K. Codell Carter - The Rise of Causal Concepts of Disease - Case Histories-Routledge (2017)

The Rise of Causal Concepts of Disease:
Case Histories
The History of Medicine in Context
Series Editors: Andrew Cunningham and Ole Peter Grell
Department of History and Philosophy of Science

University of Cambridge
Department of History
The Open University
Titles in this series include:
‘The Battle for Health’: A Political History of the

Socialist Medical Association, 1930–51
John Stewart
Medicine and Charity in Georgian Bath:

A Social History of the General Infirmary, c. 1739–1830
Anne Borsay
The Nurse Apprentice, 1860–1977

Ann Bradshaw
The Return of Epidemics:

Health and Society in Peru during the Twentieth Century
Marcos Cueto
Reinventing Hippocrates
Edited by David Cantor
The Rise of Causal Concepts
of Disease
Case Histories
K. CODELL CARTER
First published 2003 by Ashgate Publishing
Published 2016 by Routledge

2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN
711 Third Avenue, New York, NY 10017, USA
Routledge is an imprint of the Taylor & Francis Group, an informa business
Copyright © K. Codell Carter, 2003
K. Codell Carter has asserted his moral right under the Copyright, Designs and
Patents Act, 1988, to be identified as the author of this work.
All rights reserved. No part of this book may be reprinted or reproduced or

utilised in any form or by any electronic, mechanical, or other means, now
known or hereafter invented, including photocopying and recording, or in
any information storage or retrieval system, without permission in writing
from the publishers.
Notice:
Product or corporate names may be trademarks or registered trademarks,
and are used only for identification and explanation without intent to
infringe.
British Library Cataloguing in Publication Data
Carter, K. Codell, 1939–

The rise of causal concepts of disease: case histories. –
(The history of medicine in context)
1. Diseases – Causes and theories of causation 2. Medicine –
History
I. Title
616’.071
Library of Congress Cataloging-in-Publication Data
Carter, K. Codell, 1939-

The rise of causal concepts of disease: case histories / K. Codell Carter.
p. cm. – (The history of medicine in context)
Includes bibliographical references.
ISBN 0-7546-0678-3
1. Diseases – Causes and theories of causation – History. I. Title. II. Series.
RB151.C37 2003
616.07’1—dc21
2002019418
ISBN 13: 978-0-7546-0678-9 (hbk)

Contents
Preface vii
Introduction 1
1 Causes of Disease in Early Nineteenth-century Practical
Medicine 10
2 Universal Necessary Causes 24
3 Etiological Characterizations 38
4 Microorganisms as Causes 62
5 The Bacterial Hypothesis 76
6 A Bacterial Theory of Disease 90
7 Proving Disease Causation 110
8 The Etiological Standpoint 129
9 An Ideational Theory of Disease 147
10 Protozoal and Viral Theories of Disease 162
11 A Nutritional Deficiency Theory of Disease 179
Some Final Thoughts 196
Bibliography 204
Index 229
This page intentionally left blank
Preface
Like many undergraduates, I did hard time in a one-year college physics
course. Later I studied the history and philosophy of the physical
sciences. By the time I graduated I had some grasp of what classical
physics was all about. By contrast, the closest I ever came to studying
medicine was a required ten-week course in personal health and
hygiene. The local physician (not a regular professor) who had been
recruited to teach the course told us to be careful with alcohol and sex,
and that chiropractic didn’t really work, but he never discussed medi-
cine as a collection of theories in the same way that physicists taught
physics. Otherwise, no course in medicine or even in the history or
philosophy of medicine was ever available to me at any point in my
entire college career. Of course I could have studied microbiology,
physiology, anatomy, or genetics, but to study medicine – theories of
what can go wrong in your body and how – one had to be admitted to
medical school. The same holds true today at most universities. One
can’t even buy a decent book on internal medicine in ordinary college
bookstores (I don’t count self-help books written on a sixth-grade level).
Medical bookstores, like dealers in pornography or illegal drugs, seem
to have their own sources of supply and to keep a close check on their
customers.
I emerged from college realizing that I knew a whole lot more about
physics than about medicine which, even at that age, seemed a little
odd. Through the years I’ve spent lots of time and money on doctors,
but I’ve never once had professional dealings with a physicist, and I’m
sure my experience is not unique. So now I find our educational priori-
ties even stranger. I’m certain that nothing in medical school is anywhere
near as hard to grasp as, say, Maxwell’s equations or tensor algebra. So
why aren’t students forced (or even allowed) to study something that
touches all their lives so profoundly and on which they will expend so
much time and money?
Within a few years of completing my education, I began to suspect a
conspiracy by the American Medical Association (after all, controlling
knowledge is a source of power just like controlling drugs), and I
became interested in penetrating the great mystery – I wanted to find
out what doctors were trained to believe. Through the years, as I’ve
slaked my curiosity by reading medical books, my suspicions about a
conspiracy have faded (although never quite vanished). I’ve become
convinced that members of the medical establishment really aren’t try-
ing to keep their views secret; in fact, most of them don’t seem to know
viii THE RISE OF CAUSAL CONCEPTS OF DISEASE
or care much about the theories that inform their own professional
activities – they are too caught up treating patients. Imre Lakatos wrote
that ‘scientists tend to understand little more about science than fish
about hydro-dynamics’ (Lakatos, 1968, p. 148n.) which, if you think
about it, is not really an insult – after all, why should fish care about
hydrodynamics? Perhaps members of most professions know how to do
what they do (and even succeed in doing it well) without thinking much
about what they do or why. In any case, this seems true of most
physicians (which may be all for the best – what would happen to a fish
that became preoccupied with theories about how to swim?).
This book reflects my own modest efforts to figure out something
about the theories of disease that ground current medical practice.
History provides one kind of understanding; conceptual analysis
(philosophy) provides another. Ideally they are complementary. Adapting
a famous sentence from Immanuel Kant,1 Lakatos observed that
‘philosophy of science without history of science is empty; history of
science without philosophy of science is blind’ (Lakatos, 1971, p. 102)
– a comment that is no less true of the history and philosophy of
medicine. Having been trained as a philosopher, not as a historian, I’m
mainly interested in conceptual analysis, but, without historical context,
philosophical analysis can become sterile (if not quite empty). So, in
trying to figure out current theories of disease, my approach is to
embody analysis in an historical context. I like to think the result is the
sort of thing Lakatos himself advocated: ‘in writing a historical case
study, one should, I think, adopt the following procedure: (1) one gives
a rational reconstruction; (2) one tries to compare this rational
reconstruction with actual history and to criticize both one’s rational
reconstruction for lack of historicity and the actual history for lack of
rationality’ (Lakatos, 1968, p. 138).
Honestly speaking, I suspect that any philosopher who happens to
read this book will find the philosophy nearly empty, and I already
know that many historians regard my attempts at history as blind. So,
ironically, my work probably resembles (and in a way confirms what
I’ve always thought about) opera: combining drama and music ends up
serving neither interest well. Hopefully someday someone will come
along who will do it better. In the meantime, I can only say that I’ve had
great fun doing what I’ve done – I can’t think of anything I’d rather
have studied.
Parts of this book are based on earlier publications several of which
appeared in Medical History (Carter, 1977, 1980, 1982b, 1985a, 1985b)
and in The Bulletin of the History of Medicine (‘The Koch–Pasteur
Dispute on Establishing the Cause of Anthrax’ (1988) and ‘The Devel-
opment of Pasteur’s Concept of Disease Causation and the Emergence
PREFACE ix
of Specific Causes in Nineteenth-Century Medicine’ (1991)) and I am

most grateful to the editors and publishers who permit me to use them.
Through the years, however, I’ve become persuaded that some of what I
thought I knew was incorrect, and, for this volume, I’ve completely
reworked all my earlier papers. Unless otherwise noted, all translations
are my own.
I’m genuinely pleased for the opportunity to thank, publicly but
mostly without listing names, the several people (including some stu-
dents and even a few historians and philosophers) who have encouraged
me in this effort. I must mention particularly Jackie Duffin (who seems
to get similar thanks from every author I read these days) and Ann
Carmichel (who, without knowing it, happened to provide encourage-
ment at one point just when I needed it most). I also very much appreciate
Andrew Cunningham, editor of the series in which this volume is honored
to find a place; it was his idea to resuscitate a moribund earlier project
and cast it in its current form. Over the years I have enjoyed very
generous research support from Brigham Young University, the stimula-
tion of interested students, the rejuvenating Aufhebung of a disputation
of congenial colleagues, and the help of innumerable librarians in many
interesting places. Most of all a sincere thanks to my wife, Barbara,
who somehow managed to keep herself occupied on Oxford Street and
Mariahilferstrasse while I read dusty books in the basements of London’s
Royal Society of Medicine and Vienna’s Allgemeines Krankenhaus.
Note
1. The sentence is ‘Thoughts without content are empty, perceptions without

concepts are blind’; it comes at B75 in the Critique of Pure Reason.
Introduction
Of the numerous changes that have occurred in medical thinking over
the last two centuries, none have been more consequential than the
adoption of what Robert Koch called the etiological standpoint (Koch,
1901, p. 905). The etiological standpoint can be characterized as the
belief that diseases are best controlled and understood by means of
causes and, in particular, by causes that are natural (that is, they depend
on forces of nature as opposed to the wilful transgression of moral or
social norms), universal (that is, the same cause is common to every
instance of a given disease), and necessary (that is, a disease does not
occur in the absence of its cause).1 This way of conceiving disease has
dominated medical thought for the last century. As new diseases like
Legionnaires’ disease or AIDS emerge, efforts to control and to under-
stand them focus immediately on the quest for such causes.2
The fertility of the etiological standpoint is obvious from the history
of diseases like anthrax, smallpox, or rickets. At the same time it is clear
from historical and anthropological literature that an interest in univer-
sal necessary causes is unique to modern western medicine. There is no
such interest in non-western medicine or in western medicine prior to
the middle of the nineteenth century. This concept is a defining charac-
teristic of modern western thinking about disease.
Adoption of the etiological standpoint and of the concept of neces-
sary causes can be seen as part of the quest for understanding and
control that characterizes science. And, as with other fundamental theo-
retical innovations in science, the ramifications of this way of thinking
transcend any specific discipline: they inform our concept of the human
situation and shape our beliefs about society and about our place in
nature. In an 1872 novel, Erewhon, Samuel Butler portrayed a society
in which illness was regarded as willfully chosen while criminality was a
natural condition that people occasionally fell into in spite of their best
efforts to resist. We respond to willful deviance by moral expostulation
and by legal sanctions; we respond to unintentional deviance by treat-
ment. With the beliefs portrayed in Erewhon, it was inevitable and
entirely rational that the sick were punished while criminals received
treatment and care. However, given the etiological standpoint, it became
difficult to think of illness as resulting from the willful violation of
religious, moral, or social norms. While the way of thinking that Butler
portrayed may still have been feasible in 1872, it was fundamentally
incompatible with the idea that each disease had a specific natural cause
– an idea that was emerging just at the time Butler wrote his novel.
2 THE RISE OF CAUSAL CONCEPTS OF DISEASE
Thus the etiological standpoint encourages us to interpret an important

part of human experience – disease – in terms of natural processes
rather than in terms of free choice or of the violation of norms. Along
with Freudian psychology and the theory of biological evolution, the
etiological standpoint promotes a mundane and naturalistic concept of
human nature.
Because of the success of the etiological standpoint, there have been
persistent efforts to expand the range of human abnormalities that are
approached by way of causes. In recent decades, criminality, substance
abuse, hyperactivity, spouse abuse, obesity, and other forms of deviance
have been interpreted as diseases (Conrad and Schneider, 1980). Such
attempts are motivated by the belief that these abnormalities, much like
tuberculosis or cholera, may be more tractable and more intelligible
within a medical model than if approached through sanctions or admo-
nitions. This approach presupposes that natural causes can be found for
these forms of deviance and, therefore, that they stem from natural
forces rather than from individual choice. Thus, accepting the etiological
standpoint fosters an ongoing re-examination of the boundary between
events deemed manageable by the exercise of individual will and those
regarded as controllable only through the manipulation of natural forces.
In respect to its pervasiveness in contemporary thought and to its
uniqueness to the modern age, the etiological standpoint resembles the
great nineteenth-century biological innovation: the theory of evolution.
And because medicine consumes a huge percentage of the world’s
economic resources and touches virtually every human life, it is imminent
in a way that no other aspect of science can ever be. Yet, in contrast to
evolution or to the theories of physics, the etiological standpoint re-
ceives remarkably little philosophical or historical scrutiny. This book is
a collection of case studies concerning the rise of the etiological stand-
point.
It is enlightening to regard the etiological standpoint as a scientific
research programme. This requires some explanation. In response to
what he saw as deficiencies in earlier attempts to characterize scientific
progress, Imre Lakatos formulated the concept of scientific research
programmes and proposed that the history of science be regarded as a
history of research programmes rather than as a history of theories
(Lakatos, 1968, p. 132). Lakatos explained his concept of research
programmes in terms of what he called ‘progressive and degenerating
problemshifts’, and he explained these concepts as follows:
Let us take a series of theories, T1, T2, T3, … where each subse-
quent theory results from adding auxiliary clauses to … the previous
theory in order to accommodate some anomaly, each theory having
at least as much [empirical] content as the unrefuted content of its
INTRODUCTION 3
predecessor. Let us say that such a series of theories is theoretically

progressive (or ‘constitutes a theoretically progressive problemshift’)
if each new theory has some excess empirical content over its
predecessor, that is, if it predicts some novel, hitherto unexpected
fact. Let us say that a theoretically progressive series of theories is
also empirically progressive (or ‘constitutes an empirically progres-
sive problem-shift’) if some of this excess empirical content is also
corroborated, that is, if each new theory leads us to the actual
discovery of some new fact. Finally, let us call a problemshift
progressive if it is both theoretically and empirically progressive,
and degenerating if it is not. We ‘accept’ problemshifts as ‘scien-
tific’ only if they are at least theoretically progressive; if they are
not, we ‘reject’ them as ‘pseudoscientific’.
(Lakatos, 1968, p. 118)
Later in the same essay, Lakatos used the preceding distinction as a
basis for characterizing scientific research programmes:
I have discussed the problem of objective appraisal of scientific
growth in terms of progressive and degenerating problemshifts in
series of scientific theories. The most important such series in the
growth of science are characterized by a certain continuity which
connects their members. This continuity evolves from a genuine
research programme adumbrated at the start. The programme con-
sists of [mostly implicit] methodological rules: some tell us what
paths of research to avoid (negative heuristic), and others what
paths to pursue (positive heuristic).
(Lakatos, 1968, p. 132)
While not immune to criticism, Lakatos’s account of scientific progress
provides a convenient framework for describing the history of causal
thinking in medicine over the last century and a half.
By providing research models for their contemporaries and succes-
sors, Louis Pasteur and Robert Koch directed research away from such
questions as whether diseases arise spontaneously or from moral trans-
gression and toward the quest for bacteria that stand in the relation of
necessary and universal causes of diseases. Subsequent researchers ‘added
auxiliary clauses’ to the initial framework to account for new bacterial
diseases. Within the etiological standpoint, subordinate programmes
arose to deal with psychological abnormalities, protozoal, viral and
rickettsial infections, nutritional deficiencies, metabolic and genetic dis-
orders, and other classes of anomalies.3 Each of these additions was
made within the context of the quest for specific causes and each
addition expanded the original research programme in a way that was
(using Lakatos’s phrase) adumbrated from the start. Attempts to iden-
tify causes for new diseases like AIDS, or especially for deviant behavior
like addiction or abuse, can be seen as further attempts to expand the
programme or to generate subordinate programmes. Each successful
assimilation of a new anomaly reinforces the credibility of the pro-

gramme as a whole. On the other hand, the credibility of the programme
is threatened by the accumulation of recalcitrant anomalies, like cancer,
that seem to defy comprehension in terms of necessary causes and by
the discovery that the etiologies of some diseases are more complex
than originally thought. These developments erode confidence in the
research programme and threaten to replace its fundamental concept by
such notions as multicausality, risk factors, or causal webs (Kunitz,
1987; Susser, 1973, pp. 22–4).
It is impossible to predict whether the quest for necessary causes will
yield further successes or be replaced by other ways of thinking. But,
whatever may ensue, there is no denying the centrality of the concept of
necessary causes in medical thought for the last one hundred years. And
even if this concept ceases to prove useful, there is no clear alternative
to approaching diseases etiologically. One can imagine a medical system
in which risk factors replace universal necessary causes – this requires
only a change within an etiological framework. However, it is virtually
impossible to imagine how any future medical system could be organ-
ized except around some causal concepts or other.
By definition, a research programme is a temporally continuous co-
operative endeavor to which different persons can contribute (Lakatos,
1968, p. 132). But what are its boundaries? In tracing a single research
programme, one would like to distinguish the programme itself from
such extraneous elements as, first, general background knowledge or
research methods shared and exploited both by contributors to the
programme and by other researchers who may be contributing to other
(even competing) programmes; and second, ideas and methods advanced
by precursors who did not contribute directly to the programme but
who merely anticipated some of its features. Lakatos defines research
programmes as collections of methodological rules, but this definition
yields no clear practical basis for individuation. However, at least in
modern scientific literature, boundaries can be approximated by taking
account of citations. There is no need for citations when drawing on
shared background knowledge. References are required only for claims
that are not universally acknowledged or that are recognized as part of
the proprietary domain of some individual or group. Moreover, a new
idea or method contributes to a continuous research programme only if
others adopt it, and one expects such borrowing to be acknowledged by
citations. Thus, a research programme can be regarded approximately
as a collection of works linked through citations, and one can individuate
programmes by constructing citation indexes. One begins with a work
central to a programme, say Koch’s first paper on tuberculosis. One
then identifies all the works that cite or are cited in that work, all the
INTRODUCTION 5
works that cite or are cited in those works, and so on. The collection of
all these works approximates a research programme. In logical terms,
over the domain of all published works, a research programme is ap-
proximately the equivalence class defined by any particular work under
the relation ‘cites or is cited by’. This criterion has obvious weaknesses;
however it provides a practical standard for the selection of materials,
and I have chosen the topics in this book with this criterion in mind.
The works here discussed are directly and indirectly linked by citations.
One final concept requires clarification. Lakatos defined ‘research
programme’ in terms of series of scientific theories, and most of this
book concerns ideas that are analytically connected to the concept of
theories. At the start, therefore, we must give some attention to how the
word ‘theory’ is to be understood. In ordinary life, ‘theory’ often refers
to any belief or opinion, but in science the term is used more narrowly.
Wim J. Van der Steen and Harmke Kamminga gave the following
succinct account (which is compatible with other recent discussions in
the philosophy of science):
A statement is a law if it satisfies the following criteria: (i) it is
general in the sense that it contains a universal quantifier, (ii) it is
general in the sense that it does not mention particular individuals,
times or places, (iii) it has empirical content, (iv) it is well-con-
firmed, and (v) it is well-entrenched (i.e. it belongs to a theory). …
Theories will be viewed as sets of interconnected laws.
(Van der Steen and Kammingen, 1991, pp. 445f.)
As this account makes clear, theories and laws are conceptually linked:
neither is possible without the other. When formalized, some of the
laws comprised by a theory typically function within the theory like
axioms in geometry or in other deductive systems. For example, here is
a set of assumptions (axioms) for the kinetic theory:
(1) A pure gas consists of a large number of identical molecules
separated by distances that are great compared with their size. (2)
The gas molecules are constantly moving in random directions
with a distribution of speeds. (3) The molecules exert no forces on
one another between collisions, so between collisions they move in
straight lines with constant velocities. (4) The collisions of mol-
ecules with the walls of the container are elastic; no energy is lost
during a collision.
(Oxtoby, 1990, p. 104).
These assumptions have the qualities identified by Van der Steen and
Kamminga (so they are laws in their sense), and from them, other laws
(for example, Boyle’s Law, Charles’ Law) can be derived mathematically
much as one derives theorems in geometry. As a second example, in
their classic book on the general theory of relativity, Charles W. Misner,
Kip S. Thorne, and John Archibald Wheeler observed that ‘of all theories
ever conceived by physicists, general relativity has the simplest, most

elegant geometric foundation (three axioms: (1) there is a metric; (2) the
metric is governed by the Einstein field equation G = 8πT; (3) all special
relativistic laws of physics are valid in local Lorentz frames of metric)’
(Misner et al., 1973, p. 302).
And there is one other important fact about laws and theories. Just as
a statement cannot be a law without being ‘well-entrenched (i.e. part of
a theory)’, it is impossible to fully justify a law abstracted from the
theory of which it is a part. As Norwood Russell Hanson explained, the
point of a theory is ‘to offer an intelligible, systematic, conceptual
pattern for the observed data. The value of such a pattern lies in its
capacity to unite phenomena which, without the theory, are either
surprising, anomalous, or left wholly unnoticed’ (Hanson, 1963, p. 44).
In short, a theory is justified by tying together and explaining observa-
tions. Laws, which can exist only within theories, are justified by being
part of justified theories.
Obviously, in medicine (or in the biological sciences generally), one
does not find formalized theories stated as elegantly as relativity or the
kinetic theory. But ultimately the nature of theories must be the same.
Even in medicine, if one is to talk of scientific theories, say, the bacterial
theory of disease or the deficiency theory, one must mean more than
simply that people opine that bacteria are, somehow or other, involved
in disease. While different people may have different beliefs (theories in
the loose sense) about the pathologicality of bacteria, if there is such a
thing as the (or a) bacterial theory of disease (where ‘theory’ is used as
it is used elsewhere in science) it must comprise a system of law-like
statements. Believing, as I do, that ‘the bacterial theory of disease’ is not
just a misnomer, in other words, believing that there are genuine theo-
ries in medicine, one of my main goals in this book is to identify some
of the laws that constitute these theories.
In this context, I must comment on the relation of my book to
Michael Worboys’ excellent recent study: Spreading Germs: Disease
Theories and Medical Practice in Britain, 1865–1900 (Worboys, 2000).
Among contemporary historians of medicine, there is precious little
interest in scientific theories,4 and Worboys’ volume is an auspicious
sign that this omission may someday be corrected. But what does
Worboys mean by ‘theory’? He never explains the term, but it seems
clear that, for the most part, he is using the word in the loose sense of
any belief or opinion rather than in the narrow sense discussed in the
three previous paragraphs. For example, he makes it a point to talk
about disease theories (plural)5 and here is how he introduces this
concept: ‘the first and most important theme to acknowledge is the
range of germ theories of disease current between 1865 and 1900. In
INTRODUCTION 7
the 1860s and 1870s, there were many views on what disease-germs
were, for example, chemical poisons, ferments, degraded cells, fungi,
“bacteria” or a class of parasites’ (Worboys, 2000, p. 2). So what
Worboys is considering is different ‘views’ about germs and their possi-
ble relation to disease. I say for the most part Worboys seems to use the
term ‘theory’ in the loose sense because in his first chapter, he does once
refer to the narrow use of the word: ‘Yet, by the 1870s, the term [germs
of disease] was firmly linked to the modern concept of the “germ theory
of disease” [singular] – the etiological construction of disease in which
external agents entered the body to produce septic, infectious and other
diseases’ (p. 22). This reveals the first difference between my interests
and Worboys’: I use the term ‘theory’ exclusively in the narrow sense
explained in previous paragraphs and tersely exemplified in this quota-
tion from Worboys. Thus, it is no accident that, in my opinion, the first
explicit formulation of a bacterial theory (Klebs’) came in 1876 – the
same decade in which Worboys sees ‘germs of disease’ finally linked to
the modern concept of the germ theory. As I use the term ‘theory’, there
was no germ theory (at least no explicit germ theory) prior to that time
while obviously there were lots of opinions about germs. I am interested
in the laws (in Van der Steen and Kamminga’s sense) comprised by the
bacterial theory of disease (in the narrow sense of ‘theory’) – something
to which Worboys gives no attention.
A second difference between Worboys’ book and my own is this: my
main interest is not really even the bacterial theory of disease. Rather, it
is a broad movement – the etiological standpoint or research pro-
gramme – of which the bacterial theory is one part. For me, the bacterial
theory is only one example (albeit an important one) of the quest for
universal necessary causes that characterizes medicine since the middle
of the nineteenth century. Thus, while I have enormously enjoyed
Worboys’ fine book and have learned a lot from reading it, there is –
unfortunately for me – almost no intersection between his work and my
own.
The etiological research programme arose in the nineteenth century,
and, while boundaries are as much created as discovered, one can
distinguish three stages in this process. First, between about 1830 and
1860 a few researchers independently, nearly simultaneously, and ap-
parently for the first time in medical history, began to think of several
different organic disorders as having causes that were universal and
necessary. A new strategy arose – the adoption of etiological characteri-
zations – that provided a way of classifying individual illnesses so that,
in principle, every recognized disease could be ascribed to a universal
necessary cause. Basic methodological principles emerged forming what
Lakatos would call the heuristic of the programme, and researchers
began focusing on a particular domain of organisms – bacteria – that

were later seen as causes of disease. As with many conceptual innova-
tions, these developments were accompanied by little philosophical
reflection or analysis – none of those who first used them may have
considered how these approaches differed from those of their predeces-
sors or why the differences were important. Indeed, such matters can
only become clear in retrospect.6 For example, we are better able to
appreciate the nature and significance of, say, Koch’s Postulates than
Koch himself could ever have been (in this regard, I freely confess a
presentist and partly Whiggish bias).7 The issues here are the structure
and subsequent use of published arguments, not how the authors of
those arguments would have characterized their own work or how they
would have compared their positions to those of other writers. It is
worth emphasizing that, during this period of its inception, the research
programme was not exclusively about bacteria or bacterial diseases.
While the bacterial theory may be the most striking single achievement
of the programme, the programme did not begin or end with the study
of germs. As explained in the previous paragraph, this book is not
primarily about the germ theory of disease.
Second, between about 1860 and 1890 researchers focused, with
enormous success, on bacterial diseases. A coherent and explicit bacte-
rial theory of disease emerged. In this period, researchers gave attention
to the concept of disease causation: they identified, discussed, and
refined various criteria for proving causation, and they deliberately
organized their experiments to satisfy such criteria. Among these crite-
ria were those now known as Koch’s Postulates. There was also an
awareness of and a conscious interest in the basic heuristic principles of
the emerging programme.
Third, from about 1890 on, as the power of the new causal model
became apparent, researchers addressed whole new ranges of disorders
that were not bacterial. This required adjustments in the basic princi-
ples and methods of the programme, for example, in causal criteria. But
collectively the results of these efforts bear such strong family resem-
blances (in Wittgenstein’s sense) to one another and to the core of the
programme that they must all be seen as falling within the etiological
standpoint.
The case studies in this volume are organized around these three
stages in the rise of the research programme: Chapters 2 through 4
concern the first stage; Chapters 6 and 7 the second; and Chapters 9
through 11 the third. The first chapter provides background, and Chap-
ters 5 and 8 are transitional.
INTRODUCTION 9
Notes
1. The usual term is ‘specific cause’. However, as this term is currently used,
it is both vague and ambiguous. When possible I avoid the term altogether,
but when it appears, it should be understood to mean a natural, universal,
and necessary cause.
2. F. Kräupl Taylor uses the term ‘monogenic’ for diseases associated with
universal and necessary causes. In a philosophically sensitive and insightful
book that has received too little attention, he writes that while most
diseases recognized in medicine today are not monogenic ‘the final hope
and aim of medical science is the establishment of monogenic disease
entities’ (Taylor, 1979, p. 21). Alfred S. Evans has traced efforts to assimi-
late numerous new diseases to a model much like the one elaborated in this
book (Evans, 1993).
3. Lakatos writes that all of science can be seen as one huge research pro-
gramme within which special programmes focus on particular problems
(Lakatos, 1968, p. 132). In the same way the entire etiological standpoint
can be regarded as a single research programme or one can think of it as a
collection of separate programmes each dealing with a class of diseases
with similar causes. This is a useful ambiguity and there is no reason to
eliminate it arbitrarily.
4. I will have more to say about this in the chapter entitled ‘Some Final
Thoughts’ at the end of the book.
5. What could it mean to talk about ‘kinetic theories of gasses’? While differ-
ent accounts of the kinetic theory may employ different basic assumptions
or state equivalent assumptions differently, the consequences of every such
set must be isomorphic with the consequences of every other. In this sense,
there can be only one kinetic theory. No doubt, early on, there were
innumerable opinions about diseases and even about the nature of germs
and about their possible relations to diseases, but, in my mind, to use
‘theory’ in these contexts verges on false advertising. Worboys does cite
one 1878 work in which the word ‘theories’ (plural) was used in this way.
6. This point is related to Lakatos’s thesis that so-called crucial experiments
can never be seen as crucial until decades after they have been performed
(Lakatos, 1968, p. 158).
7. Back when they may have hoped there was still a chance of saving my
immortal soul, Adrian Wilson and Andrew Cunningham kindly presented
me with a copy of Butterfield’s The Whig Interpretation of History. But I
found Butterfield’s strawman arguments only about 40 per cent persuasive
so, while confessing, today, to be 60 per cent Whig is about like calling
oneself a witch in 1692 Salem, there is little point in denying what will,
soon enough, be apparent to the reader.
CHAPTER ONE
Causes of Disease in Early

Nineteenth-century Practical
Medicine
Early nineteenth-century medical texts are superficially similar to our
own. Textbook discussions of most diseases include sections with such
titles as pathology, etiology, therapy, and prognosis, and the contents
may seem readily intelligible to a modern reader. One could surmise
that our medical system is part of a continuous tradition extending
back through these decades and beyond – that changes have come
mostly through the accumulation of facts within a shared framework of
fundamental beliefs and objectives (King, 1982, pp. 5–15). However,
earlier medical texts contain statements that now seem strange, and
these statements sometimes point to fundamental discontinuities sepa-
rating earlier thinking from our own.
In 1845 James L. Bardsley, a prominent British physician, wrote that
diabetes ‘has been traced by some patients to sleeping out the whole of
the night in a state of intoxication’ (Bardsley, 1845, p. 609). This
statement seems curious not simply because Bardsley identified a cause
we no longer find plausible; more fundamental issues are involved.
Bardsley assumed that patients’ opinions were important or even deci-
sive in identifying the causes of their illnesses, and this assumption
suggests that different cases of the same disorder may have entirely
different causes.
In the same year, Wilhelm Friedrich Scanzoni, director of the Prague
maternity clinic, proposed to study the etiology of childbed fever, an
often fatal disease that struck women a day or two after delivery. As
part of his study, Scanzoni urged the authorities to require local physi-
cians to report each case of childbed fever that occurred in their practices.
In their reports they were to give ‘particular attention to the causal
factors of the disease’ (Scanzoni, 1850, p. 32). Scanzoni intended to
study the etiology of childbed fever by determining the frequency of the
different causes to which local physicians ascribed individual cases.
Like Bardsley, Scanzoni assumed that different cases of a given disease
could have completely different causes.
Medical texts contain the results of surveys like the one Scanzoni
proposed to conduct. Seven years earlier, in 1842, A. F. Chomel, a
DISEASE IN NINETEENTH-CENTURY PRACTICAL MEDICINE 11
leading French internist, explained that one could investigate the causes
of pneumonia ‘by interrogating carefully a certain number of individu-
als struck by this affliction, and by directing one’s questions to the
causes that produced it’ (Chomel, 1842, pp. 165f). In his own study,
‘made with great care on seventy-nine pneumonia patients’, he found
the following: fourteen patients reported experiencing some form of
cooling, five had consumed too much wine, two had worked exces-
sively, one had experienced a lively emotion, and another had inhaled
carbon vapor. The 56 remaining patients were unable to explain how
they had become ill. Chomel also reported an earlier study of 125
patients in which the following causes had been established: contusions
of the throat, 2; cooling, 38; violent effort and fatigue, 12; depression,
4; excess of drink or upset regimen, 3; and in 66 cases no cause could be
identified.
Proceeding in this case-by-case way, physicians accumulated exten-
sive lists of possible causes for each disease. For example, in his account
of diabetes, Bardsley identified the following causes: frequent exposure
to sudden alterations of heat and cold, indulgence in copious draughts
of cold fluid when the system has been over-heated by labor or exercise,
intemperate use of spirituous liquors, poor living, sleeping out the
whole of the night in the open air in a state of intoxication, checking
perspiration suddenly, and mental anxiety and distress (Bardsley, 1845,
p. 609). Similar lists can be found for virtually any disease in most
German, English, or French medical texts from the period.
In order to characterize more precisely the causes included in these lists,

one must first be clear about the meaning of the terms ‘necessary’ and
‘sufficient’. A set of conditions C is sufficient for events of type E if the
presence of C always precedes or accompanies an event of that type (or,
equivalently, if the absence of such an event always accompanies or
follows the absence of those conditions); a set of conditions C is neces-
sary for events of type E if the absence of C always precedes or
accompanies the absence of E (or, equivalently, if such an event always
accompanies or follows those conditions). In connection with these
definitions, several points require emphasis:
1. Necessity and sufficiency apply only to classes of events. To say

certain conditions are necessary or sufficient for a single event can
only be understood to mean that the event belongs to a class of
events that stand in the specified relation to the conditions.
2. Necessity and sufficiency are strictly empirical relations. This im-
plies that conditions C may be necessary or sufficient for event E
without either being the cause of the other: my having swallowed a

vitamin pill may always and only precede my saying grace over
breakfast – having taken the pill may be necessary and sufficient
for saying grace, but neither is the cause of the other. Thus, proving
causation requires more than simply demonstrating necessity or
sufficiency (or both). (Some of what more is required in a proof of
causation will emerge in subsequent chapters.)
3. While we often talk of events as being necessary or sufficient for
other events, as Kant taught us, causation is fundamentally a rela-
tion between sets of conditions and events.
Since we will generally be interested in necessity and sufficiency in the

context of causation, Kant’s insight is reflected in the above defini-
tions.
Causal concepts are rooted in everyday interests in controlling and
understanding the world. Our interests vary, and so does our language
about causes. We talk about causes of individual events (what caused
the accident) and of classes of events (what causes lightening); we use
causes to explain what has happened or is happening now (what caused
the power failure) and to predict what will happen in the future (adding
iron will cause the leaves to turn dark green); causes insure that events
will happen (what will cause the tree to bear fruit) and prevent events
from happening (what causes wilt). In discussing causes, the terms
‘necessary’ and ‘sufficient’ (as defined above) mark two broad kinds of
causes both of which interest us in daily life. In general when we want
to explain why some event or class of events happens or to insure or
predict that such events will happen, we look for a cause whose occur-
rence will be accompanied by the event – that is we look for a sufficient
cause. However when we want to explain why some event or class of
events doesn’t happen or to insure or predict that such events won’t
happen, we look for a cause whose absence will be accompanied by the
absence of the event – that is we look for a necessary cause. So,
depending on our interests, we sometimes seek sufficient and sometimes
necessary causes. Because our interests in the two kinds of causes are
different, most causes are not both necessary and sufficient.
What was the nature of the causes identified in early nineteenth-
century practical medicine? Physicians actually distinguished different
kinds of causes. There were proximate and remote causes and remote
causes were further classified as predisposing or exciting (or occasion-
ing). The proximate cause of a disease was the anatomical abnormality,
if any, with which the disease was associated. One writer defined ‘proxi-
mate cause’ as ‘the essence of disease, its intimate nature, the special
alteration of the solids or fluids. … [the effect of which is] to fix the
new relations that occur in the economy of the ill body’ (Lagasquie,
1849, p. 313). This made good sense until, under the influence of
pathological anatomy, some diseases were identified with specific le-
sions. Given such characterizations, the proximate cause and the disease
became one and the same: ‘The proximate cause is nothing else than the
actual disease itself – the actual condition of that part of the body, from
which the whole train of morbid phenomena essentially flows’ (Watson,
1858, p. 76). Writers acknowledged that this use of the term ‘cause’ was
confusing. ‘It is … a puzzling term, and tends to give to the study of
disease a scholastic and repulsive aspect’ (Watson, 1858, p. 76). Some
physicians dismissed all talk of proximate causes as redundant and
urged that the term be abandoned. ‘If we retain the term proximate
cause, it signifies … the pathological condition of the throat, the pleura,
or the joints; but we think the term may be advantageously banished
from medicine’ (Conolly, 1845, p. 677; Watson, 1858, p. 76).
In contrast to proximate causes, remote causes were factors normally
external to the patient that explained the onset of disease. This was the
most common sense in which the word was used, and all the causes
identified in the first section of this chapter were remote causes.
What was the relation between some particular remote cause, say
anxiety, and the disease it occasioned? Clearly, no one cause was neces-
sary for the onset of any specific disease since, as we have seen, different
cases of each disease were ascribed to different remote causes. Physi-
cians explicitly denied that the same cause was present in every case of
any disease: ‘No inference can fairly be drawn from the identity of the
effect, for certain diseased states are produced by a variety of remote
causes, this expression being used in the sense usually annexed to it by
medical men’ (Brown, 1845, p. 505). It is also clear that the remote
cause identified in a particular disease episode was not necessary even
for that episode because, if a given patient did not contract her disease
from one cause, she might contract the same disease from other causes.
Or, conceivably, several causes (no one of which was necessary) might
contribute equally to the onset of a particular episode. Thus no single
cause was regarded as necessary for any one of the diseases it could
cause.
However, particular remote causes seem not to have been thought of
as sufficient either. A typical factor, say anxiety, could be listed as a
possible cause for many diseases. For example, one widely used British
medical encyclopedia identified anxiety as a possible cause of dozens of
diseases including acne (1:48),1 catalepsy (1:379), diabetes (1:609),
ecthyma (1:739), fever (2:177), herpes (2:449), hydrocephalus (2:490),
impetigo (2:594), senile dementia (3:74), psoriasis (3:733), scrofula
(4:139), and tetanus (4:168). Thus anxiety was not expected always to
result in the same disease – one anxious person might become diabetic
but others might contract impetigo or tetanus. So remote causes were
not sufficient either. Christopher Hamlin, who has discussed accounts
of disease causation in this period, recognizes that ‘the presumption of a
single exciting cause as sufficient could not be sustained’ and seems to
infer that ‘all remote causes could plausibly be thought of as necessary
causes’ (Hamlin, 1992, p. 51). But clearly this doesn’t follow and (as we
have seen) it is inconsistent with what contemporary physicians them-
selves reported about how they used the words. Hamlin gives no further
evidence for this surprising claim.
However, while individual causes were not thought of as sufficient,
they could be part of a combination of factors that was sufficient. To
understand this, we must take account of another distinction important
in nineteenth-century etiology. Remote causes were classified as predis-
posing or exciting (or occasioning). Predisposing causes ‘render the
body liable to become the prey of something, which has a tendency to
excite the disease. The exciting cause of the disease might have no
effect, unless the body had been predisposed; and the predisposition
might not have had the effect, unless the exciting cause had occurred’
(Elliotson, 1844, p. 43). Typically, in an account of any disease, both
predisposing and exciting causes were listed. One discussion of tetanus
mentioned these predisposing causes: warm climate, humid situations,
bad or insufficient nutriment, close and ill-ventilated habitations, inat-
tention to cleanliness, and neglect of the bowels. The same discussion
listed these exciting causes: mechanical wounds, application of cold and
damp, the irritation of worms, terror, sympathy, mental anguish, the
suppression of perspiration, the accumulation of cherry-stones in the
intestines, suppression of lochia, and gastric inflammation (Symonds,
1845, p. 368). Contemporary writers observed that, over time, predis-
posing causes could become exciting, and what were usually exciting
causes could become predisposing (Elliotson, 1844, p. 45). Moreover
under suitable circumstances, either kind of cause alone could seem to
bring on a disease (Conolly, 1845, p. 678). One writer observed that
‘the same occasional cause can provoke the development of all the
diseases, and the same disease can be created by every kind of occa-
sional cause’ (Chomel, 1835, p. 425). If, as this suggests, a typical
remote cause is neither necessary nor sufficient, how can distinguishing
between predisposing and exciting causes help provide a complete suffi-
cient causal explanation?
Several discussions of disease causation included some version of this
hypothetical case:
Of several individuals exposed to the same exciting cause, scarcely
two will be affected alike. From exposure to cold, for instance, one
will be attacked with catarrh, another with rheumatism, a third

with inflammation of the bowels, a fourth with sore throat; while
by far the greater number will escape altogether. Were the exciting
cause solely chargeable with these several effects, they would un-
questionably be marked with greater uniformity. The truth is, that
the exciting cause produces its effect because the body exposed to
it is prone to be morbidly affected in consequence of its own
previous derangement; and the specific form of the disease is deter-
mined, partly by the operation of the exciting cause, but chiefly by
the predisposition of the parts affected to undergo those morbid
actions to which the general indisposition of the system and their
own partial weakness render them liable.
(Barlow, 1845b, p. 555)
Thus it was believed that any single remote cause could, in principle, be
filled out into a fully sufficient explanation of the onset of a case of
illness with the help of other predisposing and exciting factors. In 1835,
while explaining why different persons react differently when exposed
to a cold damp wind, a French writer observed that
This experiment proves only that which observation confirms every
day, namely, that the same cause can give rise to different effects,
according to the particular state of the individual on which it acts.
Thus the same frozen drink given to several perspiring persons
produces in one a simple loss of voice, in another a cold, in a third
a very serious laryngitis, etc.
(Trousseau, 1835, p. 338)
Fourteen years later, another French physician wrote that no one could
predict what would happen when persons were exposed to an exciting
cause, say, sudden chilling – some become rheumatic, some develop
pulmonary catarrh or diarrhea, and others remain healthy. He then
observed that ‘a previous examination of each [victim] would no doubt
solve the problem’ (Lagasquie, 1849, p. 313). Of course, in practice, no
examination could be thorough enough to identify all the relevant
‘indispositions and partial weaknesses’ to enable one to predict the
exact outcome of some trauma. However, in principle, the collection of
all the remote causes – both predisposing and exciting – would be
sufficient for the particular effect. Moreover, presumably, any two persons
affected by the same set of predisposing and exciting causes would
suffer the same effect.
All of this suggests that a remote cause, such as those identified at the
beginning of this chapter, could be any trauma that seemed especially
striking to the patient or physician and that helped explain the onset of
one case of a disease. Such a cause was thought of as one of a combina-
tion of conditions that, together, constituted a sufficient cause for the
particular instance of disease. There is nothing unreasonable about such
causal explanations: they are exactly what one would look for, today, if
one were interested in explaining why some individual became ill. It

was also reasonable to think patients could be aware of salient trau-
mata that might be part of such an explanation. From our point of
view, no less than from that of nineteenth-century physicians, patient-
reported experiences such as overexertion, sleep deprivation, or poor
diet could be part of a sufficient explanation of the onset of a particular
case of illness. This is how one must understand Bardsley’s comment
that some patients had traced illness to ‘sleeping out the whole of the
night in a state of intoxication’.
In the early nineteenth century, there was an elegant symmetry between

causes of disease and causes of death. In 1839 the Registrar General of
Great Britain began publishing annual reports of births, deaths, and
marriages. The first report was mostly compiled by William Farr who
explained that ‘the registration of births and deaths proves the connec-
tion of families, facilitates the legal distribution of property, and answers
several other public purposes, which sufficiently establish its utility’
(Farr, 1839, p. 86).
Farr illustrated what was to be understood by a cause of death: ‘A
man falls from a height, and breaks his neck; a woman takes arsenic,
which corrodes the coats of the stomach, and in both cases death is the
result. The arsenic and the fall, or the fracture of the neck and the
corrosion of the stomach, may be viewed as the causes of death’ (Farr,
1839, p. 89). While Farr recommended that causes of both kinds (the
external factor and the internal morbid process) be registered and con-
sidered as causes of death, he acknowledged that in many cases only
one or the other could be found, and, in practice, usually only one was
reported. In a letter accompanying his second annual report, he ob-
served that some identified causes of death were ‘diseases, which
terminate in the extinction of existence’, but at other times ‘the atten-
tion of the observer was less attracted to this class of facts, and
overlooking the proximate, passed directly on to the more impressive
external causes’ (Farr, 1840, p. 75). He then observed that a ‘vast
number of the deaths … are referable to cold, and to effluvial poisons
… and it is scarcely necessary to state that many of the diseases had
their origin in intemperance, and in the want of proper food’, and he
then discussed each of these as a cause of death. These quotations
illustrate not only that, like causes of disease, causes of death were
thought of and classified as proximate and remote, but also that one
could pass back and forth between causes of death and causes of
disease, often mentioning the same factors – cold, intemperance, starva-
tion, effluvial poison – sometimes as one and sometimes as the other.
Farr’s idea was this: suppose a woman is exposed to some trauma, say
cold, and that an ensuing morbid process (that is, a disease) causes her
death. Then there was no distinction between the remote cause of the
disease and the remote cause of death.
Not only were particular causes of disease and of death similar and
sometimes the same, but causal explanations were also similar in kind
or form: for both disease and death, the object was to explain what
happened to one individual at one time. This called for identification of
a sufficient cause for a particular event. As we have seen, having found
causes of many individual disease episodes, physicians like Scanzoni
and Chomel tried to discern patterns. In the same way, Farr sought
patterns among the causes of individual deaths; for example, he thought
that the record of causes of death could help one decide which parts of
the country would be safe retreats from serious diseases like consump-
tion or whooping cough (Farr, 1839, p. 87).
The causes of death we identify, today, are still sufficient causes for
particular events. ‘The certificate wants to find out not why every
patient died but why this patient died’ (King, 1982, p. 213). And so,
although more precise and detailed, they are of the same general form
as the causes of death identified 150 years ago. Moreover, we still
accumulate information about causes of death by surveys like those that
Scanzoni and Chomel proposed for investigating the causes of childbed
fever and pneumonia. However, as will become clear in the course of
our investigation, the symmetry between causes of disease and causes of
death has been destroyed by a change in the kind of causes we now
identify for diseases (Carter, 1997).
We can now appreciate one important discontinuity between etiological
discussions in the 1840s and causes of death, on the one hand, and the
causes of disease that most interest us, today, on the other. Today, in
both ordinary everyday conversation and in technical medical discourse,
one speaks of ‘the cause of X’ where X is the name of some disease; for
example, ‘the cause of diabetes’ or ‘the cause of tuberculosis’. And the
causes that are typically identified in such contexts are among those
that nineteenth-century physicians would have called remote. However,
in literature from the early decades of that century, one looks in vain for
any talk of ‘the cause of disease X’. Among remote causes, one finds
only exciting and predisposing causes of individual cases such as those
we have listed and characterized. One reason for the apparent strange-
ness of James L. Bardsley’s discussion of the etiology of diabetes,
considered in the first section of this chapter (pp. 10–11), is that, in a
general discussion of the etiology of diabetes, such as he was purporting
to give, we now expect an account of the cause of diabetes (so far as it is
understood) not just a list of factors contributory to individuals becom-
ing diabetic. But in the early 1800s, there were no such accounts; there
simply was no concept of ‘the cause of disease X’. Why might this have
been?
One reason is this: given how diseases were then defined, it was
reasonable and virtually inevitable that different episodes of any disease
would be attributed to a variety of unrelated causes. In the early nine-
teenth century, individual diseases were almost always characterized in
terms of prominent signs and symptoms (or, under the influence of
pathological anatomy, of particular morbid alterations in the body). For
example, in a lecture delivered at the University of Paris in 1832 and
reprinted in the British journal, Lancet, Gabriel Andral defined ‘hydro-
phobia’ as ‘complete horror of fluids, reaching to such a degree, that
their deglutition becomes almost impossible’ (Andral, 1832–33, p. 806).
Thus, hydrophobia was defined in terms of one prominent symptom: an
extreme inability to swallow. However, if hydrophobia is a horror of
swallowing, then, as Andral cheerfully acknowledged, fully authentic
cases could be caused by blows to the throat or by psychological
problems as well as by the bites of rabid dogs.2 Given a symptomatic
definition like Andral’s (or even a definition in terms of a specific
lesion), it was logical and inevitable that hydrophobia, or any other
disease so defined, would have a range of different remote causes that
varied from case to case. So it is not simply that earlier physicians failed
to notice a commonality among individual cases that we have since
managed to discern. There was no commonality: within their system of
ideas, the phrase ‘the cause of disease X’ could have no meaning.
There are ways of characterizing diseases other than in terms of symp-

toms. Early nineteenth-century anatomical studies revealed that phthisis,
a prominent disease at the time, was often associated with distinctive
caseating tumors or tubers in the lungs. By contrast, the presence of
non-caseating tubers was called tuberculosis. At the beginning of the
century, phthisis and tuberculosis were often regarded as separate con-
ditions because each was known to occur without the other. However,
when phthisis symptoms and tubers occurred together the tubers seemed
to explain the symptoms, and phthisis could be regarded as a particular
manifestation or development of tuberculosis. In 1832, Johann Lucas
Schönlein, a prominent German pathologist, wrote: ‘In recent times
phthisis has been regarded, not as a unique disease, but as a direct
sequel and higher development of tuberculosis. This opinion arose first
in [Marie François Xavier] Bichat’s school and has spread through
France, England, and even part of Germany’ (Schönlein, 1832, p. 134).
Schönlein did not agree with this way of thinking; he objected on the
grounds that some genuine cases of phthisis did not involve tubers.
However, he wrote, ‘one cannot deny credit to the pathologists who
have provided this material basis for phthisis’. By providing a ‘material
basis’, the pathologists explained symptoms and made diagnosis more
objective and precise. For these reasons, and because it seemed so
enlightening, phthisis was ultimately recharacterized in terms of tubers;
non-tuberous cases, which had formerly been genuine instances of
phthisis, were reclassified in other categories.
Because the new characterization of phthisis seemed superior and
because, at the time, phthisis was such a prominent disease, this
approach was followed in characterizing other diseases. Physicians
dissected corpses, sought morbid remains, and, when possible, redefined
diseases in terms of what they found. The new characterizations were
greeted as important advances. ‘The great improvements which have
taken place of late years in pathology, by enabling practitioners to
connect symptoms with their organic causes more accurately, have nec-
essarily diverted attention from the artificial [symptomatic] combinations
of the old nosology’ (Forbes, 1845, p. 106). However, the influence of
pathological anatomy went further than just providing new ways of
defining diseases. Thomas Watson began a lecture on morbid anatomy
by noting that the topic
is not one of merely curious interest, but has a direct bearing upon
the proper treatment of diseases. It will teach us what we have to
guard against, what we must strive to avert, in different cases. In
speaking of particular diseases, I shall constantly refer to the facts
and reasonings which I am now about to lay before you.
(Watson, 1858, p. 68)
Thus, the examination of corpses became a major source of knowledge
about the proper and improper functioning of the body – about health
and disease. Physicians were inevitably influenced not only by the content
– the specific observations and explanations – but also by the form of
the explanations that pathologists provided.
Of what form were the causes that interested pathologists? The causes
pathologists identified (and that physicians took as a basis for their
‘facts and reasonings’ about diseases) were causes of death and causes
of the successive stages of morbid processes. As we have seen, it was
natural to associate causes of disease with causes of death. Indeed, it
could not have been otherwise: whenever possible, diseases were
defined as morbid processes, and unchecked morbid processes end in
death. Thus, in contemporary medical texts, the discussion of indi-
vidual diseases was often immediately preceded by a discussion of
causes of death (Watson, 1858, pp. 67–76). But, as we have seen, a
cause of death or the cause of a certain morbid alteration, like the cause
of an accident or of the malfunction of a watch, is a sufficient cause for

a particular event; one is asking: ‘Why did this person die now?’ From
their symptomatically oriented predecessors, anatomically oriented phy-
sicians inherited a preoccupation with sufficient causes for particular
events; the fascination with pathology only reinforced this preoccupa-
tion. Thus, etiology was almost unaffected by the change from
symptomatic to anatomical characterizations of diseases. The lists of
remote causes of diseases quoted above spanned, essentially unchanged,
the shift from symptomatic to anatomical characterizations.
Insofar as one is interested in explaining why an individual becomes
sick or dies, it is (from our point of view no less than from that of an
1840s physician) logical and correct to look for a sufficient cause for
that single event and such a cause would inevitably include a variety of
both predisposing and exciting factors. The result will be an historical
explanation of a sort familiar in the biological sciences (Kamminga,
1993).3 However, in the early nineteenth century, given symptomatic or
anatomical characterizations of diseases, there simply was no other way
of thinking about causes. In respect to remote causes, the nearest a
nineteenth-century physician could come to our concept of ‘the cause of
disease X’ was a collection of different causes found in individual cases
– just the sort of collection Scanzoni and Chomel sought to accumulate
in studying childbed fever and pneumonia. Thus, in the 1840s physi-
cians focused on a kind of cause that receives relatively little medical
attention today, while causes of the sort that dominate our etiological
thinking were almost entirely absent. To help understand this striking
contrast, we will now consider one final aspect of early nineteenth-
century etiology.
Even a casual reader of the medical literature from the early nineteenth
century must be struck with the extent to which medical theory and
practice were bound up with traditional morality.4 Etiological discus-
sions of most diseases included references to such factors as drunkenness
(1:79),5 intemperance (2:595), gluttony (1:89), luxury (1:219), indul-
gence (4:331), debauchery (2:209), dissipation (4:137), vicious habits
(2:559), solitary vice (4:139), excessive venereal indulgences (1:72),
lustful excesses (2:621), indolence (1:48), sloth (1:739), envy (2:631),
jealousy (2:631), and anger (1:116). Physicians observed that excess
was the foundation of most disease (Barlow, 1845a, p. 751). Indeed, in
the etiological discussions of one multi-volumed and multi-authored
medical encyclopedia, such terms as ‘excessive diet’, ‘gross and luxuri-
ous diet’, ‘gluttony’, ‘high feeding’, and ‘full and stimulating diet’ appear
four times as frequently as terms like ‘insufficient food’ or ‘want of
stimulating diet’ (Dunglison, 1845). Through more than 3000 pages of

text, ‘excessive venereal indulgence’ is repeatedly mentioned as a cause
of disease, but celibacy is mentioned only once – as a possible cause of
insanity (Prichard, 1845b, p. 48). The specific choice of terms in these
etiological discussions is crucial – by identifying such factors as indo-
lence (as opposed to inactivity), drunkenness (as opposed to alcohol
intoxication), and solitary vice (as opposed to masturbation), moral
concepts are invoked that necessarily permeate all subsequent theory
and practice. If immorality causes disease, virtue becomes an important
prophylaxis: ‘Moderation in the pursuits of pleasure, of study, and of
business; strict temperance and virtuous habits; may be said to comprise
all that is most likely in our mode of living to give protection through-
out life against the occurrence of scrofulous disease’ (Cumin, 1845, p.
140). ‘Few things are better preservatives against infection than forti-
tude and equanimity. Nothing, we are informed by those who voluntarily
exposed themselves to the contagion … was found so great a preserva-
tive against its effects, as a steady adherence to what they believed their
duty, banishing from their minds … all thoughts of danger, and avoid-
ing every kind of passion’ (Tweedy, 1845, p. 177).
Adherence to traditional standards was not only a protection against
disease, it was central to treatment. One frequently reads that the
‘moral management of the patient’ constitutes an important or even the
‘most efficacious’ part of therapy (Prichard, 1845a, p. 560). Thus, in
both prophylaxis and therapy, the nineteenth-century physician incul-
cated and defended traditional moral and social norms; ‘his main
endeavor was to see that individuals were capable of playing their social
roles successfully in a traditional structure of social position. Illness was
for him a mark of undue deviation from the norm’ (Turner, 1967, p.
392). In this quotation, Victor Turner is discussing the social role of one
chimbuki (which Turner translates as ‘doctor’ although he points out
that ‘“ritual specialist” or “cult-adept” would be equally appropriate’)
among the Ndembu, a tribe in Zambia (Turner, 1967, p. 359), but,
without much of a stretch, the quotation also applies in the present
context.
Understanding this aspect of the physician’s role, may shed light on
the dominance of medical interest in sufficient causes of particular
events. Suppose one is interested in discouraging some practice (for
example, solitary vice). A natural response is to portray that practice as
liable to bring on undesirable consequences like disease (for example,
blindness). However, many immoral people do not become sick (so
moral infraction by itself cannot be sufficient) and some people become
sick without being immoral (so the infraction cannot be necessary or
universal) – it can only be part of a complex cause that is sufficient for
some particular events. Mary Douglas drew attention to the frequency

with which conformity is encouraged (both in primitive societies and in
our own) by portraying deviancy as a threat to such assets as time,
money, nature, or divine favor (Douglas, 1975). Douglas does not
mention health as one such asset but, clearly, endangerment of health is
used at least as frequently for the same purpose. Representing infrac-
tions of moral and social norms as liable to bring on disease is a
natural, almost universally espoused, and (possibly) even mildly effec-
tive strategy for influencing behavior. One must expect any culture
whose physicians are concerned with reinforcing norms to focus on
sufficient causes for particular disease episodes (Carter, 1991).
Nineteenth-century physicians were more interested in questions like
‘Why does this person now have mumps?’ than in questions like ‘How
does mumps come about?’ Exactly the same priority of interests has
been repeatedly noted in explanations of misfortune in primitive socie-
ties (Foster, 1976; Frankfort et al., 1949, p. 25; Evans-Pritchard, 1976,
pp. 19–23). The reason is always the same: the primary objective is not
controlling or explaining misfortune but reinforcing norms. Of course,
this should not be construed as minimizing the significance or social
utility of chimbuki-medicine:6 Whether in twentieth-century Zambia or
in nineteenth-century Europe, the wilful violation of norms is far more
destructive than illness.
Years ago Erwin Ackerknecht warned that ‘a fundamentally rational

therapeutic method may be misunderstood because of its magic, seman-
tic cloak’ (Ackerknecht, 1946, p. 474). A sort of reverse confusion is
also possible: What appears to be an essentially scientific vocabulary
may cloak beliefs, concepts and objectives totally alien from our own.
In the early nineteenth century, physicians spoke so much as we now
speak that we see continuity where there was fracture and we overlook
strands of their language that bind them inextricably to other systems
of thought. Castiglioni describes early nineteenth-century physicians as
scientists (Castiglioni, 1947, p. 760) and Foucault celebrates the rise of
pathological anatomy as the origin of a science of the individual
(Foucault, 1973, p. 197). In fact, scientific medicine is a more recent
development that began with the rise of a research programme focusing
on causes of disease.7 One superficial clue marking the origin of this
programme is a change in talk about causes of diseases. This change is
connected to more fundamental changes in the concept of disease, in
the organization of medical knowledge, in the selection of therapies,8
and in professional expectations of physicians (Carter, 1993; Schlich,
1996).
Earlier causal talk was not wrong or irrational. It was part of a

reasonable, coherent, and remarkably tenacious system of thought and
action that had enormous social utility (Rosenberg, 1979, pp. 1–5).
However, it was profoundly different from our own and was driven by
interests quite different from those that have dominated the practice of
medicine since about the middle of the nineteenth century.
Notes
1. Through this sentence, such numbers as the preceding are volume and page
references to Dunglison (1845).
2. As this case illustrates, while early nineteenth-century physicians used
many of the same disease names we use today, such names applied to
dramatically different sets of cases, and they had altogether different mean-
ings than they now bear. Michel Foucault (Foucault, 1973) and others
(Schlich, 1994; Wilson, 2000) have made this point, but the lesson is
difficult to learn.
3. For an independent analysis of causes of disease in individual patients that
reaches similar conclusions see Wulff (1984). Michael Worboys has, ap-
propriately, emphasized that ‘germ theories of disease, explicitly or implicitly,
always included ideas about the interactions between germs and bodies …
the seed and soil analogy, or some variant, was routinely used to explain
both disease and its absence from the first uses of modern germ theories of
all types’ (Worboys, 2000, p. 281). John Harley Warner has discussed an
exactly parallel focus on individual cases in the therapy of the period
(Warner, 1986, pp. 58–80).
4. Charles E. Rosenberg has given attention to this issue but from a different
perspective (Rosenberg, 1989).
5. Through this sentence, numbers such as the preceding are volume and page
references to Dunglison (1845).
6. I occasionally use the phrase ‘chimbuki-medicine’ as a convenient label for
any medical system in which immorality, per se, is regarded as a cause of
disease in the sense that it is taken as part of a sufficient cause for particu-
lar episodes of illness. No disrespect is intended either to nineteenth-century
practitioners or to the chimbuki.
7. Lakatos pointed out that, in the absence of a research programme, there is
‘a mere patched up pattern of trial and error’ (Lakatos, 1968, p. 175) – so
something less than mature science. Arguably the first research programme
in medicine (that is, the first attempt to understand and control diseases by
the formulation of scientific theories) is the one we are investigating.
8. ‘Between the 1820s and the 1880s medical therapeutics in America was
fundamentally altered. Traditional medical practices, founded on assump-
tions about disease shared by doctor and patient and oriented toward
visibly altering the symptoms of sick individuals, began to be supplanted
by strategies grounded in experimental science that objectified disease
while minimizing differences among patients’ (Warner, 1986, p. 1; see also
Carter, 1982a).
CHAPTER TWO
Universal Necessary Causes

In 1844, Jacob Henle and Carl Pfeufer founded the Zeitschrift für
rationelle Medizin. The first issue contained a long essay in which
Henle distinguished three approaches to medicine: a speculative or
theoretical approach, strict empiricism, and rational medicine (a combi-
nation of theory and observation). In his essay, Henle was sharply
critical of existing causal thinking in medicine. He wrote that typical
etiological discussions were so fallacious that medicine appeared ridicu-
lous in comparison to the exact sciences, and he continued:
Only in medicine are there causes that have hundreds of conse-
quences or that can, on arbitrary occasions, remain entirely without
effect. Only in medicine can the same effect flow from the most
varied possible sources. One need only glance at the chapters on
etiology in handbooks or monographs. For almost every disease,
after a specific cause or the admission that such a cause is not yet
known, one finds the same horde of harmful influences – poor
housing and clothing, liquor and sex, hunger and anxiety. This is
just as scientific as if a physicist were to teach that bodies fall
because boards or beams are removed, because ropes or cables
break, or because of openings, and so forth.
(Henle, 1844a, p. 25)
This paragraph contains three different criticisms of existing causal
thought in medicine and, by implication, three ideals that physicians
should strive for in identifying causes. (1) In the first sentence Henle
objected to accepting, as causes, factors that can have different effects
or even no effect at all. Thus to be acceptable to Henle, a cause must
always have the same effect; in other words it must be sufficient for its
effect. (2) In the second sentence Henle objected to accepting, as causes,
factors that need not be present for a given effect to occur (because the
same effect could flow from other possible causes); in other words, he
objects to factors whose absence does not insure the absence of the
effect. Thus, to be acceptable to Henle, the absence of a cause must
always accompany the absence of the effect; in other words, the cause
must be necessary for its effect. In Henle’s view, only factors that are
both sufficient and necessary should be accepted as causes in rational
medicine.
(3) The last two sentences contain a third criticism. Henle objected
that for most diseases a ‘horde of harmful influences – poor housing
and clothing, liquor and sex, hunger and anxiety’ – were recognized as
UNIVERSAL NECESSARY CAUSES 25
possible causes of each disease. This is exactly what we saw in our

account of remote causes in the last chapter. Henle continued: ‘This is
just as scientific as if a physicist were to teach that bodies fall because
boards or beams are removed, because ropes or cables break, or be-
cause of openings, and so forth.’ Henle believed physicists seek a single
common cause (perhaps gravitational attraction) to explain each in-
stance of a class of similar events (instances of falling) instead of
explaining those events in terms of the unique circumstances of each
event itself (for example, the removal of a beam). His idea was that
rational medicine should follow this example. In medicine, the counter-
parts to the different instances of falling are episodes of illness. So
Henle is calling for the identification of universal causes – causes that
are also necessary and sufficient – to explain every episode of each
different disease. Later in the same essay, Henle observed: ‘explanation
is always the unification of the special under universal law’ (Henle,
1844a, p. 31). Henle was interested in causes that were universal,
necessary, and sufficient.
In fact, by 1844, when Henle published his essay on rational medicine,

a few researchers had already identified causes that achieved his ideal.
In 1835, Simon-François Renucci and Philippe Ricord in France and
Agostino Bassi in Italy associated scabies, syphilis, and muscardine with
a parasitic insect, a ‘special ferment’, and a minute fungus respectively.1
Each of these causes was universal, necessary, and sufficient. In the
same year, James Paget discovered encapsulated worms in human mus-
cle tissue that, by 1860, were recognized as the universal, necessary and
sufficient cause of trichinosis. In 1837 Alfred Donné described a para-
sitic protozoan, Trichomonas vaginalis, associated with inflammation
of the vagina, and one year later Angelo Dubini discovered the hook-
worm parasite (Kunitz, 1988). In 1839 a German pathologist, Johann
Lucas Schönlein, discovered that favus was always due to a minute
fungus, and in the early 1840s David Gruby, an Hungarian microscopist
working in Paris, described this fungus more adequately and traced
several other human skin disorders to other fungi. These nearly simulta-
neous discoveries, most of which were known to Henle,2 were nearly all
independent of one another. How did each of these persons conclude –
in stark opposition to the usual ways of thinking – that a particular
cause was universal, necessary and sufficient for a certain disease? As a
sample we will consider Renucci, Bassi, Gruby, and early research that
culminated in the recognition of trichinosis.
The background to Renucci’s claim to have discovered the cause of

scabies is long and complex.3 For more than 200 years, various persons
had associated mites with the disease. However, the connection remained
unclear and controversial.
In 1828, Jean Lugol, a French medical professor, having taught for
many years that scabies was due to acari, decided that it was ‘unworthy
of an hospital professor to call on his pupil to believe anything in the
absence of ocular demonstration’ (J.F.S., 1836, p. 59). He and his
students undertook a thorough quest for acari, but several days of
intense effort failed to reveal the elusive parasite. Thereupon Lugol
‘became … a decided skeptic, and so fully was he convinced of the
impossibility of finding the insect that … he declared he would give a
prize of three hundred francs to the first student who should extract an
acarus in his presence’. Over the next six years, several persons sought
the insect, but none was successful and the prize remained unclaimed.
Then, in August 1835,
a girl presented herself at the consultation room of the Hôpital St.
Louis, to be treated for what she called the itch. Some doubt
arising as to the exact nature of the eruption, M. Renucci, an
Italian student, offered to remove all difficulty as to the diagnosis,
by ascertaining the presence or absence of the acarus, which he
said was so commonly found in cases of itch in his country, that
the peasants extracted them from each other with pins or needles
(J.F.S., 1836, p. 60).
Renucci quickly located a specimen that, to the delight of onlookers,
‘exhibited the power of locomotion, scampering about with activity,
unaware of the noise it might make in the annals of science’. Upon
hearing of Lugol’s challenge, Renucci claimed the prize. But Lugol had
not witnessed the extraction of the acarus and remained unconvinced.
Renucci offered to produce live acari in Lugol’s presence and, if neces-
sary, before the entire medical faculty. ‘The affair now began to cause
some excitement; nothing was talked of, or looked for, in the hospital,
but the acarus; the wards allotted to itch patients, heretofore so quiet,
were now thronged with students and visitors, anxious to discover or
view the long-disputed insect.’ On the appointed day, Renucci produced
more acari. Moreover, under Renucci’s tutorship, ‘the searchers for
acari soon produced them at last en masse, and during repeated sittings,
and thus forced conviction on the unbelievers’ (J.F.S., 1836, p. 60).
Renucci was awarded the prize as well as a gold medal and, on 6
September 1835, his work was reviewed before the Academy of Science.
While existence of the acarus was no longer in doubt, its exact
relation to scabies remained controversial. A British observer proposed
that, instead of causing scabies, the acarus might ‘show itself in the
individual merely in consequence of the attraction produced by the itch

matter or the filth attached to the person’ (J.F.S., 1836, p. 62). A French
physician wondered whether the insect might only be an accidental
complication of the disease, and pointed out that, even if it were a
cause, it might still not be the only possible cause (Biett, 1836, p. 547).
To clarify the relation between acari and scabies, Albin Gras, another
medical student at St Louis Hospital,
submitted his arm to a troop of these parasitical insects, and ob-
tained a development of some characteristic vesicles. A subsequent
intolerable itching, combined with the external characters, left lit-
tle doubt as to the power of these insects to communicate the
disease. But still the question is not decided, because the matter
adhering to the insects may have been the cause of the vesicles,
instead of the irritation simply produced by its presence. It has
indeed been proposed by one of the professors (seriously?) to sub-
mit the insect to the action of a warm-bath before inserting it
under the epidermis, and to pay particular attention to washing,
brushing, and drying its feet!
(J.F.S., 1836, p. 62)
According to the author of this account, experiments were still in
progress.
A short time later, Gras reported the results of his self-inoculations
(Gras, 1836). He claimed that acari appeared always and only in sca-
bies, and that the insects were the only cause of the disease. In support
of this claim, Gras reported that he had given himself scabies by insert-
ing acari under his skin and that the ensuing disease had all the
characteristics of ordinary scabies and could be further transmitted. On
the other hand, he observed that the serous fluid associated with scabies
could be introduced under the skin without effect. He judged that the
insect was too complex and too perfect to have come about by sponta-
neous generation. This implied that scabies resulted only when living
acari invaded a new host from an earlier one. If acari were not the cause
of scabies, Gras asked, how could it happen that they were always
present even in early stages of the disease. In the discussion of Gras’s
paper, respondents seemed willing to accept acari as one cause of sca-
bies, but they questioned whether they were the only cause.
While some critics remained unpersuaded, to most, the central facts
of the etiology of the disease seemed clear. In 1840, five years after
publication of Renucci’s first account, Jakob Henle concluded that sca-
bies was the only disease in which human infection through a plant or
animal parasite had been conclusively demonstrated (Henle, 1840, p.
971).
Agostino Bassi began studying muscardine, a fatal disease of silkworms,

as early as 1807. Born in Lombardy on 25 September 1773, he worked
as a civil servant in Lodi, but also managed his father’s farm at Mairgo,
four miles out of town. Bassi explained that, at the beginning of his
research, he believed that the disease ‘was due to some difference in the
atmosphere, the food, or the method of breeding, or rather to the
various fumes emanating from the fermenting litter’(Bassi, 1835, p. 4).4
Bassi tried to produce muscardine by subjecting silkworms to various
harmful influences. He reported breeding ‘silkworms in all ways, even
subjecting them to the most barbarous treatment: the wretched crea-
tures died by thousands and in a thousand ways’, but, when dead, none
displayed the hard white crust that typified the disease. Bassi reported:
I inflicted the cruellest treatment on [the worms], and used several
kinds of poisons, mineral, vegetable and animal; I employed irri-
tant, corrosive, and caustic substances, now simple, now compound;
acids, alkalis, earths, and metals, in short the most noxious sub-
stances, deadly to the animal organism, in the solid as well as in
the liquid or gaseous states; but everything proved useless for my
purpose. There is no chemical substance nor any product of the
perverted animal economy which can cause the terrible muscardine
in the silkworm. (p. 9)
Given existing beliefs about disease causation, one would have expected
these measures to have provoked various disorders including muscardine.
However, Bassi had no success. He reported that he continued these
experiments for five years, from 1808 until 1813, ‘wasting money and
journeys and labor in vain, without ever in the least succeeding in my
purposes’.
As he later described his work, Bassi seems then to have changed his
approach: he began trying to simulate the symptoms of muscardine.
Assuming that a disease is a group of symptoms, by producing the
symptoms one could produce the disease. Bassi believed that the typical
white crust consisted of an excess of ‘earths and phosphoric acid’, and
he fed the worms these substances. His results seemed favorable: ‘the
corpses of the worms into which I had introduced these substances with
their food were preserved from putrefaction or decayed much more
slowly than the others’ (p. 4). He found that by frequently soaking
silkworms in liquified phosphoric acid and lime ‘their corpses remained
uncorrupted like the worms killed by calcinaccio and almost as white’
(p. 5). But the dead worms were not as hard as those that died natu-
rally; so the analogy was not yet perfect.
Then Bassi tried this procedure:
I hung little paper bags, each containing a large silkworm near
maturity, at various heights in the chimney of a fireplace in which a
fire was burning continuously. When, after several days, I opened

the bags, … I found several that were solid and hard like calcified
worms. I maintained a given degree of moisture on the surface of
these by putting some in the cellar and placing the others under
small glasses, taking care to wet the surface of these last every day:
under these conditions, some became covered with a white erup-
tion exactly similar to that found on muscardined silkworms; and,
since they were as firm and hard as these, they had all the appear-
ance of true calcified worms, and a large number of silk worm
experts to whom I showed them without saying anything of their
origin, judged them all without exception to be true calcified worms.
(p. 5)
Bassi was encouraged, but he soon realized that these results were also
imperfect: although the corpses were superficially indistinguishable from
those of muscardine worms, they lacked ‘the power of contagion’.
Healthy worms exposed to these hard white corpses remained healthy,
while worms exposed to true muscardine corpses died of the disease.
Bassi saw this as a crucial difference between his artificial cases and real
muscardine. Coming at the end of nine years of unsuccessful attempts
to produce the disease, this realization was a major disappointment:
Bassi tells us that he fell into ‘the deepest dejection … Humiliated in the
extreme, silent and idle, I wept over my lost laurels and bitterly la-
mented the adverse fate which had put me to so much study, expense,
and labor in vain’ (p. 5).
Then came a turning point. Bassi described what happened:
It was the year 1816: oppressed by a terrible melancholy, which
had already beset me for many months in various ways and from
different directions,5 I found the courage one day to shake off its
yoke, and, challenging adverse fortune once more, I began to inter-
rogate nature afresh in various ways, with the firm resolve never to
abandon her until she had been tamed and answered my questions
honestly. Having failed by so many different processes to produce
muscardine in the silk worm except by the use of the true calcified
worm, it occurred to me that it did not originate spontaneously in
the insect, and that it needed an extraneous germ which entered the
insect from outside and caused the disease. (pp. 5f.)
Thus Bassi concluded that muscardine arose only when a germ was
conveyed from a diseased worm to a healthy one. ‘No product of the
living body or of the perverted animal economy, no simple or com-
pound substance, whether of the animal or vegetable kingdom, is capable
of causing the disease. The organism that I am about to describe alone
has the power of achieving this result’ (p. 10). Bassi went on to describe
what is now known as the fungus Beauvaria bassiana: ‘This murderous
creature is organic, living, and vegetable. It is a cryptogamic plant, a
parasitic fungus.’
Bassi concluded that, after the fungus penetrated the skin of dead
worms, it released germs that could infect new worms.
The contagion is communicated by food, by inoculation, and by
the mere contact of insects that have died of the terrible disease, of
any infected object, and even of air contaminated by the disease-
bearing germs. These are so abundant in a single calcified or
efflorescent insect, and so minute, that they spread with extreme
rapidity and in infinite numbers, cling fast even to the smoothest
and most polished objects, such as glass, metal, etc., and even rise
up into the air which they partially infect so long as they remain
suspended in it. (p. 8)
Bassi found that simply touching a healthy worm with a contaminated
pin would usually (but not always) bring on the disease (p. 9). Espe-
cially in high temperatures or in high humidity, the disease did not
always occur (pp. 31, 36). He also noted that other larvae were vulner-
able to muscardine and that, in forests, one occasionally found dead
worms covered with the same fungus; from these larvae he was able to
infect healthy silk worms (p. 9). Later Bassi demonstrated that blood
and tissue from diseased worms could also infect other worms. This
proved that the internal tissues contained elements of the contagious
fungus even before the appearance of the white excrescence.
Bassi’s understanding of the role of the fungus enabled him to explain
various facts about muscardine. The ease with which fungus germs
were carried through the air or became attached to animals explained
how the disease could spread to areas far from any known cases; the
vast numbers of germs released from each corpse explained how the
disease could suddenly destroy all the densely concentrated inhabitants
of a silkworm facility; the growth of the fungus explained the hardening
of the worm’s corpse (p. 14). Bassi’s discoveries also suggested practical
ways of controlling muscardine. He recommended strict cleanliness to
prevent the dissemination of germs. He found that, when exposed to
fresh air, muscardine germs gradually became inert, so he recommended
that breeding facilities be well ventilated. Bassi also identified chemical
washes that could disinfect eggs and even live silkworms after contami-
nation.
As Bassi recognized, his work hinted at the possible role of parasitic
organisms in other diseases. ‘My book … may perhaps resolve some of
the many anomalies which the doctrine of contagious disease in general
contains’ (p. 3).
The germs of the various animal and vegetable contagions travel
about, being transported hither and thither by so many bodies,
dead and alive, organic and inorganic, and on the wings of the
wind – especially the lightest germs and those that can live without
a liquid or an animal mucus to envelop and preserve them. …
These are the effects produced in general by the contagions that

afflict man and other animals, and that live for a long time, or at
least over a year. (p. 34)
Bassi speculated that parasitic organisms could cause a variety of conta-
gious diseases: ‘Who knows whether some species of contagions that
afflict man are not also vegetable? Indeed I suspect that the Cholera
morbus is of this nature’ (p. 39). In a later paper, Bassi compared the
spread of muscardine to the spread of infectious human diseases. Such
analogies suggested that several human diseases could also be caused by
‘parasitic beings, either animal or vegetable, which germs pass from one
individual to another and become more harmful and fatal to the poor
patient whom they invade’ (Major, 1944, p. 102). Bassi conjectured
that smallpox, spotted fever, bubonic plague, gonorrhoea, rabies, and
syphilis may be produced by minute vegetable or animal parasites.
Bassi’s account, published in 1835, did not achieve immediate suc-
cess. In 1836 the French naturalist Jean Victor Audouin reported further
studies of muscardine. Audouin observed that when white excrement
material from a muscardine worm was inoculated under the skin of a
healthy worm, the inoculated worm invariably contracted a malady
that resembled muscardine ‘in its symptoms and consequences’ (Audouin,
1836a, p. 231). But Audouin found it ‘extraordinary’ that a plant could
parasitize a living animal, and he sought proof that the inoculated
fungus actually developed within the living worm rather than simply
remaining inert and flourishing only after the worm died. He examined
worms microscopically at different intervals after they had been inocu-
lated, and he described the stages by which the inoculated fungus
spread through the living worm’s body, gradually destroyed its tissues,
and finally caused its death (Audouin, 1836a, p. 241). Audouin re-
peated his experiments several times, sometimes in the presence of
observers, and published precise accounts of his work. In contrast to
Bassi’s rambling and impressionistic book, Audouin’s publications satis-
fied existing standards of scientific proof.
Like some of Bassi’s other critics, Audouin believed that muscardine
could occur spontaneously under conditions that seemed to preclude
any chance of contagion (Audouin, 1836b, p. 269). In allowing for
spontaneous muscardine, Audouin seems not to have meant that the
disease could occur independently of the fungus (because he identified
victims of spontaneous muscardine by the characteristic excrescence on
their corpses). Audouin tried to demonstrate the spontaneous origin of
muscardine by maintaining silkworms at suitable temperatures and
humidity in sealed, moss-filled containers; he found that such worms
could become diseased, die, and produce a characteristic excrescence.
Thus, Audouin believed that the fungus itself could arise by spontane-
ous generation. He reported that he caused true muscardine by inocu-

lating healthy silkworms with fungus taken from such spontaneous
cases of the disease. He also confirmed Bassi’s conclusion that muscardine
was not unique to silkworms but could attack other larvae (Audouin,
1836b, p. 265). The excrescence from these worms could also propa-
gate the disease. This suggested that muscardine could arise spontaneously
among the larvae indigenous to some location and then spread, by
contagion, to any nearby silkworm facility.
In 1838 a commission reviewed the muscardine literature for the
French Academy (Duméril et al., 1838). The commission concluded
that inoculating healthy worms with pins that had been injected into
muscardine worms, as Bassi had done, did not show that the fungus
was the causal agent – perhaps body fluids adhering to such pins
induced the disease by acting on the fluids of the inoculated worms as a
ferment (Duméril et al., 1838, p. 90). Bassi had also failed to show that
the fungus actually killed the worms by developing within them while
they were yet alive. Thus Bassi had ‘divined’ many facts without prov-
ing them, but ‘science does not consist of divinations’ (Duméril et al.,
1838, p. 100). The commission viewed Audouin’s work more favorably.
Audouin had established that the fungus actually caused muscardine
and killed worms by growing within them. The commission also ac-
cepted Audouin’s evidence that the fungus could arise spontaneously.
Two years later, Jacob Henle observed that ‘one must agree with [the
commission] that the facts that, according to Bassi’s presentation, still
permitted many objections – objections that Bassi actually experienced
– were first proved by Audouin, who combined anatomic and micro-
scopic studies with the experiments concerning contagiousness’ (Henle,
1840, p. 945).
In 1839, four years after Bassi published his book on muscardine,

Johann Lucas Schönlein, a Berlin physician and clinician, reported that
favus was due to a minute fungus (Schönlein, 1839). Schönlein
acknowledged Bassi’s influence in his work. The fungus that Schönlein
described is now known as Achorion schönleinii.
In July 1841 independently of Schönlein, and without knowing of his
work, David Gruby also reported discovery of the fungus that causes
favus.6 Gruby was a Hungarian physician and microscopist who spent
most of his professional life in Paris. His publication was more detailed
and more accurate than Schönlein’s, and, once he became aware of
Schönlein’s letter, he pointed out errors in the latter’s sketchy account.
Over the next few years, Gruby published reports on other human
mycoses – reports still deemed substantially correct. We will review a
sample of the brief causal arguments scattered through Gruby’s publica-

tions.
In the first paper, which was on tinea or favus, Gruby argued that
‘since we have not yet found any minute particle of the true tinea not
being heavily loaded with this mycoderma, they constitute a true, essen-
tial characteristic of this disease’ (Zakon and Benedek, 1944, p. 158). In
his third paper, on thrush, Gruby argued for causality as follows:
If one puts a small particle of this substance [of thrush] under the
microscope, one sees it composed solely from a mass of crypto-
gamic plant … Since we have not found … in the white substance
of the thrush anything but the vegetables and epithelial cells and
no products of inflammation, we believe it is right to conclude that
the thrush is nothing but a cryptogamic plant vegetating on the
living mucous membrane.
(Zakon and Benedek, 1944, p. 161)
In a fourth paper, Gruby described another human mycosis – probably
sycosis barbae; he wrote:
In previous communications I demonstrated that two diseases, the
tinea favosa and the thrush of children, are due to development of
certain cryptogams in the tissues of living individuals. Today I have
the honor to submit to the judgment of the Academy my investiga-
tions on a third species of cryptogam which fixes itself in the hair
sheath of the human beard causing a disease not yet sufficiently
characterized up to this time. … In examining the scales [of dis-
eased tissues] under the microscope, one recognizes they are
composed purely of epidermal cells; the microscopic examination
of the hair, however, reveals its entire dermal portion is enveloped
by cryptogams.
(Zakon and Benedek, 1944, p. 162)
In the fifth paper, Gruby reported research on a disease he called
phyto-alopecia decalvans. He observed that ‘when one examines with
attention this white dust under the microscope which covers the skin in
porrigo decalvans, one will be surprised to find the whole formed
entirely by cryptogams’ (Zakon and Benedek, 1944, p. 164). He
proposed naming the fungus ‘Microsporum audouini’ in honor of
Audouin’s ‘fine research work on muscardine [that] has contributed to
direct attention to the vegetable parasites that destroy the live tissue of
the animals’. Later, without having given any further evidence of causa-
tion, he observed: ‘Microsporum audouini which causes the phyto-alopecia
(that is the name I suggest for distinguishing this affection) has many
analogies with the cryptogams causing the disease I described under the
name of phytomentagra’ (Zakon and Benedek, 1944, p. 165).
Finally, in a sixth paper, Gruby observed, ‘In examining with atten-
tion under the microscope hair fragments of the tinea tonsurans, one
recognizes that their entire tissue is filled with cryptogams’ (Zakon and
Benedek, 1944, p. 166).
Gruby’s reasoning in these passages is easy to characterize: in each
case he inferred causality simply from finding a cryptogam to be present
in (indeed to mostly constitute) the morbid materials of all of the
observed cases of each disease. For six different diseases, he inferred
causality after observing that disease products consisted mostly of the
fungi.
Through the 1820s, English and German pathologists occasionally

reported small, white specks in the muscle tissues of corpses. Patholo-
gists speculated about the nature of the concretions, but were unable to
associate them with any disease process or with any living organism.
On 2 February 1835 small, hard bodies – called bone spiculae – were
found in the muscles of a corpse being dissected in St Bartholomew’s
Hospital in London. James Paget, a medical student at the time, reported
that, in examining some of the structures under a lens, he ‘found that
they were cysts, and almost directly afterwards ascertained that nearly
every cyst contained a small worm coiled up’ (Paget, 1866).7 The worms
were called trichinae.
In 1852 a Göttingen physician, G. Herbst, made chance observations
about the spread of trichinae. Herbst owned a pet badger that he fed
partly with scraps from his dissecting table (Foster, 1965, p. 73). When
the badger died, Herbst found that its muscles contained encysted worms
indistinguishable from human trichinae. He fed the badger to three
dogs, and later found that the muscles of all three contained encysted
worms. Since worms were distributed throughout the dogs’ muscles, he
inferred that they were spread by the blood, but, apparently assuming
that worms would already be encysted in the intestine, he was unable to
imagine how they had entered the blood stream (Foster, 1965, pp. 72f).
Herbst’s feeding experiments, a novelty at the time, revealed much
about how trichinae arrived in the muscle tissues of carnivores, but he
reported his observations in a minor journal (Herbst, 1851–52) and his
work was overlooked.
Through the 1850s there were isolated reports of trichinae, but the
worms were believed to be harmless and rare. While there was interest
in discovering how they arrived in the muscles and in tracing their
subsequent development, these matters were thought to have no medi-
cal significance. One observer noted that, at this time, most physicians
had never even heard of trichinae (Zenker, 1865, p. 103).
In 1859 Rudolf Virchow, the famous Berlin pathologist, fed to a dog
human flesh infected with trichinae and, four days later, discovered
unencysted sexually mature male and female trichinae in its intestine.

At about the same time Rudolf Leuckart observed that, after tissues
containing trichinae were digested, the worms were released from their
cysts and matured in the intestine. He also saw that mature females
were viviparous, which helped explain how worms entered the blood
stream (Leuckart, 1860). These discoveries showed that the parasites
could only enter a host through the consumption of encapsulated worms.
With these discoveries, the essential steps of the life cycle of trichinae
was clear. Leuckart noticed disease symptoms in a dog invaded by
trichinae (Leuckart, 1860) and speculated that, instead of being mere
harmless guests within the human body (as they were widely believed to
be), trichinae may prove to be ‘among his most dreadful enemies’
(Foster, 1965, p. 74). However, pathologists were still unable to associ-
ate trichinae with any specific human or animal disease.
In January 1860 a young woman was admitted to the Dresden hospi-
tal suffering from fever, abdominal and muscular pains, oedema and
pneumonia; the diagnosis was typhoid fever (Zenker, 1860, p. 563). She
was expected to recover but soon died. Friedrich Alberti Zenker imme-
diately began an autopsy looking for remains typical of typhoid; instead
the first glance immediately revealed the astounding picture of the
most thorough penetration of the muscles with unencapsulated
trichinae. Because for several years I had often been occupied with
trichinae and had already personally observed and thoroughly stud-
ied a series of cases of encapsulated trichinae and was familiar with
all the associated questions, the uniqueness and novelty of these
findings and their significance were immediately clear
(Zenker, 1865, p. 104).
Zenker still assumed that the diagnosis had been correct and that
trichinae were a coincidental complication. However, as the examina-
tion progressed, he found no morbid changes typical of typhoid, and he
became persuaded that the multiplication and dissemination of the
trichinae had itself caused the symptoms and the patient’s death. ‘The
worms, which until then had been regarded as entirely harmless, were
recognized as the cause of a horrible disease. Thus the almost purely
zoological issue became an eminently practical and pathological one’
(Zenker, 1865, p. 104). Zenker discovered free mature worms in the
woman’s intestine. He also learned that, one month earlier, she had
assisted in slaughtering a pig. He visited her village, obtained samples of
the pig’s flesh and confirmed it was infested with trichinae. He could
not establish that the woman had actually consumed raw pork, but this
seemed likely. He found that other persons in the village, who had eaten
raw meat from the same pig, had suffered similar symptoms but recov-
ered (Zenker, 1865, p. 111).
As the clinical picture of the disease became clear, physicians began

to recognize epidemics throughout Europe. It soon became obvious that
infection was more common and the disease more deadly than anyone
had guessed.
We now have a sense of the different paths by which independent lines

of research culminated in the conclusion that several particular diseases
had universal, sufficient, and necessary causes: Renucci may have con-
ducted inoculation experiments (Ghesquier, 1999, p. 49), but his belief
that acari caused scabies seems ultimately to have stemmed from a
conviction widespread among the peasants of his homeland (J.F.S.,
1836, p. 60). Bassi surmised that Beauvaria bassiana was necessary
because he was unable to induce muscardine except by contagion.
Gruby saw that the pathological materials typical of his mycoses were
actually composed of fungi. And, in 1860, Friedrich Zenker abstracted
the concept of trichinosis from the clinical and pathological picture
presented by one patient infested with unencapsulated trichinae. Thus,
by different paths and while dealing with mostly unrelated diseases,
each of these lines of research pointed to a cause similar to those Henle
called for.
However, while these innovative efforts approached Henle’s ideal of
universal necessary causes, they were all subject to a crucial limitation:
no one of them clearly indicated how (or even if) universal necessary
causes could be identified for other disorders. This was true for two
reasons: First, on practical grounds, each of the paths reviewed above
was dictated by the circumstances of the disease in question. For most
diseases, common opinion did not point to a single cause; for most
diseases, it would be impossible to subject healthy subjects to every
conceivable trauma and note the results; most disease products were
not composed of whatever was the cause; and for most diseases, research
could not begin with a known organism whose presence could later be
correlated with clinical and epidemiological observations. While each of
these paths led to a universal necessary cause for one disease, no one of
them was open to research on most other diseases.
Second, and more fundamentally, as diseases were typically character-
ized, they simply did not and could not have universal necessary causes.
If hydrophobia is an extreme inability to swallow, it really can be
caused by blows to the throat, by psychological factors, or by the bites
of rabid dogs (Andral, 1832–33). As long as diseases were defined in
terms of symptoms, different episodes of any one disease simply did not
share a common necessary cause. And no research, however brilliant,
can find what isn’t there.
A universal necessary cause can be identified for a given disease only

if the disease is characterized so that it has one. But how is this to be
achieved? Not by defining the disease in terms of symptoms or of
internal morbid alterations, since both approaches allow for various
remote causes. How else might diseases be defined? By 1860 the first
clear example had been given.
Notes
1. The most complete statement of Ricord’s theory appeared in 1838 (Ricord,

1838), but he first announced his results in 1835 (Ricord, 1835). Refer-
ences are given below for Renucci, Bassi and others mentioned in this
paragraph. I am grateful to Alex Dracobly for drawing my attention to
Ricord (Dracobly, 2000).
2. Henle observed that his own thinking about infectious diseases was influ-
enced by Bassi (Henle, 1844b, p. 302). He knew of contemporary studies
of other parasitic diseases like favus and scabies (Henle, 1844b, p. 310).
He discussed Gruby at a time when Gruby’s work was almost universally
ignored (Henle, 1844b, pp. 381f). He knew of the discovery of Trichomonas
vaginalis, and speculated that it may be causally associated with syphilis
(Henle, 1840, p. 969). Henle himself was among the first to report seeing
the small structures in human flesh that Paget discovered to contain
Trichinella spiralis, and Rudolf Leuckart’s important observations about
the life cycle of trichinae were published in a journal that Henle edited.
3. Danièle Ghesquier has recently examined this interesting story (Ghesquier,
1999). I agree with Ghesquier’s conclusion that ‘the reappearance of the
itch-acarus was the very beginning of [the] change in concept from suffi-
cient to necessary cause of disease’ (p. 54). However, one must bear in
mind that Bassi and Ricord also reported discovering universal necessary
causes in 1835, the same year that Renucci persuaded the medical authori-
ties of the cause of scabies.
4. Through this section, page numbers in parentheses are references to Bassi
(1835).
5. Just at this time, Bassi was forced into early retirement from his govern-
ment post because he was losing his sight.
6. For translations of Gruby’s main articles and useful references, see Zakon
and Benedek (1944).
7. William C. Campbell has given an account of the discovery of trichinae
(Campbell, 1979).
CHAPTER THREE
Etiological Characterizations
The only way of characterizing diseases so that each disease will have
a universal necessary cause is by defining them in terms of their
causes. For example, the tubercle bacillus is universally necessary for
tuberculosis only because ‘tuberculosis’ is defined as infestation by
this bacillus. One cannot be certain that no one (perhaps Fracastoro,
Paracelsus, Ugo Benzi of Siena, or Sinan ibn Thabit ibn Qurra) em-
ployed such characterizations before 1850 although, to date – loose
talk about disease specificity notwithstanding – no textual evidence
has been presented that anyone did. One reads that Fracastoro envi-
sioned defining diseases in terms of what he called ‘seeds’. But
unambiguous evidence has yet to be provided and, even if this had
been his intent, in the absence of any empirical means of distinguish-
ing among seeds (or, indeed, even of demonstrating their existence)
any such attempt would have been exactly as useless as defining
diseases in terms of assorted evil spirits; vacuous definitions explain
precisely nothing. However, this much is absolutely certain: Ignaz
Semmelweis was the first person to have adopted such definitions in
publications that are linked by citations directly into the etiological
research programme. The only way to dispute this fact would be to
exhibit citations in core works of the programme to writers before
Semmelweis who used this approach, and there simply are none.1
Semmelweis may or may not have been aware of the significance of
what he did; others may or may not have consciously followed his
example; it matters not. The facts are that Semmelweis employed this
strategy clearly and unequivocally, he did so without any identifiable
precedent (none of his teachers in Vienna ever proposed such charac-
terizations and neither did any of the horde of marginally significant
eighteenth-century English physicians who wrote on childbed fever),
and his work figured in the etiological research programme. For these
reasons we take his use of etiological characterizations to exemplify
this approach – an approach without which the systematic identifica-
tion of universal necessary causes is absolutely unthinkable.
Semmelweis’s story has been told and retold many times (although
usually by persons oblivious of the nature and significance of his contri-
bution).2 For our purposes only four claims require discussion:
1. In the mid 1840s, typical accounts of the etiology of childbed fever

ETIOLOGICAL CHARACTERIZATIONS 39
(also called puerperal fever) conformed to similar accounts for

other diseases as discussed in Chapter 1 above.
2. At least by 1858, and probably as early as 1850, Semmelweis
advanced an etiological definition of ‘childbed fever’ that guaran-
teed a universal and necessary cause.
3. This definition was the basis for a theory of childbed fever that,
while seriously flawed, was unquestionably superior to any other
existing account.
4. In spite of a persisting tradition to the contrary, Semmelweis’s work
was widely known and frequently cited between 1865 and 1890 by
many of the very persons who actively contributed to the etiological
research programme – the programme that ultimately replaced
chimbuki-medicine.
Representative accounts of childbed fever from the 1840s
Childbed fever, a particularly horrible disease, was common in the

eighteenth and early nineteenth centuries.3 A few days after delivery,
victims suffered a variety of symptoms, typically including abdominal
pain and intense fever, and they usually died within hours of the onset.
Pathologists tried to understand childbed fever by studying morbid
remains in the corpses of victims. They gave particular attention to the
uterus since the disease was associated with birth, and autopsies often
revealed morbid changes there. However, no clear picture emerged. In
some victims no morbid alterations were discernible and in others the
changes took a variety of forms and did not always focus on the uterus.
According to the principles of pathological anatomy, this implied that
childbed fever was not a single disease, but a group of superficially
similar diseases. Some obstetricians abandoned the term ‘childbed fever’
and fell back on anatomical concepts like endometritis, peritonitis, or
metrophlebitis.
In 1842, Paul-Antoine Dubois, a leading French obstetrician, summa-
rized his views about puerperal fever in a long essay in a French
‘dictionary’ (encyclopedia) of medicine (Dubois, 1842). Dubois dis-
cussed the inconsistent pathological remains. He acknowledged that
some pathologists had given up looking for a primary alteration and
had concluded that childbed fever was actually a group of several
different but symptomatically similar diseases. Against this view, Dubois
hypothesized that the victim’s blood could be the site of a common
primary alteration. If so, he reasoned, a given patient could die so
quickly that other tissues could remain unaffected. On the other hand,
if a patient lived long enough, secondary morbid changes could occur.
Thus, he speculated, in spite of the variations, puerperal fever could still

be a single disease stemming from a common corruption of the blood.
He could not identify or describe this supposed corruption, and he
admitted that there may be cases in which other alterations came first
(Dubois, 1842, p. 339).
Dubois believed the corruption of patients’ blood could be initiated
by a variety of remote causes, for example, by a miasm originating in a
maternity clinic or by the putrefaction of blood clots retained in the
uterine cavity after delivery. He also admitted that childbed fever could
sometimes be contagious, and he urged physicians to prevent direct
contact between sick and healthy patients and even indirect contact by
way of healthy third persons such as physicians (Dubois, 1842, p. 343).
Dubois knew that childbed fever sometimes occurred in epidemics. At
the time, epidemics were usually ascribed to harmful atmospheric influ-
ences that could not otherwise be detected, and he concluded that
childbed fever sometimes seemed to arise from such influences. He
noted that these influences could provoke the disease in predisposed but
otherwise healthy women. As predisposing causes, Dubois mentioned
the disruption of pregnancy, the shock of labor,
a soft and effeminate life, a milieu of luxury and abundance, the
absence of exercise sufficient to induce physical vigor and complete
vitality of the blood, … weakness produced by poverty, an un-
healthy place of residence, a diet inadequately restorative, the abuse
of spirituous liquors, various kinds of grief, debauchery, and exces-
sive work.
(Dubois, 1842, p. 348)
Because of such predisposing conditions, Dubois observed, some women
‘carry within themselves the disease rather than contracting it in the
hospital’. Exciting causes of childbed fever included
the length of labor, an abundant uterine loss, a grave eclampsia,
various maneuvers and surgical operations, something attendant
on parturition or the accidents that complicate it, more or less
serious genital lesions … the affect of cold air, the use of cold and
humid linen, cold water lotions, departures from a [normal] regi-
men, exciting drinks … , the imprudence of some women who rise
in the first days [after delivery], … and the emotional states of the
newly delivered.
(Dubois, 1842, p. 349)
Dubois observed that many of these causes had no effect except during
an epidemic. Of course, Dubois’s account of the causes of childbed fever
was entirely consistent with contemporary beliefs about disease causa-
tion in general.
In 1843, one year after Dubois’s essay appeared, the American physician
Oliver Wendell Holmes published an essay entitled ‘The Contagious-
ness of Puerperal Fever’. In 1855 the essay was reprinted with an
introductory note and an appendix containing additional references but
with no changes in the body of the text. The purpose of the essay was to
show that ‘the disease known as puerperal fever is so far contagious as
to be frequently carried from patient to patient by physicians and
nurses’ (Holmes, 1843, p. 131).4 Holmes’ conclusions and most of his
case histories were drawn from earlier British sources; he admitted,
both in the essay and again in the later introductory note, that his was a
majority view. ‘A few writers of authority can be found to profess a
disbelief in contagion – and they are very few compared with those who
think differently’ (p. 129). Indeed, one year before publication of
Holmes’s essay, Lancet reported a discussion of puerperal fever in the
London Medical Society in which ‘the chief apparent circumstance is
the diversity of opinion … as to the nature … the symptoms and the
treatment of the affection. … One fact only respecting the disease was
generally admitted, namely, its unquestionable contagiousness’ (Lancet,
1842–43). But, while Holmes had little new to say about childbed fever,
he felt that the existence of the minority, who questioned its contagious-
ness, justified his essay.
Holmes cited about twenty cases in which physicians examined or
treated patients with puerperal fever or in which they performed autop-
sies on women who died from puerperal fever and in which their other
patients subsequently contracted the disease. He warned obstetricians
against taking ‘any active part in the post-mortem examinations of
cases of puerperal fever’ and that ‘on the occurrence of a single case of
puerperal fever in his practice, the physician is bound to consider the
next female he attends in labor … as in danger of being infected by him’
(p. 168). Holmes considered cases suggesting that puerperal fever could
be produced ‘by an infection originating in the matter or effluvia of
erysipelas’. But he found the relation of puerperal fever to other dis-
eases ‘remote and rarely obvious’. Thus, while he mentioned reports
that ‘puerperal fever has appeared to originate from a continued prox-
imity to patients suffering with typhus’, these cases were so rare that
they ‘hardly attract our notice in the midst of the gloomy facts by which
they are surrounded’ (p. 165). Holmes never suggested that patients
with other diseases or corpses of persons who died from other diseases
constituted a risk to maternity patients. Indeed, he observed that
the number of cases of serious consequences ensuing from the
dissection of the bodies of those who had perished of puerperal
fever is so vastly disproportioned to the relatively small number of
autopsies made in this complaint as compared with typhus or
pneumonia (from which last disease not one case of poisoning

happened), and still more from all diseases put together, that the
conclusion is irresistible that a most fearful morbid poison is often
generated in the course of this disease (p. 162).
Each time Holmes stated precisely his main point he claimed only that
the disease is sometimes (or frequently) carried from one patient to
another (pp. 112, 129, 131). He always allowed for the same range of
other possible causes that all his contemporaries acknowledged: ‘It is
not pretended that the disease is always, or even, it may be, in the
majority of cases, carried about by attendants; only that it is so carried
in certain cases’ (p. 123). Following the British, Holmes distinguished
cases arising from infection from epidemic or sporadic cases. ‘It is
granted that the disease may be produced and variously modified by
many causes besides contagion, and more especially by epidemic and
endemic influences’ (p. 133). In the chronologically later introductory
note, he wrote that his theory ‘makes full allowance for other causes
besides personal transmission, especially for epidemic influences’ (p.
107).
When Oliver Wendell Holmes’ essay was published, Eduard Lumpe was
an assistant physician in the first section of Vienna’s charity obstetrical
clinic. In 1845, two years after publication of Holmes’ essay, Lumpe
wrote a carefully documented and thorough survey of contemporary
views of puerperal fever (Lumpe, 1845). It is a good sample of Viennese
thinking about the disease.
Lumpe gave sympathetic attention to the view, dominant among
pathologists, that what was called puerperal fever was actually a group
of symptomatically similar but anatomically distinct disorders; for this
reason, he acknowledged, many preferred the term ‘puerperal diseases’
rather than ‘puerperal fever’ (Lumpe, 1845, p. 342).5 Lumpe also men-
tioned that, while inconsistent with the meaning of ‘puerperal’, women
sometimes contracted these disorders even before delivery. He accepted
the common view that puerperal diseases usually began in the uterus.
This could be the point of attack even if autopsy disclosed only minor
changes in the uterus itself. Lumpe recognized that the morbid remains
in puerperal fever resembled those found when decaying organic matter
was accidentally introduced into the blood of a healthy patient or of a
surgeon or pathologist (pp. 350f.).
Lumpe gave considerable attention to the causes of puerperal fever.
He felt that maternity patients may be vulnerable because ‘lochial secre-
tions, the purpose of which is to remove waste matter, can be retained
or absorbed and thereby induce decay of the blood’ (p. 345). The
likelihood of absorbing harmful matter was greater following delivery

because the epithelium of the womb had been discharged and was being
regenerated and the veins were lacerated. Women were predisposed to
the disease by general deprivation, worry, fear, shame, attempted abor-
tion, and by the fear of death associated with admission to a hospital.
Exciting causes included difficult delivery, damage to tissues through
the use of mechanical devices, and the retention and decomposition of
the placenta (p. 348). Lumpe was aware of the possibility that the
disease could be contagious (pp. 345f.), and he mentioned an unnamed
British physician who ‘had gone so far as to claim that, through the
contagion of childbed fever, non-maternity patients and even men could
contract similar diseases’ (p. 348). However, he discounted this possi-
bility and seems to have regarded contagion as a minor problem.
Because of the accurate morbidity records of Vienna’s General Hospi-
tal, Lumpe made several astute observations about the incidence of
epidemic puerperal diseases. He felt the existence of epidemics was
‘demonstrated by occasional increases and decreases in the number of
cases without any change in the factors most commonly recognized as
causes, by the simultaneous occurrence of numerous cases, and by the
similarity of the course of simultaneous cases’ (pp. 342f.). He observed
that the disease was more prevalent in winter than in summer; this also
suggested that the disease may be epidemic and caused by atmospheric
conditions. He knew that the disorder was almost always associated
with maternity clinics – women who delivered at home, or even on the
street as they were on their way to the hospital, seldom became ill. He
concluded, as others had before him, that the disease was usually caused
by harmful miasms that, once generated, persisted in maternity clinics
because of inadequate ventilation.
To Lumpe, the incidence of disease at Vienna’s General Hospital also
suggested a miasmatic origin. The hospital’s maternity clinic was
divided into two sections and patients were assigned to the sections
according to the day on which they happened to arrive. Since 1840 all
the male obstetrical students had been trained in the first section and
female student midwives in the second. From then on, mortality had
consistently been higher in the section for obstetricians than in the
midwives’ section. ‘At the height of an epidemic, when ten to twelve
patients die in the second section, the number of deaths in the first
section will be four or five times greater’ (p. 347). He could imagine
only one explanation for this difference: since there were more births
each year in the first section, the second could be cleared and ventilated
more frequently. Thus, Lumpe supposed, the difference in mortality was
due to local miasms that were dispelled from the second section but
retained in the first. Lumpe wrote, ‘let anyone who doubts the power of
local miasmatic influences explain this difference in any other way’

(p. 347).
Semmelweis’s etiological characterizations
On 1 July 1846 Ignaz Semmelweis became assistant physician in the

obstetricians’ section of the Viennese maternity clinic – the same section
in which Lumpe had served two years earlier.6 At the time, mortality
from childbed fever averaged about 7 per cent in the first section and
about 2 per cent in the second. The Viennese maternity facility was
relatively healthy – even 7 per cent mortality compared favorably with
similar clinics around Europe. But medical personnel knew and were
troubled by the difference in mortality between the two Viennese
sections.
After a few months, during which Semmelweis was tormented by the
relatively higher mortality rate in his section, it occurred to him that the
difference could be due to contamination on the hands of students in
the first section. Student obstetricians in the first section, but not the
student midwives in the second, were required to perform dissections.
In May 1847 Semmelweis began requiring all the medical personnel in
his section to wash in a chlorine solution. The washings were intended
to destroy decaying organic matter from the morgue with which their
hands may have become contaminated. Mortality in the first section
immediately fell to about 1 per cent – a level at or below that main-
tained in the second section.
Other experiences soon convinced Semmelweis that decaying organic
matter capable of causing puerperal fever could also be generated within
the patients themselves. This could happen because of disease, because
of retained lochia or placental remains, or through the decay of body
tissues damaged in delivery. He estimated that these sources of decaying
organic matter caused fever in approximately 1 per cent of patients.
This explained the deaths in the second section and those that occurred
in spite of the prophylactic washings, and he regarded these cases as
beyond his control.
Word of Semmelweis’s work gradually spread through Europe. This
came about partly by letters written by his students (Carter and Tate,
1991) and partly by two notices, written by Ferdenand Hebra, that
were published in a Viennese medical journal. Three years elapsed
before Semmelweis himself publicly reported his work. Finally, on 15
May 1850, Semmelweis delivered a lecture before a Viennese medical
society. Discussion of the lecture continued through three successive
meetings in June and July. Unfortunately, the lecture itself was not
published; we know it only from the secretary’s summary, from the

report of the ensuing discussions, and from a published critical response
by Eduard Lumpe who attended the meetings.
According to the secretary’s account, Semmelweis criticized earlier
attempts to explain the imbalance in mortality between the two sec-
tions. He noted that the pathological remains in childbed fever bore ‘the
greatest possible similarity to those found in pyaemia as it occurs in
anatomists and surgeons after [suffering accidental wounds while] dis-
secting corpses’ (Györy, 1905, p. 49). He described how decaying organic
matter could be conveyed on the hands of medical personnel, and he
pointed out that the disease was rare in schools for midwives except
where students also participated in dissections. According to the secre-
tary’s account, Semmelweis concluded that
puerperal fever is neither contagious nor even a specific disease.
Rather it develops when putrefied animal-organic matter, from any
disease whatsoever, regardless of whether from a living organism
or from a corpse, is taken into the blood system of a recently
delivered woman and generates the puerperal (pyemic) disintegra-
tion of the blood. … Thus, one can protect against this disease by
cleaning the fingers, the utensils, and the air. If this is done, there
will be only isolated cases of puerperal fever resulting from the
retention of decaying remains, decidual or placental.
(Györy, 1905, pp. 50f)
While less explicit than one might like, these sentences can best be
understood as advancing a new definition of childbed fever (as we will
see, this strategy becomes more explicit in later publications). The
implication is that every case of the disease is due to decaying organic
matter. However unclear the secretary’s account may be, there was no
question in the minds of those who responded to Semmelweis’s lecture
that this was his position.
In his response to this lecture, Eduard Lumpe criticized several of
Semmelweis’s claims. In his own paper, published five years earlier,
Lumpe mentioned virtually all the evidence that Semmelweis now used
to support his new concept of the disease. Yet Lumpe was astounded by
Semmelweis’s claim that every case of childbed fever had the same one
cause. While granting that chlorine washings may be useful, Lumpe
emphatically denied that every case of the disease was due to decaying
organic matter:
When one thinks how, since the first occurrence of puerperal fever
epidemics, observers of all times have sought for its causes and the
means of preventing it, Semmelweis’s theory takes on the appear-
ance of the egg of Columbus. I was myself originally overjoyed as I
heard the fortunate results of the chlorine washings; … However,
during my two years as assistant in the first section, I observed
incredible variations in the incidence of illness and death. Because

of this … any possibility is more plausible than one common and
constant cause.
(Lumpe, 1850, pp. 392f.)
At the conclusion of the discussions, Karl Rokitansky, who presided in
the meetings, ‘emphasized the incontestable utility of chlorine washings
which is admitted even by opponents of Semmelweis’s opinions’ (Györy,
1905, p. 58). This sentence shows clearly that the usefulness of the
chlorine washings was never in dispute. Everyone was more or less
willing to accept the washings, and, of course, for years the British had
advocated disinfection. However, there is no evidence in the minutes
that anyone supported or even accepted Semmelweis’s claim that childbed
fever was always due to one constant cause. That claim was the focus of
the dispute.
Before the 15 May lecture, those who responded to the second-hand
accounts of Semmelweis’s work objected only that it was unoriginal;
after his lecture, no one complained of a lack of novelty – instead they
attacked his remarkable claim that the disease had a universal necessary
cause. In an editorial published in 1850, Joseph Hermann Schmidt,
Professor of Obstetrics in Berlin, recommended that obstetrical students
have ready access to morgues in which they could profitably spend time
dissecting while waiting for the labor process (Semmelweis, 1861b, pp.
225f.). He asked how Semmelweis’s hypothesis could be reconciled with
the observation that childbed fever seldom occurred in normal deliver-
ies. He then wrote that while the resorption of decaying organic matter
could be ‘one path that leads to childbed fever, it is certainly not the
only path’. In an 1855 text, Friedrich Scanzoni, professor of obstetrics
in Prague, rejected Semmelweis’s opinion. He acknowledged that puer-
peral fever could originate as Semmelweis had described, and that
chlorine washings may be useful in such cases (Scanzoni, 1855, vol. 3,
p. 1010). Yet Scanzoni rejected Semmelweis’s views on the grounds that
the disease was usually caused by atmospheric or miasmatic influences
or by emotional trauma. In medical texts published in 1855 and 1857,
Carl Braun, professor of obstetrics in Vienna, gave extensive attention
to the etiology of childbed fever (Braun, 1855, 1857). Braun identified
30 possible causes of the disease including conception itself, the general
plethora of pregnancy, the shock of delivery, pressure of milk secretions,
the individuality of the patient, emotional disturbances, errors in diet,
defective ventilation, chilling, and the inappropriate arrangements of
some maternity hospitals. Cadaverous poison is the twenty-eighth
possible cause in Braun’s list. Braun claimed, inaccurately, that Semmelweis
regarded cadaverous poisoning as the main cause – perhaps the only
cause – of childbed fever. While admitting that cadaverous poisoning
could account for some cases of disease, Braun denied that it could have
the significance that Semmelweis (supposedly) attributed to it. In an
1858 letter to Semmelweis, D. Everkin of the Paderborn maternity
clinic wrote: ‘I could not imagine that this circumstance is the universal
cause, but I was led [by your communication] to avoid undertaking any
procedures on maternity patients after examining corpses.’ He then
warned Semmelweis that nowhere outside medicine is one so frequently
tempted by the post hoc ergo proper hoc fallacy (Semmelweis, 1861b,
p. 228). In the same year, Paul-Antoine Dubois wrote that while ‘one
could not dispense with precautionary measures to guard against conta-
gion, the contagious element is neither as effective nor as pervasive as
Semmelweis has claimed, … even before delivery other factors predis-
pose women to the disease’ (Semmelweis, 1861a, p. 458). Through the
early months of 1858, the French Academy of Medicine hosted a series
of 26 lectures on puerperal fever by prominent French obstetricians
(l’Académie, 1858). Semmelweis was mentioned in six of the lectures,
but there was little discussion of his work. Some speakers associated his
views with those of the British (l’Académie, 1858, p. 228). Some
acknowledged that chlorine washings may be useful. Only one speaker,
M. Danyau, considered Semmelweis’s claim to have identified a univer-
sal cause; Danyau rejected this claim but urged physicians to do
everything possible to avoid infecting patients (l’Académie, 1858, p.
228). In 1859 Hermann Lebert, Professor at Breslau, wrote: ‘It is ques-
tionable whether those who have died of this disease can have been
directly inoculated by poison from corpses. Semmelweis has elevated
this possibility into an entire system. In any case this would be only one
of many possibilities of conveyance’ (Semmelweis, 1861b, p. 221).
Thus, through the 1850s, several physicians acknowledged that chlo-
rine washings could be useful; some mistakenly associated Semmelweis’s
concept of childbed fever with that of the British; some mistakenly
assumed that Semmelweis was concerned only with cadaveric poison-
ing. However, Semmelweis’s view that the disease always had one
common cause was invariably either ignored or rejected. In this respect,
there was greater agreement between Semmelweis’s opponents and those
he regarded as supporters than there was between Semmelweis and his
supposed supporters. Those who realized that Semmelweis was making
this strong claim, rejected it. The few who thought they agreed with
Semmelweis did so only because they ignored what was unique and
original in his position. There is no textual evidence that, as late as
1862, anyone in Europe had accepted Semmelweis’s claim.
In 1858 Semmelweis published a short essay entitled ‘The Etiology of
Childbed Fever’; two years later he published a second essay entitled:
‘The Difference in Opinion between Myself and the English Physicians
Regarding Childbed Fever’ (Györy, 1905, pp. 61–83, 83–94). In Octo-

ber 1860 he published his book, The Etiology, Concept, and Prophylaxis
of Childbed Fever (Semmelweis, 1861a). When his book did not have
the impact he had expected, Semmelweis began publishing open letters
explaining his opinions and bitterly attacking various prominent obste-
tricians (Györy, 1905, pp. 431–511). He also explained his views in a
letter, written in English, that appeared in a British periodical.
The account of childbed fever in his publications rests on two crucial
definitions. In his first essay and in his book Semmelweis defined ‘childbed
fever’ as ‘a resorption fever dependent on the resorption of decaying
animal-organic matter’ (Semmelweis, 1861b, p. 114). He also defined
‘pyaemia’ as ‘disintegration of the blood through decaying animal-
organic matter’ (Semmelweis, 1861b, p. 117). Semmelweis used the
uncommon word ‘resorption’ in the first definition because it precisely
expressed his opinion that childbed fever occurred when decaying or-
ganic matter was resorbed – literally absorbed back – into a living
organism.
Semmelweis’s new characterizations of childbed fever and of pyaemia
have several important consequences:
1. Given these characterizations, it is true, by definition, that every

case of childbed fever is caused by decaying organic matter – there
cannot be a single exception. Semmelweis was clear about this: ‘In
order for childbed fever to occur, it is a conditio sine qua non that
decaying matter is introduced into the genitals’ (Semmelweis, 1861b,
p. 149). Semmelweis frequently asserted that, without exception,
every case of childbed fever came about through resorption of
decaying organic matter (Semmelweis, 1861b, pp. 114, 115, 120,
149). This, of course, was the claim to which everyone objected
following his 1850 lecture, but at least by the time of his publica-
tions it was a matter of definition.
2. It followed that there were no spontaneous, epidemic, or miasmatic
cases of the disease (Semmelweis, 1861b, pp. 120f, 201).
3. Given the characterizations, it is true, by definition, that every case
of childbed fever is a case of pyaemia.
4. Given the characterizations, it follows that infected surgeons and
maternity patients all died from the same disease, namely, from
pyaemia or blood poisoning.
5. Given the definitions, Semmelweis could explain precisely how
childbed fever differed from contagious diseases: ‘Childbed fever is
a conveyable but not a contagious disease. By “contagious disease”
one understands a disease spread by material generated from and
only from cases of that same disease’ (Györy, 1905, p. 72). By
contrast, on his view, childbed fever could be caused by decaying

organic matter derived from any source.
6. Finally, given the characterizations, Semmelweis could also ex-
plain precisely how his views differed from those of the British
contagionists:
The important difference between my opinion and the opinion
of English physicians consists in this: I hold that every case of
childbed fever, without a single exception, has only one cause,
namely incorporation of decaying organic matter. Of this I am
convinced. The English physicians, while they also believe that
childbed fever can be caused by decaying matter, still recog-
nize all the old epidemic and endemic causes that have been
believed to play a role in the origin of the disease.
(Györy, 1905, p. 94)
These etiological definitions and their consequences are the core of

Semmelweis’s theory of childbed fever. He himself emphasized that they
alone made his teachings unique.
By the time Semmelweis published the Etiology, many accounts of his
ideas had been circulated. Thus, by the time his own definitive accounts
appeared, everyone had an opinion about his views and few bothered to
read his book. There were only two published reviews. In 1862 Carl
S.F. Crede reviewed the Etiology in a prominent German journal of
obstetrics. According to Crede, Semmelweis called everyone who disa-
greed with him an ignoramus and a murderer. Crede wrote that
Semmelweis’s assertions
go too far and are too one-sided. In any case Semmelweis owes his
reader proof that only the one etiological condition he identifies is
always responsible. Nearly every obstetrician is still of the opinion
that a large number of cases of illness remain that originate from a
different cause, a cause admittedly as yet unknown.
(Crede, 1861, p. 407)
August Breisky, an obstetrician in Prague, published a more careful re-
view of the Etiology, but even Breisky referred to Semmelweis’s book as
‘naive’ and as ‘the Koran of puerperal theology’ (Breisky, 1861, p. 1).
Breisky objected that Semmelweis had not proven that puerperal fever
and pyemia were identical, and he insisted that other factors beyond
decaying organic matter had to be included in the etiology of the disease.
The advantages of Semmelweis’s account of childbed fever
There were substantial defects in Semmelweis’s theory of childbed fever

(Semmelweis, 1861b, pp. 42–46), and all things taken into account his
critics were probably justified in remaining skeptical. However, his

approach had two remarkable advantages, one practical and one theo-
retical. First, so long as the disease was defined to allow for various
independent causes, measures that were effective in one case could
conceivably have exacerbated other cases of the same disease.7 This
made prophylaxis and therapy so confusing it was virtually impossible
to find effective measures for prevention or treatment.8 An etiological
definition meant that every case of the disease had the same cause, thus
any prophylactic or therapeutic measures directed against that cause
that were effective in one case would be effective in every case. Thus,
the definition made possible consistent and uniform strategies for pre-
vention and treatment.
Second, Semmelweis’s approach provided a way of explaining an
amazing range of facts. From his account, in spite of its weaknesses,
Semmelweis drew explanations for dozens of facts many of which had
already been observed and recorded but never explained. For example,
Semmelweis explained why the mortality rates in the two clinics dif-
fered (Semmelweis, 1861b, p. 118),9 why those rates had changed through
history as they had (pp. 118f), why chlorine washings were effective,
why disease was rare during pregnancy or after a week or more follow-
ing delivery (p. 115), why the disease appeared to be contagious
(p. 118), why it exhibited seasonal patterns (in summer ‘the charming
surroundings of Vienna [were] more attractive than the reeking morgue
or the sultry wards of the hospital’ so students spent comparatively less
time dissecting corpses and examining patients) (p. 122), why the dis-
ease focused on teaching hospitals (p. 123), why some non-teaching
hospitals had higher mortality rates than others (p. 125), why women
who delivered prematurely or on the way to the hospital had a lower
mortality rate (in order to delay labor, women who seemed likely to
deliver prematurely were never examined by students) (p. 101), why the
disease often appeared in particular sequential patterns among patients
(p. 101), why infants seldom died of pyaemia while their mothers
remained healthy (p. 99), and why the disease appeared with different
frequencies in different countries and in different historical periods
(p. 133). Semmelweis gave coherent explanations for all these observa-
tions and more.
In this respect, the difference between Semmelweis and everyone else
who wrote on the disease – including Dubois, Holmes, and Lumpe – is
the difference between day and night. In their essays, however astute
and even beneficial, Dubois, Holmes, and Lumpe explained almost
nothing. Holmes (like others before him) observed that medical person-
nel sometimes endangered their patients; he explained this by the
hypothesis of contagion and recommended that physicians disinfect
themselves before visiting patients. Lumpe observed that mortality in

the first section was several times greater than in the second; he
explained this by the hypothesis of local miasms and recommended that
maternity clinics have adequate ventilation. Both observations were
correct; both recommendations beneficial. However, appeals to
undetectable contagions or miasms were inherently defective: the only
evidence for the existence of either were the morbid processes that the
hypothesized factors were intended to explain. Thus, both explanations
were circular and vacuous – perfect examples of what a later nine-
teenth-century researcher referred to as ‘the empty generalities of the
past [that] only appeared, superficially, to satisfy the need for causes’
(Strümpell, 1893, p. 22f). Moreover, everyone else, including Dubois,
Holmes, and Lumpe, left almost all the observations unexplained. For
example, they could not explain why women who delivered prema-
turely were safer than those who went full term or why the disease was
especially prevalent in teaching hospitals. Finally, at least as the epidemic
and miasmatic hypotheses were commonly stated, they were refuted by
events in the Viennese clinic since atmospheric and miasmatic influences
had to be the same in both sections (which were adjacent and even
shared some rooms).
All of the explanations Semmelweis gave and that others couldn’t
give rested on the existence of a single universal cause. This is because,
if the disease were so defined that its cases had many different causes, it
would be impossible to provide a unified explanation (for example, of
the reduced incidence of morbidity among premature deliveries) that
would, in effect, cut across all the different explanations that contempo-
raries gave for each of the different individual cases. The superiority of
Semmelweis’s account was soon obvious to contemporaries: In 1868,
one Berlin obstetrician noted that Semmelweis’s theory ‘provides a uni-
fied, clear, and entirely intelligible meaning for a whole series of
anatomical and clinical facts and for the disinterested experiences and
discoveries of reliable observers’ whereas ‘none of the earlier or alterna-
tive hypotheses or theories … has this characteristic to the same degree’
(Boehr, 1868, p. 403). That says it all.
In a recent book on childbed fever, Irvine Loudon criticizes Semmel-
weis’s theory in several ways and, while some of the criticisms are valid,
others are not. Some of Loudon’s failed criticisms reflect a serious
misconstruction of the logic of Semmelweis’s position. Semmelweis’s
theory rests on definitions and, while empirical observations certainly
motivated the definitions, such observations cease to have effect once
the definitions are in place. Loudon wonders how Semmelweis would
explain ‘the awkward fact’ that ‘throughout Europe … deaths from
puerperal fever [were] between 20 and 30 per 10,000 births [i.e., 0.2
per cent to 0.3 per cent] – a much lower figure than the [1 per cent] rate
in the early days of the Vienna Lying-in Hospital [that was the basis for
Semmelweis’s estimate of the rate of self-infection]’ (Loudon, 2000,
p. 98). Presumably, Loudon thinks Semmelweis would find this fact
awkward because the discrepancy suggests deaths due neither to self-
infection (which, according to Loudon’s figures, could only account for
a 0.2 per cent mortality) nor to contamination from external sources
(which, according to Semmelweis, could only account for deaths above
1 per cent). So how would Semmelweis explain the other deaths? Very
easily. Most likely Semmelweis would have admitted that his estimate of
the rate of self-infection was too high (his figure, after all, was arrived
at only empirically and even in the early days of the hospital some
decaying organic matter would have been conveyed – for example, from
patients with superficial infections or medullary carcinomas that gener-
ated decaying organic matter in his time as well (Semmelweis, 1861b,
pp. 93f.). But he could also have simply denied that the excess deaths
(any above 0.2 per cent but below his 1 per cent standard) were really
childbed fever.10 For example, suppose, from our present point of view,
a woman contracted a fever from fomites bearing strep organisms and
that no decaying matter was involved. Such a woman simply did not
have childbed fever in Semmelweis’s sense – nothing awkward about
that at all.
Loudon also finds it ‘extraordinary’ that Semmelweis denied that
‘puerperal fever was ever contagious, or that it ever occurred in the
form of epidemics’ (Loudon, 2000, p. 98). But clearly it all depends on
what one means by ‘contagious’ and ‘epidemic’: Semmelweis never
denied that the disease could spread from patient to patient (which, he
agreed with the British, happened only rarely) or that it could be spread
by way of other healthy persons such as medical personnel (which, he
agreed with the British, was the means by which it was usually con-
veyed). Semmelweis also never denied that the disease occurred more
frequently sometimes than others – that observation, after all, was the
main basis for his argument. However, given his definitions, the only
way the disease could spread at all or occur more frequently sometimes
than others was through conveyance of decaying organic matter. If
illness arose under other circumstances, once again, Semmelweis would
simply deny that the cases were really childbed fever. But how could he
deny that such other cases (which everyone knew to be childbed fever)
really were childbed fever? In exactly the same way that, once Robert
Koch had defined tuberculosis as infestation by the tubercle bacillus, he
could deny that all other cases (which everyone knew to be tuberculo-
sis) really were tuberculosis (namely, symptomatically and pathologically
indistinguishable cases ‘due to leprosy bacilli, brucella bacteria, histo-
plasma fungi, beryllium poisoning, or the unknown agent of sarcoido-

sis’ (Taylor, 1979, p. 18) but lacking the tubercle bacillus). Or, to take a
more obvious example but one from outside medicine, given current
physics, glass is not classified as a solid (that is, it does not have a
crystalline molecular structure) but rather as a highly viscous liquid
(Schrödinger, 1969, p. 63). How could the initial advocates of this
classification propose anything so contrary to what everyone knew as
that glass isn’t solid? In all such changes in scientific nomenclature, the
justification is the explanatory power of the result. However, such
recharacterizations inevitably leave a residue of cases that were in-
cluded under the old concept but do not fall under the new one. That is
why (as Foucault and others have emphasized11), the reference of dis-
ease names changes much more radically than one might suppose given
that we continue to use the same names in spite of fundamental changes
in their sense (to use Gottlob Frege’s terminology). But that is what
scientific progress is all about. ‘Every scientific advance marks some
departure from the common sense that preceded it’ (Copi, 1979, p. 195).
So, to summarize, given Semmelweis’s definitions, if there were cases
that resembled childbed fever but that were due neither to self-infection
nor to contamination from external sources (and it is not clear to me
exactly how we are to envision such cases), they were not really childbed
fever after all.
Loudon also wonders whether Semmelweis simply failed ‘to under-
stand the British view of contagion, or [chose] to misrepresent them?
[sic]’ (Loudon, 2000, p. 99), and he thinks that, because James Young
Simpson recognized that ‘puerperal fever was [sometimes] transmitted
to the patient by the fingers of the birth attendant’, he somehow antici-
pated or simultaneously originated Semmelweis’s theory (Loudon, 2000,
pp. 85–87). But (excuse me, gentle reader, for saying it again and again)
the issue is not chlorine washings or (what comes to exactly the same
thing) contaminated fingers. Everyone from Semmelweis on has admit-
ted that the British were first to recognize the danger of the latter and to
extol the virtues of the former. The issue is: how are we to define
diseases?12 If we want the theoretical and practical coherence afforded
by universal necessary causes, then diseases must be defined so that they
have them. That is what the British contagionists failed to do, and that
is what James Young Simpson failed to do; so, as they themselves were
eager to admit, Simpson, like the British generally, continued to allow
for other possible causes (as we saw, Holmes repeatedly emphasized this
point) and they thereby excluded, in principle, the very possibility of
systematic measures against the disease or coherent explanations of it.
Yet, for all the defects in his theory (and they were legion), Semmelweis
gave an etiological definition, and he recognized that the resulting
theory was better than anything else around. Now, to be sure, the subtle
shift in meaning that differentiated Semmelweis from the British would
not have mattered much to the hundreds of thousands (possibly mil-
lions) of young women sacrificed in the maternity clinics, and it may
not matter much today to physicians busy trying to keep their patients
healthy. But if one claims to understand Semmelweis’s theory (and
especially if one undertakes to explain how it is wrong or unoriginal),
one must, at the very least, be clear about this rather subtle but pro-
foundly important distinction. Because, for the last 150 years, all those
who have failed to grasp it end up either rejecting Semmelweis’s theory
for reasons that are totally irrelevant or saying that his theory wasn’t all
that original after all. Somehow, Loudon manages to do both.13
Semmelweis probably had little interest in the purely methodological
significance of his definitions; he was interested only in saving lives. But
methodologically they were singularly brilliant. The definitions effec-
tively created a new disease, albeit one called by the same name and
encompassing most (but not all) instances of what had previously been
called childbed fever. The definitions excluded from consideration all
cases that failed to have the one common cause, and that cause became,
thereby, universal and necessary. By what could be seen as nothing more
than a verbal trick – which Lumpe astutely, correctly, and precisely
characterized as an egg of Columbus (Lumpe, 1850, pp. 392f) –
Semmelweis made it possible, at least in principle, to identify reliable
strategies against the disease and to give coherent explanations of dis-
ease phenomena. Semmelweis provided not merely new facts about the
fever or practical advice for avoiding it; he offered a coherent and
empirically verifiable explanatory theory. In this respect, his account
was vastly superior to anything else that had been written on the
disease. As Semmelweis’s great contemporary Charles Darwin observed,
in reference to his own (similarly flawed yet powerful) theory of evolu-
tion, it would be impossible for an altogether false theory to explain so
much (Darwin, 1876, p. 239).
Semmelweis’s influence in the research programme
There is a long-standing tradition that Semmelweis and his work were

essentially forgotten from about the time of his death (1865) until
about 1890 when his memory was, supposedly, resurrected in connec-
tion with an international medical conference in Budapest: ‘In the twenty
years after his death, Semmelweis’s name was mentioned only on rare
occasions, and usually in uncomplimentary terms’ (Loudon, 2000,
p. 108). This view is simply false. It has been refuted conclusively and
repeatedly (Manninger, 1904, pp. 74–8; Murphy, 1946; Böttger, 1955;

Carter, 1985a), but it dies hard. To appreciate the extent to which
Semmelweis was tied to the etiological research programme, it is neces-
sary only to look at the evidence.14
During the late 1860s, a year or two after Semmelweis was beaten
and died in an insane asylum in Vienna (Carter et al., 1995), three
important Viennese obstetricians converted to Semmelweis’s views: Carl
Mayrhofer, Joseph Späth and A. G. C. Veit. All three had initially
opposed Semmelweis but each seems to have become converted by his
own independent research (Carter, 1985a). Mayrhofer was particularly
noteworthy; he sought, albeit without complete success, to trace childbed
fever to infection by vibrions (microorganisms). As we will see, through
the following decades, he was frequently mentioned together with
Semmelweis.
In an 1864 paper, Späth noted that, whatever anyone might say,
every obstetrician now believed that Semmelweis was correct and that
even attempts to control childbed fever by improved ventilation were
ultimately based on an awareness that Semmelweis was right (Späth,
1864, pp. 160f.). Späth continued, ‘even Mayrhofer’s theory, if it turns
out to be correct, can only be regarded as a further confirmation of
Semmelweis’s view, because in that theory the vibrions that originate in
decomposed animal matter are taken as the infective agents’. In 1865
A.C.G. Veit observed that in addition to confirming Semmelweis’s find-
ings, Mayrhofer’s work refuted the view that ventilation significantly
influenced the incidence of childbed fever (Veit, 1865, pp. 195f.).
Investigations of childbed fever were more prominent in the develop-
ment of the bacterial theory than one might now think. Before 1880,
studies of cholera, tuberculosis, the exanthemata, and most other infec-
tious diseases produced only meager and equivocal results. In the early
1870s, surveys of literature on the role of vibrions in disease etiology
usually focused on anthrax, relapsing fever, and especially wound infec-
tions (Steudener, 1872; Birch-Hirschfeld, 1872, 1875). Of these, wound
infections received more attention in medical literature than anthrax
and relapsing fever combined. Indeed, in this period more than half of
all publications on the etiological significance of bacteria concerned
wound infections.15 Thus, it is not surprising that both German and
French writers in the period regarded the bacterial theory of disease as
having originated from the study of wound infections (Ravitsch, 1872;
Jeannel, 1880). By the middle of the nineteenth century, childbed fever,
and so-called surgical fever were the most frequently discussed forms of
wound infections. Semmelweis was recognized as having been among
the first to connect these two diseases; by 1860, even before publication
of his book, the idea that childbed fever was a septic wound infection
seems to have been generally associated with him (Roser, 1860). ‘Before
the time of Semmelweiss [sic], the idea of the identity of puerperal fever
with pyemia was very far from the minds of the obstetricians, puerperal
fever being regarded as something essentially inherent in the relations
presented by the puerperal woman’ (Hoag, 1887, p. 832). Perhaps
because clinical cases of puerperal fever and surgical fever were readily
available in nearly every hospital, discussions of these two diseases, and
especially of childbed fever, became prominent in the literature on
wound infections. Thus, given the historical situation and the knowl-
edge available at the time, studies of puerperal fever were central to the
development of the bacterial theory.
In 1865, Mayrhofer’s views were mentioned in connection with a
paper on epidemic puerperal fever which was presented by a doctor
Kaufmann in the annual meeting of the German Natural Scientists and
Physicians (Kaufmann, 1866). After mentioning Semmelweis, Kaufmann
rejected the possibility of septic infectious matter, and endorsed, in-
stead, a miasmatic explanation of childbed fever. However, according
to the published report:
Veit (Bonn), [Franz Ludwig Wilhelm] Winckel (Rostock), Pernice
(Greifswald), and [Eduard] Martin (Berlin) were inclined to the
infectious theory. … Martin mentioned Mayrhofer’s theory of
vibrions, and Mankiewitz (Mühlhausen) referred to the injuries to
genitals, particularly in operations with forceps.
(Kaufman, 1866, p. 424)
Thus all who are reported as having responded to Kaufmann’s lecture
favored the infection theory, and Mayrhofer was explicitly mentioned.
This report appeared in the same issue of the periodical that contained
Veit’s article (cited above) in which Mayrhofer was associated with
Semmelweis.
In 1866, F.L. Winckel published an important text on obstetrics. His
detailed historical account of various theories of childbed fever included
a summary of Semmelweis’s position (Winckel, 1866, p. 264). Winckel
mentioned that Lange of Munich had been among the first to adopt
Semmelweis’s views. He also discussed and agreed with Veit’s conclu-
sion that prevention of fever depended less on ventilation than on
absolute cleanliness – a view Winckel had already supported in the
discussion of Kaufmann’s paper one year earlier (Kaufmann, 1866,
p. 424). Near the end of his account, Winckel gave detailed and positive
attention to Mayrhofer’s work and to other publications based on it
(Winckel, 1866, pp. 274f.).
Also in 1866 Leon Coze and Victor-Timothee Feltz mentioned
Mayrhofer in an historical survey that initiated one of their series of
highly influential studies of wound infections (Coze and Feltz, 1866,
p. 62). In 1866 and 1867, Semmelweis’s animal experiments were cited

by German scientists investigating infected wounds (Schweninger, 1866,
pp. 591, 597; Roser, 1867, pp. 17, 20f). In one essay, W. Roser ob-
served that Semmelweis’s concept of childbed fever had become the
accepted position among obstetricians and that even many surgeons
had been converted to the doctrine (Roser, 1867, p. 20). In a long study
of diseases of the female sexual organs, published in 1867, Veit noted
that ‘the old idea of miasmatic contagion was first challenged by
Semmelweis. He taught that childbed fever was a resorption fever occa-
sioned by infection with decaying organic matter. This view is penetrating
ever greater circles and, in a short time, will find no more opponents’
(Veit, 1867, p. 678). He also noted that, since childbed fever was the
result of a septic infection, it was essential to establish the exact nature
of the septic poison, and he mentioned Mayrhofer’s work as a possible
solution to this problem. In 1867 Mayrhofer was also mentioned in a
general review of literature on the etiological significance of micro-
organisms (Richter, 1867, p. 95).
In 1868, Max Boehr published an essay entitled ‘On the infection
theory of puerperal fever and its consequences for public health offi-
cials’. The paper, which was originally presented before the Society for
Obstetrics in Berlin, drew heavily on Semmelweis’s work and conclu-
sions. Boehr noted that Semmelweis’s theory
has the characteristic of all good pathological and physiological
theories; it provides a unified, clear, and entirely intelligible mean-
ing for a whole series of anatomical and clinical facts and for the
disinterested experiences and discoveries of reliable observers during
epidemics. None of the earlier or alternative hypotheses or theories
regarding the occurrence of childbed fever has this characteristic to
the same degree.
(Boehr, 1868, p. 403)
Boehr mentioned that in Semmelweis’s Etiology, ‘the superstitions of
our predecessors, who believed in unknown cosmic-telluric-atmospheric
influences, were dealt a severe blow, as was the belief in miasmata’
(Boehr, 1868, p. 404). Boehr cited Veit, Späth, and Mayrhofer, as well
as others who had supported the infection theory. The same periodical
reported the discussion of Boehr’s paper. The report included comments
by Max Wegscheider, David Haussmann, Eduard Martin, and three
other physicians identified as Scharlau, Krieger, and Ebell. Only Ebell is
reported as having said anything opposed to the infection theory; his
comment was that ‘in addition to contagiousness, there are still other
causes of puerperal fever’ (Boehr, 1868, p. 433). According to the
report, ‘Boehr responded that in addition to artificial infection from
external sources, he himself recognized only self-infection through foul
matter from the patient herself’ – which, of course, was exactly

Semmelweis’s view.
Over the next few years, Semmelweis and Mayrhofer were mentioned
favorably in several publications (Ferber, 1868, p. 318; Haussmann,
1870, p. 46). Beginning in 1872 they were cited in papers, both French
and German, that were among the most influential of the early contri-
butions to the bacterial theory of disease. During the 1860s, Coze and
Feltz conducted studies of infectious diseases that resulted in several
publications and culminated in a book that was cited by both Pasteur
and Koch. The chapter on puerperal sepsis began with a sympathetic
and favorable discussion of Semmelweis (Coze and Feltz, 1872, p. 259).
Friedrich Wieger and Franz Hektor Arneth, two of Semmelweis’s stu-
dents in the first clinic and eyewitnesses to his discoveries, published the
first French accounts of Semmelweis’s work; both were cited by Coze
and Feltz (Coze and Feltz, 1872, pp. 259, 261). One page later, Coze
and Feltz praised Mayrhofer as the first to identify vibrions in the
lochial discharge of puerperal fever victims. Also in 1872, in an article
in a German periodical on gynecology, Wilhelm Waldeyer of Breslau
mentioned that Mayrhofer had been the first to identify bacteria in
puerperal fever victims (Waldeyer, 1872, p. 294). Three years later,
Felix Victor Birch-Hirschfeld identified Waldeyer’s paper as one of four
that had been most influential in awakening general interest in wound
infections (Birch-Hirschfeld, 1875, p. 171). In 1873 Johannes Orth
published an important discussion of puerperal fever in Virchow’s Ar-
chives (Orth, 1873, p. 437). The paper began with a careful and
sympathetic review of Mayrhofer’s work. Orth mentioned that Mayrhofer
had attempted to use pure cultures to show that the vibrions caused the
disease. ‘Remarkably enough,’ Orth observed, ‘this obvious and conclu-
sive experiment was not sufficiently appreciated and in the next few
years there was almost no subsequent research.’ In the mid-1870s promi-
nent French writers also mentioned Mayrhofer and Semmelweis favorably
(Billet, 1872, pp. 20, 28; D’Espine, 1873, pp. 13, 139).
In 1874 Haussmann published accounts of animal experiments in-
tended to confirm the results of the similar animal experiments conducted
by Semmelweis and by Mayrhofer (Haussman, 1874, pp. 311f, 315). In
the same year Carl Rokitansky, Jun., published a study of the micro-
scopic constituents of lochial discharges. He mentioned Mayrhofer as
the first to have examined the occurrence of bacteria in lochial dis-
charges (Rokitansky, 1874, p. 172).
In 1877, Mayrhofer’s results were mentioned in Birch-Hirschfeld’s
widely used textbook on pathological anatomy (Birch-Hirschfeld, 1877,
pp. 1141f.). One year after the appearance of this textbook, Robert
Koch published his essay on wound infections. Koch relied heavily on
Birch-Hirschfeld’s review of the literature and on his textbook (Koch,

1878a, pp. 20–23, 27, 29). Koch also cited the papers by Waldeyer and
Orth mentioned above (Koch, 1878a, pp. 22f.). In his studies of puer-
peral fever, Pasteur also cited Waldeyer and Orth (Pasteur, 1879a, p.
133). Thus, both Pasteur and Koch cited both French and German
authors who discussed Mayrhofer’s work. While neither Pasteur nor
Koch ever mentioned either Semmelweis or Mayrhofer, Mayrhofer was
mentioned in a history of bacteriology written by Friedrich Loeffler,
Koch’s associate and colleague (Loeffler, 1887, p. 89).
In 1888 M. Wertheimer wrote,
The earlier theory of the miasmatic nature of [childbed fever], as
taught and developed by Eisenmann, Helm, Litzmann, Kiwisch,
Scanzoni, and others, was first put on the right track by Semmelweis.
… His theory was soon supported, expanded, and confirmed by a
series of authors such as Hegar, Buhl, Winkel, Fischer, Veit, [and]
Mayrhofer.
(Wertheimer, 1888, pp. 5f)
Semmelweis and Mayrhofer continued to be cited in German, French,
and English sources into the 1880s (Bar, 1883, p. 28; Lomer, 1884, pp.
366, 369; Jaggard, 1884, p. 443) although by that time (15 years after
Semmelweis’s death and Mayrhofer’s publications), more recent devel-
opments in bacteriology were eclipsing the older work.
In 1882 – at the very time when, according to conventional wisdom,
Semmelweis had been completely forgotten throughout Europe – Wilhelm
Fischel wrote: ‘although long contested, the opinions of the genial
Semmelweis … have become part of the common property of a whole
generation of medical personnel’ (Fischel, 1882, p. 1). The phrase ‘the
genial Semmelweis’ – poignant in view of the treatment he received
from his peers and of his ultimate fate – appears again in the same year
in an obituary for Carl Mayrhofer (r, 1882).16
There is no reason to think Henle saw any relation between his call for a
rational system of medicine built on universal necessary causes and
Semmelweis’s definition of ‘childbed fever’ in terms of decaying organic
matter. For all we know, Henle, like most of his contemporaries, regarded
Semmelweis as an aberration. Moreover, Semmelweis almost certainly
never thought of his theory in terms of Henle’s aspirations for rational
medicine. Yet the two are inseparably linked: only by way of etiological
characterizations can the range of human illnesses be systematized into
diseases with universal necessary causes, and only in the creation of such
a system can the potential strength of Semmelweis’s innovation be real-
ized. However, Semmelweis showed no interest in generalizing from his
new approach – he was content to exploit the practical ramifications of

his theory, and Henle never mentioned Semmelweis in print.
To be sure, even given the possibility of etiological definitions, there
was a serious obstacle to reorganizing all of medicine in terms of
necessary causes: it was first essential to find causes in terms of which
all (or at least many) diseases could be defined. In this respect
Semmelweis’s theory was something of a dead end. Of course decaying
organic matter did cause childbed fever and, given a suitable definition,
it could be made into a universal and necessary cause. However, decay-
ing organic matter – like acari, Beauvaria bassiana or trichinae – could
be causally linked only to a limited variety of diseases. Where was one
to look for causes in terms of which to define other diseases? Imagine a
philosophical physician (1) committed to Henle’s goal of making medi-
cine an explanatory science, (2) aware of the universal necessary causes
being identified for such unrelated diseases as scabies, muscardine,
trichinosis, and childbed fever, and (3) sensing the power of Semmelweis’s
tactic as a potential tool for creating a rational explanatory system.
Had such a person thought of all this in just the right way (and certainly
no one ever did) that person could still have despaired of achieving a
rational system of medicine on the grounds that there simply were no
causes of the sort required to bring it all together.
Yet, at the very time Semmelweis was writing his book on childbed
fever, Louis Pasteur was becoming fascinated with the enormous variety
of natural organic processes caused by microscopic living organisms.
Some of these processes – for example, different kinds of spoilage in
wine (diseases of wine, as Pasteur called them) – were caused by a
variety of different microorganisms. Moreover, Pasteur traced organic
decomposition to the operation of germs and, as Semmelweis and oth-
ers had recognized, childbed fever and other wound infections involves
tissue decomposition. The possible connection was too striking to be
overlooked: Pasteur’s discoveries motivated both Carl Mayrhofer’s search
for microorganisms in the lochial discharges of childbed fever victims
and Casimir Davaine’s profoundly important investigations of anthrax.
Notes
1. Although, to be completely honest, some writers came close. For exam-

ple, in one paper, David Gruby observed of a particular fungus that it
constituted ‘a true, essential characteristic of [favus]’ (Zakon and Benedek,
1944, p. 158). It is a small step from this recognition to an etiological
definition, and there are similar passages in Agostino Bassi’s discussions
of muscardine. However, as far as I can see, Fracastoro never came close
to anything like this.
2. For references to some of the standard sources see Semmelweis (1861b,

pp. 1–58).
3. For a fine study – informative, scholarly and yet moving – of the history
of childbed fever see Loudon (2000). Below, I disagree with certain as-
pects of Loudon’s discussion of Semmelweis. These points are crucial in
the context of our investigation, but are less significant from the perspec-
tive of Loudon’s purposes and so (in spite of the emphasis they here
receive) they constitute at most a limited criticism of his book.
4. Through this sub section, page numbers in parentheses are references to
Holmes (1843).
5. Through this sub section, page numbers in parentheses are references to
Lumpe (1845).
6. The present account is based on the Translator’s Introduction to
Semmelweis’s Etiology (Semmelweis, 1861b).
7. For example, like many other diseases, among other possible causes,
childbed fever was ascribed both to eating too much and to eating too
little. In cases of the first kind, physicians removed blood; in the second,
they sustained patients with rich and nourishing foods. Letting blood in
cases of the second kind would have been regarded as life threatening.
8. David S. Barnes has recently pointed out that one can also reduce the
incidence of a disease by attacking causes that are not universal (Barnes,
2000, p. 438). Of course. One can prevent some house fires by insuring
that wiring in new houses conforms to building codes even though other
houses will still burn down if struck by lightning. I am not denying that
the chlorine washings were beneficial. Some cases of TB can be avoided if
we prevent people from breathing the miasms of TB wards, but other
cases may still occur if people drink infected milk. By contrast, if we
somehow block invasion of the tuberculosis bacillus, we avoid the disease
altogether – end of story, no matter what one breaths or eats or drinks.
But this latter possibility depends on having tuberculosis (or childbed
fever, or any other disease) defined in terms of a cause that is universal.
9. Through this paragraph, page numbers in parentheses, such as this one,
are references to Semmelweis (1861b).
10. Eight sentences further on I will return to and justify this dubious claim.
11. See Chapter 1, note 2.
12. The reader may wish to review the quotation on page 45.
13. As I wrote 20 years ago: ‘The problem [with criticisms of Semmelweis’s
theory] is that no one bothers to determine what the theory in question is,
before deciding [it is wrong or] that Semmelweis didn’t originate it’
(Semmelweis, 1861b, p. 52n). The situation has not changed appreciably.
14. What follows is an abbreviated version of one part of an earlier publica-
tion (Carter, 1985a). Here I include only a representative sample of
sources that cited Semmelweis and Mayrhofer during the years from
about 1865 through 1890 – there are many more.
15. For example, 46 of the 86 papers cited by Birch-Hirschfeld in his 1875
review of the role of bacteria in infectious diseases concerned infected
wounds (Birch-Hirschfeld, 1875).
16. The author of the obituary was identified only as ‘r’ which suggests, at
least, that this author was someone other than Fischel who used the same
phrase in the same year.
CHAPTER FOUR
Microorganisms as Causes
In this chapter we will review some of Louis Pasteur’s early attempts
to prove that germs cause certain organic processes. Today, it seems
obvious that microorganisms cause various natural phenomena such
as fermentation, decomposition, and disease, but in the early nine-
teenth century this was not clear. There were inherent obstacles to
accepting germs as causes, and overcoming some of these obstacles
required more than simply making the right empirical observations.
Causes are not discovered in the same way that one might discover
shells on a beach.
As described by Imre Lakatos, a ‘research programme consists of
methodological rules: some tell us what paths of research to avoid
(negative heuristic), and others what paths to pursue (positive heuris-
tic)’ (Lakatos, 1968, p. 132). Such rules are often left implicit. However,
if stated, they may resemble very general claims about the world –
claims that exceed any possible empirical confirmation.1 Lakatos says
that such principles are made ‘“irrefutable” by the methodological
decision of its protagonists’ (Lakatos, 1968, p. 133). However, strictly
speaking, being rules and not empirical claims, they are neither true nor
false and they are not subject to proof or refutation. Like any rules, they
can be shown to be useful in achieving certain objectives (for example,
in explaining and controlling nature) and, once seen as useful, they may
become standard conventions (one might say made irrefutable) by the
consensus of the researchers who adopt them.
Three such heuristic rules emerged in early work leading to the
etiological research programme. We will refer to these rules as the
Distinguishability, Dissemination, and Bacterial Hypotheses.2 These hy-
potheses cannot be confirmed empirically yet they were required before
microorganisms could be accepted as causes of the natural processes
with which they are now universally associated. In this chapter we will
encounter the first two of these, the Distinguishability and Dissemina-
tion Hypotheses; the third, the Bacterial Hypothesis, will appear at the
end of the next chapter. In this chapter we will also examine several of
Pasteur’s early arguments to see how he attempted to prove that germs
are causes.
MICROORGANISMS AS CAUSES 63
The distinguishability and dissemination hypotheses
In 1854, Louis Pasteur, a chemist by training, was appointed professor

of chemistry and dean of the Faculty of Sciences in Lille. The faculty at
Lille had been formed, in part, to find scientific solutions to the techni-
cal problems challenging local industry (Geison, 1974, p. 361) – a task
perfectly compatible with Pasteur’s practical disposition. One important
industry in Lille involved the production of alcohol from fermented beet
roots. However, the process often failed because the beet pulp spoiled
rather than fermenting. In seeking to avoid such failures, Pasteur began
studying fermentation.
In the 1830s Charles Cagniard-Latour, Theodor Schwann and Friedrich
Traugott Kützing had independently concluded that yeast was a living
organism and that it alone could ferment sugar to produce alcohol. This
implied that fermentation was essentially a biological process. Other
scientists, including Jons Jacob Berzelius, Friedrich Wöhler, and Justus
von Liebig, regarded fermentation as a chemical process having no
necessary connection with living organisms.
Pasteur began publishing on fermentation in 1857; by the time his
first papers appeared, he seems already to have been convinced that
fermentation was always due to what he called living ferments. Gerald
L. Geison argued that this conviction was likely a result of strongly held
convictions that originated much earlier while Pasteur was still doing
research in chemistry (Geison, 1995, pp. 95–103). By this time, Pasteur
had become preoccupied with the life processes of microorganisms.
From then on, except for occasional reports and special lectures, all his
scientific publications related directly to what he later called the theory
of organized ferments (Pasteur, 1875a, p. 43). As he conceived it, this
theory encompassed a wide range of natural processes such as putrefac-
tion and fermentation as well as many animal and human diseases.
Pasteur observed that he had become fascinated with the immense role
of microscopic organisms in nature (Pasteur, 1880a, p. 147), and all the
activities of such organisms fell under the theory of organized ferments.
In a letter written to Pasteur in 1874, Joseph Lister used the phrase
‘theory of germs’. Pasteur included Lister’s letter in a book published in
1876 (Pasteur, 1876a, pp. 40f). Pasteur himself first used Lister’s phrase
somewhat later in the same book (Pasteur, 1876a, p. 101). Thereafter,
rather than ‘theory of organized ferments’, Pasteur used ‘theory of
germs’ to identify his central research interests.
Although from an early period Pasteur seems to have been interested
in human infectious diseases (Pasteur, 1863a, pp. 8f.), he did not con-
duct original research in this branch of the theory of germs until the
early 1870s. But, as he himself explained, when he began studying
human infections, his work was based on the same ideas and methods
that had guided him in his earlier studies (Pasteur, 1873a, p. 415). In
this chapter we will consider only Pasteur’s early publications on the
theory of germs – those appearing before 1870.
In his first paper on fermentation, Pasteur claimed that, just as a par-

ticular ferment is always present when sugar breaks down to form
alcohol, so too a different ferment is always present when sugar forms
lactic acid (Pasteur, 1857a, p. 6).3 In his early papers, Pasteur frequently
made assertions of this kind, and it is important to understand the
nature of such claims. Asserting that the lactic ferment is present when-
ever sugar is converted into lactic acid is equivalent to saying that, in
the absence of the ferment, no lactic acid forms, and therefore, that the
ferment is necessary for the production of lactic acid. Having inferred
from his own observations that the ferment was necessary for fermenta-
tion, Pasteur seems simply to have concluded that the ferment caused
fermentation. Pasteur also reported seeding appropriate media with
lactic and alcoholic yeasts. He reported that either ferment would mul-
tiply in the liquid in which it was seeded (Pasteur, 1857a, p. 9). This
experiment suggested that, under suitable conditions, either ferment
was sufficient for its fermentation products; it also persuaded Pasteur
that the two ferments were distinct organisms. In this early paper
Pasteur asserted that all fermentation was correlative with organic life
(Pasteur, 1857a, p. 13).
By the end of 1857, as part of an attack on Justus von Liebig’s
chemical theory of fermentation, Pasteur reported growing yeast and
sustaining fermentation in a medium free of organic nitrogen (Pasteur,
1857b). In 1861 Pasteur identified another ferment that he described as
an animal that could live without free oxygen (Pasteur, 1861a). This
ferment produced butyric acid. Again, Pasteur inferred causality from
the observation that the ferment was always present in (and so neces-
sary for) the production of butyric acid. He also reported that the
organism produced butyric acid whenever it was seeded in a suitable
medium.
Pasteur’s early work on fermentation persuaded him that there were
several different ferments. He later admitted that he was not always
able to distinguish ferments morphologically and sometimes identified
them physiologically (Pasteur, 1866a, p. 155). By this he can only have
meant that he distinguished them by their metabolic byproducts, that is,
by the different fermentation processes with which they were associ-
ated. By 1861 Pasteur was convinced that different forms of fermentation
were associated with the life processes of distinct organized ferments
(Pasteur, 1861b, p. 142) and that these ferments could not be trans-
formed into one another (Pasteur, 1862). Given that he could not
always distinguish ferments other than by their fermentation processes,
this conviction could not be confirmed universally without begging the
question. However, it could be illustrated in some cases (since some
organisms could be distinguished morphologically) and thereby made
plausible, and it was an absolutely essential presupposition if one wished,
as Pasteur did, to explain different fermentation processes strictly in
terms of the ferments themselves. If, as some researchers then believed,
all ferments were ultimately the same and transformable into one an-
other (that is if they were not truly distinct), then different fermentation
processes could not be explained simply in terms of the organisms
themselves – such explanations would require an appeal to whatever
other factors accounted for their different effects in different circum-
stances. Thus, insofar as the ferments themselves were to be the basis
for explaining fermentation, it was essential to assume that different
fermentation processes were due to distinct ferments. Of course the
same assumption would be equally essential if one attempted, as Pas-
teur and others later did, to explain different kinds of spoilage or
different diseases in terms of organisms. Here and throughout his ca-
reer, therefore, Pasteur required a crucial assumption that could not be
conclusively demonstrated – an assumption that would constitute part
of what Lakatos would call the negative heuristic of his research pro-
gramme. Pasteur had to assume that different organic processes, whether
fermentation, spoilage, or disease, are caused by distinct organisms. As
we will see later (p. 105) this assumption need not be understood as a
claim about the world although it is natural to think of it in this way. As
Imre Lakatos explained regarding other basic scientific hypotheses, this
assumption may be best regarded as a rule that guides research: always
suppose that different organic processes are due to distinct organisms.
We will refer to this as the Distinguishability Hypothesis4 and I will
have more to say about it later on.
Pasteur’s attempts to reduce failures in industrial fermentation led di-

rectly to the ancient problem of spontaneous generation. If, as Pasteur
believed, fermentation and putrefaction depended on the multiplication
of living organisms, one could prevent spoilage by excluding undesir-
able ferments. But how was this to be accomplished? The answer
depended on knowing how organisms gained access to fermenting liq-
uids. If organisms could arise spontaneously, there was little hope of
preventing spoilage; but if there was no spontaneous generation, spoil-
age could only occur when fermenting liquids were contaminated from
external sources. This meant that preventing contamination would in-

sure consistent success. Of course, the question of spontaneous generation
had broader significance than merely insuring the success of industrial
fermentation. Joseph Balsamo-Crivelli and others argued that, given
suitable nutrients and proper atmospheric conditions, Beauvaria bassiana
could arise spontaneously and do so simultaneously with the onset of
muscardine (Balsamo-Crivelli, 1839, p. 123). Even Henle felt that the
possibility of such a miasmatic origin could not be ignored (Henle,
1840, p. 945). However, if this were to happen, even if the fungus and
muscardine always occurred together (so each was necessary and suffi-
cient for the other), the pre-existing materials together with the crucial
atmospheric conditions, not the fungus itself, would be the cause. This
possibility must be excluded if organisms are to be identified as causes
of organic processes and thereby explain those processes. Thus, as
Pasteur recognized, the question of whether living organisms could
arise independently of antecedent organisms had profound implications
for the germ theory.
In his early studies of fermentation, Pasteur occasionally used the
term ‘spontaneous generation’ in reference to the unintended growth of
organisms in a culture medium (Pasteur, 1857a, p. 9). However, he
explained that he had too little evidence to determine whether such
developments were truly spontaneous or were due to accidental con-
tamination (Pasteur, 1859a, p. 628). Then, in about 1860, Pasteur
began a careful study of the subject. Geison has shown that Pasteur’s
research on spontaneous generation, like his work on fermentation, was
driven by deeply held convictions – preconceived ideas – that can be
traced to his early work in chemistry (Geison, 1995, pp. 110–42), so, as
he began this phase of his research, he was not really neutral with
respect to the issue. By the time Pasteur began studying spontaneous
generation, earlier scientists had already devoted considerable effort to
the problem without conclusive results, and it was not clear that Pas-
teur would have greater success. However, within a few months, he
concluded that pure air and a suitable medium were never sufficient to
generate living organisms, and that the organized corpuscles regularly
found in airborne dust were ‘the exclusive origin – the first and neces-
sary condition – for life in infusions, in putrescible bodies, and in all
bodies capable of fermentation’ (Pasteur, 1860, p. 204). That is, the
presence of living organisms is always due to their having spread from
some source of contamination and is never spontaneous. Pasteur re-
ferred to this idea as ‘the hypothesis of the dissemination of germs’
(Pasteur, 1860, p. 202). We can shorten Pasteur’s title to ‘the Dissemi-
nation Hypothesis’. Thus, as we use this term, the Dissemination
Hypothesis is the denial of spontaneous generation.
In 1861 Pasteur published an extensive study of spontaneous genera-

tion (Pasteur, 1861c). He criticized earlier attempts to exhibit the process;
all such attempts, he argued, had failed to exclude contamination. He
also reported his own research on the topic. Pasteur’s experiments fell
into two classes. In the first, he showed that ordinary airborne dust
contained particles indistinguishable from the germs of living organ-
isms. When fermentable liquids were seeded with such dust, colonies of
living organisms soon flourished there. However, the same liquids re-
mained sterile if exposed to dust-free air or even to dust that had been
heated sufficiently to destroy living organisms (Pasteur, 1861c, pp.
238–59). Pasteur concluded that ‘all organized productions of the pre-
viously heated infusions have no other origin than the solid particles
which the air always carries and allows constantly to be deposited on
all objects’ (Pasteur 1861c, pp. 259f.) – the Dissemination Hypothesis.
Pasteur’s second series of experiments involved the use of flasks with
long contorted stems. Pasteur filled such flasks with milk, blood, and
urine; he sealed the stems and sterilized the flasks and their contents. He
then broke the seals and exposed the liquids to ambient air. However,
air could reach the liquids only by passing through the contorted stems
where it deposited all the dust it normally bore. The liquids remained
sterile (Pasteur, 1861c, pp. 259–64). Indeed even today, more than a
century later, some of Pasteur’s flasks and their unaltered contents can
be inspected in museums throughout France. As the months and years
passed, without any change in the liquids, it became progressively less
plausible to think that living organisms of any kind would ever arise
spontaneously.
Partly because of Pasteur’s great rhetorical skills, his experiments
were psychologically compelling (Geison, 1995, pp. 132f.), but were
they conclusive? Did he prove absolutely that life never arises spontane-
ously – that contamination is ‘the exclusive origin’ of living germs? He
freely admitted that he did not:
I must hasten to repeat here what I have often said: one cannot
prove a priori that there exists no spontaneous generation. All one
can do is demonstrate (1) that there have been unperceived sources
of error in the experiments [of those who claim to have observed
spontaneous generation], and (2) that by eliminating these sources
of error, without affecting the basic conditions of the trials, the
inferior beings cease to appear.
(Pasteur, 1866b, pp. 354–5)
On another occasion, Pasteur wrote that ‘in the empirical sciences, in
contrast to mathematics, it is impossible to demonstrate a negative’
(Pasteur, 1874a, p. 16). In fact, Pasteur’s experiments were less compel-
ling than he himself made them appear. One problem was that Pasteur’s
own experiments often failed, in some series of experiments, perhaps

more than 90 per cent of the time, ‘but rather than draw the seemingly
obvious conclusion that this microbial life had originated spontane-
ously, Pasteur refused to accept this experimental evidence at face value
and pressed relentlessly toward an alternative explanation’ (Geison,
1995, p. 130). On the other hand, of course, advocates of spontaneous
generation could never prove, absolutely, that it had occurred, that is,
that there had been no contamination. Thus, neither the Dissemination
Hypothesis nor its denial were susceptible of strict empirical proof or
confirmation: for either side, and for the public at large, it was a
question of which experiments were to be taken as successful and which
as having failed.5 Pasteur’s arguments ‘ingenious and wonderfully theat-
rical’ (Geison, 1995, p. 118), carried the day; his opponents retired in
confusion.
While Pasteur’s theory of germs requires the absolute exclusion of
spontaneous generation, this could never be accomplished merely by
observation or experiments. At this point, one might despair of ever
giving a rational justification for the Dissemination Hypothesis. And, in
fact, Geison himself seemed to teeter on the very brink of this abyss: he
invited his reader to share the ‘liberating effect’ of ceasing to pretend
‘that we believe in the Scientific Method … [for] it is not Pasteur who
has fallen short; it is this Scientific Method’ (Geison, 1995, p. 132). On
a causal reading, these remarks suggest the pathos of a lost soul luring
others to the illusory emancipation of the damned. However, by ‘this
Scientific Method’ (my emphasis), Geison may have been referring to
(and so discrediting) only ‘the simplistic and passé notion’ that the
impossibility of direct empirical confirmation means there is no rational
justification whatsoever. That would still leave open the possibility of
some other sort of justification (that is, of a more sophisticated and up-
to-date understanding of how science works). Unfortunately, instead of
going on to explain such an alternative justification, Geison cited Bruno
Latour (the great bugbear who does seem to argue against scientific
justification in general6) and then passed directly to an account of
Pasteur’s rhetorical and theatrical skills (which, obviously, could pro-
vide no sort of justification whatsoever). This leaves the reader perplexed:
can we or can we not justify the exclusion of spontaneous generation?
To be sure there was (and could be) no conclusive empirical justifica-
tion for the Dissemination Hypothesis. And Geison made a persuasive
argument that Pasteur’s own rejection of spontaneous generation flowed,
not from evidence, but from what Pasteur himself admitted were ‘pre-
conceived ideas’. However, this is no surprise. As explained above (p. 6)
isolated laws, being universal in scope, can never be justified empirically
(as we have seen, Pasteur even admitted this); at best they can be
illustrated, made plausible, or shown to apply in certain cases. Laws are

justified by being part of a theory which, in turn, is justified by its
explanatory power. Who would be so foolish as to try to justify, in
isolation from the kinetic theory as a whole, the law (which is one part
of the kinetic theory) that molecules in a gas ‘exert no forces on one
another between collisions’ (Oxtoby and Nachtrieb, 1990, p. 104) –
stated like this, naked and alone, it doesn’t even seem plausible, much
less verifiable, no matter what or how many experiments one performs,
but so what? At this point in our story (as in the historical context as
presented by Geison) the Dissemination Hypothesis – the denial of
spontaneous generation – stands as an isolated claim, and as such, it
can not be fully justified by any means whatsoever. However, as we will
see in Chapter 6, this hypothesis was soon assimilated to an explicit
bacterial theory, and in that context it, along with the other laws
comprised by the theory, was justified by the explanatory power of that
theory as a whole. By the 1880s, everyone saw that the bacterial theory
could better explain a whole range of facts than any other theory that
was then (or is now) available (seeing this did not rest primarily on
rhetorical or theatrical tricks), and the Dissemination Hypothesis was
accepted as an indispensable part of that theory. Of course, as Geison
brilliantly demonstrated, even in the 1860s Pasteur regarded the Dis-
semination Hypothesis as part of his own system of ideas (which he
called the germ theory), but, at that point, it remained to be seen how
successful Pasteur’s system would ultimately be. Thus, his system could
provide little or no help in justifying the hypothesis.
I agree that in the 1860s, when Pasteur advanced the Dissemination
Hypothesis, it could not be justified – not by any means whatsoever (so
its acceptance at that point can only be explained in terms of force),
and, from a logical point of view, it remained an open question for
some time whether the hypothesis would ever be justified. From one
point of view, justification did not really begin until about 1876 when
Klebs made the hypothesis part of an explicit bacterial theory of disease
and it was completed only in the early 1880s when the bacterial theory,
as a whole, finally won out. At that point, as Lakatos would put it, the
Dissemination Hypothesis became ‘irrefutable’ by the consensus of the
researchers who used it. But understanding things in this way does not
expose a defect in the Scientific Method (whatever that may be) –
instead it illustrates the poverty of arguments based on simplistic and
passé conceptions of science.
Pasteur’s early attempts to prove microorganisms are causes
In 1863 Pasteur began publishing papers on the spoilage of wine. In his

earliest papers he assumed the Distinguishability Hypothesis, that is, he
assumed different forms of spoilage, like different forms of fermenta-
tion, were caused by distinct organisms (Pasteur, 1864, p. 396).
As in his work on fermentation, Pasteur seemed not to have felt that
establishing causality presented any particular difficulties. He did not
identify causal criteria and occasionally simply asserted that a particu-
lar form of spoilage was caused by a specific organism without giving
any justification whatsoever (Pasteur, 1864, p. 402). When he did give
reasons, the reasons were usually of the form exemplified by this
statement: ‘Mycoderma aceti causes the acidity that attacks the vats of
red and white wine in the Jura. I have identified it on the surface of all
the wines, a considerable number, that have been pointed out to me as
acidic’ (Pasteur, 1864, pp. 396f.). As in his work on fermentation,
Pasteur seems to have assumed that finding a particular organism in
all cases of acid wines proved that it was the cause. In discussing
another form of spoilage, Pasteur wrote that a certain ferment ‘deter-
mines the disorder known as bitterness of wine … I have studied bitter
wines from all the provinces, and have constantly recognized this
curious vegetable in a quantity that varies according to the bitterness
of the wine in question’ (Pasteur, 1864, p. 401). He made similar
observations for other forms of spoilage (Pasteur, 1864, p. 400). Here
again we see the same kind of reasoning: Pasteur argued that, since a
given organism was always present in wine that had spoiled in a
certain way (that is, since the organism was necessary for that kind of
spoilage), it was the cause of the disorder. The only evidence for
causality one finds in these early essays is observational evidence of
necessity.
In 1865, Pasteur concluded that ‘the diseases of wine, at least all
those presently known to me, are determined by the development of a
microscopic vegetable of the nature of a ferment’ (Pasteur 1865a, p.
409). One year later, in 1866, in his main work on wine spoilage,
Pasteur claimed: ‘there is never a souring of an alcoholic liquid except
in the presence of the microscopic mushroom known as Mycoderma
aceti’ (Pasteur, 1866a, p. 125). He asserted that ‘without exception’
certain other wine disorders were due to the presence of other organ-
ized ferments (Pasteur 1866a, pp. 138, 161f.). In all this work, the
sketchy evidence for causation Pasteur actually gave was of the form ‘in
the absence of the organism, there is no spoilage’.
While Pasteur continued to write on the spoilage of wine and to seek

methods for preventing it, colleagues urged him to study a new prob-
lem: In the 1860s vast numbers of silkworms were dying throughout
Europe from unknown causes, and the French silk industry was suffer-
ing serious losses (Geison, 1974, p. 372). Pasteur agreed to investigate.
By 1865, when he began studying silkworms, a certain corpuscular
structure had already been found in the intestines of many dead worms,
and some researchers believed the structures were parasitic organisms
that infected and killed their hosts. Pasteur began by confirming the
existence of the structures (Pasteur, 1865b). He soon found that if
healthy worms were exposed to dust containing the corpuscles they also
became diseased (Pasteur, 1866c, pp. 442f.). Yet the situation was
different from what he had found in spoiled wines because, while
worms with corpuscles were always diseased, diseased worms did not
always contain corpuscles – we would say that corpuscles appeared to
be sufficient but not necessary for the disease. Thus, the situation was
not the same as in his earlier research, and Pasteur was clearly reluctant
to infer causality (Pasteur, 1866c, p. 447). Moreover, he found no
evidence that the corpuscles reproduced (Pasteur, 1866d, p. 472) and
was unsure they were alive. Instead, he suggested, they may be similar
to corpuscles of blood or pus (Pasteur, 1866d, p. 472; Pasteur, 1866c
pp. 442, 447). Pasteur suspected that, rather than causing the disease,
the corpuscles may form in existing tissues as a result of the disease
process (Pasteur, 1867a, p. 466). He also thought the disorder may be
constitutional, and he compared it to tuberculosis (Pasteur, 1866c, p.
445). Given contemporary beliefs, this may have meant that Pasteur
suspected the silkworm disease resulted from unhealthy environmental
factors rather than from a specific cause.
In 1867 Pasteur reported several important new conclusions. First, he
discovered that silkworms were dying from two distinct diseases. One
disease, called pebrine, was associated with the corpuscles; the other,
flacherie, involved no corpuscles, but was characterized by the appear-
ance of ferments in the digestive tracks of the worms.7 Of course, this
way of thinking enabled Pasteur to hold that corpuscles were present in
(and therefore necessary for) every case of pebrine. In other words, this
change brought the situation into line with the kinds of experimental
results he had found both in fermentation and in the spoilage of wine.
This increased his confidence that the relation was truly causal. Second,
he claimed to have discovered that the corpuscles reproduced by a kind
of scission – a discovery that removed one obstacle to recognizing them
as alive (Pasteur, 1867c, p. 499). Finally, he concluded that both dis-
eases could be prevented, or at least controlled, by suitable breeding
techniques. In several papers that appeared in the next few months,
Pasteur claimed to have mastered pebrine; he justified this claim on the

grounds that he could ‘give and prevent it at will’ (Pasteur, 1868a, p.
528; 1869a, p. 19).
Pasteur soon identified two different species of ferments associated
with flacherie. He described one species as a kind of vibrion and the
other as having the structure of a string of grains (Pasteur, 1869b, pp.
591f.) – a form of organism already becoming familiar to microbiologists
in some human diseases. However, he insisted that no one had, as yet,
proved that these organisms caused flacherie, and he acknowledged the
organisms could actually be a consequence of prior digestive disorders
(Pasteur, 1868b, p. 568). Pasteur’s reluctance to draw a causal conclu-
sion – so different from the attitude displayed in earlier papers on
fermentation and spoilage – may reflect the fact that, in contrast to
what he had found in his earlier work and even ultimately in pebrine,
neither species of ferment was present in all cases of flacherie (that is
neither species was necessary).
By 1869, when his book on silkworm diseases appeared, Pasteur was
convinced the corpuscles caused pebrine; however, he remained unsure
about the significance of the ferments in flacherie. Pasteur identified
four organisms that could often be found in the intestines of worms
dying of flacherie and that did not appear in healthy worms (Pasteur,
1870, p. 205). Against the objection that the organisms may be a
product rather than the cause of the disease, Pasteur wrote: ‘if I can
take very healthy worms and communicate flacherie to them by intro-
ducing vibrions, … is it possible not to relate the cause and first origin
of the evil to the vibrions and to the fermentation they determine?’
(Pasteur, 1870, p. 216) This argument, like the observation that organ-
isms are never present in healthy worms, supports sufficiency rather
than necessity. Thus, the evidence for causation in these studies is
different from that in his earlier discussions and in his work on pebrine.
This change was forced on him because, in flacherie, he was unable to
identify a particular microorganism that was present in all cases.
Beginning with his studies of fermentation and spontaneous generation

and continuing through his work on wine disorders and diseases of
silkworms, Pasteur’s arguments for causality took this form: he found
some organism regularly present in an organic process and concluded it
was the cause of that process. When he was unable to reason in this
way, he fell back on experiments suggesting sufficiency, but he was
clearly reluctant to infer causality from that relation.
From a logical point of view, Bassi, Gruby, Semmelweis, and Pasteur
all reasoned in much the same way. Bassi concluded that a fungus was
the cause of muscardine after becoming persuaded that the disease

never occurred in the absence of the fungus. Gruby saw that the morbid
remains he studied were actually composed of fungi; in their absence,
therefore, there would be no morbidity. Once Semmelweis eliminated
decaying organic matter by chlorine washings, childbed fever occurred
only in cases he identified as self-contamination; thus, in the absence of
the purported cause, there would be no disease. Pasteur inferred causal-
ity by observing that the same ferment was present in each case of each
organic process and, therefore, in the absence of that ferment, the
process would not occur. Bassi, Semmelweis, and Pasteur sometimes
exposed test animals to the operation of purported causes, but they
seem to have conducted these tests without much conviction and gave
little emphasis to their results. The evidence that each seemed to find
most compelling was evidence of causal necessity. So far as one can tell,
each reasoned in this way without much awareness that doing so de-
parted from the way physicians usually thought about causes. Yet the
closer their work approached human diseases, the more vigorously they
were opposed. Gruby was ostracized and his work dismissed as insig-
nificant (British Medical Journal, 1898). Bassi and Semmelweis were
immediately attacked by those who continued to think in terms of
spontaneous disorders and of pluralities of possible causes. Pasteur’s
early work, which concerned problems relatively remote from human
disease, seems not to have attracted much critical attention from con-
temporary physicians. However, after 1870, as he began studying human
diseases using the same methods he had used earlier, physicians and
pathologists criticized him in the same ways and with the same animos-
ity experienced by Bassi, Gruby, and Semmelweis.
We now move into a new phase in the development of the etiological

programme. Bassi, Gruby, Semmelweis, and Pasteur conducted their
research and claimed to have identified causes, but they did so in the
absence of clear and acknowledged standards for proving causality.
Each seems to have assumed without justification (apparently without
even sensing a need for justification) that one could demonstrate causal-
ity simply by identifying a factor always present in some organic process.
The research we will now consider differs in two ways from what we
have seen so far: First, we will find a growing awareness of, and interest
in, the concept of disease causation. This interest is revealed in disputes
about whether given sets of experiments are or are not adequate to
prove causality. These disputes differ from those we have so far encoun-
tered. While critics of Bassi and Semmelweis challenged their claims
that Beauvaria bassiana or decaying organic matter were the only causes
of muscardine or of childbed fever, no one objected to the idea that

each was a cause. Early in the century, causal concepts were sufficiently
vague that virtually anything could be accepted as a cause of almost any
pathological process. However, by the late 1860s there was a sense that
causal claims required evidence, and by the 1870s disputes about how
to prove causality were partly resolved with the elaboration and almost
universal acceptance of explicit causal criteria. Thereafter, causal claims
could be accepted only if they were supported by experiments intention-
ally organized to satisfy those criteria.
Second, whereas most of the research we have considered so far was
conducted by scientists working alone, we now find research that builds
on the work of predecessors and contemporaries. By the 1880s, in place
of relatively independent research of relatively isolated scientists each
following his own intuitions about proving causation, we find shared
techniques, assumptions, and nomenclature, an interest in confirming
the observations and in repeating the experiments of others, and re-
search subordinated to the shared goal of finding specific causes of
diseases. In short, by the 1880s, we find a full-blown research pro-
gramme focusing on the identification of universal and necessary causes.
Casimir Davaine was a transition figure in this change. Rather than
working in relative isolation, his work connected solidly into the re-
search of others in the two ways just mentioned: he helped secure a
consensus of expectations about how one could prove causation, and he
frequently drew on and motivated the research of others. However
Davaine’s work never fully persuaded his contemporaries – a failure
due, in part, to the very lack of the consensus he was helping to achieve.
In particular, his work was an important step in securing acceptance of
a vital heuristic principle, the Bacterial Hypothesis; but his contempo-
raries remained unconvinced by his work, in part, precisely because
they did not as yet accept that principle.
Notes
1. For example, the first two Laws of Thermodynamics can be interpreted as

stating that the total energy in the universe is constant and that entropy is
always increasing. If understood as empirical claims about the world,
neither could possibly be confirmed.
2. All three hypotheses are laws in the sense defined by Wim J. Van der Steen
and Harmke Kamminga (Van der Steen and Kamminga, 1991) and they
function within theories of disease in the same way that, say, the Laws of
Motion function in Newtonian mechanics or the Laws of Thermodynamics
function in thermodynamics. However, in common usage they are not
referred to as ‘laws’, and ‘hypothesis’ would seem to be a suitable term.
3. Geison observed that this paper is ‘astonishing in its audacity and scope’
and that it introduced nearly all ‘the basic convictions and techniques that
would thereafter guide [Pasteur’s] study of fermentation in general’ (Geison,
1995, p. 90).
4. Geison wrote that, among the basic convictions introduced in Pasteur’s
first (1857a) paper on lactic fermentation was ‘his notion of specificity,
according to which each fermentation could be traced to a specific mi-
crobe’ (Geison, 1995, p. 90).
5. This situation was made even more complex by the fact, recently un-
earthed by Geison, that Pasteur actually believed in a kind of spontaneous
generation and conducted numerous (unsuccessful) experiments to verify
it. However, Pasteur was not quite inconsistent; his view seems to have
been that spontaneous generation could not come about using the ordinary
techniques available in chemical or biological laboratories (Geison, 1995,
pp. 136–42). This aspect of Pasteur’s work does not concern us in the
present context.
6. I will have more to say about Latour under ‘Some Final Thoughts’ at the
end of this book.
7. Pasteur first reported this discovery in remarks before a special meeting of
the Comice agricole d’Alais in 1867 (Pasteur, 1867b). Pebrine is now
attributed to a protozoan, Nosema bombycis, and flacherie to a virus that
predisposes the silkworm to infection by the bacterium, Bacillus bombycis,
that Pasteur had observed (Lechevalier and Solotorovsky, 1975, p. 40).
CHAPTER FIVE
The Bacterial Hypothesis

During the 1860s, while Pasteur studied fermentation, spontaneous
generation, and silkworm diseases, other French scientists were becom-
ing interested in an infectious disease known as anthrax.
Anthrax usually attacks livestock, and especially sheep, in wet or
marshy areas. Infected animals occasionally drop dead without any
signs of disease, but, more commonly, over the space of a few hours or
days, afflicted animals become progressively weaker until they are un-
able to walk. In the final stages of the disease, dying animals simply lie
on the ground, panting and hemorrhaging through all body openings;
death comes by asphyxiation. Anthrax spreads rapidly among livestock
– within a few days of the first fatalities, entire herds can be lost; it can
also attack humans who handle meat or skins from infected animals.
Among humans, anthrax usually occurs either as cutaneous boils (known
as malignant pustules) or as a gastro-intestinal disorder. Anthrax is
sometimes fatal to humans, but livestock are at much greater risk: in
early nineteenth-century Europe, a few hundred people may have died
of anthrax in a given year, but deaths among livestock often numbered
in the hundreds of thousands. Obviously the disease had profound
economic significance.
By the end of the eighteenth century it was generally known that
anthrax was contagious (Bollinger, 1875, p. 449), and by 1824 experi-
ments had shown it could be conveyed by inoculating blood from
diseased animals (Théodoridès, 1966, p. 157). However, it was still
believed that animals could contract the disease if they were overfed,
allowed to graze where the soil composition was inappropriate, or
exposed to malarial poisons (Bollinger, 1875, p. 449). In addition to
catching anthrax from animals, people were deemed vulnerable if they
consumed indigestible or poor quality food, or were exposed to acid
substances or to skin irritants. One writer reported having seen it
‘several times in persons who had been attacked by mange or scaling
and who had employed energetic repercussions to make the eruptions
disappear’ (Marjolin, 1833, p. 194). Another wrote that anthrax could
occur in hot, humid weather, among the poor who had inadequate diet,
after drinking corrupt water, by contagion, or spontaneously (Beaude,
1849, vol. 1, p. 129).
In the summer of 1850, Pierre-François-Olive Rayer conducted an-
thrax inoculation experiments on sheep in the region of Chartres (Rayer,
THE BACTERIAL HYPOTHESIS 77
1850). In a footnote in his published report, Rayer noted that an

associate, Casimir Davaine, had seen ‘small filiform bodies about twice
the length of a blood corpuscle’ in the blood of anthracoid sheep. There
is no indication Rayer or Davaine regarded the structures as having any
particular significance. Five years later, in 1855, a German physician,
Franz Aloys Antoin Pollender, reported finding immobile rod-shaped
bodies in the blood of anthracoid animals (Pollender, 1855). Pollender
noted that the bodies resembled bacteria and, in contrast to Davaine,
speculated that they could be causally associated with anthrax. Pollender
claimed to have seen the bodies as early as 1849, one year before
Davaine, but his account was first published five years after Rayer’s
essay. Contemporary German writers generally credited Pollender rather
than Davaine with discovery of the anthrax bacillus (Bollinger, 1875, p.
450; Birch-Hirschfeld, 1875, p. 205). Because anthrax was studied
simultaneously by German and French scientists in the years surround-
ing the Franco-Prussian war, their efforts were animated by keen
nationalistic rivalry.
In 1857 and 1858 Friedrich August Brauell published the results of
inoculation experiments with the blood of anthracoid animals. Brauell
confirmed Pollender’s observations of rod-shaped microscopic bodies in
anthrax blood. He reported the bodies did not result from the decom-
position of blood after death because he had seen them in the blood of
living animals. According to Brauell, before the death of an anthracoid
animal, and even during the first two or three days after death, the rod-
shaped bodies remained immobile. However, he reported that, about
three days after an animal died, the rods began to move and gradually
became active vibrions. He also reported finding similar vibrions in
animals that had not died of anthrax (Brauell, 1857, pp. 142f.).1
In his 1858 paper Brauell claimed that the immobile rod-shaped
bodies characterized, exclusively, anthracoid animals. He reported that
the rods never appeared in the blood of healthy animals or of animals
suffering from other diseases (Brauell, 1858, p. 462). Brauell continued
to hold that, within three or four days after death, the previously
immobile rods began to move and ultimately became active vibrions.
He also found that, even if the blood of a pregnant anthracoid sheep
contained the characteristic rods, the blood of its fetus may not, and
inoculations with such fetal blood did not convey anthrax to healthy
animals (Brauell, 1858, p. 466). Brauell was able to identify rod-shaped
bodies in the blood of diseased animals only within the last few hours
before they died, but he found that blood from anthracoid animals in
which these bodies had not yet appeared could still convey anthrax.
From these observations he inferred that, while the rods indicated the
imminent death of diseased animals, they were neither the contagium of
anthrax nor even the bearers of the contagium (Brauell, 1858, p. 463).
Given that Brauell relied on visual inspection to determine whether
blood samples contained rod-shaped bodies, he may simply have over-
looked isolated organisms. Since he did not distinguish anthrax bacilli
from ordinary decomposition organisms, it is also possible that some of
his test animals died from other diseases. In any case while Brauell
believed the rod-shaped bodies conclusively indicated that a diseased
animal was about to die, he was not persuaded that they caused
anthrax.
In 1863 Davaine resumed his study of anthrax. He later explained that,

in 1850, when he first saw the rod-shaped bodies, he had intended to
confirm his observations and to determine whether the development of
the bodies could cause animals to deteriorate and die. However, he had
not found the time for these studies (Davaine, 1863a, pp. 220f). He
wrote that ‘the opportunity had still not presented itself, and other
concerns had not permitted me to seek it actively when, in February
1861, Pasteur published his remarkable work on the butyric ferment, a
ferment consisting of small cylindrical rods possessing all the character-
istics of vibrions or bacteria’ (Davaine, 1863a, p. 221).
Davaine wrote that the similarity between the butyric ferment and
the bodies he had seen in 1850 revived his interest in anthrax. He
conjectured that the rod-shaped bodies could play the role of a ferment
in the blood. This would explain, he wrote, ‘the rapid infection of the
blood mass of an animal that had received into its veins, accidentally or
experimentally, a certain number of these bacteria, that is to say, of this
ferment’ (Davaine, 1863a, p. 221). Although his interest in anthrax was
revived, two more years passed before Davaine obtained a sample of
anthrax blood. He then resumed inoculations.
In his first report on anthrax, Davaine described the rods in detail
and emphasized that they were incapable of independent movement. He
also observed that they disappeared completely once blood began to
decompose. This, he wrote, provides an easy way of distinguishing
anthrax infusoria from those involved in ordinary putrefaction (Davaine,
1863a, p. 222). Davaine did not specifically mention Brauell, but the
implication was obvious: Brauell had mistakenly believed he saw immo-
bile anthrax rods changing into moving organisms, whereas in fact he
had seen anthrax rods replaced by putrefaction vibrions. Davaine was
unable to decide whether anthrax bacteria functioned as minute
animalcules that poisoned their host or as ferments that decomposed
the host’s blood, but he hoped future research would shed light on such
questions. ‘For the moment’, he concluded, ‘I limit myself to announc-
ing one fact I believe to be new. The examination of six animals suffer-
ing or dead of anthrax has six times revealed in their blood the same
microscopic beings. These corpuscles evidently develop during the life
of the infected animal, and their relation to the disease that leads to
death cannot be in doubt’ (Davaine, 1863a, p. 223).
Whether he realized it or not, Davaine’s reasoning followed a familiar
pattern. It was like Pasteur’s arguments that various organisms cause
fermentation and wine spoilage, it was like Gruby’s arguments that vari-
ous fungi cause human mycoses, and it was like Bassi’s argument that
Beauvaria bassiana causes muscardine: having observed a particular or-
ganism regularly present in some process (fermentation, putrefaction,
disease), each inferred that the organism caused the process. Each author
inferred causality from observational evidence of necessity. Partly in re-
sponse to his critics (whom he identified as pathologists), Davaine soon
provided other evidence of causality. However, in the preceding argu-
ment, which was his first attempt to prove a causal connection, and in
others that came later, the evidence that Davaine seems to have found
most compelling was evidence of necessity. Like Bassi, Gruby, Semmelweis,
and Pasteur, Davaine began by thinking in terms of necessary causes.
Given Davaine’s interest in Pasteur’s work, it is possible that this prefer-
ence reflects Pasteur’s influence.
In his second paper Davaine confirmed his earlier findings and re-
ported new experiments in which he inoculated rabbits with fresh blood
containing anthrax bacteria. He reported the organisms were present in
the rabbits’ blood before the onset of morbid changes and that very
small quantities of infected blood could kill healthy rabbits. Davaine
also found that anthrax blood, after being rapidly dried, could be
heated to 100 degrees without losing its contagious powers. On the
other hand, putrefied blood from which the organisms had disappeared
was incapable of conveying anthrax. Davaine concluded that there was
no reason to seek some further ‘contagious element, mysterious and
unknown’, that presumably developed and destroyed when and only
when the bacterium was also present. The bacterium ‘is visible and
palpable; an organized being, endowed with life, that develops and
propagates as a living being. By its presence and rapid multiplication in
the blood, it brings about, in the constitution of this liquid, without
doubt after the manner of a ferment, those modifications from which
the infected animal quickly dies’ (Davaine, 1863b, p. 387).
In 1863, apparently thinking he was confirming Davaine’s work, a

French physician named Signol reported his own investigations of the
blood of diseased livestock. Signol reviewed French and German litera-
ture on anthrax and mentioned particularly the work of Pollender and

Brauell. He confirmed Davaine’s claim that blood from anthracoid
sheep contained bacteria, and he reported that, if infected blood was
inoculated into healthy animals, the blood of those animals soon con-
tained large numbers of bacteria. However, Signol also reported that
the same bacteria could be found in the blood of horses with non-
anthracoid diseases. Thus, he concluded, the bacteria were not unique
to anthrax (Signol, 1863).
By 1864 when he presented his next paper, Davaine had conducted
inoculation experiments on more than 150 animals; his experiments
consistently confirmed his earlier results. Davaine was convinced the
anthrax rods differed from the organisms Signol and others found in
other diseases and, to help mark the distinction, he proposed calling
the anthrax rods bacteridia (Davaine, 1864a, p. 393). Without men-
tioning Brauell’s similar experiments, Davaine reported that the blood
of a pregnant anthracoid sheep did not infect its fetus, and that
inoculating healthy animals with blood from the fetus did not convey
anthrax. He found that anthrax blood that had been rapidly air-dried
and then conserved for as long as 11 months was still infectious. He
also confirmed that anthrax could be induced by the consumption of
contaminated food. These arguments were intended to provide experi-
mental evidence that, if animals were suitably exposed to bacteridia,
they could contract anthrax – that is, under certain conditions,
bacteridia were sufficient for anthrax. Davaine argued that the conta-
gious element of anthrax is completely different from the toxic element
of putrefaction: ‘the toxic agent of putrid matter does not regenerate
itself as does that of anthrax blood. In a word, putrefaction acts on
the animal economy as a poison whereas anthrax acts as a virus’
(Davaine, 1864a, p. 396).
In another paper given later in the same year, Davaine reported that
bacteridia were also prominent in cutaneous boils, known in human
victims as malignant pustules, and that, if tissue fragments from these
boils were placed under the skin of test animals, the animals soon died
of anthrax (Davaine and Raimbert, 1864). Malignant pustules had long
been associated with anthrax, and they were widely believed to follow
from exposure to the blood and hides of anthracoid sheep. By showing
that bacteridia were present in both diseases, Davaine supported the
claim that bacteridia caused both.
Davaine also proposed using the presence of bacteridia as a diagnos-
tic criterion for use in classifying questionable cases of anthrax (Davaine
and Raimbert, 1864, p. 431). Although Davaine did not make this
explicit, this proposal amounted to a redefinition of ‘anthrax’ in terms
of its causal agent because, given this criterion, the presence of bacteridia,
rather than any symptomatic or pathological considerations, became

the ultimate criterion for deciding what was and what was not anthrax.
Four years later, convinced that anthrax and malignant pustule both
depended on the presence of bacteridia, Davaine just assumed that
malignant pustule was simply a form of anthrax (Davaine, 1868a, p.
145). Davaine was now assuming that anthrax was to be defined in
terms of the presence of its necessary cause.
Davaine’s research was soon challenged by two French physicians, Emile-

Claude Leplat and Pierre-François Jaillard (Leplat and Jaillard, 1864).
Leplat and Jaillard cited Signol’s paper as well as research conducted in
Italy suggesting that bacteria were sometimes found in non-anthracoid
disorders (Tigri, 1863). They criticized Davaine for using whole blood
in all his experiments. Blood was a complex liquid containing many
ingredients that could not be identified by microscopic examination.
Thus, they reasoned, if one conveyed anthrax by injecting infected
blood, it was impossible to decide whether the contagious element was
the bacteridia or some other ingredient.
Like Davaine, Leplat and Jaillard saw no hope of isolating bacteridia
from the blood. Instead they proposed inoculating vibrions contained in
other media. Their idea was this: assuming the vibrions were the same
regardless of the medium in which they flourished, if inoculating vibrions
from different cultures produced different results, the differences had to
be ascribed to the media rather than to the vibrions. ‘Nothing is easier’,
they wrote, ‘than to provide these small microscopic beings, that have
the greatest resemblance to one another and, without doubt, the same
properties’ (Leplat and Jaillard, 1864, p. 251). In place of anthrax
blood, Leplat and Jaillard injected ‘certain vegetable infusions, some
liquids charged with decomposing animal matter, putrefied urine, [and]
the serum of decomposed blood’. In each case, before carrying out the
injections, the authors determined ‘the identity and the vitality’ of the
vibrions by microscopic examinations (Leplat and Jaillard, 1864, p.
251). Most of the injected animals remained healthy; deaths occurred
only when the medium containing the vibrions was putrid and was
injected in quantities so large that the animals died of septic poisoning.
Leplat and Jaillard concluded that the vibrions themselves did not cause
anthrax and were harmless. This was clearly intended as an argument
that bacteridia (assumed to be the same regardless of their origin) were
not sufficient for anthrax whereas, at least as he himself stated it,
Davaine’s position was that the organisms were necessary for the dis-
ease. Thus, to some extent, Davaine and his critics were speaking past
one another. Of course, there were other enormous differences as well.
To many of Davaine’s contemporaries, these and other criticisms

seemed persuasive. In 1865, when Davaine was awarded a prize for his
research, the citation stated that ‘either the bacteridium is the transmit-
ting agent of anthrax or else it is the corpuscle which invariably
accompanies the necessary condition for inoculation and development
of the disease’ (Théodoridès, 1966, p. 161). Even the author of this
citation, while honoring Davaine for his work on anthrax, was unwill-
ing to accept without reservation Davaine’s view that the organism was
the only active agent in anthrax blood.
Davaine responded quickly to Leplat and Jaillard. He reported re-
search indicating that even vibrions that appeared to be identical could
have significantly different properties. Vibrions that are visually indis-
tinguishable could require totally different media and may quickly die if
introduced into an unsuitable medium. ‘It follows’, Davaine concluded,
‘that these species cannot be substituted for one another. Thus research-
ers seeking to elucidate certain pathological questions, must not hope to
determine identical phenomena by the introduction of vibrions drawn
from different sources’ (Davaine, 1864b, p. 633).
In 1865 a German physician named Huppert reviewed the anthrax

literature in an essay in a prominent periodical (Huppert, 1865). Huppert
was enthusiastic about Davaine’s research; in his opinion, it conclu-
sively demonstrated that bacteridia cause anthrax. He claimed that
Brauell’s work anticipated Davaine’s results and had been inadequately
appreciated only because, when Brauell conducted his studies, Pasteur
had not yet shown that each fermentation process was associated with a
unique ferment.
Brauell immediately responded to Huppert; he wanted, as he put it,
to defend German soil from the encroachment of French theories
(Brauell, 1866, p. 292). Brauell insisted the causal role of the anthrax
bacillus had not yet been established: ‘ignoring the fact that nothing is
known about the development and spread of the rod-shaped bodies, I
find no logically compelling reason for the opinion that these bodies
cause anthrax’ (Brauell, 1866, p. 293). Brauell pointed out that, even
given Davaine’s experiments, it was still possible that some other
component of anthrax blood was the true contagium. He mentioned
that Leplat and Jaillard had raised similar objections in France. Ac-
cording to Brauell, the only way to prove causality would be by
inoculating isolated vibrions. However, he doubted that such a test
could ever be conducted because, as he insisted, one could never
insure that the bacteridia had been completely freed from every other
ingredient in blood. Thus it would always be possible that the real
cause was some other entity – an otherwise unknown virus – that was
regularly associated with bacteridia.
Brauell reported that blood from anthracoid animals sometimes con-
tained no trace of rod-shaped bodies, and he also confirmed that
inoculations with such blood could sometimes induce anthrax. He men-
tioned other researchers, including even Davaine himself, who had
acknowledged similar results. Thus, he concluded, far from proving
causality, the research Huppert cited actually contained evidence against
it (Brauell, 1866, p. 294). Brauell wrote that he would be pleased if
Pasteur’s theory of fermentation could be applied to the contagious
diseases and if the rod-shaped bodies could be shown to cause anthrax
by blood fermentation. However, he regarded this as unlikely: he doubted
that all fermentation was due to living organisms, but even if it were, he
observed, it would still be doubtful that anthrax was a kind of fermen-
tation. Brauell wrote that, even if Pasteur’s work had been available
when he had conducted his own study of anthrax, his conclusions
would have been precisely the same.
In 1868 Davaine published a paper intended to establish, once and for

all, the role of bacteridia in anthrax. He began by observing that for
more than five years his research had revealed that
bacteridia are found in all anthracoid disorders and in all victims
of the vulnerable species; the appearance of these small beings in
the spleen, the liver, and in the blood precedes the onset of morbid
phenomena; finally, anthrax blood ceases to be contagious once
bacteridia disappear. These facts, and several others it would take
too long to review, appear to me to be adequate reasons for affirm-
ing that the development of bacteridia is the cause of anthrax.
However, these facts do not have the same value in the eyes of all
pathologists, and in recent times, some very esteemed authors have
expressed reservations in speaking of the role of the bacteridia in
producing anthrax.
(Davaine, 1868a, p. 144)
In this passage Davaine listed three reasons for accepting the causal role
of bacteridia: First, the organisms were always present in the disease
(this, of course, is the same observational evidence of necessity that he
mentioned in his first argument for causality). Second, the organisms
are in the blood before the onset of morbid alterations. Third, blood
ceases to be contagious once bacteridia have disappeared (in other
words, absence of bacteridia is associated with absence of the disease –
further evidence of necessity). Davaine felt this evidence justified the
claim of causality, but he acknowledged that some pathologists re-
mained unpersuaded.
After emphasizing the importance of the issue, Davaine explained

that, in the present paper, he intended ‘to establish in a definite way, the
role of bacteridia in anthrax’. He proposed to do this, first, by answer-
ing the objections his critics had raised, and second, by advancing an
additional new argument. He considered three objections. First, that
bacteridia were sometimes found in non-anthracoid animals (if true this
would show bacteridia are not sufficient for anthrax). In response, he
pointed out that those who made this claim were confusing bacteridia
with other organisms. Second, bacteridia were not always found in the
blood of animals that died after inoculation with anthrax blood (this
would show that bacteridia are not necessary). He wrote that these
observations could be explained either by assuming the animals in
question had died of some other disease before the bacteridia had
incubated or by assuming the observers were looking for bacteridia
only in blood from the major arteries whereas bacteridia often collected
‘in rather small fibrose clots, either white or semi-transparent, that one
may neglect to examine’ (Davaine, 1868a, p. 145). The third objection
was this: ‘it has been claimed that the vibrions are the effect and not the
cause of the alteration of the blood’ (Davaine, 1868a, p. 146). Recent
observers had maintained, Davaine continued, that the only way of
establishing that bacteridia are the anthrax virus would be by filtering
bacteridia from anthrax blood and by inoculating the isolated organ-
isms into healthy animals. However, he doubted that any man-made
filter could achieve this result, so he repeated his earlier experiments
using blood filtered through the placenta of a pregnant guinea-pig. He
found (as he and Brauell had both reported earlier) that blood from the
pregnant guinea-pig conveyed fatal anthrax while the blood of its fetus
did not (Davaine, 1868a, pp. 146f).
After responding to these three criticisms, Davaine gave a new indi-
rect argument in support of the causal role of bacteridia. Davaine
observed that ‘the invasion and subsequent destruction of an animal by
an infusorium is, as yet, a fact without analogue, and because it is so
singular, one requires a multitude of proofs before it can be admitted’
(Davaine, 1868a, p. 147). Davaine supported his claim by an analogical
argument based on his discovery that certain plants could also contract
bacterial diseases. His discovery had some inherent interest, but there is
no evidence his critics found it persuasive.
Later in the same year Davaine reported further experiments with
anthrax blood. He demonstrated that even minute quantities of infected
blood could be fatal. He also reported that the time required for the
onset of symptoms varied inversely with the quantity of inoculated
organisms (Davaine, 1868b). Both results made it appear that anthrax
was not simply a consequence of physical or chemical poisoning, and he
concluded that the cause of the disease had to be a living organism that
could multiply within the victim.
In 1870 Davaine published yet another important paper on anthrax.

The paper was read before the French Academy of Medicine and both
the paper and the ensuing discussion were published. In this paper
Davaine addressed one of the main problems raised by Brauell, namely,
if one assumed bacteridia cause anthrax, how could one account for the
spread of the disease. He began by listing various observations about
the spread of anthrax that seemed difficult to explain – for example,
anthrax sometimes appeared to spread without direct contact between
diseased and healthy animals, anthrax epizootics sometimes ended when
herds moved from one pasture to another, and changes in the weather
sometimes affected the course of an epizootic. He then criticized a few
of the inconsistent hypotheses that had been advanced to explain these
facts. Finally he presented his own hypothesis: while acknowledging
that other means of transference were possible, he proposed that an-
thrax was often conveyed by flying insects carrying minute traces of
bacteridia-bearing blood from diseased animals to healthy ones (Davaine,
1870a, p. 219). He supported this claim with several experiments, and
he showed the hypothesis could account for some observations that
were otherwise difficult to explain (Davaine, 1870a, p. 225). He sug-
gested how farmers could curb the effect of flying insects and, thereby,
protect their herds.
The discussion following Davaine’s paper provides insight into medi-
cal thinking in 1870 – seven years after he had begun arguing that
bacteridia cause anthrax. Of the eight persons whose comments are
reported, five specifically insisted that epizootic anthrax could better be
explained by the assumption that anthrax was often spontaneous
(Davaine, 1870a, pp. 231–5). The comments of a physician named
Leblanc were typical. Leblanc allowed that anthrax could be contagious
and could perhaps spread as Davaine had hypothesized, but he was
convinced that ‘the more usual cause of the great epizootics is the
spontaneous development of anthrax under the influence of [un]hygienic
and morbid influences’ (Davaine, 1870a, p. 232). Judging from the
reported discussion, even in 1870 many French physicians still accepted
the idea that diseases could originate spontaneously and most were
skeptical of the claim that anthrax had a single universal cause.
In a subsequent meeting of the Academy, Davaine defended his ac-
count of the spread of anthrax. He observed that the concept of
‘contagion is sufficient to explain the indefinite transmission of anthra-
coid afflictions without appealing to the outdated doctrine of heterogony’
(Davaine, 1870b, p. 490). He insisted that, whenever possible, medicine

should learn from its sister sciences. Without mentioning Pasteur by
name, he observed that recent work in natural history had shed new
light on the role of microscopic parasites and had so thoroughly dis-
credited the notion of spontaneous generation that the doctrine could
now be taken to apply, at most, to amorphous leukocytes – entities
without identifiable characteristics that seemed to be neither plants nor
animals. By contrast, he insisted, bacteridia had a fixed nature and
could be precisely classified among other living beings. Davaine ob-
served that, in the face of all that was known about bacteridia, it was
impossible to believe they arose spontaneously. Those who still accepted
this possibility were forced into the impossible situation of explaining
how bacteridia or their germs could suddenly appear in malignant
pustules, in anthracoid tumors, and in the liver, the spleen, and the
blood of every anthracoid animal (Davaine, 1870b, p. 491). It seemed
more plausible, he concluded, to regard them as spreading from one
host to another and as multiplying in the hosts’ bodies. Given this point
of view, the only remaining question was how they spread, and this, he
insisted, was usually by way of flying insects.
Davaine admitted that liquid anthrax blood quickly putrefied and,
once putrefaction began, it no longer contained bacteridia. However,
even if anthrax blood dried rapidly before the onset of putrefaction, it
continued to be contagious for months and possibly even for years. This
explained why anthrax occurred again and again in the same areas even
though the outbreaks were separated by long periods (Davaine, 1870b, p.
492). In conclusion, he pointed out that, on his theory, one could under-
stand how a drop of anthrax blood could dry and remain protected in a
stable for years until it happened to be deposited in the open wounds of
an animal, how this one animal could become ill and die, and how, by
way of flying insects, the diseased animal could be the source of infection
for a murderous epizootic. One could understand how, among all the
herds grazing in a given pasture, the animals sheltered in a given stable
might become diseased while those in other stables might not. One could
understand how the disease could be controlled by measures that would
prevent this from happening. Moreover, he insisted, ‘this way of thinking
was compatible with experimental results and with current knowledge of
the nature of organized ferments and of their means of generation either
within or outside the animal economy’ (Davaine, 1870b, p. 496). Davaine
was convinced he was right because his account could explain the facts
and was compatible with other current scientific beliefs; other accounts
did not meet these conditions.
In spite of Davaine’s emphatic defense, respondents continued to
insist that there still remained cases that could most plausibly be
accounted for in terms of spontaneous generation (Davaine, 1870b, p.

497). Thus, while acknowledging that the spread of bacteridia could
explain some cases of anthrax, they continued to believe that the dis-
ease did not always depend on the dissemination of bacteridia and their
germs.
Even after all of Davaine’s work, both French and German critics
remained unpersuaded. In their minds, he had not excluded the possibil-
ity that anthrax could arise spontaneously or through some combination
of factors not involving bacteridia (that is he had not proven causal
necessity), and he had not proven that bacteridia, when isolated from
every other ingredient of anthrax blood, would themselves bring on the
disease (that is, he had not proven causal sufficiency). A German writer
observed that Davaine had been unable to perform ‘the one experiment
that would have overcome all objections to his opinion on the signifi-
cance of bacteridia: namely, freeing the blood from bacteridia by filtration
through clay and inoculating the filtrate’ (Steudener, 1872, p. 304).
Explained in this way, the deficiency in Davaine’s argument appears to
have been purely technical – he seems only to have needed a more
reliable way of isolating bacteridia from the other ingredients in anthrax
blood. Indeed, Davaine and other researchers devoted considerable
effort to progressively more rigorous methods of separating bacteridia
from every contaminant that could have been causally relevant. How-
ever, no such effort could be completely successful until there was
agreement as to exactly which potential contaminants could be causally
relevant. In the absence of such agreement, no matter what steps were
taken to insure the purity of a given culture, skeptics could always
maintain there could still be some unknown substance that had not
been excluded and that could be the true cause of anthrax.
The only way of silencing this objection would have been by appeal-
ing to some hypothesis such as this: every disease is caused by bacteria.
We will call this the Bacterial Hypothesis, and we will defer until the
next chapter the objection that it is obviously false. No one questioned
that bacteridia were the only bacteria involved in anthrax. Thus if
Davaine’s critics had accepted something like the Bacterial Hypothesis,
they would clearly have accepted the bacteridium as the cause, rather
than continue holding out for some possible unknown entity. This is
like saying (1) if Davaine’s critics had accepted what Pasteur called the
Dissemination Hypothesis – the assumption that bacterial cultures never
arise spontaneously but only from antecedent bacteria – they would not
have taken seriously the idea that anthrax could occur spontaneously,
or (2) if Leplat and Jaillard had accepted the Distinguishability Hypoth-
esis they would not have attempted to induce anthrax by injecting

‘certain vegetable infusions, some liquids charged with decomposing
animal matter, putrefied urine, [and] the serum of decomposed blood’
(Leplat and Jaillard, 1864, p. 251). Regardless of whether the Bacterial
Hypothesis is defensible, this suggests that Davaine’s problems could
have been due as much to the lack of a shared theoretical framework as
to technical imperfections in his experiments.
Causation is ultimately a theoretical relation, so causal claims can
never be justified in the absence of a theory. To say a set of conditions
causes an event is to say our theories connect the conditions with the
event in certain ways – it does not mean the event is somehow, once and
for all, bonded to those conditions by ‘cosmic glue’ (Hanson, 1969, p.
64). Without a theoretical context to warrant these connections it is
impossible even to conceive of causation, and it is impossible, in princi-
ple, to advance beyond such empirical correlations as necessity or
sufficiency to justify a causal claim. Regardless of how much empirical
evidence Davaine may have accumulated, in the absence of a shared
theoretical framework his critics could never have been persuaded.
Davaine’s problem is suggested by the following passage from a
prominent and sympathetic nineteenth-century researcher, Otto Bol-
linger:
From the almost constant presence of the characteristic bacteria (or
bacterial germs) in the blood of anthracoid animals, in view of the
inoculation experiments … and of the unmistakable character of
anthrax as a blood disease, if one succeeds in explaining the clini-
cal and pathological appearances in terms of the properties and
operations of the bacteria, one could attempt to regard these small
organisms as the anthrax poison.
(Bollinger, 1875, p. 461)
In this summary, Bollinger mentioned nearly all the empirical evidence
one could imagine having for causality. But, the most Bollinger was
willing to hazard was that, given this evidence, ‘one could attempt to
regard’ bacteridia as the causal agent. Even given such evidence, in the
absence of what we have called the Bacterial Hypothesis (or something
like it) Bollinger was unwilling to conclude, categorically, that bacteridia
cause anthrax.
The purpose of the Bacterial Hypothesis is to certify that, given some
case of disease, of all the infinitely many conditions that could conceiv-
ably cause it (evil spirits, yesterday’s cauliflower salad, miasms,
astrological configurations, gazelles grazing on the Serengeti), only bac-
teria must be taken into account. By appealing to such a principle, and
only by such an appeal, empirical evidence can finally amount to a
proof of causation.
Note
1. French observers later dismissed these last two observations as resulting

from a failure to distinguish immobile anthrax bacilli from mobile organ-
isms that flourished in putrefying blood (Davaine, 1863a, p. 222).
CHAPTER SIX
A Bacterial Theory of Disease

Edwin Klebs was born 6 February 1834 in Königsberg. Klebs com-
pleted his medical studies in 1858 at the University of Berlin and, after
graduating, he practiced medicine for one year in Königsberg. Perhaps
because of an interest in Semmelweis’s work (Röthlin, 1962, p. 6),
Klebs was initially attracted to obstetrics but, in response to Rudolf
Virchow’s encouragement, he turned to pathology (Hirsch, 1935, p.
539). Klebs worked for a time as Virchow’s assistant and, in the early
1860s, published several unexceptional essays on pathology in a journal
edited by Virchow.1 In 1866, he accepted a position at the University of
Bern. Five years later, in the first volume of a periodical he founded and
edited, Klebs reported investigations he had conducted on anthrax and
on infected gunshot wounds (Tiegel, 1871; Klebs, 1871). These papers
mark a significant and permanent change in his research orientation.
At the outset of his work he had followed Virchow’s general research
approach (Röthlin, 1962, pp. 6–7). In his early essays, like Virchow and
other pathologists, he focused on internal morbid processes and dis-
played little interest in remote causes. His discussion of an outbreak of
epidemic meningitis was typical: Klebs described in detail the sequence
of pathological changes characteristic of the disease, but regarding causes
he observed only that ‘the local influence of cold and the pressing
together of the population into a small area must be taken into account’
(Klebs, 1865, p. 379). However, his studies of anthrax and wound
infections seem to have persuaded Klebs that these diseases occurred
only under the influence of microorganisms. Thereafter he became pro-
gressively more interested in parasitic organisms as causes of disease.
Within a few years, Klebs was vigorously criticizing the pathologists for
their narrow concern with internal disease processes and their indiffer-
ence to what, in his opinion, mattered most about disease causation.
Klebs noted that, in his later years, Henle ‘recognized, as causes of
disease, only general physical and chemical influences, the life impulses,
the same factors that have often been identified by others who wrote
before and after him. The concept of a specific cause of a disease, which
is absolutely destructive of life, is alien to him as to most other patholo-
gists’ (Klebs, 1878a, p. 46). Klebs also objected to Rudolf Virchow’s
uncritical adoption of antiquated and fruitless causal notions (Klebs,
1878a, p. 46; Klebs, 1878b, pp. 133f.) and Virchow’s few scattered
remarks about remote causes suggest that this criticism was entirely
A BACTERIAL THEORY OF DISEASE 91
justified.2 In response to his former student’s telling criticisms, Virchow

admitted that pathologists were less interested in external remote causes
than in internal disease processes, but he also warned that, by ignoring
internal dispositions, Klebs risked falling into the opposite error: ‘evi-
dently Klebs is … of the opinion the internal arrangement of the tissues
is irrelevant [in determining the effect of a remote cause], or, expressed
in terms of universal pathology, that the external cause is an Agens
Causa sufficiens for all the consequences’ (Virchow, 1880, p. 9). Klebs
ignored this feeble response. Because the pathologists had limited their
attention to internal phenomena, in his opinion, they were unable to
contribute significantly to practical medicine. Klebs wrote that he had
abandoned Virchow’s interests and devoted himself to the ‘progressive
expansion of those teachings that seek to find the causes of many
diseases outside the cell boundaries and thereby make into scientific
truth the opinions of Schönlein, Remak, and Henle (in his earlier years)’
(Klebs, 1874a, p. 207).
Through the late nineteenth century, Klebs published important stud-
ies on the etiologies of a variety of diseases; an excerpt from Fielding H.
Garrison’s history of medicine shows the breadth of his interests:
He saw the typhoid bacillus before Eberth (1881), the diphtheria
bacillus before Löffler (1883), used solid cultures of sturgeon’s glue
(1872), and investigated the pathology of traumatic infections be-
fore Koch (1871). The priority of his inoculations of syphilis in
monkeys was recognized by Metchnikoff (1878), and, in his ex-
periments on anthrax (1871) and other diseases, he was the first to
filter bacteria and to experiment with the filtrates. In 1874, he
invented the fractional method of obtaining pure cultures of bacte-
ria … which was followed by Lister’s method of dilution (1877)
and prepared the ground for Koch’s work. He wrote two text-
books on pathology (1869–76, 1887–89), monographs on the
bacteriology of gunshot wounds … (1872), on tumors (1877), and
gigantism (1884), made many investigations on tuberculosis, and
he was, with Gerlach, the first to produce bovine infection or
Perlsucht by feeding with milk (1873). In his studies on gunshot
wounds (1872) he showed that the filtrate of the wound-discharges
is non-infectious, whence he reasoned that traumatic septicemia is
of bacterial origin.
(Garrison, 1929, p. 581)
Garrison also mentioned that Klebs probably did more than any other
single researcher to win the pathologists over to the bacterial theory of
infection – no mean accomplishment (Garrison, 1929, p. 580). It was
typical of Klebs that he gave cursory attention to many common dis-
eases, but made thorough studies of few – this inevitably led to mistakes:
Garrison does not mention that Klebs also claimed to have identified
bacterial causes of smallpox, malaria, and goiter. On the other hand,
the very breadth of his interests gave Klebs a perspective from which to
approach philosophical issues. He gave more extensive and more ex-
plicit attention to causal criteria than any of his contemporaries, and he
stated explicitly some of the basic assumptions underlying the bacterial
theory of disease to which he was contributing. These are the two
aspects of his work that will be of interest to us, but before we examine
Klebs’ views, we must briefly review some developments in bacteriology
that provided the context of Klebs’ own investigations.
Bacteriology in the 1870s
Since the late eighteenth century, pathologists had discussed the morbid
alterations typical of infected wounds. In infected tissues one found soft
gray spots the size of pin heads. In advanced cases the spots became
larger and more numerous, they merged, and their centers degenerated
into a substance resembling pus. These spots seemed to be new sites to
which infection was spreading, and they were called metastases. In
1866 Eduard Rindfleisch, a German pathologist, found that metastases
did not contain pus corpuscles, but rather microorganisms (Rindfleisch,
1867–69, pp. 183, 204). In harmony with existing terminology, he
identified the organisms as vibrions. Rindfleisch reported that metastases
originally contained vibrions densely packed between strands of muscle
tissue; later the organisms spread into surrounding tissues and the
muscle fibers decomposed. Another common morbid product of wound
infections was a gray film that gradually spread over wound surfaces.
Similar structures developed in the throats and lungs of diphtheria
patients. As early as 1868 researchers reported finding clusters of small
spherical organisms in these gray films. These organisms were called
micrococci. However, finding living organisms in infected wounds did
not prove they caused the morbid processes, and the exact role of the
organisms remained a matter of dispute.
By the beginning of the 1870s, many different observations of micro-
organisms had been reported, but the reports were of uneven quality
and, in the absence of photographs, they were supported at best by
freehand drawings of the organisms in question. Under these conditions
there was confusion about who had seen what and this uncertainty was
compounded by the lack of a consistent nomenclature. Many writers
followed Christian Gottfried Ehrenberg’s 1824 taxonomic system, but
as new organisms were discovered, this scheme became progressively
less adequate and researchers simply made up names of their own (like
Davaine’s ‘bacteridium’). Often there was considerable variation in ter-
minology even within the works of a single author.
In 1872, Ferdinand Julius Cohn, a German botanist at Breslau, noted

‘anyone who studies the literature, even from recent years, knows there
is nearly chaotic confusion in the names of the bacteria’ (Cohn, 1872, p.
128). In 1872 Cohn published an extensive review of the literature and
recommended a new nomenclature and scheme for classifying bacteria.
This scheme was widely adopted and became especially popular in
Germany. Cohn characterized bacteria as ‘cells lacking chlorophyll whose
shape is spherical, oblong, or cylindrical, occasionally twisted or stunted,
which proliferate exclusively by division and which vegetate either in
isolation or in cell families’ (Cohn, 1872, p. 136). He distinguished
spherical, rod-shaped, thread-like, and spiral-shaped bacteria. Each group
included various organisms. However, Cohn mentioned only those that
were clearly documented in the literature, and he himself confirmed the
existence and characteristics of most of the organisms he described.
Cohn’s spherical bacteria included the smallest organisms that had
then been observed; all were incapable of independent movement. Be-
cause they were so small, some were difficult to locate and to identify.
However, small spherical bacteria were obviously important since they
were the organisms most frequently associated with human diseases, and
Cohn gave them more attention than those of any other group. The
spherical bacteria included some of Pasteur’s ferments and the micrococci
commonly observed in diphtheria and in puerperal fever. Rod-shaped
bacteria were also small, but they differed from spherical bacteria in
that they were capable of independent motion. This group included the
organisms usually associated with putrefaction as well as the ferments
to which Pasteur ascribed the souring of milk and the creation of
vinegar. None of the rod-shaped bacteria that Cohn identified were
known to be pathological. Thread-like bacteria included the anthrax
bacillus and Pasteur’s butyric ferment. According to Cohn, all the or-
ganisms in this group were filaments that could, under suitable
conditions, be seen to form extended cylindrical elements. Bacteria in
this group varied in length, in thickness, and in flexibility. Spiral-shaped
bacteria included, as separate subgroups, the organisms Ehrenberg re-
ferred to as the spirillum and the spirochaete.
At the time of Cohn’s work, micrococci were regularly reported in a
variety of diseases especially in several forms of wound infections. In
1871 Friedrich Daniel von Recklinghausen found micrococci in the
kidneys, veins, urinary tracts, and lungs of childbed fever victims
(Recklinghausen, 1871). In further studies of puerperal fever, Wilhelm
Waldeyer found spherical bacteria in the kidneys and heart muscles
(Waldeyer, 1872) and Felix Victor Birch-Hirschfeld observed micrococ-
cal masses in vaginal ulcers, in paravaginal cellular tissues, and in
victims’ blood, spleens, and livers (Birch-Hirschfeld, 1872, p. 105). In
France, Gustave Nepveu found micrococci in the blood of erysipelas

patients; he reported that the organisms were more numerous in blood
drawn from the area of infection than in blood drawn from remote
body parts (Nepveu, 1872).
In 1864, while studying an epidemic of relapsing fever, Otto Obermeier
noticed fine thread-like structures in the blood of some victims. The
epidemic ended at about the time he first observed the filaments, and,
having no opportunity to confirm his discovery, he did not report it
(Obermeier, 1873, p. 33). A few years later, in a new epidemic of
relapsing fever, Obermeier found the same structures in the blood of
new patients. In an 1873 paper, he reported that the structures were
present only during the fever stage of the disease or within a short time
before or after the fever stage. He also established that, while filaments
could be found in the blood of most patients, they were not present in
the blood of four healthy persons or of three persons suffering from
other diseases (Obermeier, 1873, p. 34f.). The filaments moved vigor-
ously and could most easily be detected by looking for certain irregular
movements among the blood corpuscles. In his published report he
drew no conclusions about whether the filaments had pathological
significance or were simply modifications of normal body tissues.
Obermeier reported showing blood samples to two professors in his
clinic, Carl Westphal and Rudolf Virchow, and to several colleagues;
each was able to identify the living filaments in the samples (Obermeier,
1873, p. 35). Within a few months of his discovery, in connection with
further research, Obermeier contracted cholera and died.
Two years later, in a paper reviewing the literature on bacteria,
Ferdinand Cohn wrote that Obermeier’s discovery was ‘clearly the most
important fact [discovered] in recent times about the occurrence of
fermenting organisms in the infectious diseases’ (Cohn, 1875, p. 196).
According to Cohn ‘the discovery has been confirmed, without excep-
tion, by all later observers’. Cohn classified Obermeier’s structures as
spirochetes. He pointed out that, prior to the discovery, only one species
of spirochete was known – a species observed in swamp water by
Ehrenberg and later by Cohn himself, but it was difficult to find and
had rarely been reported. According to Cohn, ‘the most important fact
about the spirochete of relapsing fever … is that the threads are found
exclusively in the blood of relapsing fever victims – never in their
secretions or in other organs – and only during the paroxysm or a short
time after the attack … , but never during the interval between fevers’
(Cohn, 1875, p. 197). The regular occurrence of the spirochetes in the
blood seemed so clearly associated with the onset of symptoms that,
even in the absence of other evidence for causality, most observers seem
to have accepted the spirochetes as the cause of relapsing fever. In
several papers Klebs cited Obermeier’s work and seems to have taken it
for granted that Obermeier had proven causation (Klebs, 1878a, p. 51).
Thus, relapsing fever became the first disease widely believed to be
caused by a spirochete.
Klebs on causal criteria
Klebs’ first bacteriological publication was an 1871 paper on anthrax

coauthored with his student, Ernst Tiegel. Tiegel and Klebs reported
new experiments with anthrax blood; the goal was to determine whether
bacteridia were the cause of anthrax (Tiegel, 1871). Tiegel filtered
anthrax blood through clay to remove the bacteridia. He then showed
that injecting the filtration residue, which included bacteridia, produced
anthrax while injecting filtered blood did not. In a note at the end of
Tiegel’s paper, Klebs explained that ‘Davaine wished to isolate the
bacteria [from body fluids] while we undertook to free the liquids from
the bacteria. Given our method, the latter is quite easy, but the former
remains absolutely impossible’ (Tiegel, 1871, p. 280). Tiegel and Klebs
sought to answer objections to Davaine’s work, but their paper seems
not to have had much impact. Neither Otto Bollinger nor Felix Victor
Birch-Hirschfeld cited the Tiegel-Klebs article in their thorough reviews
of the anthrax literature that appeared only four years later (Bollinger,
1875; Birch-Hirschfeld, 1875).
Klebs’ work on wound infections attracted more attention. Klebs
studied infected gunshot wounds while serving in a military hospital,
the Bahnhofslazarette, in Karlsruhe during the Franco-Prussian war. In
his first paper on the topic, Klebs reported that a particular minute
spherical organism, which he called Microsporon septicum, seemed to
cause both fever and wound putrefaction (Klebs, 1871). He observed
that wounds did not putrefy simply because tissues were damaged or
even because they contained foreign matter; instead the health of a
wound seemed to vary inversely with the quantity of Microsporon
septicum it contained. He also reported that body fluids that had been
filtered to remove the organisms did not cause fatal diseases in rabbits
whereas unfiltered fluids produced local festering and death. Klebs
obviously believed that specific organisms had unique causal roles in
anthrax and in wound infections; the organisms were not simply one
possible remote cause among many.
Klebs’ work in Karlsruhe culminated in a book entitled Contributions
to the Pathological Anatomy of Gunshot Wounds (Klebs, 1872), a
work frequently cited by later researchers. Koch praised it as ‘the first
attempt to demonstrate a causal connection between bacteria and infected
wound diseases’ (Koch, 1878a, p. 22). In Contributions Klebs first

stated explicit criteria for disease causality. He observed that ‘tracing
the invasion and the course of microorganisms can make causality
probable, but the crucial experiment is isolating the efficient cause and
allowing it to operate on the organism’ (Klebs, 1872, p. 105). We can
state these conditions as follows:
E1. The presence of the organism correlates with disease phenomena.

E2. Inoculation of isolated organisms is followed by the disease.
It is unclear exactly how E1 is to be understood. Klebs could have

meant, starting with disease phenomena, always find organisms to be
present (which was Pasteur’s approach and would suggest the organ-
isms are necessary), or, starting with advancing organisms, show that
disease phenomena always follow (which would suggest sufficiency);
most likely actual practice would have involved a combination of the
two. In any case, E2, which Klebs regarded as more conclusive, clearly
provides evidence of sufficiency. Bassi, Semmelweis, Davaine, and Pas-
teur all reported what happened when healthy hosts were exposed to
purported causes, but they all seem to have been most persuaded by
evidence of necessity. Perhaps Klebs’ different preference reflected his
background as an experimental pathologist – as we saw in Chapter 1,
the general orientation of pathology favored thinking in terms of suffi-
cient causes.
In 1872 Klebs moved to Würzburg and shortly thereafter gave a
lecture on micrococci as causes of disease (Klebs, 1874b). Klebs noted
that examination of blood samples revealed that micrococci, while
normally absent from healthy tissues, were present in all cases of certain
diseases. He observed that the next task was proving these bodies were
alien parasites. Klebs described a new method for isolating and cultur-
ing microorganisms, a method that reduced the possibility of accidental
contamination; he called his new technique fractional cultivation (Klebs,
1874b, p. vi). Body fluids from diseased animals were filtered and the
residue washed in distilled water. Organisms from the residue were then
purified by passing them through a series of cultures each of which was
seeded with a minute sample from the preceding culture. By passing the
organisms through numerous subcultures, Klebs sought to remove every
possible disease product except the living organisms that multiplied in
each successive medium. Davaine had used similar methods to study
anthrax, but Klebs carried the approach to new levels of technical
sophistication. He reported that inoculations after fractional cultivation
still produced new artificial cases of certain diseases. He concluded ‘that
specific disease processes were due to various specific organisms’ (Klebs,
1874b, p. vii). Klebs’ process of fractional cultivation received favorable

attention from Koch and others.
In 1874 Klebs was appointed a professor of pathology at the Univer-
sity of Prague. In Prague he was a founding editor of another journal,
and he published a series of essays on microorganisms as disease agents.
His first essay reviewed some of the reasons for associating microorgan-
isms with specific diseases. He wrote that, if a disease could be induced
by exposing an animal to fully isolated microscopic bodies, ‘their sig-
nificance as causes of disease could no longer be doubted’ (Klebs, 1873,
p. 33). Thus, satisfaction of E2 is conclusive. He discussed examples of
research that approached this ideal including Davaine’s work on an-
thrax.
In 1875 Klebs wrote that if one can show ‘inflammation and other
reactive changes follow, step by step, the spread of the schizomycetes, it
is logical to infer a causal relation rather than a simple coincidence’
(Klebs, 1875–76, p. 321).3 The criterion Klebs here identified for prov-
ing the schizomycetes are pathological is E1, but as here described it
seems clearly to support sufficiency, not necessity. Klebs noted that
experimental evidence could yield the same conclusion. To obtain such
evidence, he wrote, one must ‘isolate substances from the body and use
them to induce further cases of infection’ (Klebs, 1875–76, 321). Klebs
reported following both approaches and obtaining mutually supporting
results. In the conclusion of the paper, Klebs observed that the goal now
was to identify more external causes and to characterize them precisely
just as, in Virchow’s cellular pathology, one characterized cellular mor-
bid changes (Klebs 1875–76, p. 324).
On 20 September 1877, Klebs delivered a controversial address to the
Society for German Natural Scientists and Physicians entitled ‘On the
Revolution in Medical Opinions in the Last Three Decades’ (Klebs, 1878a).
Klebs observed that pathology had contributed almost nothing to practi-
cal medicine, and that recent advances resulted from associating particular
diseases with specific remote causes. He identified three areas of research
as relevant to demonstrating disease causation: the study of diseased
organs, the isolation and cultivation of germs, and the generation of new
cases of disease by exposing healthy animals to germs (Klebs, 1878a, p.
49). Near the end of his lecture, he described, in more detail than usual,
two criteria for establishing the causal significance of germs: ‘(i) If organ-
isms are identified that are well characterized and that are found exclusively
in the given disease process, anatomical evidence can be conclusive. (ii) If
the form of the organisms provides no certain point of departure, it can
be decisive to convey the disease by means of organisms that have been
isolated and cultivated outside the body’ (Klebs, 1878a, p. 51). These
criteria correspond to E1 and E2 respectively. Klebs’ remarks suggest that
correlating disease phenomena with advancing organisms can prove cau-

sation, but only for organisms that are conclusively identifiable and
found exclusively in the given disease. Thus, as here stated, the first way,
E1, requires the satisfaction of three conditions: (1) the organisms must
be distinguishable from other organisms; (2) given the parasitic organism,
one always observes the disease process; and (3) the presence of the
organism must correlate with disease phenomena. Supposedly, the second
method, E2, can work even if the organisms in question are not conclu-
sively distinguishable from organisms associated with other diseases. Klebs’
lecture was reported in medical periodicals (for example, Berliner klinische
Wochenschrift, 1877), and was widely known in German medical circles.
As we will see later (pp. 141–2), it may have been the immediate source
for the first version of what we now know as Koch’s Postulates (Carter,
1985b).
In 1878, Klebs responded to Louis Pasteur’s claim to have been the
first to have established conclusively the cause of anthrax (Klebs, 1878c).
Klebs discussed ‘the beautiful work of Davaine’ that made it ‘extremely
probable that anthrax is caused by bacteridia’. He noted that ‘the
conclusive proof of this opinion is provided by the ineffectiveness of
anthrax fluids that have been deprived of bacteridia’, in other words,
proof that inoculation of liquids not containing the organisms is not
followed by the disease. Satisfaction of this test provides evidence the
organism is necessary (but not that it is sufficient) for the disease. This
passage is the closest Klebs ever came to stating that evidence of neces-
sity was relevant to proving causation, and here he suggested that such
evidence was conclusive. However, surrounded as it was by repeated
appeals to evidence of sufficiency, this passage should probably be
discounted as anomalous. According to Klebs, in showing that fluids
deprived of bacteridia have no effect, Pasteur had merely repeated the
experiments that he and Tiegel had reported in their 1871 anthrax
paper.
In a long paper published in 1879, Klebs argued that a specific
bacillus may cause malaria. He wrote that ‘to convert this hypothesis
into a scientific theory one must decide, first, whether the bacillus
occurs in malarial soils or in the surrounding air, and second, whether
this organism itself, without the support of any other disease causing
agent, can cause true intermittent fever’ (Klebs and Tommasi-Crudeli,
1879a, p. 326). In a summary of the paper, Klebs explained that the
second of these steps consisted of propagating bacilli through frac-
tional cultivation and then inoculating both filtered liquids and the
cultivated organisms to determine the effect of each (Klebs and
Tommasi-Crudeli, 1879b). Thus, once again, the crucial test is E2 –
isolation and inoculation.
In an 1881 paper on abdominal typhoid fever, Klebs observed that ‘in

every carefully examined case, one and the same form of schizomycete
has been identified in the fresh and intensively developed changes’
(Klebs, 1880, pp. 232f.); here Klebs is invoking E1. Later in the same
year, in another paper on typhoid, he observed that ‘obtaining the same
results in repeated attempts gave us a certain confidence that the result
was not merely accidental. … In the meantime I have provided experi-
mental proof the organism is the carrier or the cause of typhoid poison.
I now regard the question of the cause of typhoid fever as closed (Klebs,
1881, p. 381). The experimental proof Klebs provided was isolating
and inoculating the organism.
Between 1872 and the early 1880s, Klebs consistently invoked one or
both of E1 and E2: these two approaches to causality require, respec-
tively, establishing that the advance of certain clearly identifiable
organisms corresponds to disease phenomena, and isolating and inocu-
lating a possible causal agent. Many of Klebs’ contemporaries deliberately
patterned their experiments on these causal criteria. In 1877 Felix Vic-
tor Birch-Hirschfeld discussed these strategies in his widely used
pathology textbook. He observed that Klebs and others tried to explain
the pathological significance of microorganisms by correlating advanc-
ing parasites with the sequence of morbid changes (Birch-Hirschfeld,
1877, p. 233). According to Birch-Hirschfeld, recognition that this was
inconclusive led to experiments in which ‘bacteria were isolated in
various ways from the liquid constituents of infectious substances. One
then compared the results of inoculating isolated bacteria and the other
liquid materials’ (Birch-Hirschfeld, 1877, p. 234). Birch-Hirschfeld called
Klebs’ use of these approaches ‘epoch making’ (Birch-Hirschfeld, 1872,
p. 98). Other early bacteriologists who deliberately adopted Klebs’
methods included Friedrich Steudener (Steudener, 1872, p. 300), Leopold
Landau (Landau, 1874, p. 529), and Max Schüller (Schüller, 1876,
p. 160).
A bacterial theory of disease
In 1875 and 1876 Klebs wrote a series of papers summarizing his

research on the pathological schizomycetes. The last of these papers
contains at least part of an explicit theory of disease.4 As when Henle
approached similar fundamental questions, Klebs drew on an analogy
with ‘the exact sciences’. According to Klebs, these sciences (and, by
following their example, pathology and physiology) make progress ‘by
simplifying conditions before one puts questions to nature’ (Klebs, 1875–
76, p. 376). The answers one obtains through such simplifications,
Klebs continued, are valid even if later experiments, conducted under

more realistic conditions, yield unexpected results. Klebs’ idea is that in
studying etiology one must adopt similar simplifications. Here is an
example of what he may have meant. Scientific theories typically in-
volve basic assumptions from which explanations can be derived; for
example, in a kinetic theory of gases one makes assumptions such as the
following:
(1) A pure gas consists of a large number of identical molecules
separated by distances that are great compared with their size. (2)
The gas molecules are constantly moving in random directions
with a distribution of speeds. (3) The molecules exert no forces on
one another between collisions, so between collisions they move in
straight lines with constant velocities. (4) The collisions of mol-
ecules with the walls of the container are elastic; no energy is lost
during a collision.
(Oxtoby and Nachtrieb, 1990, p. 104)
Some or all of the assumptions associated with a particular theory may
be literally false. For example, molecules are not identical; and, like all
physical bodies, they exert gravitational forces on one another so they
do not move in straight lines or with constant velocities. However, such
false assumptions are justified because, by simplifying calculations, they
enable us to predict (within ordinary requirements for accuracy) what
will happen in the physical world. Without such assumptions, if one
sought to take account of the actual shape of each molecule, of gravita-
tion and other forces, and of all other distorting influences, the
calculations necessary even for relatively simple problems would be
impossibly complex. This may be what Klebs meant in writing that in
the exact sciences one simplifies conditions before putting questions to
nature. Klebs also wrote that the answers obtained in this way are still
valid even if later experiments, conducted in the complex conditions of
the physical world, yield unexpected results. Suppose, by using the
simplified properties of an ideal gas, one calculates what will happen if
one heats a sealed container. Perhaps Klebs meant there is a sense in
which this prediction is still valid even if it turns out to be incorrect
(perhaps the container explodes or dissipates heat more quickly than
one expected).
What does this mean for etiology? Klebs identifies four Grundversuche.5
On a superficial reading the Grundversuche may appear to be yet more
causal criteria (Carter, 1987b), but this is surely incorrect. In this context
Klebs does not mention E1 or E2, the criteria that consistently appear in
his other papers, both before and after this one, and in none of the papers
in which he uses and discusses causal criteria does he mention the
Grundversuche. Moreover, in this context his interest is different. He is
not arguing that some particular organism causes some disease (an argu-
ment that would require E1 or E2); instead, he is addressing the
fundamental and more general issue of establishing a context within
which any causal demonstration can be given. He is asking: what are the
basic conditions on which any such demonstration depends? Rephrasing
the question in terms of his analogy with the exact sciences, Klebs is
asking: what conditions must be met in order successfully to ask nature
(by applying E1 or E2) whether a particular schizomycete causes a certain
disease? Read in this way, the Grundversuche are ideal or simplifying
conditions basic to any investigation of disease causation.
Klebs writes:
In respect to the pathological influence of the schizomycetes, as
such fundamental tests [Grundversuche], I give heed to the follow-
ing: first, the proof that they never occur in the body fluids or in
some part of the secretions of a healthy organism; second, that they
do not develop spontaneously in liquids appropriate to their nour-
ishment … third, the fact that with the mechanical separation or
the natural death of the schizomycetes, the potency of the liquids
that contained them ceases; and fourth, the fact that local or gen-
eral infections of the [different] individual forms of pathogenic
schizomycetes, which have now been at least partially character-
ized, bring forth changes that are precisely characterized, uniformly
similar, and adequately distinguishable from [those associated with]
the operation of schizomycetes of other forms.
(Klebs, 1875–76, pp. 376f.)
As an idealized basis for a theory of disease, the Grundversuche tell
us schizomycetes have these properties: first, they are absent from the
body fluids and secretions of healthy animals; second, they never arise
spontaneously, not even in fluids that satisfy their nutritional needs;
third, once removed (artificially or otherwise) from some liquid, the
liquid alone has no pathogenic effect; and fourth, distinct schizomycetes
cause different pathological processes. In this passage, Klebs speaks of
the third and fourth Grundversuche as ‘facts’ and of the first as ‘a
proof’. However, the second resembles what we have called the Dis-
semination Hypothesis (p. 66 above), and as Pasteur realized, this
hypothesis can be neither proved nor disproved. While such principles
can be illustrated and made plausible, strictly speaking they cannot be
proven and they are not facts in the ordinary sense of the word. It seems
most plausible to refer to all four as hypotheses, and this is consistent
with Klebs’ idea that the Grundversuche are simplifying conditions
under which one can ‘put questions to nature’.
Four hypotheses emerge from the Grundversuche that are of particu-
lar interest. In stating these, we use the more common term ‘bacterium’
in place of Klebs’ term ‘schizomycete’.
1. The first Grundversuch assures us that healthy animals are free

from all bacteria. This hypothesis is basic to the use of test animals
in inoculation experiments because it insures that any organisms
found in a test animal, after inoculation, are the result of the
inoculation. We will see in Chapter 7 that Pasteur gave explicit
attention to this hypothesis.
2. The second Grundversuch insures that bacteria never arise sponta-
neously. This differs from the Dissemination Hypothesis only in
being restricted to bacteria (or schizomycetes) rather than being
about all microorganisms.
3. The third Grundversuch stipulates that any liquid that is free from
bacteria has no harmful effect; in other words, pathologicality is
always due only to bacteria. Thus, in general, every pathological
process is caused by bacteria. In discussing Davaine’s research, we
called this the Bacterial Hypothesis.
4. The fourth Grundversuch tells us that distinct bacteria cause differ-
ent pathologies. As stated, this hypothesis is inessential: many distinct
organisms can cause the same pathological processes (as for example
the common cold). However, the converse of this hypothesis (which
Klebs, an unclear and sometimes imprecise writer, could have had in
mind) is more important; it tells us that different pathological proc-
esses are caused by distinct bacteria. By virtue of being limited to
bacteria (or schizomycetes) this is a special case of the more general
Distinguishability Hypothesis that we encountered in Pasteur.
A few explanatory observations are necessary: First, the Grundversuche

or the hypotheses we have derived from them, are clearly laws in the
sense discussed by Wim J. Van der Steen and Harmke Kammingaa (Van
der Steen and Kamminga, 1991) and, given their definition of ‘theory’
as a set of interconnected laws, the Grundversuche clearly constitute a
theory – so far as I know, the first, and perhaps the only, explicit theory
of disease in the history of medicine.
Second, a complete theory may require additional assumptions, and
these four may require modification; they are intended only as a first
approximation to illustrate how a bacterial theory of disease looks and
works.
Third, the Bacterial Hypothesis refers to the cause of pathological
processes. For use as a technical term within the theory, one can define
the word ‘cause’ as anything within the domain specified by the Bacte-
rial Hypothesis (that is, any bacterium) that satisfies acceptable criteria
(perhaps E1 and E2) and explains disease phenomena. Given this defini-
tion, if some bacterium (say Davaine’s bacteridium) satisfies the causal
criteria in respect to a pathological process and explains the disease
phenomena, then, by this definition, it is the cause of that process.

Thus, we see how this theoretical assumption warrants inferences from
empirical evidence to causal claims.
Fourth, these hypotheses do not exclude the possibility that more
than one bacterium will satisfy the causal criteria with respect to the
same pathological process and, therefore, qualify as its cause. In other
words, these hypotheses do not insure that causes are universal and
necessary. Universal necessity is achieved by defining each disease in
terms of its cause. Having first established that some bacterium causes a
given pathological process (in the sense of ‘cause’ just defined), one
characterizes a particular disease as infestation by that bacterium. This
definition comes after one has identified a sufficient cause for at least
some cases of the process. By such a definition, one insures that the
cause of each disease, as so defined, is universal and necessary even if
two or more etiologically defined diseases are symptomatically indistin-
guishable. For example, different instances of the common cold may be
symptomatically indistinguishable, but, once defined in terms of distin-
guishable organisms, they count as different diseases. Then since, by
definition, every case of each disease has the same one cause, disease
phenomena can be consistently explained in terms of that cause and one
can seek means for controlling the disease by manipulating the cause.
Fifth (and finally), the phrase ‘pathological process’ in the last two
hypotheses can mean either internal organic lesions or collections of
signs and symptoms. However, once a disease is defined etiologically –
say as infection by a certain bacterium – strictly speaking, the phrase
‘pathological process’ cannot be understood in the sense of a bacterial
infection. Otherwise, we end up saying that bacterial infections cause
bacterial infections; but nothing can cause itself. This means (again,
strictly speaking) that we cannot both define a disease in terms of a
causal factor and continue to hold that the disease is caused by that
factor. However, both in technical medical treatises and in everyday
conversation, we ignore this problem and slip between ‘disease’ in the
sense of ‘a collection of symptoms’ and disease in the sense of ‘infection
by a particular organism’. For example, a medical text might define
‘leprosy’ as infection by Mycobacterium leprae and yet identify Myco-
bacterium leprae as the cause of leprosy (and sometimes even add that a
certain percentage of cases are clinically inapparent).6 However, this
ambiguity is useful and there is no point in trying to avoid it.
We have already discussed the Dissemination and Distinguishability
Hypotheses in connection with Pasteur’s early studies. No further com-
ments are required for the Dissemination Hypothesis. We will now
consider the Bacterial Hypothesis and then return briefly to the Distin-
guishability Hypothesis.
The Bacterial Hypothesis tells us that each disease is caused by some

bacterium; as it stands, understood as a claim about the world, this is
obviously false. As we saw in discussing Davaine’s research, the purpose
of the Bacterial Hypothesis is to limit the field in which causes are to be
sought. The need for such limitations is clear from nineteenth-century
etiological discussions. At about the same time that Klebs wrote the
Grundversuche, another prominent researcher, Julius Cohnheim, observed
that existing etiology is ‘an unbounded domain’ including ‘cosmic
physics, meteorology and geology no less than the social sciences and
chemistry as well as botany and zoology’. He observed that etiological
discussions in common texts include everything ‘from temperament to
beds, from air electricity to fungi and fleas, from inheritance to drinks’
(Cohnheim, 1877a, vol. 1, pp. 8f.). Only the imposition of limitations,
such as those insured by the Bacterial Hypothesis, could produce order
in this chaos. However, one might object that the Bacterial Hypothesis
is too limiting – there are certainly causes of disease besides bacteria,
and some (for example, acari, trichinae, and various poisons) were
known before Klebs wrote the Grundversuche. Robert Koch criticized
unnamed contemporaries for assuming that all diseases were bacterial
(Koch, 1881a, p. 119), and it is imaginable that Koch had Klebs in
mind. On the other hand, as Klebs himself made clear, the Grundversuche
are only a simplification or an idealization and, as this suggests, he may
not have regarded them as literally true.
Ultimately, it matters little what Klebs himself believed. A more serious
question is: what is the value of adopting an assumption that is factually
false? In line with Lakatos’s account of research programmes, all four
Hypotheses should probably be regarded, not as factual claims, but as
strategic directives or rules of method. Taken in this sense, the Bacterial
Hypothesis tells us: when confronted by some new and unexplained
disease, assume it is caused by a bacterium. If understood in this way and
if addressed to researchers who have techniques for studying bacteria but
who are unprepared to study anything else (as would have been true of
Klebs’ intended readers), this directive not only makes good sense, it is
absolutely indispensable. And this is true even if one knows in advance
that, as a claim about the world, the Hypothesis is literally false – that is,
even if one knows some diseases are non-bacterial. To act otherwise
would be to begin by refusing contemporary researchers access to the
only tools they were prepared to use. Rational inquiry must begin with
the expectation that new problems can be solved by existing methods –
that the unknown is homogeneous with the known. Only when this
expectation is defeated is it rational to abandon proven methods and
look for a new approach. From this point of view, understood as an
idealized strategic rule, the Bacterial Hypothesis is entirely reasonable.
Now we return to the Distinguishability Hypothesis. Davaine, Koch,

and Pasteur were all forced to rebut the idea that all bacteria were
ultimately of one species – that, given suitable conditions, any bacte-
rium could assume the characteristics of any other. Suppose this were
true; what would it mean for a germ theory of disease? We have already
seen (in Chapter 3) that organisms must be distinct if they are to
explain different organic processes, but in the context of theories of
disease it is worth sharpening that claim by considering two ways in
which organisms could fail to be distinct in the sense required. First,
even if bacteria were indistinguishable, it could happen that animals
exposed to doses of bacteria usually contract some disease or other.
Given such experiences, one might conclude that bacteria are a general
health hazard (like malnutrition or smoking) and that one should mini-
mize exposure to them as far as possible. However, it would clearly be
impossible to regard bacteria as causes that could, themselves, explain
different particular disease phenomena. Second, suppose bacteria were
only distinguishable in respect to their pathogenic effect. It could then
happen that animals exposed to bacteria from a particular culture
regularly contract one disease while animals exposed to different (but
physically indistinguishable) cultures regularly contract other diseases.
This one difference among bacteria (their pathogenic effect) could not
explain the differences in the diseases they provoke because these differ-
ences would be one and the same, and nothing can explain itself. Thus,
if bacteria are to explain diseases and their differences – that is, to be
their causes – then one must assume that differences in diseases can be
explained by differences in the bacteria that cause them. Bacteria can
qualify as causes of different diseases only if the bacteria themselves are
distinct.
Here is a possible counter-example to the Distinguishability Hypoth-
esis: it is currently impossible to distinguish directly between the
spirochetes that cause syphilis, yaws, pinta, and endemic syphilis.
The pathogenic treponemas cannot be distinguished by morphology,
biochemical capabilities, physiologic criteria, or DNA homology.
Specific antigenic differences have not yet been identified, and sero-
logic reactions positive for one disease are also positive for the other
three. Differentiation of the treponematoses is based on geographi-
cal location, modes of transmission, and clinical manifestations.
(Fitzgerald, 1991, p. 496)
Since the diseases are distinguishable, and since the causal bacteria must
somehow explain these differences, one must assume that the organisms
themselves are also different, and they are given different names. It is
reasonable to assume that they are distinct because other distinguish-
able bacterial diseases are known to be caused by organisms that are
distinguishable in one way or another. Thus, one assumes there are

relevant (although unknown) differences between the causal organisms
of these four diseases and that one’s inability to distinguish between
them is only a temporary and technical inconvenience. As this example
shows, ‘distinguishability’ in the phrase ‘Distinguishability Hypothesis’
means distinguishable in principle not necessarily in fact. If one believed
the four causal organisms were really identical, one would probably be
forced to regard syphilis, yaws, pinta, and endemic syphilis as different
manifestations of the same underlying disease – just as we regard sys-
tematic anthrax and malignant pustule (or what were once called scrofula
and phthisis) as different manifestations of the same disease.
Thus we see that systematic explanations of disease phenomena de-
pend on accepting the Distinguishability Hypothesis. There is another
way of making this point. The purposes of any scientific theory are
explanation and control. In a theory of disease, achieving either pur-
pose requires universal necessary causes, and such causes depend on
etiological characterizations. But etiological characterizations are possi-
ble only if the causes in terms of which different diseases are defined are
themselves distinct. Thus, the formulation of etiological definitions (and
hence the adequacy of any theory of disease) requires distinguishability.
In logical terms, the Bacterial and Distinguishability Hypotheses in-
sure that ‘the cause of disease X’ is a partial function whose domain is
species of bacteria and whose range is all diseases; given the first
Grundversuch, which implies that all bacteria are pathogenic, the func-
tion would be total and one to one.
Through the last five chapters we have traced the beginnings of an

etiological research programme based on the quest for universal and
necessary causes. We are now in a position to appreciate this develop-
ment from a slightly different point of view. As we saw in Chapter 1,
early nineteenth-century physicians distinguished between remote and
proximate causes. So far, our entire discussion has involved only what
would then have been thought of remote causes. As they each became
prominent, acari, Beauvaria bassiana, and decaying organic matter were
all thought of as remote causes, and bacteria would have been classified
in the same way.7 Given this continuity in content, it is natural to think
that, insofar as any aspect of the earlier doctrine of causes has survived,
it was remote causes. In fact, this is false.
Early in the nineteenth century, proximate causes were typically de-
fined as original and internal morbid alterations from which the
symptoms of each disease were thought to follow (Watson, 1858, p.
76). All talk of remote causes, now regarded as a vestige of earlier
medical thinking, was rejected by contemporaries as ‘scholastic and

repulsive’. One problem was that such language implied that the disease
itself (the essence of the disease: the morbid alteration) was its own
(proximate) cause. And this made no sense since nothing can cause
itself (Watson, 1858, p. 76). However, the increased explanatory power
of pathological anatomy was due precisely to the use of morbid altera-
tions – proximate causes – to explain the sequence of symptoms
associated with each disease. This was possible because, given patho-
logical definitions, every case of any one disease had the same one
cause. And everyone recognized that such explanations raised medical
theory to a new level of sophistication.
The introduction of etiological definitions simply made what had
been specific remote causes into new proximate causes: whenever
possible the essence of each disease became infection by a certain
organism (say, tubercle bacilli) instead of a morbid alteration (for
example, the formation of tubers). And the new proximate causes
achieved in spades the same benefits as the old ones. Instead of merely
explaining symptoms in terms of lesions (as the pathologists had
done), researchers were now able to explain the lesions themselves
(and thereby symptoms and everything else lesions could ever explain)
as well as many clinical and epidemiological facts that the patholo-
gists could never begin to account for. This is exactly what we saw in
contrasting Semmelweis’s account of childbed fever with the patho-
logically based accounts of Dubois, Holmes, and Lumpe. The new
proximate causes also provided much more effective targets for therapy
and prophylaxis than did internal lesions.
However, thinking of bacterial infections as proximate causes, brings
into prominence certain similarities between earlier causal thinking and
our own – similarities one might otherwise overlook. First, the ‘scholas-
tic and repulsive’ problem of identifying the essence of a disease with its
own cause has not gone away: it now shows up in the tendency (noted
on p. 103), to define diseases as bacterial infections while continuing to
speak of the bacteria themselves as causing those diseases. Second, the
old proximate causes were universal and necessary in exactly the same
way as the new ones, and their universality and necessity were achieved
in exactly the same way as in the case of the new ones – namely, by
redefining diseases in terms of causes. At the beginning of the nine-
teenth century, whenever possible, diseases (which had earlier been
construed as collections of symptoms) were redefined in terms of spe-
cific lesions; with the rise of the etiological standpoint, whenever possible,
diseases (then construed as lesions) were redefined in terms of bacterial
infections. In both cases, the new definitions created proximate causes
that were universal and necessary.
We no longer use the term ‘proximate’ and we are now interested,

almost exclusively, in universal and necessary causes rather than in
explaining the onset of particular instances of disease (which was the
whole point of nineteenth-century talk about remote causes). This makes
it easy to overlook the fact that, in respect to how causes function in the
logic of disease explanations, the one aspect of earlier etiology that has
vanished is actually remote causes and not proximate causes. Misled by
the superficial similarity in content between earlier remote causes and
the causes we now identify, and oblivious of how remote causes actually
functioned in earlier medical thought, historians get this exactly back-
wards. For example, Christopher Hamlin portrays what he sees as
Thomas Watson’s increased emphasis on ‘single exciting [remote] causes’
as a kind of transition to ‘the kinds of specific agents that we think of as
causes’ today (Hamlin, 1992, pp. 51f.). But there was no transition –
there could be none: one either accepted new etiological definitions or
one did not.8 As with any true revolution in science, the change is all or
nothing.
In spite of obvious shortcomings, the Grundversuche were (and still

remain) elegant, powerful, and amazingly bold. Bassi, Mayrhofer,
Davaine, Obermeier, and others argued that all cases of a few particular
diseases were caused by specific microorganisms; Bassi and Henle specu-
lated that several other diseases may also be caused by germs. So far as
one can tell from the literature, they all regarded such diseases as
exceptional. No one made claims about all diseases. And even their
relatively guarded claims about particular diseases were strenuously
opposed by contemporary physicians. Few researchers would have seri-
ously entertained, even as a working hypothesis, the idea that every
disease was caused by a bacterium.
With the Grundversuche, Klebs adopted as a new norm what every-
one else regarded as exceptions (a few diseases believed to be caused
only by specific bacteria); and what had been the recognized norm
(diseases believed to be attributable to almost any chance trauma)
thereby became new anomalies. The implicit challenge to contemporary
bacteriologists was to assimilate these new anomalies (the old norm) to
the new norm (the old anomalies). It was as though Klebs had seized on
wound infections, anthrax, and relapsing fever, and by pulling on them,
turned the entire world inside out.
Notes
1. Klebs’ papers appeared in volumes 16, 32, 33, 34, and 38 of Archiv für
pathologische Anatomie, Physiologie, und klinische Medizine (Virchows
Archiv).
2. For example, in an 1861 discussion of puerperal fever, Virchow insisted
that ‘just as a man who is overheated will contract facial erysipelas if he
exposes his face to a draft, so too the puerpera, overheated by the process
of delivery, can become ill if her uterus is chilled’ (Monatsschrif für
Geburtskunde … , 1861, p. 380).
3. The term ‘schizomycete’ was coined by the Swiss microscopist Carl Naegeli,
as a generic term including both bacteria and fungi. Ferdinand Cohn
objected to the term as implying a false connection between bacteria and
fungi (Cohn, 1872, p. 191; 1875, p. 201), but Klebs seems to have liked
the term and used it regularly.
4. For an enlightening discussion of the nature and possibility of theories and
laws in the biological sciences see Van der Steen and Kamminga (1991).
5. The roots of this compound plural noun are Grund (ground, basis or
foundation) and Versuch (test, trial or attempt); literally, the word means
something like ‘basic tests’. However, the combination itself is uncommon,
perhaps unique to this context, and Klebs does not explain what he means.
6. For another example, in discussing phenylketonuria (PKU), F. Kräupl Taylor
writes: ‘From such ambiguity, there is only a small step to apparent self-
contradiction. The parents of PKU patients, being heterozygous for the
pathological gene, can have PKU (the clinical sign) after a meal rich in
phenylalanine, though they do not have PKU (the [disease] or the morbus).
Similarly, a PKU patient, when on a diet poor in phenylalanine from his
earliest days, still has PKU (the morbus) without PKU (the illness or the
clinical sign) (Taylor, 1979, p. 84).
7. For example, as early as 1855 Semmelweis’s bitterest enemy, Carl Braun,
speculated about the possible causal role of Henle’s ‘botanical parasites’ in
childbed fever – he even mentioned Beauvaria bassiana (then called Botrytis
bassiana) in muscardine as a possible analogy. Braun classified ‘botanical
parasites’ among 30 kinds of possible remote causes of the disease (Braun,
1855, pp. 477–85).
8. In the same way, glass is either a solid or it is a viscus liquid – there is
nothing in between, and there can be no possibility of gradually moving
from one definition to the other. See p. 53 above.
CHAPTER SEVEN
Proving Disease Causation

Causes are made universal and necessary by adopting suitable disease
characterizations. However, in discussing Klebs’ Grundversuche, we
noted that such characterizations require the prior identification of
sufficient causes – one can only define diseases in terms of causes after
causes have been found. But identifying causes requires a shared theory
including (among other things):
a. a hypothesis – such as the Bacterial Hypothesis – specifying a

domain of entities that can count as causes (Davaine’s inability to
convince opponents was due, in part, to the absence of such a
hypothesis),
b. other assumptions as may be needed to insure that these entities
can explain the diseases in question (disputes such as those between
Davaine and Laplat-and-Jaillard, and between Bassi and Andouin
were due to the absence of these hypotheses), and
c. criteria whose satisfaction establishes that some particular entity
from the domain of possible causes generates the morbidity in
question. Of course Klebs gave considerable attention to identify-
ing such criteria.
Thus definitions to establish universal necessary causes, and the control

and explanatory power that stem from such definitions, require shared
acceptance of a theory of disease with at least these elements.
As we have seen, Pasteur started out inferring causality from the mere
fact that distinguishable and non-spontaneous bacteria were found in
different organic processes. However, in the mid-1870s, when he used
similar reasoning to identify causes of infectious diseases, his contempo-
raries were unconvinced, and he was forced to broaden the evidence he
provided. As we will see in the first part of this chapter, Pasteur came to
rely most heavily on evidence of sufficiency of the sort required by
Klebs’ criteria. Then, over the next few years, Pasteur became involved
in a bitter dispute with Robert Koch. Each claimed to have been the
first to complete Davaine’s proof that bacteridia cause anthrax. The
second part of this chapter concerns this dispute. We will see that
Pasteur and other contemporaries criticized Koch’s arguments for the
same reasons that Pasteur’s own arguments had initially seemed inad-
equate. Koch sought to correct this deficiency by using causal criteria
PROVING DISEASE CAUSATION 111
(his so-called Postulates) that seem to have been based directly on

Klebs’ E1 and E2. In this way, Pasteur and Koch ended up using
essentially the same criteria for establishing causation. Finally, in the
third part of this chapter, we will consider developments that, in the
early 1880s, led to general acceptance of the bacterial theory (including,
of course, the causal criteria that Koch and Pasteur were, by then,
sharing). As explained in the preceding paragraph, general acceptance
of the bacterial theory cleared the way for the systematic characteriza-
tion of diseases in terms of bacteria and for the understanding and
control of bacterial diseases that we associate with modern medicine.
Pasteur’s changing arguments for causation
By 1859 Pasteur had become interested in infectious diseases, and he

believed that many such diseases were caused by microorganisms (Pas-
teur, 1859b, p. 481). Through the 1860s he occasionally discussed
issues relating to human infectious diseases: In 1865, at about the time
he began studying silkworms, Pasteur reported and commented on the
anthrax research conducted by Leplat and Jaillard (Pasteur, 1865c)1;
and three years later he discussed the nature of vaccine (Pasteur, 1868c).
However, Pasteur did not begin a systematic study of human infectious
diseases until the early 1870s. The delay was partly because his atten-
tion had been focused on other projects, but Pasteur himself explained
that he had been reluctant to study human diseases since he had never
been trained in medicine (Pasteur, 1878a, p. 197; Geison, 1974, pp.
384f.).
When Pasteur did turn to human diseases, his work was based on the
same principles that had guided his earlier research.2 And, as one would
expect, his approach to disease causation was similar to that exempli-
fied in his earlier work. In the 1870s, Pasteur seems to have assumed
that each infectious disease – like each disorder of wine or each disease
of silkworms – would have a distinct cause present in every case of the
disease. By contrast, contemporary French physicians continued to think
in terms of sufficient causes that could vary from case to case. As
Pasteur began lecturing on human infectious diseases, he did not dem-
onstrate that the causes he identified were sufficient and he assumed
that they were universal; in both respects he failed to meet contempo-
rary expectations.
In March 1873 Pasteur was elected to the French Academy of Medi-
cine. In the following month, in his first published comments before
that body, Pasteur actively defended some of his causal claims. He
discussed the spoilage of beer – a topic to which he was then giving
attention – and ascribed the commonest form of beer spoilage ‘to the
presence of a microscopic filiform organism’:
This fact is so true, the relation of cause and effect is so evident,
that it is absolutely impossible to produce an alteration of beer
when it does not contain the germs of this inferior organism. On
the other hand, I can demonstrate that no beer, whether from
England, France, or Germany, can be conserved beyond a certain
time; beer always becomes altered because it contains the germs of
the organized ferments of which I speak. The correlation between
the disease and the presence of the organism is certain and indis-
putable.
(Pasteur, 1873b, p. 5)
Members of the Academy objected to Pasteur’s claim that fermentation
could only occur because of a living ferment. Pasteur responded sarcas-
tically that only physicians could take seriously the possibility of
putrefaction without a putrefying agent, he demanded proof that such a
phenomenon had ever been observed, and he immediately associated
spoilage without organic agents with spontaneous generation (Pasteur,
1874b, pp. 6–7; 1874a). He responded to critics by reviewing his argu-
ments for the Dissemination Hypothesis; however, this response was
not quite to the point. One could maintain, as his critics apparently did,
that spoilage could result from a suitable combination of inert factors
without regarding those factors as sufficient to generate living organ-
isms. For some time Pasteur and his critics argued at cross purposes:
Pasteur recounting his contributions to science and denouncing sponta-
neous generation; his opponents insisting that putrefaction could occur
without living organisms (Pasteur, 1922–39, vol. 6, pp. 90–99, 20–26,
26–8, 28–37, 37–58 and vol. 5, pp. 344f.). The debate ceased in 1879
when other members of the Academy finally pointed out that most
physicians had given up spontaneous generation even before Pasteur’s
experiments, and that he could not answer his critics by this evasion
(Pasteur, 1879b, pp. 235–7). In these discussions Pasteur effectively
assumed that the cause of spoilage must be universal; his opponents did
not share this assumption.
In 1874 and 1875, Pasteur gave more lectures on putrefaction, and he
began reporting studies on ammoniacal urine. He continued making
assertions exactly in line with his earlier research and he continued to
encounter opposition. Pasteur’s opponents rejected his claim that an
organism was ‘the sine qua non of ammoniacal urine’ (Pasteur, 1875b,
p. 78) and insisted the disorder could occur without a living ferment.
Perhaps because of these discussions, in the second chapter of his
1876 book on the spoilage of beer, Pasteur gave increased attention to
establishing causality. Pasteur insisted that ‘all the alterations to which
wort and beer are subject have as their exclusive cause the development
of organized ferments whose germs are carried by dust’ (Pasteur, 1876a,

p. 21). Pasteur’s argument for this claim fell into two parts, the first of
which begins with the assertion that: ‘all undesirable alterations of beer
coincide with the development of microscopic organisms that are alien
to yeast properly called’ (Pasteur, 1876a, p. 22). Because the phrase
‘coincide with’ is vague, it is unclear exactly what Pasteur intended. He
could have meant that whenever there are alterations there are organ-
isms (organisms are necessary), or that whenever there are organisms
there are alterations (organisms are sufficient), or both. His ensuing
discussion does not clarify matters. Pasteur recounted having heated
some bottles of beer while leaving others unheated. He found the un-
heated bottles spoiled while the heated ones did not. Moreover, the
spoiled beer contained alien organisms.
The second part of the causal argument begins with the assertion that
‘the absence of alterations in wort or beer coincides with the absence of
the alien organisms’ (Pasteur, 1876a, p. 26). This claim is ambiguous in
exactly the same way as the earlier one, and again, the discussion does
not remove the obscurity. In this section, Pasteur considered the possi-
bility that germs were the effect rather than the cause of spoilage. He
tried to dispose of this possibility by showing that, regardless of tem-
perature, beer never spoiled when exposed to pure air, and he used
experiments similar to those he had used against spontaneous genera-
tion. Pasteur summarized the chapter as follows:
the absence of microscopic organisms alien to the yeast corre-
sponds invariably to a beer that remains healthy indefinitely, even
in contact with pure air; … on the other hand, the presence of
these organisms corresponds always to a disordered beer … It
would be difficult to go further in proving a correlation between
the organisms and the alterations of beer … No proof can more
decisively establish a relation of cause and effect in the succession
of physical phenomena.
(Pasteur, 1876a, p. 31)
For all the ambiguity of the preceding arguments, two facts are clear:
first, in comparison with his earlier studies on wine spoilage, Pasteur
now recognized the need for explicit arguments and experiments to
demonstrate causality; he no longer inferred causation simply from
having regularly observed organisms in altered substances. Second, he
was now using a concept of causation that included both necessity and
sufficiency. While it may not be clear exactly how each of the halves of
his argument is to be understood, several times Pasteur claims that beer
becomes disordered if and only if it contains organisms. Thus, as a
whole, Pasteur saw his argument as supporting the claim that organ-
isms are both necessary and sufficient for spoilage.
In the summer of 1876 – just months after publication of his book on

beer – Pasteur suddenly adopted a new strategy for demonstrating
causality, one which dominated most of his subsequent writings on
infectious diseases. This strategy first appeared in an essay on ammo-
niacal urine that began with the observation that outside the bladder,
urine can only be converted into ammonia through the influence of
organized ferments. ‘It remains to be seen,’ Pasteur continued, ‘whether
things are not similar in the bladder’ (Pasteur, 1876b, p. 82). Pasteur
then described the following strategy for proving causation:
When one seeds the pure organized ferment in question in a nutri-
tive liquid, … the ferment multiplies. One filters and precipitates it
with alcohol. The collected precipitate contains the soluble ferment
… ready to transform a solution of urine and water into carbonate
of ammonia. The conditions of this experiment permit one to
demonstrate that urine is not essential for production of the organ-
ized ferment, and that the ferment can form in a medium completely
different from urine. … It is difficult to go further in an experimen-
tal proof of the facts we have announced.
(Pasteur, 1876b, pp. 85f.)
The method that Pasteur here outlined is a way of showing sufficiency:
one isolates the ferment by growing it in an artificial medium and then
introduces it into a new solution where it produces ammonia.
In the course in his early papers, Pasteur occasionally reported seeding
experiments: he introduced dust containing organized corpuscles into
culture media or into healthy silkworms. But these earlier experiments
were different in two respects: they involved no attempt to insure that
only pure strains were used and Pasteur did not use them to justify
causal claims. Only when one wishes to establish that an organism itself
causes a disease – that the organism is sufficient – must one isolate it
from contaminating influences in a pure culture. Pasteur’s 1876 paper
on ammoniacal urine contains his first explicit description of isolation
and inoculation, and his first clear use of these techniques to prove
causal sufficiency. This contrasts sharply with his earlier seeding experi-
ments and also with the confused and ambiguous causal arguments in
his book on beer – a context in which such experiments would have
been exactly appropriate.
A few months later, in 1878, in discussing diseases of silkworms,
Pasteur observed that ‘when flacherie occurs, one finds organisms in the
digestive track of the immense majority of sick worms’ (Pasteur, 1878b,
p. 692).3 This is evidence of necessity. Next Pasteur described the same
experiment he had recently used in studying ammoniacal urine: ‘Con-
versely when one makes leaves ferment in the intestinal canal by
introducing germs or adult organisms, one induces flacherie.’ Pasteur
observed that ‘two phenomena are in a relation of cause and effect if
when one of the two exists the other follows. This is the case here; if the
ferments are present in the intestinal canal, there is flacherie, and if
there is flacherie there are ferments in the intestinal canal’ (Pasteur,
1878b, p. 692). These remarks show clearly that Pasteur sought to
establish necessity and sufficiency, that he understood the difference,
and that he recognized his inoculation test as a means of demonstrating
the latter. Pasteur’s concept of causality and his account of how to
prove necessity and sufficiency are more explicit here than in any of his
earlier publications on flacherie.
The Pasteur–Koch dispute regarding anthrax
Through the late 1870s anthrax was studied extensively by several

researchers including Louis Pasteur and Robert Koch. All of these studies
were based on Davaine’s research and, in turn, the work by Pasteur and
Koch provided models for those seeking to trace other diseases to para-
sitic organisms. Koch claimed to have been the first to prove causality,
and he is sometimes supposed to have done this in his 1876 anthrax
paper using the criteria later known as Koch’s Postulates (Susser, 1973, p.
23). However, Pasteur denied that Koch ever proved causality and claimed
that he himself had been the first to do so in a paper published in 1877.
These contradictory claims initiated an acrimonious dispute that quickly
took on nationalistic overtones and was never clearly resolved (Geison,
1974, pp. 397–400; Dolman, 1974, p. 424; Mollaret, 1983). Only an-
thrax was a major research interest for both Koch and Pasteur and this
was the only area of investigation in which they came into direct competi-
tion. Our interest in this dispute focuses on the arguments Koch and
Pasteur used to show bacteridia caused anthrax.
The circumstances of the publication of Koch’s 1876 anthrax paper
have often been recounted (Brock, 1988, pp. 38–53); it is sufficient to
note that Koch’s research was carried out under primitive conditions
and in virtual isolation from others who were studying the disease. Koch
introduced and exploited technical innovations including the suspended-
drop method for studying cultures. Using this method Koch was able to
trace the complete life cycle of the bacteridium which he called Bacillus
anthracis. He showed the bacillus formed spores that remained viable
for long periods even in hostile environments. The discovery of spores
removed some of the obstacles to recognizing bacilli as the cause of
anthrax; in particular, it explained Davaine’s observation that quickly
dried anthrax blood could remain infectious for long periods.
Koch’s conclusions were based almost exclusively on artificially in-
duced anthrax in test animals. He claimed he had often examined
animals that died of natural anthrax (Koch, 1876, p. 1), but of those
examinations, he reported only that he found bacilli in the spleen of one
anthrax horse – the only horse he had examined (Koch, 1876, p. 13).
Koch cited earlier researchers who identified bacilli in natural cases, but
he also mentioned that other researchers had been unable to find them.
As Koch explained in the introduction to his paper, his study grew out
of Davaine’s research that was widely accepted as presenting the strong-
est evidence that anthrax was caused by bacteria. However, Koch did
not have access to Davaine’s papers (Koch, 1878a, p. 51); he knew
Davaine’s work only from abstracts in which Davaine’s arguments were
described as supporting the necessity of the anthrax organism (Steudener,
1872). In harmony with the abstracts, Koch noted that Davaine ‘as-
serted that the rods were bacteria and that [anthrax] could occur only
when these rods from anthrax blood were present’ (Koch, 1876, p. 1).
One objection to Davaine’s view was that anthrax sometimes seemed to
have been caused by inoculations with bacteridia-free blood. Koch
answered this objection by arguing that the bacilli were often difficult
to find and could easily be overlooked.
However, explaining away apparent counter evidence did not consti-
tute a proof that the bacilli caused anthrax and, in spite of what Koch
himself later claimed, his paper contained remarkably little direct evi-
dence of causation. Moreover, the evidence it did contain was not
persuasive to contemporaries. Koch’s argument for causation appears
toward the end of his paper. He observed that anthrax and septicemia
bacilli were distinct because the latter never caused anthrax (Koch,
1876, p. 11) – an argument that presupposes the Distinguishability
Hypothesis. He mentioned he had been unable to cause anthrax by
injecting hay infusion bacilli or a species of bacillus resembling the
anthrax bacillus that developed accidentally in one of his cultures. On
the other hand, he reported that animals inoculated with spore masses
‘derived from entirely pure cultures of Bacillus anthracis … invariably
died of anthrax’ (Koch, 1876, p. 11). However, at this point Koch’s
inoculation cultures were grown in liquid media, as Davaine’s had been,
and his only evidence that the cultures were pure was that microscopic
examination revealed no contamination. For more than a decade, Davaine
had tried, unsuccessfully, to persuade critics with arguments at least as
strong as these. Yet Koch insisted his argument was conclusive:
‘Anthrax substances, whether they are fresh, decayed, or dried, can only
cause anthrax if they contain Bacillus anthracis or its viable spores.
This fact removes all doubt that Bacillus anthracis is the actual cause
and contagium of anthrax’ (Koch, 1876, p. 13). In an 1881 paper, Koch
claimed he had established causality by this argument. In that later
paper, after reviewing the earlier argument and without any additional
evidence, he concluded ‘thus, anthrax never occurred without viable

anthrax bacilli or spores. In my opinion, no more conclusive proof can
be given that anthrax bacilli are the true and only cause of anthrax’
(Koch, 1881b, p. 64). However, this argument failed on two counts:
first, it was not clear that Koch had truly isolated the bacillus; and
second, while the argument provided evidence that the absence of the
bacilli was followed by the absence of the disease (necessity), his audi-
ence expected inoculation experiments showing that the presence of
isolated bacilli insured the presence of the disease (sufficiency). In spite
of his brilliant technical innovations, which everyone acknowledged to
be of major value, critics were unpersuaded (Bert, 1876). Koch’s argu-
ment was judged to have failed for precisely the same reason that
Pasteur’s earlier arguments were inconclusive – both presented evidence
of necessity and both assumed universality. In spite of his impressive
technical innovations, there is no documentary evidence that, at the
time, anyone found Koch’s causal argument conclusive.
A few months later, in the spring of 1877, Pasteur published his first
major paper on anthrax (Pasteur, 1877a). After mentioning that Davaine
had preceded Pollender in reporting anthrax bacteridia, Pasteur claimed
that he himself, while studying diseases of silkworms, had been the first
to identify spores and to recognize that they remained viable for long
periods.4 Pasteur mentioned Koch’s paper favorably and acknowledged
that Koch had been the first to trace the life cycle of the bacteridium
and to describe the formation of anthrax spores. However, he pointed
out that Koch’s work had not persuaded the skeptics, and stated that
his own goal was to provide a conclusive proof. As Pasteur explained it,
the main problem was showing that, of all the components of anthrax
blood, the bacteridia themselves were the causal agent. Pasteur ob-
served that the body of a healthy animal is closed to all bacteria –
having reached this important conclusion, which is equivalent to Klebs’
first Grundversuch, in a paper published several years earlier.5 Thus the
bodies of healthy animals provide sterile media in which one can pro-
duce successive cultures of anthrax bacteria. He observed that, if no
other organisms show themselves, one could be sure the original culture
was pure. He also produced subcultures in inorganic media and filtered
the remains to remove impurities. The isolated and filtered bacteridia
consistently caused anthrax. Toward the end of his paper, Pasteur con-
sidered some possible alternative interpretations of his results: the true
cause of anthrax could be (1) virulent matter adhering to bacteridia
even after isolation and filtration, (2) a substance secreted by the organ-
isms rather than the bacteridia themselves, or (3) some organism other
than bacteridia that also multiplied in culture media (Pasteur, 1863b, p.
170). Pasteur felt the first alternative could be eliminated by using
numerous successive subcultures to eliminate all extraneous material

from the diseased animal. He discounted the second alternative on the
grounds that, if it were true, the bacteridia would still be necessary for
the onset of disease. He tried to counter the third alternative by using a
medium (urine) in which other organisms could easily be detected. Of
course, ultimately, this argument rests on the Bacterial Hypothesis – the
hypothesis that only bacteria are possible causes of disease.
In a second 1877 anthrax paper, Pasteur claimed that ‘the experi-
ments I have reported … have demonstrated without reply that there is
a microscopic organism, the unique cause of … anthrax; it is the
bacteridium, first perceived by Davaine in 1850’ (Pasteur, 1877b, p.
172). He claimed Davaine had come closest to proving causality but
had been unable to provide a conclusive argument. The issue could only
be resolved by passing bacteridia through numerous subcultures and by
filtering the remains to remove all other materials. Pasteur claimed to
have been the first to have done this. This argument for causality
obviously conformed more closely to contemporary expectations than
did Koch’s and some critics found it conclusive.6
Pasteur continued to study anthrax: seven more publications appeared

in 1878, six in 1879, and five in 1880 (V.R. Pasteur, 1922–39, vol. 6,
pp. 197–270, 452–512). Most of these reported new results. Pasteur
found one could induce anthrax in chickens by artificially lowering
their body temperatures (Pasteur, 1878c), and that spores from buried
cadavers could be carried to the surface by earthworms where they
could infect grazing animals (Pasteur, 1880b). Although Pasteur be-
lieved he had conclusively proven causality in 1877, he continued to
refine his argument. In one paper he claimed to have passed bacteridia
through more than 100 subcultures to insure they had been completely
isolated from all extraneous disease products (Pasteur, 1878a, pp. 197f.).
In 1878 Pasteur observed that his research on anthrax ‘left the etiology
of the putrid diseases or septicemia less advanced than the etiology of
anthrax’ (Pasteur, 1878d, p. 112). He described his techniques for
isolating bacteridia, and observed that ‘in the current state of science’
this procedure is the only way to prove ‘the organism is the real agent of
disease and of contagion’ (Pasteur, 1878d, pp. 112f.). He proposed
undertaking similar studies of septicemia. A few months later, in report-
ing research on plague, Pasteur again outlined the strategy of isolation,
purification, and inoculation, and observed that, in the current state of
science, the only conclusive proof of causality was passing an organism
through successive animals and inert media and showing that isolated
organisms could produce disease and death (Pasteur, 1879c, p. 494). He
also claimed to have used this method to study a disease called chicken
cholera (Pasteur, 1879c, p. 495), and he discussed the applicability of
the method to puerperal fever (Pasteur, 1879d, pp. 498f). On two
occasions during these months, Pasteur observed: ‘everything is hidden,
obscure, and open to discussion when one ignores the cause of some
phenomena, but everything is clear when one possesses it’ (Pasteur,
1922–39, vol. 6, pp. 110, 114).
In contrast to Pasteur, after his initial paper, Koch seems not to have
continued studying anthrax. Apart from incidental references, between
1876 and 1880 Koch discussed anthrax in only one paper – his long
1878 work on wound infections (Koch, 1878a, pp. 46–48).7 The only
new result Koch announced was his use of methyl violet as a stain.
Having observed that anthrax was not recidivous, in 1880 Pasteur
began looking for a vaccine (Pasteur, 1880c). This observation, together
with similar results for other diseases, suggested what Pasteur called
‘the general law of the non-recidiviousness of virulent diseases’ (Pasteur,
1881a, p. 339). Early in 1881, Pasteur reported he could attenuate the
virulence of the bacilli of chicken cholera; this led to his important
discovery that one could also attenuate anthrax bacilli (Pasteur, 1881b).8
Pasteur began using a vaccine in tests on thousands of livestock in
various European countries. His results, along with other findings, were
announced in a series of papers. Like many of Pasteur’s earlier studies,
his anthrax inoculations promised profound economic benefits and
were widely publicized.
In 1881, Koch published a detailed criticism of Pasteur’s work (Koch,
1881b) – this was only Koch’s second major paper on anthrax. Accord-
ing to Koch, Pasteur’s work had contributed nothing to the etiology of
the disease. ‘Only a few of Pasteur’s beliefs about anthrax are new, and
they are false’ (Koc, 1881b, 65).9 He accused Pasteur of confusing
anthrax bacilli with unrelated contaminants (p. 62) and wrote that
Pasteur had probably never observed anthrax uncomplicated by other
diseases (p. 60). Koch rejected some of Pasteur’s experiments as worth-
less and naive (p. 70) and ridiculed his work with chickens and
earthworms. Koch also reviewed his own argument for causality and
asserted that ‘no more conclusive proof can be given that anthrax
bacilli are the true and only cause of anthrax’ (p. 64). In respect to the
demand to isolate bacilli from every contaminant – the condition that
most observers then expected in any proof of causality – Koch wrote:
‘this is impossible … no one can take seriously such an undertaking’.
Koch acknowledged that, even in the face of his experiments, ‘one
could still object that … some other substance associated with the
bacteria, rather than the bacilli themselves, causes the disease’. But, he
continued, ‘for all practical purposes, this objection is meaningless’
(p. 64). However, critics did not find the objection meaningless. In
reviewing Koch’s paper, a Professor Zuber observed: ‘how opinions can
differ! We think, and much of the world with us, that this objection has
a capital importance and we reproach precisely those authors who …
subject complex liquids to complicated procedures and thereby reach
dubious results’. Zuber praised those who tried to isolate the bacillus by
subcultures (Zuber, 1882, pp. 104f).
In September 1881 Koch and Pasteur met in an international medical
congress in London. Pasteur had not yet seen Koch’s recent scathing
criticism and veiled insults, and their relations were cordial (Geison,
1974, p. 397). Pasteur praised some of Koch’s new techniques. In a
plenary session, Pasteur summarized the results of his inoculation pro-
gramme (Pasteur, 1881c).
In March 1882, in the midst of his debate with Pasteur, Koch an-
nounced discovery of the tuberculosis bacillus (Koch, 1882a) – the
announcement created an immediate sensation throughout Europe. Af-
ter reporting he had identified a specific organism in hundreds of natural
and artificial cases of tuberculosis, Koch turned to the question of
causality. He began by explaining that in order to prove causality,
it is necessary to isolate the bacilli from the body, to grow them in
pure culture until they are freed from every disease product of the
animal organism, and, by introducing isolated bacilli into animals,
to reproduce the same morbid condition known to follow from
inoculation with spontaneously developed tuberculosis materials’.
(Koch, 1882a, p. 87)
Koch’s paper was organized around attempts to meet these conditions.
His proof of causality by isolation and inoculation is remarkable for
two reasons. First, it is his earliest use of the causal criteria later known
as Koch’s Postulates. Second, by claiming that this strategy was essen-
tial for proving causality – a strategy he had rejected as impossible just
a few months earlier – Koch was effectively admitting the inadequacy of
his own attempt to prove causality for anthrax. Moreover, his claim
implies that Pasteur’s demonstration, although conceivably inexact or
even unoriginal, at least followed the general procedure that Koch now
described as essential for proving causality. Of course, Koch never
explicitly admitted either that endorsing the new strategy undercut his
earlier claims or that Pasteur’s work had been more relevant to proving
causality than his own, yet both conclusions follow from claims he
made in the tuberculosis paper and in several papers that followed.
Koch’s adoption of isolation and inoculation probably resulted from a
combination of factors: his dispute with Pasteur; discussions with his
friend, Edwin Klebs, during the 1878 meetings of the German Natural
Scientists and Physicians; and his introduction of solid culture media
that enabled him to isolate bacteria much more reliably than had been
possible using only liquid media.10
Pasteur answered Koch’s criticisms in a lecture at a medical congress
in Geneva in the fall of 1882 (Pasteur, 1882). Pasteur pointed out that
critics had rejected Koch’s initial work on anthrax as inconclusive. He
explained that he himself had followed the only method for proving
causation that satisfied existing standards, and he mentioned Paul Bert
who had rejected Koch’s argument but had been persuaded by his own.
He also answered some of Koch’s specific objections. Given the tone of
Koch’s attack, Pasteur’s response was remarkably restrained. Yet Koch,
who attended Pasteur’s lecture, took offense and registered a formal
protest (Schwalbe, 1912, vol. 1, pp. 207f.; Ellenberger, 1970, p. 269).
Koch responded quickly with a long and polemical attack on Pasteur’s
inoculation programme (Koch, 1882b). Koch insinuated that Pasteur was
falsifying results (Koch, 1882b, pp. 106, 111), and he misrepresented
Pasteur’s work. Near the beginning of his paper he explained how his
methods differed from Pasteur’s. Koch claimed he first became oriented
by undertaking a thorough investigation of all the diseased tissues of the
body to determine the distribution of the organism.11 Koch then claimed
to use isolation and inoculation, which he described as ‘the only method
that meets current scientific standards’ (Koch, 1882b, p. 98). As usual, he
did not mention that Pasteur had preceded him in using this strategy or
that his initial papers on anthrax, in which he still claimed to have proved
causality, were defective for not using it.
This time Pasteur responded to Koch in an emotional open letter
(Pasteur, 1883). Pasteur expressed surprise at the virulence of Koch’s
attack. He reviewed his contributions to medicine and to science gener-
ally and mentioned that several prominent researchers, including both
Davaine and Joseph Lister, had based their studies on his own early
work. Pasteur summarized earlier research on anthrax, reiterated his
claim to have been the first to have identified spores (while admitting
Koch had first seen anthrax spores), and reviewed his argument for
causality. He again cited researchers who had rejected Koch’s causal
argument but had been persuaded by his own.
One year later, in 1884, Koch published another long criticism of
Pasteur’s attempt to attenuate anthrax bacilli (Koch, 1884a). In 1887 he
again criticized Pasteur’s inoculation program (Koch, 1887). However,
by this time, Pasteur’s programme had achieved significant success and
had been dramatically corroborated by his work on rabies; he re-
sponded to Koch’s criticism with a brief note (Pasteur, 1887). These
were the last publications by either Koch or Pasteur on anthrax.
One can appraise the controversy between Koch and Pasteur in dif-
ferent ways. On a personal level, Koch comes off poorly: his arguments
were sarcastic, polemical, unfair, and often included personal attacks.
While Koch regularly minimized the significance of Pasteur’s work and
misrepresented his opinions, Pasteur’s appraisals of Koch were fair and
consistent with opinions expressed by others.
In respect to their arguments for causality, Pasteur’s work obviously
came closer than Koch’s to satisfying existing expectations. Since most
contemporary researchers expected evidence of sufficiency, Koch’s argu-
ments were not seen as compelling, in spite of his technical achievements.
He seems to have tried to obscure the perceived inadequacy of his argu-
ment by ad hominem attacks on Pasteur. From what we have seen, there
is a certain irony in using the phrase ‘Koch’s Postulates’ to refer to the
causal strategy that Koch ultimately adopted and that continues with us
today. If one takes the strategy to focus on isolation and inoculation, both
Klebs and Pasteur used it before Koch and both applied it to a wider range
of diseases than Koch ever did. Moreover, Koch’s early work was deemed
inconclusive precisely because he was slow to adopt this approach.
In any case, by the mid-1880s, in France and in Germany, a consen-
sus had emerged as to how one must go about proving that specific
bacteria cause particular diseases. From this time on, everyone accepted
isolation and inoculation as crucial to proofs of causality.
General acceptance of the bacterial theory of disease
In proving that some factor, say, bacteridia, causes some pathological

process, one requires consensus as to what sorts of things must be taken
into account as possible causes (that is, one requires something like the
Bacterial Hypothesis) and consensus as to how one proves that some
particular factor of the appropriate sort causes the morbidity in ques-
tion (that is, one requires causal criteria). We have seen how a consensus
emerged with respect to causal criteria. We now turn to what was, in
some degree, a more basic problem: establishing a consensus among
medical researchers with respect to the fact that bacteria were the sorts
of things that could cause disease in the first place. Of course, Klebs,
Pasteur, Koch, and other contributors to the research programme were
already committed to the pathologicality of bacteria – the problem was
converting skeptics.12 To see how this was achieved, we must review
another dramatic development that emerged from Pasteur’s laborato-
ries: discovery of a vaccine against rabies.
Rabies has never been a common disease, but it is so terrible and,
once symptoms begin, death so certain that nineteenth-century physi-
cians were willing to attempt almost anything by way of treatment.

During the course of the century, Lancet reported the use of more than
300 assorted drugs and chemicals to treat victims; these included am-
monia (as an ointment, orally, and injected into the blood stream ‘until
the patient begged that treatment might be discontinued’), antimony,
arsenic, fresh blood (orally), Spanish fly, carbolic acid (orally and by
injection), chlorine (applied to the wound and taken orally), chloroform
(inhaled, rubbed on the skin, and orally), creosote, curare, ether (in-
haled or injected), hydrochloric acid, iodine, lead, mercury, strychnine,
prussic acid, rhubarb, steel, sulphurous acid, nicotine, turpentine (orally
or by enema), viper’s poison, vinegar, and zinc.13
Conventional wisdom dictated that a bite by a rabid animal should
be cauterized as quickly as possible. Usually this was done by applying
a strong acid such as nitric acid or sulfuric acid, but some physicians
recommended physical cauterization by boiling oil or by a red-hot iron.
Once the wound was attended to, various methods of treatment were
employed. In England, sea dipping was common, and in some areas
fishermen advertised their expertness in dipping – dippers used long
poles to hold patients underwater for three or more minutes at a stretch.
In some respects, strychnine poisoning resembled hydrophobia; Marshall
Hall, a respected British physician, observed that frogs injected with
strychnine died if they were agitated, but recovered if left in peace. He
recommended that rabies patients be placed on spring beds and sur-
rounded by gauze curtains in a dark and quiet room. By contrast, one
patient seemed to have been greatly benefited by escaping from a hospi-
tal and running around the town; this suggested exercise and the use of
tonics. Another physician recommended excision of the uvula as an
excellent shock to the system.
Some writers attributed rabies to inadequate sexual release or, by
contrast, to sexual incontinence in dogs. This led one physician to
recommend castration as a radical but effective treatment. Other physi-
cians advocated hot air baths with temperatures as high as 200°F, or
vapor baths in which patients were wrapped in blankets or flannel and
placed on a wicker chair over hot bricks, live coals, or a spirit lamp.
Others advocated cold effusions, large doses of ice, or the application of
ice along the whole length of the spine. Some combined warm baths
with immersion in cold water. ‘To reduce to a lesser or negative state’
the great excitement under which the victim labors, one physician rec-
ommended wrapping the patient in blankets and dashing the spine with
cold water every 15 minutes. Another physician reported that electrocu-
tion cured patients of hydrophobia but that they occasionally died of
exhaustion in the process. Yet another recommended that patients be
fed massive doses of asparagus; he reported trying this on a patient who
went mad and died, but the patient seemed to have been cured of
hydrophobia. Some writers noticed that rabies is spontaneous only in
animals that do not perspire; to them, vigorous perspiration seemed to
offer hope. Others wrote that in animals that do not perspire, saliva
must carry off large quantities of effete matter; they advocated vigorous
salivation. One physician recommended that patients be poisoned by
curare – the South American arrow poison – and then be revived by
artificial respiration. For a time it was believed that lead, mercury, and
turpentine would cure rabies, but that victims were sometimes poisoned
in the process. One case was reported in which massive doses of croton
oil and prussic acid had apparently converted hydrophobia into a fatal
case of typhoid fever. Some physicians recommended that patients be
kept in a constant state of nausea or given small but continuous
enemata. One physician tried soaking a rennet in water, saturating it
with savanilla and forcing it down the patient’s throat; after being
placed in the sun and sleeping for 48 hours the patient awoke com-
pletely recovered.
For several reasons nineteenth-century research on rabies was par-
ticularly frustrating. First, it is difficult to convey rabies reliably among
test animals. Even deep bites by rabid animals produce new cases of
rabies only infrequently. Second, when the disease is conveyed, there
may be an incubation period of several weeks or even months (Pasteur,
1881d, p. 574). Finally, while many physicians were fairly confident
that rabies was caused by a microorganism, the best efforts of the
leading experts failed to identify it. For several years Pasteur searched
for rabies microbes – ‘On two occasions he suspected he had seen them
– in 1881 in the form of encapsulated bacteria and in 1883 as fine
particles in the brain tissue of rabid animals’ (Hughes, 1977, p. 32).
From a modern point of view, the failures were due, in part, to the fact
that the causal agent is a virus and, therefore, too small to be detected
given the equipment and methods then in use. However, Pasteur was
not deterred by his inability to identify a causal organism.
Pasteur tried injecting saliva from rabid dogs directly into the brains
of live rabbits. He found that brain fluids taken from these rabbits were
maximally virulent. Inoculating these fluids into healthy dogs usually
produced rabies within days. Although he could not identify a causal
organism, Pasteur may have wondered if the virulence of the supposed
organism could be reduced in the same manner he had attenuated the
bacteria of anthrax and chicken cholera. Developing results that had
been announced by other researchers, Pasteur confirmed that the viru-
lence of the rabies virus for animals of one species could be affected (in
some cases increased, in others decreased) by passage through series of
animals of a different species. Early on, Pasteur also concluded that the
elusive virus was most likely to be found in brain tissue rather than in
blood or saliva. These insights led to experiments involving spinal cords
from rabbits that had been dried in sterile air. He found that the longer
the spinal cords were allowed to dry, the longer the incubation period
for inoculations using matter taken from those cords. Pasteur inferred
that the causal agents contained in the dried cords had become less
virulent (Pasteur, 1884, pp. 593f.). Beginning with injections of tissues
that had dried for two weeks and were nearly inert, he injected dogs
with matter that was progressively more virulent. He hypothesized that
at some stage test animals would be immune. This, he believed, could
be shown if injecting full-strength brain tissue into immunized dogs had
no ill effects (Pasteur, 1885, p. 603).14
While still engaged in this research and before he had conclusive
evidence his methods would work, Pasteur was urged to try his proce-
dure on a human being. On 4 July 1885, a nine-year-old Alsatian boy,
Joseph Meister, had been badly bitten by a rabid dog, and the boy had
been brought to Pasteur.15 Pasteur knew that only a small percentage of
similarly bitten humans would contract rabies. He also acknowledged
that his experiments were incomplete and that he had never tried inocu-
lations on test animals after they had already been bitten (Pasteur,
1885, p. 606). Yet Pasteur took a chance that could be described as
irresponsible: on 6 July he began inoculating Meister with tissues from
the dried spinal cords of diseased rabbits. On 16 July 1885, having
given Meister 13 shots in 10 days, Pasteur used fluids containing fresh
and fully virulent rabies material. When this produced no evidence of
the disease, Pasteur pronounced the boy cured. Given that inoculations
did not always produce rabies and that the disease had an uncertain
incubation period, the announcement was premature. However, Joseph
Meister remained healthy, and Pasteur’s work was proclaimed around
the world as a brilliant success.
Within months thousands of exposed persons began flocking to Pas-
teur’s laboratory from all over the world. Between 1886 and 1934,
34,111 persons who had been bitten by confirmed rabid animals were
treated at the Pasteur Institute. Of these thousands, only three persons
died; among untreated victims, the average mortality was 16 per cent
(Malkin, 1986, p. 45). Thus, even though the cause of rabies had not
been identified (and even though it turned out, ultimately, not to be
bacterial), since the inoculations clearly emerged from what Pasteur
called the theory of germs, the success of the inoculations was taken as
further support for that theory.
In their work on anthrax, both Koch and Pasteur accumulated empiri-

cal evidence of causality and their evidence was of different kinds.
However, it is impossible to prove causation by empirical evidence
alone. This is reflected in the standard objection – which continued to
be raised throughout the century – that, no matter how carefully organ-
isms were isolated, it was always possible that an unknown virus or
some totally different entity in the blood was the true cause of the
disease. Proving bacteria cause a particular disease, such as anthrax,
was impossible in the absence of a theory within which bacteria were
recognized as possible causes – one needed an assumption like the
Bacterial Hypothesis.
However, it would have been at least as difficult to adopt the bacte-
rial theory as to accept Beauvaria bassiana as the cause of muscardine
or decaying organic matter as the cause of childbed fever. It was not just
a matter of accepting a new kind of cause for a particular disease; the
bacterial theory was a whole new way of thinking about diseases in
general – it meant exchanging what were the norms and what were the
anomalies that required explanations. There is inevitable resistance to
such a revolution. As in all such cases, innovators were confronted by a
vicious circle: observations (such as those made by Davaine, Pasteur,
and Koch) could count as evidence that bacteridia cause anthrax and,
therefore, as evidence for the bacterial theory in general, but only if one
accepted something like the Bacterial Hypothesis in the first place, and
this, of course, was itself a central part of the bacterial theory. Thus,
any attempt to accumulate direct evidence for the bacterial theory
depended on accepting the theory to begin with – a circular argument.
In the absence of direct evidence, researchers fell back on analogical
reasoning involving diseases known to be caused by minute but visible
parasites like acari and trichinae. Yet, although these analogies had
some heuristic value, they were never entirely persuasive. So how were
skeptics to be won over?
Where rational persuasion fails, the only alternative is conversion,
and converting opponents to the bacterial theory required successes
that were ‘supernatural’ in the sense of being outside what could
reasonably have been expected within the framework of traditional
medicine. To achieve conversion, bacteriologists required, not mere
evidence, but something akin to miracles.16
According to Pasteur, the theory of germs became generally accepted
in the early 1880s, at about the time of the London International
Medical Congress (Pasteur, 1881c, p. 370). What miraculous successes
at about that time would have made the theory of germs dramatically
more plausible than it had previously been? Among the sensational
accomplishments of this period were Pasteur’s anthrax and rabies in-
oculations and Koch’s discovery of tuberculosis and cholera bacilli. The

cumulative effect of such developments was a sea change in medicine –
a fundamental reorientation in how physicians thought about disease.
Given the new standpoint, which encompassed the Bacterial Hypoth-
esis, empirical evidence about the presence or absence of bacteria in
disease processes became evidence for the theoretical relation of disease
causation.
In respect of winning support for the bacterial theory, however differ-
ent and even opposed their work may appear, Koch and Pasteur were
mutually supportive. Since proving that bacteridia caused anthrax was
possible only within the bacterial theory, if one asks which of Koch or
Pasteur ultimately proved that the anthrax bacillus causes anthrax, the
best answer may be: both together.
Notes
1. This was the paper we considered in Chapter 5 in which Leplat and

Jaillard attacked Davaine.
2. Pasteur himself pointed this out in the preface to Études sur la bière
(Pasteur, 1876a, p. 5; cp. the diary entry reprinted in Pasteur, 1922–35,
vol. 6, pp. vf.).
3. In the next year Pasteur strengthened his claim: ‘It has been impossible
for me to discover a worm struck by flacherie, advanced or beginning,
without establishing immediately under my own eyes or those of my
collaborators, that the intestinal matter contained one or several of the
microscopic organisms that one finds in fermenting mulberry leaves’ (Pas-
teur, 1878e, p. 702).
4. Pasteur cited his Études sur la maladie des vers à soie (Pasteur, 1870, pp.
231f.).
5. Pasteur frequently referred to his argument for this principle which first
appeared in 1863 (Pasteur, 1863b).
6. For example, Paul Bert, who had rejected Koch’s argument, was per-
suaded by Pasteur’s (Bert, 1877).
7. Koch’s collected works also contains a letter on anthrax written to the
Secretary of the Interior (Schwalbe, 1912, vol. 2, pp. 831f.).
8. Geison’s account of events surrounding this development is essential read-
ing for anyone seeking to understand this crucial chapter in the rise of the
bacterial theory of disease (Geison, 1995, pp. 146–76).
9. Through this paragraph, page numbers in parentheses are references to
Koch (1881b).
10. This will be discussed in more detail in the next chapter.
11. This may also be an attempt to discredit Pasteur since Koch, having been
trained in pathology, was qualified to undertake such a study whereas
Pasteur, a chemist, was not.
12. Bruno Latour has shown that, in France, Pasteur’s theories were accepted
in some groups far more readily than his evidence would warrant (1984).
In part, this resulted from a kind of bilateral leveraging made possible by
the intersecting interests of Pasteur and the Hygienists. However, as

Latour admits, Pasteur’s ‘opponents were numerous enough in the
Académie de Médecine’ (p. 29). In my discussion, ‘the critics’ who re-
quired conversion were these members of the Académie – physicians,
pathologists, physiologists, researchers and scientists. It was they who
raised rational arguments against Pasteur; and it was they against whom
Pasteur’s rational arguments were directed. I will have more to say about
Latour in the concluding chapter entitled ‘Some Final Thoughts’.
13. The first four paragraphs of this section are based on an earlier publica-
tion (Carter, 1982b); documentation for all the treatments mentioned
here and below can be found in that source.
14. The story of Pasteur’s rabies research is more complex than this superfi-
cial account suggests. For what is, as yet, the most complete and
enlightening discussion see Geison (1995).
15. As Geison has revealed, before Joseph Meister, Pasteur had treated two
other persons with only equivocal results (Geison, 1995, pp. 195–203).
16. After writing this, I discovered that Latour uses a similar metaphor (1984,
pp. 87f.).
CHAPTER EIGHT
The Etiological Standpoint

With a bacterial theory in place, it is possible to identify causes of certain
pathological processes, to define diseases in terms of those causes, and to
secure the practical and theoretical benefits of a system of universal and
necessary causes. However, the etiological standpoint – the research pro-
gramme we are investigating – was never limited to the bacterial theory. It
began with diseases caused by fungi, by minute but visible parasites, and
by decaying organic matter. And while the bacterial theory eclipsed (and
in one case assimilated) these earlier accomplishments, it seems always to
have been clear that the research programme was not limited to bacterial
diseases. Koch spent most of his professional life investigating diseases he
knew to be non-bacterial and he clearly regarded all this work as falling
within what he called the etiological standpoint. More revealing yet was
a comment by Adolf von Strümpell, a physician writing in the 1880s
about psychological disorders: ‘one can justly claim that the scientific
treatment of the etiology of diseases constitutes the most characteristic
thrust of modern pathology, and … the secure establishment of the
doctrine of organized, externally invading disease agents is until now the
most beautiful and important achievement of this effort’ (Strümpell,
1884a, p. 2). Implicit in this comment is the recognition that the ‘scien-
tific treatment of the etiology of diseases’ is not limited to ‘externally
invading disease agents’. Indeed, Strümpell’s article discusses the etiology
of non-bacterial psychological disorders. So, as he conceived it, the
etiological reorganization of medicine was not just about bacteria.
Koch’s causal criteria (his so-called Postulates) can be seen as the
culmination of one aspect of the bacterial theory: The Postulates in-
clude a thoughtful and explicit formulation of causal criteria within
that theory. However, the Postulates also point beyond the bacterial
theory to the possibility of establishing alternative theories in which
factors other than bacteria are recognized as causes. Seen in this light,
the Postulates begin to hint at the full breadth of the etiological stand-
point.
The origin of Koch’s postulates
Koch used explicit causal criteria in fewer than ten of his more than 100
published papers, and, with the exception of one 1890 lecture, all of
these papers appeared between 1878 and 1884. There are important
differences in the causal criteria that Koch stated: later versions in-
cluded conditions absent from the earlier ones, and some of the conditions
were given in several non-equivalent versions. We will examine Koch’s
discussions of causation in his first anthrax paper, in two 1878 papers
on wound infections, and in several papers published between 1882 and
1884. However, his later accounts are most interesting, and discussion
of individual criteria will be postponed until we have reviewed all the
publications.
We have already considered Koch’s 1876 anthrax paper; in this con-
text, a few additional comments are necessary. In this first paper, Koch
neither discussed nor even mentioned causal criteria; he gave little explicit
attention to proving that anthrax bacilli cause anthrax, and the few
arguments he put forward focused on necessity. By the time Koch wrote
this paper, he knew that the mere presence of anthrax bacilli in a suscep-
tible host was not sufficient to insure onset of anthrax: ingesting bacilli
did not usually provoke the disease (Koch, 1876, p. 11), some inocula-
tion procedures were unreliable (Koch, 1876, p. 2), and among exposed
susceptible animals, vulnerability varied (Koch, 1882b, pp. 101f.). Koch
never claimed the bacilli were sufficient to cause anthrax. After some
early failures, he developed a reliable inoculation procedure; thus, this
procedure (as opposed to the organisms alone) was sufficient for the
disease. However, in describing the procedure, he wrote only that it
provided a useful test for the viability of his cultures (Koch, 1876, p. 2).
Nowhere did he suggest that the inoculations constituted direct or con-
clusive evidence that bacilli caused anthrax. Instead, here is his only
argument for causation: ‘anthrax substances, whether they are fresh,
decayed, or dried, can only cause anthrax if they contain Bacillus anthracis
or its viable spores. This fact removes all doubt that Bacillus anthracis is
the actual cause and contagium of anthrax’ (Koch, 1876, p. 13). Thus, as
he himself presented his argument, he seems to have been most influenced
by his unsuccessful inoculations with substances lacking bacilli and spores.
Of course, such experiments provided evidence of necessity.
In 1878, Koch published two papers on wound infections. The first
was a summary lecture; the second was a long account of his research.
Koch began the summary by observing that the regular identification of
microorganisms in infected wounds did not prove that these diseases
come about only when organisms enter the body and proliferate
there, in other words, that the diseases are parasitic. This is be-
cause in many cases of unquestionably infectious diseases, either
no organisms are found or too few are found to explain the symp-
toms or the fatal termination of the disease.
(Koch, 1878b, p. 58)
THE ETIOLOGICAL STANDPOINT 131
According to this passage, regularly finding parasites in infected wounds

does not prove that wound infections are parasitic because, in other
cases, no parasites or too few parasites are found. Thus, presumably, to
support the parasitic concept, one must show that a sufficient number
of parasites are always present in infection. As he here explains it, his
goal was, therefore, to show that infections do not occur without
parasites – in other words, that the parasites are necessary for disease.
By using accepted methods, Koch was able to identify organisms in only
a few test animals, so his results were inconclusive. However, he re-
ported that by using new staining and illumination techniques, which
he had developed, he could invariably locate microorganisms (Koch,
1878b, pp. 58f). This effectively removed objections to the parasitic
account. Koch also conducted inoculation experiments, but, as in the
first anthrax paper, seems not to have regarded them as central to the
causal argument; he reported only that inoculations were useful for
insuring the availability of diseased animals in which organisms could
be identified. He also wrote that inoculations proved wound diseases
were indeed infectious. He did not cite these experiments in his argu-
ments for causality – therefore, in the short paper on wound infections,
Koch’s concept of causation seems to have been the same as in his first
anthrax paper.
In his long paper on wound infections, Koch first explicitly stated
causal criteria, although they are quite different from those now attrib-
uted to him. He wrote that a conclusive proof of the parasitic origin of
some disease ‘would require that we find parasitic microorganisms in
all cases of the disease, that their presence is in such numbers and
distribution that all the symptoms of the disease can be explained, and
finally that for every individual traumatic infective disease, a morpho-
logically distinguishable microorganism is identified’ (Koch, 1878a, p.
29). We can summarize these conditions as follows:
Rw1. The organism must be exhibited in every examined case of the

disease.
Rw2. The distribution of the organism must correlate with and ex-
plain disease phenomena.
Rw3. For each different disease, a morphologically distinguishable
organism must be identified. (This is a strong version of the
Distinguishability Hypothesis.)
Koch stated the same criteria in another passage, and he mentioned

Rw1 and Rw2 in a third passage as well (Koch, 1878a, pp. 48, 27).
These criteria do not include isolation and inoculation – procedures
central to his later attempts to prove causation.
Koch’s discussions of particular wound infections were loosely organ-

ized around these criteria. For each disease, Koch identified a particular
organism and showed that its distribution in diseased animals could
explain the pathological changes, some of the symptoms, and the even-
tual death of the animal: to some extent, this confirmed Rw2 and Rw3.
As in the summary paper, Koch attributed earlier failures in finding
organisms to technical problems and suggested that his improved meth-
ods showed organisms were always present (Koch, 1878a, p. 48). Exactly
as in the anthrax paper, these observations provided evidence of neces-
sity and also supported necessity indirectly by explaining apparent
counter evidence. However, Koch’s evidence for Rw1 and Rw3 was
relatively weak. Here, as in his anthrax paper, Koch’s discussion rested
almost entirely on artificial cases of disease: he did not locate parasitic
organisms in natural infections, and he applied his results to human
infections only by an analogical argument (Koch, 1878a, pp. 19, 48).
This may help explain why Koch never claimed his work on infected
wounds conclusively established causality.
In his early research, Koch introduced several important technical
innovations. Recognizing that the identification of microorganisms was
too subjective, he developed new methods for staining and illuminating
slides and for photographing microorganisms. His photographs were so
well done that, according to one modern microbiologist, they ‘are as
good as many that are published in scientific journals today’ (Collard,
1976, p. 20). At the time, bacteriologists relied almost exclusively on
liquid media for cultures. Solid media such as cooked potatoes were
known, but their significance was not widely recognized. With liquid
media it was impossible to separate various strains of organisms that
happened to invade a given culture and this made pure cultivation diffi-
cult. To overcome this problem, Koch pioneered the so-called ‘poured
plate method’ of cultivation. In using this method, one inoculates a liquid
medium with the organisms one wishes to cultivate and then pours the
mixture onto plates where the medium solidifies. Individual bacteria
grow into discrete colonies that can be protected from contamination.
In the decades preceding the end of the nineteenth century, what we

now call tuberculosis consisted of various anatomically and symp-
tomatically distinct disorders the most prominent of which were called
scrofula and consumption or phthisis. In Europe these disorders prob-
ably accounted for about one-seventh of all deaths and for about
one-third of all adult deaths (Koch, 1882a, p. 83). Consumption, the
most prominent disease of the age, became an artistic and literary
symbol and significantly influenced contemporary thought about the
human condition (Sontag, 1979). There were different opinions about

the etiology of the disease. Some believed consumption was due to
general social causes and could only be controlled through social reforms;
others believed consumption, scrofula, and diseases like syphilis and
hysteria were inevitable consequences of the degeneration of the human
race. Virchow, like many other pathologists, believed the morbid changes
in terms of which consumption was defined were non-specific tissue
reactions that could not have a unique causal explanation.
Some physicians suspected that consumption may be contagious, but
the evidence was inconclusive. In the 1860s Jean Antoine Villemin
published a series of studies supporting the contagious concept by
showing the disease could be conveyed between susceptible animal
species (Villemin, 1868). Villemin believed consumption was due to a
living contagium, but he could not identify it. One problem obstructing
early research was the difficulty of distinguishing reliably between natu-
ral and artificial cases of the disease. There was always the possibility
that animals used in inoculation tests had simply contracted natural
tuberculosis that was not due to experimental inoculations. Julius
Cohnheim and Carl Julius Salomonsen solved this problem by selecting
the anterior ocular chamber of rabbits as the inoculation site (Cohnheim
and Salomonsen, 1877). Since rabbits did not contract spontaneous
tuberculosis of the eye, this seemed to provide conclusive evidence that
at least some forms of tuberculosis could be conveyed by inoculation.
Cohnheim and Salomonsen believed tuberculosis was parasitic, but
they, too, were unable to identify the parasite.
Koch began studying tuberculosis in August 1881 (Brock, 1988, pp.
117–39). His work was carried out in great secrecy – apparently not
even Ferdinand Cohn, who had been his friend and mentor since his
first work on anthrax, was informed of the study. Koch and his associ-
ates conducted their work at a frantic pace and after only a few months,
on 24 March 1882, they announced their initial results. The announce-
ment was made at an evening medical meeting in Berlin; it was held in a
small reading room, 30 by 45 feet in size.
A total of seventy-two chairs were available, and table space for
200 preparations was also required. … The room was filled to the
last available seat. Virchow was conspicuously absent. He had
recently taken a strong stand against the ‘juvenile work’ of the
‘youngsters’ of the Imperial Health Office.
(Lechevalier and Solotorovsky, 1965, p. 84)
Koch’s paper was marvelously clear – the most compelling and persua-
sive essay he ever wrote.
As usual, technical innovations were central to his argument. Koch
found that adding potassium to the standard methylene-blue stain
intensified the effect and revealed a new kind of bacillus in tubercu-

lous materials. He then grew cultures of these bacilli in coagulated
serum.
His genius was revealed by his patience in waiting several weeks
for growth to occur. All known bacteria, at the time, appeared in
culture media within a few days; if nothing was visible, the cultures
were usually discarded. Just what went on in Koch’s mind to make
him preserve his cultures of tuberculous material and patiently
wait for several weeks he does not say. Perhaps he forgot to discard
them, or perhaps he did not know himself; therein lay his genius.
(Bloomfield, 1958, p. 212)
Extensive inoculation tests supported the conclusion that the bacillus in
question was the causal agent for tuberculosis.
Koch’s presentation was stunning; ‘the audience was left spellbound,
and for a time after he had ended the presentation not a word was
uttered’ (Lechevalier and Solotorovsky, 1965, p. 84). In later years, Paul
Ehrlich wrote of the meeting: ‘all who were present were deeply moved
and that evening has remained my greatest experience in science’
(Lechevalier and Solotorovsky, 1965, p. 84). Koch’s paper created an
immediate sensation and probably did as much as any single accom-
plishment to establish the bacterial theory of disease. Later in 1882,
Koch published a second paper on tuberculosis. Other papers appeared
over the next two years culminating in his 1884 paper on the etiology
of the disease. In these papers, Koch meticulously followed specific
criteria for establishing disease causality – criteria we now know as
Koch’s Postulates. The centrality of the Postulates in Koch’s famous and
eminently influential tuberculosis papers ensured that they would pro-
foundly influence subsequent etiological research.
In the first tuberculosis paper, Koch observed that ‘the first goal of
the investigation was to exhibit certain parasitic forms foreign to the
body and that could cause the disease’ (Koch, 1882a, p. 84). He re-
ported finding tuberculosis bacilli in materials from dozens of natural
cases of human and animal tuberculosis and in approximately 200
animals that had been inoculated with tuberculous materials. He con-
cluded that ‘tubercle bacilli occur in all tuberculosis disorders and …
are distinguishable from all other microorganisms’ (Koch, 1882a, 87).
Up to this point, Koch had presented evidence that the bacilli are
necessary for the disease. Next he observed that the bacilli ‘are distin-
guishable from all other microorganisms’. In both 1882 tuberculosis
papers, Koch observed that establishing a regular coincidence between
a disease and a specific organism did not prove causality; he wrote that
even identifying the parasite in the organs where the disease is known to
originate is not conclusive (Koch, 1882c, p. 446). The next step was to
isolate the suspected disease agent in pure culture and to confirm it was
alive. Even after all of this,
it remains to prove the isolated parasite is really the cause of the
disease. To accomplish this, one must show that animals inoculated
with pure culture contract the original disease. The inoculation
must succeed not only sometimes, but in every attempt as is achieved
in such infectious diseases as anthrax.
(Koch, 1882c, pp. 446f.)
In substantial sections of both 1882 papers, Koch reported attempts to
inoculate materials from pure cultures. The effort devoted to this part
of the argument, together with his own explicit assertions, show clearly
that he regarded this as the decisive step in proving causation. After
reporting his inoculation experiments, he concluded, ‘All these facts,
taken together, show that the bacilli in tuberculous substances are not
merely coincidental with tuberculosis, but cause it. These bacilli are the
real tuberculosis virus’ (Koch, 1882a, p. 93). Koch reclassified, as one
disease, all the symptomatically and pathologically distinct cases in
which the organism could be identified and excluded all other cases as
not true tuberculosis (Koch, 1882a, pp. 93f.) thereby, in effect, giving a
new definition to ‘tuberculosis’ (Taylor, 1979, pp. 18f.).
In 1882 Koch also published a paper criticizing Pasteur’s attempts to
immunize animals against anthrax (Koch, 1882b). In this paper, he
described what he called his own method of studying infectious disease,
and he contrasted this method with what was supposedly Pasteur’s
method. Koch claimed to begin by examining ‘all the body parts that
are altered by the disease. In this way, one can establish the presence of
the parasites, their distribution in the diseased organs, and their relation
to body tissues’ (Koch, 1882b, p. 98). However, this study provides
only ‘a complete orientation’ to the disease – it still remains to prove
causation. For this purpose, the organisms ‘must be cultured in pure
form. Then, after they have been freed from all the parts of the diseased
body with which they were originally associated, they must be inocu-
lated back into animals of the same species as those in which the disease
was originally observed’ (Koch, 1882b, p. 98). Koch mentioned tuber-
culosis as a disease for which these criteria had been fully satisfied.
Thus, in this 1882 anthrax paper, Koch endorsed the same steps for
proving causality as in the tuberculosis papers appearing in the same
year. As we have seen, these were not the steps he himself followed in
his original work on anthrax; in fact, Pasteur’s study of anthrax came
closer to following these steps than did Koch’s.
Koch’s monumental 1884 paper on the etiology of tuberculosis con-
tains his most complete discussion of causation. He wrote:
First, it was necessary to determine whether the diseased organs

contained elements that were not constituents of the body or com-
posed of such constituents. If such alien structures could be
demonstrated, it was necessary to determine whether they were
organized and showed any sign of independent life. Such signs
include motility – which is often confused with molecular motion –
growth, propagation, and fructification. Moreover, it was neces-
sary to consider the relation of such structures to their surroundings
and to nearby tissues, their distribution in the body, their occur-
rence in various states of the disease, and similar other conditions.
Such considerations enable one to conclude, with more or less
probability, that there is a causal connection between these struc-
tures and the disease itself. Facts gained in these ways can provide
so much evidence that only the most extreme skeptic would still
object that the microorganisms may not be the cause, but only a
concomitant of the disease. Often this objection has a certain jus-
tice, and, therefore, establishing the coincidence of the disease and
the parasite is not a complete proof. One requires, in addition, a
direct proof that the parasite is the actual cause. This can only be
achieved by completely separating the parasites from the diseased
organism and from all products of the disease to which one could
ascribe a causal significance. The isolated parasites, if introduced
into healthy animals, must then cause the disease with all its char-
acteristics.
(Koch, 1884b, p. 131)
Here, and in his 1882 papers, Koch outlines a series of steps for proving
causation. These steps can be summarized as follows:
Rt1. An alien structure must be exhibited in every case of the disease.

Rt2. The structure must be shown to be a living organism and must
be distinguishable from all other organisms.
Rt3. The distribution of organisms must correlate with and explain
disease phenomena.
Rt4. The organism must be cultivated outside diseased animals and
isolated from all disease products that could be causally signifi-
cant.
Rt5. The pure isolated organisms must be inoculated into test animals
and these animals must then display the same symptoms as the
original diseased animal.
Several of these steps require comment:
1. Rt1 corresponds approximately to Koch’s earlier Rw1. In Rt1 Koch

refers to the structures as ‘alien’ by which he presumably meant
that they are not normally found in healthy bodies. If so Rt1 (in
contrast to Rw1) implies that structures are found in all and only
cases of the disease. Thus Rt1 requires observational evidence both
of necessity (structures present in all cases of pathology) and of

sufficiency (structures present only in cases of pathology).
2. Rt2 corresponds approximately to Rw3. Rw3 requires that the
organism is morphologically distinguishable from all other organ-
isms. However, by the time he wrote the tuberculosis papers, Koch
realized that this requirement was too strong. In 1884 he wrote
that ‘morphological characteristics are not normally sufficient to
distinguish bacteria’ (Koch, 1884c, p. 171), and later, in 1890, he
urged that every possible characteristic of different strains of or-
ganisms be considered before identifying them (Koch, 1890, pp.
180f.). As finally stated in the 1884 tuberculosis paper, Rt2 is the
Distinguishability Hypothesis.
3. Rt3 is the same as Rw2. As we have stated this condition, it
requires that the purported cause both correlate with and explain
the observed disease phenomena, and this is how Koch usually (but
not always) phrased it. Occasionally, Koch followed Klebs who (in
E1) required correlation but did not mention explanatory adequacy.
Explanatory adequacy is essential in proving causation, so Koch’s
way of putting this condition is superior to Klebs’ formulation.
4. As we have seen, Rt4 and Rt5 have no counterpart in the 1878
papers on infected wounds or in Koch’s first anthrax paper. How-
ever in a later paper on anthrax, Koch spoke of Rt4 and Rt5 as
providing the real proof of causality – he regarded the earlier steps,
Rt1 through Rt3, as providing only an orientation preliminary to
the actual proof (Koch, 1882b, p. 98). Rt4 and Rt5 provide strong
evidence of sufficiency. Together they correspond approximately to
Klebs’ E2 (p. 96); later we will review evidence that Klebs was the
source of Koch’s isolation and inoculation criteria.
Through the 1880s, Koch and his fellow workers discovered the
causal agents for a range of bacterial diseases. They also published
important papers on applying bacteriology to public health, they an-
nounced techniques for measuring the bacterial content of air, soil, and
drinking water, they tested new methods for disinfection, and they
developed a procedure for steam disinfection. Later, Koch discussed the
methods that produced these results:
These new methods proved so helpful and useful in dealing with
various problems that one could regard them as the keys for the
further investigation of microorganisms, at least insofar as they
relate to medicine. Once I had developed them and gained some
experience in their use, I utilized them in the study of the patho-
genic microorganisms. In a rapid sequence, my colleagues and I
were successful in discovering the cause, and thereby the etiology,
of a number of infectious diseases. These included the wound
infections, tuberculosis, cholera, typhoid, and diphtheria. Once the

appropriate methods had been found, these discoveries fell into our
laps like ripe fruit, and they were then used for practical purposes
such as the control of such plagues as cholera, typhoid and malaria
(Koch, 1909, p. 4).
Two further matters require attention before we turn to a broader

interpretation of Koch’s Postulates. First, we will give more attention to
the sufficiency criteria, Rt1, Rt4 and Rt5. Second, we will review the
evidence that Koch derived the Postulates – in particular Rt4 and Rt5 –
from Edwin Klebs.
Can we be sure that any organism that satisfies Rt4 and Rt5 is always
followed by a particular disease? Of course, one could insure sufficiency
in the same way one insures necessity, namely, by defining the disease so
that the organism is sufficient as well as necessary. This would mean
stipulating that a person has a disease, say, tuberculosis, if, and only if,
that person harbors the causal bacterium. There are diseases for which
this approach seems appropriate: embedded fertilized female acari mites
are both sufficient and necessary for scabies; embedded live trichinae
are sufficient and necessary for trichinosis. We define both diseases so
that hosting these parasites is having the diseases. Henle proposed that
causal agents should be sufficient and necessary, and Pasteur sometimes
claimed that causal organisms were sufficient and necessary (Pasteur,
1881e, p. 554).
However, in many cases, practical considerations count against taking
the mere presence of an organism as sufficient for having a disease, and
some of these considerations were apparent in early research. Semmelweis
reported a case in which decaying organic matter from a patient with a
discharging medullary carcinoma spread fatal childbed fever throughout
her ward while she herself remained healthy (Semmelweis, 1861b, pp.
43f). Davaine, Pasteur, and Koch all knew that sheep could consume
anthrax bacilli without becoming diseased and that inoculations with
virulent anthrax blood sometimes had no effect. Koch admitted that
tubercle bacilli could exist in a suitable host without causing symptoms
(Koch, 1882a, pp. 94f.). If, as a matter of definition, the presence of
causal organisms is made sufficient for these diseases, the preceding
examples all become asymptomatic cases of the respective diseases. How-
ever, in practice it is more useful to label as diseased only persons who
require medical intervention, rather than everyone infected with the causal
agent, so there are practical advantages to restricting ‘tuberculosis’ to
conditions involving more than the simple presence of the cause. While
we do recognize clinically inapparent cases of many diseases, most dis-
ease names are so used that, in order to have the disease, victims must
both harbor the causal agent and manifest symptoms of a certain inten-
sity or otherwise require medical intervention. Thus, for practical reasons,
we often decide against making causes definitionally sufficient for their
diseases.
Exactly what evidence is furnished by Rt1, Rt4 and Rt5? One can
think of Rt1 as providing observational evidence of sufficiency; alterna-
tively, Koch suggested that one could also regard this criterion as
providing motivation (what he called an orientation (Koch, 1882b, p.
98)) for investigating causality rather than as part of the actual proof.
What about Rt4 and Rt5 – the isolation and inoculation test? In a
tuberculosis paper, Koch insisted that ‘inoculation must succeed not
only sometimes, but in every attempt as is attained in such infectious
diseases as anthrax’ (Koch, 1882c, pp. 446f.). Even if this is achieved, it
does not mean the organism is sufficient for the onset of symptoms. As
we have seen, Koch knew that the mere presence of anthrax bacilli was
not sufficient for anthrax: ingesting them did not usually provoke symp-
toms (Koch, 1876, p. 11), even some inoculation procedures were
unreliable (Koch, 1876, p. 2), and vulnerability varied among suscepti-
ble animals (Koch, 1882b, p. 101). Koch never claimed bacilli alone
were sufficient for anthrax, and current definitions of ‘anthrax’ allow
organisms to occur in test animals in the absence of the disease. Thus,
even the invariable success of inoculations does not – and, given most
definitions, can not – demonstrate that the organism alone is strictly
sufficient: universal success shows only that the organism can be used in
an inoculation procedure that is sufficient. This claim is obviously
weaker than the claim that the factor itself is sufficient. Because criteria
like Rt4 and Rt5 support only this weaker claim, they have been called
weak sufficiency criteria (Carter, 1985b, pp. 358f.). As here defined, a
weak sufficiency condition requires that there is some means by which
the organism can be introduced into a host in such a way that the host
will invariably become diseased.
Are weak sufficiency criteria essential to proving causation? As we
have seen, both Koch and Pasteur started out arguing for causation
using only evidence of necessity. However, after criticism by contempo-
rary physicians who expected causal arguments to establish sufficiency,
both adopted weak sufficiency criteria. Indeed, both insisted that isola-
tion and inoculation was the only way of proving causation. Yet within
months of making this claim, Koch admitted he was unable to find an
animal vulnerable to experimental cholera; he also acknowledged that
no susceptible animals had been found for leprosy or for abdominal
typhus. He concluded: ‘we must therefore be content that we have
established the constant occurrence of a specific species of bacteria in
some disease, and that these bacteria do not occur in other diseases. …
This is acceptable because we already know of other infectious diseases

that are caused by pathogenic organisms and that resemble these in
every respect’ (Koch, 1884d, p. 161). Thus, unable to satisfy his own
weak sufficiency criterion, Koch fell back on observational evidence of
necessity and an analogical argument based on what was known of
other diseases.
Even when inoculations are successful, they may not meet Koch’s
conservative standard – namely, of succeeding ‘not only sometimes, but
in every attempt’ (Koch, 1882c, pp. 446f.). One may be able to show
only that exposure increases the likelihood of the disease. In fact, while
he never made this explicit, this is all Koch achieved in his first anthrax
inoculations. If this is the most that can be shown, the organisms could
be thought of only as a risk factor – like smoking or malnutrition.
Alternatively, the success of inoculations may depend on meeting spe-
cial ancillary conditions. Pasteur found that chickens could contract
experimental anthrax, but only if one lowered their body temperatures
so bacteridia could flourish in their blood. If such special conditions are
essential for success, a causal agent may be thought of only as a cofac-
tor in the disease. Risk factors and cofactors both fall short of strict
causality, but they do involve causation in a broad sense, and causal
criteria have been proposed that accommodate them. For example,
rather than Koch’s conservative Rt4 and Rt5, a more liberal criterion
may require only that ‘incidence of the disease is significantly higher
among those exposed to the putative cause than among those not so
exposed’ (Evans, 1976, p. 192). We now accept, as evidence of kinds of
causation, weak sufficiency conditions that are considerably more lib-
eral than the conservative criterion Koch himself claimed to follow.
Indeed for some disorders, we recognize causation even when there is
no possibility of ever satisfying any weak sufficiency criterion whatso-
ever; we sometimes identify specific causes in the absence of all
experimental evidence of sufficiency. At present, this is true for all
human genetic disorders; for example, trisomy 21 is the most common
cause of Down’s Syndrome but there are no experimental tests for
sufficiency. This suggests that weak sufficiency criteria, whether liberal
or conservative, may not actually be essential even in proving causation.
Nevertheless, for at least some kinds of diseases, especially for bacterial
diseases, since the middle of the nineteenth century, evidence of weak
sufficiency has been part of what one normally expects in causal argu-
ments.
Henle’s ideal was that rational medicine should accept only causes
that are necessary and sufficient. However, the causes we now identify
are not usually sufficient, and there is seldom any point in defining
diseases so that their causes will be. Moreover, while (for at least some
kinds of diseases) we expect demonstrations of causation to include the

satisfaction of weak sufficiency criteria, even sufficiency in this sense
may not be absolutely essential to the concept of a specific cause.
In 1840 Jacob Henle speculated about proving causation by separating

living organisms from the fluids in which they are located and by
observing the powers of each separately (Henle, 1840, p. 948). Since
Henle was Koch’s teacher at the University of Göttingen, Koch is often
thought to have adopted the Postulates, or at least Rt4 and Rt5, from
Henle (Evans, 1976, p. 175; Richmond, 1978, p. 85). There are reasons
for doubting this view. First, while admitting that he owed Henle a
great debt of gratitude (Koch, 1909, p. 3), Koch never suggested Henle
had influenced his thinking about disease causation or even his interest
in bacteriology. Indeed Koch wrote that he received no encouragement
to study bacteriology while he was a student at Göttingen since ‘bacteri-
ology did not exist at that time’ (Koch, 1909, p. 3). The recollections of
Elie Metchnikoff, who worked with Henle in Göttingen just at the time
Koch was graduating, confirm Koch’s remark. ‘When, in 1866, I worked
under Henle, in Göttingen, at a time when there were serious investiga-
tions on the microscopic agents of infectious disease he remained
indifferent and … at no time did the question of contagious diseases
come up in his laboratory’ (Lechevalier and Solotorovsky, 1974, p. 65).
Second, in his early work on anthrax, Koch isolated relatively pure
cultures of anthrax bacilli and inoculated them into test animals. How-
ever, at that time, he seems not to have regarded this procedure as
particularly significant in establishing causality. The causal arguments
in his early anthrax papers stressed necessity rather than sufficiency.
The first weak sufficiency criterion appeared in his long 1878 paper on
infected wounds, and Koch first used isolation and inoculation as a way
of proving causality in his 1882 tuberculosis papers. If, as a student,
Koch had adopted Rt4 and Rt5 from Henle, one would have expected
these criteria to figure in his early causal arguments.
When Koch began studying anthrax as a district physician in Wollstein,
he was isolated from everyone else who was studying microorganisms.
In his first anthrax paper, he admitted he did not have access to several
important publications (Koch, 1876, p. 7) and he relied on reviews and
abstracts rather than original papers. His first anthrax papers and his
summary version of the 1878 paper on infected wounds, delivered
before the Society for German Natural Scientists and Physicians, con-
tain no explicit causal criteria. The Society was the same organization
before which Klebs had presented his famous 1877 lecture on the
revolution in medical thought, and, in the 1878 meetings, again criti-
cized Virchow’s research programme (Klebs, 1878b). In 1878 Koch was

engaged in his priority dispute with Pasteur regarding the role of
bacteridia in anthrax. Klebs and Koch were often together at the 1878
meetings (Heymann, 1932, p. 236), and, in light of the priority dispute
and Klebs’ own earlier attempt to prove causation for anthrax by
isolation and inoculation, they certainly could have discussed methods
for proving causation. Shortly thereafter Koch published his long paper
on wound infections – the paper containing his first explicit use of
causal criteria. In this paper Koch cited several of Klebs’ essays (Koch,
1878a, pp. 22, 26) and he mentioned Klebs’ method of fractional
cultivation (Koch, 1878a, p. 26). Koch credited Klebs with having been
the first to attempt to prove that microorganisms cause wound infec-
tions (Koch, 1878a, p. 22). In his paper on infected wounds, Koch also
frequently cited Felix Victor Birch-Hirschfeld’s textbook on pathology.
In this text, Birch-Hirschfeld credited Klebs with developing the two
main strategies for proving disease causation and he referred to Klebs’
work as ‘epoch making’ (Birch-Hirschfeld, 1872, p. 98).
Koch cited Klebs in several papers that he published in the early
1880s (Schwalbe, 1912, vol. 1, pp. 133, 158f., 183). In his famous
tuberculosis papers of 1882 and 1884, he cited Klebs but did not
associate Rt4 and Rt5 with him (Schwalbe, 1912, vol. 1, pp. 433, 437,
468, 525, 529). However, in his less famous and less commonly read
second 1882 paper on tuberculosis, Koch explicitly attributed the strat-
egy to Klebs. He mentioned that, while various methods had been used
to prove causality, ‘the best method, the method used by everyone who
has been seriously occupied with these investigations, was introduced
and refined by Klebs’ (Koch, 1882c, p. 446). He then described this
method as, first, producing successive pure cultures to separate the
parasite from all disease products and, second, inoculating isolated
parasites into test animals. Koch wrote that his own study of tuberculo-
sis followed this procedure.
Thus, it seems likely that Klebs was the immediate source for Koch’s
Rt4 and Rt5. Koch first invoked isolation and inoculation at about the
time of the 1878 meetings. Perhaps, at that time, he was concerned with
causal arguments because of his dispute with Pasteur, and perhaps
discussions with Klebs helped him see the necessity of this approach.
The etiological standpoint
By the second half of the nineteenth century, medicine was beginning to

prize causes of a particular kind: universal necessary causes that met
recognized criteria, that explained disease phenomena and held out the
promise of control. One contemporary described the quest for such

causes as ‘the chief science of medicine’ (Stamm, 1865, p. 477). Of
course the greatest success of this effort was the bacterial theory of
disease.
However, through the 1890s it began to appear that many common
disorders were not bacterial. Careful study failed to reveal bacterial
causes for hysteria, rabies, smallpox, measles, beriberi, and other disor-
ders. At this point, one could simply have concluded that only certain
pathologies – bacterial diseases plus the few diseases attributed to non-
microscopic parasites – could be associated with specific causes. This
would have allowed for the possibility that other disorders could only
be traced to the familiar varieties of causes (miasms, anxiety, gluttony,
dissipation, self-abuse) that had been acknowledged for centuries. In-
deed, some fifty years earlier, this is exactly how everyone had responded
to the realization that acari and Beauvaria bassiana caused scabies and
muscardine, and to Semmelweis’s discovery that decaying organic mat-
ter caused childbed fever. Earlier in the century, such diseases were
anomalous, and no one inferred, from these anomalies, that other dis-
eases could also be organized in terms of universal and necessary causes.
By contrast, in the 1890s, no one seems to have inferred, from the
failure to identify causal bacteria for some disorders, that those
pathologies must lack specific causes – one searches in vain for any such
idea in the medical literature of the period. While traditional causes
continued to be mentioned perfunctorily in connection with such ill-
nesses as hysteria,1 everyone seems to have been open to the possibility
that specific causes would be identified for these disorders as well. By
this time there had been a reversal in the accepted norms and in what
was regarded as anomalous. As it began to appear that a particular
disease may not be bacterial, researchers simply turned to other possible
specific causes.
Of course, expanding the research programme to include non-bacte-
rial diseases encountered obstacles: new domains of causes had to be
identified, new experimental techniques had to be developed, and new
theories of disease – including new causal criteria – had to be con-
structed. But conceptually there was no significant hesitation, and
researchers moving in new directions encountered nothing like the re-
sistance that Bassi, Semmelweis, Davaine, Mayrhofer, and even Pasteur
had encountered through the middle decades of the century. While the
medical profession accepted new discoveries only as evidence was forth-
coming, everyone was open to the idea that universal necessary causes
could be found even for non-bacterial diseases. The reason for the
difference is clear: by the 1890s, medical researchers had been con-
verted, not just to the bacterial theory, but to the research programme
from which the bacterial theory had emerged – to the etiological stand-
point. Even in the face of substantial obstacles, no one seems to have
questioned that all pathological processes could be so organized that
each disease would have a specific cause. By the 1890s the quest for
such causes drove virtually all medical research.
How do Koch’s Postulates relate to identifying causes for non-bacterial
diseases? While there is little textual evidence that Koch envisioned using
the Postulates beyond bacterial diseases, he would likely have been open
to the possibility. He consistently resisted the idea that all diseases were
bacterial, and his concept of an etiological standpoint surely encom-
passed more than bacterial diseases. But, whatever Koch himself may
have thought, in fact, one can easily interpret Rt1 through Rt5 as a
heuristic for finding causes generally rather than merely as causal criteria
within the bacterial theory. Indeed given that the whole point of the
etiological research programme – Koch’s etiological standpoint – is un-
derstanding and controlling diseases by means of causes, it would be
surprising if the Postulates could not be generalized in this way.
By reorganizing the order slightly, by supplementing the Postulates
with other considerations we have found to be necessary, and by ab-
stracting from bacteria (the particular causes in terms of which Rt1
through Rt5 were originally stated), we can recast the Postulates as a
series of steps that constitute a general strategy for finding causes:
1. One or more theories of disease must be in place. Each theory will

include at least
a. some basic assumptions such as:
● a hypothesis (for example, the Bacterial Hypothesis) speci-
fying a domain of potential causes,
● the Distinguishability Hypothesis (Koch’s Rt2) to insure
there are differences in the potential causes to account for
differences in the pathological processes to which the theory
is applied, and
● such other assumptions as dictated by the nature of the
causes in the domain (for example, theories dealing with
living organisms will require the Dissemination Hypoth-
esis); and
b. criteria that a particular factor must satisfy to qualify as the
cause of a specific pathological process. Which possible criteria
are included will be determined by the nature of the causal
factors in the domain; normally the criteria will include one or
more weak sufficiency conditions (for example, for bacterial
theories, Klebs’ E1 or Koch’s Rt4 and Rt5).
2. Given some unexplained pathological process, one determines which
available theory (if any) to invoke. One makes this judgment by

considering analogies between the morbidity in question and other
pathological processes previously assimilated to the different avail-
able theories. Of course, some pathologies may defy assimilation
and call for the creation of altogether new theories.
3. From among the potential causes in the domain, one seeks a possible
cause that is present in most cases of the pathological process and
absent otherwise (Koch’s Rt1).
4. One establishes that the presence of the potential cause can explain
the observed pathological phenomena (Koch’s Rt3).
5. One demonstrates that the potential cause satisfies the criteria en-
compassed by the theory. Together, steps (4) and (5) show that the
factor in question is the cause of the pathological process under
investigation.
6. Following Semmelweis’s tactic, one characterizes a new disease in
terms of the cause. This new characterization makes the cause
necessary and universal, and, if practical considerations warrant
such a definition, sufficient as well.
7. Since every case of the newly characterized disease shares a com-
mon necessary cause, one now seeks ways of controlling the disease
by manipulating the cause and explanations of disease phenomena
in terms of that cause. Success in these efforts confirms the causal
claim, supports the specific theory within which the claim is ad-
vanced, and even reinforces the entire etiological research programme
of which the theory is a part.
It is important to distinguish in principle between theories of disease

(with their associated assumptions and sets of causal criteria) and the
strategy for finding causes that we have here outlined. Each theory,
along with its causal criteria, will be constructed to address a particular
domain of potential causes (one domain being bacteria), and each do-
main of potential causes will require its own theory. This is true even
though theories of disease often include assumptions analogous to the
assumptions in other theories. For example, each theory will include the
Distinguishability Hypothesis and some assumption (such as the Bacte-
rial Hypothesis) that identifies a domain of possible causes. By contrast,
the strategy for finding causes is central to the heuristic of the research
programme as a whole. Thus it is not restricted to any one domain of
potential causes, and it encompasses the bacterial theory and every
other theory aimed at understanding and controlling disease through
the identification of causes.
Described in these terms, the research programme virtually calls out
for theories of disease beyond the bacterial theory (in Lakatos’s words,
such applications were ‘adumbrated at the start’ (Lakatos, 1968, p.

132)). In the remaining chapters, we will now see how new domains of
causes and new theories of disease were assimilated by the programme.
Note
1. And, to repeat a point made in Chapter 1, insofar as one is seeking a

sufficient cause for the onset of a particular case of illness, such factors
were and still are today potentially relevant.
CHAPTER NINE
An Ideational Theory of Disease

It is generally acknowledged that late nineteenth-century medical re-
searchers were preoccupied with the bacterial theory of disease. ‘With
the work of Pasteur and Koch, … there penetrated rapidly into all fields
of medicine the idea that infinitely small beings, endowed with special
pathogenic qualities, played a pre-eminent role in producing many dis-
eases. The new concept made such a great impression that for a while it
was believed that the cause of all diseases could be ascribed to microbes
alone. … Almost completely dominant, bacteriology at this period be-
came the center and goal of medical investigation’ (Castiglioni, 1947, p.
809). However, this account is misleading because the change that
swept medicine in the late nineteenth century involved more than just
acceptance of a particular theory of disease.
In 1884 (in a passage we encountered earlier), Adolf von Strümpell
wrote, ‘one can justly claim that the scientific treatment of the etiology
of diseases constitutes the most characteristic thrust of modern pathol-
ogy, and … the secure establishment of the doctrine of organized,
externally invading disease agents is until now the most beautiful and
important achievement of this effort’ (Strümpell, 1884a, p. 2). Strümpell
distinguished clearly between scientific etiology and the bacterial theory,
which he described as ‘until now [scientific etiology’s] most beautiful
and important achievement’. Implicit in this description is the recogni-
tion that non-bacterial etiologies could also be studied scientifically and
that such studies could yield achievements as impressive as the bacterial
theory itself. Strümpell was clear – clearer perhaps than some modern
historians – that the etiological standpoint was not exhausted by the
bacterial theory. It was inevitable that the methods of the research
programme, having successfully accounted for many bacterial diseases,
would be brought to bear on other diseases as well.
Sigmund Freud’s early work in psychopathology was among the earli-
est attempts to assimilate a class of non-bacterial diseases to the etiological
standpoint. Yet this aspect of Freud’s early work has been entirely
ignored. To take one prominent example, a great part of Ellenberger’s
The Discovery of the Unconscious is ‘devoted to authors and systems of
thought, which … could be called sources or precursors of Freud’. In a
12-page summary, he gives ‘a succinct list of these sources, insofar as
they are known today’. The list includes more than two dozen persons
and movements, but neither the etiological standpoint nor even its first
outstanding success – the bacterial theory of disease – is mentioned

(Ellenberger, 1970, pp. 34–46).1 By ignoring the connection between
Freud’s investigations and the great research programme of late
nineteenth-century medicine, one overlooks aspects of Freud’s work
that are both different from the research of many of his contemporaries
who wrote on psychopathology and fundamentally allied to work that
was being carried out at the same time in other areas of medicine.
In the late nineteenth century, hysteria was among the most widely
discussed diseases. Partly because Freud’s early work focused on hyste-
ria, nineteenth-century discussions of that disorder have been the subject
of continuing interest. Unfortunately, certain misconceptions, partly
initiated by Freud himself, have been perpetuated in contemporary
accounts. Because of these misconceptions, Freud’s own contribution to
the discussion has been misunderstood. We begin by reviewing standard
medical opinions about hysteria in the 1880s, the decade in which
Freud began his work.
Medical opinions about hysteria in the 1880s
In the 1880s, hysteria was generally regarded as a functional nervous

disorder where ‘functional’ meant a disturbance of function without an
identifiable organic foundation. It was usually classified as a psychosis
or a neurosis. In 1883 Martin Cohn classified hysteria as a functional
psychosis, that is, a disease ‘such that, given the current state of knowl-
edge, no organic alteration of the central organs can be exhibited’
(Cohn, 1883, p. 44). Cohn noted that Emanuel Mendel, a medical
professor at Berlin, distinguished hysteria from other psychoses as a
neurotic condition from which psychotic states follow. In 1884, the
publications by Cohn and Mendel were reviewed in a comprehensive
survey by a Dr Schäfer of Berlin (Schäfer, 1884). Schäfer adopted the
same definition of ‘hysteria’. The first edition of Adolf von Strümpell’s
influential text, Diseases of the Nervous System, appeared in the same
year; Strümpell defined ‘hysteria’ as a functional disturbance without
gross changes in the anatomy of the nervous system (Strümpell, 1884b,
vol. 2, p. 417). Also in 1884, J. Weiss, docent for psychiatry in Vienna,
wrote that hysteria has symptoms and a course of development that
could only belong to a psychosis and could only be treated by psychiat-
ric methods (Weiss, 1884, pp. 457f.). He wrote that the disease is
functional and that attributing the disease ‘to a palpable disorder of the
central nervous system is entirely unthinkable’. Schäfer’s survey, to-
gether with the independent article by Weiss, provided the basis for an
1885 essay by Maximilian Herz, docent for childhood diseases in Vi-
AN IDEATIONAL THEORY OF DISEASE 149
enna (Herz, 1885). Herz adopted the definition of ‘functional psycho-

sis’ from Schäfer and Cohn, and agreed that hysteria should be so
classified. Herz added that numerous attempts to trace hysteria to
anatomical lesions – he mentioned Theodor Meynert, Hermann
Nothnagel, and others – had produced no significant results (Herz,
1885, col. 1371). By 1886, even pathological anatomists described
hysteria as a functional disorder.
In this respect there were no significant differences between the Vien-
nese and Jean Martin Charcot in Paris. Charcot observed that there are
‘a great number of morbid states, evidently having their seat in the
nervous system, which leave in the dead body no material trace that can
be discovered. Epilepsy, hysteria, even the most inveterate cases, [and]
chorea … deny the most penetrating anatomical investigations’ (Char-
cot, 1892–95, vol. 3, pp. 14f.). Charcot identified such disorders as
neuroses – a term defined by his English translator as ‘diseases of the
nervous system apparently due to functional or dynamic causes; which
are not, so far as we know, attended by any organic lesion’ (Charcot,
1889, vol. 3, p. 13). This, of course, is essentially the same definition
that appears so frequently in the German and Austrian medical litera-
ture of the period.
Hysteria was generally regarded as irregular, both because symptoms
could change dramatically in a given patient and because symptoms
could vary fundamentally from patient to patient. Emanuel Mendel
cited Sydenham as having referred to the disease as a veritable Proteus
displaying as many colors as the chameleon (Mendel, 1884, p. 241). In
1883, Heinrich von Bamberger described hysteria as a collection of
‘disturbances in different parts of the body, often contradictory in na-
ture and highly variable, without any anatomical foundation being
discovered in necropsy’ (Bamberger, 1883). Weiss noted that one is
justified in thinking of a hysterical condition whenever one encounters a
group of symptoms that resembles some definite organic illness, but
that departs in some respects from the nature or course of development
of that disease. ‘There is hardly a symptom, whether or not we are in a
position to ascribe it to a particular anatomical foundation, that can-
not, either alone or with other symptoms, belong to the picture of
hysteria’ (Weiss, 1884, p. 452). But, in spite of these irregularities,
physicians tried to detect patterns.
Charcot characterized hysteria by the use of five ‘stigmata’ which he
felt were always present in a greater or lesser degree: (1) sensorial
hemianesthesia, ‘that stigma which almost surely characterizes the hys-
terical condition’; (2) the ovarian phenomenon, that is, the fact that in
many female hysterics an attack could be provoked or arrested by direct
pressure on an ovary; (3) the existence of hysterogenic points that
function like the ovary in provoking and arresting attacks but whose
location varies from patient to patient; (4) the manifestation of a defi-
nite series of stages in hysteric attacks; and (5) paraplegic or hemiplegic
paralysis (Charcot, 1892–95, vol. 3, pp. 115f.). Charcot first presented
this scheme in 1883 (Charcot, 1883, p. 39) and the first German trans-
lation (Freud’s) appeared in 1886. Charcot’s work was known and
followed in Austria and Germany. In 1882, Moriz Rosenthal, who
maintained personal contact with Charcot, characterized hysteria in
terms of Charcot’s stigmata, and distinguished three major classes of
hysterics depending on which symptoms were most pronounced
(Rosenthal, 1879). Eduard Heinrich Henoch’s text on childhood dis-
eases, which went through 11 editions beginning in 1881, contained a
discussion of hysteria based on Charcot. Henoch, like Rosenthal, distin-
guished classes of hysterics depending on which symptoms were most
apparent: psychotic hallucinations; convulsions; motor disturbances in-
cluding paralysis; and sensory disturbances including hemianesthesia
(Henoch, 1881). Henoch’s system was adopted by both Herz and Cohn
in works cited above. Strümpell’s account of hysteria in his text on
diseases of the nervous system was heavily dependent on Charcot. In
the second edition of his text, published in 1885, Strümpell cited Char-
cot ten times in a 22-page account (Strümpell, 1885, vol. 2, pp. 450–71).
He discussed and adopted Charcot’s five stigmata and the stages that
Charcot identified as characteristic of hysteric attacks. Strümpell’s dis-
cussion reflects a thorough grasp of Charcot’s main ideas; the text was
among the most widely used in the field, and would certainly have been
well known in Vienna.
Standard characterizations of hysteria were symptomatic: ‘hysteria,
like neurasthenia, is only a symptom or a complex of symptoms’ (Herz,
1885, col. 1305); ‘we seek the constituent elements of hysteria, the
hysterical symptoms’ (Tuczek, 1886, p. 511); and ‘hysteria designates a
series of the most variable symptom-complexes’ (Cohn, 1883, p. 51). In
each of these passages, hysteria was identified with certain combina-
tions of symptoms. This seemed particularly appropriate (indeed
necessary) given that no organic lesions could be conclusively demon-
strated in autopsies of hysterics – what could hysteria be besides
symptoms? Charcot’s stigmata were obviously of this nature. Weiss,
Henoch, Herz, Cohn, and Oppenheim (Oppenheim, 1890, p. 554) all
adopted symptomatic characterizations. Ludwig Seeligmuller argued
that chorea should be regarded as a form of hysteria since choreatics
invariably display hysterical symptoms (Seeligmuller, 1881, p. 584);
later F. Tuczek argued that hysteria could only be defined symptomatically
(Tuczek, 1886, p. 511). Given that other nervous disorders were also
characterized symptomatically and that, in some cases, it was difficult
or impossible to make differential diagnoses, these disorders seemed to

blend together. Physicians regularly suggested that the nervous disor-
ders were ultimately all one or that they differed only in degree.
As early as the 1850s, cases of male hysteria were regularly described
in European medical literature; 30 years later, in the period we are
considering, it was common knowledge in Vienna and throughout Eu-
rope that either sex was vulnerable (Carter, 1980). In this period there
was great interest in infantile hysteria; this interest, together with the
long recognition of male hysteria, completely exploded the old idea that
hysteria was connected with movements or irritation of the uterus.
Writers in the 1880s occasionally began essays on hysteria by noting
that this connection had been totally abandoned. Tuczek asserted that
‘associating hysteria with the uterus is like associating melancholy with
black bile’ (Tuczek, 1886, p. 511).
Hysteria was generally regarded as caused by the usual combination
of predisposing and exciting factors. Heredity and such conditions as
chronic illness, malnutrition, emotional instability, ethnic origin, ad-
verse climate or meteorological conditions, sexual abnormality, and
persistent irritations (either physical or emotional) were mentioned as
predisposing factors. Even more exciting causes were mentioned; these
included (but were by no means limited to) sexual trauma, illness,
various infections, emotional shocks, inadequate or excessive exercise,
intellectual exertion, and fear. In this respect Charcot was entirely
typical. He distinguished predisposing and provocative causes (Charcot,
1892–95, 2:312); the former included heredity in a broad sense as well
as other factors;3 provocative causes included dog bites, lightning bolts,
unrequited love, and alcoholic and lead poisoning. Charcot explicitly
insisted both that different cases of a single nervous disease such as
hysteria could have a variety of different causes and also that different
diseases – for example, hysteria and epilepsy – could have exactly the
same exciting cause (1:371f., 2:32f.). He also discussed cases of hysteria
that ‘could be assigned to no cause’ (Charcot, 1892–95, vol. 1, p. 366).
Charcot explained that, while he could not ignore causes, as a clinician
his task was to portray and to treat the disease as he presently saw it
(2:360).
There were obvious similarities between the concept of hysteria in the
1880s and the concepts of most diseases at the beginning of the nine-
teenth century. Hysteria was defined and classified symptomatically;
etiological accounts were vague and inconsistent, and included the usual
range of possible factors contributory to sufficient causes of particular
cases. The causes of hysteria were not used to explain other aspects of
the disease. As the contrast between this confusion and the orderly
scientific explanations of the infectious diseases became progressively
more apparent, it was inevitable that neurologists and psychiatrists

would look to the bacterial theory for a model.
In 1884 Adolf von Strümpell advocated a new approach to hysteria and

to the other nervous disorders (Strümpell, 1884a). Strümpell observed
that symptomatology and pathological anatomy could not significantly
advance the comprehension of any disease. Even a complete microscopical
description of a diseased organ could not satisfy the standards for
comprehension that had been established for the bacterial diseases.
Such comprehension, Strümpell observed, could be achieved only when
the symptoms and the anatomical lesions could themselves be explained
as necessary developments from the original cause of the disease and
this required following the model of the bacterial theory (Strümpell,
1884a, p. 3).
In 1888, P. J. Möbius offered an etiological characterization of hyste-
ria and attempted to give causal explanations for its symptoms. ‘All
those diseased modifications of the body are hysterical that are caused
by ideas’ (Möbius, 1888, p. 66).4 Möbius admitted he was unable to
trace all hysterical symptoms to ideas – indeed, even patients may not
be able to give an account of all their mental processes. It is, however, a
common experience that hysterical symptoms often come and go be-
cause of ideas. In what he called an argument by analogy, Möbius
alluded to Charcot’s findings that all hysterical symptoms could be
induced by hypnotic suggestion and concluded that all these symptoms
were caused by ideas. He observed that this definition was confirmed
by clinical experience, but he also mentioned the definition’s theoretical
and practical advantages: it provided conceptual clarity and unity by
realigning the boundaries between hysteria and other nervous disorders,
and a conceptual basis for existing psychiatric therapies; and suggested
new therapies as well (Möbius, 1888, p. 67).
Möbius’s definition was mildly influential; it was given serious criti-
cal attention in European medical literature and was adopted by some
writers. In a later essay (Möbius, 1892), Möbius made it clear that the
definition was intended to bring unity and coherence into discussions of
hysteria by using the same strategy employed in defining the infectious
diseases. He also gave interesting arguments that pathological anatomy
could not provide an adequate basis for understanding or for classifying
diseases, that the nervous disorders must be treated along exactly the
lines exemplified by contemporary work in infectious diseases, and that
this approach, which he regarded as essentially new, would totally alter
the concept of psychological medicine (Möbius, 1892, p. 299). Later, in
their publications on hysteria, Freud and Josef Breuer gave more critical
attention to Möbius than to anyone else (Strachey, 1955–74, vol. 2, pp.

8n., 186–91, 215, 243, 248n.). In 1894 several of Möbius’s essays were
reprinted in a volume entitled Contributions to Neurology. Freud wrote
to his friend, Wilhelm Fliess, that Möbius’s essays were ‘very well done;
they are important on the subject of hysteria. His mind is the best
among the neurologists; fortunately,’ Freud continued, ‘he is not on the
track of sexuality’ (Freud, 1954, p. 101).
In 1892, Strümpell delivered a lecture entitled ‘On the Origin and
Healing of Diseases through Ideas’ (Strümpell, 1893), which carried
one step further the project of explaining the nervous diseases by ap-
pealing to their causes. After some introductory comments, Strümpell
noted that the most characteristic thrust of contemporary medicine was
the quest for causes of diseases. ‘The empty generalities of the past only
appeared, superficially, to satisfy the need for causes; this need can only
be met through the discovery of causes that operate in every single case,
only through a knowledge of their nature, of the manner of their
operation, of the site of their influence, and of the necessity of their
consequence’ (Strümpell, 1893, pp. 22f.). Everyone knows, he contin-
ued, how much our opinions have been enriched and deepened in these
respects in the last 20 years, particularly through work in the area of
the infectious diseases (Strümpell, 1893, p. 23). Strümpell then consid-
ered the influence of psychiatric processes in the generation and healing
of disease. Also in this area the quest for insight into causes has achieved
a level from which the physician, freed from earlier prejudices, can
obtain a clear and realistic perception of the actual situation. Strümpell
considered some of the specific neuroses and showed how regarding
them as diseases of ideas could explain them and the therapeutic meas-
ures employed in treatment.
Between 1884 and 1892, both Möbius and Strümpell took steps
toward etiological accounts of hysteria; however, neither had the
persistence or the imagination to generate a theory with lasting impact.
By the following year, 1893, Freud was beginning to develop just such a
theory.
Freud on hysteria
Freud’s earliest medical studies emphasized neurology and anatomy. We

know that Josef Breuer called Freud’s attention to one remarkable
hysteric, the woman referred to as Anna O., before Freud went to Paris
in the autumn of 1885. Freud reported this case to Charcot, but, he
wrote, ‘the great man showed no interest in my first outline of the
subject, so that I never returned to it and allowed it to pass from my
mind’ (Freud, 1925, pp. 19f.). Apparently Freud began studying hyste-
ria when he was unable to obtain adequate laboratory facilities for the
neurological research that had been his first interest (Freud, 1886, pp.
8f.). James Strachey estimates that this momentous shift occurred in
early December 1885 (Strachey, 1955–74, vol. 1, p. 4).
After Freud returned to Vienna, he presented a paper on male hyste-
ria before the Viennese Society of Physicians. This paper has not survived.
The contents of his presentation and the events of the meeting are
known only from reports published in Viennese and German medical
journals and through a preliminary report that Freud presented to the
Viennese Medical Faculty (Freud, 1886). In this preliminary report,
Freud discussed ‘what was completely novel’ in the studies of Charcot:
he claimed that, prior to Charcot, hysteria had not been well defined,
and that no definitive symptomatology had been assigned to it (Freud,
1886, pp. 10f.). Freud objected to the ‘widespread prejudices’ that
hysteria was attributable to genital irritation, and he credited Charcot
with having refuted this prejudice by demonstrating the unsuspected
frequency of male hysterics. He further attributed to Charcot the dis-
covery of special somatic signs by which hysteria could be conclusively
diagnosed. ‘Thus,’ Freud concluded, by Charcot’s efforts, ‘hysteria was
lifted out of the chaos of the neuroses, was differentiated from other
conditions with a similar appearance, and was provided with a symp-
tomatology which makes it impossible any longer to doubt the rule of
law and order’ (Freud, 1886, p. 12). These claims were certainly not
impressive to Freud’s audience: Charcot’s attempts to systematize the
symptomatology of hysteria were neither unique nor unknown. The
‘widespread prejudices’ to which Freud objected had, in fact, been
abandoned years earlier, and Briquet’s estimation of the frequency of
male hysteria, which formed the basis of Charcot’s opinions, had been
accepted by the Viennese and Germans for years (Carter, 1980, p.
265n.). Freud, who first became seriously interested in hysteria while in
Paris, may simply not have been familiar with existing literature on the
disease. In any case, Freud’s misconceptions, together with his un-
restrained admiration for Charcot, no doubt contributed to the
disappointing reception his paper received (Freud, 1925, p. 15).
Between 1886 and 1893 Freud published little about hysteria. Apart
from an unsigned article in one medical encyclopedia, which Freud
almost certainly wrote (Freud, 1888), there are only items of minor
interest. Assuming that the article is Freud’s, it shows that as late as
1888 Freud had not departed significantly from the position taken in
his report: hysteria is still described as an orderly disease accurately
characterized by Charcot’s stigmata; there is no indication that Freud
had become interested in an etiological account of hysteria; the causes
of the disease are still the familiar predisposing and exciting factors
(Freud, 1888, pp. 50f.). Perhaps the most significant difference between
Freud’s preliminary report and the 1888 article is that the article reveals
a greater familiarity with the literature on hysteria.
Freud’s publications in 1893 – the year following the publications by
Möbius and Strümpell considered above – depart, for the first time,
from Charcot’s symptomatic characterization of hysteria. Freud’s trans-
lation of Charcot’s lectures, which was published in that year, contains
a series of footnotes clearly indicating the new direction of his thought.5
In one note, he objects that Charcot’s etiology did not separate the
disposition to neuroses from the disposition to organic nervous disor-
ders (Charcot, 1892–95, 1:237n.6). Given a purely symptomatic concept
of the neuroses there would be no reason to expect etiology to make
this separation. Indeed such a distinction would be impossible if, as
most people assumed, symptomatically-defined functional disorders could
be caused by the same factors that, on other occasions, caused related
organic diseases. After adopting symptomatic definitions of hysteria,
various writers explicitly denied that hysteria could be causally distin-
guished from other organic or functional disorders. One would expect
etiology to reflect this distinction only if one assumed that distinguish-
able diseases must have distinguishable causes – that is, only if one
assumed the Distinguishability Hypothesis which we encountered in the
bacterial theory. In fact, in an 1892 publication, Möbius rejected as
nonsense the idea that different diseases could have the same cause
(Möbius, 1892, p. 290). But this would be nonsense only to those who
viewed diseases from the etiological standpoint. At the time he was
translating Charcot’s lectures, Freud knew Möbius’s publication and, in
a footnote, recommended it to Charcot’s readers (1:149n). In the same
publication, Möbius observed (as Strümpell had before him) that rede-
fining the nervous diseases in causal terms would entail reclassifying
them. Several of Freud’s footnotes express objections to Charcot’s scheme
for classifying the nervous disorders, the so-called famille neuropathique.
In one note Freud explained that his objections to Charcot’s scheme
were at least partially the result of his work on the etiology of tabes
(1:8n). In another note, he observed that his own theory of ‘hysterical
counterwill’, connected together various hysterical symptoms and thereby
cast light on the mechanism of the hysterical condition (1:137n). At
about the same time, in a preliminary draft for a subsequent joint
publication, Freud and Josef Breuer objected that Charcot had only
described hysteria, and that ‘this description throws no light at all on
any connection there may be between the different phases, on the
significance of attacks in the general picture of hysteria, or on the way
in which attacks are modified in individual patients’ (Freud, 1893,
p. 151). Strümpell and Möbius had pointed out that the etiological
definitions of the bacterial theory had shed light on just these factors in
the infectious diseases.
Thus, Freud’s writings from 1893 and 1894 show he was moving
away from Charcot’s symptomatic treatment of hysteria and toward an
etiological approach. It has been universally recognized that Freud be-
gan criticizing Charcot in 1893, but the significance of the criticism has
been generally overlooked. Ernst Jones wrote: ‘What Freud maintained
as the result of his observations was that, whenever a thorough investi-
gation of the patient could be carried out, sexual etiological factors
would be found which were different in [hysteria and the anxiety neu-
roses]; this was his justification for separating them’ (Jones, 1953, vol.
1, p. 256). Jones mentions Möbius only as one ‘from whom … Freud
could have derived but very little’ (Jones, 1953, vol. 1, p. 370). There is
no indication that Jones sees any novelty or special importance in this
new strategy, and Freud’s other commentators generally overlook the
change altogether.7 Yet these steps, which Möbius and Strümpell had
advocated, fundamentally severed Freud’s work from the ideas of most
of his other predecessors; he thereby adopted an orientation, never to
be abandoned, that brought his work on psychopathology into har-
mony with the prevailing orientation of medical research in his time.
These facts can be ignored only at the cost of failing to understand the
true nature of Freud’s contribution.
We must now consider Freud’s work in the few years following 1893.
It is unnecessary to follow the evolution of Freud’s thought or to sum-
marize his ultimate views. Our goal is only to trace Freud’s quest for
etiological characterizations of the nervous disorders, especially hyste-
ria, and his use of those characterizations to provide explanations that
were exactly analogous to the explanations that were, at the same time,
being derived from the etiological definitions of the infectious diseases.
To accomplish this it will be necessary only to review certain prominent
themes in Freud’s writings through 1896, the year in which Freud
published both ‘Heredity and the Etiology of the Neuroses’ and ‘The
Etiology of Hysteria’.
In ‘Heredity and the Etiology of the Neuroses’ Freud asks: ‘Is it
possible to establish a constant etiological relation between a particular
cause and a particular neurotic effect, in such a way that each of the
major neuroses can be attributed to a specific etiology?’ (Freud, 1896a,
p. 149). His answer is that each neurosis ‘has as its immediate cause one
particular disturbance of the economics of the nervous system’ and, in
particular, disturbances of ‘the subject’s sexual life, whether they lie in a
disorder of his contemporary sexual life or in important events in his
past life’. After considering the specific causes of some of the other
neuroses, he writes: ‘A passive sexual experience before puberty: this,

then, is the specific etiology of hysteria’ (Freud, 1896a, p. 152). In
‘Further Remarks on the Neuro-Psychoses of Defence’ he writes:
In order to cause hysteria, it is not enough that there should occur
… an event that touches [the subject’s] sexual existence and be-
comes pathogenic through the release and suppression of a
distressing affect. On the contrary, these sexual traumas must have
occurred in early childhood (before puberty), and their content
must consist of an actual irritation of the genitals (of processes
resembling copulation).
(Freud, 1896b, p. 163)
And finally, in ‘The Etiology of Hysteria’:
I therefore put forward the thesis that at the bottom of every case
of hysteria there are one or more occurrences of premature sexual
experience, occurrences which belong to the earliest years of child-
hood but which can be reproduced through the work of psycho-
analysis in spite of the intervening decades. I believe that this is an
important finding, the discovery of a caput Nili in neuropathology.
(Freud, 1896c, p. 203)
How are such passages to be understood? Freud sometimes suggests
that these claims are simply empirical discoveries from clinical observa-
tion (Strachey, 1955–74, vol. 3, pp. 52, 99). Indeed, it is possible that
the theses originated in just that way. However, their logical role in
Freud’s thought is not that of simple empirical generalizations.
Freud started out believing that Charcot’s symptomatic characteriza-
tion of hysteria was essentially correct. Perhaps for this reason Freud
presented his etiological account as a discovery based on Charcot’s
symptomatic definition. However, there are indications that, at an early
stage, Freud regarded the etiological discovery as more fundamental
than a simple empirical generalization: In his An Autobiographical
Study, he explained that Breuer’s discoveries in the treatment of Anna
O. ‘seemed to me to be of so fundamental a nature that I could not
believe it could fail to be present in any case of hysteria if it had been
proved to occur in a single one’ (Freud, 1925, p. 21). In letters to Fliess
written in 1892 and 1893 – the first years in which he departed from
Charcot’s symptomatic characterization and only four years after the
entirely orthodox article in the encyclopedia – Freud insisted that ‘no
neurasthenia or analogous neurosis can exist without a disturbance of
the sexual function’, and ‘the contention which I am putting forward
and desire to test by observations is that neurasthenia is always only a
sexual neurosis’ (Freud, 1954, pp. 65f.).
In ‘The Neuro-psychoses of Defence’, Freud identified two ‘extreme
forms of hysteria’ that did not conform to a characterization of hysteria
given by one famous French researcher, Pierre Janet. Both forms are
defined etiologically (Freud, 1894, pp. 46f.). Similarly, in a long paper
on anxiety neuroses, Freud distinguished six forms of neuroses in women,
four in men, and two found in both sexes; all 12 forms were defined
etiologically (Freud, 1895). It is clear that Freud used his etiological
account of the nervous disorders to generate a nosology more coherent,
rational, and precise than had been possible before.
By 1896, the sexual etiology of hysteria had become definitional.
Freud repeatedly exploited the tactic of redefining diseases in terms of
causes. In ‘Heredity and the Etiology of the Neuroses’, he set forth a
‘nosographic innovation’ resulting from his researches into the etiology
of the major neuroses (Freud, 1896a, p. 146). His innovation was a
fourfold scheme in which each specific neurosis was attributed to a
particular abnormality in the subject’s sexual behavior. ‘What gives its
distinctive character to my line of approach,’ he wrote, ‘is that I elevate
these sexual influences to the rank of specific causes, that I recognize
their action in every case of neurosis, and finally that I trace a regular
parallelism, a proof of a special etiological relation between the nature
of the sexual influence and the pathological species of the neurosis’
(Freud, 1896a, p. 149). In the next few pages, Freud discussed neuras-
thenia, other anxiety neuroses, hysteria, and obsessional neuroses; each
of these was differentiated from the others by the specific etiology of the
symptoms.
However, as Freud saw, the etiological definitions and nosological
innovations were not an end in themselves. In an early draft of their
book on hysteria, Freud and Breuer objected that Charcot’s account
explained virtually nothing about the disease (Freud, 1893, p. 151).
By contrast, in his own writings, Freud used the etiological account of
the nervous disorders to explain an impressive variety of phenomena
including certain hysterical symptoms (Strachey, 1955–74, vol. 3, p.
348), the incidence of hysteria (p. 153) and the hysterogenic zones (p.
163), the response of hysterics to hypnosis (p. 59), certain similarities
among the neuroses (p. 99), patterns of incidence of anxiety neuroses
among married couples (p. 101), neurasthenia occurring in some cases
of sexual abuse (p. 113), the suppression of those events that cause
specific cases of hysteria (p. 154), the predominance of hysteria among
women and of obsessional neurosis among men (p. 169), the apparent
familial neurotic disposition and various pathological symptoms (p.
209), habits (p. 169), and phobias (pp. 130f.), the course of develop-
ment of obsessional neuroses (p. 146), the success and failure of
various therapeutic measures (p. 195), the rare occurrence of hysteria
in the lower social orders (p. 211) and much much more. Moreover,
Freud considered observed facts that could not be explained as possi-
ble weaknesses in his theory (Charcot, 1892–95, vol. 1, p. 314n.). In

one discussion of organic lesions, Freud commented that the physi-
cian’s representation of the causes and alterations of these lesions
must be right ‘since they allowed him to understand the details of the
illness’ (Freud, 1909, pp. 11f.). By 1896 he wrote that ‘the symptoms
of hysteria can only be understood if they are traced back’ to etiological
factors (Freud, 1896b, p. 163). In a lecture of the same year he
asserted, ‘In the sole attempt to explain the physiological and psychi-
cal mechanism of hysteria which I have been able to make in order to
correlate my observations, I have come to regard the participation of
sexual motive forces as an indispensable premise’ (Freud, 1896c, p.
200). Thus, at least by 1896, Freud found the explanatory force of his
etiological account most compelling.
Because the theoretical advantages of his new approach were so
apparent, Freud persisted even in the face of apparently incompatible
clinical evidence. So far as Freud knew at this time, the case of Anna O.
was an exception to his theory (Freud, 1925, p. 26), and there were
other likely exceptions as well (Breuer and Freud, 1893, p. 14). In an
1893 letter to Fliess, Freud admitted that it required courage to insist on
his etiological theories in the face of intractable clinical evidence, and,
in another letter, he confessed that ‘the connection between obsessional
neurosis and sexuality does not always lie so near the surface … if it
had been sought for by anyone less obstinately wedded to the idea, it
would have been overlooked’ (Freud, 1954, pp. 78, 81).
We now see striking parallels between Freud’s post-1893 approach to
psychopathology and work that was being done at about the same time
on the infectious diseases. At least initially, Freud and Breuer saw their
work as closely associated with the positions of Möbius and Strümpell
(Strachey, 1955–74, vol. 2, pp. 7f., 187, 215; vol. 3, p. 51), and both
Möbius and Strümpell, in turn, saw their own work as modelled on the
bacterial theory. Indeed, Möbius and Strümpell explicitly set out to do
for the nervous disorders what had been accomplished in the infectious
diseases by using an etiological approach. Freud was less candid about
the ultimate source of the strategy he followed, but he used the contem-
porary infectious account of tuberculosis as an analogy in explaining
and justifying some aspects of his views about the causes of anxiety
neuroses (Strachey, 1955–74, vol. 2, p. 187; vol. 3, pp. 137, 209) and
certain of the metaphors chosen by Freud and Breuer also suggest that
they were aware of the connection between their work and the bacterial
theory (Carter, 1980, p. 273n.). In any event, whether intentionally or
not, Freud’s work on psychopathology ended up exactly in harmony
with the main orientation of the medical research of his time. At the
very least, given the successes of bacterial theory, this fact must have
made his account more appealing than other simultaneous accounts

that were not etiological and did not have the same explanatory force.
Why have Freud’s commentators ignored the parallels between his early
work and simultaneous research on the bacterial diseases? In part, this
may reflect a simple lack of historical perspective. We have become so
accustomed to etiological characterizations and classifications that Freud’s
etiological approach may appear to require no special consideration. In
his time, however, the situation was different: this approach was still
relatively new, even in accounting for the infectious diseases. Until 1884
no one seems to have envisioned characterizing the nervous disorders
etiologically, and until the end of the 1880s no one actually tried to do
so. Freud was certainly the first to use this approach to provide coher-
ent explanations for the nervous diseases.
In recent years, Freud’s work has been eulogized as revolutionary – as
the introduction of a new paradigm in science (Mujeeb-ur-Rahman,
1977). Freud’s admirers (beginning with Freud himself) have compared
him with Darwin and Copernicus.9 Viewed in relation to any of his
recognized sources, his work does seem revolutionary and without
precedent. Charcot – like most other late nineteenth-century physicians
who dealt with nervous disorders – started with symptoms and ended
up with total chaos in the discussion of causes. As a result there were no
coherent explanations of anything. By starting with causes, Freud was
able to explain the symptoms as well as many other facets of the
nervous diseases – and, ultimately, everything from jokes and dreams to
spelling errors. In the context of his recognized sources, this was indeed
a revolution. However, if we inquire into the nature of this revolution,
and if we view Freud against the background of nineteenth-century
medicine, from which his thinking emerged, his work takes on the
appearance of an ingenious application of a strategy that was already
being employed with enormous success in other areas – a strategy Freud
applied with considerable ingenuity but did not create. From this point
of view, therefore, we must be more cautious about describing Freud as
a paradigm initiator or scientific revolutionary. These facts, too, may
relate to his commentators’ inability to see what must be among the
crucial factors that guided his work and helped to secure its acceptance.
It now appears that the model of comprehension exemplified by the
bacterial theory underlies our way of thinking about other classes of
diseases that have nothing to do with bacteria. The current pervasive-
ness of this way of thinking is illustrated by the fact that none of Freud’s
commentators sees any change at all when Freud completely reversed
himself and adopted it. Thus, while taking account of the relation
between Freud’s work and the etiological standpoint may challenge his
role as a paradigm initiator, by so much the more does it confirm, in
that role, those who first adopted the basic principles of the etiological
standpoint.
Notes
1. Hannah S. Decker (Decker, 1977), Ernest Jones (Jones, 1953) and Ola
Andersson (Andersson, 1962) also ignore the bacterial theory in their
discussions of the origins of Freud’s doctrines.
2. Through this paragraph, references such as this one are volume and pages
references to Charcot (1892–95).
3. Under heredity Charcot included such factors as a father abandoning his
family (2:6), alcoholic aunts (2:227), and a suicidal father (2:319) in
addition to hysteric, epileptic, or nervous relatives. Even Freud objected
when Charcot called an arthritic tendency in relatives a hereditary neuro-
pathic disposition (1:237).
4. This definition presupposes what we could call an Ideational Hypothesis.
It implicitly identifies a domain of causes for a new theory of disease, and
in this respect it is parallel to the Bacterial Hypothesis discussed in earlier
chapters.
5. For an enlightening study of ‘Freud’s rebellion’ from Charcot’s point of
view, see Gelfand (1989).
6. Through this paragraph, references such as this one are volume and page
references to Charcot (1892–95).
7. Ola Andersson writes ‘In the late 1880s and 1890s P. J. Möbius and A. v.
Strümpell had published papers on the traumatic neuroses and on hysteria
in which they espoused views very similar to those of Charcot [!]’
(Andersson, 1962, p. 115).
8. Through this paragraph, page numbers in parentheses are references to
volume 3 of Strachey (1955–74).
9. For a review and references see, Nigel D. Walker, ‘A New Copernicus?’
and especially David Shakow and David Rapaport, ‘Darwin and Freud: a
Comparison of Receptions’; both reprinted in Mujeeb-ur-Rahman (1977,
pp. 35–42 and 43–63).
CHAPTER TEN
Protozoal and
Viral Theories of Disease
A research programme ‘consists of methodological rules: some tell us
what paths of research to avoid … , and others what paths to pursue’
(Lakatos, 1968, p. 132). The result of following such rules is a sequence
of theories, ‘T1, T2, T3 … where each subsequent theory results from
adding auxiliary clauses to … the previous theory in order to accommo-
date some anomaly’ (Lakatos, 1968, p. 118). By technical innovations
and dedicated research, nineteenth-century scientists successfully ex-
plained an impressive array of bacterial diseases; each success could be
regarded as yielding a new member in a Lakatosian sequence of bacte-
rial theories of disease. A set of basic assumptions such as Klebs’
Grundversuche was common to all such theories.
However, as medical research continued, various obviously conta-
gious diseases became prominent precisely because, while superficially
similar to known bacterial diseases, their causal agents eluded every
attempt at identification. In 1894, William Henry Welch observed:
We have a large number of infectious diseases which have thus far
resisted all efforts to discover their specific infectious agents. Here
belong yellow fever, typhus fever, dengue, mumps, rabies, Oriental
pest, whooping cough, smallpox and other exanthematous fevers,
syphilis, and some other infectious diseases in human beings. It will
be noted that many of these are the most typically contagious
diseases, which it might have been supposed would be the first to
unlock their secrets.
(Hughes, 1977, p. 25)
One way of explaining these failures is to say that the diseases in
question were caused by such entities as protozoa, viruses, or rickettsia
that were different from the bacteria studied by classical bacteriologists.
However, this explanation suggests that discovering new causes is like
finding previously overlooked shells on a beach – it obscures the theo-
retical innovations required in recognizing new kinds of specific causes.
A theory of disease must specify a domain of entities over which the
theory is to be applied and it must include causal criteria that indicate,
in part, how the word ‘cause’ is to be used within the theory. Given
such a theory and some new illness, there may simply be nothing within
the domain of the theory that satisfies the criteria with respect to that
PROTOZOAL AND VIRAL THEORIES OF DISEASE 163
illness. One cannot expect always to find missing causes simply by

looking more carefully – as Davaine’s work illustrates, in the absence of
an appropriate theory, no amount of empirical evidence may be suffi-
cient to establish causal relations. With respect to the diseases mentioned
by Welch, part of what was needed was the creation of new theories
defined over new domains of possible causes and adoption of new
criteria that could be satisfied by those entities.
Driven by a shared commitment to the etiological research programme,
late nineteenth-century scientists worked toward meeting these require-
ments. In Lakatos’s terminology, we can think of this work as the
creation of new research programmes subordinate to the overriding
programme that we have called the etiological standpoint. In this chap-
ter we will consider two such subordinate programmes – one focusing
on protozoa and the other on viruses.
A protozoal theory of disease
Malaria, also known as intermittent fever, has been prominent in Eu-

rope for millennia, and there were serious attempts to assimilate it to
the bacterial theory. In 1879, Edwin Klebs and Corrado Tommasi-
Crudeli described a bacterium they had isolated from swamp-ooze and
from the urine of malarial patients. When these bacteria were cultured
and injected into rabbits, the rabbits died with symptoms resembling
malaria. Upon dissection, the rabbits were found to have spleens that
were enlarged and contained black pigment similar to that found in the
spleens of human victims of malaria (Klebs and Tommasi-Crudeli, 1879b,
p. 125). Klebs and Tommasi-Crudeli named the organism Bacillus ma-
lariae. They seemed to have met all the conditions for proving causation,
and other researchers confirmed their results. Their account of the
disease was widely accepted – in 1881 Giuseppe Cuboni and Ettore
Marchiafava began a paper by observing: ‘knowledge of the parasitic
nature of malaria has received the most brilliant confirmation in the
recent work of Klebs and Tommasi-Crudeli’ (Cuboni and Marchiafava,
1881, p. 265).
On 6 November 1880, a French physician named Charles Louis Alphonse

Laveran, noticed unusual ‘elements’ of three kinds in blood drawn from
a young soldier suffering from malaria. Laveran regarded the elements
as parasitic and suspected they were causally involved in the disease. He
communicated his discovery to the Academy of Medicine in Paris (Colin,
1880a). Laveran’s report was presented by Léon Colin, but Colin him-
self was skeptical about Laveran’s conclusions. He did not believe the
elements were parasites, partly because malaria was not contagious; he
recommended a commission be established to appraise the validity of
future communications. One month later, Colin presented another pa-
per on Laveran’s behalf, and this time Colin was even more skeptical.
He suggested that what Laveran took to be parasites were actually only
leukocytes that had ingested pigment (Colin, 1880b).
In October 1881, Laveran himself presented a paper before the Acad-
emy of Sciences. Laveran reported that
the blood of malaria patients contains parasitic elements which
present themselves under the following aspects: (1) cylindrical ele-
ments, tapered at their extremes, almost always curved into a
crescent. … (2) transparent spherical elements containing movable
grains of pigment … on whose boundaries one can often observe
very fine filaments, [the spherical elements] are animated, in every
sense, with a very rapid movement. … (3) spherical or irregularly
shaped transparent elements that contain pigment [but with prop-
erties] that enable one to distinguish them clearly from pigmented
leukocytes. (4) spherical transparent pigmented elements smaller
than those mentioned above.
(Laveran, 1881, pp. 627f.)
According to Laveran ‘the spherical bodies containing mobile grains of
pigment and armed with actively mobile peripheral filaments are un-
doubtedly alive’ (Laveran, 1881, p. 629). By the time he presented this
paper, Laveran had examined the blood of 192 victims of different
forms of malaria and had identified the parasites in blood samples from
148 of them. Most of the patients in whose blood the elements could
not be found had been treated with quinine, and, as Laveran confirmed,
quinine killed the parasites. He noted that parasitic elements were only
found in the blood of malaria patients and that the parasites were most
likely to be present just before the onset of fever. Laveran believed that,
between attacks, the parasites were concealed in such organs as the
spleen and liver. He concluded that ‘attacks of malaria are caused by
the introduction, into the blood, of parasitic elements that present
themselves under the aspects described above; sulfate of quinine leads
to the termination of malarial attacks because it kills these parasites’
(Laveran, 1881, p. 630).
Laveran’s conclusions were challenged both by those who still ques-
tioned the role of living parasites in any disease and by those who
believed malaria was caused by a bacterium. However, M. Richard, a
physician who was one of Laveran’s friends and who was also stationed
in a hospital in Algeria, confirmed Laveran’s findings, and suggested
that the presence of the parasites could be taken as a diagnostic crite-
rion for malaria. This, in effect, meant recharacterizing the disease in
terms of the presence of the parasites. Other researchers also confirmed

Laveran’s findings.
In 1882, Laveran travelled to the Roman Campagna where malaria
was endemic. He confirmed his earlier discoveries and demonstrated his
results to the leading experts in Rome, but the Italians remained skeptical.
They proposed that the supposed parasites were only degenerating
blood cells. Many observers continued to believe that ‘algae or
“microphytes” present in marshy waters, were the true pathogens and
particular attention was given to the claims of Klebs and Tommasi-
Crudeli’ (Bruce-Chwatt, 1988, p. 24). In 1883, Marchiafava and A.
Celli, who had witnessed Laveran’s presentations in Rome, published a
paper describing bodies similar to those Laveran had exhibited (al-
though without mentioning Laveran or his work). They pointed out
that some of the bodies resembled micrococci, and admitted, guardedly,
that the bodies could be parasitic (Marchiafava and Celli, 1883, p.
574). However, the authors favored a different hypothesis: in a letter to
Laveran, Marchiafava wrote, ‘We believe that the pigmented forms
which you have described are nothing but degenerated and pigmented
red cells’ (Bloomfield, 1958, p. 350). In his account of these events,
Arthur L. Bloomfield observed
It would be interesting if one could reconstruct the medical politics
of the time. Klebs and Tommasi-Crudeli were big names. The latter
was head of the Bacteriologic Institute in Rome; Marchiafava
worked in the same institution. There must have been great pres-
sure to take sides in this controversy with a relatively insignificant
doctor from Val-De-Grace.
(Bloomfield, 1958, p. 350)
Meanwhile, Laveran found a more sympathetic audience in Paris when
he demonstrated the malarial organisms to Pasteur and Roux.
In 1884, C. Gerhardt published the results of experimental inocula-
tions with malarial blood (Gerhardt, 1884). From observations about
the spread of the disease, Gerhardt inferred that malaria must be due to
a ‘poison’ that multiplies within the human body and is carried by a
lower organism. If so, he reasoned, it should be possible to convey the
disease by the inoculation of infected blood. He mentioned that
Marchiafava and Cuboni had inoculated dogs with blood from malarial
patients and that other researchers had conducted inoculation experi-
ments on human volunteers, yet these experiments had produced only
equivocal results. ‘The conclusive experiment,’ he wrote, ‘appeared to
be the inoculation of blood from a malarial patient to a healthy person
… using blood drawn during an attack of fever’ (Gerhardt, 1884, p.
373). Gerhardt also listed four conditions that, he felt, should be satis-
fied in performing the test: (1) the experiment must be performed in a
region free from malaria, (2) the patient who supplies the blood must
have no other infectious diseases, (3) the subjects to be inoculated must
volunteer and must not be required, by the experiment, to sacrifice an
opportunity for gainful employment, and (4) the subjects’ temperature
curves must have been recorded since long before the inoculation to
insure that they were free from fever (Gerhardt, 1884, pp. 373f.).
Gerhardt was able to induce artificial malaria in two volunteers who
met these conditions and, subsequently, to control both cases by admin-
istering quinine. Gerhardt did not mention Laveran’s work, and he did
not report any microscopic examinations of the subjects’ blood before
or after the inoculations.
By 1885, Marchiafava and Celli had been converted to the idea that
malaria was parasitic. They cited Laveran, Richard, and Gerhardt and
reported further microscopic studies of malarial blood (Marchiafava
and Celli, 1885a). They described an attempt to transmit malaria by
injecting a volunteer with blood freshly drawn from a malarial patient;
the injection produced typical fever symptoms. Marchiafava and Celli
identified the characteristic organisms in the blood of the volunteer, and
controlled the disease by administering quinine. ‘All of these condi-
tions,’ they concluded, ‘make probable our hypothesis that these forms
are living agents’ (Marchiafava and Celli, 1885a, p. 353). Marchiafava
and Celli also reported attempts to cultivate the parasites in various
media. The attempts failed, but they reported finding an organism in
mud from swamps that, when cultured, appeared ‘morphologically simi-
lar to the initial forms’ of the supposed malarial parasite, and, they
reported, ‘the multiplication of these bodies was especially obvious in
slide cultures of malarial blood’ (Marchiafava and Celli, 1885a, p.
354). According to Bloomfield, this comment reveals ‘that Marchiafava
and Celli were still unclear about the significance of their findings’; he
concludes that they ‘did little more than confirm Laveran’s findings’
(Bloomfield, 1958, pp. 350f.). However, their paper had this heuristic
significance: for the first time, it brought together observational reports
of the parasite and Gerhardt’s reports of successful inoculation experi-
ments on human volunteers.
By the end of the nineteenth century, it was becoming progressively
more clear that Laveran’s protozoan played a role in malaria. Moreo-
ver, the study of malaria had been supplemented by research on other
diseases that also appeared to be caused by protozoa. In 1890, Koch
summarized the situation as follows:
In many areas … indeed precisely where one would least have
expected it, bacteriological investigation has been of no avail. This
has been the case in the study of several infectious diseases which,
because of their pronounced infectiousness, appeared to be easy
objects of investigation. … I suspect that these diseases involve

organized disease agents that are not bacteria but rather belong to
completely different groups of microorganisms. This opinion is all
the more justified by the recent discovery that the blood of many
animals – as for example the blood of malaria victims – contains
the unique animal parasites known as protozoa. Of course, as yet
one can do no more than demonstrate the presence of these note-
worthy and important parasites. Apparently, no further progress
will be made until protozoa, like bacteria, can be grown in an
artificial medium or under natural conditions outside the body …
Once this is done, and there is no reason to doubt that it will be,
the investigation of pathological protozoa and related microorgan-
isms will probably become part of bacteriology. Hopefully this
investigation will explain the etiology of these as yet mysterious
diseases.
(Koch, 1890, p. 184)
Koch underestimated the changes involved in accepting non-bacterial
parasites as causes – there was no reason to think that they could satisfy
causal criteria (for example, flourishing in artificial media) that had
been developed to accommodate bacteria or that they could be studied
by the same techniques used to study bacteria (thereby assimilating
their study to bacteriology). However, scientists became persuaded that
malaria was caused by Laveran’s protozoan, and research on malaria
attracted so much attention that, for a time, it was believed that all non-
bacterial diseases may be caused by protozoa (Joest, 1902, p. 380).
Within a few years, it became clear that, while malaria, Texas fever, and
a few other diseases were protozoal, many common non-bacterial dis-
eases definitely were not.1
A viral theory of disease
In 1879, Adolf Eduard Mayer, a bacteriologist at a Dutch experimental

agriculture station, began studying a disorder that affected local to-
bacco plants. He reported his research in a remarkable paper published
seven years later (Mayer, 1886). The tobacco disease was not widely
known in Europe, and where it was known, it was called by different
names. In the absence of a standard nomenclature, Mayer called it
tobacco mosaic disease after the mottled appearance of the leaves of
diseased plants. According to Mayer, the many causes to which the
disease had been ascribed constitute ‘a virtual chaos, likely to induce
dizziness, and useful only to confirm the old idea that humans cannot
exist without theories’ (Mayer, 1886, pp. 453f.2). In harmony with
earlier beliefs about disease causation generally, various meteorological
conditions as well as planting and fertilizing practices had been men-
tioned as possible causes. Mayer wrote: ‘Many regard the disease as

entirely inexplicable – as a kind of witchcraft, and have pressed on me
the warning: “You will never find it[s cause]! Never!”’ (p. 454).
Mayer analyzed diseased and healthy plants and the soils in which
they were cultivated, and patiently excluded dozens of purported causes.
In reporting this research, he warned his reader against becoming so
bored with the negative results as to be inattentive and thereby miss the
positive conclusions that would follow (p. 455). He pointed out that, by
showing the inadequacy of the many purported general causes, he had
increased the likelihood that the disease was due to a specific agent – an
argument like one used by Agostino Bassi nearly a century earlier.
Mayer explained that, while ruling out many purported general causes,
he and his colleagues had also examined diseased plants for fungi and
animal parasites. At first, they found nothing. ‘Then,’ he wrote, ‘I
suddenly discovered that the sap that one could rub out of diseased
plants was a reliable infectious material for healthy plants’ (p. 461).
According to Mayer, if one applied a small quantity of diluted sap to an
incision in the stem of healthy plants, those plants usually became
diseased, whereas sap from healthy plants had no effect. Mayer redou-
bled his efforts to discover some harmful ‘body’ within the sap of
diseased plants but found nothing: ‘I tried to culture the supposed
organisms using the methods of Koch and others; in many cases I
obtained bacterial vegetation, but, when used as inoculation material,
none of this vegetation infected healthy tobacco plants’ (p. 463). Mayer
tried inoculating various bacteria and substances known to contain
bacteria but all his attempts were without effect.
Mayer then tried to decide whether the success of the inoculations
was due to an ‘unformed or to a formed ferment’ (p. 464). He admitted
that a ferment of the first kind would be ‘unusual as the cause of a
disease’ and that no ‘enzyme’ was known that had the capacity of
‘multiplying itself’ within a plant as seemed to happen in his inoculation
experiments. However, ‘a formed ferment could be either a fungus or a
bacterium, and, elements of these groups can be distinguished mechani-
cally [for example, by filtration] as well as by the microscope’ (p. 464),
and he resolved to try filtration. Mayer found that infectious agents
passed through one sheet of filter paper although filtration seemed to
reduce the concentration of causal agents in diluted sap. He inferred
(incorrectly, as critics pointed out) that the agent could not be a fungus
since all fungi would be stopped by the filter. He then found (also
incorrectly, according to critics) that the infectious agent would not
pass through two sheets of filter paper, and he concluded that the agent
could not be an ‘enzyme’ since any such substance would pass through
both sheets. Learning at the same time that infectious sap could be
made harmless by heating it to 80 degrees, he concluded that the

infectious substance was an organic ferment – a bacterium (p. 465).
Mayer felt he had learned everything one could learn from filtration
experiments: ‘Further knowledge of the form and way of life of the
guilty bacterium cannot be expected to follow from experiments of this
kind, and must await future research’ (p. 466).
A Russian student, Dimitri Iosifovitch Ivanovski, continued Mayer’s
work (Ivanovski, 1892). Ivanovski discovered that Mayer was wrong in
thinking that the cause of tobacco mosaic disease could be removed
from sap by two or more sheets of filter paper. Because of this realiza-
tion, Ivanovski is often mentioned in connection with the discovery of
filterable viruses. However, he himself believed his main contribution
was recognizing that the disease Mayer studied actually included two
independent disorders. In summarizing the relation of his work to
Mayer’s, he completely ignored the filtration experiments (Ivanovski,
1902), and he showed little interest in filtration until later researchers
revealed its significance (at which time he claimed priority for its use).
Like Mayer, Ivanovski believed tobacco mosaic disease was caused by a
bacterium, and, while he entertained the possibility that the causal
bacterium was filterable, he found it more likely that the bacterium
generated a soluble poison that made sap infective (Ivanovski, 1892, p.
69, 1903, pp. 202f). He conducted no serial inoculations and never
suggested that the substance that passed through the filters was itself
capable of multiplying in inoculated plants. Moreover, his research
‘does not appear to have had an effect on the development of the
concept of the virus. His papers of 1892 had been published only in
Russian journals and his work was virtually unknown in the West
before 1899’ (Hughes, 1977, p. 56). By then, other researchers had
explicitly postulated the existence of discrete non-bacterial causal agents
that replicated and were filterable.
Another scientist whose research was based on Mayer’s work was
Marinus Willem Beijerinck. Beijerinck was an original thinker and made
significant contributions in several areas (Wilkinson, 1976, p. 117).
Mayer and Beijerinck worked together at the Wageningen agricultural
school, and Mayer showed Beijerinck his experiments with the sap of
diseased tobacco plants. Over the next few years, Beijerinck occasion-
ally looked for parasites in the sap of diseased plants, but he found
none. For a time he suspected that the disease may be caused by anaero-
bic organisms, but he later saw that this idea was incorrect. Early on
Beijerinck conducted diffusion experiments from which he concluded
that the infective substance, which he referred to as a virus, ‘must be
regarded as a liquid or as dissolved, and not corpuscular’ (Beijerinck,
1898, p. 6). He conducted serial inoculations from plant to plant. Given
the small quantity of diluted sap that was sufficient to infect healthy
plants, his serial inoculations persuaded him that ‘the contagium, al-
though liquid, replicates itself in the living plant’ (Beijerinck, 1898, p.
5). Beijerinck used the term ‘contagium vivum fluidum’ to refer to this
unusual replicating yet liquid disease agent whose existence seemed to
be implied by his experiments.
Beijerinck made the remarkable observation that injections of infec-
tive sap did not affect mature leaves, but only new growing buds and
shoots in which cell division was in progress. He astutely inferred that
the contagium vivum fluidum, being ‘unable to grow independently, is
drawn into the growth of the dividing cells and multiplies powerfully
therein but without losing its own individuality’ (Beijerinck, 1898, p.
9). Of course, contemporary scientists who were studying viral diseases
in animals found no counterpart to these discoveries, and Beijerinck’s
observation and the inference he drew from it were inadequately appre-
ciated. In a sequel to this paper, Beijerinck pointed out that the
impossibility of culturing the contagious substance in artificial media
suggested it was a liquid that could not multiply outside living proto-
plasm. However, he admitted that this concept was not easy to
comprehend:
The reproduction or growth of a dissolved body is not unthinkable
but it is difficult to imagine. It is not easy to accept a process of
division in the molecules which would lead to their multiplication,
and the idea of self-sustaining molecules, which is thereby presup-
posed, seems unclear to me, if not contrary to nature. It is to some
extent an explanation to consider that, in order to reproduce itself,
the contagium must be incorporated into the living protoplasm of
the cell into whose multiplication it is, so to speak, passively in-
cluded. … However, there is no denying that even if the incorporation
of a virus into the living protoplasm is confirmed as factual, it
cannot be regarded as a clearly comprehensible process.
(Beijerinck, 1899, p. 31)
Lise Wilkinson suggested that Beijerinck’s early interest in chemistry
may have ‘enabled him in later life to consider chemical molecules and
even their possible role in biological systems in a more realistic way
than could most of his fellow biologists and pathologists’ (Wilkinson,
1976, p. 117).
By the time Beijerinck published his first work on tobacco mosaic
disease, Friedrich Johannes Loeffler and Paul Frosch had begun pub-
lishing studies of foot-and-mouth disease (Loeffler and Frosch, 1898).
Beijerinck knew of their work, and believed that Loeffler and Frosch
were dealing with causal factors related to those he was studying, but
he wrote that he could not support their conclusions ‘in respect to the
corpuscular nature of the virus’ (Beijerinck, 1898, p. 7n). While Beijerinck
proposed that other plant diseases may be due to other contagium

vivum fluidia, he did not attempt to explain foot-and-mouth disease or
any other animal disorders.
In respect to the etiological research programme, the most important
early work on what are now called viruses was reports by Loeffler and
Frosch on foot-and-mouth disease. Mayer, Ivanovski, and Beijerinck
never doubted that tobacco mosaic disease had a specific cause; indeed,
they all began (and Mayer and Ivanovski ended up) believing that the
cause in question was a bacterium. Their work fostered the growth of
an important ramification of the research programme into plant pathol-
ogy – a ramification we will not pursue further. However, in respect to
subsequent developments in medicine, the studies of Loeffler and Frosch
were more important. In 1908, John McFadyean, an English bacteriolo-
gist, wrote that discovering the filterability of the causal agent of
foot-and-mouth disease ‘at once attracted general attention, and gave a
great impetus to the investigation of the nature of the virus in those
diseases which had hitherto baffled investigation conducted on ordinary
bacteriological lines’ (McFadyean, 1908, p. 66).
As early as 1840 Jakob Henle assumed, as common knowledge, that

foot-and-mouth disease was contagious, and it was among the diseases
for which one would have expected an explanation in terms of parasitic
organisms. Both Loeffler and Frosch were bacteriologists who had been
trained by Koch, and who, like Mayer, Ivanovski, and Beijerinck, began
their work using standard bacteriological techniques.
Loeffler and Frosch submitted three reports to the German Ministry
of Culture (17 April 1897, 14 August 1897, and 8 January 1898); the
reports were published in 1898. By the time they conducted their work,
technical developments had raised bacteriology to a level from which it
was apparent that foot-and-mouth disease was not due to an ordinary
bacterium: no such organism was found in microscopic studies of stained
or unstained tissues and fluids, and none grew in culture tests using any
of the standard solid or liquid media. However, Loeffler and Frosch
found that apparently sterile lymph from diseased animals produced
typical foot-and-mouth disease after being injected into healthy cattle
(Loeffler and Frosch, 1898, p. 372). Moreover, the disease could be
further conveyed from artificially infected animals, so the cause – which
they referred to as a virus – was not simply a toxin produced in diseased
animals.
Loeffler and Frosch emphasized the practical consequences of their
work; however, near the end of their last report, they turned to certain
results that seemed to have little immediate practical significance but
were of interest ‘not only for the further study of foot-and-mouth

disease, but also for numerous other infectious diseases of humans and
animals’ (Loeffler and Frosch, 1898, p. 388). Independently of Mayer
and Ivanovski (whose works they did not know) and of Beijerinck
(whose first paper on tobacco mosaic disease had not yet appeared),
they conducted filtration experiments using infectious lymph. Later, in
recounting the course of their research, Loeffler attributed their use of
filtration to a research tradition that began with Edwin Klebs’ study of
Microsporon septicum (Loeffler, 1911, p. 1). To insure that the filtered
lymph was free from bacteria, Loeffler and Frosch added Bacillus
flourescens to lymph before filtration. Since these bacteria were not
present in the filtered lymph, they were confident it was also free from
any bacterium as large or larger than Bacillus flourescens. Yet inocula-
tions proved that the filtered lymph was still effective. They recognized
this result as important, and they repeated their experiments several
times. They could imagine only two possibilities: ‘either the bacteria-
free filtered lymph contains an exceptionally powerful dissolved poison,
or the hitherto undiscovered causes of the disease are so small they can
pass through the pores of a filter that reliably retains the smallest
known bacteria’ (Loeffler and Frosch, 1898, p. 389). From the quantity
of filtered lymph sufficient to infect a calf, they calculated that any
dissolved poison would be enormously more virulent than tetanus toxin
– a possibility they dismissed as incredible (Loeffler and Frosch, 1898,
p. 390). The only alternative was that the effectiveness of the filtrate
was due to an agent that multiplied in living animals. Loeffler and
Frosch passed foot-and-mouth disease through a series of six animals
each of which died just as quickly as did the first. ‘This series of trials,
in each of which one fiftieth of a cubic centimeter of lymph, drawn
from the previous animal, then diluted and filtered, was used to infect
the succeeding animal, proves the virus of the disease is a living agent
that multiplies within the body of diseased animals’ (Loeffler, 1911, pp.
3f.).
The influenza bacillus was the smallest known bacterium at that
time. Loeffler and Frosch calculated that if the causal agent of foot-and-
mouth disease was one-tenth or even one-fifth as large, it would fall
outside the theoretical limits of what could be made visible by the best
microscopes. This could easily explain why the agent could not be seen
in virulent lymph. They conjectured that these results could be signifi-
cant for research on such other infectious diseases as smallpox, cowpox,
scarlet fever, measles, typhoid, and rinderpest (Loeffler and Frosch,
1898, p. 391).
At about the same time that Loeffler and Frosch published their
reports, Edmund-Isidore-Etienne Nocard and Emile Roux discovered
structures in bovine pleuropneumonia that, at maximum magnification,

appeared as minute points. They cultured these structures in collodion
sacs inserted into the abdominal cavities of rabbits and passed the
disease to healthy animals (Nocard and Roux, 1898). At a medical
conference in 1898, Loeffler asked Nocard whether the minute struc-
tures were filterable. Nocard said they were not, but within a few
months he advised Loeffler that, if suspended in liquids that were
adequately diluted, the structures did pass through filters. In discussing
these developments, Loeffler later observed that, ‘a second agent was
thereby recognized to be filterable’ (Loeffler, 1911, p. 4). The causal
agent of pleuropneumonia was at the limit of visibility with light micro-
scopes, but everyone recognized that the existence of such a filterable
organism supported the existence of other, still smaller, causal agents
(Nocard and Roux, 1898, p. 248).
In an account of his discovery, Loeffler mentioned that, in the same
year, Beijerinck reported conveying tobacco mosaic disease to healthy
plants after filtering sap. Loeffler then noted:
Having demonstrated the existence of infinitesimally small micro-
organisms, which elude microscopic demonstration and pass through
filters, one must now believe that the same methods of study
should be applied to all diseases for which only unsuccessful at-
tempts had been made to find causal agents, in the hope of
discovering, in the disease products of conveyable diseases, similar
minute living entities.
(Loeffler, 1911, p. 5)
By 1911, he reported that filterable viruses had been discovered for 17
different diseases (Loeffler, 1911, pp. 5f).
Klebs’s Grundversuche apply only to bacteria and, even assuming that

his theory accommodates all bacterial diseases, disorders caused by
other kinds of entities require different theories. One could take ac-
count of protozoal and viral diseases simply by restating the basic
assumptions of the bacterial theory as disjunctions. For example, in
place of the Bacterial Hypothesis one could write: every disease is
caused either by bacteria or by protozoa or by viruses, and so forth.
However, it is simpler to achieve the same result by formulating a
disjunction of theories, each more or less parallel to the bacterial theory
and designed to accommodate one kind of pathogenic organism. Each
of these theories will contain an assumption, similar in form to the
Bacterial Hypothesis, that specifies the domain of the theory. In seeking
the cause of a new disease, one will, in effect, consider each of these
theories one after another until a cause is identified. Of course, known
facts about the new disease may lead one to prefer some theories over
others.
Recognizing new kinds of causes also requires one to adopt new
causal criteria.3 In 1937, Thomas M. Rivers wrote:
The idea that an infectious agent must be cultivated in a pure state
on lifeless media before it can be accepted as the proved cause of a
disease has … hindered the investigations of certain maladies, inas-
much as it denies the existence of obligate parasites, the most
striking phenomenon of some infections, particularly those caused
by viruses
(Rivers, 1937, p. 5).
In the same paper, Rivers also discussed possible causal criteria for
viruses. About 20 years later, R.J. Huebner, proposed another set of
causal criteria for viruses (Huebner, 1957). Both Rivers and Huebner
took account of developments that came after Koch – for example, they
included criteria that involved immunological reactions to pathogens.
However, the only respect in which their criteria differed from those
suitable for use in the study of bacteria was that neither Rivers nor
Huebner had a counterpart to Koch’s Rt4; they did not require that a
purported causal agent be cultivated in inert media outside diseased
animals. Instead, the corresponding weak sufficiency criterion stipu-
lated only that inoculations of fluids drawn from a diseased animal
(and possibly filtered) were followed by the same disease in test
animals.
The discovery of filterable viruses was a greater accomplishment than

one might suppose. Looking back from our point of view (which as-
sumes a century of productive research on viruses and takes for granted
tissue cultures and electron microscopes) it is easy to see a smooth path
of research that required only certain empirical discoveries; but there
were important conceptual obstacles to be overcome. One such obstacle
concerned the distinguishability of viruses. We can approach this prob-
lem by way of the following quotation:
Microscopy had not proved useful in positively identifying the
cause of foot and month disease, although negative findings had
led Löffler and Frosch to conclude that the agent was submicro-
scopic. They also emphasized the use of inoculation experiments
on animals whose pathological reactions were the only positive
evidence then obtainable for the presence of the viruses. … Löffler
and Frosch thus established an experimental methodology which
in the early 20th century was widely adopted in research on human
and animal viral diseases. It was based primarily on three tech-
niques: microscopy and filtration to determine the absence in
infectious materials of microscopically visible organisms, and serial

animal inoculation experiments to provide evidence for the pres-
ence of infectious agents in the filtrates.
(Hughes, 1977, p. 64; my emphasis)
Thus, one condition for recognizing the presence of a virus was that no
causal agent could be detected by microscopy or by culturing. The only
positive evidence for the existence (or of the identity or nature) of the
virus was the pathological reaction that followed inoculation. Contempo-
rary researchers were clear about this fact. In 1915, one scientist observed
that varieties of non-pathogenic bacteria outnumber the pathogenic ones,
and one might expect the same to be true of the ‘ultra-microscopic
viruses’; however, he continued, ‘it is difficult to obtain proof of their
existence, as pathogenicity is the only evidence we have at the present
time of the presence of an ultra-microscopic virus’ (Twort, 1915, p.
1241). Nocard and Roux also admitted that animal inoculations were the
only way of demonstrating the presence of microscopically invisible dis-
ease agents (Nocard and Roux, 1898, p. 248). Yet the existence of the
viruses was postulated precisely to explain these very pathological reac-
tions – thus, the reasoning is circular.
Strümpell, Möbius, and Freud confronted exactly the same problem
with respect to the psychological disorders: because of the nature of
ideas, the only evidence for the existence or identity of a pathogenic
idea seemed to be the pathological reaction that followed it. So long as
Freud attributed hysteria to a sexual assault in childhood, there was a
chance for independent identification and differentiation of (at least the
source of some of the) supposedly pathogenic ideas. However, he later
admitted that the same pathogenic ideas could stem from fantasized
assaults – a concession forced on him by recalcitrant facts that would
otherwise have refuted his central hypothesis. And this left him with a
causal factor the only evidence for whose existence was the very patho-
logical reaction it was intended to explain. An 1898 critic could have
responded to Loeffler (or to Freud) as follows: ‘You assume the exist-
ence of a virus (an idea) to explain foot-and-mouth disease (hysteria),
but your only evidence for the existence of the virus (idea) is that
animals (people) contract the disease.’ How would this differ from the
following possible response (in 1847) to Eduard Lumpe: ‘You hypoth-
esize the existence of a miasm to explain childbed fever, but your only
evidence for the existence of the miasm is that women contract the
disease’?4 In all three cases, the reasoning approaches circularity. We
have seen that the Distinguishability Hypothesis was necessary to avoid
just this problem in dealing with bacterial diseases, but bacteria could
generally be identified and distinguished other than by their patho-
genicity. Since (at this point) this was no more true for viruses or for
ideas than for miasms, attempts to include ideas and viruses within the
research programme threatened to become what Lakatos called degen-
erating problem shifts (Lakatos, 1968, p. 118).
At this point, researchers could simply have abandoned these two
attempts to expand the programme. By 1898, childbed fever, diphthe-
ria, anthrax, cholera, tuberculosis, and several other bacterial diseases
had been assimilated to the etiological standpoint. However, for other
diseases, rather than hypothesizing the existence of viruses or ideas
(each undetectable except through the very pathologicality it was in-
tended to explain), one could have fallen back on traditional causal
thinking: foot-and-mouth disease, tobacco mosaic disease, hysteria, anxi-
ety neuroses, and the other disorders in both groups, could have been
regarded as having no necessary causes and as being due to the tradi-
tional factors to which conventional wisdom had long ascribed them.
One could simply have acknowledged that inoculations from diseased
sources or sexual attacks before puberty (respectively) were among the
many ways in which diseases of these two kinds could be caused. This
approach would not have undercut the success of the bacterial theory
and would have been compatible with all the observed facts. Moreover,
in contrast to hypothesizing unknown viruses or pathogenic ideas, this
would not have been question begging. However, this approach meant
abandoning the quest for specific causes – the very core of the research
programme. And, as Lakatos explained, in situations of this kind, the
negative heuristic of the programme shields the core from refutation:
In Newton’s program the negative heuristic bids us to divert the
modus tollens [of apparent counter-examples] from Newton’s three
laws of dynamics and his law of gravitation. This ‘core’ is ‘irrefuta-
ble’ by the methodological decision of its proponents: anomalies
must lead to changes only in the ‘protective’ belt of auxiliary,
‘observational’ hypotheses and initial conditions.
(Lakatos, 1968, p. 133)
At the core of our research programme was the assumption that every
disease could be characterized in terms of a universal necessary cause,
and this assumption simply could not be given up. Instead, researchers
began looking for non-circular ways of establishing the existence and
identity of the new purported causes.
Distinguishing viruses was a formidable task. At least by 1902, scien-
tists had concluded that viruses were too small ever to be revealed by
light microscopes, and some realized that, being obligate parasites, they
could never be cultured in an inert medium (Joest, 1902, p. 378). This
meant that, in principle, it would be impossible ever to gain knowledge
about viruses through the means that had proven most enlightening in
the study of bacteria. Moreover, since most viral diseases were species
specific, there was no possibility of ever using animals in test inocula-

tions to study most human viral diseases. Since inoculations were the
only known way of learning about or even demonstrating the presence
of viruses, this came close to implying that there was no possibility of
any positive knowledge about the hypothesized causal agents of these
diseases. However, the momentum of the research programme was
sufficient that even these daunting obstacles were seen only as tempo-
rary technical inconveniences. One scientist, asked whether, facing a
complete lack of positive knowledge about viruses, one should simply
be resigned to ignorance? ‘No!’ he answered, ‘for the results of previous
investigation prove only that the goal cannot be achieved by way of the
ordinary broad and familiar paths of bacteriological exploration. This
means only that we must invent other ways and means of approaching
and of learning about the unknown enemy’ (Joest, 1902, p. 417). It
seems never to have occurred to anyone to question the feasibility of the
goal itself.
If finding ways of distinguishing viruses presented major technical
difficulties, distinguishing ideas proved to be absolutely impossible. On
the assumption that one’s speech reflects one’s ideas, Möbius, Freud,
and others were driven, in one way or another, to the analysis of
patients’ language and to what came to be called talking cures. And,
while some of these attempts have been interesting and may have con-
tributed results of lasting value, the original goal of the project, namely,
the identification of pathogenic ideas as specific causes for psychologi-
cal disorders, seems to have evaporated. After a century of effort, most
psychological disorders are still not associated with specific causes (ideas
or otherwise), and the hypothesis of pathogenic ideas has been widely
discredited as nothing more than regressive question begging. In the
absence of feasible alternative hypotheses, the psychological disorders
remain unassimilated anomalies and, therefore, a distraction from the
credibility of the research programme as a whole.
By contrast, through the early decades of the twentieth century tech-
nical developments soon provided independent ways of identifying and
differentiating viruses. By 1957, Huebner’s causal criteria for viruses
required one to show, by passage in laboratory animals or in cell
tissues, that some new and purportedly pathogenic virus was a real
entity and to fully characterize the virus and to compare it with other
known disease agents (Huebner, 1957). This requirement includes a
counterpart to the bacterial Distinguishability Hypothesis. Within dec-
ades, viral diseases had been fully assimilated to the etiological standpoint,
and their discovery emerged as a major achievement of the programme.
Notes
1. Victoria A. Harden (1987) has given an account, compatible with the

general position of this book, of elaboration of the etiology of rickettsial
diseases, and Alfred S. Evans (1993) has recounted attempts to modify
Koch’s Postulates to encompass many different kinds of diseases.
2. Through the first three paragraphs of this section, page numbers in paren-
theses are references to Mayer (1886).
3. Both Gerhardt (1884, pp. 373f.) and Marchiafava and Celli (1885, pp.
805f.) made remarks that can be construed as attempts to identify such
criteria for protozoa.
4. Another example of pseudoexplanation is Fracastoro’s hypothesized ‘seeds’
(1546) which could be differentiated only by the diseases they were in-
tended to explain and which, therefore, explained exactly nothing.
CHAPTER ELEVEN
A Nutritional
Deficiency Theory of Disease
By the beginning of the twentieth century, many diseases had been
explained as infestations of microscopic parasites or of filterable vi-
ruses. However, other important diseases defied such explanations. One
such disease was beriberi which, in some populations, was responsible
for more deaths than all known parasitic diseases combined (Carter,
1977, p. 1191). Identifying the cause of beriberi required a new theory
of disease: a deficiency theory.
In conventional wisdom, the bacterial theory of disease is seen only
as a positive obstacle to development of a deficiency theory: Aaron J.
Ihde and Stanley L. Becker claim that ‘the germ concept proved a major
barrier to the recognition and study of deficiency diseases’ (Ihde and
Becker, 1971, p. 16). C. P. Stewart writes, ‘one factor which undoubt-
edly held up the development of the concept of deficiency diseases was
the discovery of bacteria in the 19th century and the consequent preoc-
cupation of scientists and doctors with positive infecting agents in
disease’ (Stewart, 1953, p. 408). Numerous writers assert that the defi-
ciency concept was difficult to accept because the bacterial theory
suggested that only a positive agent could cause something. In his
history of nutrition, Emer Verner McCollum notes that ‘improved mi-
croscopes and the staining and other technics applicable to the study of
the problems of pathology so monopolized the attention of investiga-
tors that they had little incentive to consider any aspect of malnutrition
as a cause of disease’ (McCollum, 1957, pp. 225f). In Toward the
Conquest of Beriberi, Robert R. Williams treats the bacterial theory
only as the source of one false explanation for the etiology of beriberi
(Williams, 1961, pp. 13–15, 18f., 35). Such claims, so frequently re-
peated, pass for common sense – never mind that they are usually
unsupported by any evidence whatsoever. Do these authors really mean
to assert that, if there had been no bacterial theory, the deficiency
concept would have been more readily accepted? What evidence could
there possibly be for a contrafactual claim of this kind? Who can say
what might have been?2
However, all these claims are based on a serious misconstruction of
the historical context. That the study of bacteria did not monopolize
late nineteenth-century etiology is obvious from the variety of causes
contemporary researchers were willing to entertain. Some diseases had

been traced to minute but visible parasites such as trichinae or various
fungi. Other diseases were ascribed to filterable viruses. On analogy
with bacteria, some may have assumed that viruses were corpuscular,
but Martinus Willem Beijerinck thought in terms of a living fluid
(Beijerinck, 1899, p. 31). Even as late as 1915, one researcher specu-
lated that viruses could be minute bacteria, tiny amoeba, or even a
‘more lowly organized’ form of life – a ‘living protoplasm that forms no
definite individuals, or an enzyme with power of growth’ (Twort, 1915,
p. 1242). Strümpell, Möbius, and Freud ascribed psychological disor-
ders to pathological ideas. Even more significantly, as Thomas Schlich
has discussed in a recent paper, by 1892, Theodor Kocher had proven
that cretinism was due to thyroid failure, that is, to a deficiency (Schlich,
1994). All these possibilities were given serious attention in medical
literature. In fact, rather than focusing on bacteria, medical research at
the time was driven by the quest for universal necessary causes (many of
which turned out to be bacteria). And, while one cannot say what might
have been had bacteria attracted less attention, it is perfectly clear that,
in the absence of the quest for universal necessary causes, no one could
possibly have sought the cause of beriberi at all because only given the
etiological standpoint could one even make sense of such a notion.
Prior to assimilation by the research programme, individual cases of
beriberi were ascribed to the same unlimited range of chance factors as
were cases of every other disease; prior to inclusion in the programme,
such causes were the sum total of the etiology of beriberi. ‘The cause of
beriberi’ – like the cause of any other disease – was first given a
meaning by the etiological research programme. Certainly beriberi was
studied by bacteriologists who used their own techniques and found
little. Given the historical context, it would have been irrational to have
proceeded in any other way. However, this was an important null
result, not just an obstacle. And to portray the rise of the deficiency
concept only in terms of its supposed competition with the bacterial
theory, rather than as part of the broader programme (of which the
bacterial theory itself was a part), is exactly to blind oneself to the
forest by gaping at a single (if impressive) tree.
By the beginning of the twentieth century, the etiological programme
drove virtually all medical research, and this programme imposed no a
priori restrictions on the nature of potential causes. While deficiency
diseases obviously could not be assimilated to the bacterial theory, the
programme itself presented no conceptual barriers to accepting defi-
ciencies, or almost anything else, as causes. The positive heuristic of the
programme provided a strategy for finding causes. Near the end of our
discussion of Robert Koch (Chapter 8), we identified some of the ele-
A NUTRITIONAL DEFICIENCY THEORY OF DISEASE 181
ments of that strategy: one requires a theory defined over a domain of

potential causes, the theory will include a distinguishability condition
and a set of causal criteria, potential causes must explain disease phe-
nomena, ordinarily there will be some test for weak sufficiency. None of
this precludes negative causes. Ultimately, both the bacterial and the
deficiency theories (like the ideational and viral theories) emerged from
this common strategy. To portray the bacterial theory simply as an
obstacle is to obscure a vital part of the story – in exactly the same way
that traditional Freudian scholarship obscures crucial aspects of Freud’s
‘new’ approach to hysteria by ignoring his indebtedness to the research
programme.
The contrast between the histories of scurvy and of beriberi provides
vivid confirmation of the dominance of the research programme in
nineteenth-century medicine. We will begin with some remarks about
the history of scurvy.
Scurvy as a disorder of diet
In the early nineteenth century, scurvy and rickets were generally be-
lieved to be diet-related. In 1830, Lancet published a series of clinical
lectures by John Elliotson, Professor of Medicine in University College,
London. In discussing scurvy, Elliotson noted that ‘the cause … is
always, I believe, a want of fresh animal and fresh vegetable food’
(Elliotson, 1830–31, p. 651). Seven years later, in a similar lecture,
Marshall Hall observed
scorbutus is generally induced by a deficiency of fresh vegetable
food. It is also occasionally referred to other errors in diet, to the
respiration of a crowded or otherwise impure atmosphere, to ex-
cessive fatigue, anxiety, etc. … The prevention and cure of scorbutus
consists in the administration of fresh and vegetable food, but,
above all, of citric acid
(Hall, 1837–38, p. 851).
Both men refer to works by Gilbert Blane (1753) and James Lind
(1785) that were, by that time, recognized sources on scurvy. Neither
Elliotson nor Hall suggested that his views were new or atypical; in-
deed, both acknowledged that lemon juice had been the standard cure
for scurvy for more than 200 years.
In 1840, George Budd summarized clinical knowledge of scurvy in an
article in Alexander Tweedie’s A System of Practical Medicine (Budd,
1840). Budd discussed various fruits and vegetables generally known to
cure scurvy. Knowing of François Magendie’s experiments demonstrat-
ing the inadequacy of certain macronutrients, Budd speculated that ‘the
study of organic chemistry and the experiments of physiologists’ would

ultimately shed light on the essential element common to the antiscor-
butic plants (Budd, 1840, p. 77). Budd’s article on scurvy was frequently
cited by English physicians through the turn of the century. One reason
for his continued influence was that he was among the last practicing
physicians to be particularly interested in the disease. By the beginning
of the nineteenth century, scurvy had been virtually eliminated from the
British navy (Dudley, 1953). Except for uncontrollable situations, such
as famine and war, where known therapies could not be applied, scurvy
gradually declined in other areas and populations (Stewart, 1953, pp.
404–12). The symptoms of scurvy were more readily controlled than
were those of most other diseases. Gilbert Blane wrote that ‘the efficacy
of lemon juice in curing scurvy [must] be stated as singular when
compared to that of any other remedy in any other disease’. Since it
prevents and cures the disease so completely and with no adverse ef-
fects, ‘it performs not only what no other remedy will perform in this
disease, but what no known remedy will effect in any known disease
whatever’ (Blane, 1785, pp. 179f.). Given this, and given that the
prevailing concept in medical circles was that each disease was a par-
ticular collection of symptoms, scurvy was believed to be completely
understood. This attitude is reflected in an editorial written in 1858
wherein the conquest of scurvy is spoken of as a leaf in the laurel
wreath on the brow of medical science (Lancet, 1858, p. 145).
One can gauge the attitude of nineteenth-century British physicians
toward scurvy by examining references to the disease in the articles,
editorials, and letters in Lancet. After Budd’s articles there are few
references to scurvy until 1848, when failure of the potato crop resulted
in numerous cases of the disease in England and Ireland. Work carried
out at this time by John Aldridge and (independently) by Alfred B.
Garrod suggested that scurvy may have been caused by a deficiency of
potash (Carter, 1977, p. 122). Through the next decade there are few
references to the disease. Beginning in the late 1850s and continuing for
about twenty years there are numerous letters and editorials decrying
the continued appearance of scurvy in the British merchant fleet (Carter,
1977, p. 122). In these notices, scurvy is consistently treated as a
disease entirely understood and completely preventable – ship owners
who allow the disease to appear are regarded as criminal and, in one
editorial, compared to murderers (Lancet, 1858, p. 146). In one of the
few original opinions on scurvy to appear in this period, the disease is
attributed to deficiency of protein (Oliver, 1863, p. 61). Fifteen years
later, in an original article, we read that ‘No fact in medicine is more
clearly established than that the exclusive cause of scurvy is the pro-
longed and complete withdrawal of succulent plants and fruits’ (Ralfe,
1877, p. 868). By the 1880s, scurvy was treated as a medical curiosity

of which very few practicing physicians had had immediate experience
(Eade, 1880, p. 992). In that period, however, letters and editorials
begin to take account of the experiences of arctic explorers who had
survived for months on fresh uncooked meat with no sign of scurvy.
The chief question in these writings is whether the dietary deficiency
that results in scurvy could be corrected by fresh meat as well as by
fresh vegetables (Carter, 1977, p. 122).
By 1883, some bacteriologists suspected that scurvy was caused by
microorganisms, but the first hint of this view in Lancet was in 1886. In
that year, a brief editorial mentions research by a Russian pathologist,
T. Stazevich, who argued that scurvy was a ‘form of septic poisoning’
(Lancet, 1886, p. 1036). Until that time Lancet contains no suggestion
that scurvy is anything other than a nutritional deficiency disorder. In
1889, Wilhelm Koch published an ambitious study of blood diseases in
which he argued that scurvy, hemophilia, and various other disorders
were variant forms of an infectious blood disease, and that the obvious
correlation between scurvy and the lack of fresh food (like the heredi-
tary pattern of hemophilia) could be explained along these lines (Koch,
1889). Koch’s book seems not to have had much impact, but it was
reviewed seriously in Lancet (Lancet, 1890). Through the end of the
century, there are scattered attempts to explain scurvy as a result of
toxins or of microorganisms; these were greeted skeptically and often
rebutted. In the first decade of the twentieth century, opinion was
clearly shifting toward such an explanation, but, as late as December
1904, an editor of Lancet could still write that ‘the general disposition
is to regard scurvy as due to the absence of certain elements in the food
which is taken, but the exact nature of those elements has not been
conclusively demonstrated’ (Lancet, 1904, p. 1660). In 1907 a letter
printed in the British Medical Journal (BMJ) bemoans the fact that ‘the
theory that want of fresh vegetable causes scurvy dies very hard’, but
the author’s point is only that fresh meat will also supply the necessary
nourishments (BMJ, 1907, p. 683).
As we have seen, the preceding quotation is part of a continuous
tradition that extends back beyond the time of George Budd. This
quotation was published less than three years before the classic paper
on scurvy by Alex Holst and Theodor Frolich that was fundamental to
the development of the deficiency theory (Holst and Frolich, 1907), and
less than eight years before that theory received its first full articulation
by Casimir Funk (Funk, 1912). Yet in 1911, seven years after the
quoted editorial, a report on vitamins published by the British Medical
Research Committee stressed the difficulty of implanting ‘the idea of
disease as due to deficiency’ (Medical Research Committee, 1919, p. 2).
Ironically, this report refers to Budd’s work on scurvy and commends

his treatment of the history of the disease (Medical Research Commit-
tee, 1919, p. 58). In 1932, a second report by the Medical Research
Council retained the language about the difficulty of thinking of nega-
tive causes, dropped the reference to Budd, and added a note indicating
that ‘it is now difficult to understand how … scurvy failed in practical
medicine to obtain recognition as a disease due to a deficiency in food’
(Medical Research Council, 1932, p. 14). It is still more difficult to
understand how these false traditions could have gained credence so
quickly. No doubt, part of the explanation is that there is little incentive
to study scientific literature outside one’s own research. For example, in
reviewing early references to the deficiency concept, Holst and Frolich
refer to James Lind (one of the few early British authorities who failed
to identify scurvy unequivocally as a deficiency disease) and August
Hirsch, and to various reports by observers of particular occurrences of
the disease (Holst and Frolich, 1907, pp. 663–9). There is no recogni-
tion of the 100-year tradition in British practical medicine in which
scurvy was consistently and almost unanimously regarded as a defi-
ciency disease.
No doubt the bacterial theory fascinated medical researchers during
the latter part of the nineteenth century; however, serious original work
on scurvy had in fact ceased decades earlier. Through the century,
practicing physicians saw fewer and fewer cases of scurvy; medical
interest in the disease declined accordingly. In this period, the few
chemists and physiologists who studied human nutrition showed almost
no interest in the disease. By 1880 bacteriology was certainly the most
promising field for medical research, but, at least in England, physicians
continued to believe that scurvy was a well-understood nutritional dis-
order that held little practical or theoretical interest. Under these
circumstances it is misleading, if not false, to speak of the bacterial
theory as an obstacle to the recognition and study of scurvy as a
nutritional disease.
The rise of Beriberi
Beriberi seems to have been known to Asian writers as early as the

second century AD; the earliest descriptions in western literature are
from the seventeenth century (Carter, 1977, p. 124). Prior to the nine-
teenth century, the disease was relatively unimportant even in the orient.
This is evident, first, from the treatment the disease received in pre-
nineteenth-century medical literature, and second, from the fact that
modern steam-milling procedures, which are now recognized as having
been substantially responsible for the rise of beriberi, did not become
common in Asia until then. By the last quarter of the nineteenth century
the disease was widespread and growing at an alarming rate.
It is difficult to gauge accurately the growth of beriberi; we must be
content with a few hints. First, in the Japanese Army in 1876 there were
3,868 cases of beriberi from among 35,300 men (11 per cent), in 1877
there were 2,687 cases among 19,600 men (14 per cent), and in 1878
there were 13,629 cases among 36,100 men (38 per cent) (Scheube,
1894, p. 14). This percentage remained roughly constant through 1884
and then declined because of dietary reforms, but during the war between
Japan and Russia (l904–05) there were nearly 100,000 cases of beriberi
(Takaki, 1906, p. 1521). Second, in 31 district hospitals in the Malay
peninsula, Chinese patients admitted for beriberi increased from 1,206 in
1881 to 3,175 in 1891, and to 6,767 in 1901. In that period, beriberi
accounted for more than 100,000 deaths among Chinese laborers in the
Malay States – over half the total death rate for that population (Braddon,
1907, pp. 1–4, 513–21). Third, in 1883 August Hirsch reported that
within the preceding 20 years, beriberi had appeared in many areas for
the first time, and that it was epidemic in coastal areas of Japan and of
other Asian countries (Hirsch, 1885, pp. 573–8). Ten years later, B.
Scheube reported that the disease had recently appeared for the first time
in Siam and in the Philippines, that it was spreading in Africa and in
South America, and that it was now endemic throughout Japan (Scheube,
1894, pp. 10–16). In 1907, Patrick Manson gave the disease an even
more extensive distribution (for example, he stated that it was then
common in southern China where, according to Hirsch and Scheube, it
had been rare), and in most areas incidence of the disease seemed to have
increased (Manson and Daniels, 1907, pp. 615–43). Finally, one can
appreciate the remarkable rise of the disease by surveying the medical
literature of the period. Scheube’s nearly complete bibliography includes
two publications on beriberi between 1800 and 1809; in subsequent
decades the publications numbered 8, 10, 11, 30, 64, 80, and between
1880 and 1889, 181 (Scheube, 1894, pp. 207–18). Leonard Braddon’s
incomplete bibliography lists nearly 200 articles and books between 1890
and 1899 (Braddon, 1907); another incomplete bibliography lists 250
publications between 1900 and 1910 (Schaumann, 1910, pp. 699–709).
A complete bibliography for this decade alone would approach 500
items. Writers in the late 1870s still regarded beriberi as an exotic and
unfamiliar topic; about 30 years later, an editorial in Lancet observed
that ‘there is probably no disease concerning which so much discussion as
to its etiology has taken place as beriberi’ (Lancet, 1911, p. 842).
We can, somewhat arbitrarily, take 1880 as the beginning of serious
western interest in beriberi. By that time there was a strong and broadly
based opinion that the disease was diet-related. It was obvious to every-
one that the disease was prominent only where rice was the staple diet.
Between 1850 and 1890, several observers attributed beriberi to an
‘insufficient diet or a diet not corresponding to the metabolisms and
bloodmaking, or to the needs of the body’ (Hirsch, 1885, p. 589). Of
those who ascribed it to insufficient diet, some seem to have believed
the problem was simply quantitative – those who got beriberi had too
little to eat; however, others believed the problem was qualitative –
sufferers had too little of certain essential foods.
Among the most important early observers to espouse a dietary ex-
planation was Kamehiro Takaki. Takaki was the first to collect systematic
evidence on a large scale that supported the deficiency concept. He first
heard of beriberi from his father who was a guard at the Japanese
Imperial Palace. Many of the guards suffered from beriberi; ‘they attrib-
uted the cause to food and called a provision box the “beriberi box”’
(Takaki, 1906, p. 1370). Takaki became a naval doctor, spent five years
studying medicine in London, and was appointed director of the Tokyo
Naval Hospital. By 1882 Takaki’s own observations led him to at-
tribute beriberi to a poor diet. His view was that ‘wide departure of
nitrogen and carbon from the standard proportion (one to fifteen)
essential to the maintenance of health, resulting from a great deficiency
of nitrogenous substances and a great excess of carbohydrates in food,
is the cause of kakke (beriberi)’ (Takaki, 1885, p. 29). Takaki persuaded
the skeptical Japanese admiralty to initiate massive dietary reforms –
crews ate more meat (especially fresh meat), more vegetables and, at
some meals, barley instead of rice. The effects were dramatic: in 1882
there were over 400 cases of beriberi for each 1,000 men, within five
years the disease had been eliminated. Takaki’s work was reported in
major European medical periodicals, he was honored at home and
abroad, and his evidence was ultimately important in understanding
beriberi.
However, Takaki’s ideas about the disease were inconsistent with a
large, if unsystematic, body of epidemiological facts widely known
before his studies even began. In 1835, John G. Malcolmson observed
that ‘the comparative cheapness of all kinds of grain in the Circars, and
the easy circumstances of many of the native soldiers who suffered, are
fatal to any supposition of the disease depending on deficient and
unhealthy diet’ (Hirsch, 1885, p. 592). By 1880 it was common knowl-
edge that those who contracted beriberi often ate more and a better
range of foods than those who did not (Hirsch, 1885, p. 592). More-
over, populations that seemed especially vulnerable often lived amid a
larger population that remained healthy, and the only apparent dietary
difference between the groups was that the larger and immune popula-
tion consumed less protein (Hirsch, 1885, p. 593). In Japan, beriberi

was most prevalent in sea ports where fish was most abundant; in the
East Indies, military garrisons contracted the disease while their serv-
ants (who ate less meat) did not (Rupert, 1880, pp. 509f.). Finally,
specific cases were known in which beriberi appeared among persons
living almost entirely on protein (Hirsch, 1885, p. 593). It was impossi-
ble to reconcile these facts with Takaki’s theory. They and similar ones
continued to count as evidence against dietary theories of beriberi until
after the turn of the century.
Because the deficiency theory seemed incompatible with the facts,
most early observers favored other explanations of beriberi. Perhaps the
most popular alternative theory was that it was miasmatic or malarial.
Early observers noted epidemiological similarities between beriberi and
malaria. These similarities were emphasized even by persons who did
not espouse the miasmatic concept. William Anderson, among the earli-
est western writers to investigate beriberi, noted that most Japanese
doctors ‘believe that the complaint is caused by some poisonous emana-
tion from the soil’. He listed similarities between beriberi and malaria
and concluded that the disease is most likely due to ‘the existence of an
atmospheric poison’ but observed that no single cause could account
for the disease (Anderson, 1876, p. 19). The Lancet review of Takaki’s
work emphasized that ‘the majority of observers … have been inclined
to attribute [beriberi] … to a specific poison which is generated in the
soil under certain unsanitary conditions of local origin, and finds its
way into the human body by means of the atmosphere, and perhaps of
the food and drinking water also’. The reviewer observed that, Takaki’s
work notwithstanding, ‘the weight of evidence is still in favor of the
miasmatic hypothesis’ (Lancet, 1887, p. 234).
As bacteriology became more prominent, many expected that beriberi
would prove to be bacterial. Researchers sought (and often claimed to
have found) the causal microorganism: Glockner identified an amoeba,
Fajardo a hematozoon, Perelra a spherical microorganism, Durham a
looped streptococcus, Lacerda a polymorphous ascomycete, Taylor a
spirillum, Pekelharing and Winkler a micrococcus, Thomas the anchy-
lostomum duodemale, Nepveu a strep bacillus, Rost a diplobacillus,
and Dangerfield an aerobic micrococcus, and there were others (Scheube,
1894, pp. 176–91). Numerous theories linked beriberi to specific or-
ganisms. Hamilton Wright attributed the disease to an organism that
entered the body by the mouth, and produced a toxin in the pyloric end
of the stomach (Wright, 1902, p. 58). Herbert Durham concluded that
the disease is similar to diphtheria and was communicated by fomites
(Durham, 1904). Patrick Manson proposed it was a form of intoxica-
tion, not unlike alcohol intoxication, in which a toxin, manufactured
outside the body (probably by microorganisms) is introduced into the

body (probably through air) (Manson, 1901–02, pp. 12–16). Research-
ers seldom agreed which microorganism was the cause, but few seriously
doubted that there was one. In 1897, M. H. Spencer wrote: ‘Little
doubt … remains that [beriberi] is a germ-borne disease and that the
microorganism which is the cause of it has a specially toxic influence
upon the peripheral nerves’ (Spencer, 1897, p. 32).
The deficiency theory of disease
In 1883, C.A. Pekelharing and A. Winkler were sent to Java by the

Dutch government to study beriberi. Christiaan Eijkman, who had
studied with Koch in Berlin, was a staff bacteriologist with the expedi-
tion. He was specifically instructed to find the organism that caused
beriberi (Williams, 1961, p. 19). When Pekelharing and Winkler re-
turned to Europe, they reported that the disease was definitely parasitic
but that they had not yet conclusively identified the agent (Van der
Berg, 1889, pp, 941f.).
Eijkman remained in Java as director of a bacteriological laboratory
connected with a military hospital. While there he made two important
discoveries, the first of which was that under certain conditions, chick-
ens spontaneously contracted a disease whose symptoms and histological
features resembled those traditionally connected with beriberi. For years,
dogs, rabbits, guinea pigs, and monkeys had been used by bacteriolo-
gists in beriberi research and, although Eijkman is not explicit on this
point, it seems likely that the chickens used in his laboratory were
intended for that purpose. He reported making this discovery, by acci-
dent, when his chickens began to show signs of polyneuritis (Eijkman,
1897a, p. 525). The chickens were examined carefully and no patho-
logical organisms were found; nor was he able to infect healthy chickens
by exposure to those with polyneuritis. Eijkman then discovered that,
whereas the chickens in his laboratory were generally fed a low-grade
uncooked rice, it had happened that for some weeks they had been fed
surplus cooked rice from the hospital kitchen.
His second important discovery was that, while ordinary chicken feed
was unpolished rice, rice from the kitchen was polished. Here is one
account of the sequence of events leading to these:
Eijkman’s original research in connection with beriberi began in a
curiously accidental way. He wished to carry out certain investiga-
tions on fowls, and in order to economize on their food he fed
them on scraps from the wards of the military hospital to which he
was attached. On these scraps, which consisted chiefly of cooked,
polished rice, the fowls developed paralyses, whose nature was at

first obscure. A clue thereto was unintentionally given by a newly
appointed director of the hospital, who refused to let Eijkman feed
his fowls any longer on scraps from the wards. Henceforward they
were fed on gaba (rice still in the husks) and on this diet they
recovered.
(Lancet, 1930, pp. 1097f.)
Eijkman soon became convinced that the consumption of polished rice
was associated with the inception of polyneuritis gallinarum, as he
called the chicken disease. Without assuming that polyneuritis gallinarum
was the same as beriberi, Eijkman wondered whether beriberi was also
correlated with the consumption of polished rice.
Rice was commonly prepared for human consumption in three ways:
1. Rice could be milled immediately after being harvested. If so, both

outer and inner husks were generally removed, leaving only the
white grain. This rice was called uncured, polished, or decorticated
rice.
2. In some areas, newly harvested rice was soaked, steamed, and dried
before milling. In this event, milling usually left some of the inner
layers of husk, called pericarp, and certain light brown inner cover-
ings of the grain. This rice was called cured or unpolished rice
(Grist, 1965, p. 404).
3. In some primitive areas, rice was not machine milled at all but was
stored unhusked and then pounded and winnowed just prior to
eating. Local custom, and the ethnic origin and the economic status
of consumers were among the factors determining which kind of
rice was eaten (Braddon, 1907, p. 137–50).
In Java, prison inmates were usually fed whichever form of rice was
commonly consumed by the local population. As it happened, in 27
prisons inmates ate unpolished rice, while in 74 others the rice was
decorticated to some extent. This provided an ideal opportunity for
determining whether beriberi was correlated with the consumption of
decorticated rice. Eijkman’s colleague, A. G. Voderman, who was then a
Civil Medical Inspector, conducted surveys of the prisons in Java. The
preliminary results were astonishing: of nearly 300,000 prisoners, only
one in 10,000 of those who ate unpolished rice had beriberi while one
in 39 of those who ate polished rice had the disease (Eijkman, 1897b).
These results exhibited a striking correlation between the incidence of
beriberi and the consumption of particular kinds of rice.
Within a few years, studies of other populations confirmed these
results. Braddon noted that in the Malay States, Chinese, who ate
polished rice, were seriously afflicted whereas Tamils, who ate unpolished
rice, and native Malays, who ate rice that was not mechanically milled,
were almost free from the disease (Braddon, 1907, pp. 150–198). Both
Voderman’s and Braddon’s evidence was demographic and not subject
to controls. Eijkman’s studies on fowl could yield only analogical argu-
ments that many found unconvincing. ‘The favorite view of most
physicians who knew of [Eijkman’s] work at all, was to dismiss the
matter with the assertion that polyneuritis in fowls had no relation to
human beriberi’ (Williams, 1961, p. 14). However, toward the end of
the decade, William Fletcher (Fletcher, 1907), and Henry Fraser and
Thomas Stanton (Fraser and Stanton, 1909) published important stud-
ies on small but carefully controlled populations. These studies probably
did more than anything else to persuade physicians that consumption of
decorticated rice was connected with beriberi.
By 1910, the evidence was clear enough for a meeting of the Far
Eastern Association of Tropical Medicine to adopt a motion stating that
‘in the opinion of this association sufficient evidence has now been
produced in support of the view that beriberi is associated with the
continuous consumption of white (polished) rice as the staple article of
diet’ (BMJ, 1910, p. 1000). In a 1911 meeting of the Society of Tropical
Medicine and Hygiene, the consensus of opinion favored a dietary
theory for beriberi, and Patrick Manson, admitting that his ideas about
beriberi ‘belong to a past age’, was almost alone in expressing reserva-
tions about such a theory (Manson, 1911–12, p. 85).
Numerous theories were proposed to explain the relation between ber-

iberi and polished rice and, not surprisingly, most of these were modelled
in one way or another on the bacterial theory. Braddon proposed that
polished rice was the locus within which a toxin was created by micro-
organisms and by which it was transferred to potential victims (Braddon,
1907, pp. 39–48). Eijkman’s original hypothesis was also a version of
the toxin theory:
under the assumption that all polyneuritis ultimately seems to be
intoxication, we must assume that the starch in these cases carries
a poison or that from it – either in the alimentary canal (under the
influence of a microorganism?) or in the nerves – a poison is
produced by the chemical process of metabolism. In the pericarp of
the grain, then, the material(s) would be present, through which
the poison is, in some way, made harmless or, perhaps, its creation
is prevented.
(Eijkman, 1897a, pp. 529f.)
There were other theories as well (Williams, 1961, pp. 14f).
In 1901, Gerrit Grijns, a bacteriologist who succeeded Eijkman as

director of the pathology laboratory in Batavia, first proposed that
there may be some unknown ingredient in the pericarp whose absence
resulted in beriberi (Grijns, 1901). Grijns asserted that recent develop-
ments were leading away from the idea that beriberi was infectious.
Moreover, in his own experiments he had induced polyneuritis in chick-
ens by feeding them concentrated protein (cooked horse meat) and this,
he felt, counted heavily against Eijkman’s hypothesis. The most likely
conclusion seemed to be ‘that there are various natural foodstuffs that
cannot be missed without particular damage in the peripheral nerves’
(Grijns, 1901, p. 45). Grijns’s hypothesis, which appeared in a relatively
minor journal, seems to have received little attention.
For the next few years, most researchers continued to favor a toxic or
infectious explanation for beriberi, and this obviously reflects the promi-
nence of the bacterial theory. However, it does not imply that facts
about beriberi were blindly or irrationally forced into the bacterial
mold. First, as we have seen, there were persuasive arguments against a
strictly dietary explanation of beriberi. Second, through the first decade
of the twentieth century, explanations of beriberi based on the bacterial
theory continued to yield significant new results. Fletcher and Braddon
both conducted their demographic studies in connection with infectious
theories of the disease. In testing Eijkman’s toxin hypothesis, Grijns
obtained important results from animal experiments. Because he be-
lieved that the pericarp contained a natural antitoxin that neutralized
the harmful influence of the starchy rice grain, Eijkman himself began
an active chemical investigation in the effort to isolate the antitoxin. He
was able to show that an aqueous extract from rice polishings cured
polyneuritis gallinarum; he also showed that, when foods were heated
above 120°C, they lost their effectiveness in preventing and curing the
disease (Eijkman, 1906, pp. 164–70). All these results were conse-
quences of the assumption that bacteria were somehow involved in the
etiology of beriberi, so it was reasonable to continue thinking that
bacteria may somehow be involved.
The success achieved by Eijkman and Grijns suggested profitable
avenues of investigation for researchers working on a different disease.
Beginning in 1894, crews on Norwegian ships were recognized as suf-
fering from a disease some of whose symptoms resembled beriberi. This
disease was therefore called ship-beriberi. In 1902, a Norwegian re-
search commission concluded that ship-beriberi was a non-infectious
intoxication from tainted foods (Lancet, 1903). In the same year, Alex
Holst visited Grijns in Batavia, and Grijns showed him the experiments
he was performing on fowls (Holst, 1907). Holst was impressed with
the work of Eijkman and Grijns, and he adopted a similar approach in
his own study of ship-beriberi. He tried to induce the disease in pigeons

and chickens, and later, having been joined by Theodor Frolich, in
guinea pigs (Holst and Frolich, 1907). The results of controlled feeding
experiments were curious: while Holst and Frolich were trying to study
ship-beriberi, the guinea pigs regularly contracted a disease that bore
every similarity to scurvy. Holst and Frolich did not verify that similar
diets would induce scurvy in humans, but they did mention that earlier
evidence connected scurvy with deficient diet, and that scurvy and
beriberi often appeared together. They also cited B. Nocht, a German
observer, who had argued that ship-beriberi was a form of scurvy.
Because of these publications, beriberi, scurvy, and ship-beriberi were
all seen as causally linked to deficient diets and, therefore, as theoreti-
cally linked to one another.
In 1910 Fraser and Stanton showed that the substance that prevented
beriberi was soluble in strong alcohol, and that the effectiveness of a
given grain in combating beriberi seemed to depend on the amount of
phosphorus it contained (Fraser and Stanton, 1910). By obtaining a
purer specimen of the ingredient in rice husks that protected test ani-
mals from beriberi, Casimir Funk was soon able to show that the
disease was not due to phosphorus deficiency, but the idea that beriberi
and scurvy were due to a deficiency of some essential nutrient or
nutrients was becoming more plausible.
By this time, several lines of work were converging: epidemiological
studies (Takaki, Voderman, Braddon, Fletcher, Stanton and Fraser) and
animal experiments (Eijkman, Grijns, Holst and Frolich) had exhibited
a connection between the consumption of milled grain and various
distinctive disorders; chemical studies (Eijkman, Fraser and Stanton,
Funk) had isolated and characterized with fair precision the particular
ingredient in rice polishings that would prevent and cure some of these
disorders; an important negative result was that, after 30 years of
searching, bacteriologists could not agree in identifying a particular
microorganism that was causally responsible for any of the diseases in
question. Given the importance of the bacterial theory, it would have
been difficult for any deficiency hypothesis to have been accepted with-
out this negative result. The strands were finally assembled in Casimir
Funk’s ‘The Etiology of the Deficiency Diseases’ (Funk, 1912). Funk
announced – somewhat prematurely as it turned out – the isolation of a
concentrated form of the protective substance for beriberi; he proposed
that it be called vitamine (Funk, 1912, p. 347). Funk classified beriberi,
polyneuritis in birds, epidemic dropsy, scurvy, infantile scurvy, experi-
mental scurvy in animals, ship-beriberi, pellagra, and rickets as deficiency
diseases. He stated that these diseases are due to different deficiencies.
Funk noted that about twenty years of experimental work had been
required to establish that these various diseases were caused by a defi-

ciency of essential nutrients. He admitted that this view was still not
generally accepted, but claimed that there was enough evidence ‘to
convince everybody of its truth, if the trouble be taken to follow step by
step the development of our knowledge on this subject’ (Funk, 1912,
p. 341).
The deficiency theory in relation to the etiological research programme
As we have seen, in 1910 the Far Eastern Association of Tropical

Medicine concluded that ‘sufficient evidence has now been produced in
support of the view that beriberi is associated with the continuous
consumption of white (polished) rice as the staple article of diet’ (BMJ,
1910, p. 1000). There are unmistakable parallels between this conclu-
sion and beliefs about scurvy that prevailed 80 years earlier in the
1830s. In both cases physicians discovered an empirical correlation
between a symptomatically defined disorder and a particular dietary
pattern. In both cases, also, the disorder could be effectively controlled
by changing what people ate. The remarkable difference is that, whereas
in the 1830s this empirical understanding of scurvy effectively satisfied
scientific interest in the disease, in 1910 recognition of the empirical
correlation between beriberi and the consumption of polished rice only
intensified research on the etiology of the disease. How can one explain
this difference? The obvious answer is the research programme to which,
by 1910, virtually all medical scientists subscribed. In the 1830s, the
understanding of scurvy enabled physicians to control symptoms (and,
as then conceived, that was all there was to the disease), but, by the turn
of the century, in comparison to the explanations emerging from the
identification of universal necessary causes, even effective control was
seen as inconclusive. As Adolf Strümpell observed:
The empty generalities of the past only appeared, superficially, to
satisfy the need for causes; this need can only be met through the
discovery of causes that operate in every single case, only through
a knowledge of their nature, of the manner of their operation, of
the site of their influence, and of the necessity of their consequence.
(Strümpell, 1893, p. 22f.)
The etiological standpoint carried a new standard for comprehension –
a standard that was obligatory for the acceptance of any etiological
account.
By 1910 the most impressive achievements in the quest for specific
causes involved bacteria. Largely because these successes were so strik-
ing, medical science in general had become converted to the etiological
standpoint. Having accepted this way of thinking, researchers looked

for universal necessary causes for all diseases. Many diseases proved to
be non-bacterial. However, since all this research was informed by the
etiological standpoint, and since the etiological standpoint itself was
accepted largely because of the striking successes of the bacterial theory,
it is clearly a mistake to regard the bacterial theory only as an obstacle
to research on non-bacterial diseases. The truth is almost exactly the
opposite. Conventional wisdom is that ‘improved microscopes and the
staining and other [bacteriological] technics … monopolized the atten-
tion of investigators’ and left them with ‘little incentive to consider …
malnutrition as a cause of disease’ (McCollum, 1957, pp. 225f.). Even
if this were true (which it is not), without the successes of bacteriology,
there would have been neither a reason to look for universal necessary
causes nor even a theoretical framework within which such causes
could have been conceived. The quest for a universal necessary cause
for beriberi makes sense only within the context of the etiological
standpoint.
In a recent book, Kenneth J. Carpenter concludes that:
[It was] the repeated failure of lime juice after 1860 that was most
important in leading investigators to doubt the whole structure of
experience and conclusions about scurvy that had been built up in
the previous two hundred years … . In this situation the investiga-
tor was induced to look at the whole subject anew, rejecting earlier
knowledge which now seemed unreliable, and accepting only the
first-hand experiences and ways of thought of his own generation.
(Carpenter, 1986, pp. 250f.)
It may well have been that the failure of lime juice after 1860 motivated
scientists to look again at the etiology of scurvy. But, quite obviously,
the ‘ways of thought of his own generation’ that the early twentieth-
century investigator found most compelling were those dictated by the
quest for universal necessary causes. Without this ‘way of thought’
there was no universal cause of scurvy (or of beriberi) and, therefore,
while there may have been piecemeal changes in what people thought
about the disease and how to control it, there was no possibility for
doubting ‘the whole structure of [earlier] experience and conclusions’
about the disease. Unfortunately, except for this one vague allusion,
Carpenter totally misses the broader story.
Thus, as with every other etiological investigation in the period,
research on the deficiency diseases was informed by the heuristic of the
research programme of which it was a part. And, more than anything
else, the stunning success of the bacterial theory insured universal
acceptance of that programme. Thus, far from merely presenting an
obstacle to researchers studying the deficiency (and other non-bacterial)
diseases, the success of the bacterial theory was the most vivid source
both of motivation for their work and of the standards for comprehen-
sion that their etiological investigations were required to satisfy.
Notes
1. In this earlier paper (‘The Germ Theory, Beriberi, and the Deficiency Theory
of Disease’), which forms the basis for this chapter, I took a position that I
now regard as incorrect. It was not ‘the germ theory [per se] that revived
interest in scurvy and … motivated researchers to seek a theoretical under-
standing of scurvy’ (Carter, 1977, p. 136), but rather what I am here
calling the etiological research programme of which the germ theory [more
precisely the bacterial theory] of disease was a part. So while I still think
that my earlier argument is essentially correct, the conclusion as stated in
the earlier paper requires modification.
2. Kenneth J. Carpenter has recently written books on scurvy and beriberi
(1986, 2000). Carpenter is much more reserved in his judgments about the
relation of bacteriology and the deficiency concept than those quoted
above so they constitute the foil for my argument. Carpenter provides
more details about research on both scurvy and beriberi, but, from the
point of view of this book, what he says neither adds to nor (with one
minor exception to be discussed below) challenges the position I advocated
in my 1977 paper or that I now present. Therefore, his interesting and
engaging work has not occasioned any substantial change in my argument.
Some Final Thoughts
In these case studies, my goals have been twofold: to present modern
medicine as a Lakatosian research programme focusing on the quest for
universal necessary causes, and to identify some of the heuristic princi-
ples required in that undertaking. The cases we have examined are by
no means exhaustive. Alfred S. Evans has shown that Koch’s Postulates
can be expanded to encompass diseases of several different kinds – for
example, those caused by ‘slow viruses’ and occupational and immuno-
logical diseases (Evans, 1993). Given the broad interpretation of the
Postulates proposed in Chapter 8, Evans’s account is obviously compat-
ible with my objectives. In the last two decades, other writers have
discussed still other diseases in ways that also fit well with the point of
view here presented (Ghesquier, 1999; Harden, 1987 and 1992; Kunitz,
1988; Schlich, 1994). The process by which causes were finally identi-
fied for chromosomal disorders provides a further example (Carter,
2002).
This way of thinking reveals mistakes in current beliefs about the rise
of modern medicine. Thinking that nineteenth-century medicine means
the germ theory, historians have almost totally ignored the broader
movement of which it was a part. As a result, most writers entirely
overlook the role of the research programme in the origins of Freudian
psychology. A related flaw is evident in accounts of the deficiency
theory: historians see only competition with the germ theory and are
oblivious of the movement that nurtured them both and endowed each
with its own objectives and research orientation. By failing to grasp the
nature of the programme, historians also misconstrue the work of those
who contributed to it. For example, Semmelweis is seen only as having
introduced chlorine washings – in spite of his own clear and repeated
claims that his opinions mattered most and distinguished him from
other early advocates of disinfection and in spite of subsequent obstetri-
cians who praised his theory of childbed fever because of its explanatory
power. Finally, as another example, there has been almost universal
neglect of early nineteenth-century studies of parasitic diseases, yet
these were arguably more important in the rise of medicine than was
pathological anatomy (which, partly through the unfortunate influence
of Foucault, has attracted far more attention than its significance war-
rants). In my view, Rennucci, Bassi, and Gruby, Semmelweis, Davaine,
Klebs, Pasteur and Koch, the early Freud, Loeffler and Frosch, Eijkman
and Funk (each of whose work was, demonstrably, focused almost
exclusively on the quest for universal necessary causes) and virtually
SOME FINAL THOUGHTS 197
every other medical researcher since the demise of chimbuki-medicine

have all contributed to the same research programme. ‘The final hope
and aim of medical science is the establishment of monogenic disease
entities’ (Taylor, 1979, p. 21); it can’t be made clearer than that.
By identifying heuristic principles essential to this programme – for
example, the Distinguishability Hypothesis – our investigation has also
underscored the importance of theory in the quest for causes. One
might expect that historians of medicine, like many who write on the
history of the physical sciences, would give careful attention to medical
theories. No such luck. Theory, and every other manifestation of rea-
son, gets short shrift these days. In William F. Bynum’s recent book,
Science and the Practice of Medicine in the Nineteenth Century, the
phrase ‘etiological theories’ appears only once; it appears in the follow-
ing subordinate index entry: ‘see also germ theory, miasmata, etiological
theories’ (Bynum, 1994, p. 262). But there is nothing else to see; the
phrase ‘etiological theories’ appears nowhere else in the index and, so
far as I can determine, Bynum never once used the phrase in the entire
body of his book (Carter, 1995, p. 496). To be sure, in a five-page
section entitled ‘bacteriology’, Bynum devotes just over one page (!) to
Koch’s Postulates (Bynum, 1994, pp. 129f.). Otherwise there is no
discussion whatsoever of what Adolf Strümpell called ‘the most charac-
teristic thrust of modern pathology (Strümpell, 1884a, p. 2) – namely,
the quest for universal necessary causes. Bynum’s neglect is typical and
symptomatic. Influenced by one brand of postmodernism – the French
Disease – historians have become suspicious of anything that smacks of
rationality, so scientific theories are marginalized. Today’s historians
look, instead, at the power structures of medicine (professional organi-
zations, medical schools) and at medicine among the unempowered
(irregular practitioners, treatment of the insane, the Poor Laws). As a
result, we know less about the nature and origin of the medical theories
that affect our lives than about, say, eighteenth-century quackery or the
average income of general practitioners in Victorian England. And that,
I think, is an abrogation of responsibility.
‘But,’ one early reviewer, asked: ‘what would Carter say about
Latour?’. Bruno Latour has shown (in a lively manner replete with the
quaint metaphors one has come to expect in Gallic ‘philosophy’) how
Pasteur and members of the early French Hygienic movement managed
to Pasteurize France – that is, managed bilaterally to leverage them-
selves into nearly absolute hegemony (Latour, 1984). But we are not
talking here about hula-hoops or polyester leisure suits. At the end of
the day, Pasteur advanced rational scientific theories in support of
which he gave evidence and made arguments. As an exercise in ‘let’s
pretend’, suppose none of Pasteur’s evidence was relevant to the process
by which his theories won out (which, by the way, reduces Pasteur
himself to little more than a fool). Suppose, as Latour seems to fancy,
power relations alone explain why any one theory surpasses its rivals
(p. 153). So what? Even if this were true, at some point scientific
theories deserve attention as examples of the best thinking the species
has yet achieved (‘There’s no point in listening to Bach’s music any
more; now we understand the system of patronage that gave him influ-
ence at court.’). So what can be the point (other than to sustain a fad
nearly on a par with leisure suits) of writing books arguing that one can
dispense with arguments? Moreover, Latour, like everyone else, sees
only bacteriology – not the quest for universal necessary causes. French
bacteriology, a local phenomenon after all, is less in need of explanation
than is the research programme of which it was one part. Perhaps
Latour should be arguing that the origin and growth of the entire
international programme can be reduced to social forces. That would
be a more interesting thesis and it has a certain plausibility – the
broader an intellectual change, the less likely it can be explained in
strictly rational terms. Moreover, that thesis might at least force one to
transcend Gallic chauvinism.
As Koch’s Postulates are usually explained in introductory textbooks,

the goal seems to be proving causation by routine procedures (isolation,
cultivation, and inoculation) that generate empirical evidence of suffi-
ciency; these accounts invariably ignore the theoretical assumptions
that, alone, can breathe life and meaning into such procedures. How-
ever, ever since Kant refuted Hume we have known that causation is
not an empirical relation. Apparently without realizing that they are re-
enacting history, some writers suggest that causes can be conjured from
the dusty boneyard that is empiricism through the judicious use of
induction (Parascandola, 1998, p. 320). But since we can’t observe
causes in individual cases, we can’t induce universal causes from par-
ticular cases either. Rather, causation is inherently theoretical and, in
the absence of an accepted theory, no amount of empirical evidence can
demonstrate causal relations – as we saw illustrated in Davaine’s dispute
with his critics.
While it is often overlooked, Koch was clear that empirical correla-
tions alone can never prove causality. In ‘Investigations of the Etiology
of Wound Infections’, he repeatedly insisted that proving causation
required showing that ‘the number and distribution of bacteria [is]
appropriate to explain completely the disease symptoms’ (Koch, 1878a,
pp. 27, 29, 48), and his papers on tuberculosis and cholera contain
similar language. More recent scientists have made analogous observa-
tions. For example, in concluding an essay on the application of Koch’s

Postulates to viruses, Thomas M. Rivers urged that research be ‘tem-
pered by the priceless attributes of common sense, proper training and
sound reasoning’ (Rivers, 1937, p. 11). But the need for such attributes
varies directly with the degree to which one’s work involves theoretical
considerations: counting requires little training or common sense. In a
reflective paper on proving the etiological significance of non-living
toxins, J. Yerushalmy and Carroll E. Palmer advised researchers to
examine the relation between similar and related possible causal factors
and the disease in question (Yerushalmy and Palmer, 1959, p. 38). But
such examinations can yield only analogical arguments that are inher-
ently theoretical and not subject to direct empirical confirmation – how,
after all, does one decide which factors are similar and related except by
an appeal to theory? At the end of his own list of mostly empirical
causal criteria, Alfred S. Evans notes that ‘the whole thing should make
biologic and epidemiologic sense’ (Evans, 1976, p. 192). He later wrote
that he added this requirement ‘with tongue in cheek’ (Evans, 1989,
p. 108). The problem with his last criterion – perhaps the reason he
regarded it as less than fully serious and why it is so seldom addressed
explicitly – may be that, while everyone recognizes intuitively that
‘making sense’ of the world is the whole point in looking for causes and
an essential standard against which purported causes must always be
judged, it is absolutely impossible to specify, in advance, what it will
take to make sense of the world or how one goes about doing it. It is
nearly impossible, sometimes until long after the fact, to be sure that the
causes we hypothesize really do make sense of anything at all. Thus, the
maximally and irremediably vague notion of ‘making sense’ lies at the
heart of the concept of causation. This shows how totally pointless,
hopeless, and downright silly it is to think one can ever state precisely
what it is for one thing to cause another.1
So, popular versions of Koch’s Postulate suggest that causation is
ultimately an empirical relation and that establishing causation resem-
bles counting. But all the while we know in our hearts – and our bright
students quickly figure out – that our hypotheses must also satisfy a
different standard, one on a completely different conceptual level, because
ultimately, making sense (biological relevance, theoretical adequacy,
explanatory power) matters at least as much as any empirical correla-
tion.2
We have traced some aspects of the rise of the etiological research

programme and we have seen the enormous increase in explanatory
power that this programme achieved. What, if anything, can we expect
of future causal thinking in medicine? Will the quest for universal

necessary causes continue to dominate medical research? We begin with
a minor philosophical dispute in contemporary epidemiology.
Nowadays epidemiologists get criticized for how they talk (Stehbens,
1992), and sometimes this seems to foster a sense of insecurity – some
epidemiologists apparently worry about whether theirs is only ‘second-
rate science’ (Parascandola, 1998). But do the criticisms make sense?
‘At the crux of the issue’, William E. Stehbens insists, ‘is the use and
meaning of cause in medicine’ (Stehbens, 1992, p. 97). He finds current
epidemiological usage teetering on the brink of imprecision, and the
fate of scientific medicine seems to hang in the balance: ‘If epidemiology
continues to disregard the misusage of terminology … it threatens the
very survival of logic and science in medicine’ (Stehbens, 1992, p. 116).
How can we save our way of life? our civilization? the self-respect of
our embattled epidemiologists? Stehbens recommends paying strict heed
to what Aristotle, Spinoza, Hume, Mill, and Bertand Russell had to say
about causes.
However, one must be cautious about using philosophy to regulate
scientific discourse, because most of philosophy is inherently retrospec-
tive. Hegel put it this way: ‘When philosophy paints its gray in gray,
then has a shape of life grown old. By philosophy’s grey in grey it
cannot be rejuvenated but only understood. The owl of Minerva spreads
its wings only with the falling of the dusk’ (Hegel, 1821, p. 7). As Hegel
saw, while philosophy describes the way words have been used, it says
little about how words will be (or should be) used in the future. Anthony
Flew (borrowing an image from Rudolf Carnap) explained that to think
otherwise would be like an explorer saying: ‘This country has its rivers
in the wrong place! Look, my map shows where they are supposed to
be’ (Flew, 1966, p. 150). Philosophy occasionally contributes new ways
of thinking or helps us refine our use of specific terms, but past philoso-
phers seldom provide much help when it comes to deciding how words
should be used in the future.
In contrast to philosophy, most important scientific advances depend
on changes in how we talk. ‘As ideas are preserved and communicated by
means of words, it necessarily follows that we cannot improve the lan-
guage of any science without at the same time improving the science
itself; neither can we, on the other hand, improve a science without
improving the language or nomenclature which belongs to it’ (Lavoisier,
1789, p. 1). Between about 1830 and 1880, medicine reorganized itself
around the concept of universal necessary causes. Robert Koch referred
to the new way of thinking as the etiological standpoint (Koch, 1901, p.
905). As we have seen, before that time, no one spoke about the cause of
anthrax or the cause of tuberculosis – such phrases literally had no
meaning. This is because, as diseases were then defined, virtually none of

them could have had that sort of cause. Recharacterizing diseases in
terms of necessary causes was a major advance that made possible coher-
ent explanations of disease phenomena as well as systematic treatment
and prophylaxis. Currently, medical science seeks such characterizations.
But this tells us nothing about whether we will continue, in the future, to
characterize diseases in this way. The reason is that one can never know
in advance what kinds of explanations will be required to make sense of
future observations. Judging from current developments in cosmology,
our observations are likely to get weirder and weirder, and our explana-
tions will probably follow suit.
One reason science is no longer done in Latin is that dead languages
cannot accommodate the linguistic adjustments on which all good sci-
ence depends. And any contemporary who favors quaint anachronisms
like causa vera sine qua non (Stehbens, 1992, p. 115) is either confused
or making a naked attempt to obfuscate the reader – probably both.
We make up words and give them meanings, and they can be used
however we like. But this does not imply that meanings should be
changed capriciously. Current scientific usage embodies what is argu-
ably the best thinking that our species has achieved through several
thousand years of intellectual groping, and it should be adjusted cau-
tiously only when there are compelling reasons for doing so. On the
other hand, as we have seen, language must be open to change if science
is to progress at all.
It isn’t a fact about the world that some cases of illness come pack-
aged together with universal necessary causes; rather, it’s a fact about
how we have learned to characterize diseases. When possible, we define
diseases so they have necessary causes because there is great utility in
doing so. But there may be diseases for which this is impossible either in
principle or simply in fact.
On the one hand, it is impossible, in principle, to explain why one
unstable atom decays at a certain time while identical atoms around it
do not. As currently understood, there are no sufficient causes for such
events; they simply happen. This is a metaphysical issue not an episte-
mological one; it stems from the way the world is, not from the weakness
of our intellects. By a loose analogy, there could conceivably be indi-
vidual cases of illness (perhaps those that start on the chromosomal
level) whose beginnings have no sufficient causes; they may simply
happen. It would obviously be impossible to group such cases in such a
way that a new disease could be so characterized that all its instances
would share a necessary cause. If this is how the world turns out to be,
the very most one could hope for would be risk factors (and even these
may ultimately prove impossible). Of course, even if there were such
illnesses, we would probably never come to know it and, in the absence
of such knowledge, we could continue seeking non-existent causes for
as long as the species shall live.
Alternatively, there could be cases of illness that do share sufficient
causes but causes whose identification requires more intellectual, finan-
cial and/or temporal assets than the human species will ever have at its
disposal. In short, there could be causes that we simply are not smart
enough to identify.
As a species, we have spent more time and money on cancer research
than on any other single problem in our history. Yet, while we have
learned a lot, we may be no closer to identifying sufficient causes in
terms of which classes of cancers can be characterized than we were
100 years ago.3 And this could be because there simply are no such
causes or it could be because, so far (and maybe forever), we are not
smart enough to find them. Perhaps at this point (and maybe forever),
the best we can do is find risk factors. So how should we respond? By
wringing our hands and insisting that no one speaks until we find
causae verae sine qua non? Of course not. We do our best to understand
and to control the world in whatever ways we can, and this may require
changes in how we talk.
We all know that risk factors are different from the causes Koch
identified. So what? At some point we may confront diseases for
which causes of his kind can never be found. But language evolves to
accommodate and to foster new ways of thinking. What form of
speech should we adopt? We cannot prescribe for the future, but here
are two obvious candidates: either we could exploit the similarity
(family resemblance) between traditional causes and theory-poor sta-
tistical correlations and simply broaden our concept of causation to
include some risk factors; or we could continue to talk about risk
factors as different from traditional causes. At this point, the choice
seems arbitrary. Scientific medicine is unlikely to collapse, whichever
course we adopt. But, however we end up speaking, this much is
absolutely certain: just as dead languages have no names for whatever
new concept of causation may emerge, dead philosophers (Aristotle,
Hume, Mill, Popper, and so on) are in no position to advise epidemi-
ologists about how to talk.
Half a century from now, historians may document when and how
we resolved these issues (assuming they finally get around to looking at
theories again), and philosophers, with their gray in gray, may clarify
why the resolution took the form it did. But how we talk about causes
will certainly be no more binding on scientists in 2060 than James L.
Bardsley’s way of talking about the causes of diabetes is binding on us

today.
Notes
1. Which, of course, doesn’t deter the bold from trying. First try: ‘Cause is
here used as the essential antecedent [i.e. the cause] or sole prerequisite
determinant [i.e. the cause] by which the disease is brought about [i.e.
caused]’ (Stehbens, 1992, pp. 99f.). In short: a cause is the cause by which
the disease is caused. Second try: ‘Cause is something that occasions [i.e.
causes], produces [i.e. causes], or effects [i.e. causes] a disease’ (Stehbens,
1992, p. 102). In short: a cause is something that causes, causes, or causes
a disease.
2. And that, by the way, seems to me to have been – at least at the beginning –
the real point of contention between Peter Duesberg and his opponents in
the AIDS establishment. While the arguments were generally stated in
terms of the necessity and sufficiency conditions of Koch’s Postulates, the
real issue was usually whether a retrovirus like HIV could possibly explain
AIDS.
3. To be sure, we may be a whole lot closer, but, until such causes are found,
there is no telling – not knowing where the goal lies, we can’t be sure
whether we are getting closer to it or not.
Bibliography
l’Académie Impériale de Médecine (1858), De la fièvre puerpérale,
Paris: Baillière et fils.
Ackerknecht, Erwin H. (1946), ‘Natural Diseases and Rational Treat-
ment in Primitive Medicine’, Bulletin of the History of Medicine,
19:467–497.
Adelon, Nicholas Philibert (ed.) (1832–46), Dictionnaire de médecine,
2nd edn, 30 vols, Paris: Bechet.
Anderson, William (1876), ‘On Kakke, or the Beri-beri of Japan’, Saint
Thomas Hospital Reports, 7:5–30.
Andersson, Ola (1962), Studies in the Prehistory of Psychoanalysis,
Stockholm: Svenska Bokförlaget.
Andral, Gabriel (1832–33), ‘Hydrophobia’, Lancet, 1:806–808.
Audouin, Jean Victor (1836a), ‘Recherches anatomiques et physiologiques
sur la maladie contagieuse qui attaque les vers à soie, et qu’on désigne
sous le nom de muscardine’, Annales des Sciences Naturelle: Zoologie,
2nd series, 8:229–245.
——— (1836b), ‘Nouvelles expériences sur la nature de la maladie
contagieuse qui attaque les vers à soie, et qu’on désigne sous le nom
de muscardine’, Annales des Sciences Naturelle: Zoologie, 2nd series,
8:257–269.
Balsamo-Crivelli, Joseph (1839), ‘Ueber den Ursprung und die
Entwicklung der Botrytis bassiana und eine andere schmorotzende
Art von Schimmel’, Linnaea, 16:118–123.
Bamberger, Heinrich von (1883), ‘Hysterie’, Allgemeine wiener
medizinische Zeitung, 28:529f.
Bar, Paul (1883), Des méthodes antiseptiques en obstétrique, Paris:
Alexandre Coccoz.
Bardsley, James L. (1845), ‘Diabetes’, in Dunglison (1845), vol. 1, pp.
606–625.
Barlow, Edward (1845a), ‘Education (Physical)’, in Dunglison (1845),
vol. 1, pp. 750–765.
——— (1845b), ‘Plethora’, in Dunglison (1845), vol. 3, pp. 553–574.
Barnes, David S. (2000), ‘Historical Perspectives on the Etiology of
Tuberculosis’, Microbes and Infection, 2:431–440.
Bassi, Agostino (1835), Del mal del segno, calcinaccio o moscardino,
trans. P.J. Yarrow, ed. G.C. Ainsworth, (1958), Ithaca, NY: Phy-
topathological Classics, Number 10.
Beaude, Jean Pierre (1849), ‘Anthrax’, in J.-P. Beaude (ed.), Dictionnaire
BIBLIOGRAPHY 205
de médecine usuelle à l’usage des gens du mode, 2 vols, Paris: Didier,

vol. 1, pp. 129f.
Beijerinck, Martinus Willem (1898), ‘Ueber ein Contagium vivum fluidum
als Ursache der Flackenkrankheit der Tabaksblätter’, Verhandelingen
koniinklijke Akademie van Wetenscharppen te Amsterdam, 65:3–21.
——— (1899), ‘Ueber ein Contagium vivum fluidum als Ursache der
Fleckenkrankheit der Tabaksblätter’, Abteilung II, Centralblatt für
Bakteriologie, Parasitenkunde und Infektionskrankheiten, 5:27–33.
Berliner klinische Wochenschrift (1877), [Review of Klebs’ ‘Ueber die
Umgestaltungen der medicinischen Anschauungen in den letzten drei
Jahrzehnten’], Berliner klinische Wochenscrift 14: 594.
Bert, Paul (1876), ‘Nouvelles recherches sur le sang de rate’, Comptes
rendus de Séances et Mémoires de la Société de Biologie, 28:380–381.
——— (1877), ‘Sur la nature du charbon’, Comptes rendus de Séances
et Mémoires de la Société de Biologie, 29:317–321.
Biett, L. (1836), ‘Gale’, in Adelon (1932–46), vol. 13, pp. 545–549.
Billet, Léon (1872), De la fièvre puerpérale et de la réforme des
maternités, Paris: H.-B. Baillière et fils.
Birch-Hirschfeld, Felix Victor (1872), ‘Die neuern pathologischana-
tomischen Untersuchungen über krankmachende Schmarotzerpilze’,
Schmidts Jahrbücher der in- und ausländischen Medizin, 155:97–109.
——— (1875), ‘Die neuern pathologischanatomischen Untersuchungen
über Vorkommen und Bedeutung niederer Pilzformen (Bakterien) bei
Infektionskrankheiten’, Schmidts Jahrbücher der in- und ausländischen
Medizin, 166:169–223.
——— (1877), Lehrbuch der pathologischen Anatomie, Leipzig: F.C.W.
Vogel.
Blane, Gilbert (1785), Observations on the Diseases Incident to Sea-
men, in Christopher C. Lloyd (ed.) (1965), The Health of Seamen,
London: Naval Records Society, pp. 1–211.
Bloomfield, Arthur L. (1958), A Bibliography of Internal
Medicine: Communicable Diseases, Chicago: The University of Chi-
cago Press.
Boehr, Max (1868), ‘Ueber die Infectionstheorie des Puerperalfiebers
und ihre Consequenzen für die Sanitäts-Polizei’, Monatsschrift für
Geburtskunde und Frauenkrankheiten, 32:401–433.
Bollinger, Otto (1875), ‘Infectionen durch thierische Gifte (Milzbrand)’,
in Hugo Wilhelm von Ziemssen (ed.), Handbuch der speciellen
Pathologie und Therapie, 4 vols, Leipzig: F.C.W. Vogel, vol. 3, pp. 447–
490.
Böttger, Herbert (1955), ‘Förderer der Semmelweisschen Lehre’, Sudhoffs
Archiv, 39:341–362.
206 BIBLIOGRAPHY
Braddon, W. Leonard (1907), The Cause and Prevention of Beriberi,

London: Rebman.
Brauell, Friedrich August (1857), ‘Versuche und Untersuchungen betreffend
den Milzbrand des Menschen und der Thiere’, Archiv für pathologische
Anatomie, Physiologie, und klinische Medizine (Virchows Archiv),
11:132–144.
——— (1858), ‘Weitere Mittheilungen über Milzbrand und
Milzbrandblut’, Archiv für pathologische Anatomie, Physiologie, und
klinische Medizine (Virchows Archiv), 14:432–466.
——— (1866), ‘Zur Milzbrand-Frage’, Archiv für pathologische Anatomie,
Physiologie, und klinische Medizine (Virchows Archiv), 36:292–297.
Braun, Carl (1855), ‘Kindbettfieber’, in J.B. Chiari, C. Braun and J.
Späth (eds), Klinik der Geburtshilfe und Gynakologie, Erlangen:
Ferdinand Enke.
——— (1857), Lehrbuch der Geburtshülfe, Wien: Wilhelm Braumüller.
Breisky, August (1861), ‘Semmelweis: Die Aetiologie, der Begriff und
die Prophylaxis des Kindbettfiebers’, Vierteljahrschrift für praktsche
Heilkunde, 18, Literärischer Anzeiger, pp. 1–13.
Breuer, Josef and Sigmund Freud (1893), ‘On the Psychical Mechanism
of Hysterical Phenomena: Preliminary Communication’, in Strachey
(1955–74), vol. 2, pp. 3–17.
British Medical Journal (1898), ‘An Eccentric Physician’, British Medi-
cal Journal, 2:1705.
——— (1907), ‘Scurvy’, British Medical Journal, 1:683.
——— (1910), ‘The Far Eastern Association of Tropical Medicine’,
British Medical Journal, 1:999f.
Brock, Thomas D. (1988), Robert Koch: A Life in Medicine and Bacte-
riology, Madison, WI: Science Tech Publishers.
Brown, Joseph (1845), ‘Contagion’, in Dunglison (1845), vol. 1, pp.
500–505.
Bruce-Chwatt, Leonard J. (1988), ‘History of Malaria from Prehistory
to Eradiction’, in Walter H. Wernsdorfer and Ian McGregor (eds),
Malaria: Principles and Practice of Malariology, 2 vols, Edinburgh:
Churchill Livingstone, vol. 1, pp. 1–59.
Budd, George (1840), ‘Scurvy’, in Alexander Tweedie (ed.), A System of
Practical Medicine, 5 vols, London: Whittaker, vol. 5, pp. 58–95.
Bynum, William F. (1994), Science and the Practice of Medicine in the
Nineteenth Century, Cambridge: Cambridge University Press.
Campbell, William C. (1979), ‘History of Trichinosis: Paget, Owen and
the Discovery of Trichinella Spiralis’, Bulletin of the History of Medi-
cine, 53:520–552.
Carpenter, Kenneth J. (1986), The History of Scurvy and Vitamin C,
Cambridge: Cambridge University Press.
BIBLIOGRAPHY 207
——— (2000), Rice, Beriberi and Vitamin B, Berkeley: University of

California Press.
Carter, K. Codell (1977), ‘The Germ Theory, Beriberi, and the Defi-
ciency Theory of Disease’, Medical History, 21:119–136.
——— (1980), ‘Germ Theory, Hysteria, and Freud’s Early Work in
Psychopathology’, Medical History, 24:259–274.
——— (1982a), ‘On the Decline of Bloodletting in 19th Century Medi-
cine’, The Journal of Psychoanalytic Anthropology, 5:219–234.
——— (1982b), ‘Nineteenth-Century Treatments for Rabies as Reported
in the Lancet’, Medical History, 26:67–78.
——— (1985a), ‘Ignaz Semmelweis, Carl Mayrhofer, and the Rise of
Germ Theory’, Medical History, 29:33–53.
——— (1985b), ‘Koch’s Postulates in Relation to the Work of Jacob
Henle and Edwin Klebs’, Medical History, 29:353–374.
——— (trans.) (1987a), Essays of Robert Koch, New York: Greenwood
Press.
——— (1987b), ‘Edwin Klebs’ Criteria for Disease Causality’,
Medizinhistorisches Journal, 22:80–89.
——— (1991), ‘Causes of Disease and Death in the Babylonian Tal-
mud’, Medizinhistorisches Journal, 26:94–104.
——— (1993), ‘The Concept of Quackery in Early Nineteenth Century
British Medical Periodicals’, The Journal of Medical Humanities,
14:89–97.
——— (1995), ‘Essay Review: Toward a Rational History of Medical
Science’, Studies in the History and Philosophy of Medical Science,
26:493–502.
——— (1997), ‘Causes of Disease and Causes of Death’, Continuity
and Change, 12:189–198.
——— (2002), ‘Early Conjectures that Down Syndrome is Caused by
Chromosomal Nondisjunction’, Bulletin of the History of Medicine,
76:528–563.
——— and George S. Tate (1991), ‘The Earliest-Known Account of
Semmelweis’s Initiation of Disinfection at Vienna’s Allgemeines
Krankenhaus’, Bulletin of the History of Medicine, 65:252–257.
———, Scott Abbott, and James L. Siebach (1995), ‘Five Documents
Relating to the Final Illness and Death of Ignaz Semmelweis’, Bulletin
of the History of Medicine, 69:255–270.
Castiglioni, Arturo (1947), A History of Medicine, 2nd edn, trans. E.B.
Krumbhaar, New York: Alfred A. Knopf.
Charcot, Jean Martin (1883), ‘Deux cas de contracture hystérique
d’origine traumatique’, Progrès médical, 11:37–39.
——— (1888–94), Oeuvres complètes, M. Babinski et al. (eds), 9 vols,
Paris: Bureaux du Progrès Médical.
208 BIBLIOGRAPHY
——— (1889), Clinical Lectures on Diseases of the Nervous System,

trans. Thomas Savill, London: New Sydenham Society.
——— (1892–95), Poliklinische Vorträge, trans. Sigmund Freud, 2 vols,
Leipzig: Franz Deuticke.
Chomel, M. (1835), ‘Etiologie’, in Adelon (1832–46), vol. 12, pp. 415–
425.
——— (1842), ‘Pneumonie’, in Adelon (1832–46), vol. 25, pp. 144–
232.
Cohn, Ferdinand (1872), ‘Untersuchungen über Bakterien’, Beiträge zur
Biologie der Pflanzen, 2:128–222.
——— (1875), ‘Untersuchungen über Bakterien’, Beiträge zur Biologie
der Pflanzen, 3:141–204.
Cohn, Martin (1883), ‘Ueber die Psychosen im kindlichen Alter’, Archiv
für Kinderheilkunde, 4:28–64, 102–107.
Cohnheim, Julius Friedrich (1877), Vorlesungen über allgemenine
Pathologie, 2 vols, Berlin: August Hirschwald.
——— and Carl Julius Salomonsen (1877) [Report of discussion]
Jahresbericht der schlesische Gesellschaft für vaterländische Kultur,
55:222f.
Colin, Léon (1880a), ‘Présentation d’ouvrages manuscrits et imprimés’,
Bulletin de l’Académie de médecine, 2nd series, 9:1235f.
——— (1880b), ‘Présentation d’ouvrages manuscrits et imprimés’, Bul-
letin de l’Académie de médecine, 2nd series, 9:1346f.
Collard, Patrick (1976), The Development of Microbiology, Cambridge:
Cambridge University Press.
Conolly, John (1845), ‘Disease’, in Dunglison (1845), vol. 1, pp. 674–
689.
Conrad, Peter and Joseph W. Schneider (1980), Deviance and
Medicalization: From Badness to Sickness, St Louis: C.V. Mosby.
Copi, Irving R. (1979), Symbolic Logic, 5th edn, New York: Macmillan
Publishing Co.
Coze, Leon and Victor-Timothee Feltz (1866), ‘Recherches expérimentales
sur la présence des infusoires et l’état du sang dans les maladies
infectieuses’, Gazette médical de Strasbourg, 2nd series, 6:61–64,
115–125, 208f., 225–229.
——— and ——— (1872), Recherches cliniques et expérimentales sur
les maladies infectieuses, Paris: J.B. Baillière et fils.
Crede, Carl S.F. (1861), ‘Die Aetiologie, der Begriff und die Prophylaxis
des Kindbettfiebers von Ignaz Philipp Semmelweis’, Monatsschrift für
Geburtskunde, 18:406f.
Cuboni, Giuseppe and Ettore Marchiafava (1881), ‘Neue Studien über
die Natur der Malaria’, Archiv für experimentelle Pathologie und
Pharmakologie, 13:265–280.
BIBLIOGRAPHY 209
Cumin, William (1845), ‘Scrofula’, in Dunglison (1845), vol. 4, pp.

125–145.
D’Espine, H.-A. (1873), Contribution a l’étude de la septicémie
puerpérale, Paris: J.-B. Baillière et fils.
Darwin, Charles (1876), The Origin of Species, Vol 49, Great Books of
the Western World, ed. Robert M. Hutchins (1952), Chicago: Ency-
clopaedia Britannica.
Davaine, Casimir (1863a), ‘Recherches sur les infusoires du sang dans
la maladie connue sous le nom de sang de rate’, Comptes rendus de
l’Académie des sciences, 57:220–223.
——— (1863b), ‘Nouvelles recherches sur les infusoires du sang dans la
maladie connue sous le nom de sang de rate’, Comptes rendus de
l’Académie des sciences, 57:351–353, 386–387.
——— (1864a), ‘Nouvelles recherches sur la nature de la maladie
charbonneuse connue sous le nom de sang de rate’, Comptes rendus
de l’Académie des sciences, 59:393–396.
——— (1864b), ‘Recherches sur les Vibrioniens’, Comptes rendus de
——— (1868a), ‘Sur la nature des maladies charbonneuses’, Archives
générales de Médicine, Series 4, 11:i:144–148.
——— (1868b), ‘Expériences relative à la durée de l’incubation des
maladies charbonneuses et à la quantité de virus nécessaire à la
transmission de la maladie’, Bulletin de l’Académie de médecine,
33:816–821.
——— (1870a), ‘Études sur la contagion du charbon chez les animaux
domestiques’, Bulletin de l’Académie de médecine, 35:215–235.
——— (1870b), ‘Études sur la genèse et la propagation du charbon’,
Bulletin de l’Académie de médecine, 35:471–498.
——— and Raimbert (1864), ‘Sur la présence des Bactéridies dans la
pustule maligne chez l’homme’, Comptes rendus de l’Académie des
sciences, 59:429–431.
Decker, Hannah S. (1977), Freud in Germany, Psychological Issues,
Vol. 11, Monograph 41, New York: International Universities Press.
Dolman, Claude E. (1974), ‘Robert Koch’ in Charles Coulston Gillispie
(ed.), Dictionary of Scientific Biography, 15 vols, New York: Charles
Scribner’s Sons, vol. 7, pp. 420–435.
Douglas, Mary (1975), ‘Environments at Risk’ in Implicit Meanings,
London: Routledge & Kegan Paul, pp. 230–248.
Dracobly, Alex (2000), ‘Therapeutic Innovation and Philippe Ricord’s
“New Doctrine” of the Venereal Diseases’, paper presented at the
American Association for the History of Medicine conference,
Bethesda, 20 May 2000.
Dubois, Paul (1842), ‘Puerpérale (Fièvre)’ in Nicholas Philibert Adelon
210 BIBLIOGRAPHY
et al. (eds), Dictionnaire de médecine, 2nd edn, Paris: Bechet, vol. 26,
pp. 336–359.
Dudley, Sheldon F. (1953), ‘The Lind Tradition in the Royal Naval
Medical Service’, in C.P. Stewart and Douglas Guthrie (eds), Lind’s
Treatise on Scurvy, Edinburgh: Edinburgh University Press, pp. 369–
385.
Duméril et al. (1838), ‘Rapport sur divers travaux entrepris au sujet de
la maladie des vers à soie, connue volgairement sous le nom de
muscardine’, Comptes rendus de l’Académie des sciences, 6:66–102.
Dunglison, Robley (ed.) (1845), Cyclopaedia of Practical Medicine,
American Edition, 4 vols, Philadelphia: Lea and Blanchard.
Durham, Herbert E. (1904), ‘Notes on Beriberi in the Malay Peninsula
and on Christmas Island (Indian Ocean)’, Journal of Hygiene, Cam-
bridge, 4:112–155.
Eade, Peter (1880), ‘Civil Practice’, Lancet, 1:992f.
Eijkman, Christiaan (1897a), ‘Eine beriberi-ähnliche Krankheit der Huh-
ner’, Archiv für pathologische Anatomie, Physiologie, und klinische
Medizine (Virchows Archiv), 148:523–532.
——— (1897b), ‘Ein Versuch zur Bekampfung der Beriberi’, Archiv für
pathologische Anatomie, Physiologie, und klinische Medizine
(Virchows Archiv), 149:187–194.
——— (1906), ‘Ueber Ernährungspolyneuritis’, Archiv für Hygiene,
58:150–177.
Ellenberger, Henri F. (1970), The Discovery of the Unconscious, New
York: Basic Books.
Elliotson, John (1830–31), ‘Clinical Lectures: Scurvy’, Lancet, 1:650–
655.
——— (1844), Principles and Practice of Medicine, 2nd edn, Philadelpia:
Carey and Hart.
Evans, Alfred S. (1976), ‘Causation and Disease: The Henle-Koch Pos-
tulates Revisited’, Yale Journal of Biology and Medicine, 49:175–195.
——— (1989), ‘Does HIV Cause AIDS? an Historical Perspective’, Jour-
nal of Acquired Immune Deficiency Syndromes, 2:107–113.
——— (1993), Causation and Disease: A Chronological Journey, New
York: Plenum Medical Book Company.
Evans-Pritchard, E.E. (1976), Witchcraft, Oracles, and Magic Among
the Azende, abridged edn, Oxford: Oxford University Press.
Farr, William (1839), ‘Letter to the Registrar-General’, in First Annual
Report of the Registrar-General of Births, Deaths, and Marriages in
England, London: W. Clowes and Sons.
——— (1840), ‘Letter to the Registrar-General’, in Second Annual Re-
port of the Registrar-General of Births, Deaths, and Marriages in
England, London: W. Clowes and Sons.
BIBLIOGRAPHY 211
Ferber, Rudolf H. (1868), ‘Die Aetiologie, Prophylaxis und Therapie

des Puerperalfiebers’, Schmidts Jahrbucher der in- und ausländische
Medizin, 139:318–346.
Fischel, Wilhelm (1882), ‘Zur Therapie der puerperalen Sepsis’, Archiv
für Gynaekologie, 20:1–70.
Fitzgerald, Thomas J. (1991), ‘Treponema’, in Samuel Baron (ed.), Medi-
cal Microbiology, 3rd edn, New York: Churchill Livingstone, pp.
491–504.
Fletcher, William (1907), ‘Rice and Beriberi’, Lancet, 1:1776–1779.
Flew, Anthony (1966), God and Philosophy, New York: Dell.
Forbes, John (1845), ‘Angina Pectoris’, in Dunglison (1845), vol. 1, pp.
103–117.
Foster, George M. (1976), ‘Disease Etiologies in Non-Western Medical
Systems’, American Anthropologist, 78:773–782.
Foster, W.D. (1965), A History of Parasitology, Edinburgh: E. & S.
Livingstone.
Foucault, Michel (1973), The Birth of the Clinic, New York: Pantheon
Books.
Frankfort, H.A., John A. Wilson and Thorkild Jacobsen (1949), Before
Philosophy, Baltimore, MD: Penguin Books.
Fraser, Henry and A. Thomas Stanton (1909), ‘An Inquiry Concerning
the Etiology of Beriberi’, Lancet, 1:451–455.
——— and ——— (1910), ‘The Etiology of Beriberi’, Lancet, 2:1755–
1757.
Freud, Sigmund (1886), ‘Report on my Studies in Paris and Berlin’ in
James Strachey (ed.) (1955–74), The Standard Edition of the Com-
plete Psychological Works of Sigmund Freud, London: Hogarth Press,
vol. 1, pp. 5–15.
——— (1888), ‘Hysteria,’ in Strachey (1955–74), vol. 1, pp. 41–59.
——— (1893), ‘On the Theory of Hysterical Attacks’, in Strachey (1955–
74), vol. 1, pp. 151–154.
——— (1894), ‘The Neuro-Psychoses of Defense’, in Strachey (1955–
74), vol. 1, pp. 45–61.
——— (1895), ‘On the Grounds for Detaching a Particular Syndrome
from Neurasthenia under the Description “Anxiety Neurosis”’, in
Strachey (1955–74), vol. 3, pp. 90–117.
——— (1896a), ‘Heredity and the Aetiology of the Neuroses’, in Strachey
(1955–74), vol. 3, pp. 143–156.
——— (1896b), ‘Further Remarks on the Neuro-Psychoses of Defence’,
in Strachey (1955–74), vol. 3, pp. 162–185.
——— (1896c), ‘The Aetiology of Hysteria’, in Strachey (1955–74),
vol. 3, pp. 191–221.
212 BIBLIOGRAPHY
——— (1909), ‘Five Lectures on Psycho-Analysis’, in James Strachey

(1955–74), vol. 11, pp. 9–55.
——— (1925), An Autobiographical Study, in James Strachey (1955–
74), vol. 20, pp. 7–74.
——— (1954), The Origins of Psycho-Analysis: Letters to Wilhelm
Fliess, Drafts and Notes: 1887–1902, ed. Marie Bonaparte, Anna
Freud and Ernst Kris, New York: Basic Books.
Funk, Casimir (1912), ‘The Etiology of the Deficiency Diseases’, Jour-
nal of State Medicine, 20:341–368.
Garrison, Fielding H. (1929), Introduction to the History of Medicine,
4th edn, Philadelphia: W.B. Saunders.
Geison, Gerald L. (1974), ‘Pasteur’, in Charles Coulston Gillispie (ed.),
Dictionary of Scientific Biography, 15 vols, New York: Charles
Scribner’s Sons, vol. 10, pp. 350–416.
——— (1995), The Private Science of Louis Pasteur, Princeton, NJ:
Princeton University Press.
Gelfand, Toby (1989), ‘Charcot’s Response to Freud’s Rebellion’, Jour-
nal of the History of Ideas, 50:293–307.
Gerhardt, C. (1884), ‘Ueber Intermittens-Impfungen’, Zeitschrift für
klinische Medizin, 7:372–377.
Ghesquier, Danièle (1999), ‘A Gallic Affair: The Case of the missing
Itch-Mite in French Medicine in the early 19th Century’, Medical
History, 43:26–54.
Gras, Albin (1836), ‘Du rôle que joue l’acarus de l’homme dans la
production de la gale’, Bulletin de l’Académie de Médecine 1:77–82.
Grijns, Gerrit (1901), ‘Over polyneuritis gallinarum’, Geneeskundig
Tijdschrift voor Nederlandsch-Indië, 41:3–110.
Grist, D.H. (1965), Rice, 4th edn, London: Longmans.
Györy, Tiberius von (1905), Semmelweis’ gesammelte Werke, Jena:
Gustav Fischer.
Hall, Marshall (1837–38), ‘Lectures on the Theory and Practice of
Medicine: Scorbutus’, Lancet 2:851f.
Hamlin, Christopher (1992), ‘Predisposing Causes and Public Health in
Early 19th-Century Medical Thought’, Social History of Medicine,
5:43–70.
Hanson, Norwood Russell (1963), The Concept of the Positron, Cam-
bridge: Cambridge University Press.
——— (1969), Patterns of Discovery, Cambridge: Cambridge Univer-
sity Press.
Harden, Victoria A. (1987), ‘Koch’s Postulates and the Etiology of
Rickettsial Diseases’, Journal of the History of Medicine and Allied
Sciences, 42:277–295.
BIBLIOGRAPHY 213
——— (1992), ‘Koch’s Postulates and the Etiology of AIDS’, History

and Philosophy of the Life Sciences, 14:249–269.
Haussmann, David (1870), Die Parasiten der weiblichen
Geschlechtsorgane, Berlin: August Hirschwald.
——— (1874), ‘Untersuchungen und Versuche über die Entstehung der
übertragbaren Krankheiten des Wochenbettes’, Beiträge zur
Geburtshülfe und Gynäkologie 3:311–421.
Hegel, G.W.F. (1821), Preface to Philosophy of Right, Vol. 46, Great
Books of the Western World, ed. Robert M. Hutchins (1952), Chi-
cago: Encyclopaedia Britannica.
Henle, Jacob (1840), Pathologische Untersuchungen, Part 2 ‘Miasmata
and Contagia’, trans. George Rosen, Bulletin of the History of Medi-
cine, 1938, 6:911–983.
——— (1844a) ‘Medicinische Wissenschaft und Empirie’, Zeitschrift
für rationelle Medizin, 1:1–35.
——— (1844b) ‘Miasmatisch-Contagiöse Krankheiten’, Zeitschrift für
rationelle Medizin, 2:287–412.
Henoch, Eduard Heinrich (1881), Vorlesungen über Kinderkrankheiten,
Berlin: August Hirschwald.
Herbst, G. (1851–52), ‘Beobachtungen über Trichina spirallis in Betreff
der Uebertragung der Eideweidewürmer’, Nachrichten Georg-August
Universität. Königlich Gesellschaft für Wissenschaft Göttingen, 2:260–
264, 183–204.
Herz, Maximilian (1885), ‘Uber Hysterie bei Kindern’, Wiener
medizinische Wochenschrift, 34:cols 1305–1308, 1338–1342, 1368–
1371, 1401–1405.
Heymann, Bruno (1932), Robert Koch, Leipzig: Akademische Verlag.
Hirsch, August (1885), ‘Beriberi’, in Handbook of Geographical and
Historical Pathologie, 2nd edn, trans. Charles Creighton, London:
New Sydenham Society.
——— (1935), ‘Klebs’, Biographisches Lexikon hervorragenden Aerzte,
2nd edn, Berlin: Urban and Schwarzenberg.
Hoag, Junius C. (1887), ‘Puerperal fever and its treatment’, American
Journal of Obstetrics and Diseases of Women and Children, 20:828–
844, 941–957.
Holmes, Oliver Wendell (1843), ‘The Contagiousness of Puerperal Fe-
ver’, reprinted in Medical Essays (1883), New York: Houghton Mifflin.
Holst, Alex (1907), ‘Experimental Studies Relating to Ship-Beriberi and
Scurvy’, Journal of Hygiene, London, 7:619–633.
——— and Theodor Frolich (1907), ‘Experimental Studies Relating to
Ship-Beriberi and Scurvy’, Journal of Hygiene, London, 7:634–671.
Huebner, Robert J. (1957), ‘The Virologist’s Dilemma’, Annals of the
New York Academy of Science, 67:430–438.
214 BIBLIOGRAPHY
Hughes, Sally Smith (1977), The Virus: a History of the Concept, New
York: Science History Publications.
Huppert (1865), [Review of Anthrax Literature], Schmidts Jahrbucher
der in- und ausländische Medizin, 132:37–43.
Ihde, Aaron J. and Stanley L. Becker (1971), ‘Conflict of Concepts in
Early Vitamin Studies’, Journal of the History of Biology, 4:1–33.
Iwanovski, Dimitri Iosifovitch (1892), ‘Ueber die Mosiakkrankheit der
Tabakspflanze’, Bulletin Académie Impériale de Science St. Petersburg,
3:67–70.
——— (1902), ‘Die Mosiak- und die Pockenkrankheit der Tabakspflanze’,
Zeitschrift für Pflanzenkrankheiten, 7:202f.
——— (1903), ‘Ueber die Mosaikkrankheit der Tabakspflanze’,
Zeitschrift für Pflanzenkrankheiten, 13:1–41.
J.F.S. (1836), ‘Discovery of an Insect in Itch’, Lancet 1:59–62.
Jaggard, W.W. (1884), ‘The Pathology, Etiology, Prophylaxis, and Treat-
ment of Puerperal Fever, from the Vienna Standpoint’, Medical News,
44:442–445.
Jeannel, Maurice (1880), L’infection purulente ou pyohémie, Paris: J.-B.
Baillière et fils.
Joest, Ernst (1902), ‘Unbekannte Infektionsstoffe’, Centralblatt für
Bakteriologie Parasitenkunde und Infektionskrankheiten, Abt. I,
31:361–384, 410–422.
Jones, Ernest (1953), The Life and Works of Sigmund Freud, 3 vols,
New York: Basic Books.
Kamminga, Harmke (1993), ‘Taking Antecedent Conditions Seriously:
A Lesson in Heuristics from Biology’, in Steven French and Harmke
Kamminga (eds), Correspondence, Invariance and Heuristics: Essays
in Honor of Heinz Post, Dordrecht: Kluwer.
Kaufmann (1866), ‘Ueber die Ursachen des epidemischen Puerperalfiebers
in Gebäranstalten’, Monatsschrift für Geburtskunde und
Frauenkrankheiten, 26:422–424.
King, Lester S. (1982), Medical Thinking: A Historical Preface, Princeton,
NJ: Princeton University Press.
Klebs, Edwin (1865), ‘Zur Pathologie der epidemischen Meningitis’,
Archiv für pathologische Anatomie, Physiologie, und klinische
——— (1871), ‘Die Ursache der infectiösen Wundkrankheiten’,
Correspondenz-blatt für Schweizer Aerzte, 1:241–246.
——— (1872), Beiträge zur pathologischen Anatomie der Schusswunden,
Leipzig: V.C.W. Vogel.
——— (1873), ‘Beiträge Zur Kenntniss der Micrococcen’, Archiv für
experimentelle Pathologie und Pharmakologie, 1:31–64.
BIBLIOGRAPHY 215
——— (1874a), ‘Beiträge Zur Kenntniss der Micrococcen’, Archiv für

——— (1874b), ‘Micrococcen als Krankheitsursache’, Verhandlungen
der physicalische-medizinische Gesellschaft in Würzburg 6:vi–viii.
——— (1875–76), ‘Beiträge zur Kenntniss der pathogenen
Schistomyceten’, Archiv für experimentelle Pathologie und
Pharmakologie, 3:305–324, 4:107–136, 207–247, 409–488.
——— (1877), ‘Pathologisch-anatomische Demonstrationen’, Prager
medizinische Wochenschrift, 2:529–535.
——— (1878a), ‘Ueber die Umgestaltung der medizinische Anschauungen
in den letzten drei Jahrzehnten’, Gesellschaft der Deutsche
Naturforscher und Aerzte in München, Leipzig: Vogel.
——— (1878b), ‘Ueber Cellularpathologie und Infectionskrankheiten’,
Gesellschaft der deutsche Naturforscher und Aerzte in Cassel, Cassel:
Baier und Lewalter.
——— (1878c), ‘Notiz über die Ursache des Milzbrandes’, Archiv für
experimentelle Pathologie und Pharmakologie, 8:269f.
——— (1880), ‘Der Ileotyphus, eine Schistomycose’, Archiv für
——— (1881), ‘Der Bacillus des Abdominaltyphus und der typhöse
Process, Archiv für experimentelle Pathologie und Pharmakologie,
13:381–460.
——— and C. Tommasi-Crudeli (1879a), ‘Studien über die Ursache des
Wechselfiebers und über die Natur der Malaria’, Archiv für
——— and ——— (1879b), ‘Einige Sätze über die Ursachen der
Wechselfieber und die Natur der Malaria’, Archiv für experimentelle
Pathologie und Pharmakologie, 11:122–126.
Koch, Robert (1876), ‘The Etiology of Anthrax, Founded on the Course
of development of the Bacillus Anthracis’, in K. Codell Carter (trans.)
(1987), Essays of Robert Koch, New York: Greenwood Press, pp. 1–
17.
——— (1878a), ‘Investigations of the Etiology of Wound Infections’, in
Carter (1987), pp. 19–56.
——— (1878b), ‘Neue Untersuchungen über die Mikroorganismen bei
infektiösen Wundkrankheiten’, in J. Schwalbe (ed.) (1912), Gesammelte
Werke von Robert Koch, Leipzig: Georg Thieme, vol. 1, pp. 57–60.
——— (1881a), ‘Zur Untersuchung von pathogenen Organismen,’ in
Schwalbe (1912), vol. 1, pp. 112–163.
——— (1881b), ‘On the Etiology of Anthrax’, in Carter (1987), pp.
57–81.
——— (1882a), ‘The Etiology of Tuberculosis’, in Carter (1987), pp.
83–96.
216 BIBLIOGRAPHY
——— (1882b), ‘On the Anthrax Inoculation’, in Carter (1987), pp.

97–115.
——— (1882c), ‘Ueber die Aetiologie der Tuberkulose’, in Schwalbe
(1912), vol. 1, pp. 446–453.
——— (1884a), ‘Experimentelle Studien über die künstliche
Abschwächung der Milzbrandbazillen und Milzbrandinfektion durch
Fütterung’, in Schwalbe (1912), vol. 1, pp. 232–270.
——— (1884b), ‘The Etiology of Tuberculosis’, in Carter (1987), pp.
129–150.
——— (1884c), ‘On Cholera Bacteria’, in Carter (1987), pp. 171–177.
——— (1884d), ‘Lecture at the First Conference for Discussion of the
Cholera Question’, in Carter (1987), pp. 151–170.
——— (1887), ‘Ueber die Pasteurschen Milzbrandimpfungen’, in
Schwalbe (1912), vol. 1, pp. 271–273.
——— (1890), ‘On Bacteriological Research’, in Carter (1987), pp.
179–186.
——— (1901), ‘Massnahmen gegen die Pest’, in Schwalbe (1912), pp.
905–907.
——— (1909), ‘Antrittsrede in der Akademie der Wissenschaften am 1.
Juli 1909’, in Schwalbe (1912), vol. 1, pp. 1–4.
Koch, Wilhelm (1889), Die Bluterkrankheit in ihren Varianten, Stutt-
gart: Ferdinand Enke.
Kunitz, Stephen J. (1987), ‘Explanations and Ideologies of Mortality
Patterns’, Population and Development Review, 13:379–408.
——— (1988), ‘Hookworm and Pellagra: Exemplary Diseases in the
New South’, Journal of Health and Social Behavior, 29:139–148.
Lagasquie, A. (1849), ‘Causes’, in Jean Pierre Beaude (ed.), Dictionnaire
de médecine usuelle à l’usage des gens du mode, 2 vols, Paris: Didier,
vol. 1, p. 313.
Lakatos, Imre (1968), ‘Falsification and the Methodology of Scientific
Research Programmes’, in Imre Lakatos and Alan Musgrave (eds)
(1974), Criticism and the Growth of Knowledge, Cambridge: Cam-
bridge University Press, pp. 91–196.
——— (1971), ‘History of Science and its Rational Reconstructions’, in
John Worrall and Gregory Currie (eds) (1974), The Methodology of
Scientific Research Programmes, Cambridge: Cambridge University
Press, pp. 102–138.
Lancet (1842–43), ‘London Medical Society’, Lancet, 1:879.
Lancet (1858), ‘Scurvy in the Merchant Fleet’, Lancet, 1:145f.
Lancet (1886), ‘Pathology of Scurvy’, Lancet, 1:1036.
Lancet (1887), ‘Kakké, or Japanese Beriberi’, Lancet, 2:233f.
Lancet (1890), [Review of William Koch, Die Bluterkrankheit in ihren
Varianten], Lancet, 1:1186f.
BIBLIOGRAPHY 217
Lancet (1903), ‘Report of the Norwegian Beriberi Committee’, Lancet,

1:378.
Lancet (1904), ‘The Etiology of Scurvy’, Lancet, 2:1659f.
Lancet (1911), ‘The Etiology of Beriberi’, Lancet, 2:842.
Lancet (1930), ‘Christiaan Eijkman’, Lancet, 2:1097f.
Landau, Leopold (1874), ‘Zur Aetiologie der Wundkrankheiten’, Archiv
für klinische Chirurgie, 17:527–554.
Latour, Bruno (1984), The Pasteurization of France, trans. Alan Sheridan
and John Law (1988), Cambridge, MA: Harvard University Press.
Laveran, Charles Louis Alphonse (1881), ‘De la nature parasitaire des
accidents de l’impaludisme’, Comptes rendus de l’Académie des sci-
ences, 93:627–630.
Lavoisier, Antoine (1789), Elements of Chemistry, Vol. 45, Great Books
of the Western World, ed. Robert M. Hutchins (1952), Chicago:
Encyclopaedia Britannica.
Lechevalier, Hubert A. and Morris Solotorovsky (1974), Three Centu-
ries of Microbiology, New York: Dover Publications.
Leplat, Emile-Claude and Pierre-François Jaillard (1864), ‘De l’action
des bactéries sur l’économie animale’, Comptes rendus de l’Académie
des sciences, 59:250–252.
Leuckart, Rudolf (1860), ‘Der Geschlechtsreife Zustand der Trichina
spiralis’, Zeitschrift für rationelle Medizin, 8:259–262, 334f.
Loeffler, Friedrich (1887), Vorlesungen über die geschichtliche
Entwickelung der Lehre von den Bacterien, Leipzig: F.C.W. Vogel.
——— (1911), ‘Ueber filtrierbares Virus’, Centralblatt für Bakteriologie,
Parasitenkunde und Infektionskrankheiten, 50: Beiheft:1–12.
——— and P. Frosch (1898), ‘Berichte der Kommission zur Erforschung
der Maul- und Klauenseuche bei dem Institut für Infektionskrankheiten
in Berlin’, Centralblatt für Bakteriologie Parasitenkunde und
Infektionskrankheiten, 23:371–391.
Lomer, M. (1884) ‘Ueber den heutigen Stand der Lehre von den
Infectionsträgern bei Puerperalfieber’, Zeitschrift für Geburtshülfe
und Gynäkologie, 10:366–397.
Loudon, Irvine (2000), The Tragedy of Childbed Fever, Oxford: Ox-
ford University Press.
Lumpe, Eduard (1845), ‘Die Leistungen der neuesten Zeit in der
Gynäkologie’, Zeitschrift der k. k. Gesellschaft der Aerzte zu Wien,
1:341–371.
——— (1850), ‘Zur Theorie der Puerperalfieber’, Zeitschrift der k. k.
Gesellschaft der Aerzte zu Wien, 6:392–398.
Major, Ralph H. (1944), ‘Agostino Bassi and the Parasitic Theory of
Disease’, Bulletin of the History of Medicine, 16:97–107.
218 BIBLIOGRAPHY
Malkin, Harold (1986), ‘Louis Pasteur and “le rage” – 100 Years Ago’,
Perspectives in Biology and Medicine, 30:40–46.
Manninger, Vilmos (1904), Der Entwickelungsgang der Antiseptik und
Aseptik, Breslau: J.U. Kerns.
Manson, Patrick (1901–02), ‘The Etiology of Beriberi’, Transactions of
the Epidemological Society, 21:1–17.
——— (1911–12), ‘[Discussion following] The Etiology of Beriberi [by
Alex Holst]’, Transactions of the Society for Tropical Medicine and
Hygiene, 5:81–90.
——— and C.W. Daniels (1907), ‘Beriberi’, in T.C. Allbutt and H.D.
Rolleston, A System of Medicine, 9 vols, London: Macmillan, vol. 2,
part 2, pp. 615–643.
Marchiafava, Ettore and Angelo Celli (1883), ‘Die Veränderung der
rothen Blutscheiben bei Malaria-Kranken’, Fortschritte der Medizin,
1:573–575.
——— and ——— (1885a), ‘Untersuchungen über die Malariainfection’,
Fortschritte der Medizin, 3:339–354.
——— and ——— (1885b), ‘Weitere Untersuchungen über die
Malariainfection’, Fortschritte der Medizin, 3:787–806.
Marjolin (1833), ‘Anthrax’, in Adelon (1832–46), vol. 3, pp. 193–200.
Mayer, Adolf (1886), ‘Ueber die Mosaikkrankheit des Tabaks’,
Landwirtschaftlichen Versuchs-Stationen, 32:451–467.
McCollum, Elmer Verner (1957), A History of Nutrition, Boston:
Houghton Mifflin.
McFadyean, John (1908), ‘The Ultravisible Viruses’, Journal of Com-
parative Pathology and Therapeutics, 21:58–68, 168–175, 232–242.
Medical Research Committee (1919), Report on the Present State of
Knowledge Concerning Accessory Food Factors (Vitamines), Special
Report No. 38, London: HMSO.
Medical Research Council (1932), Vitamins: a Survey of Present Knowl-
edge, London: HMSO.
Mendel, Emanuel (1884), ‘Ueber Hysterie beim mannlichen Geschlecht’,
Deutsche Medizinische Wochenschrift, 10: 241–244.
Misner, Charles W., Kip S. Thorne and John Archibald Wheeler (1973),
Gravitation, San Francisco: W.H. Freeman and Company.
Möbius, P.J. (1888), ‘Ueber den Begriff der Hysterie’, Centralblatt für
Nervenheilkunde und gerichtliche Psychiatrie, 11:66–71.
——— (1892), ‘Ueber die Eintheilung der Krankheiten’, Centralblatt
für Nervenheilkunde und Psychiatrie, 15:289–301.
Mollaret, H.H. (1983), ‘Contribution à la connaissance des relations
entre Koch et Pasteur’, NTM-Schriftenreihe für Geschichte der
Naturwissenschaften, Technik, und Medizin, 20:57–65.
Monatsschrift für Geburtskunde und Frauenkrankheiten, (1861),
BIBLIOGRAPHY 219
‘Deutsche Naturforscher und Aerzte in Speier: Dritte Sitzung’,

Monatsschrift für Geburtskunde und Frauenkrankheiten, 18:376–382.
Mujeeb-ur-Rahman, M. (ed.) (1977), The Freudian Paradigm: Psychoa-
nalysis and Scientific Thought, Chicago: Nelson-Hall.
Murphy, Frank P. (1946), ‘Ignaz Philipp Semmelweis: an Annotated
Bibliography’, Bulletin of the History of Medicine, 20:653–707.
Nepveu, Gustave (1872), ‘Note sur la présence des bactéries dans le
sang des érysipélateux’, Comptes rendus Société de biologie, 5th
series, 2:164–168.
Nocard, Edmund-Isidore-Etienne and Emile Roux (1898), ‘Le Microbe
de la Péripneumonie’, Annales de l’Institut Pasteur, 12:240–249.
Obermeier, Otto (1873), ‘Vorkommen feinster, eine Eigenbewegung
zeigender Fäden im Blute von Recurrenskranken’, in Heinz Zeiss (ed.)
(1926), Otto Obermeier: Die Entdeckung von fadenförmigen Gebilden
im Blut von Rückfallfieberkranken, Leipzig: Johann Ambrosius.
Oliver, W.S. (1863), ‘Scurvy: Its Cause’, Lancet, 1:61.
Oppenhelm, Herman (1890), ‘Thatsächliches und Hypothetisches über
das Wesen der Hysterie’, Berliner klinische Wochenschrift, 27:553–
556.
Orth, Johannes (1873), ‘Untersuchungen über Puerperalfieber’, Archiv
für pathologische Anatomie, Physiologie, und klinische Medizine
(Virchows Archiv), 58:437–460.
Oxtoby, David W. and Norman H. Nachtrieb (1990), Principles of
Modern Chemistry, Chicago: Saunders College Publishing.
Paget, James (1866), ‘On the Discovery of Trichina’, Lancet, 1:269.
Parascandola, M. (1998), ‘Epidemiology: Second-Rate Science?’, Public
Health Reports, 113:312–320.
Pasteur, Louis (1857a), ‘Mémoire sur la fermentation appelée lactique’,
in Vallery-Radot Pasteur (ed.) (1922–39), Oeuvres de Pasteur, 7 vols,
Paris: Masson, vol. 2, pp. 3–13.
——— (1857b), ‘Mémoire sur la fermentation alcoolique’, in V.R. Pas-
teur (1922–39), vol. 2, pp. 18–22.
——— (1859a), ‘Lettre manuscrite de Pasteur à Pouchet’, in V.R. Pas-
teur (1922–39), vol. 2, pp. 628–630.
——— (1859b), ‘Note remise par Pasteur au Ministre de l’instruction
publique et des cultes’, in V.R. Pasteur (1922–39), vol. 3, pp. 481f.
——— (1860), ‘Suite à une précédente communication relative aux
générations dites spontanées’, in V.R. Pasteur (1922–39), vol. 2, pp.
202–205.
——— (1861a), ‘Animalcules infusoires vivant sans gaz oxygène libre et
déterminant des fermentations’, in V.R. Pasteur (1922–39), vol. 2, pp.
136–138.
220 BIBLIOGRAPHY
——— (1861b), ‘Expériences et vues nouvelles sur la nature des

fermentations’, in V.R. Pasteur (1922–39), vol. 2, pp. 142–147.
——— (1861c), ‘Mémoire sur les corpuscules organisés qui existent
dans l’atmosphère. Examen de la doctrine des générations spontanées’,
in V.R. Pasteur (1922–39), vol. 2, pp. 210–294.
——— (1862), ‘Quelques faits nouveaux au sujet des levures alcooliques’,
——— (1863a), ‘Lettre au Colonel Favé’, in V.R. Pasteur (1922–39),
vol. 7, pp. 8f.
——— (1863b), ‘Examen du rôle attribué au gaz oxygène atmosphérique
dans la destruction des matières animales et végétales après la mort’,
——— (1864), ‘Études sur les vins. Deuxième partie: des altérations
spontanées ou maladies des vins, particulièrement dans la Jura’, in
V.R. Pasteur (1922–39), vol. 3, pp. 396–406.
——— (1865a), ‘Procédé pratique de conservation et d’amélioration
des vins’, in V.R. Pasteur (1922–39), vol. 3, pp. 409–412.
——— (1865b), ‘Observations sur la maladie des vers à soie’, in V.R.
Pasteur (1922–39), vol. 4, pp. 427–431.
——— (1865c), ‘Note au sujet de la communication de MM. Leplat et
Jaillard’, in V.R. Pasteur (1922–39), vol. 6, pp. 161–163.
——— (1866a), Études sur le Vin, in V.R. Pasteur (1922–39), vol. 3,
pp. 111–386.
——— (1866b), ‘Observations verbales présentées après la lecture de la
note de M. Donné’, in V.R. Pasteur (1922–39), vol. 2, pp. 352–355.
——— (1866c), ‘Nouvelles études sur la maladie des vers a soie’, in
V.R. Pasteur (1922–39), vol. 4, pp. 436–448.
——— (1866d), ‘Observations au sujet d’une note de M. Balbiani rela-
tive à la maladie des vers à soie’, in V.R. Pasteur (1922–39), vol. 4,
pp. 471–472.
——— (1867a), ‘Nouvelle note sur la maladie des vers à soie, présentée
à la commission impériale de sériciculture, dans sa séance du 12
janvier 1867’, in V.R. Pasteur (1922–39), vol. 4, pp. 454–468.
——— (1867b), [report of discussion held 24 June 1867], in V.R. Pas-
teur (1922–39), vol. 4, pp. 505–510.
——— (1867c), ‘Sur la nature des corpuscles des vers à soie. Lettre à M.
Dumas’, in V.R. Pasteur (1922–39), vol. 4, pp. 498–499.
——— (1868a), ‘Éducations précoces de graines des races indigènes
provenant de chambrées choisies. Lettre à M. Dumas’, in V.R. Pasteur
(1922–39), vol. 4, pp. 524–528.
——— (1868b), ‘Rapport à S. Exc. M. le Ministre de l’Agriculture sur
la mission confiée à M. Pasteur, en 1868, relativement à la maladie
des vers a soie’, in V.R. Pasteur (1922–39), vol. 4, pp. 547–576.
BIBLIOGRAPHY 221
——— (1868c), ‘Remarques à propos de la note de M. Chauveau sur la

nature du virus-vaccin’, in V.R. Pasteur (1922–39), vol. 6, p. 469.
——— (1869a), ‘Note adressée à l’empereur sur la sériciculture’, in V.R.
Pasteur (1922–39), vol. 7, pp. 18–20.
——— (1869b), ‘Résultats des observations faites sur la maladie des
morts-flats, soit héréditaire, soit accidentelle’, in V.R. Pasteur (1922–
39), vol. 4, pp. 590–594.
——— (1870), Études sur la maladie des vers à soie, in V.R. Pasteur
(1922–39), vol. 4, pp. 1–284.
——— (1873a), ‘Réponse à une note de M. Trécul’, in V.R. Pasteur
(1922–39), vol. 2, pp. 411–415.
——— (1873b), ‘Observations sur la putréfaction et la fermentation’, in
V.R. Pasteur (1922–39), vol. 6, pp. 3–5.
——— (1874a), ‘Discussion sur la fermentation putride’, in V.R. Pas-
teur (1922–39), vol. 6, pp. 13–16.
——— (1874b), ‘Discussions sur la putréfaction’, in V.R. Pasteur (1922–
39), vol. 6, pp. 6–12.
——— (1875a), ‘Discussion sur la fermentation’, in V.R. Pasteur (1922–
39), vol. 6, pp. 37–58.
——— (1875b), ‘Sur les urines ammoniacales’, in V.R. Pasteur (1922–
39), vol. 6, pp. 77–80.
——— (1876a), Études sur la bière, in V.R. Pasteur (1922–39), vol. 5.
——— (1876b), ‘Sur la fermentation de l’urine’, in V.R. Pasteur (1922–
39), vol. 6, pp. 80–84.
——— (1877a), ‘Étude sur la maladie charbonneuse’, in V.R. Pasteur
(1922–39), vol. 6, pp. 164–171.
——— (1877b), ‘Charbon et septicémie’, in V.R. Pasteur (1922–39),
vol. 6, pp. 172–188.
——— (1878a), ‘Sur les découvertes relatives à la maladie charbonneuse’,
——— (1878b), ‘Congrès international séricicole tenu à Paris du 5 au
10 Septembre 1878’, in V.R. Pasteur (1922–39), vol. 4, pp. 691–696.
——— (1878c), ‘Discussion sur l’étiologie du charbon; poules rendues
charbonneuses’, in V.R. Pasteur (1922–39), vol. 6, pp. 210–214.
——— (1878d), ‘La théorie des germes et ses applications à la médecine
et à la chirurgie’, in V.R. Pasteur (1922–39), vol. 6, pp. 112–130.
——— (1878e), ‘Discussion sur la flacherie’, in V.R. Pasteur (1922–39),
vol. 4, pp. 698–725.
——— (1879a), ‘Septicémie puerpérale’, in V.R. Pasteur (1922–39), vol.
6, pp. 131–135.
——— (1879b), ‘Étiologie du charbon’, in V.R. Pasteur (1922–39), vol.
6, pp. 232–238.
222 BIBLIOGRAPHY
——— (1879c), ‘Discussion sur la peste en orient’, in V.R. Pasteur

(1922–39), vol. 6, pp. 493–497.
——— (1879d), ‘Commission dite de la peste’, in V.R. Pasteur (1922–
39), vol. 6, pp. 497–502.
——— (1880a), ‘De l’extension de la théorie des germes a l’étiologie de
quelques maladies communes’, in V.R. Pasteur (1922–39), vol. 6, pp.
147–158.
——— (1880b), ‘Sur l’étiologie du charbon’, in V.R. Pasteur (1922–39),
vol. 6, pp. 254–263.
——— (1880c), ‘Sur la non-récidive de l’affection charbonneuse’, in
V.R. Pasteur (1922–39), vol. 6, pp. 316–322.
——— (1881a), ‘De la possibilité de rendre les moutons réfractaires au
charbon par la methode des inoculations préventives’, in V.R. Pasteur
(1922–39), vol. 6, pp. 339–343.
——— (1881b), ‘De l’atténuation des virus et de leur retour à la viru-
lence’, in V.R. Pasteur (1922–39), vol. 6, pp. 332–338.
——— (1881c), ‘Vaccination in relation to chicken-cholera and splenic
fever’, in V.R. Pasteur (1922–39), vol. 6, pp. 370–378.
——— (1881d), ‘Sur la rage’, in V.R. Pasteur (1922–39), vol. 6, pp.
573f.
——— (1881e), ‘Expériences faites avec la saline d’un enfant mort de la
rage’, in V.R. Pasteur (1922–39), vol. 6, pp. 553–555.
——— (1882), ‘De l’atténuation des virus’ in V.R. Pasteur (1922–39),
vol. 6, pp. 391–411.
——— (1883), ‘La vaccination charbonneuse: réponse à un mémoire de
M. Koch’, in V.R. Pasteur (1922–39), vol. 6, pp. 418–440.
——— (1884), ‘Microbes pathogènes et vaccins’, in V.R. Pasteur (1922–
39), vol. 6, pp. 590–602.
——— (1885), ‘Méthode pour prévenir la rage après morsure’, in V.R.
Pasteur (1922–39), vol. 6, pp. 603–612.
——— (1887), ‘A propos de la vaccination charbonneuse’, in V.R. Pas-
teur (1922–39), vol. 6, pp. 460f.
Pasteur, Vallery-Radot (ed.) (1922–39) Oeuvres de Pasteur, 7 vols,
Paris: Masson et cie.
Pollender, Franz Aloys Antoine (1855), ‘Mikroskopische und
mikrochemische Untersuchungen des Milzbrandblutes sowie über
Wesen und Kur des Milzbrandes’, Vierteljahrschrift für gerichtliche
Medizin und Oeffentliches Sanitätswesen, 8:103–114.
Prichard, J.C. (1845a), ‘Hypochondriasis’, in Dunglison (1845), vol. 2,
pp. 554–562.
——— (1845b), ‘Insanity’, in Dunglison (1845), vol. 3, pp. 26–76.
r (1882), [Obituary of Karl Mayrhofer], Wiener medizinische Blätter,
5:col. 725.
BIBLIOGRAPHY 223
Ralfe, Charles Henry (1877), ‘General Pathology of Scurvy’, Lancet,

1:868–71.
Ravitsch, J. (1872), Zur Lehre von der putriden Infection und deren
Beziehung zum sogenannten Milzbrande, Berlin: August Hirschwald.
Rayer, Pierre-François-Olive (1850), ‘Inoculation du sang de rate’,
Comptes rendus des séances et mémoires de la Société de biologie,
2:141–144.
Recklinghausen, Friedrich Daniel von (1872), ‘Ueber Pilzmetastasen’,
Verhandlungen der Würzburger physikalischemedicinische Gesellschaft,
2:xiif.
Richard, M. (1882), ‘Sur le parasite de la malaria’, Comptes rendus de
Richmond, Phyllis A. (1978), ‘The Germ Theory of Disease’ in Abraham
M. Lilienfeld (ed.), Times, Places and Persons, Baltimore, MD: Johns
Hopkins University Press.
Richter, Hermann Eberhard (1867), ‘Die neuern Kenntnisse von den
krankmachenden Schmarotzerpilzen’, Schmidts Jahrbucher der in-
und ausländische Medizin, 135:81–98.
Ricord, B. [sic] (1835), ‘Correspondance médicale: Lettre de M. Ricord
sur la syphilis, adressée au président de la Société royale académique
de Nantes’, Gazette médicale de Paris, 2nd series 3:540–541.
Ricord, Philippe (1838), Traité pratique des maladies vénériennes, Paris:
Just Rouvier et E. Le Bouvier.
Rindfleisch, Eduard (1867–69), Lehrbuch der pathologischen Gewebelehre
zur Einführung in das Studium der pathologischen Anatomie, Leipzig:
Engelmann.
Rivers, Thomas M. (1937), ‘Viruses and Koch’s Postulates’, Journal of
Bacteriology, 33:1–11.
Rokitansky, Carl Jun. (1874), ‘Untersuchungen der mikroskopischen
Zusammensetzung der Lochien’, Medizinische Jahrbücher, 30:161–
178.
Rosenberg, Charles E. (1979), ‘The Therapeutic Revolution: Essays in
the Social History of America’, in Morris J. Vogel and Charles E.
Rosenberg (eds), The Therapeutic Revolution: Medicine, Meaning,
and Social Change in 19th-Century America, Philadelphia: University
of Pennsylvania Press.
——— (1989), ‘Body and Mind in 19th-Century Medicine: Some Clini-
cal Origins of the Neurosis Construct’, Bulletin of the History of
Medicine, 63:185–197.
Rosenthal, Moritz (1879), ‘Untersuchungen und Beobachtungen über
Hysterie’, Wiener medizinische Presse, 20:569–805 (passim).
Roser, W. (1860), ‘Die specifische Natur der Pyämie’, Archiv der
Heilkunde, 1:39–50.
224 BIBLIOGRAPHY
——— (1867), ‘Zur Verständigung über den Pyämiebegriff’, Archiv für

Heilkunde, 8:15–24.
Röthlin, Otto Mario (1962), Edwin Klebs, Zürich: Juris.
Rupert, J. (1880), ‘Uber Beriberi’, Deutsches Archiv für klinische Medizin,
27:95–110, 499–519.
Scanzoni, Wilhelm Friedrich (1850), ‘[Review of Josef] Skoda[’s], Ueber
die von Dr. Semmelweis entdeckte wahre Ursache der in der Wiener
Gebäranstalt ungewöhnlich häufig vorkommenden Erkrankungen der
Wöchnerinen’, Vierteljahrschirft für das praktische Heilkunde,
Literarische Anzeiger, 26:25–33.
——— (1855), Lehrbuch der Geburtshilfe, 3rd edn, 2 vols, Vienna:
L.W. Seidel.
Schäfer (1884), ‘Ueber Hysterie bei Kindern’, Archiv für Kinderheilkunde,
5:401–428.
Schaumann, H. (1910), ‘Die Aetiology der Beriberi’, Archive für Schiffs-
und Tropenhygiene, 14: Beiheft 8: 325–329.
Scheube, B. (1894), Die Beriberikrankheit, Jena: Gustav Fischer.
Schlich, Thomas (1994), ‘Changing Disease Identities: Cretinism, Poli-
tics and Surgery (1844–1892)’, Medical History, 38:421–443.
——— (1996), ‘Die Konstruktion der notwendigen Krankheitsursache:
Wie die Medizin Krankheit beherrschen will’, in Cornelius Borck
(ed.), Anatomien medizinischen Wissens, Frankfurt am Main: Fischer
Taschenbuch.
Schönlein, Johann Lucas (1832), Allgemeine und specielle Pathologie
und Therapie, 2nd edn, Würzburg: C. Etlinger.
——— (1839), ‘Zur Pathologie der Impetigines’, Archiv für Anatomie,
Physiologie und wissenschaftliche Medicin, 6:82.
Schrödinger, Erwin (1969), What is Life?, Cambridge: Cambridge Uni-
versity Press.
Schüller, Max (1876), ‘Experimentelle Beiträge zum Studium der
septischen Infection’, Deutsche Zeitschrift für Chirurgie, 6:113–190.
Schwalbe, J. (ed.) (1912), Gesammelte Werke von Robert Koch, 2 vols,
Leipzig: Georg Thieme.
Schweninger, Franz (1866), ‘Ueber die Wirkung faulender organischer
Substanzen auf den lebenden thierischen Organismus’, Aertzliches
Intelligenz-blatt 13:590–672 (passim).
Seeligmuller, Ludwig (1881), ‘Ueber Chorea magna und ihre Behandlung’,
Deutsche Medizinische Wochenschrift, 7:584.
Semmelweis, Iganz (1861a), Die Aetiologie, der Begriff, und die Prophy-
laxis des Kindbettfiebers, Pest: C.A. Hartleben.
——— (1861b), The Etiology, Concept, and Prophylaxis of Childbed
Fever, trans. K. Codell Carter (1983), Madison, WI: University of
Wisconsin Press.
BIBLIOGRAPHY 225
Signol (1863), ‘Présence des bactéries dans le sang’, Comptes rendus de

Sontag, Susan (1979), Illness as Metaphor, New York: Vintage Books.
Späth, Josef (1864), ‘Statistische und historische Rückblicke auf die
Vorkommnisse des Wiener Gebärhauses während der letzten dreissig
Jahre mit besonderer Berücksichtigung der Puerperal-Erkrankungen’,
Medizinische Jahrbücher, 20:145–164.
Spencer, M.H. (1897), ‘Notes on beriberi as observed at the Seamen’s
Hospital, Greenwich’, Lancet, 1:30–32.
Stamm, August Theodor (1865), ‘Grösse und Einrichtung von
Gebäranstalten’, Gesellschaft der Deutsche Naturforscher und Aerzte
in Giessen, Biessen: Keller.
Stehbens, William E. (1992), ‘Causality in Medical Science with Par-
ticular Reference to Heart Disease and Atherosclerosis’, Perspectives
in Biology and Medicine, 36:97–119.
Steudener, Friedrich (1872), ‘Ueber pflanzliche Organismen als
Krankheitserreger’, Volkmann’s klinische Vorträge, Innere Medizin,
1:283–308.
Stewart, C.P. (1953), ‘Scurvy in the 19th Century and after’, in C.P.
Stewart and Douglas Guthrie (eds), Lind’s Treatise on Scurvy, Edin-
burgh: Edinburgh University Press.
Strachey, James (ed.) (1955–74), The Standard Edition of the Complete
Psychological Works of Sigmund Freud, 24 vols, London: Hogarth
Press.
Strümpel, Adolf von (1884a), ‘Ueber die Ursachen der Erkrankungen
des Nervensystems’, Deutsche Archiv für klinische Medizin, 35:1–17.
——— (1884b), Krankheiten des Nervensystems, Leipzig: F.C.W. Vogel.
——— (1885), Krankheiten des Nervensystems, 2nd edn, 2 vols, Leip-
zig: F.C.W. Vogel.
——— (1893), ‘Ueber die Entstehung und die Heilung von Krankheiten
durch Vorstellungen’, Berliner klinische Wochenschrift, 30:22–25.
Susser, Mervyn (1973), Causal Thinking in the Health Sciences, Ox-
ford: Oxford University Press.
Symonds, J.A. (1845), ‘Tetanus’, in Dunglison (1845), vol. 4, pp. 364–
376.
Takaki, Kamehiro (1885), ‘On the Cause and Prevention of Kakke’,
Trans. Sei-i-kwai, 4:29–37.
——— (1906), ‘The Preservation of Health Amongst the Personnel of
the Japanese Navy and Army’, Lancet, 1:1369–1374, 1451–1455,
1520–1523.
Taylor, F. Kräupl (1979), The Concepts of Illness, Disease and Morbus,
Cambridge: Cambridge University Press.
226 BIBLIOGRAPHY
Théodoridès, Jean (1966), ‘Casimir Davaine (1812–1882): a Precursor

of Pasteur’, Medical History, 10:155–165.
Tiegel, E. (1871), ‘Die Ursache des Milzbrandes’, Correspondenzblatt
für Schweizer Aerzte, 1:275–280.
Tigri (1863), ‘Sur la présence d’infusoires du genre Bacterium dans la
sang humain’, Comptes rendus de l’Académie des sciences, 57:633.
Trousseau, A. (1835), ‘Croup’, in Nicholas Philibert Adelon (ed.),
Dictionnaire de médecine, 2nd edn, Paris: Bechet, vol. 9, pp. 334–
401.
Tuczek, F. (1886), ‘Zur Lehre von der Hysterie der Kinder’, Berliner
klinische Wochenschrift, 31:511–515, 534–537.
Turner, Victor (1967), The Forest of Symbols, Ithaca, NY: Cornell
University Press.
Tweedy, Alexander (1845), ‘Fever (Continued)’, in Dunglison (1845),
vol. 2, pp. 153–201.
Twort, F.W. (1915), ‘An Investigation on the Nature of Ultra-Micro-
scopic Viruses’, Lancet, 2:1241–1243.
Van der Berg, C.L. (1889), ‘Reviews and Notices of Books: C.A.
Pekelharing and A. Winkler, Onderzoek naar den aard en de Orzaak
der Beriberi’, Lancet, 1:892f, 941f.
Van der Steen, Wim J. and Harmke Kamminga (1991), ‘Laws and
Natural History in Biology’, British Journal of the Philosophy of
Science, 42:445–467.
Veit, A.C. Gustav (1865), ‘Ueber die in der geburtshilflichen Klinik in
Bonn im Sommer 1864 und 1864–65 aufgetretenen puerperalen
Erkrankungen’, Monatsschrift für Geburtskunde und Frauenkrank-
heiten, 26:127–155, 161–208.
——— (1867), Krankheiten der weiblichen Geschlechtsorgane, 2nd edn,
Erlangen: Ferdinand Enke.
Villemin, Jean Antoine (1868), Études sur la tuberculose, Paris: J.B.
Baillière.
Virchow, Rudolf (1880), ‘Krankheitswesen und Krankheitsursachen’,
Medizine (Virchows Archiv), 79:1–19, 185–228.
Waldeyer, Wilhelm (1871), ‘Ueber die pathologische Bedeutung der
Bacterien, Vibrionen, etc.’, Jahresbericht der schlesische Gesellschaft
für vaterländische Kultur, 49:205–208.
——— (1872), ‘Ueber das Vorkommen von Bacterien bei der
diphtheritischen Form des Puerperalfiebers’, Archiv für Gynaekologie,
3:293–296.
Warner, John Harley (1986), The Therapeutic Perspective, Cambridge,
MA: Harvard University Press.
BIBLIOGRAPHY 227
Watson, Thomas (1858), Lectures on the Principles and Practice of

Physic, Philadelphia: Blanchard and Lea.
Weiss, J. (1884), ‘Die infantile Hysterie’, Archiv für Kinderheilkunde,
5:451–461.
Wertheimer, M. (1888), Von dem Verhalten der Lochialsecretion zur
Pathogenese des Kindbettfiebers, Freiburg: H.M. Poppen & Sohn.
Wilkinson, Lise (1976), ‘The Development of the Virus Concept as
Reflected in Corpora of Studies on Individual Pathogens; 3. Lessons
of the Plant Viruses – Tobacco Mosaic Virus’, Medical History, 20:111–
134.
Williams, Robert R. (1961), Toward the Conquest of Beriberi, Cam-
bridge, MA: Harvard University Press.
Wilson, Adrian (2000), ‘On the History of Disease-Concepts: The Case
of Pleurisy’, History of Science, 38:271–319.
Winckel, Franz Karl Ludwig Wilhelm (1866), Die Pathologie und
Therapie des Wochenbettes, Berlin: Hirschwald.
Worboys, Michael (2000), Spreading Germs: Disease Theories and Medi-
cal Practice in Britain, 1865–1900, Cambridge: Cambridge University
Press.
Wright, Hamilton (1902), An Inquiry into the Etiology and Pathology
of Beriberi, Singapore: Kelly and Walsh.
Wulff, Henrik R. (1984), ‘The Causal Basis of the Current Disease
Classification’, in Lennart Nordenfelt and Ingemar B. Lindahl, Health,
Disease, and Causal Explanations in Medicine, Dordrecht: D. Reidel,
pp. 169–177.
Yerushalmy, J. and Carroll E. Palmer (1959), ‘On the Methodology of
Investigations of Etiologic Factors in Chronic Diseases’, Journal of
Chronic Diseases, 10:27–40.
Zakon, S.J. and T. Benedek (1944), ‘David Gruby and the Centenary of
Medical Mycology 1841–1941’, Bulletin of the History of Medicine,
16:155–168.
Zenker, F.A. (1860), ‘Ueber die Trichinen-Krankheit des Menschen’,
——— (1865), ‘Beiträge zur Lehre von der Trichinenkrankheit’,
Deutsches Archiv für klinische Medizin, 1:90–124.
Zuber (1882) [Review of Koch’s 1881 anthrax paper], Revue d’hygiène,
2:104f.
Index
acari, 26–7, 37, 50, 104, 106, 126, research programme, 143–4, 147–8,
138, 143 159–60, 162, 180, 194–5
Achorion schönleinii, 32 bacteridium, 92, 95, 98, 102, 110,
Ackerknecht, Erwin, 22 115, 117–19, 122, 126, 140
AIDS, 1, 3, 203 Davaine on 80–88
Alderidge, John, 182 bacteriology in the 1870s, 92–5
ammoniacal urine, 112, 114 Balsamo-Crivelli, Joseph, 66
Anderson, William, 187 Bamberger, Heinrich von, 149
Andersson, Ola, 161 Bar, Paul, 59
Andral, Gabriel, 18, 36 Bardsley, James L., 10, 11, 16, 17,
anomalies, in relation to theories 3–4, 202–3
6, 108, 126, 143, 162, 176, 177 Barlow, Edward, 15, 20
anthrax, 55, 90, 93, 106, 108, 110, Barnes, David S., 61
137–41, 176; Davaine on, 60, Bassi, Agostino, 25, 28–32, 36–7, 60,
76–89, 95, 98; Koch and Pasteur 72–3, 79, 96, 108, 110, 143, 168,
on, 115–22, 124, 130–32, 135; see 196
also malignant pustules Beaude, Jean Pierre, 76
Aristotle, 200, 202 Beauvaria bassiana, 29, 36, 60, 66,
Arneth, Franz Hektor, 58 73–4, 79, 106, 109, 126, 143
attenuation, of bacterial virulence, Becker, Stanley L., 179
119, 124 beer, spoilage of, 111–14
Audouin, Jean Victor, 31–2, 110 Beijerinck, Marinus Willem, 169–73,
axioms, basic theoretical assumptions 180
as, 5 Benedek, T., 33–4, 37, 60
beriberi, 143, 179–81, 184–95
Bach, Johann Sebastian, 198 Bert, Paul, 117, 121, 127
Bacillus anthracis, 115–16, 130 Berzelius, Jons Jacob, 63
Bacillus flourescens, 172 Bichat, Marie François Xavier, 18
Bacillus malariae, 163 Biett, L., 27
bacteria, 6, 8, 58, 61, 77–8, 92–5, Billet, Léon, 58
102, 106, 109, 145, 168–9, 171, biological sciences, laws and theories
175, 187–8; nomenclature for, in 2, 6, 109; see also science,
92–3; see also Bacillus anthracis; theories of disease, theory of germs
Bacillus flourescens; Bacillus Birch-Hirschfeld, Felix Victor, 55,
malariae; bacteridia; ferments; 58–9, 61, 77, 93, 95, 99, 142
germs; microbes; microorganisms; Blane, Gilbert, 181, 182
vibrions; virus blood letting, 61
bacterial hypothesis, 62, 74, 76–89 Bloomfield, Arthur L., 134, 165, 166
(esp. 87–8), 102–4, 106, 118, 122, Boehr, Max, 51, 57–8
126, 144–5, 161, 173 Bollinger, Otto, 76, 77, 88, 95
bacterial theory of disease, 6–8, Böttger, Herbert, 55
55–9, 69, 76–88, 90–108, 111, Boyle’s Law, 5
126–7, 129, 155–6, 159–60, 173, Braddon, Leonard, 185, 189, 190,
176, 190–92; general acceptance of, 191, 192
122–7, 134; in relation to the Brauell, August, 77–8, 80, 82–3, 85
deficiency theory, 179–81, 184, Braun, Carl, 46–7, 109
190–92; in relation to the etiological Breisky, August, 49
230 INDEX
Breuer, Josef, 152–3, 155, 157, 83–4, 87, 96, 98, 103, 110, 113–15,
158–9 117, 130, 134–5, 138–41, 145; of
Brock, Thomas D., 115, 133 childbed fever 10, 40–43, 45–51,
Brown, Joseph, 13 109, 126, 143, 175; of death 16–17,
Bruce-Chwatt, Leonard J., 165 19–20, 31; of diabetes 10, 11,
Budd, George, 181–4 13–14, 17; of individual events 12,
Butler, Samuel, 1 17, 19–20, 22, 23, 146; of pneu-
butyric acid and ferment, 64, 78, 93 monia 11; predisposing, 12, 14–15,
Bynum, William F., 197 17, 20, 40, 43, 47, 155; proximate,
12–13, 16, 106–8; remote, 12–17,
Cagniard-Latour, Charles, 63 20, 25, 40, 90, 106–8; specific, 9,
Campbell, William C., 37 108, 143–4, 156, 158, 171, 176; see
cancer research, 4, 202 also causes, natural, universal, and
Carnap, Rudolf, 200 necessary
Carpenter, Kenneth J., 194–5 Celli, A., 165–6, 178
Carter, K. Codell, 22, 23, 44, 55, 61, characterizations of diseases, 7–8,
98, 100, 128, 139, 151, 154, 159, 18–20, 36–7, 38–61, 80–81, 103,
179, 182, 183, 184, 195, 196, 197 106–8, 109–10, 129, 138–9, 145,
Castiglioni, Arturo, 22, 147 151–2, 155–6, 158, 160, 164–5,
causal arguments, 31, 33–4, 36, 193, 200
73–4, 81–3, 87–8, 110, 122, 126, Charcot, Jean Martin, 149–61
138–9, 163–4, 172, 190, 198; Charles’ Law, 5
Davaine’s, 79, 82, 83–5; Klebs’, childbed or puerperal fever, 38–61,
95–8, 101, 163; Koch’s, 115–17, 93, 109, 119, 176, 196
119–22, 130–31, 134–5, 141; etiology of 10, 40–43, 45–51, 74,
Pasteur’s, 64, 70–73, 111–15, 109, 126, 143, 175; Semmelweis on,
117–18, 119–22 44–54; opinions about, in the
causal criteria, 8, 70, 73–4, 102–3, 1840s, 39–44
110–11, 122, 140, 162–3, 167, 174, chicken cholera, 119, 124
177, 178, 181; Klebs on, 92, 96–9, chickens: anthrax in, 118, 119, 140;
100; Koch’s use of 111, 131–2, 142, polyneuritis in, 188–90, 192
144; see also Koch’s Postulates Chimbuki, 21; medicine 22, 23, 39,
causal explanations, 30, 53, 54, 57, 197
60, 65, 105, 107, 119, 142, 144, chlorine washings, 44–7, 53
152, 196, 199, 200; in Davaine cholera, 55, 176, 198; discovery of
85–6; in Freud 158–60; in Koch bacillus, 127, 138, 139
132, 137, 198; in Semmelweis Cholera morbus, 31
48–51, 57 Chomel, A.F., 10–11, 14, 17, 20
causal webs, 4 chorea, 149, 150
causes, 88, 198, 203; exciting, 12, circularity of arguments for theories,
14–15, 17, 20, 40, 43, 108, 151, 126
155; in nineteenth-century medicine citations in scientific literature, 4–5,
10–23, 28, 104, 106–8, 151, 167–8, 38
176, 180; moral transgressions as, cofactors in disease, 140
20–22; natural, universal, and Cohn, Ferdinand Julius, 93–4, 109,
necessary, 1–4, 7, 24–37, 38, 51, 133
53–4, 59–60, 103, 106–8, 110, 112, Cohn, Martin, 148–50
118, 129, 142–3, 176, 180, 193–4, Cohnheim, Julius, 104, 133
196–9, 201 (see also causes, spe- Colin, Léon, 163–4
cific); necessary or sufficient 11–15, Collard, Patrick, 132
20–22, 23, 24–5, 48, 66, 70–73, 81, Columbus, egg of, 45, 54
INDEX 231
Conolly, John, 13, 14 Dunglison, Robley, 21, 23

Conrad, Peter, 2 Durham, Herbert, 187
consumption or phthisis, 18–19, 106,
132–3; see also tuberculosis earthworms, in the spread of
Copernicus, 160 anthrax, 118–19
Copi, Irving M., 53 egg of Columbus, 45, 54
Coze, Leon, 56, 58 Ehrenberg, Christian Gottfried, 92–4
Crede, Carl S.F., 49 Eijkmann, Christian, 188–92, 196
cretinism, 180 Ellenberger, Henri F., 121, 147–8
crucial experiments, 9 Elliotson, John, 14, 181
Cuboni, Giuseppe, 163 enzymes as possible causes of disease,
Cumin, William, 21 168, 180
Cunningham, Andrew, 9 epidemic childbed fever, 40, 42, 43,
48, 51, 52
Danyau, M., 47 epidemiologists, contemporary, under
Darwin, Charles, 54, 160 threat, 200
Davaine, Casimir, 60, 74, 76–89, 92, epilepsy, 149
95–8, 102, 104–5, 108, 110, etiological disease characterizations,
115–18, 121, 126, 127, 138, 143, 7–8, 38–61, 106–8, 152, 156, 158,
163, 196, 198 160, 164–5; see also characteriza-
death, causes of, 16–17, 19–20, 31 tions of diseases
Decker, Hannah S., 161 etiological standpoint or research
deficiency theory of disease, 6, programme, 1–4, 7–8, 9, 22–3, 38,
179–95, 196 73–4, 106–7, 122, 129–46, 147,
definitions of diseases; see characteri- 161, 163, 171, 176–7, 180, 193–4,
zations of diseases 198, 199–200
D’Espine, H.-A., 58 etiology, 24–5, 55, 100, 147, 151,
diabetes, 10, 11, 13–14, 17 155; of childbed fever 10, 40–43,
diphtheria, 92–93, 138, 176 45–54, 109, 126, 143, 175
disease characterizations; see charac- Evans, Alfred S., 9, 141, 178, 196,
terizations of diseases 199
diseases of silkworms, 25, 28–32, 60, Evans-Pritchard, E.E., 22
66, 71–4, 75, 76, 79, 109, 111, Everkin, D., 47
114–15, 117, 126–7, 143 evolution, 2, 54
Dissemination Hypothesis, 62, 65–9, exciting or occasioning causes, 12,
87, 101–3, 112, 144; see also 14–15, 17, 20, 40, 43, 108, 151, 155
spontaneous generation explanations, based on causes; see
Distinguishability Hypothesis, 62, causal explanations
64–5, 70, 87–8, 102–3, 105–6, 116,
131, 137, 144–5, 155, 175, 177, 197 Farr, William, 16–17
distinguishability of possible disease favus, 25, 33, 37, 60; see also
agents, 175–7, 181; of viruses, 174–7 mycoses
Dolman, Claude E., 115 Feltz, Victor-Timothee, 56, 58
Donné, Alfred, 25 Ferber, Rudolf, H. 58
Douglas, Mary, 22 fermentation, 63–6, 75, 76, 82, 112
Down syndrome, 140 ferments 63–5, 70, 72, 82, 86, 93,
Dracobly, Alex, 37 112, 114, 168–9; see also Bacillus
Dubini, Angelo, 25 anthracis, Bacillus flourescens,
Dubois, Paul-Antoine, 39–41, 47, Bacillus malariae, bacteria,
50–51, 107 bacteridia, germs, microbes, micro-
Duesberg, Peter, 203 organisms, vibrions, virus
232 INDEX
filtration experiments, 84, 95–6, 98, Grijns, Gerrit, 191–2

114, 117, 168–9, 172–3 Gruby, David, 25, 32–4, 36, 37, 60,
Fischel, Wilhelm, 59, 61 72–3, 79, 196
Fitzgerald, Thomas J., 105 Grundversuche, 100–109, 110, 117,
Flacherie, 71–2, 75, 114–15, 127; see 162, 173
also diseases of silkworms Györy, Tiberius von, 45, 46, 48
Fletcher, William, 190–92
Flew, Anthony, 200 Hall, Marshall, 123, 181
Fliess, Wilhelm, 153, 157, 159 Hamlin, Christopher, 14, 108
flying insects, as anthrax vectors, 85–6 Hanson, Norwood Russell, 6, 88
foot-and-mouth disease, 170–72, Harden, Victoria A., 178, 196
174–6 Haussman, David, 57, 58
Forbes, John, 19 Hebra, Ferdinand, 44
Foster, George M., 22 Hegel, Georg Wilhelm Friedrich, 200
Foster, W.D., 34–5 Henle, Jacob, 24–5, 27, 32, 36, 37,
Foucault, Michel, 22, 23, 53, 196 59–60, 66, 90, 91, 99, 108, 109,
Fracastoro, 38, 60, 178 138, 140–41, 171
fractional cultivation, 96, 98, 142 Henoch, Eduard Heinrich, 150
Frankfort, H.A., 22 Herbst, G., 34
Fraser, Henry, 190, 192 Herz, Maximilian, 148–50
Frege, Gottlob, 53 heuristic, of research programme, 3,
French Disease, 197 7–8, 62, 65, 144–5, 176, 180, 194,
Freud, Sigmund, 147–8, 153–61, 196–7
175, 177, 180–81, 196 Heymann, Bruno, 142
Freudian psychology, 2, 196 Hirsch, August, 90, 184–7
Frolich, Theodor, 183–4, 192 history and philosophy, viii, 200,
Frosch, Paul, 170–72, 174, 196 202
functional disorders, 148 Holmes, Oliver Wendell, 41–2,
fungus, 109, 129, 168, 180; see also 50–51, 53, 61, 107
mycoses Holst, Alex, 183–4, 191–2
Funk, Casimir 183, 192, 196 Hueber, R.J., 174, 177
Hughes, Sally Smith, 169, 174–5
Garrison, Fielding H., 91 Hume, David, 198, 200, 202
Garrod, Alfred B., 182 hydrophobia or rabies, 18, 31, 36,
Geison, Gerald L., 63, 66–9, 71, 75, 121–6, 143, 162
111, 115, 120, 127, 128 hypotheses, 102; see also bacterial
Gelfand, Toby, 161 hypothesis, dissemination hypoth-
geometry, 5 esis, Distinguishability hypothesis,
Gerhardt, C., 165–6, 178 Ideational hypothesis
germ theories of disease, 6–7, 8, 195, hysteria, 133, 143, 148–61, 175–6,
196; see also bacterial theory of 181
disease
germs, 6–7, 9, 30, 31, 60, 62, 108; ideational: hypothesis, 161, 177;
see also bacteria, bacteridia, theory of disease, 147–61, 181
ferments, microbes, micro- Ihde, Aaron J., 179
organisms, vibrions, virus induction, of universal causes, 198
Ghesquier, Danièle, 36, 37, 196 influenza bacillus, 172
glass, not a solid, 53, 109 inoculation experiments, 77, 80–85,
Gras, Albin, 27 96–99, 114–18, 120–122, 124–5,
gravitational attraction, as a univer- 130–31, 133–42, 165–6, 168–71,
sal cause, 25 174, 176
INDEX 233
inoculations, to promote immunity, Lechevalier, Hubert A., 75, 134, 141

119–21, 124–7, 135 legionnaires’ disease, 1
itch, 26–7 Leplat, Emile-Claude, 81–2, 87–8,
Ivanovski, Dimitri Iosifovitch, 169, 110–11, 127
171–2 leprosy, 103, 139
Leuckart, Rudolf, 35, 37
Jaggard, W.W., 59 leukocytes, 86, 164
Jaillard, Pierre-François, 81–2, 87–8, Liebig, Justus von, 63, 64
110–11, 127 Lind, James, 181, 184
Janet, Pierre, 158 Lister, Joseph, 63, 91, 121
Jeannel, Maurice, 55 Loeffler, Johannes Friedrich, 59, 91,
Joest, Ernst, 167 170–75, 196
Jones, Ernst, 156, 161 Lomer, M., 59
Loudon, Irvine, 51–4, 61
Kamminga, Harmke, 5, 7, 20, 74, Lugol, Jean, 26
102, 109 Lumpe, Eduard, 42–4, 45–6, 50–51,
Kant, Immanuel, viii, 12, 198 54, 61, 107, 175
kinetic theory, 5–6, 9, 69, 100
King, Lester S., 10, 17 Magendie, François, 181
Klebs, Edwin, 7, 69, 90–92, 95–101, Major, Ralph H., 31
110, 121, 122, 137–8, 141–2, 162, malaria, 98, 163–7, 187
163, 165, 172, 173, 196; and Koch Malcolmson, John G., 186
91, 95–8, 104, 137, 141–2; on malignant pustules, 76, 80–81, 86,
causal criteria 95–9 106; see also anthrax
Koch, Robert, 1, 4, 58–9, 105, Malkin, Harold, 125
129–46, 166–8, 171, 174, 180, 188, Manninger, Vilmos, 55
196, 198, 200, 202; and Klebs, 91, Manson, Patrick, 185, 187–8, 190
95–8, 104, 137, 141–2; and Pasteur, Marchiafava, Ettore, 163, 165–6, 178
3, 110–11, 115–22, 126–7, 135, Martin, Eduard, 56, 57
138, 142, 147; on tuberculosis, 52, maternity clinics, 40, 43, 44, 54
127, 132–7, 141–2, 198 Mayer, Adolf Eduard, 167–9, 171–2,
Koch’s postulates, 8, 98, 110–11, 178
115, 120, 122, 129–45, 174, 178, Mayrhofer, Carl, 55–60, 108, 143
196–9, 203; see also causal criteria McCollum, Emer Verner, 179, 194
Koch, Wilhelm, 183 McFadyean, John, 171
Kocher, Theodor, 180 Meister, Joseph, 125, 128
Kunitz, Stephen J., 4, 25, 196 Mendel, Emanuel, 148–9
Kützing, Traugott, 63 Metchnikoff, Elie, 141
methodological rules, 62, 162; see
lactic acid, 64 also research programmes
Lagasquie, A., 12–13, 15 Meynert, Theodor 149
Lakatos, Imre, viii, 2–5, 7, 9, 23, 62, miasms, 40, 43, 46, 48, 51, 56, 57,
65, 69, 104, 145–6, 162–3, 176, 196 61, 66, 143, 175–6, 187
Landau, Leopold, 99 microbes, 147; see also bacteria,
Latour, Bruno, 68, 75, 127–8, 197–8 bacteridia, ferments, germs micro-
Laveran, Charles Louis Alphonse, organisms, vibrions, virus
163–7 micrococci, 92–4, 96, 165, 187
Lavoisier, Antoine Laurent, 200 microorganisms, 57, 60, 62–75, 90,
laws of science, 5–6, 25, 68–9, 74, 96, 111, 142, 183; see also bacteria,
102, 109 bacteridia, ferments, germs,
Lebert, Hermann, 47 microbes, vibrions, virus
234 INDEX
microsporon septicum, 95, 172 165, 198; and Koch, 3, 110–11,

Mill, John Stuart, 200, 202 115–22, 126–7, 135, 138, 142, 147;
Misner, Charles W., 5–6 on rabies, 124–5
Möbius, P.J., 152–3, 155–6, 159, pathological anatomy, 13, 18–20, 22,
161, 175, 177, 180 39, 90–91, 97, 107, 149, 152, 196
Mollaret, H.H., 115 pebrine, 71–2, 75; see also diseases of
moral transgressions as causes of silkworms
disease, 20–22 Pekelharing, C.A., 187–8
Mujeeb-ur-Rahman, M., 160–61 pellagra, 192
multicausality, 4 Pfeufer, Carl, 24
Murphy, Frank P., 55 phenylketonuria, 109
muscardine, 25, 28–32, 60, 66, 73–4, philosophy and history, viii, 200,
79, 109, 126, 143; see also diseases 202
of silkworms photographing microorganisms, 132
mycobacterium leprae, 103 phthisis or consumption, 18–19, 106,
mycoderma aceti, 70 132; see also tuberculosis
mycoses, 25, 32–4, 36, 60, 79 plague, 118
pneumonia, 11, 20
Nachtrieb, Norman H., 69, 100 Pollender, Franz Aloys Antoin, 77,
Naegeli, Carl, 109 80, 117
necessary or sufficient causes; see polyneuritis, in chickens, 188–90,
causes: necessary or sufficient 192
Nepveu, Gustave 94, 187 Popper, Carl, 202
neurosis, in relation to hysteria, 149 postmodernism see French disease
Ndembu of Zambia, 217 predisposing causes, 12, 14–15, 17,
Nocard, Edmund-Isidore-Etienne, 20, 40, 43, 47, 155
172–3, 175 presentist interpretation of history,
Nocht, B., 192 8, 9
Nothnagel, Hermann, 149 Prichard, J.C., 21
nutritional deficiency, theory of problem shifts, 2–3, 176
disease, 6, 179–95, 196 programme, etiological research see
etiological standpoint or research
Obermeier, Otto, 94–5, 108 programme
occasioning or exciting causes, 12, programmes, research, 2–5, 7–8, 62,
14–15, 17, 20, 40, 43, 108, 151, 104, 106, 142, 162
155 progress, scientific, 3, 99, 200–201;
Oliver, W.S., 182 see also Whiggish interpretation of
Oppenheim, Herman, 150 history
Orth, Johannes, 58, 59 prophylaxis, 21, 30, 50, 54, 107, 201
Oxtoby, David W., 69, 100 protozoa, 162, 166–7, 178
protozoal, theory of disease, 162–7,
Paget, James, 25, 34, 37 173
Palmer, Carroll E., 199 proximate causes, 12–13, 16, 106–8
Paracelsus, 38 puerperal or childbed fever, 38–61,
Parascandola, M., 198, 200 93, 109, 119, 176, 196
parasitic organisms, 24–36, 71, 86, etiology of, 10, 40–43, 45–51, 74,
90, 129, 131, 134, 138, 143, 163–4, 109, 126, 143, 175
179, 196 Semmelweis on, 44–54
Pasteur, Louis, 58, 59, 60, 62–75, 76, 1840’s opinions about 39–44
78, 79, 82–3, 86, 87, 93, 96, 98, putrefaction, 63, 80, 86, 95, 112
101–3, 105, 110–28, 138, 143, 147, pyemia, 48, 56
INDEX 235
rabies or hydrophobia, 18, 31, 36, schizomycetes, 97, 99, 101, 109
121–6, 143, 162 Schlich, Thomas, 22, 23, 180, 196
Ralfe, Charles Henry, 182–3 Schmidt, Joseph Hermann, 46
Rapaport, David, 161 Schneider, Joseph W., 2
rational medicine, Henle’s program Schönlein, Johann Lucas, 18–19, 25,
for, 24–5, 59, 140 32, 91
Ravitsch, J., 55 Schrödinger, Erwin, 53
Rayer, Pierre-François-Olive, 76–7 Schüller, Max, 99
Recklinghausen, Friedrich Daniel Schwann, Theodor, 63
von, 93 Schweninger, Franz, 57
relapsing fever, 55, 94–5, 108 science, 1, 2–6, 9, 22, 23, 24–5, 31,
relativity, theory of, 5–6 68–9, 99–100, 104, 108, 109, 121,
Remak, Frederick, 91 200–201; see also research pro-
remote causes, 12–17, 20, 25, 40, 90, grammes
106–8 scientific: laws, 5–6, 25, 68–9, 74,
Renucci, Simon-François, 25–7, 36, 102, 109; see also heuristic of
196 research programme, hypotheses;
research programme, etiological see method, 68–9; progress, 3, 99,
etiological standpoint or research 200–201; see also Whiggish inter-
programme pretation of history; theories, 2–3,
research programmes, 2–5, 7–8, 62, 5–7, 54, 69, 88, 100, 102, 105–6,
104, 106, 142, 162; see also science 109, 162, 197–9; see also theories
resorption, of decaying organic of disease; medicine, 104, 147, 196,
matter, 48, 57 200–202; see also etiological
Richard, M., 164–6 standpoint or research programme,
Richmond, Phyllis A., 141 rational medicine
Richter, Hermann Eberhard, 57 scrofula, 106, 132–3; see also
ricidivousness of disease, 119 tuberculosis
rickets, 181, 192 scurvy, 181–4, 192–5
rickettsial diseases, 162, 178 Seeligmuller, Ludwig, 150
Ricord, Philippe, 25, 37 Semmelweis, Iganz, 38–9, 44–61,
Rindfleisch, Eduard, 92 72–3, 79, 90, 96, 107, 109, 138,
risk factors, 4, 140, 201–2 143, 145, 196; influence of, in
Rivers, Thomas M., 174, 199 etiological research programme, 39,
Rokitansky, Karl, 46 54–60
Rokitansky, Karl Jr., 58 septicemia, 116, 118
Rosenberg, Charles, 23 Sharkow, David, 161
Rosenthal, Moriz, 150 silkworms, diseases of, 25, 28–32,
Roser, W., 56–7 60, 66, 71–4, 75, 76, 79, 109, 111,
Röthlin, Otto Mario, 90 114–15, 117, 126–7, 143
Roux, Pierre Paul Emile, 165, 172–3, Simpson, James Young, 53
175 smallpox, 143, 162, 172
Rupert, J., 187 solid culture, media introduced, 121,
Russell, Bertrand, 200 132
Solotorovsky, Morris, 75, 134, 141
Salomonsen, Carl Julius, 133 Sontag, Susan, 133
scabies, 25–7, 37, 60, 143 Späth, Joseph, 55, 57
Scanzoni, Wilhelm Friedrich, 10, 17, specific causes, 9, 108, 143–4, 156–8,
20, 46, 59 171, 176; see also causes, natural,
Schaumann, H., 185 universal, and necessary
Scheube, B., 185, 187 Spencer, M.H., 188
236 INDEX
Spinoza, 200 theory of germs, 63–4, 66, 69,

spirochete, 94–5, 105 125–6; see also germs, germ theory
spoilage of beer, 111–14 of disease
spoilage of wine, 60, 70–71, 72, 111, therapy, 21–3, 50, 54, 107, 200
113 thermodynamics, laws of, 74
spontaneous: cases of disease, 29, Thorne, Kip S., 5–6
31–2, 48, 73, 85; generation, 27, thyroid failure and cretinism, 180
31–2, 65–9, 75, 76, 86–7, 101, Tiegel, E., 90, 95, 98
112–13; see also dissemination tobacco mosaic disease, 167–73, 176
hypothesis Tommasi-Crudeli, Corrado, 163,
spores, 115, 117–18, 121 165
staining of microorganisms, 119, trichinae, 34–6, 37, 60, 104, 126,
132–4 138, 180
Stamm, August Theodor, 143 trichinosis, 25, 34–6, 60
Stanton, Thomas, 190, 192 Trichomonas vaginalis, 25, 37
Stazevich, T., 183 Trousseau, A., 15
Stehbens, William E., 200–201, 203 tubercle bacillus, 38, 52–3, 61, 107,
Steudner, Friedrich, 55, 87, 99, 116 120, 127, 133–5, 138
Stewart, C.P., 179, 182 tuberculosis, 18, 38, 52–3, 55, 61,
stigmata, in Charcot’s characteriza- 71, 138, 159, 176; Koch’s work on,
tion of hysteria, 149–50, 154 52, 127, 132–7, 141–2, 198
Strachey, James, 154 Tuczek, F., 150–51
Strümpell, Adolf von, 129, 147–8, Turner, Victor, 21
150, 152–3, 155–6, 159, 161, 175, Tweedie, Alexander, 181
180, 193, 197 Twort, F. W., 180
sufficient or necessary causes see typhoid fever, 99
causes, necessary or sufficient
surgical fever, 55–6 vaccine, 111, 119
Susser, Mervyn, 4, 115 Van der Berg, C.L., 188
Sydenham, Thomas, 149 Van der Steen, Wim J., 5, 7, 74, 102,
symmetry between causes of disease 109
and causes of death, 16–17 Veit, A. G. C., 55–7, 59
Symonds, J.A., 14 vibrions 55, 72, 77–8, 81, 92; see
symptomatic characterizations of also bacteria, bacteridia, ferments,
diseases, 18, 36–37, 155; see also germs, microbes, microorganisms,
characterizations of diseases virus
symptoms, 18, 28, 36–7, 107, Villemin, Jean Antoine, 133
149–50, 158, 160, 182, 188, 193 viral theory of disease, 162, 167–78,
syphilis, 25, 31, 37, 105–6, 133, 181; see also theories of disease
162 Virchow, Rudolf, 34–5, 90–91, 94,
97, 109, 133, 142
Takaki, Kamehiro, 186–7, 192 virus, 124–6, 135, 162, 169–78,
talking cures, 177 179–80; see also bacteria,
Tate, George, 44 bacteridia, ferments, germs,
Taylor, F. Kräupl, 9, 53, 109, 197 microbes, microorganisms, vibrion
Texas fever, 167 vitamine, 192
Théodoridès, Jean, 76, 82 Voderman, A.G., 189–90, 192
theories: of disease, 39, 49, 51–2, 54,
55–9, 69, 88, 90–108, 110, 145, Waldeyer, Wilhelm, 58–9, 93
147–61, 162–78, 179–95, 197; Walker, Nigel D., 161
scientific; see scientific theories Warner, John Harley, 23
INDEX 237
Watson, Thomas, 13, 19, 106–8 wine, spoilage of, 60, 70–71, 72,
weak sufficiency criteria, 139–41, 111, 113
181 Winkler, A., 187–8
Wegscheider, Max, 57 Wittgenstein, Ludwig, 8
Weiss, J., 148–50 Wöhler, Friedrich, 63
Welch, William Henry, 162–3 Worboys’ Michael, 6–7, 9, 23
Wertheimer, M., 59 wound infections, 55–6, 58, 60, 90,
Westphal, Carl, 94 92, 95–6, 108, 119, 130–32, 137–8,
Wheeler, John Archibald, 5–6 141–2
Whiggish interpretation of history, 8, Wright, Hamilton, 187
9; see also scientific progress Wulff, Henrik R., 23
Wieger, Friedrich, 58
Wilkinson, Lise, 169–70 yeast 63–4, 113; see also fermentation
Williams, Robert R., 179, 190 Yerushalmy, J., 199
Wilson, Adrian, 9, 23
Winckel, Franz Ludwig Wilhelm, 56, Zakon, S.J., 33–4, 37, 60
59 Zenker, Friedrich Alberti, 34–6

K. Codell Carter - The Rise of Causal Concepts of Disease - Case Histories-Routledge (2017)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

K. Codell Carter - The Rise of Causal Concepts of Disease - Case Histories-Routledge (2017)

Uploaded by

Copyright:

Available Formats

The Rise of Causal Concepts of Disease:

Series Editors: Andrew Cunningham and Ole Peter Grell

Department of History and Philosophy of Science

Titles in this series include:

‘The Battle for Health’: A Political History of the

Medicine and Charity in Georgian Bath:

The Nurse Apprentice, 1860–1977

The Return of Epidemics:

Published 2016 by Routledge

Routledge is an imprint of the Taylor & Francis Group, an informa business

Copyright © K. Codell Carter, 2003

All rights reserved. No part of this book may be reprinted or reproduced or

British Library Cataloguing in Publication Data

Carter, K. Codell, 1939–

Library of Congress Cataloging-in-Publication Data

Carter, K. Codell, 1939-

ISBN 13: 978-0-7546-0678-9 (hbk)

of Specific Causes in Nineteenth-Century Medicine’ (1991)) and I am

1. The sentence is ‘Thoughts without content are empty, perceptions without

Thus the etiological standpoint encourages us to interpret an important

predecessor. Let us say that such a series of theories is theoretically

assimilation of a new anomaly reinforces the credibility of the pro-

ever conceived by physicists, general relativity has the simplest, most

began focusing on a particular domain of organisms – bacteria – that

Causes of Disease in Early

In order to characterize more precisely the causes included in these lists,

1. Necessity and sufficiency apply only to classes of events. To say

without either being the cause of the other: my having swallowed a

Since we will generally be interested in necessity and sufficiency in the

will be attacked with catarrh, another with rheumatism, a third

one were interested in explaining why some individual became ill. It

In the early nineteenth century, there was an elegant symmetry between

There are ways of characterizing diseases other than in terms of symp-

of an accident or of the malfunction of a watch, is a sufficient cause for

stimulating diet’ (Dunglison, 1845). Through more than 3000 pages of

some particular events. Mary Douglas drew attention to the frequency

Years ago Erwin Ackerknecht warned that ‘a fundamentally rational

Earlier causal talk was not wrong or irrational. It was part of a

Universal Necessary Causes

possible causes of each disease. This is exactly what we saw in our

In fact, by 1844, when Henle published his essay on rational medicine,

The background to Renucci’s claim to have discovered the cause of

individual merely in consequence of the attraction produced by the itch

Agostino Bassi began studying muscardine, a fatal disease of silkworms,

fire was burning continuously. When, after several days, I opened

These are the effects produced in general by the contagions that

ous generation. He reported that he caused true muscardine by inocu-

In 1839, four years after Bassi published his book on muscardine,

sample of the brief causal arguments scattered through Gruby’s publica-

Through the 1820s, English and German pathologists occasionally

unencysted sexually mature male and female trichinae in its intestine.

As the clinical picture of the disease became clear, physicians began

We now have a sense of the different paths by which independent lines

A universal necessary cause can be identified for a given disease only

1. The most complete statement of Ricord’s theory appeared in 1838 (Ricord,

1. In the mid 1840s, typical accounts of the etiology of childbed fever

(also called puerperal fever) conformed to similar accounts for

Representative accounts of childbed fever from the 1840s

Childbed fever, a particularly horrible disease, was common in the

Thus, he speculated, in spite of the variations, puerperal fever could still

pneumonia (from which last disease not one case of poisoning

likelihood of absorbing harmful matter was greater following delivery

local miasmatic influences explain this difference in any other way’