsuf) and 70 (Carmel Beach,
530 (ca. 100-25 B.C., blue glass ingots) and 274, no. Haifa). Glassmaking at Apollonia Arsuf is mentioned by
691 (ca. 200-250, cakes of glass, blue frit). See also D. Gorin-Rosen (supra n. 208) 15. Two large chunks of blue-
Foy and M. Picon, "Lingots de verre en M6diterran6e oc- green glass from Apollonia/Arsuf, shown in the exhibition
cidentale (3e siecle avant J.-C.-5e siecle apres J.-C.)," Ann- (supra n. 29), are labeled "second century BCE-first Cen-
AIHV 14, 1998 (forthcoming). tury CE."
214 M.-D. Nenna, "Les ateliers de verriers dans le monde
217 Bass (supra n. 195) 64-69, esp. 64; Parker (supra n.
grec aux 6poques classique et hell6nistique," Topoi 8 (1998) 49) 398-99, no. 1070.
693-701, esp. 696: shipwrecks Sanguinaire A (near Ajac- 218 Goitein 421, n. 65; on the weight of a bale, ibid. 335.
cio) and Lequin 2 (Ile de Porquerolles), both carrying

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
a significant
the glass, upon arrival inrole, as suggested to
port, by the middlem
large number
of Roman glass cupsSmall
resentatives of workshops. and beakers excavated
amounts in settle-
ments and
ized, colored raw glass graves in Scandinavia
could perhaps and north-central
be or
Europe.221partner
rectly from a business It is not clear to whatnear
extent the glass
the ves- f
sels from Scandinavia, north-central
transported as one merchant's consignmen Europe, and
northern Britain are evidence of regular (barter?)
trade patterns or individual gift exchanges.222 The
Trade beyond the Frontiers
glass fromtrade
Before speaking about Scandinavia and Germania libera consists
beyond the f
it is useful to remember
almost exclusively of that
luxury drinking"Roman
vessels. This
fixation
catchall term for glass on one function
made appears consistent
during the withfir
a
fourth centuries. It greater
includes objects
pattern of exchange made
that included a similarly
borders of the Roman empire
specialized array of importedand beyond,
metal wares, the major-
ple, in Mesopotamia ityand
of whichthewere largeHellenized
buckets and basins. The im- cit
north coast of the Black
ports Sea.
included very little "Roman
pottery. The limited num- g
come to light in excavations far
ber of functions associated withbeyond
the imports from thethe
in north, central, Roman empire
and suggests to me thatEurope
southeast these objects
rica, India, and eastern Asia.
were indeed part of an organized trade pattern in
In western Europe, which
the Roman
a specific military
demand determined a specific sup- p
important role in introducing glass
ply. One is reminded of the seemingly and
exotic copper gl
to the provinces. Roman legionaries,
cauldrons stamped "made in Germany" thatcompo
were of-
nic groups from all fered
corners of
for sale in markets theGreece
throughout empire
in the
the borders along the Rhine
1960s. The and glass
demand for imported Donaudrinking ves-riv
legions were transferred they
sels persisted after the demisewere
of the Romanusuall
empire
in the West.
nied by craftsmen and Excavations have yielded
artisans, numerous
including
ers to produce windowglass
fourth-to-sixth-century(appreciated
beakers of a high quality, col- in
mate) and supply soldiers
orless glass thatwith
was rare in tableware.
the (former) Roman
the expansion of the Roman
provinces empire
of Europe. The in of
shapes and decoration the
tury saw important these
glass centers
glasses became spring
increasingly "unRoman." They up
provinces northwestmayandhave beennortheast of glass
made in an as yet unidentified Italy
diers came from the eastern
center outside the formerMediterranea
Roman empire.223
glass was part of the One of the most surprising findspots
instrumentum of Roman
domestic
glass services imported from
glass is the Libyan desertItaly
(the Sahara). graced
Italian exca-
ers' tables.219 The vations
civilian population
unearthed several deposits of glass in the in
demanded the accoutrements of Roman life. Glass area known as Fezzan.224 Among the earliest glass
tablewares became symbols of culture as well as re- were fragments of ribbed bowls and other vessels
alistic indicators of wealth and social diversity.220 from a mausoleum at Germa (perhaps ancient Ga-
Diplomatic and commercial contacts with peoples rama) that also yielded fine pottery dated to the late
living beyond the borders in Germany, Scandinavia, first century. Associated with the burial of a nonlocal
and central Europe created further interest in Ro-person, these objects probably reached the site in
man utilitarian and luxury products. In this cultural the wake of an expedition to the Garamantes that
and economic exchange, glass drinking vessels playedtook place in the Flavian period.225 Most Roman
219 On the important role of the Roman army in intro- gration Periods. A Study on Glasses Found in Eketorp-II, Oland
ducing glass to western Europe: S.M.E. van Lith, "First-cen- Sweden (Uppsala 1984); E. Straume, Gldser mit Facettenschliff
tury Cantharoi with a Stemmed Foot: Their Distribution aus skandinavischen Grdbern des 4. und 5. Jahrhunderts n. Chr.
and Social Context," in Two Centuries 99-110; Stern 1995, 96. (Oslo 1987).
220 Van Lith and Randsborg 437-45. 224 G. Caputo, "Scavi Sahariani: ricerche nell' Uadi el-
221 H.J. Egger, Der r6mische Import imfreien Germanien (At- Agial e nell' Oasi di Gat," Mem. Accademia Nazionale dei Lin-
las der Urgeschichte 1, Hamburg 1951); U. Lund Hansen, cei 41 (Rome 1951) 151 ff. (not available to me); EAA 6.
R6mischer Import im Norden. Warenaustausch zwischen dem Ri-1014-15, s.v. Romana Arte (F. Coarelli).
mischen Reich und demfreien Germanien (Copenhague 1987). 225 J. Desanges, Recherches sur l'activite des Miditerrandens
See also infra n. 223.
aux confins de l'Afrique (Ecole Francaise de Rome 1978)
222 On the difficulty of interpreting the evidence: Harris 197-211 describes two Roman expeditions that passed
(supra n. 66) 15, 16. through the area under the Flavians.
223 U. Nisman, Glass and Trade in the Late Roman and Mi-

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
glass from Fezzan was excavated in monumental

trade good, both as finished objects and raw glass.
tombs in the vicinity of the mausoleum. The first-century
An anonymous glass manuscript written in
vessels included a one-handled cylindrical bottle
Greek gives precise instructions on the commodities
decorated with incised geometric designs, two
that could coni-
be sold and bought along two sea routes,
cal beakers decorated with large dotsone(glass scholars
following the East coast of Africa as far south as
call them blue 'blobs'), a shallow bowlMadagascar,
with a wheel-the other using seasonal monsoon
cut inscription, and numerous engraved
windsfragments,
to cross the ocean to India. The manuscript,
all of which date to the fourth century. TheMaris
Periplus same date (hereafter PME), is a com-
Erythraei
has been proposed for the glass excavated
bined navigationkm
ca. 360 guide and merchant's compen-
farther southwest in the cemetery of dium. FromThe
Gat. this text we learn that several sorts of
most
spectacular find here was a painted beaker.glass
multicolored These
objects, thought to be mosaic
fourth-century finds may suggest (temporary?) trade.
glass inlays, were exported to the north coast of So-
Many sites in Sudan, south of Egypt,
mali,yielded
glass vesselslux-
to north India, and raw glass (hya-
ury glass objects dating to the latelos
Hellenistic
arge "unworkedand glass") to south India. Excava-
Roman Imperial periods.226 Although some
tions at various of
sitesthe
in these areas brought to light
glass was probably local,227 much wasexactly
imported
the kindsfrom
of glass mentioned in the PME.231
or through Egypt along the Nile. When theroutes
The trade Romansmentioned in the PAME began in
annexed Egypt the kingdom of Meroe remained
Egypt, in-
where some-but not all-of the glass ex-
dependent, but relations between ported
Egypt to theand itshave been produced. Several
East may
southern neighbors remained closeReduntil
Sea portsthe sec-
took part in this trade; the most im-
ond half of the fourth century.228 Meroe
portant was an im-and Myos Hormos (recently
were Berenike
portant trading center through which many
identified African
as Quseir al-Qadim).232
luxury goods were channeled to the Not Mediterranean
all exotic findspots are necessarily evidence
and beyond. Many fine glass vessels that found
for regular their
trade in glass. The fragments of glass ves-
way to Meroe may have played a role in this
sels excavated ex- on the east coast of south
at Arikamedu,
change.229 A similar exchange took place
India, withtothe
are thought represent personal possessions
ancient kingdom of Aksum through rather than items
its port Adulisof trade.233 Some of the most in-
on the Red Sea.230
triguing findspots in this connection are the island
In East Africa and India, Roman glass was a regular
of Bahrain234 and ed-Dur (Umm Al-Qaiwain, United
226 D. Dunham, Royal Cemeteries of Kush 4: Royal Tombs at Text with Introduction, Translation and Commentary (Prince-
Meroe and Barkal (Boston 1957) and 5: The West and South ton 1989). On the economics of the trade: S.E. Side-
Cemeteries at Meroe (Boston 1963); J. Leclant, "Glass from botham, Roman Economic Policy in the Erythra Thalassa 30
the Meroitic Necropolis at Sedeinga," JGS 15 (1973) 52- B.C.-A.D. 217 (Leiden 1986).
68; R. Brill, "Scientific Investigations of Some Glasses from 232 S.E. Sidebotham and W.Z. Wendrich eds., Berenike 1994.
Sedeinga,"JGS 33 (1991) 11-28. Preliminary Report of the 1994 Excavations at Berenike (Egyptian
227 Stern 1979 (supra n. 14) 46, 47 (locally made inlays); Red Sea Coast) and the Survey of the Eastern Desert (CNWS,
H.E.M. Cool, "Sedeinga and the Glass Vessels of the King- Leiden 1995); Sidebotham and Wendrich's 1996 Report in-
dom of Meroe," AnnAIHIV 13, 1995 (Lochem 1996) 201- cludes a preliminary report on the glass finds by P.T. Nichol-
12 (vessels).
son, 279-88. On the Roman glass from Quseir: Meyer 1992
228 Desanges (supra n. 225) 307-66. (supra n. 231). On the identification of Quseir al-Qadim as
229 Cool (supra n. 227) 211. Myos Hormos: A. Billow-Jacobsen et al., "The Identifica-
230 H. Morrison, "Glass and Trade of the Ancient Ak- tion of Myos Hormos. New Papyrological Evidence," BIFAO
sumite Kingdom," AnnAIHV 9, 1983 (Liege 1985) 113- 94 (1994) 27-42; D.P.S. Peacock, "The Site of Myos Hor-
126; Morrison, "The Glass," in S.C. Munro-Hay, Excava- mos: A View from Space," JRA 6 (1993) 226-32.
tions at Aksum (London 1989) 188-209. 233 Stern 1991 (supra n. 231); Stern, BibO 52 (1995)
231 On the glass mentioned in the PME: E.M. Stern, 833-40 (with lit.). In the medieval Indian trade, glass ves-
"Early Exports Beyond the Empire," in Two Centuries 141- sels feature primarily as household items, not as trade
54; Stern, "Early Roman Export Glass in India," in V. Beg- goods: S.D. Goitein, Letters of MedievalJewish Traders (Prince-
ley and R.D. De Puma eds., Rome and India: The Ancient Sea ton 1973) 185-92, letter 38, sent from Aden to India in
Trade (Madison 1991) 113-24; C. Meyer, Glass from Quseir 1139; the glass appears in Section D "household goods."
al-Qadim and the Indian Ocean Trade (Oriental Institute of
On long distance trade in the early Middle Ages, see also
the University of Chicago Studies in Ancient Oriental Civi- Goitein and Meyer 1992 (supra n. 231) 98-103.
lization 53, Chicago 1992) 43-74;J. Desanges et al., Sur les 234 E.C.L. During Caspers, The Bahrain Tumuli An Illus-
routes antiques de l'Azanie et de l'Inde (Memoires de l'Ac- trated Catalogue of Two Important Collections (Uitgave van het
ademie des Inscriptions et Belles-Lettres N.S. 13, Paris Nederlands Historisch Instituut te Istanbul 47, Leiden
1993). On the PME: L. Casson, The Periplus Maris Erythraei. 1980).

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
Arab Emirates) on the Arabian Gulf. Both loca- Iraq." Moreover, much of the glass from ed-Dur is sim-
tions yielded first-century ribbed bowls and other ilar to that found at Dura Europos.
fine glassware. The finds from the Belgian excava- Other important findspots of Roman glass in Asia
tions at ed-Dur are of particular importance be- include Taxila and Begram.237 The glass finds from
cause they were recently unearthed during con- Begram are so diverse that individual groups of
trolled excavations.235 The catalogue of the glass glasses have been assigned to periods 200-300 years
vessels includes 122 objects ranging from 25 B.C. apart, from the first to the third/fourth centuries.
to A.D. 75. The manufacturing techniques include This is not the place to enter the controversy, how-
casting, mold-forming, free-blowing, and mold- ever one unique group of Begram glasses, decorated
blowing; decorative techniques include mosaic glass, with applied wavy coils for which no comparisons were
tooled vessels, flecked wares, threaded, and wheel-cut known until recently, now has an exact parallel exca-
vessels. The finds also include fine tableware (drink- vated in the vicinity of Padua in a tomb assigned to the
ing vessels, jugs, a plate), luxury containers (a mold- late first century. If that date is correct, all the glass
blown pyxis, polychrome and mold-blown bottles) as from Begram may well date from the first century.238
well as utilitarian bottles and unguentaria. Most of Roman and Sassanian glass travelled as far as
these vessels were imported from the Roman empire, China, Korea, and Japan. While some of the finds
but some of the utilitarian bottles could have been may have reached these areas by sea, through India
made in Parthia. and Sri Lanka, others may have travelled overland in
Do these vessels represent items of regular trade a camel's load.239
or were they the personal belongings of Roman or
Mediterranean traders who settled in the area? In GLASS IN A SOCIAL CONTEXT
view of the well documented glass trade with India,Differences in quality, size, and workman
and the fact that smaller assemblages of Roman glass
proof that Roman glass vessels, like textile
have come to light at other sites on the Arabian Gulf,
the entire spectrum from expensive luxury
it is tempting to interpret the glass from ed-Durthe
as simplest of utilitarian goods. The archa
trade goods. Three trade routes have been consid-
evidence is complemented by anecdotal
ered: 1) by sea from an Egyptian port like Quseir sources.
al- While some drinking cups were so
Qadim; 2) the overland route from Syria to the Eu-sive they could be bought "for a copper
phrates, passing through Charax at the head of the16.2.25), two particular glass drinking cups
Arabian Gulf; and 3) export from Egypt to India erate size sold for 6000 sesterces in the time
from where it was reexported to the Gulf.236 The (Plin. HN 36.194).
most compelling reason for suggesting the glass wasA detailed analysis of individual topog
exported first to India and then reexported from In-
units at Augst showed that the quantity of
alone cannot be used as a definitive criterion to de-
dia to the Gulf: the PME scarcely mentions the Gulf,
whereas Indian ships are known to have frequented termine social status. Large numbers of fine table-
the area. Similarities between the glass excavated wares
at excavated in middle-class houses "contrasted
Quseir al-Qadim and the Arabian Gulf suggests much sharply with the real rarity of finds in the luxurious
of the glass could have reached the Gulf through the
residences of the upper class." This did not mean
Red Sea. However several observations suggest that"that
at the socially well-off disdained the use of table-
least some of the glass came overland: the presenceware
of glass";240 the most expensive tablewares, made
Parthian glass bottles, the presence of foreign coins
of mosaic glass and colorless vessels with wheel-cut
minted at Charax, and the fact that "40% of the diag-
decoration, were found predominantly in the better
nostic pottery from ed-Dur was made in southern
residential quarters. The highest concentration of
235 D. Whitehouse, Excavations at ed-Dur 1: The Glass Ves-

burg 1996), review by D. Whitehouse, AJA 102 (1998)
sels (Leuven 1998). 639-41. The glass beaker from Padua has not yet been
236The following discussion summarizes Whitehouse published.
(supra n. 235) 65-67. 239 E.J. Laing, "A report on Western Asian Glassware in
237J. Marshall, Taxila (Cambridge 1951) 683-90. P. the Far East," Bulletin of the Asia Institute 5 (1991) 109-20
Hamelin, "Verreries de Begram," CahByrsa 2 (1952) 11- provides a readily accessible survey of foreign glass vessels
25; 3 (1953) 121-28; 4 (1954) 153-83. excavated in China, Korea, andJapan. See also: E.M. Stern
238 On the date of the glass from Begram most recently: 1995 (supra n. 233) 836 with n. 11; E.R. Knauer, The
M. Menninger, "Untersuchungen zu den Glisern und Gip- Camel's Load in Life and Death (Kilchberg 1998) 117-21.
sabgiissen aus dem Fund von Begram (Afghanistan)," (Wiirz- 240 Riltti 342.

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
windowglass was found in public buildings and elegantiae

Nero's arbiter the Petronius (Sat. 5
houses of the wealthy. Changes of"I prefer
usage glass-that's
over time got no taste at
were observed in individual houses and insulae.241
didn't break I'd prefer it to gold."
The material's
Most of the uses for blown glass known until the impermeability and cle
19th century were explored during the first
probably the 100
reasons glass became the p
terial for
years of commercial glassblowing, roughly large utilitarian containers. F
coincid-
ing with the first century A.D. Glass ond
was half
often of the in
used first century glass stor
antiquity for purposes that, since thegan to compete
Middle Ages, with clay amphoras for
perhaps
have been commonly served by porcelain, a transport
material of liquids and solids.2
unknown in the Roman empire.242 The Glass was
last non-
major contribution of glassbl
ciety wasTrans-
porous, it did not contaminate the contents. the development of new ve
parent glass even allowed inspection through
specifically the
for lighting purposes.246 Unli
wall of a closed container. Although ancient
which werelitera-
filled entirely with oil, gla
ture contains many references to filledthe functions
with water ofupon which the oil and
glass vessels, it is not always easy toby a wick
assign holder) floated. The choice
a specific
function to each individual type of vessel.243
pended on the type of light that was actu
To judge from excavated finds, glass vessels
Light be- studies show that glass
emission
came widely accessible to all levels of society
the during
advantage of burning longer and al
Tiberius's reign (A.D. 14-37). The brightly
physician as Scri-
terracotta lamps. In terracot
bonius Largo, active in the time type of Claudius (41- no difference. Filled
of fuel made
54), mentioned glass containers for oilmedicines as a
and a V2 cm-wide cotton twine wick
matter of course.244 Small bottles for scented
lamp oils,a steady candle power of
produced
cosmetics, and medicines were amonghours the earliest
and 23 minutes. In glass lam
blown glass vessels. Tableware was proved
a second area
to be of satisfactory than oliv
more
early usage. Initially, blown-glass tableware
lamp filledserved withthe350 ml water and 100
burned with
same purpose as high quality Hellenistic glassa smokeless
table-flame for 2 hours and 45
ware: drinking and serving. Very minutes,
soon pouring
needing only be-
one adjustment of the wick.
came a major function. The great Roman tradition
The output ofpower of 1.36.248 Pliny at-
was a candle
glassjugs, flasks, and decanters begantests
before
the usethe mid-
of castor oil in antiquity (HN 15.25-26)
dle of the first century. Some of theclassifying
earliestjugs
it as anwere
oleum ficticium "artificial oil." He
high quality mold-blown vessels made by Ennion.
described two different methods for extracting the oil
In 70, Pliny (HN 36.199) noted that glass
from the table-
kiki plant, which grew wild in Egypt. Writing
ware had almost ousted silver andlong
gold. The
before prob- of glass lamps in which the
the invention
lem may have begun earlier, under oil
Tiberius,
floated on who
top of is
water, Pliny had no high opin-
said to have destroyed the workshop
ionofof a glassblower
castor oil. He called it disgusting for food and
in order to avoid a devaluation oflucernis
precious
exile "ofmetals
thin quality for burning in lamps."
(HN 36.195). Glass continued to graceThethe
glass tables of to have been an eastern
lamp appears
the wealthy into the fifth century because the mate-
Mediterranean, perhaps Syro-Palestinian, innova-
rial had the one great advantage thattion.249
it didThe
not affect
earliest glass lamps were flat-bottomed
the taste of wine, as metal does, an hemispherical
advantage which
bowls with small loop handles that
241 Riitti 170-287.

width of ca. 4.5 cm, these bottles appear to have been rela-
242 In 1134, a religious authority in Cairo was asked tively small.
"whether translucent Chinese porcelain could be regarded 246 On glass lamps: G.M. Crowfoot and D.B. Harden,
as glass for ritual purposes" (Goitein 421, n. 67). "Early Byzantine and Later Glass Lamps," JEA 17 (1931)
243 On the functions of glass vessels mentioned in an-196-208; D.B. Harden, Roman Glass from Karanis (Ann Ar-
cient literature: Trowbridge (supra n. 3) 150-93; W. Hil-bor 1936) 155-64; E.L. Higashi, Conical Glass Vessels from
gers, Lateinische Gefiissnamen (Dusseldorf 1969) passim. Karanis (Diss. Univ. of Michigan 1990); S. Hadad, "Glass
244 Taborelli (supra n. 171) 148-56. Lamps from the Byzantine through Mamluk Periods at Bet
245 The contents of a square bottle from Pompei provedShean, Israel,"JGS 40 (1998) 63-76.
to be vegetable oil: Scatozza et al. (supra n. 145) 557-58. 247 Higashi (supra n. 246) 378-79 notes that at Karanis,
Cylindrical and prismatic bottles have recently been re-glass and clay lamps were often found in the same room of
viewed by Cool and Price 184-85. The earliest square bot-one house.
tles come from Cosa and the Port Vendres II wreck, both 248 Higashi (supra n. 246) 380-81.
of which are firmly dated early to mid 40s. With a base 249 Stern (in prep.).

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
Fig. 29. Synagogue, Ham

menorah with burning gl
Weinberg pl. 4:C)
Fig. 28. Conical lamp. H. 17.1 cm, wt. 317 g. Se
of fourth century. No pontil scar. Probably P
(Courtesy Ernesto Wolf tures known as polycan
Collection)
illumination. According
century), Roman church
could serve for individual
and gold suspension,
coronae (poly a
cones that could be mounted inthe
as early as holders (fig
pontificat
lamps inserted
A great technical improvement was in these
the inv
made
the late fifth or early sixthof precious
century ofmeta
the
with glass.252
a built-in internal The earliest
glass wick gl
holder; th
lamp was not common notin the West.
predate the late fourt
Glass lamps were not widely used before t
ond half of the fourth century.
Forms By the end
and Function of of
Gl
century shapes In were
various the West, common the in secon m
tries bordering the was Mediterranean
by far the most and pr B
They could be used architecturally
tity, variety of for forms, draa
vessels. This has been demonstrated
phasis. The water collected for Britain and
and intensified
of the flame so thatthe theEuropean continent.254 vessel
entire This may have glowe
also been
true multiple
was lit.250 In churches of the Pontic cities, but comparable data
lamps, set are not
in
250 On lighting effects: C. Steckner, "Pharokantharoi candela (10th century): F. Valdes Fernandez, "Kalifale
und Kylikeia," AnnAIHVIt11, 1988 (Amsterdam 1990) Lampen," 257- MM 25 (1984) 208-15, pl. 71.
70. 253 M. Uboldi, "Diffusione delle lampade vitree in eta
251 H. Geertman, "L'illuminazione della basilica paleo- tardoantica e altomedievale e spunti per una tipologia,"
cristiana secondo il liber pontificalis," RACrist 64 (1988) Archeologia Medievale 22 (1995) 93-145.
135-60. 254 Van Lith and Randsborg; Cool and Price 221-23.
252 At Madinat az-Zahra, ceramic lamps were set in poly-

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
Table 2. Numbers of Vessel Forms from Colchester Grouped
Food Liquid Storage Total

Period Drinking Presentation Serving Storage Vessels for Number
(A.D.) Vessels Vessels Vessels Vessels Liquid only Other of Forms
ca. 43-60/1 5 5 3 - 5 2 20
60/1-ca. 100 6 5 4 3 5 1 24
100-170 3 2 4 5 5 - 19
170-230 4 2 3 5 5 - 19
230-300 2 1 4 - 3 - 10
300-350 4 1 5 - 1 - 11
350-400+ 4 2 4 - 1 1 12
After Cool and Price 223, table 13.1.
readily accessible for that area. In Palestine and Syria Beginning in the third century the forms of glass
as well as Egypt, the floruit of blown vessel glass was the storage vessels for solids and liquids became less var-
late Roman and early Byzantine periods (see below). ied. In the fourth century, the demand for variety in
In the first century glass vessels were used in the glass drinking vessels increased sharply. From then
West for a wide variety of purposes: as tablewares, on drinking vessels were to remain the vessels with
for storage and transport of solids and liquids, for the largest variety in forms throughout Frankish and
personal use (e.g., for cosmetics, scents, bath oils, medieval times and drinking was to be the main
and medicines), and various miscellaneous pur- function of glass vessels.256
poses. It is not clear why the functions served by
glass declined sharply during the second to fourth Forms and Function of Glass Vessels in the East
centuries. By the fourth century almost all the glass The most prolific period of glass production in
made in the northwest provinces was tableware, in the eastern Mediterranean was the late Roman pe-
particular for drinking and serving liquids. An analy- riod. Egypt, Palestine, Syria, Cyprus, Asia Minor, and
sis of the glass excavated at Colchester (table 2) ap- the north Pontic cities all had flourishing glass in-
pears to reflect a pattern of use that was broadly sim- dustries, and those of Syria and Palestine experi-
ilar across the western part of the Roman empire.255 enced a prolonged period of growth. Glassware
Whereas the number of forms for jugs and flasks played an increasingly important role in the daily life
remained approximately unchanged from the sec- of all levels of society, to such an extent that in the
ond century on, showing a slight increase with re- fourth century glass vessels all but ousted pottery for
spect to the variety available in the first century, certain functions.
there were fewer forms of drinking vessels to choose With increased demand regionalism became a
from, although there is no evidence to suggest that dominant factor in production. There are marked
the actual number of drinking vessels decreased. On
differences between the vessel glass made in Syria,257
the other hand, the reduced choice in dishes for Jordan,258 and Palestine, and also between different
serving foods does seem to have been accompanied parts of Palestine: inland versus coast, western Gali-
by a decrease in use of glass for this purpose. lee and southern Phoenicia versus Judea and areas
255 On the functions of glasswares available in northwest (forthcoming). E. Baumgartner and I. Krueger, Ph6nix aus
Europe: van Lith and Randsborg passim. On the glassSand und Asche: Glas des Mittelalters (Munich 1988); Foy and
from Colchester: Cool and Price 211-36. Compare also: B.Sennequier (supra n. 65); H.E. Henkes, Glass without Gloss:
Ruitti, Vitudurum 4: Die Gldser (Berichte der Zfircher Denk-Utility glass from five centuries excavated in the Low Countries
malpflege Monographien 5, Zfirich 1988) 109-23; Rutti1300-1800 (Rotterdam Papers 9, Rotterdam 1994).
170-264; M.R. DeMaine, "The Northern Necropolis at 257 C.W. Clairmont, Excavations at Dura Europos. Final Re-
Emona. Banquet Burials with Ladles," AnnAIHV 11, 1988 port 4, part 5: The Glass Vessels (New Haven 1963); S. Abdul
(Amsterdam 1990) 129-44; G. Sennequier, "Roman GlassHak, "Contribution d'une decouverte archeologique re-
Found in Upper Normandy,"JGS 36 (1994) 56-66. cente A l'etude de verrerie syrienne A l'epoque romaine,"
256 L. Sagui, "Produzioni vetrarie a Roma tra tardo-anticoJGS 7 (1965) 26-34. On the date of the tomb at Homs:
e alto medioevo," in L. Paroli and P. Delogu eds., La storiaStern 1977 (supra n. 76) 83-84.
economica di Roma nell' alto medioevo alla luce dei recenti scavi
258 0. Dussart, Le verre enJordanie et en Syrie du Sud (Insti-
archeologici (Biblioteca di Archeologia Medievale, Firenzetut Frani;ais d'Archeologie du Proche-Orient, Biblio-
1993) 113-36; Sagui, "Produzioni vetrarie a Roma tra V e theque archeologique et historique 152, Beirut 1998).
VII secolo: nuovi dati archeologici," AnnAIHV 14, 1998

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
farther south.259 Remains of glassworking

spite economic instability. Palestine benefited from
have been identified at several
Constantine's decision aftersites intheIsrael;
324 to target Holy
Land for his half
nace, active in the second building programs.
of the Exempted from
fourth
has been excavated atJalame.260
personal taxes by the imperial edict of 337, archi-
Palestine and southern Phoenicia
tects, painters, were
and sculptors, as well as selected cate- als
primary glass producers. Sixteen
gories of highly skilled rectang
craftsmen including glass-
furnaces, dating from
workers,theprofitedfirst
greatly fromhalfthe economic of and the
century, were excavated cultural boom.
at Glassblowers
Hadera, created an abundance
halfway
Tel Aviv and Haifa. Each furnace measured ca. 2 x 4 of new shapes and styles, many surviving into the Is-
m and had a capacity of 8-10 tons per firing. The lamic period.
process took up to two weeks and the temperature in New types introduced in the late Roman period
the furnaces reached ca. 11000C. The furnaces ap- include: glass kohl tubes (fig. 30) widely used in Pal-
pear to have functioned only for a short time, per- estine,265jars (fig. 31) used in Palestine and Syria but
hapsjust one glassmaking season, until fuel supply in with different styles, and sprinklers (fig. 32) used in
the vicinity was exhausted. Very little glass remained Syria. Closely associated with cultural traditions of
at the site, suggesting all of it was sold.261 Syria and Palestine, these types remained largely un-
The rise of the Severan dynasty (193-212) known in the West. Also new in this period was the
brought a higher level of prosperity to Syria.262 Close creation of specific glass vessel forms designed for in-
connections with Rome benefited the native glass interior lighting (see above). Lamps were destined to
dustry. Syrian glassworkers appear to have moved become one of the most important products of the
freely between East and West, with some artisans set- Byzantine glass industry.
tling in the Rhineland. Certain types of luxury glass The forms and functions of vessel glass repre-
excavated in Cologne and Syria, such as glasses with sented in assemblages from eastern Mediterranean
snake-thread decoration263 and flasks within flasks,264 sites have not yet been analyzed, but some general
required technical know-how that was probably trans- observations can be made, albeit with the reserva-
mitted from craftsman to craftsman. The difficulties tion that the conclusions are preliminary and will
encountered by modern glassblowers attempting to need to be adjusted when detailed analyses of indi-
duplicate these wares are an indication that such ves- vidual sites become available. With regard to the di-
sels could probably not be copied simply by studying versity of forms serving one function, the situation in
a finished object. We may credit migrating glassblow- Syria and Palestine developed opposite to the West.
ers with the transfer of these specialized techniques. A cursory count of common blown forms used in
The floruit of the Palestinian glass industry was Palestine266 serves only to indicate the trend (table
the fourth to early fifth century. Following the rule 3). In the first and second centuries 13 forms of
of Diocletian (284-305), when civil strife plagued bowls and dishes were available for serving food, 4
most of the western part of the empire, the eastern forms of drinking vessels, and 3 forms for serving liq-
Mediterranean enjoyed an era of relative peace, de- uids, a total of 20 forms for the table. There were
259 D. Barag, Hanita Tomb XV A Tomb of the Third and Atiqot Hebrew Series 6 (1970) 74-75 [Hebrew], English sum-
Early Fourth Century CE, (Atiqot Engl. Ser. 13, Jerusalem mary p. 8*. On flasks within flasks made in Western Euro-
1978) 54-56; Gorin-Rosen (supra n. 208) 22-23. pean workshops: F. Fremersdorf, Rimische Gldser mit Faden-
260 Weinberg; Stern 1992 (supra n. 44) 490-94.
auflage in Ki'ln (Die Denkmdiler des r6mischen K61n 5,
261 Y. Gorin-Rosen, "Hadera, Bet Eli'ezer," Excavations Cologne 1959) pls. 76-79.
and Surveys in Israel 13 (1995) 42-43; Gorin-Rosen (supra 265 On function, chronology, and typology of kohl tubes:
n. 208) 17 with additional glassmaking sites, 14-17. A slab Stern 1977 (supra n. 76); W.D. Blanchard et al., "Analysis
of raw glass from Beth Shearim was probably made in a of Materials contained in mid-4th to early 7th century A.D.
similar installation: R.H. Brill, "A Great Glass Slab from Palestinian Kohl Tubes," in Materials Issues in Art and Archae-
Ancient Galilee," Archaeology 20 (1967) 88-95. I. Freestone ology 3 (Materials Research Society Symposium Proceedings
has recently redated the slab to the ninth century, based 267, Pittsburgh 1992) 239-54; Stern (in prep.).
on its chemical composition: I.C. Freestone and Y. Gorin-
266 Based on a rough count of individual forms depicted
Rosen, "The Great Glass Slab at Bet She'arim, Israel: An by D. Barag, Glass Vessels of the Roman and Byzantine Periods
Early Islamic Glassmaking Experiment?" JGS 41 (1999). in Palestine (Diss. Hebrew University, Jerusalem 1970) 2:
262 Stern 1977 (supra n. 76) 155-58.
pls. 30-47; decorative variations within one shape are not
263 On snake-thread decoration: Harden et al. (supra n. counted, nor are forms not included in Barag's survey but
16) 105-108 and nos. 55-67. known from excavated finds in Israel.
264D. Barag, "Syro-Palestinian Flasks within Flasks,"

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
Fig. 31. Jar. H. 7.7 cm; Weight 67.4 g.

century. Palestinian. (Courtesy Toledo M
1923.1032. Gift of Edward Drummond
due primarily to their use in chur

drinking, as lighting apparatus, or
If the stemmed
Fig. 30. Kohl tube with two compartments. H. 12.85goblet
cm; was not the
wt. 87.8 g. Pontil scar. Mid-fourth to vessel
ing mid-fifth century.
of the Byzantine period,
Probably Palestinian. (Courtesy Toledo Museum cup
the ordinary of Art,
looked like. In
no. 1923.1272. Gift of Edward Drummond Libbey.)
quantities of glass available it is n
pothesize that vessels of other mate
also 4 forms for storage and/or eclipsed glass drinking
transportation, 37 vessels. D
an important
forms of unguentaria, and 4 miscellaneous function of glass t
forms.
preceding
In the third to early fifth centuries centuries
about 30 indi-and was to be
vidual forms were available for lamic
servingperiod;
foods ininbowls
the West drinki
and dishes (not included are 20 function
Palestinian
of forms of from the fi
glassware
jars that may or may not have modern times.
been used asI table-
suggest that in th
pire goblets
wares), 22 forms of cups and beakers, were
and 46 not only used for
forms
ofjugs and flasks for serving liquids (i.e., a total of 98
forms of tablewares; 118, if thejars are included), 44
general purpose unguentaria (including 11 forms of
kohl tubes), and 7 miscellaneous forms, including 1
form of spouted bottle and 2 forms of lamps. Glass was
not common for bulk storage: perhaps four forms,
probably less were available. These numbers do not in-
clude typical Syrian types such as sprinklers.267
Perhaps the most salient feature of the Byzantine
period (fifth to early seventh centuries) is a tendency
toward uniformity apparent throughout the empire.
Certain types such as the ubiquitous stemmed gob-
let, also known as wineglass from its similarity to the
modern wineglass, were used far beyond the borders
of the empire. Likewise, the stemmed, footed flask
enjoyed widespread distribution. Both types appear
to have originated around the middle of the fifth
century, presumably in Syria or Palestine. The func-
tions of these types suggest that their popularity was
Fig. 32. Sprinkler. H. 6.9 cm; wt. 37.4
First half of third century. Syrian. (D
267 On sprinklers: Stern 1977 (supra n. 76) 95-100; Toledo Museum of Art, after inv. no. 1923.1334. Gift of
Stern 1995, 187 and nos. 129-33; Stern (in prep.). Edward Drummond Libbey.)

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00
484 E.M. STERN, ROMAN GLASSBLOWING IN A CULTURAL CONTEXT
Table 3. Numbers of Vessel Forms from Palestine bottomed cups and beak
Grouped by Likely Function come obsolete.
Period (cent. A.D.) CONCLUSIONS
3rd-
Philological, archaeological, and tec
1st- Early 5th- dence shows that crucial improvements
Vessel Forms 2nd 5th 7th
ing, now taken for granted, occurred in It
Food Presentation during the first century. Physical restric
Bowls, Jars, Dishes 13 50* 1 by size and construction of the Roman
Drinking 4 22 9** limited the number of glassblowers wh
PouringJugs, Flasks 3 46 21
Storage Liquids 4 4 - in one shop, setting glassblowing apart
Storage Unguentaria 37 44 7 fire-based crafts such as pottery and m
Lighting Lamps - 2 17 Almost every glassblower depended on
Other 4 5 1
functioning of long distance trade since
Total Tablewares 20 118* 31 needed for blowing was made at only a
Total All Forms 65 173 56 analysis of the maximum prices for gla
Total Functions 6 7 6** the PE shows that the glassblower wh
* Includes 20 'jars."
in vessel production would have had gre
ties in making a profit. A separate secti
** Includes 9 "wineglasses."
mosaic cubes may have been included
Due to their unique association with
for drinking. In addition,
ture,it is conceivable
blown that
glass tableware played an ce
im
tain types of flasks doubled as drinking vessels.
in bringing Roman culture to the new p
In Palestine, the three most
Europe.common
Glass was anfunctions re
important item
served for glass vessels yond
in the Byzantine
the frontier. period
Trade with Africa in
an
clude pouring (21 forms,brisk. 18
Glassof which
vessels were
were objects fla
of daily
without handles), lighting (17 forms of glass lamp
segments of society but their function a
not including 5 forms ofberswineglass that differed
of forms available were pro con
bly used for drinking West
as well asWhereas
and East. lighting), and
in the West the
local type of cosmetics container, the Palestini
functions and individual forms was by f
kohl tube for which a minimum of
est in the first seven
century, individua
with a marked
forms were available. Glass
versitybowls and
beginning dishes
at the end of proba
the seco
bly remained in use but the
they are of
diversity not easily
individual recogn
forms in use
able among the common blown
Palestine Palestinian
reached its peak in thefor
four
of the period. To judge fifth
from theand
century glass excavated
was unrivalled in anya
Sardis, the same may have
of history. held true in other part
of the Byzantine empire. Classical Roman typ
such as the footed wine-jug with
2040 PARKWOOD AVE.trefoil mouth, sh
low and deep bowls for serving foods, as well as fl
TOLEDO, OH 43620

193.255.211.64 on Tue, 19 Dec 2023 10:29:44 +00:00

This Content Downloaded From 193.255.211.64 On Tue, 19 Dec 2023 10:29:44 +00:00

Uploaded by

Document Information

Original Description:

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

This Content Downloaded From 193.255.211.64 On Tue, 19 Dec 2023 10:29:44 +00:00

Uploaded by

Copyright:

Available Formats

Roman Glassblowing in a Cultural Context

Author(s): E. Marianne Stern

Stable URL: https://www.jstor.org/stable/506970

Archaeological Institute of America is collaborating with JSTOR to digitize, preserve and

This content downloaded from

Abstract between glassblowers are explored. Diocletian's Pr

This content downloaded from

Caesar hand it back to

This content downloaded from

glassblowing sites include Avenches, corners Lyon, and

This content downloaded from

Fig. 1. Modern technique: Gathering hot glass from a pot

ters of his day; nor is the city mentioned by the Flavian

This content downloaded from

Fig. 3. Modern technique: Widening the mouth of the ves

lier cultures.'7 It is my contention,

tempts of ancient techniques-was not invented un-

Fig. 4. Modern glassblower's bench. The glassblower rolls

This content downloaded from

Fig. 5. Core-forming furnace with vertical

20 Glass tubes pinched closed at the lower end and di officina

one del vetro soffiato nell' antichita," JGS 29 (1987) 22-29.

This content downloaded from

Fig. 8. Blowing with a

Fig. 7. Clay oil lamp ation

This content downloaded from

Fig. 10. Cylindrical bottle. H. 14.7 cm; wt. 572 g. No pon-

This content downloaded from

This content downloaded from

against the furnace wall.29 The shelf, als

vessels excavated at Magdalensberg. They are very rare

Molten Glass and Recycling

This content downloaded from

This content downloaded from

Fig. 16. Using a second bubble to flare open th

stan), Damascus, Hebron, and Cairo still fu

This content downloaded from

Fig. 19. Marver(ing surface) adjacent to furnace. Hebron glass

Fig. 20. Picking up a preheated chunk of glass at the tip of a

Fig. 21. Heating a chunk of glass at the tip of a ceramic blowpipe

This content downloaded from

be blown. The advantage

44 P Oxy. 50, no. 3536, line 3. On the literary and archae-

This content downloaded from

Roman glassworking furnace was small52 (figs. 7,

Although excavations often find more than one

This content downloaded from

of glassblowing still show a solitary master at

It is not known how many days per year the an-

60 R.J. Charleston, "Glass Furnaces through the Ages,"

This content downloaded from

.... ... ..... ..

.. .. ....... ..... . . . . . 'D iv

..... ..... ... ....... ...... ... ........ ... ........

....... .. .. ... .. ...

YX. ..... ..........

A number of glassblowers of the Roman period

Greek-speaking eastern Mediterranean.68 Ennion specialized in mold-blown tablewares.

This content downloaded from

Glass vessels showing

This content downloaded from

This content downloaded from