General
Agricultural

Industrial
Microbiology

d-^^^^'^^i

MBL/WHOI

Library

sssis

-

^g

^ r-=1

^^^

J

1 ^^^^

,3"

5^^

m
J]

X

ui

^

o

_j^

o

D

m

SB

-— -

D

^

^^^

MORPHOLOGIC VARIATION
AND THE

RATE OF GROWTH

OF BACTERIA

MONOGRAPHS ON GENERAL, AGRICULTURAL AND
INDUSTRIAL MICROBIOLOGY

Edited by

R. E. Buchanan, Ph.D.

KB. Fred

S. A. Waksman, Ph.D.

The subject of Microbiology, including the bacteria, fungi and
other microorganisms, is reaching such broad outlines at the pres-
ent time, that no single book can ever hope to cover the various
phases of the subject, such as classification, morphology, physiology,
numerous applications, etc. No one person, however versed and
active in the field, can ever expect to become a master of this rapidly
growing subject, which has already branched out in a brief period
of time into a number of special fields.

One way to present this subject in a most authoritative and service-
able manner is through the publication of the material in the form
of a series of monographs by authors of repute and authority. Such
a series will enable those who are interested in the subject as a whole
or in certain phases of it to obtain authoritative books at a reason-
able cost. Monographs will be published concerning certain phases
of microbiology in general or concerning personal investigations which
can be treated in a monographic manner. Each author will be
primarily an authority on the subject with which he deals. A com-
plete bibliography will tend to make each monograph an independent
contribution.

Each monograph will consist of one hundred to three hundred
pages and the price will range from $1.50 to $3.50; in many in-
stances the editions will be bound both in cloth and in paper.

R.E.B.
E.B.F.
S.A.W.

MICROBIOLOGY MONOGRAPHS
GENERAL • AGRICULTURAL • INDUSTRIAL

EDITED BY

R. E. Buchanan, Iowa State College
E. B. Fred, University of Wisconsin
S. A. Waksman, Rutgers University

Volume One

MORPHOLOGIC VARIATION

AND THE RATE OF GROWTH

OF BACTERIA

BY

Arthur T. Henrici, M.D.

Professor of Bacteriology
The University of Minnesota

1928

CHARLES C THOMAS, PUBLISHER

Springfield, Illinois Baltimore, Maryland

Copyright, 1928, by Charles C Thomas

First Published, 1928
Manufactured in the United States of America

To my father
who first showed me
the marvels revealed
by the microscope.

PREFACE

This book makes no pretense of being a treatise on the morphol-
ogy of bacteria, but is rather a record of personal researches under-
taken with the hope that by the "magic of numbers" some order
might be brought out of the chaos which has so far filled that field
of bacteriology w^hich has to deal with the form and structure of
bacterial cells. I must state frankly at once that I am no mathema-
tician. The expert in biometrics will therefore find here no such
carefully fitted curves and elaborate formulae as delight his heart,
but the ordinary person may follow my argument with only a very
elementary knowledge of statistics. With so poor a foundation I
should have been rather hesitant about publishing this work were I
not actuated by a conviction that the use of even such simple statis-
tical methods as I could handle has yielded truth of profound signfii-
cance not only for bacteriology, but for general biology as w^ell, and
by the hope that such publication may stimulate others more thor-
oughly equipped to enter this field of investigation.

The history of the problem of morphologic variation in bacteria
is rather curious. The ideas prevalent in the early days of the science
— that bacteria possessed complex life cycles, that the various species
were freely transmutable, and that they were all perhaps but stages
in the development of some higher fungi — were completely upset by
the discovery of methods of pure culture, and replaced by the doc-
trine of monomorphism. This dogma has dominated the field ever
since, being opposed only in recent years by a minority of new
pleomorphists. So firmly has the monomorphistic viewpoint been
established that research on the problem has been almost completely
discouraged; bacteriologists have frequently noted that their organ-
isms appeared different from day to day, but have blindly con-
sidered all variations from the textbook picture as pathological, as
evidence of injury or death of the cells, or else have considered the
variations noted to be of such a haphazard character as to be of no
significance. The development of specific biochemical and sero-
logical reactions has made possible the identification of organisms

ix

PREFACE

without having recourse to their morphologic characters, and the
carrying on of many types of bacteriological investigation without
ever using the microscope.

In this work I shall show that, contrary to the orthodox teach-
ing, the cells of bacteria are constantly changing in size and form and
structure ; but that instead of these changes occurring in a haphazard
or meaningless fashion, or instead of being phases in a rather vague
and complex life cycle, they occur with great regularity and are
governed by relatively simple laws which, after more data have
been accumulated and analyzed, may probably be very precisely
formulated.

Bacteria have been rather neglected by the general biologists, and
in works on general biology one rather gains the impression that
bacteria are to an extent organisms apart from the rest of the living
world, more or less exceptions to the general biologic laws. This is
due in part to the fact that the study of bacteria requires a peculiar
technique, but more particularly to the fact that they are so minute
in size and simple in form. There is no multiplicity of structures
to explain the complexity of function; their very simplicity has re-
pelled rather than invited research. More particularly the absence
of a demonstrable nucleus and of sexual reproduction has discouraged
investigation. Now this would seem on the face of it illogical; the
simplest forms of life should be studied before the complex. Here
are the most elemental organisms that can be visually observed, yet
we know practically nothing concerning their fundamental biology.

My investigations indicate that the growth of bacteria in artifi-
cial cultures is governed by the same laws as govern the develop-
ment of a multicellular organism; that their cells during growth pass
through exactly the same sort of a developmental cycle as the cells
of a plant or animal, exhibiting in turn an embryonic form during the
period of rapid growth, a mature or differentiated form during the
period of slow growth or rest, and a senescent form during the pe-
riod of death; that in short we may speak of a "cytomorphosis" in
populations of free unicellular organisms differing only in degree
from that of muticellular individuals. If this hypothesis be accepted
it must prove of great importance in general biology, for it carries
with it the implication that multicellular organisms are after all but

PREFACE

populations of cells, whose organized individually is a result, not a
cause, of the differentiation of those cells.

These researches have been supported by grants from the Grad-
uate School of the University of Minnesota, for which I am deeply
grateful. I wish also to express here my gratitude to the Depart-
ment of Dairy Industry of Cornell University, which has provided
stenographic assistance in the preparation of the manuscript; to Dr.
W. P. Larson, head of the Department of Bacteriology at the Univer-
sity of Minnesota, who has made every effort to provide the special
equipment necessary for my work; to Dr. J. M. Sherman, who has
kindly read the manuscript of Chapter I; and to Dr. Otto Rahn,
who has similarly read Chapter II for me. Above all I must ac-
knowledge my obligation to Dr. R. E. Scammon, of the Department
of Anatomy of the University of Minnesota, without whose sugges-
tions this work would not have been undertaken, and without whose
kindly assistance and friendly criticism it could not have been carried
through.

I am indebted to the publishers of the Journal of Infectious Dis-
eases, the Journal of Experimental Zoology, the Journal of Experi-
mental Medicine, the Journal of Hygiene, the Proceedings of the
Society for Experimental Biology and Medicine, and of Pearl's
"Biology of Population Growth" for permission to reproduce certain
of the illustrations. Due acknowledgement will be found in the
legends accompanying the various figures used.

XI

TABLE OF CONTENTS

I. The Problem of Morphologic Variation of Bacteria ... 1

II. The Rate of Growth of Bacteria 17

III. On Technique 47

IV. The Size of the Cells of Bacillus Megatherium 59

V. The Size and Form of the Cells of the Colon Bacillus . 87

VI. Some Observations of a Diphtheroid Bacillus 99

VII. A Note on Spore Formation Ill

VIII. Morphologic Variations of the Cholera Vibrio 117

IX. The Senescent Forms of the Colon Bacillus 125

X. On Cytomorphosis in Bacteria 139

Appendix:

Tables 151

References 175

Index 187

xui

CHAPTER I

THE PROBLEM OF MORPHOLOGIC VARL\TION
OF BACTERIA

Bacteriology has developed more as an applied science than as
a theoretical one, and the fundamental biological problems of the
bacteria have been neglected. There are many reasons for this, but
perhaps one of the most important has been the blind acceptance
by the majority of bacteriologists of the Cohn-Koch dogma of the
constancy of cell forms and the immutability of bacterial species,
which has discouraged all investigation of problems of morphology,
inheritance, and variation in bacteria for a good many years. The
progress of science is measured not so much by the wealth of fact
established as by the veins of unsolved problems uncovered, and the
moment a field of knowledge comes to be considered complete, at that
moment this branch of science is dead. It is therefore a healthy
sign that within recent years, through the activities of a small but
persistent group of modern pleomorphists, bacteriology is definitely
breaking away from the Cohn-Koch tradition and is seriously re-
opening for discussion and investigation the old problem of morpho-
logic variation in bacteria.

But this movement has within it that danger inherent in all revo-
lutionary movements, the tendency to go too far in the opposite
direction. It does not necessarily follow that because Koch was
wrong Naegeli was right; and yet in the writings of at least some
of these new pleomorphists, their protests to the contrary not-
withstanding, there is clearly evident a tendency to return to the
old idea that any bacterium may transmute to the form of any other.
Von Niessen, referring to Xohnis' work, exclaims, "Im Prinzip
Naegeli redivivus, Bestdtigung deiner Befunde, vollig unabhdngig
und rein wissenschajtlich, und — was wird die jiihrende Kochshe
Schule dazu sagen?" and indeed the dominant school of bacterio-
logists has so far had very little to say. A hypothesis endlessly
repeated without criticism becomes a dogma, and it is high time there-

1

2 MORPHOLOGIC VARIATION

fore that we stop to evaluate the data and critically analyze the
logic of these disciples of the "newer biology" of bacteria.

While occasional contributions have been made by a number
of other authors, the principal members of this modern school of
pleomorphists are Almquist, Hort, Lohnis, Mellon, and Enderlein.
It is noteworthy that, excepting Almquist, all of these men made
their initial contributions to the subject in the years 1915-1917.
For details of their work the reader must consult their published
papers, the most important of which are listed at the end of this
work. But we may with profit summarize briefly here the ideas
of each concerning the nature of the morphologic variations of
bacteria.

Almquist has made observations of various members of the
colon-typhoid-dysentery group of organisms, especially in old
cultures kept at low temperatures on drying media. In his earlier
work he has particularly emphasized the formation of exogenous
globular forms which are produced from the normal rod forms by
lateral or terminal budding. If these give rise again to rod forms
directly they are called conidia; if they form new globular forms
endogenously they are sporangia, and the cells so formed are
gonidia; this process, however, may result from a sexual conjuga-
tion in which case the mother cell is an oospore, the male element
being a very small conidium called an antheridiospore. From
cultures containing these conidia a new type which bred true was
separated by filtration through Berkefeld filters, the obvious con-
clusion being that the cells which passed through were minute
conidia (1911). These new types when used as antigens gave
rise to agglutinins for the parent culture. In later papers (1924)
he claims to have proved the occurrence of sexual reproduction
by the production of hybrid crosses between typhoid and dysen-
tery organisms, some of the colonies from mixed cultures of these
two organisms being agglutinated by both serums; and by the dis-
covery of haploid and diploid phases in the nuclei, the diplonts be-
ing large globular forms with a nucleus twice as large as in the
normal, haplont rod forms.

Hort, believing that all of the phenomena observed in epi-
demic cerebrospinal meningitis could not be explained by the cur-

MORPHOLOGIC VARIATION OF BACTERIA 3

rent views as to its etiology, was led to make some filtration ex-
periments, and found that from filtered urine, blood, and cere-
brospinal fluid from cases of meningitis, there could be cultivated
a pleomorphic organism appearing as large and small rods and
cocci (1915). He concluded that the meningococcus is but a late
non-virulent stage in the life cycle of £in organism which has a
filterable, virulent phase. Microscopic observation of cultures of
meningococci revealed great variation in size and the forma-
tion of small cells from larger ones by budding, also refractile
globular structures were found forming within some of the
larger cells which he says are ascospores. He therefore considers the
meningococcus to be a pleomorphic fungus belonging to the ascomy-
cetes (1917a). Multiple lateral and terminal buds were also found
to develop on typhoid and dysentery bacilli grown in acid media
(1917b), and these were observed to give rise again to normal rod
forms (1920).

Lohnis (1916) observed a great variety of cell forms in cul-
tures of Azotobacter, and particularly the occurrence of slender,
spore-forming rods, from which he concluded that the normal
morphology of this organism is but a stage in the life cycle of a
spore forming bacillus. Further study of this organism led to the
conception of a complex cycle which was extended to apply to all
bacteria by observations of some forty-two other species. Later
(1921) appeared a very comprehensive monograph, which in spite
of its bias is an extremely valuable review of the literature, in which
he showed that the structures described by him had been repeatedly
observed, described, and illustrated by many of the previous bac-
teriologists, among them the most orthodox.

Lohnis' hypothesis is so complicated that it cannot be readily sum-
marized. A few quotations may serve to outline its general scope.

Contrary to the monomorphistic theory which only knows one constant
type of vegetative cell for every species, and not more than one type of repro-
ductive organ, the endospore, for merely one group of bacter'a, it has been
ascertained by numerous independent investigations that in reality all bacteria
are not only distinctly pleomorphous in their vegetative growth, but are also
able to produce various organs of reproduction. These are : Gonidia, regenerative
bodies, exo- and endospores, arthrospores, and microcysts. All of them are made
up of nuclear substances, which are reinforced by smaller or larger amounts of

4 MORPHOLOGIC VARIATION

reserve material and protected by a more or less resistant membrane. Go-
nidia and regenerative bodies participate actively in the process of multiplica-
tion, whereas the other reproductive organs are in the first place resting forms.
Gonidia, regenerative bodies, and probably microcysts are produced by all
bacteria, while arthrospores, exo- and endospores are less common; though
there are indications that they will be discovered in many more cases, as soon as
these problems are more thoroughly investigated (1921, p. 162).

All bacteria live, in vivo as well as in vitro, alternately in an organized
and in an amorphous stage. By the partial or complete dissolution of the
vegetative and reproductive cells, a plasmatic mass, the symplasm, is formed,
which, after a period of rest, according to circumstances may transform itself into

new cells of the same or of a more or less modified character The

formation of the bacterial symplasm proceeds always in two phases: First, the
cells agglomerate to smaller or larger clumps; second, a more or less complete

dissolution of the cells takes place, resulting in a crumbly or slimy mass

The reconstruction of new cells from the bacterial symplasm follows various
lines according to the internal as well as the external conditions (quality of
the symplasm, environment, and technique). At first always regenerative
units become visible, which either grow up separately to new cells, or of which
several may agglomerate and surround themselves with a uniting membrane,
thus forming at once full-grown cells. (1921, p. 195).

He also describes another process as conjunction, where two or
more cells combine without previous disintegration. Lohnis is care-
ful to point out that "by discussing the life cycles of bacteria we do
not intend to revive any of those unclear theories concerning bac-
terial polymorphism or pleomorphism. The development of bacteria
is characterized not by the irregular occurrence of more or less abnor-
mal forms but by the regular occurrence of many different forms and
stages of growth connected with each other by constant relations"
(1916, p. 677). But the diagram which he presents to show these
constant relations, with its numerous arrows connecting each type
with many others, and with double points indicating their reversi-
bility, leaves one without any clear notion of regularity, but rather
with the idea, in vogue in Naegeli's day, that any bacterial type
may develop in a hap-hazard way from any other. Indeed, from
neither text nor illustrations can one gain any exact impression of
the sequence of events in the life cycle he postulates.

Mellon, in a long series of papers extending over ten years, has
described morphologic variations in diphtheroid bacteria, typhoid and
colon bacilli, and fusiform organisms; many of his papers deal with

MORPHOLOGIC VARIATION OF BACTERIA 5

variations in serologic or cultural characters rather than morphology,
but the origin of the former is traced to the latter in every case;
some of his work has had to deal with pathogenic fungi rather than
bacteria. Again the work is too voluminous to be readily summarized
in a few words. But the main idea seems to be, as in the case of
Lohnis, that bacteria go through a more or less complex life cycle
in which various types, rods, spheres, branching forms, etc., appear
in sequence; further, that any one of these types may be stabilized,
i.e., continuously reproduce itself, without going through the rest of
the cycle. The origin of these stabilized types is sought in sexual
reproduction, which is found particularly in the case of the colon
bacillus in the formation of large globular swellings, previously de-
scribed and illustrated by Scales, at the point of fusion of two rod
forms, and designated by Mellon as zygospores, though conidia and
even endospores are thought to involve at least a process of autogamy.
Mellon believes that the name applied to these structures is relatively
unimportant as long as the essential factors of conjugation and nu-
clear reorganization are recognized. He also describes the separation
of variants from cultures by filtration which differ markedly from the
parent strain, especially in virulence, and implies, as have the pre-
ceding authors, that these variants have been derived from minute,
perhaps ultramicroscopic, conidia.

The latest, and perhaps most remarkable contribution to this
new pleomorphism is the work of Enderlein. He unfortunately has
found it necessary to coin a vocabulary of nearly two hundred new
words to express his ideas, which proves a serious obstacle to the
reading of his work. Once this vocabulary is mastered, however, one
finds that the hypothesis is presented with a clearness and coherence
that is annoyingly lacking in the papers previously reviewed. I say
hypothesis, for one gains the impression that the ideas expressed are
pure theory having only the most slender basis in actual observation.
Only slight details of technique and brief descriptions of actual ob-
servation are given. The drawings, made with extremely high mag-
nifications, are not convincing. His description of the centrosome
with dimensions of 0.01 pi, which, as is well known, is quite below
the limits of resolution of the microscope, seriously impairs one's
faith in the accuracy of his other observations. This impression

6 MORPHOLOGIC VARIATION

that the work is a philosophical one rather than the result of labora-
tory observation is supported by the fact that the author is a zoolo-
gist who has made no other contribution to bacteriology.

Enderlein considers the cytology, life cycles, and taxonomy of
bacteria. The basis of his cytology is the primitive nucleus or
''mych." This with its accompanying protoplasm forms the "mychit,"
which is the building stone of bacterial cells. It may exist by itself
as in the micrococci, or may combine with others in various multi-
ples to form "pliomychits," the cells of larger or more complex bac-
teria. Asexual reproduction is accomplished through equational cell
division, "monogony," in the case of a mychit; "isomorphous ar-
throgony" in the case of more complex cells; or by unequational cell
division, "heteromorphous arthrogony," which can occur only with
pliomychits. The latter may give rise to a number of different repro-
ductive bodies: Gonidia, which contain a single mych; the "cystit," a
gonidium with a compound nucleus ("polydynamen mych") ; the
''cystoid," an unusually large gonidium with much reserve foodstuff;
structures behaving like embryos in that they grow at the expense
of a neighboring or enclosing mother cell; and structures like the
oidia of higher fungi formed by the multiple fragmentation of a cell.
Sexual reproduction is accomplished by the development of gameto-
cytes from gonidia. These undergo a reduction division of the mych,
the haploid mychit resulting being called a "gonit." The latter may
develop into a male cell ("spermit") or a female cell ("oit"). The
former, in the case of the cholera vibrio, is smaller, curved, and ac-
tively motile by means of a polar flagellum ; the latter is larger, glob-
ular, and but sluggishly motile. These unite, followed by a fusion
of their respective nuclei, giving rise again to the diploid phase. En-
derlein also confirms Lohnis' observation of the symplasm.

The life cycle of a bacterium, according to Enderlein, consists of
two simultaneous, parallel, and co-ordinated processes, a multiplicative
development through simple cell division, and a progressive develop-
ment {'^fortschreitenden Entwicklung'^) , very slow and characterized
by morphologic variation. The former is "auxanogeny," the latter
"probseogeny." As the ontogeny of an individual repeats the phy-
logeny of the race, so the life cycle of a bacterium repeats the evo-
lution of the species; and as a multicellular organism cannot avoid

MORPHOLOGIC VARIATION OF BACTERIA 7

in its life cycle of development the one-celled stage, the egg, so the
bacteria are derived from and return to the elementary unit, the
mychit. Each complete life cycle begins with the fusion of two hap-
loid mychits, goes through progressive changes in cell complexity,
reaches a maximum fixed for the species, and returns again to the
haploid mychit. But cycles may be incomplete, shortened, or com-
pletely arrested. The progressive development may be completely
eliminated and only the multiplicative may occur, and this condition,
"mochlosis," is usually the case when bacteria are maintained in a
uniform environment as is true when artificially cultivated in the
laboratory.

Space will not permit of a more complete review of this remark-
able work, but enough has been given to indicate that, while the
whole thing sounds like a strangely complex product of the imagina-
tion, the hypothesis is a very orderly and systematic one compared with
the rather vague and confused ideas of Almquist, Lohnis, Mellon,
and Hort.

The work of Bergstrand should be noted here, though it should
not be considered in the same class with that of the authors pre-
viously mentioned, since he does not draw such radical conclusions.
From a very extensive and careful study of morphologic variations
of a diphtheroid organism (1919), he concludes that morphologic
variations of bacteria are much more extensive than is generally be-
lieved, and is opposed to the idea that all variations from the normal
are to be considered involution or degeneration forms. Bacteria may
reproduce by budding or by abstriction. The pleomorphism of
bacteria may probably be explained by the fact that they are fungi
which may form mycelium. He considers the large globular forms
produced by some bacteria, identified as sexual forms by Mellon
and Almquist, as resting cells analogous to or identical with the
chlamydospores of higher fungi (1923). He finds no evidence of
sexual spore formation (1923) or of a nucleus (1921).

Bergstrand does not go nearly so far as the others, but is satisfied
to emphasize the extent of the variation which occurs and the im-
portance of the problem. The reader will bear in mind, therefore,
that most of the criticism which follows does not apply to his work.

Notice should also be taken of the papers of Fuhrmann on the

8 MORPHOLOGIC VARIATION

life cycles of bacteria. He recognizes two cycles, a minor one con-
sisting of a single generation, i.e., the changes involved in the devel-
opment of a cell from one division to the next, and a major one con-
sisting of the regular changes occurring during the growth and aging
of a culture. This conception was developed from observations of
morphologic variations of an organism isolated from beer, Pseudo-
monas cerevisice. The young forms are motile rods; with increasing
age of the culture these tend to lose their motility and to form
chains; with still greater age certain of them become swollen at one
end and accumulate numerous metachromatic granules. These lat-
ter dauerjormen are the more resistant to various agents than are the
young cells, but not as resistant as the endospores of spore-forming
bacteria. With the latter the endospore formation occupies the same
stage of the major cycle as do the dauerjormen of the beer organ-
ism. When the latter are transferred to fresh media they rupture
and discharge several small cells which immediately develop into
motile rods. He is rather inclined to believe that the metachro-
matically stained granules of the dauerjormen are of a nuclear nature,
and that the schwdrmer develop from them.

Fuhrmann lays considerable stress upon the regularity with which
the various changes in morphology occur in succession during the
aging of a culture. Various external factors tend to influence this
cyclical development only in degree. The age of the culture seems
to be the most important factor. To this extent his ideas anticipate
somewhat and are in agreement with my own. His papers have only
come to my attention during the preparation of this manuscript.

While in most of these papers no exact details of the nature and
sequence of the supposed life cycles are given, and the various authors
are not in perfect agreement on all points, the same general idea
clearly runs through all of their writings. This idea has perhaps been
most clearly formulated by Mellon, who states that "bacteria in
their fundamental biology are in reality fungi that have been tele-
scoped down, as it were, to a somwhat lower order; but this order
is not so low as to preclude the preservation by the bacteria of the
fundamental organization characterizing the fungi and higher plants.
.... In point of fact, instead of reproducing themselves by transverse
fission alone, we know that under conditions none too specialized

MORPHOLOGIC VARIATION OF BACTERIA 9

bacteria may bud, branch, form gonidia, and even structures com-
parable with asci. The latter characteristic implies a complicated
life cycle sexual in nature. . . In any event, it is clear, and now beyond
dispute, that Koch's law of morphologic type specificity cannot longer
be regarded in the absolute sense which has characterized it in the
past. Bacteria do change their morphologic type and within very
wide limits; and with this change may go at times important physio-
logical modifications" (1927, p. 107).

Much can and will be said in criticism of both the work and
the hypotheses of these men, but this criticism should not in any
way minimize the real value of the service which they have rendered
to bacteriology in reopening for discussion and investigation this
fundamental problem. Confronted with the cold skepticism towards
questions of morphologic variation in bacteria common to bacteri-
ologists a decade ago, it required real courage to publish papers of the
type here reviewed. And while these papers have probably
not convinced more than a very few that bacteria exhibit complex
fungoid life cycles, they have demonstrated beyond question of a
doubt that bacteria do regularly show pronounced morphologic vari-
ations, the nature and significance of which must be determined
before we can make any real progress towards understanding the
fundamental biological problems of the group.

One can no longer satisfactorily answer their arguments by the
old formula that they are dealing with irnpure cultures; for while un-
doubtedly in isolated cases conclusions have been drawn from obser-
vations of contaminated cultures, this is by no means general, and
anyone who will patiently study with the microscope his own cul-
tures which he knows to be pure can quickly confirm the general
observation that rod forms may appear in cultures of cocci, spherical
forms in cultures of bacilli, lateral buds and branches, and internal
globular bodies in both. In undertaking a critical analysis of this
work, then, one cannot find iault so much with the actual data as
with the logic followed in erecting the hypothesis.

But the data are not beyond criticism. Throughout these papers
one is impressed with the meagerness and haphazard character
of the observations as compared with the widespread importance
of the conclusions. The small amount of data submitted may per-

10 MORPHOLOGIC VARIATION

haps be explained in part by the present day difficulty in obtaining
sufficient space in scientific journals for complete reports; but the
descriptions of observations offered do not give a hint of any wealth
of observation withheld from publication. The observations are for
the most part casual, not the result of systematically planned experi-
ments. Cultures are examined today and then set in the icebox for
weeks or months, at the end of which time they are found to have
undergone a transformation; but a systematic step by step observa-
tion of this transformation with a record of all the transitional stages
is lacking. Enderlein states in criticism of the monomorphists, "Der
methodische Hauptfehler war anfangs im Hinblick auf die Cyclo-
genie der Mangel einer methodische Beobachtung der Veranderungen
im Laufe der Zeit; meist wurde nur mit jungen Kulturen gearbeitet."
But if one substitutes "old" for "young" in this statement it applies
equally well to the pleomorphists.

The failure to make continuous observation and record of all
stages of the transformations described is the most serious criticism
of the factual material in the work under discussion. It is due to
the technical difficulties involved ; but certainly much more could have
been done in this direction than has been recorded. It is only fair
to state that a certain amount of continuous microscopic observation
has been recorded: Hort (1917a) has illustrated the formation and
separation of buds and the formation of the intracellular structures
which he calls ascospores in meningococci, and the formation (1917b)
and later development into normal rods (1920) of buds in typhoid
bacilli; Almquist records the transformation of spherical conidia by
germination into a "probacterium" in typhoid bacilli; Enderlein
claims to have followed the fertilization of the "oit" by the "spermit"
in cultures of the cholera vibrio. Gardner has clearly observed the
origin of normal rods from branched cells, without, however, con-
cluding that they form part of a complex life cycle. But these iso-
lated cases only serve to emphasize the general lack of that continu-
ous step by step observation which must be made before any clear
idea of a life cycle can be established. It should also be stated in all
fairness that frequently it is impossible to tell from the author's
statement whether in particular cases he is recording hypothesis or
stating observed fact; statements of hypothesis are made so positively

MORPHOLOGIC VARIATION OF BACTERIA 11

that they sound like established fact, particularly in the works of
Lohnis and Enderlein. I have assumed in such cases where illus-
trations or actual descriptions of the method are not furnished that
the statement has not been verified by actual continuous observation.
In most cases, however, it is quite clear that no continuous observa-
tion has been made, and that the author's idea of the sequence of
events has developed from an attempt to trace the transitions in
different individuals in the same preparation. This clearly involves
an element of error. The structures, for instance, which Mellon in-
terprets as zygospores, namely two cells with swollen ends in contact,
might just as well be interpreted as a cell with a central bulging
undergoing division.

This lack of systematic observation, this attempt to patch to-
gether the life cycles from haphazard observation of cultures in
widely different media without regard to the age of the culture or
the phase of growth, is responsible for the vague and incoherent
nature of the descriptions of the life cycles presented. In practically
no case can one obtain from the published papers any clear notion
as to the author's ideas regarding the actual sequence of events in
the cycle; as to when gonidia are formed, and when asci and zygo-
spores; under what conditions simple conjunction occurs and under
what circumstances a symplasm results. Lohnis' complicated dia-
gram is absolutely unintelligible, and the one published by Mellon
to illustrate the ontogeny of a diphtheroid organism (1926, v), while
simpler, is far from being clear.

The fact that practically all of these observed variations in mor-
phology have been found in old cultures, and especially on unusual or
unfavorable media, of course at once leads one to believe that they
are "involution forms" i.e., dead or degenerating cells. This argu-
ment is answered at once by the statement that the cultures contain-
ing them are perfectly viable. But the fact that growth occurs when
such old cultures are transferred to a new medium does not prove
that this growth has resulted from the cells showing the morphologic
variation, even when such variant cells preponderate. Only careful
quantitative observation of the proportion of viable cells and of mor-
phologically variant cells, or single cell isolations of variants to new
media, or continuous observation of particular variant cells trans-

12 MORPHOLOGIC VARIATION

ferred to new media will answer this question of viability. None
of the authors has followed the first procedure; Mellon has used
the second in a few instances, and as has been mentioned above, Gard-
ner, Hort, and Almquist have followed the third in the case of bud-
ding and branching cells.

Hort states that an involution form is, "strictly speaking, a sterile
organism which is not only incapable of maintaining its reproductive
activity, but is also incapable of maintaining its integrity of form"
(1920, p. 370). Whatever may be the dictionary definition of the
word, this idea is erroneous, for senescence and death are processes
requiring some time and the terms are relative. It may well be
that an injurious agent or process has led to the deformity of a cell
long before it has actually been killed, and that such a cell, an in-
volution form in the sense that it has undergone a modification of
form as part of a degenerative process, is still capable of "reviving"
and multiplying when transferred to a more favorable environment.
Even the proof of viability of single variant cells does not, therefore,
prove that their variation is not the result of a degenerative process.

Throughout these works is the assumption that bacteria possess
discrete nuclei; and Almquist, Mellon, and Enderlein have definitely
interpreted structures which they have seen as nuclei. It would take
too much space to debate here this much disputed question
concerning the presence or absence of nuclei in bacteria, but I feel
safe in stating that no one has so far demonstrated certainly that
any of the true bacteria possess a discrete nucleus. The structures
described by Meyer and Paravicini are the only ones that have been
reported which fulfill the requirements of a nucleus, but their find-
ings have not been confirmed by anyone except in the case of endo-
spore-forming bacteria during the stage of spore formation. The
assumption that bacteria possess nuclei is therefore unwarranted, and
the casual way in which this or that intracellular structure is desig-
nated a nucleus without any clear evidence is characteristic of all
of the data. Almquist mentions no attempt to differentiate his nu-
clei from volutin, neither does Mellon, though Meyer has clearly
pointed out the danger of mistaking this material for chromatin and
the means of differentiating it. Mellon lays stress on the skein-like
structure as evidence of a nuclear nature, but this structure may be

MORPHOLOGIC VARIATION OF BACTERIA 13

assumed by volutin in old cultures of diphtheroid bacilli (personal
observation). Lohnis accepts the older uncritical works on the cy-
tology of bacteria, and almost any internal structure may be a nu-
cleus. Enderlein claims to have differentiated his "mych" from re-
serve foodstuffs, but I doubt that his observations can be substan-
tiated. Almquist's recognition of haploid and diploid phases on size
alone hardly deserves serious consideration. In this connection it
should also be noted that Lohnis has apparently made no attempt by
micro-chemical methods to determine the nature of his regenerative
granules, i.e., whether or not they might be volutin, and Hort has
similarly overlooked this possibility with regard to the so-called
ascospores of the meningococci. In short such differential staining
processes as are at our disposal for identifying intracellular ergastic
substances have not been used on the various structures interpreted
as nuclei, gonidia, regenerative bodies, etc., and until they have been
used we are justified in suspecting that these structures may be fat
or volutin or other material.

The most serious criticism of this new pleomorphism, however,
has to do not so much with the data or the technique, as with the
line of reasoning which has led to the development of the theories of
life cycles. Nowhere does one find a clear argument starting from
proved facts, but only either vague analogies or conclusions drawn
from unwarranted assumptions. An example from Almquist will
serve to illustrate the most important fallacy. He says:

Die mannlichen Antheridien habe ich zweifelsohne oft in meinen Prapara-
ten gesehen, obgleich ich die Kopulation nie sicher beobachten konnte. Wohl
sah ich manchmal geschwollene Formen mil feinsten, an der Seite festsitzenden
Komchen oder Nadelchen. Dem naheren verlauf ich konnte aber nicht folgen.
Ich nehme doch an, dass die zahlreichen, 0.5m. oder darunter messenden Mikro-
konidien und auch andere feine Kornchen zu den Antheridien gehoren.

Henceforth these structures are referred to as antheridia, though
their identity as such is only an assumption on the part of the author.
The same thing has been done repeatedly by Mellon, Lohnis,
Hort, and Enderlein, though never so frankly admitted. Structures
look like conidia or ascospores or zygospores or what not; therefore
they are named such and in further reasoning are considered as such,

14 MORPHOLOGIC VARIATION

as though the mere naming of them had somehow or other proved the
author's opinion as to their identity.

This persistent naming of all the structures described with names
borrowed in part from mycology, in part from cytology, or as in the
case of Enderlein, coined for the occasion, only serves to obscure the
problem and to make it more difficult for the reader to follow the
author. One cannot help but suspect that the necessity for using
a new terminology has its origin in the obscurity of the thought
which the author is trying to express. The confusion introduced by
this terminology is all the greater because the same structures are
referred to by the various authors under different names, and the
same names are used to designate different structures. Thus the
lateral and terminal projections from cells are referred to by Hort
simply as buds, by Almquist as conidia, by Lohnis and Enderlein as
gonidia, by Mellon sometimes as conidia, sometimes as exospores.
Almquist speaks of a single bacterial cell of irregular or ameboid
form as a plasmodium, but states that this is identical with Lohnis'
symplasm; while Mellon applies the same term to the connecting
filaments of protoplasm described by Meyer under the name plas-
modesmids. Mellon refers to the globular structures arising in colon
bacilli as zygospores, but to illustrate their nature offers for compari-
son a picture of the ascospores of Endomyces; and in a footnote
implies that the two words have the same meaning. In fact, to
Mellon the terms chlamydospore, arthrospore, dauerzellen, gonidium,
and zygospore all apply to essentially the same sort of a process; "they
may be viewed as branches of the same reorganization tree." But
these terms have very distinct meanings in mycology, and refer to
structures physiologically as well as morphologically different, and
their confusion can only result from an imperfect knowledge of the
nature of the processes under consideration.

There is another unwarranted assumption in the reasoning of
these various authors, namely that permanent variation is an evidence
of sexual reproduction, that all variants are hybrids. No matter
how well we may consider this to be established for higher organisms,
we are certainly not justified in accepting it for unicellular organisms
where non-sexual reproduction is common and where there is no
differentiation into somatic and germ cells. Jennings has shown that

MORPHOLOGIC VARIATION OF BACTERIA 15

by artificial selection quite different races may be obtained from
cultures of protozoa in which no sexual reproduction has occurred.
No clear evidence of the occurrence of hybridization in bacteria has
as yet been furnished, Almquist's experiments with so-called typhoid-
dysentery crosses being too incomplete. It is not enough to show that
variation has occurred; it must also be shown that the variation fol-
low the Mendelian law. This has recently been claimed for the muta-
bile variant of the colon bacillus by Stewart, but his article is not con-
vincing since it was possible to obtain from the "hybrid" only one
of the assumed parent factors. His assumption that the other char-
acter was associated with a lethal factor that suppressed growth is
hardly warranted, since the same conditions hold with practically
all bacterial mutants, and it is hardly likely that in every hybrid-
ization one of the characters is associated with a lethal factor. Topley
and Ayrton have pointed out that these variations can be as well
explained by a non-equational cell division; that we may have segre-
gation without syngamy. With most bacterial variations it is
rather difficult to distinguish between temporary modifications and
true mutations, and the possibility of inheritance of acquired modi-
fications in microorganisms has not been disproved. In short, our
know^ledge of the mechanism of variation in unicellular forms is quite
too incomplete to permit us to use such variations as a premise from
which to argue the occurrence of sexual reproduction.

With these tw^o unwarranted assumptions, that similarity of form
indicates indentity of function, and that variation is necessarily an
evidence of conjugation, we are led into a sort of circle of false logic
which runs something like this: Certain cell forms of bacteria look
like conjugating cells or products of conjugation. Let us call them
such; they are ascospores (or what you like). Since they are sexual
spores they should at times give rise to hybrids. Variations are
found; therefore sexual reproduction occurs; therefore our assumption
of the sexual nature of the observed structures was valid. But there
is no proved fact anywhere in this argument, save the existence of
both morphologic and cultural variation.

In fact, neither the data nor the logic of these new pleomorphists
are adequate to convince us that bacteria possess complex fungoid
life cycles. We may sum up the present status of the problem by

16 MORPHOLOGIC VARIATION

stating that bacteria in pure culture do show wide variations in the
size and form and structure of their cells, but that the nature and
significance of these variations are not known.

"Was ist nun zu tun?" says Almquist. "Wir miissen die Natur
objektiv untersuchen und den Tatsachen unterordnen."

The problem is one particularly adapted to statistical analysis,
since it deals with large numbers of individuals varying among
themselves at any point in time, and varying continuously with time.
This work has to deal with an attempt to apply such statistical
methods to the problem of morphologic variation in bacteria, with
the hope that by such methods the investigation will be truly kept
objective, and that the facts will eventually truly be placed in order.

CHAPTER II

THE RATE OF GROWTH OF BACTERIA

". . . .It is perfectly true that all changes in form, inasmuch

as they necessarily involve changes of actual or relative magnitude,

may in a sense be properly looked upon as phenomena of growth; and

it is also true, since the movement of matter must always involve an

element of time, that in all cases the rate of growth is a phenomenon

to be considered." _,, _

D Arcy Thompson

The outstanding fact established by my studies is that bacteria
are continually varying in morphology with increasing age of the cul-
ture in which they are growing, and that the morphologic variations
are correlated with the rate of growth, the transition from one type
to another occurring at the points of inflection between phases of
the growth curve. Before discussing these morphologic variations,
therefore, it is desirable to devote some space to a review of what
is known concerning the growth curves of bacteria and the factors
which determine them.

When we place viable bacteria in a suitable medium they gen-
erally do not commence to grow at once, but after a longer or shorter
period of quiescence they begin to divide, at first rather slowly, then
with greater and greater rapidity until a maximum growth rate is
attained, after which they continue to divide more and more slowly
until growth ceases altogether. After a longer or shorter period
during which the number of cells remains practically constant, they
begin to die, the rate of death becoming greater and greater until
a maximum death rate is attained, then becoming slower and slower,
some cells remaining viable in most cultures for long periods of
time. Thus if we plot a curve of the number of living cells in the
culture, against time, there is obtained a characteristic S-shaped
curve of growth, which is repeated in reverse during the period of
death.

Such growth curves for populations of microorganisms in cul-
tures are essentially similar to the curves obtained for the growth

17

18

MORPHOLOGIC VARIATION

of other populations, as fruit flies in a bottle or human populations
in a restricted area; and also to the growth curves of those co-
ordinated populations of cells which we call multicellular organisms.
This has been emphasized by Pearl (1925), who concludes that a
common law governs the growth of all such populations, which he
states as follows:

"Growth occurs in cycles. Within one and the same cycle, and
in a spatially limited area or universe, growth in the first half of the

dlO.

•r*

? (J

S 9-

b-

X

^

^ 7-

^/

\

m

/

/

1 5'

V. k .

o

<« 3-

t /

\

s

\

•5 2 •

\

\

a 1 .

V

^ 0.

i" M H

Id

15

2d

2i

3rf

Units of Time.

Fig. 1. The Growth Curve of Bacteria.
From "Life Phases in a Bacterial Culture" by R. E. Buchanan. Reproduced
through the courtesy of the Journal of Infectious Diseases. (1918, 23, 109).

cycle starts slowly, but the absolute increment per unit of time in-
creases steadily until the midpoint of the cycle is reached. After
that point the increment per unit of time becomes steadily smaller
until the end of the cycle. In a spatially limited universe the amount
of increase which occurs in any particular unit of time, at any point
of the single cycle of growth, is proportional to two things, viz.:
(a) The absolute size already attained at the beginning of the unit
interval under consideration, and (b) the amount still unused or

RATE OF GROWTH OF BACTERIA 19

unexpended in the given universe (or area) of actual and potential
resources for the support of growth."

Growth curves of bacteria are obtained by counting the number
of cells in a culture by one or another method, usually plating, at
regular time intervals during the growth of a culture. The results
may be plotted in various ways. The actual number of cells may
be plotted against time, as has been done in Fig. 9. This serves to
show the absolute increase or decrease, but not the rate of change,
and has the disadvantage that where the numbers are small, at the
beginning and end, slight but relatively important changes are not
shown. The usual method, illustrated in Fig. 1, is to plot the loga-
rithms of the numbers of cells against time. Such a curve shows the
proportional increase or decrease in the number of cells, and the slope
of the curve is a measure of the rate of change; it possesses the ad-
vantage that variations at the beginning and the end, when the
numbers are small, are plotted on a scale such that these variations
are clearly distinguishable. Since, after the maximum growth period
has passed, changes in the rate of increase or decrease occur at an
increasingly slower rate, it is sometimes advantageous to plot the
logarithms of the numbers of cells against the logarithms of the time
intervals. This serves to magnify the scale of the abscissae during
the early stages of growth when changes are occurring rapidly. Such
a curve cannot be used to measure the rate of growth, but still serves
to show clearly the division into phases. It has been necessary to
follow this procedure in Fig. 25 and Fig. 33 simply to obtain space
for the cells forms and the growth curve on the same graph.

The rate of growth has also been expressed in terms of generation
time, i.e., the time required for the cells to double in number. The
mean generation time may be computed for any time interval by
dividing the time by the number of generations, and the latter quan-
tity may be determined from the cell counts at the beginning and
the end of the time interval according to the formula

log. b — log. a

n = ;

log. 2

where n is the number of generations and a and b are the first and
last cell counts, respectively. The mean generation times for the

20 MORPHOLOGIC VARIATION

various intervals of observation may then be plotted against time.
Such a curve has the advantage that it expresses more clearly than
the others changes in the rate of growth; it has the serious disad-
vantage that it tends to magnify small experimental errors in count-
ing, so that most data must be iirst smoothed by some method before
they can be used in this way.

Mathematical analyses of the growth curves of bacteria have
been made by McKendrick and Pai, Ledingham and Penfold, Slator
(1917), and Buchanan; of yeasts by Slator (1913, 1918, 1919), and
by Carlson; of growth curves in general by Pearl and by Robertson;
and various empirical formulae have been derived to express these
curves. Where these are not interpreted in terms of organisms, sub-
strate, products of metabolism or other definite factors, they do not
seem to be very helpful to an understanding of the phenomena to
one who is not mathematically minded.

For purposes of description and analysis it has been generally
found convenient to divide the growth curve of bacteria into phases.
This is in a sense unfortunate in that it makes what is essentially a
continuous process appear somewhat discontinuous, but is almost
necessary for intelligent discussion. Lane-Claypon described four
phases, an initial lag period during which the number of cells re-
mains constant or very slowly increases, a period of maximum growth,
a stationary or resting period, and a period of death. This phasing
has been followed by most succeeding authors, but Buchanan further
subdivided these phases, making seven in all, illustrated in Fig. 1,
which is taken from his paper. They are the initial stationary phase,
from the origin to a, during which the number of cells shows no
increase (it may actually decrease, according to Chesney) ; the lag
phase, or period of positive acceleration in growth rate, a-b; the
logarithmic growth phase when the rate of growth is constant and
maximum, the logarithms of the number of cells falling on a straight
line, b-c; the phase of negative acceleration, c-d] the maximum sta-
tionary phase, d-e\ the phase of accelerated death, e-j ; and the
logarithmic death phase, during which the cells are dying at a
constant rate, and the logarithms of the number of cells again fall
on a straight line, f-g. A final phase of negative acceleration in death
rate might have been added, for in many cases at least the death

RATE OF GROWTH OF BACTERIA 21

rate decreases after a time, and some cells may be found alive after
relatively long periods.

This division into phases is more obvious in cultures which grow
rather slowly. Most organisms when grown on the surface of agar
develop much more rapidly than in liquid media (which have been
used almost exclusively in growth studies of bacteria) and the growth
curves of such agar cultures, as will be seen later, do not show such
a sharp separation into phases, the so-called logarithmic growth
phase being practically absent, the maximum rate being maintained
for only a brief interval, so that the logarithmic curve is nearly
S-shaped; this S-shaped curve is then repeated in reverse during
the period of death, but extends over a much longer interval of time.
The total growth curve thus has the appearance of a rather skewed
frequency distribution curve. One can then distinguish a period of
accelerating growth, a period of negative acceleration in growth, a
period of accelerating death, and a period of negative acceleration
in death.

Various factors, as temperature; the size, age, and previous history
of the inoculum; and the composition and nutrient value of the
medium, influence the form of the growth curves of bacteria. It
is of course well known that growth is slower on either side of the
optimum temperature, the decrease in growth rate being greater
per degree with higher than with lower temperatures. It is apparent
that the effect of temperature is due both to a shortening of the lag
phase (Tenfold, Chick) and to an actual increase of growth rate
during the maximal growth period (Lane-Claypon, Barber). Lane-
Claypon interprets the effect of temperature on the growth rate
as an operation of the Van't Hoff law. Zikes (1919b) has made
the interesting observation that the growth rate-temperature curve
for yeast is modified by the temperature at which the strain has
previously been cultivated. After being continuously subcultivated
for some time at a low temperature (8 degrees), it was found that
growth was more rapid at all lower temperatures than before, and
that there was a tendency to develop two optima; this might be
interpreted as indicating that the growth-temperature curve is in
reality an expression of the frequency distribution of the cells with
regard to their optimum temperature, and that by growth at low

22

MORPHOLOGIC VARIATION

temperatures a new strain with a lower optimum was being selected
out. Graham-Smith has made the rather surprising observation that

Ik

16

0 2 k 6 S 10 12

Initial Seeding, — Cells per 1/4000 Ccm.
Fig. 2. Influence of Size of Seeding on Rate of Growth of Yeast.
Plotted from data of Adrian Brown (Jour. Chem. Soc. Lond., 1905, 87,
1395.)

while growth is slower at lower temperatures, it continues longer
and the final level of the culture is actually considerably higher

RATE OF GROWTH OF BACTERIA

23

8 12 IG 20 2k 32 3G kO kk ^

Hours of Growth.
Fig. 3. Influence of Size of Seeding on Growth Curves of Yeast.

24 MORPHOLOGIC VARIATION

than in cultures grown at a higher temperature; the death rate is
also correspondingly slower at lower temperatures. This was not
confirmed by Tanner and Wallace with thermophiles.

Of the various factors which influence the rate of growth and
the form of the growth curve, the initial number of cells introduced
into a unit volume of medium seems to be one of the most important.
Rahn first showed that the duration of the lag phase was greater
with smaller seedings, and Penfold, although finding fault with Rahn's
data, was able to confirm his conclusion. Montank has demonstrated
the same thing for yeasts (Figure 3). It should be noted that the
occurrence of a lag phase depends upon the age of the culture used
for seeding, and that if the material used for inoculation is in the
active growth phase no lag will occur in the subculture no matter
how small the seeding (Sherman and Albus, 1926).

McKendrick and Pai noted that the final yield in cultures fourteen
hours old was practically independent of the size of the seeding,
and that therefore the rate of growth is higher for smaller seedings.
This was confirmed by Graham-Smith as long as the seedings are
smaller than the normal maximum yield of the medium. Adrian
Brown has calculated the mean rates of growth of yeast cultures
with various sized seedings, and derived a simple formula to express
the relationship. Figure 2 is a curve plotted from his data, in which
the yield from one cell in eighteen hours is plotted against the mitial
density of the population. It will be seen that the curve is hyper-
bolic, and that with very small seedings a very slight difference
in the number of cells introduced makes a very great difference in
the rate of grow^th, while with larger seedings the effect is slight.
In these observations only the mean rate of growth for relatively
long periods is available, i.e., it is impossible to separate the effects
upon the lag phase and the growth phase.

Montank, working in my laboratory, has, however, obtained data
in which such a separation is possible.* His results are shown graphi-
cally in Figure 3, the data for which are presented in Table I. In

*I am greatly indebted to Mr. Montank for the privilege of using here
his as yet unpublished observations, part of an investigation still in progress.
The data from which his curves have been plotted will be found in Tables 1
and 2.

RATE OF GROWTH OF BACTERIA

n

25

?o "30 Ho 50 ^0 70

Mean Number of Flies per Bottle.

Fig. 4. Influence of Density of Population on Reproduction Rate of.

Fruit Flies.
Reprinted from 'The Biology of Population Growth" by Raymond Pearl,
by and with permission of and special arrangement with Alfred A. Knopf, Inc.

26 MORPHOLOGIC VARIATION

these curves the logarithm of the number of yeast cells is plotted
against time. It may be clearly seen that with smaller seedings,
while the lag phase is prolonged in proportion to the decrease in
the size of the seeding, in all cases (except with the smallest seed-
ing), when once growth has been initiated, the actual rate of growth
is higher, as indicated by the slope of the curve.

It is interesting to note that the size of the "seeding" has a simi-
lar effect upon the growth of individuals and populations of higher
organisms. Figure 4 shows a curve for the mean rate of increase
of a population of fruit flies from the data of Pearl. It is essen-
tially the same as the curve plotted from Brown's data for yeasts
(Figure 2). Figure 5 presents curves for the rate of growth of re-
generating tadpoles' tails with different sized stumps, taken from the
data of Ellis. The logarithms of total tail length have been plotted
against time, and the graph is comparable with the curves for yeast
growth in Figure 3. It will be noted that the same relationships
hold. With a large stump (seeding), only a small part of the tail
having been removed, there is no lag, and the rate of regeneration
is slow, as indicated by the slope of the line; with decreasing size
of the stump, the rate of regeneration becomes higher, and with the
smallest there is a distinct lag before regeneration commences.

The age of the culture from which the inoculated cells are de-
rived also has an influence upon the rate of growth in the new
medium. Miiller first demonstrated that the duration of the lag
period was greater with older cultures. While Rahn and Coplans
found that this was true, they still observed some lag when the
culture used for seeding was young, and concluded that the lag
phase could not be entirely eliminated by using actively growing
cells. But Barber, isolating single cells, found that if these were
growing at their maximum rate when transplanted, they would con-
tinue to grow at that rate in the new medium, and Tenfold observed
that the same thing was true with ordinary culture transplants, point-
ing out that probably Coplans and Rahn were dealing with cultures
which had passed the maximum growth phase even though they
were young as measured in hours. Chesney has demonstrated this
effect of age of the parent culture most clearly in his work on the
lag phase of the pneumococcus. Figure 6 is a graph copied from

RATE OF GROWTH OF BACTERIA

27

his paper, the lower curves showing the initial rates of growth
of subcultures from the parent culture, the growth curve of which

g

1.3

U

V 1

1 1

^ — "

^^^^^^ ^____,.,,.— —

^/

^-

,

O 3 6 9 1

t 15 18

Days of Growth.

Fig. 5. Influence of Size of Stump on Rate of Regeneration of Tadpoles'

Tails.
Plotted from data of M. M. Ellis (Jour. Exp. Zool., 1909, 7, 421).

is shown immediately above. It will be noted that when sub-
cultured during the lag phase, the subcultures show no immediate

28 MORPHOLOGIC VARIATION

growth, but when subcultured during the logarithmic growth phase,
they grow at once at the same rate as in the parent culture, while
after the resting phase has been reached in the parent culture, the
subcultures show slight or no growth during the first two hours.

The previous history of the culture used for seeding apparently
has some influence upon the rate of growth or the form of the
growth curve. Tenfold states that transfer from one medium to

5 7—-

I'

to

Ix
So

=F=f=f=f

I I I I I I I I I I I I I I I I I I I

15 20 25

Hours.

Fig. 6. Influence of Age Of Parent Culture on Initial
Growth Rates of Pneumococci.
From "The Latent Period in the Growth of Bacteria" by A. Chesney.
Reproduced through the courtesy of the Journal of Experimental Medicine.
(1916, 24, 387).

another of different composition may give a longer lag, the organism
requiring a period for adaptation to the new environment, although
he presents no data of his own. I have had an occasion to test
this theory, having at hand two cultures of the colon bacillus de-
rived from the same parent culture which had been carried for over
five years continuously, one on ordinary nutrient agar, the other on
a "synthetic" medium composed of various salts and glycerine.
These were subcultured simultaneously into flasks of nutrient broth
and the above mentioned medium, taking care to keep the size
of the seedings approximately the same. While the growth rate was
somewhat higher in the case of the broth cultures, both strains
behaved approximately the same, and there was no appreciable

RATE OF GROWTH OF BACTERIA 29

difference in the duration of the lag phase. Graham-Smith has
observed that the frequency with which the parent strain has been
subcultured influences the rate of growth, the growth rate being
slightly higher and the final yield somewhat greater if the parent
culture has been rapidly transplanted for some time.

McKendrick and Pai concluded from their mathematical analysis
of the growth of bacteria that the concentration of nutrients in the
medium was a limiting factor for growth, and that the rate of growth
was proportional to this concentration, and the same conclusion was
arrived at by Carlson for yeasts. Penfold and Norris observed the
effect of varying the concentration of peptone on the rate of growth,
taking care to eliminate the lag phase from their data. Their re-
sults are presented graphically in Figure 7. In reading this graph
it should be noted that the ordinates are the generation times, the
reciprocal of the actual rate of growth. With this in mind it is
interesting to compare their curves with that of Brown for the
effect of the size of seeding in yeasts. The two curves are very
similar. It would seem, therefore, that the amount of available
foodstuff bears a reciprocal relation to the number of cells in its
effect on the growth curve. With small concentrations, slight dif-
ferences have a profound influence on the rate of growth; with
larger concentrations the effect is much less, the critical concentra-
tion in this experiment being about 0.4 per cent peptone. Salter,
using higher concentrations of peptone, also found that increasing
the concentration increased the rate of growth, the effect being most
pronounced in the lag phase. Curran, on the other hand, found
very little difference in the rate of growth in cultures in
1 per cent and 4 per cent peptone, and this difference was mani-
fested in the later stages of growth, not at all in the lag phase.
Zikes (1919a) also found that while the final yield of yeast was
proportional to the concentration of wort, and the rate of growth
must be correspondingly proportional, the initial rate of growth was
actually somewhat greater with smaller concentrations of foodstuffs.
Montank has observed the rate of growth of yeast in media of
decreasing concentrations of dextrose and peptone. His results are
shown graphically in Figure 8, the data in Table II. There was
practically no effect on the duration of the lag phase, but the rate

30
800

MORPHOLOGIC VARIATION

6oc-

4oc« .

w

20(

M "TbO

Concentration of Peptone.

Fig. 7. Influence of Concentration of Nutrients on Rate of Growth

OF Bacteria.
From "The Relation of the Concentration of Food Supply to the Genera-
tion-Times of Bacteria" by W. J. Penfold and D. Norris. Reproduced through
the courtesy of the Journal of Hygiene. (1913, 12, 527).

RATE OF GROWTH OF BACTERIA

31

0

12

15 20 2k 2S 32 36
HOURS OF GBQWTH

Uo 14+ i;s

Fig. 8.

Influence of Concentration of Nutrients on Growth Curves
OF Yeast.

32 MORPHOLOGIC VARIATION

of growth was roughly proportional to the concentration of the
medium, as was the final yield.

Graham-Smith has contributed some valuable information on
the relationship between the concentration of foodstuff and the
growth of bacteria, though his observations are confined to the
later stages of growth and are therefore not useful in determining
the relative effects upon the lag and logarithmic phases. Figure
9 is copied from his paper, and shows curves for the number of
cells in five cultures in media containing graded concentrations of
meat extract. It will be noted that the maximum number attained
is proportionately greater with increasing concentrations of nutrient,
and also that growth is continued for a longer period with increas-
ing amounts of extract. Further experiments showed that after
growth had come to a standstill and the number of cells was de-
creasing, the addition of new meat extract in concentrated solution
led to the development of a new cycle of growth, and that this could
be repeated a number of times. By adding small quantities of ex-
tract in concentrated solution daily to the cultures, the death phase
could be postponed for some time, the number of living cells remain-
ing at a constant level. The addition of distilled water, diluting
the medium, decreased the number of cells per unit volume but not
the total number of cells in the flask.

It would appear, therefore, that the concentration of available
foodstuff is a factor of prime importance in determining not only the
rate of growth but also the final yield, the duration of the growth
period, and possibly also the rate of death ; and that the concentration
of nutrients bears a reciprocal relation to the initial density of the
population in determining the form of the growth curve, in con-
formity with Pearl's "law of growth" stated at the beginning of this
chapter.

Some information is available on the influence of various other
factors upon the rate of growth of bacteria. Salter found that the
inhibitory effect of crystal violet on B. coli was more pronounced
on young cells than old ones, and acted more in prolonging the lag
phase than in reducing the rate during the logarithmic phase. Cur-
ran found that a large series of compounds added in small quanti-
ties tended to prolong the lag phase. Graham-Smith similarly found

RATE OF GROWTH OF BACTERIA

33

ro

flours.

Fig. 9. Influence of Concentration of Nutrients on Growth

Curves of Bacteria.

From "The Behavior of Bacteria in Fluid Cultures as Indicated by Daily

Estimates of the Numbers of Living Organisms" by C. S. Graham-Smith.

Reproduced through the courtesy of the Journal of Hygiene. (1920, 19, 131).

34 MORPHOLOGIC VARIATION

that the addition of acid and alkali to the medium tended to retard
growth during the earlier phases, though subsequently growth pro-
ceeded at the normal rate. These results may be explained by the
fact demonstrated by Sherman and Albus (1923) that "physio-
logically young" cells of bacteria are much more sensitive to various
injurious agents than are older ones. Sherman, Holm, and Albus
have shown that the accelerating effect of small quantities of neutral
salts added to the medium is due to both a shortening of the lag
period and an actual increase in rate during the growth phase.

In considering the growth curves of bacteria, two main problems,
more or less interrelated, present themselves, viz: Why do the
organisms stop growing, and why do they not start to grow at once
with maximum rapidity when introduced into fresh medium? These
have both been the subject of much theorizing.

Three explanations have been offered for the first problem:
prowth ceases because the foodstuff has been exhausted and the
organisms are starving; or because certain by-products of growth,
either excretory materials or specific "autotoxines" are injurious to
the organisms; or because actual physical crowding limits them.

A little thought will show that this last theory is quite untenable.
When grown on the surface of agar, the organisms are more closely
crowded together at all stages of growth than is ever the case in
liquid media, and during the later stages are packed together as
closely as they can be, yet with practically all aerobic bacteria the
growth rate is higher on agar than in broth, and the final yield is
higher than is the case with an equal volume of broth of the same
nutrient value. Clark has shown that by growing yeast in a paper
filter with the wort continually added, the yield was practically as
high as that obtained when allowed to grow free in the wort, though
the yeast cells were actually packed in a solid paste. I have simi-
larly found that when yeasts are grown in a paper extraction thimble
through which they could not pass, suspended in a large flask of
dextrose broth, the total yield closely approached that in the control
flask where the cells were free to diffuse through the medium; a
similar experiment with colon bacilli contained in a collodion sack
did not give a yield so closely approaching the expected amount,
but still the actual density was much greater than when grown

RATE OF GROWTH OF BACTERIA 35

free in the broth. Finally Rogers and Whittier have found in the
case of Streptococcus lactis that if the effect of crowding is eliminated
by repeatedly filtering off the organisms during growth and return-
ing the filtrate to the parent culture, the period of multiplication
and fermentation is not prolonged.

The idea that growth is limited by the accumulation of some
toxic substance is the one that seems to be most generally accepted,
though the evidence for it is far from being convincing. It is diffi-
cult to say who first suggested this theory, but Eijkman was one
of the first to clearly formulate it. He grew the colon organism
in gelatine for fourteen days at 37 degrees, divided the gelatine into
two parts, one of which was heated to boiling, the other left un-
heated, and allowed both to solidify in culture dishes, after which
they were inoculated on the surface with the same strain. No growth
occurred in the unheated culture; a good growth occurred on the
heated gelatine. Three possible explanations are offered: (1) That
the medium was exhausted of nutrients which were set free again
in the medium by heating; (2) that during growth antagonistic
products of metabolism accumulated which were destroyed by heat;
(3) that a diffusible thermolabile inhibitory substance was formed
during growth.

The first possibility was excluded because after removing the
bacteria from the unheated gelatine culture by filtration, it. again
yielded a good growth; therefore the nutrients were not exhausted.
Herein lies a fallacy which has been repeated by almost all of the
workers on this problem. It is assumed that growth should continue
as long as any nutrient remains, taking no account of the amount
necessary to merely sustain existence. The amount of food utilized
by a cell during any given period of time will depend not only upon
the amount of food present, but also on the number of cells present,
and a culture which contains only sufficient food to maintain exist-
ence without growth for a population of say a million cells: per c.c.
will have enough to permit a very active growth of a population
starting with say ten cells per c.c; and similarly a medium might
yet contain enough nutrient to support an active growth of a small
number of cells, when with a very large number they would be
actually starving and dying. That is, the occurrence or absence of

36 MORPHOLOGIC VARIATION

growth, or the relative rates of growth, depend not on the absolute
concentration of nutrient, but on the amount of food available per
cell. The occurrence of growth in a medium in which organisms
have grown and from which they have been removed is therefore
no proof that their cessation of growth in that medium was not
due to a partial exhaustion of the foodstuff present.

Eijkman's filtration experiment indicated that if a toxic substance
were active it would not pass a porcelain filter, even if the latter
were paper thin. Further experiments showed that the growth in-
hibiting substance was destroyed by any temperature continued long
enough to kill the organisms, and by volatile antiseptic substances
— ether and ammonium sulphide — which could be evaporated from
the medium after they had acted. Similar results were obtained
with a number of other species of bacteria.

A number of further interesting experiments were made with the
gelatine plate technique. If the "coli-gelatine" (gelatine in which
coli had grown, resolidified in a plate) were inoculated with coli and
covered with a layer of fresh gelatine, no growth occurred. If a
fresh gelatine plate was streaked with coli and covered in part with
fresh gelatine, in part with coli-gelatine, the former part grew,
the latter did not. This was interpreted as indicating that a diffusible
inhibiting substance passed out of the coli-gelatine into the fresh
medium. Similar experiments were made, covering the coli-gelatine
with paper dipped in agar; inoculation of the latter failed to give a
growth unless the coli-gelatine had been heated. This experiment
was made to prevent the possibility of organisms migrating from coli-
gelatine to the upper layer of medium.

From these experiments Eijkman concludes that growth inhibi-
tion is due to a diffusible, very thermolabile substance formed by
the bacteria themselves. Further experiments showed that it pos-
sessed a certain degree of specificity, inhibiting some closely related
organisms but not other species. He failed to explain its failure
to pass through porcelain filters in spite of the fact it would diffuse
through agar and gelatine, and also found that it could not be
separated from the bacteria by centrifuging.

It will be noted that in none of the above experiments was the
hypothetical inhibiting substance separated from the living bacteria

RATE OF GROWTH OF BACTERIA 37

save by a relatively thin layer of agar or gelatine, and in all cases
the results can just as well be explained on the ground that no growth
occurred because there were too many cells present for the amount
of nutrient available. As Roily pointed out, the apparent diffusion
of the toxic substance into the fresh gelatine laid on the coli-gelatine
might be explained by a diffusion of nutrients from the former to
the latter.

Conradi and Kurpjuweit criticized Eijkman's technique on the
ground that he used gelatine and agar as media, colloidal substances
which might adsorb the toxic substance and so decrease its activity.
They explained the failure of the substance to pass a porcelain filter
as being due to this adsorption. They conducted experiments similar
to those of Eijkman, growing their organisms in broth which, after
varying periods of incubation, was added to molten agar which was
poured into culture dishes. They claimed that no growth occurred
in these agar plates even though the living broth culture had been
added to the agar, but this was not true, the number of cells intro-
duced being so great that only microscopic colonies occurred, which
were apparently overlooked by Conradi and Kurpjuweit, but were
demonstrated in similar experiments by Manteufel and by Roily.
After the agar had solidified it was inoculated on the surface with
the same or some other organism, and the presence or absence of
growth determined. They concluded that all organisms produce
more or less specific inhibitory substances which they called "auto-
toxines," which are formed in quantities parallel with the growth
of the organisms themselves, reaching a maximum at twenty-four
hours and not decreasing until after some days, comparable with
phenol in potency, highly thermolabile, diffusible, and very easily
adsorbed. The same criticism will apply to their experiments as
to those of Eijkman, namely, that the inhibition of growth is due to
an excess of living organisms within the medium. They claimed
to have separated this autotoxine from the bacterial cells by di-
alysis. This could not be confirmed by Oebius or by Kauffman, and
I also have failed to note any inhibition of growth in broth in which
had been suspended a collodion sack containing a broth culture of
the same bacillus.

Rahn, working with B. fluorescens liquejaciens, also concluded

38 MORPHOLOGIC VARIATION

that the inhibition of growth was due to a non-filterable thermolabile
substance. Unlike Eijkman, however, he found that the organisms
could be killed with ether (subsequently evaporated off) without
destroying the inhibitory effect. A heated forty-eight-hour culture,
re-inoculated, gave a good growth; a similar culture treated with
ether gave no growth, but when heated later again allowed growth
to occur; a tube of sterile broth treated with ether as above gave
a good growth, showing that the ether treatment itself was not
inhibitory; the same was true when a heated forty-eight-hour culture
was treated with ether. Here is a clean-cut experiment which would
seem to allow but one conclusion, namely that some substance inhibi-
tory to growth had been formed, which was destroyed by heat but
not by ether. The failure to pass a porcelain filter was explained
by its adsorbability ; small pieces of a filter placed in an ether-treated
tube served to destroy the inhibiting effect. A number of other
organisms were tried, but most of them could not be killed by the
ether treatment; only three others, B. lactis erythro genes, Vibrio lac-
tis, and Micrococci's grossus, gave positive results. One might sus-
pect that here we are dealing with a special case not generally
applicable to bacteria at large, that perhaps the substance is pyocya-
nase or a related substance.

Chesney observed that curves of death of pneumococci in cultures
appeared to follow the monomolecular law, as had previously been
shown to be the case when bacteria are killed by disinfectants; and
that when pneumococci are centrifuged out of an actively growing
culture the remaining cells continue to grow actively, but if the
culture be centrifuged after the period of maximum growth the re-
maining cells in the supernatant liquid show a prolonged lag and
may even die for a time. Culture filtrates from twenty-four-hour
cultures tended to inhibit the growth of actively growing pneumo-
cocci inoculated into them, the inhibitory action being lost if the
filtrates were held for a time at incubator temperature. Actively
growing pneumococci exposed to the action of a twenty-four-hour
culture filtrate at low temperature showed a longer lag than controls.
Filtrates of four-day-old cultures failed to inhibit. He concludes
that the inhibition of growth is due to some unstable toxic product
of growth which injures the cells. M'Leod and Govenlock also be-

RATE OF GROWTH OF BACTERIA 39

lieved that the limitation of growth of pneumococci was due to
the accumulation of a thermolabile toxic substance which they named
"bacterocidin," but in later papers M'Leod and Gordon showed that
this substance was hydrogen peroxide and that it was produced by
several other organisms, notably the streptococci. This was con-
firmed by Avery and Morgan, and Morgan and Avery concluded
that the limitation of growth in pneumococcus cultures is due to
three factors, the accumulation of acid, the production of peroxide,
and the exhaustion of the nutrients. These findings would indicate
that here again we are dealing with a special case not generally appli-
cable to all bacteria.

Hajos grew various members of the colon-typhoid group in broth
until a maximum growth was attained, centrifuged out the bacteria,
re-inoculated, and incubated again, and repeated this process until
no further growth occurred. This exhausted medium was distinctly
inhibitory to bacteria; a few drops added to 10 c.c. of fresh broth
was sufficient to produce a definite inhibition of growth. The toxic
substance was merely inhibitory, failing to kill bacteria suspended
in it, was not specific, was not destroyed by heating to boiling, and
passed through a Bechold ultrafilter. This substance is quite dif-
ferent from those previously described. Hajos considers it to be a
product of metabolism rather than a specific "autotoxine." This
substance can, however, hardly be of great importance in the cessa-
tion of growth in normal cultures, since it could only be obtained
in quantities sufficient for demonstration after the medium had been
completely exhausted by repeated subcultures.

Curran, observing that growth occurred in media in which bac-
teria had previously reached a maximum growth, and from which
they had been removed by heating or by filtration, also concluded
that the cessation of growth is due to a thermolabile substance
readily adsorbed by bacterial filters. To test the latter theory more
carefully, he performed the following experiment. A three-day old
culture of B. coli was passed through a filter, the first and last 50
c.c. portions of the 200 c.c. culture being retained. These two por-
tions were tested separately for their ability to support growth; the
culture in the last portion grew more slowly, only half as many
organisms being present after ten hours as in the first portion, though

40 MORPHOLOGIC VARIATION

both cultures were equally and simultaneously seeded. He con-
cludes, therefore, that there is a toxic substance which is adsorbed
by the filter, but that after a certain period of filtration the filter
becomes saturated and allows the substance to pass through. This
experiment, like the filtration experiment of Chesney, is very sug-
gestive, but cannot be considered conclusive until it has been re-
peated a sufficient number of times to show that it is universally true.

With the relatively recent discovery of the bacteriophage, the
question has naturally arisen whether or not these supposed auto-
toxic substances formed in cultures of bacteria might not be the
d'Herelle principle. Hajos says that the inhibitory products of
metabolism which he obtained from exhausted cultures have nothing
in common with the bacteriophage. Otto and Munter, apparently
accepting as valid Conradi and Kurpjuweit's autotoxine, showed that
the latter was not neutralized by antilysin. Kauffman, denying the
existence of the autotoxine, naturally concludes that it is not identi-
cal with the lytic substance of d'Herelle.

While many of the experiments reported above are very sug-
gestive of the occurrence of toxic products which tend to limit growth,
and one cannot, therefore, categorically deny their existence, there
is nowhere any clearcut evidence that such substances are universally
formed, sufficient to warrant the prevalent acceptance of the idea.
The observations of Penfold and Norris, Salter, and Graham-Smith
with bacteria, of Brown, Zikes and Montank with yeasts, that the
rate of growth and the final yield are proportional to the concen-
tration of nutrient in the medium, seem to offer a serious obstacle
to the toxic theory, for it seems more likely that toxic products of
growth would accumulate more rapidly and thus inhibit growth
sooner with a medium of higher nutrient value, rather than the re-
verse. The crucial experiments of Graham-Smith, in which he
showed that a culture of staphylococci could be revived at any time
after the maximum growth period and caused to grow again by
simply adding some more meat extract in concentrated form, and
that the death phase could be postponed indefinitely by small daily
additions of meat extract, would seem sufficient to completely upset
the toxic theory, unless one can believe that meat extract is an anti-
dote to the autotoxine!

RATE OF GROWTH OF BACTERIA 41

If the cessation of growth is due to a partial exhaustion of the
medium, repeated subcultures on the same medium, with removal
of the organisms after each culture, should behave as cultures in
media decreasing in concentration of nutrients, and one should ob-
tain a series of growth curves similar to those in Figure 8. While
there is much evidence in the literature that such repeated sub-
cultures do yield steadily diminishing amounts of bacteria, there is
very little quantitative data available. Rahn measured the rate
of growth in successive cultures of B. fluorescens in parallel series
heated to 100 degrees and 68 degrees, but the various cultures were
not equally seeded, and growth was not continued to the maximum
in each case, so that the figures are not directly comparable. The
68 degree series gave the following mean generation times for the
first twenty- four hours of growth in five successive crops: 98, 96, 108,
108, and 103 minutes, certainly not enough difference to be signifi-
cant. In these cultures, of course, the dead cells remained in the
medium, and it is probable that on heating some nutrient was
liberated from them and returned to the medium, and that during
growth further food was made available by the action of enzymes
secreted by the living cells upon the dead ones. The amount of
nutrient thus made available, however, must be small. Similar
studies with filtered or centrifuged cultures would be more signifi-
cant.

Curran made an experiment similar to the above, using B. coU,
the medium being autoclaved and reseeded every four days. The
first culture yielded 500,000,000 cells per c.c; the four subsequent
crops yielded about 100,000,000 each time. He also measured the
rate of growth hour by hour in a culture which had been heated, with
a control in fresh broth, and another culture from which the bacteria
had been removed by filtration with a similar control. The results
of the two experiments were very similar; in neither case was there
any difference in the duratian of the lag phase between the used
broth and the fresh medium, while in both cases there w^as a decrease
in the rate of growth during the late logarithmic phase in the used
broth.

The apparent contradiction in the findings of various authors
with regard to the limitation of growth in cultures may be explained

42 MORPHOLOGIC VARIATION

if it be agreed that different factors may limit growth in different
cases, i.e., that every culture is a special case like the pneumococci,
and that there is no generally applicable law. This is probably
true. Thus Brown found that in ordinary flask cultures of yeast,
oxygen was definitely the limiting factor; if the cultures were well
aerated, food seemed to be the limiting factor, the yield being greater
in wort of higher specific gravity. Clark believed that in aerated
yeast cultures the accumulation of alcohol was the limiting factor,
but the work of Balls and Brown would indicate that under such
circumstances the concentration of dextrose determines the rate of
growth. Rogers and Whittier have found that in cultures of Strepto-
coccus lactis the accumulation of acid is normally the limiting factor;
if this be removed by keeping the hydrogen ion concentration con-
stant, then growth proceeds to a higher level which is determined
by the oxygen concentration, a still higher level being reached if
the culture is aerated. Probably similar different limiting factors
will be discovered in each case, and if they are removed the final
factor will be found to be the concentration of foodstuff. At least
there is no sound basis for believing in the production of specific
autotoxines.

Many theories have been proposed to explain the period of
latency at the beginning of the growth curve. Chesney has well
divided these into two groups, those which seek the origin of lag in
the medium, and those which find the cause in the inoculated cells.

Rahn found that cultures seeded into a medium in which the
same organisms had previously grown and had been killed by heat,
had little or no lag, although there was a definite latent period in
control cultures on new medium. There was a definite lag, however,
when organisms were recultivated in a medium from which the
preceding crop had been removed by filtration. He concluded that
the maximum growth rate could only be attained at a definite con-
centration of some heat stable, non-filterable substance formed by
the bacteria themselves.

This idea that the lag is due to the lack of some substance
secreted by the bacteria has also been held by various other authors,
but more particularly by Robertson, who has referred especially to
a similar phenomenon in the growth of protozoa. He lays especial

RATE OF GROWTH OF BACTERIA 43

stress upon the fact that the lag period is shortened by increasing
the size of the seeding, and assumes that this is due to a mutual
stimulation of the cells by each other. Since the cells are separate,
this can only be accomplished by their shedding something into
the medium, and this, Robertson states, is a catalytic agent liberated
during the process of cell division. This mutual stimulation of the
cells Robertson refers to as the "allelocatalytic effect."

But Cutler and Crump, and Greenleaf, failed to observe any
allelocatalytic effect with protozoa, and Peskett had similar negative
results with yeasts. Curran failed to observe any shortening of lag
when bacteria were seeded into broth separated from an actively
growing culture by a collodion membrane; therefore if there is an
extracellular catalyst it is not diffusible. But, as was demonstrated
by Chesney, there is no lag if the organisms are taken from a culture
in the maximum growth phase, and this was found to be true regard-
less of the size of the seeding by Sherman and Albus (1924b),
growth continuing at a maximum rate if the inoculum was but one
cell per c.c. It would seem, therefore, that if lag is due to the
lack of some extracellular catalyst, this catalyst is required only
by cells which had passed the logarithmic growth phase in the
preceding culture. As will be pointed out later, the effect of the
size of the seeding can be readily explained on the basis of variability
in the inoculated cells with regard to their ability to grow^ the
larger the number introduced, the greater the chance for obtaining
rapidly growing individuals.

Since actively growing cells show no lag and cells taken from
a culture which has stopped growing do, Chesney rightly concluded
that the cause of the lag is to be found notin the medium, but in
the cells, and seeks this cause in an injury which the cells have
sustained in the preceding medium, and from which they must
recover before they can start to grow. He believes the cause of
this injury to be the toxic substance above mentioned. But proof
of the general existence of such a toxic substance has so far not
been furnished. However, this theory will still hold if we believe
that the cessation of growth is due to a lack of nutrients, and
that the cells have been injured by starvation. It would seem, at
least, that whatever causes a slowing and final cessation of growth

44 MORPHOLOGIC VARIATION

in the culture also somehow changes the cells so that they are noi
immediately able to respond to fresh medium by growth; whether
this change is of the nature of an injury, and just what is the cause,
is not clear.

The latter view is practically that of Buchanan, namely that
the cells of a culture which has ceased growing are changed into
"resting" cells physiologically analogous to spores or seeds, and that
when introduced into fresh media these must "germinate" before
they can grow. The lag phase is the time required for this germina-
tion. Buchanan's theory postulates, therefore, a functional difference
between growing cells and resting cells, and it is interesting to note
that such physiological differences have been demonstrated by Sher-
man and Albus (1923), and to anticipate somewhat by stating that
I have found corresponding morphological differences. Sherman and
Albus (1924a) also explained lag on this basis, observing that the
increased sensitivity of "physiologically young" cells could be demon-
strated before the maximum growth rate had developed. It is also
interesting to note in this connection that Durbin explained the
latent period in the growth curves of regenerating tadpoles' tales
as being due in part at least to the time required for the adult cells
of the stump to transform to embryonic cells capable of growing.

Buchanan first divided the lag phase into two parts, an initial
dormant period and the true lag phase or period of accelerating
growth. The former may be quite prolonged, as has been observed
by Burke and his co-workers, who believe that such prolonged
periods of dormancy, especially observed with spores, have nothing to
do with the ordinary lag phase. But I believe that the difference is
only one of degree, and that the short period of dormancy commonly
observed is the same sort of a phenomenon. During this period
of dormancy some of the cells may actually die, and as will be
seen later, they may also show morphological changes of the same
sort as they did in the parent culture. This period of initial dor-
mancy, then, is not entirely occupied by a transformation, either
physiological or morphological, to a new type of growing cell. It
may perhaps best be explained in the terms of Tenfold, as being due
to an "inertia", i.e., to the bacteria continuing to react for a time
in the new medium to a stimulus received in the old.

RATE OF GROWTH OF BACTERIA 45

Penfold has suggested another theory that deserves mention,
namely that "the inoculum consists of organisms having individually
different powers of growth, and that during the lag the selection
of a quick growing strain occurs in response to some selecting agent
in the peptone". This explains the gradual acceleration of growth
in the true lag phase. It will be seen from later chapters that such
individual variations in power of growth occur, and that a selection
of a rapidly growing strain may take place.

CHAPTER III

ON TECHNIQUE

In the work here reported, two things were desired: To obtain
as accurate as possible a measurement of the rate of growth of the
cultures studied, and also to obtain, as far as possible clear micro-
scopic preparations of the cells, from which reasonably accurate
measurements of their size and form could be made, since the pur-
pose of the work has been to study morphologic variations of bac-
teria quantitatively, and to correlate these variations as far as
possible with the rate of growth. For various reasons it was neces-
sary to use solid media for the growth of the organisms. First,
because it is difficult to obtain clear microscopic preparations from
liquid media, because of the presence of peptone or other material
which obscures the picture; and second, because it is necessary to
obtain considerable numbers of cells for the purpose of microscopic
observation during the early hours of growth when they are relatively
few in number in the medium. If liquid media were used, it would
be necessary to centrifuge, wash, and recentrifuge, in order to clear
the organisms of the accompanying debris of the medium, and to
sufficiently concentrate them. In addition to being time consuming,
it was thought likely that this process would be apt to lead to some
artificial variations in the morphology of the cells.

Broth cultures were, however, used for certain studies of B.
megatherium described in Chapter IV (Cultures II-VI). Since this
organism grows very slowly in liquid media because it grows only
at the bottom and therefore lacks oxygen, some method had to
be used for providing aeration. In Cultures II and III this was
accomplished by inoculating 1,500 c.c. quantities of broth, and
then distributing this in 60 c.c. amounts into Kolle flasks, so that
the broth was disposed in a thin layer with a large surface exposed
to the air. A flask was removed for sampling after each hour, assum-
ing that the conditions of growth were identical in all of the flasks.
In Cultures IV, V, and VI, 1,500 c.c. flasks of broth were also inocu-

47

48 MORPHOLOGIC VARIATION

lated, but aeration was accomplished by thoroughly shaking the flasks
every half hour during growth. Samples were removed from the
flasks with sterile pipettes.

The difficulties met with in broth cultures can be obviated by
using solid media, agar slants, the cells being sufficiently concen-
trated and obtained practically free of products of the medium itself
by simply scraping them off with a wire loop. The use of agar,
however, introduced certain other difficulties. In an ordinary agar
slant the medium varies in depth from the bottom to the top, and the
conditions of growth are therefore not uniform throughout the tube.
As a matter of fact, growth is heavier and probably also more rapid
in the lower part of the tube. In order to obtain a representative
sample from agar slant cultures, then, it is necessary to suspend
the entire growth in water and remove a sample from this suspension
for examination. This necessitates using at least one agar slant cul-
ture for each observation, and therefore the preparation of a large
series of agar slant cultures as uniform as possible in size of seeding
and conditions of growth for the entire period of observation. This
was approximately accomplished by using a large series of measured
test tubes of practically uniform diameter. These were obtained by
first boring two holes into a piece of sheet brass, differing in diameter
by about one-half millimeter. Test tubes as they came in the com-
mercial packages were then selected by means of this piece of brass,
only those tubes being chosen which would pass through the larger
opening but not through the smaller one, the variation in diameter
of the tubes thus being practically negligible. These tubes were
then filled, each with 5 c.c. of agar carefully measured, and were
also sloped, as far as possible, at the same angle. The tubes, there-
fore, were probably sufficiently uniform for the purposes, the diffi-
culty being in obtaining uniform inoculation.

Seeding was accomplished by preparing a suspension in sterile
water, or by using a peptone or broth culture, and one loopful of
such a suspension was spread as uniformly as possible over the sur-
face of each slant. Undoubtedly the loopfuls varied somewhat in
size, and more particularly it was difficult to maintain uniformity
of distribution of the seeding over the surface of the agar slant.
At any rate there was a certain amount of unavoidable variation in

ON TECHNIQUE 49

the results, which could only be explained as being due to a lack
of uniformity in the various tubes. This was most noticeable in
the early hours of growth.

In all of the work standard beef extract agar and bouillon were
used as culture media, adjusted to pH 7.

The cell counts and microscopic preparations for purposes of
measurement were made from the growth on the above mentioned
agar slant cultures suspended in sterile water. For the later stages
of growth, where the number of cells was large, generally one or two
tubes were removed from the incubator and the growth washed off
with water and used for preparations for counting and measuring.
Samples for counting were immediately killed by adding formalin.
In the early stages of growth, however, the number of cells is so
small that it is necessary to remove a number of tubes in order to
obtain a sufficient number of cells for reasonably accurate counting,
and particularly to obtain a sufficient number of cells for micro-
scopic preparation and measurement.

The procedure followed was to remove a number of tubes and
add to the first 2 or 5 c.c. of water. The bacteria were then scraped
off into this water, which was poured into a second tube, and this
process repeated from tube to tube until enough bacteria had been
removed to give a faint but perceptible turbidity to the liquid. The
number of tubes used and the amount of water used was varied
according to the stage of growth and the number of bacteria actually
present in the cultures, the final calculations being made so that the
number per c.c. is expressed as though the growth from one tube
had been suspended in 2 or 5 c.c. of water, as the case may be.

Of the various methods which offered themselves for obtaining
the cell counts necessary for plotting growth curves, one of two
microscopic methods was used in each case. There were various
reasons for this, the most important being that with some of the
organisms studied the organisms tend to clump or to form long
chains, and the degree of clumping and chain formation varies with
the age of the culture. Plate cultures, therefore, would introduce a
serious source of error, which can be more or less obviated by the
microscopic method. With Bacillus megatherium, which is relatively
large, counting was done by the use of the Helber counting chamber.

50 MORPHOLOGIC VARIATION

With this organism all of the cell counts were made with the same
counting chamber and cover glass, so that whatever absolute error
there may have been, there is no error in the course of the growth
curves due to this technique. Wilson has concluded that when the
number of cells is lower than fifty million per ex. the error of the
counting chamber method becomes too great to permit accurate
results, but it is evident that this depends entirely on the number
of counts made and averaged for the final results, i.e., on the total
number of cells counted regardless of their concentration. With the
agar slant cultures the cells were actually concentrated during the
early hours of growth until the liquid showed some turbidity, which
generally indicates about a million cells per c.c, so that the count
is the average of a number of tubes. This method involves a certain
amount of error in that not all of the cells are washed off in the
collecting process. With the broth cultures no such concentration
was possible, and the counts during the early hours of growth had
to be made by patiently going over one preparation after another
until a sufficient number of cells was counted; in general a thousand
cells were chosen as the limit, but in the very early hours of the
lightly seeded broth cultures this was quite impractical, and the
counts here are admittedly not of a high degree of accuracy.

With the other organisms, however, because of their small size
and their tendency to remain floating in the suspension, the counting
chamber method was too difficult and too prone to error to be
useful. For all of these other organisms, therefore, I have used a
direct microscopic counting method devised by myself, but in prin-
ciple essentially the same as that introduced by Breed and Brew
for counting bacteria in milk. The suspensions of organisms were
mixed with an equal volume of a 2 per cent solution of Congo red,
and .01 c.c. of this mixture was spread as uniformly as possible over
an area of four square centimeters on clean glass slides. During
later phases of growth, where the number of organisms is larger,
eight square centimeters were sometimes used. The Congo red
preparations, after drying, were then placed a moment in a solution
of 1 per cent hydrochloric acid in alcohol, which serves to fix the
film and also to change the Congo red to a deep blue. The count-
ing of the organisms is then accomplished by counting the number

ON TECHNIQUE 51

of cells in a number of representative fields on the slide, and comput-
ing the area of each field, and thus obtaining the number in the
total preparation. This method presents certain difficulties. The
films do not dry uniformly; that is, they are much thinner at the
edge than at the center, and the cells are correspondingly more
numerous in the center than at the periphery. It is necessary, there-
fore, to count a number of fields chosen at regular intervals along
a line extending from one edge through the center to the opposite
edge of the slide. In general, twenty fields were chosen, ten along
each of two lines at right angles to each other, the spacing of the
fields being easily accomplished by means of a mechanical stage.
There is also some difficulty in accurately discharging .01 c.c. pipettes
onto a slide, and leaving behind on the tip of the pipette the same
quantity each time. It is therefore very desirable to prepare a num-
ber of slides, in most of the work here reported — five. The total
number of cells counted in most of the cases, therefore, was the
number contained in one hundred microscopic fields; that is twenty
fields from each of five different preparations. This is tedious but
necessary to obtain uniform results. Where the number of cells is
small, as during the early stages of growth, frequently many more
fields had to be counted to obtain a count of accuracy comparable
to that during the later stages of growth.

Since practically all of the cell counts were made by direct micro-
scopic methods, death of the cells is not recorded in the curves
until the dead cells have completely autolyzed. Wilson has pointed
out that some cells are dying at practically all stages of growth,
but during the growth cycle proper the number of these is small,
and the error is practically negligible. In the death phase, however,
the microscopic counts do not give a true measure of the rate of
death, since of course the curve does not begin to decline until some
time after the cells have actually begun to die; and further, as will
be shown in Chapter IX, because the rate of autolysis is not depen-
dent upon the death rate. The death phase therefore began much
earlier and the rate was probably much higher than is indicated in
the curves given, particularly with the cholera vibrio as described
in Chapter VIII. No particular attention has been paid to the death
phase, however, save in studies of the involution forms of the colon

52 MORPHOLOGIC VARIATION

bacillus described in Chapter IX, and here the death curves were
determined by plate counts.

The measurements of morphologic characters, size, and form of
the cells, were accomplished in various ways. With Bacillus mega-
therium, because of its large size, these measurements were made
fairly easily. Microscopic preparations were projected, using a 1.5
mm. apochromatic objective and a 1 2 X periplan eye piece, with an
arc lamp as illuminant, onto a screen at such a distance that the
magnification was approximately 3000 X. The individual cells were
then easily measured simply by spreading dividers over the pro-
jected image and holding these dividers to a ruled scale. In some
cases measurements have also been made by the use of the Abbe
camera lucida, simply spreading the dividers along the table over
the apparent image of the cells. For the smaller organisms, how-
ever, it is necessary to obtain a higher magnification, in order to
have reasonably accurate results; and since projection at magnifi-
cations beyond 3000 X is impractical, because of the difficulty in ob-
taining sufficient light, the following procedure was adopted:

Photomicrographs were prepared of numerous fields, using tne
above named lenses, the magnifications varying from 2000X to
3000 X, according to the bellows draw, and the photographic nega-
tives were again projected at a magnification of seven to ten times.
All photographs were made on panchromatic film, using a Wratten
red filter, which gave the maximum contrast. The projects images of
the cells were then either measured directly, or more frequently they
were traced onto individual cards and the drawings on these cards
were then measured for area by use of a planimeter, and for length by
means of dividers. This last procedure may be considered as involv-
ing considerable error, because of the extreme magnification, the final
magnification being about 20,000 X. Naturally with such magnifi-
cations the edge of the image is far from being sharp, there being
a zone of fuzziness corresponding in dimensions to the limits of reso-
lution of the microscope, that is, to about 0.15[jl in terms of the
original microscopic preparation. This error is, however, a com-
pensating one, for in trying to place the dividers or trying to trace
with the pencil as far as possible through the center of this zone of
unsharpness, the chances are as great for going inside the middle

ON TECHNIQUE 53

of this zone as outside of it; and where relatively large numbers
of measurements are made, this error tends to cancel itself out.
While on theoretical grounds none of these methods of measurement
are as accurate as could have been obtained, for instance, by the use
of a filar micrometer the large number of measurements made serves
to give at least a reasonably close approximation to the true measure-
ment, and because of the much greater length of time required, it
would have been practically impossible to make all these measure-
ments by use of a micrometer.

In the case of the colon bacillus (Chapter V) and the cholera
vibro (Chapter VIII) I have attempted to illustrate the average size
and form of the cells by means of composite photographs. These
were prepared from the tracings of the cells on cards as described
in the paragraph above. The proper exposure for one card was de-
termined, and then all of the cards were photographed on one plate,
each receiving a proportional share of the entire exposure. The
cards were all so placed that the center of each cell came at the same
point on the plate, and that the long axes of the cells coincided sis
nearly as possible. If the cells were curved, the cards were so placed
that the major curvature always came on the same side.

Nearly all of the measurements have been made from slides pre-
pared as above described, using the negative staining method of Beni-
ans, where the organisms are surrounded by a field of dark blue
Congo red. This procedure was chosen because it seems to give
the least amount of shrinkage of any of the methods used for making
permanent microscopic preparations of bacteria. The image is not
that of the bacterium itself as it appears after it has been dried and
fixed, but of the space which was occupied by the cell before it
underwent the shrinkage which occurs during drying and fixing. This
fact, previously stated by Eisenberg, can be easily demonstrated, for
with the large organisms, at least, one can see the shrunken cell
within the space which is not occupied by the dye, there being a
considerable zone between the cell and the edge of the dye. In a
preliminary series of measurements of a preparation of Bacillus mega-
therium, measuring first the living cells suspended in water, then the
clear spaces in the Congo red preparation, and finally a dried fixed
slide stained with methylene blue, it was found that the measure-

54 MORPHOLOGIC VARIATION

ments of the Congo red preparation were practically identical with
those of the living cells, whereas the methylene blue preparation
gave dimensions about one-third less. With the Congo red method,
therefore, the image of the cell is larger than with other staining
methods, and the measurements are correspondingly more accurate.

Other negative staining methods could be used, as India ink or
nigrosin; but India ink is more granular and tends to give a less
sharp image, and at least some samples of nigrosin tend to become pre-
cipitated by certain substances associated with the bacteria, and
similarly give a granular background with an unsharp image. The
negative staining method also appears to give a much sharper image
than is the case with slides stained in the ordinary way.

This negative staining procedure with Congo red gives such clear
preparations for certain types of morphologic investigation that a
word or two further concerning it will not be out of place. A per-
fect image of the cell is only obtained when the film of stain comes
up to the very edge of the cell but does not extend over it. It is
therefore not useful with organisms which accumulate any appreciable
amount of slime about them, and only those parts of the film are to
be used where the layer of stain is not so thick as to extend over the
top of the cell. This latter point may be readily determined by ex-
amination of the preparation ; if one observes the cells in a thick por-
tion it will be found that they are relatively slender and not clear
white, but as one proceeds to the thinner parts it will be found that
they become larger, and that one soon arrives at a place where no
decrease in the intensity of the blue is accompanied by an increase
in the size and brilliancy of the image of the cells, indicating that
the film is now so thin that it has not extended over the top of the
cells. The Congo red is precipitated by peptone, and with other
liquid media artifacts are produced by the precipitation of substances
dissolved in the medium, so that the method is not satisfactory
with organisms from liquid cultures unless they have first been
centrifuged out and washed with water. For this reason cell
measurements for the broth cultures described in Chapter IV have
been made from preparations stained with methylene blue.

While in young actively growing cultures none of the cells show
any portion stained with the Congo red, the entire cell appearing as

ON TECHNIQUE 55

a colorless space in the film, this is not true in older cultures. Eisen-
berg has noted that with the Gram-negative bacteria there frequent-
ly appears a colored spot in the middle of the cell, which he inter-
prets as being due to plasmolysis, the shrinkage of the protoplasm
to the poles of the cell leaving a concavity in the center which be-
comes filled with dye solution; such colored portions were not ob-
served with Gram-positive organisms which could not be readily plas-
molyzed. I have also observed this, but in addition have noted that in
B. megatherium blue stained granules appeared in the cell during the
period of spore formation, corresponding in size and position with
the sporogenous granules which could be demonstrated in prepara-
tions stained with methylene blue, and also that stained granules
appeared in the preparations of the diphtheroid organism described
in Chapter VI, corresponding with the volutin granules demonstrated
by Ponder's stain. It would appear, therefore, that such structures
are stained by the Congo red with the technique employed. It is
hard to understand why such intensely basophilic granules should
stain with an acid dye like Congo red, while the remainder of the
protoplasm remains unstained. Ernst has also observed that the
granules which bear his name are stained by Collargol (which, ac-
cording to Eisenberg, behaves like Congo red), free granules con-
tained within vacuoles in mold mycelium becoming stained a dark
brown. There is, then, no question but that these colloidal nega-
tively charged dyestuffs can penetrate within the cells.

With increasing age of the cultures it is also found that some of
the cells become uniformly stained throughout by the Congo red.
Because these stained cells become increasingly more numerous with
increasing age of the culture, I have concluded that they are dead
cells. But cultures killed by formalin or other fixative solutions do
not become stainable by the Congo red, and cultures killed by heat-
ing to boiling show only a faint staining after some time. Cul-
tures killed by heating to 60 degrees, however, show all of the cells
distinctly stained after standing several hours following the heat-
ing. It would appear, therefore, that the mere death of the cells
does not cause them to become stainable, but that the change in their
reaction towards Congo red is due to some process which occurs after
death, i. e., to beginning autolysis.

56 MORPHOLOGIC VARIATION

Various other staining methods which will differentiate living
from dead cells have been described. As is well known, Gram-posi-
tive bacteria become Gram-negative in old cultures, and R. and W.
Albert have shown that the loss of the Gram-staining property in
yeasts is correlated with the degree of autolysis of the dead cells.
Fraser studied the action of various dyes on living and dead yeast
cells, and found that these could be differentiated readily by methy-
lene blue; his procedure was to suspend the unfixed cells in the dye
solution. This method was used by Fulmer and Buchanan to meas-
ure the death rate of yeasts acted upon by disinfectants; they
pointed out that the ability of the cells to absorb stains occurs
ajter the cells have lost the ability to grow when transferred to a
new medium. Burke has shown that dead spores may be differen-
tiated from living ones by staining with carbol fuchsin, and Koser
and Mills also found that the ability of the spores to absorb the
dye was acquired later than the period when they lost the power
to germinate. Seiffert found that when suspended in Congo red dead
cells of bacteria absorbed the dye, living ones not. But the differ-
entiation by this technique is not so sharp as with the procedure I
have used.

It has generally been assumed that the staining of the dead cells
is due to increased permeability of the membrane to the dyestuff,
and this view would seem to be supported by the observation of
Fraser that dead yeast cells stain much more slowly with the less
diffusible Congo red than with methylene blue. With Fraser's tech-
nique, however, it is found that with neutral red the dead yeast cells
become stained a uniform deep red, and that in the living cells the
volutin granules and vacuoles are also stained. The cell wall of the
living cells must therefore be permeable to the stain, and the failure
of the protoplasm to become stained must be due to a difference
in its adsorption properties from those of the dead cells. As I have
pointed out above, granules are also stainable in the living cells
of bacteria when Congo red is used, so the membrane must also be
permeable to this dyestuff.

The fact that dead, partially autolysed cells may be detected by
the use of Congo red gives us a valuable procedure for measuring the
rate of death and autolysis of bacteria which has been used in con-

ON TECHNIQUE 57

nection with a study of morphologic variations during the death
phase described in Chapter IX.

The statistical part of the work was greatly facilitated by the
use of an adding machine, measurements being recorded on that
instrument as they were observed, so that the total and consequently
the average w^ere immediately available. In the earlier w^ork with
B. megatherium recorded in Chapter IV, 300 cells were measured
from each sample except during the period of great variability when
500 were measured (1,000 in the case of the five and one-half-hour
sample from Culture I) ; 300 cells were also measured from each
sample in the case of the diphtheroid organism described in Chapter
VI; but it was apparent that nothing was gained in the way of
smoother curves by using such large samples, and the number of cells
measured was reduced to 250 in the case of the colon bacillus (Chap-
ter V), and to 200 in the succeeding studies. Cell measurements
were made to the nearest millimeter of the projected image, and this
unit was chosen in most cases for the class intervals of the fre-
quency curves; the class intervals are therefore smaller with the
higher magnifications. The various statistical constants have been
computed by standard methods save the modes and medians. Modes
have been determined simply by inspection of the smoothed frequency
curves, tnedians from the (unsmoothed) cumulative percentile curves.
The frequency curves presented in the succeeding chapters have all
been smothed twice by the method of three-point averages. None
of the other curves have been smoothed save those shown in Figure
25 and Figure 27, which have been fitted by simple inspection.

CHAPTER IV

THE SIZE OF THE CELLS OF BACILLUS MEGATHERIUM

While it has been well known for a long time that in old cul-
tures the form of the cells is different from that which they ex-
hibit in younger cultures, it has been only relatively recently de
monstrated that in very young cultures, that is during the active
growth period, there is also a difference in the cells from those in the
standard twenty-four-hour culture. This was demonstrated by Clark
and Ruehl, who have shown that during the early hours of growth
there occurs a marked increase in size of the cells with most species
of bacteria. They observed this in Bacillus typhosus, B. paratypho-
sus, B. colt, B. influenzcc, B. pyocyaneus, B. pertussis, B. anthracis
B. subtilis, B. vulgatus, B. avisepticus, and the cholera vibrio. The
same thing was observed with Streptococcus pyogenes, but not with
the glanders organism, the diphtheria bacillus, Hoffman's bacillus or
Bacillus xerosis, the last four showing a decrease rather than an in-
crease in the size of the cells. They note that the change in the
size of the cells is in some instances so great "as to render the organ-
ism unrecognizable when viewed by the ordinary standards of the
twenty-four-hour culture." They found that with a large number
of the cultures studied there was some increase in size apparent after
two hours incubation, the maximum size being reached on the aver-
age from four to six hours after inoculation. They state that "the
correlation in time of the occurrence of the long, coarse forms, with
the period when the maximum growth of the culture is taking place,
is obvious. That the cross section of the culture with its very short
average generation time should show a majority of forms dividing
and nearly ready for further^ division, is also obvious." This latter
statement, however, cannot be accepted. If the increase in size is
to be explained merely by the growth required for the cells to attain
a sufficient length for division to occur, that is twice their original
size according to our previous conceptions of the process, then the
maximum size attained should never be more than one and one-half

59

60 MORPHOLOGIC VARIATION

times the normal size, since in any given sample there will be just as
many cells recently divided as cells about to divide. As a matter
of fact, in their own observations, in a good many cases the maximum
size attained was considerably more than this amount, and according
to my own studies may be as much as six times that of the resting
cells. Moreover, if this is the explanation, it should apply equally
well to all bacteria, including the diphtheria group, which, as they
observed, decreased, rather than increased during the maximum growth
period. As will be seen later, there is much evidence that this change
in the size of the cells is correlated with other changes in their mor-
phological characters, and also in their physiology, and that the
real significance of the phenomenon is that the actively growing
cells are fundamentally different from the resting cells, and that
we have in a young culture a special morphological type which dif-
fers just as much from the standard type that we know from obser-
vations of twenty-four-hour cultures, as do the so-called involution
forms which we find in old cultures.

I have studied this phenomenon of increase in size in some detail,
particularly with Bacillus megatherium, chosen because it is the lar-
gest bacterium that can be easily grown in artificial culture media,
such a large organism making measurements of size relatively easy
and more accurate than is the case with smaller bacteria. The strain
used was one isolated from soil, which corresponds with the published
descriptions of Bacillus megatherium in every respect save that it
is non-motile. This quality of motility in the megatherium group
seems to be one which is extremely variable. In addition to its
large size this organism is also characterized by the occurrence of
numerous intracellular granules, rather more refractile than the pro-
toplasm which contains them. They do not give the staining re-
actions of either volutin or fat; in fact they do not stain with any
of the dyes. But in the fact that they disappear during active
growth and reappear when growth slows up they behave like volutin
and possibly serve a similar function.

The general character of the morphologic variations exhibited
by this organism may be seen in Plate 1 , which shows camera lucida
drawings of living cells removed from a culture at various stages of
growth. There was no appreciable variation in the cells in the first

1 1 1 i 1 1

„. 0 10

20

30

40 50 fJ

%> #^

''-'■' _ •■& ..-.•,' -V

^^^r^- ,:. -t---

o c>> %

%*^ ■.:.. ... ® :^v

6> ^.* ^^ "^^=

^ ■' O 0 ' ':^^ ^.

^. ^ .^

•: •• -.. ^ ;v" © '■' ..

, ^ ' ^<; ^^^ t

■••■: .•-■ o ,'■•'''

F'^^ u'-" 0 - ■;'' ®

^'■■^ ^^r-.v"^;® -^^^' •■

Tii* -■■'\,",

."> ..::;■• ■•'• :'.

0

o <S )'. '•■{;,

:}.°o

2hr5

4hrt3.

» * **

• •■■ _ '■{.

y '.

"" . "': ^•

.; p

^.;>.^ • ".-■ v.

ft- .♦ ' •

■■••■■' ■'■-,

^

\ 'r^-: ':'>

/»

< & '■ '' ~^t ""'■'"* '

i3 . . . ^

^ '-•,

;.

'-v^

• •• o O

y

'. • •* J

' )

6 ho

dhn

'»,,

'-''■■

0' .' o

■ * " ■ '. •

o o

c

o

'-^^., \

10 hr.5

l£br.5.

Plate I. Morphologic Variations of B. megatherium at Different Stages

OF Growth.

I I \ \ \ 1

0 10 20 30 40 50^

lehrs

^ r, °

AQ bri

Plate I (continued) Morphologic Variations of B. megatherium at Differ-
ent Stages of Growth.

CELLS OF BACILLUS MEGATHERIUM 61

2 hours, so that the drawing of the two-hour sample serves to in-
dicate the character of the cells used for seeding. These were taken
from a rather old culture on dextrose agar. It will be observed that
three kinds of cells were present in the seeding, namely, spores,
sporangia, and cells of irregular form containing numerous granules.
As growth proceeded these all rapidly beca^me transformed into long
cylindrical cells, practically free of granules. The spores germinated
by simply gradually swelling and assuming elongated form. The
same thing happened to the irregular cells that were present in the
seeding, so that by six hours there were to be found only these long,
sausage-shaped, relatively hyaline cells. As growth proceeded these
long, clear cells became smaller and more and more filled with
granules. The number of spores present in the various samples was
somewhat irregular, probably due to variations in the distribution of
spores throughout the agar slant, as all of these preparations were
taken from the same agar slant. The large number of free spores
present, for instance in the twelve hour sample, were probably un-
germinated spores that had been present in the original seeding, and
there is no evidence of active spore formation again until after thirty-
six hours. By this time it will be noticed that all of the cells are
relatively small; the difference between the thirty-six and the six-
hour sample is, for instance, quite striking. With increasing age,
more and more spores develop within the cells, and the cells become
more and more irregular in form and filled with granules, until when
the culture is a week old we have nothing left but free spores and
bizarre, irregular, granular involution forms. Thus it will be seen
that at all stages there is a continuous and regular metamorphosis
taking place.

For the present we will concern ourselves with the changes which
take place during the early stages of growth, with the development
of the large, clear cells characteristic of the active growth phase.
Variations in the size of the cell have been studied in micro-colonies
by continuous observation, and in a series of cultures on agar slants
and in broth, from samples removed at regular intervals of time dur-
ing growth.

The micro-colonies were prepared by inoculating a loopful of
culture into a tube of molten agar. A loopful of this agar sus-

62 MORPHOLOGIC VARIATION

pension was then placed on a cover glass, which was inverted over
a hanging drop slide and sealed with vaseline to prevent evaporation.
These were incubated in an electric warm stage at 30° C. With the
low-power lens a cell or small group of cells was found. This was then
focused upon with the oil immersion lens, and watched continuously
for some hours, the progress of events being recorded by camera
lucida drawings. Such micro-colony observations are not easily
made, for if the preparation is heavily seeded, so that the cells are
easy to find, then the growth rate is low and the morphologic varia-
tions are slight. If, however, the preparation is lightly seeded, so
that there are only a few cells in the whole preparation, then these
cells are difficult to find, and in such preparations isolated cells fre-
quently failed to grow for some hours. Many such preparations
were made and observed at intervals all day long, without any evi-
dence of growth being seen, and then on the following morning it
was found that a good sized colony had developed. Most successful
results were obtained when small groups of cells, generally in chains,
were found well isolated from any others.

The results of such continuous observations of growing micro-
colonies are presented in Figures 10, 11, and 12. The first developed
from two cells, both of which commenced to grow shortly after the
observation was begun and grew at approximately the same rate
throughout the period of observation. Over three hours was re-
quired for the first cell division to take place. During this period,
however, the cells had greatly elongated, to considerably more than
twice their original length, and the contained granules had gathered
at the poles of the cell, and had perhaps also decreased somewhat
in number. After the first cell division, which occurred approximately
simultaneously in the two cells, the four resulting cells again increased
markedly before they divided, increasing once more to more than
twice the size at the preceding cell division. The second micro-colony
developed from three cells, of which only one grew, the other two
showing no changes in morphology during the period of observation.
The cell which grew showed its first evidence of growth in an in-
crease in length, with, at the same time, a decrease in the number of
granules and an accumalation of these granules at that end of the
cell which was not in contact with its sister cell. Cell division oc-

CELLS OF BACILLUS MEGATHERIUM

63

y.fd

4:00

J: 10
5:30

6:00

6:15

-S^

6:>0

6:4>

7tl5

7:?0

7:45

8:00

8:^$

8:50

7:00

8:45

I 1 I i i I

0 10 20 JO 40 JO/'

Fig. 10. Morphologic Variation in a Growing Microcolony of B. megathi
ium: Microcolony I.

64

MORPHOLOGIC VARIATION

S 7:30

12:00

12:10

12:25

12:^0

12:55

^

— a

1:05

^

2:35

3:05

3:35

1__J \ L_J I

0 10 20 30 W 50U

Fig. 11. Morphologic Variations in a Growing Microcolony of B.
megatherium: Microcolony II.

CELLS OF BACILLUS MEGATHERIUM

65

3:^5

4:1R

5:30

6:00

6:15

6:30

6:U5

.<S>

7:00

7:15

7:30

S:00

0' fo — db — 3^ — ^ — 5ttx

Fig. 12. Morphologic Variations in a Growing Microcolony of B.
megatherium: Microcolony III.
From "Morphologic Variations of Bacteria in the Lag Phase." Reproduced
through the courtesy of the Journal of Infectious Diseases (1926, 38, 54).

66 MORPHOLOGIC VARIATION

curred shortly after this increase in length, and in repeated cell divi-
sions it is observed that the cells become progressively longer before
each division. The new cells formed are free of granules, but the
granules present in the original tip cell remain, and can be seen
throughout the subsequent periods of observation. After some five
hours growth, the number of cells began to increase greatly so that
it was difficult to follow individual cells, but the original tip cell was
still easily distinguished, and repeated divisions of this portion of the
colony were followed for several hours longer.

In studying these developing micro-colonies, two rather interesting
features are observed, other than the changes in size and form of
the cells. One particularly evident in the second micro-colony, is
the apparent tendency for growth to be restricted to the end of a
chain of cells as long as the cells remain in a chain. This is reminis-
cent of the terminal growth of the filaments of mycelium in molds.
It is not clearly distinguished in micro-colonies or in cultures from
solid media, where the chains, durmg the period of active growth
at least, quickly break up; but in preparations from broth cultures
this is very frequently clearly evident; long chains of cells are formed
in which the two terminal cells are found to be much longer than
other cells in the chain. The second feature is the mechanism by
which the chain is broken and the cells are separated. This is par-
ticularly evident in the first two micro-colonies. It is seen that
after a cross wall has formed, the two cells bend at right angles,
so that the cells are finally torn apart by the leverage of their pres-
sure against each other. This mechanism has been previously des-
cribed in spore forming bacteria. It is also noteworthy that the ac-
tively growing cells are free of granules, though those granules which
were present in the original cell at the beginning of growth may be
retained at the tip, which serves to identify it through the subse-
quent periods of observation.

The third micro-colony developed from five cells. It will be ob-
served that for a period of three hours there was practically no evi-
dence of growth. One of the cells developed a cross wall, and one
of the daughter cells so formed proceeded to disintegrate. Another
became deformed, developing a constriction at one end, with an ac-
cumulation of granular material in the constricted portion. At the

CELLS OF BACILLUS MEGATHERIUM 67

end of three hours, however, some of the cells showed evidences of
growth by an increase in size and a decrease in the number of gran-
ules contained in the cells, and at an increasingly more rapid rate
this process continued, the cells becoming longer and freer of gran-
ules; and after they did so they also divided. Three of the cells re-
mained throughout the period of observation without any signs of
growth, actually becoming smaller and more densely packed with
granules.

It is evident from these micro-colonies that two kinds of growth
are occurring — an increase in the size of the cells, and an increase in
the number. A true measure of the rate of growth, then, is obtained
only by considering both, i.e., by obtaining the product of the number
and size of the cells. The relationship of these conditions to each
other is shown graphically in Figure 13, and the values from which
these graphs have been plotted are presented in Tables III, IV, and
V. In order to plot the three quantities, namely the number of cells,
the mean length of the cells, and the total cell length, on the same
scale, they have been converted to index numbers, the index in each
case being the mean value for the entire period of observation; that
is, the absolute number of cells has been converted into a percent;
age of the average number of cells, and the same process has been
carried through with the other tw^o values. In the first colony, where
the two cells both grew simultaneously, the curve for the average
length of the cells is interesting. It will be noted first of all that
there was a definite lag in this curve, active increase not being ap-
parent until after 180 minutes; that the cells showed no change in
size for a short period preceding each cell division; that the size
of the cells after cell division was greater than the size at the be-
ginning of observation, indicating that the increase in size is definitely
more than twice; and that the size attained after the first cell di-
vision was greater than the size just previous to the first cell division.
There is, then, a sort of rhythmical character to the curve for cell
size. The cells increase to a certain length, divide, then grow again to
reach a still higher length, divide, and so on. The increase in size
of course precedes the first cell division, so that if we plot the ab-
solute rate of growth in terms of total cell length, the lag is very
slightly shorter than the lag in the curve for numbers of cells, but

68

MORPHOLOGIC VARIATION

320

2 2U0

^160 -

go

0

1 I I I ' I I n I I I i I I i i I I I I I I I 1 I I I I I 1/

ii

Number of Cells
Average Length of Cells
Total Cell Length

V /

» 'M I I I i I I i I I I I I I I I I I I I I I I I I I I I

60

120 180
Minutes

Micro-colony I

2^10

300

£160

I I I I I I M I I I I I I I I I I I M M I I I I I I I I I

'«' "I I I I I I I I I I I I I I I I I I I I I I I I I M

60

120 ISO

Minutes

2kQ

300

Micro-colony II.
^^ M I I I I I I I I I I I I I I I I I I I M

ol I I II I I I I I I I II I I M I M

0 60 120 150

Minutes
Micro-colony III.

Fig. 13. Growth Curves and Variations in Cell Size in Growing Micro-
colonies OF B. megatherium.
From "Morphologic Variations of Bacteria in the Lag Phase." Reproduced
through the courtesy of the Journal of Infectious Diseases (1926, 38, 54).

CELLS OF BACILLUS MEGATHERIUM 69

not enough so to be of great significance. In the second colony only
one of the inoculated cells grew; the others actually decreased some-
what in size. We have again a long lag period in the curve for cell
size, during which they actually decreased somewhat. In the third
micro-colony the changes in cell size are not nearly so pronounced.
It will be remembered that this colony started from five cells, of
which three grew, and as will be seen later, the tendency is for the
change in cell size to be less with increased seedings.

From these micro-colonies we may draw the following tentative
conclusions. The increase in number of cells is accompanied by an
increase in size. This is apparently more marked with the small
seeding (first colony, starting from two cells) than with the larger
seeding (third colony, developing from five cells). The lag period
before the beginning of cell multiplication is not entirely occupied by
a transformation of the mature cells to large growing cells. There
is also a lag phase in the curves for cell size. Not all of the cells
respond equally to the new medium, some remaining dormant for
considerable periods of time. There does, therefore, occur a selec-
tion of a rapidly growing strain in some cases. The cells which remain
dormant may continue to change in morphologic characters as they
would have if they had remained in the old medium, becoming smaller
in size, distorted in form, or dying and disintegrating.

Further studies have been made in a culture grown on agar
slants (designated Culture I in the tables), following the technique
described in Chapter III. The agar slant tubes were inoculated
from an agar culture which had been transferred every twelve hours
for several days to get rid of spores, since it was desirable to study
the morphologic variations during the early stages of growth uncom-
plicated by spore germination. In all of these studies of B. megather-
ium the cultures used for seeding were free from spores as far as
could be determined by microscopic examination. Samples were
removed every half hour for the first twelve hours of growth, every
hour for the next twelve hours, every four hours for the second
twenty-four-hour period, and again at sixty and seventy-two hours.
Cell counts were made beginning with the first half-hour period,
and are presented in Table VI. The number of free spores was
observed aftd bjjore formation commenced, and is recorded in the same

70

MORPHOLOGIC VARIATION

table. Not a sufficient number of cells was obtained before the
first two and one-half hours to make quantitative studies of the size
of the cells, but from then on such observations of morphologic

o

\ Average Length of Cells

10

12

Ik 16
Hours.

Fig. 14. Growth Curve and Variations in Size of Cells o? B. megatherium:

Culture I.

characters were made. The results were subjected to statistical analy-
sis, and are presented in Table VIL In Figure 14 the rate of growth,
as indicated by the logarithms of the number of cells, is plotted

CELLS OF BACILLUS MEGATHERIUM 71

for the first sixteen hours, at which time growth had practically
ceased. It will be noticed that there was a lag period of about four
hours, after which time a maximum growth rate was quickly attained,
which continued for another hour, and then the growth rate began
to slow up, and there was a gradual decrease until the end of the
growth period.

In this same graph the mean length of the cells is plotted. It will
be seen that this curve also presents a latent period, which is, how-
ever, shorter, ending at two and one-half hours. Then the
cells rapidly increase in length, and the increase is surprisingly
great, the maximum size reached at five and one-half hours being
approximately six times that of the original cells introduced into the
medium. Following this increase the cells as rapidly decrease again,
and at the end of nine hours they have practically reached the orig-
inal size. This period corresponds fairly closely with the point of
inflection in the growth curve between the period of active growth
and the resting period. From then on the size of the cells shows
slight fluctuations, but is relatively constant throughout the remain-
ing period of observation. x\lthough both curves show little fluctua-
tions due to experimental error, so that it is impossible to use the
data without smoothing for purposes of studying correlation, it is
quite evident from " the curves that there is a definite correlation
between the size of the cells and the rate of growth. The size of the
cells increases as long as the growth rate is increasing. It reaches
a maximum at a point corresponding to the point where the rate of
growth changes from positive to negative acceleration, and returns
to a minimum at a point corresponding with the point of inflection
between the phase of growth and the resting phase.

It will be noted that while the curve for cell size shows a lag,
this lag period is shorter than the lag period in the curve for cell
numbers, and that, therefore, a change in cell size precedes actual
cell division, but that the lag period is not entirely occupied by this
change in cell size. Since the cells are increasing both in size and in
number, a true measure of growth is obtained only by plotting the
total cell length (assuming, as will be shown later is nearly the case,
that the cells do not change much in diameter, and that the length
of the cells is therefore a true measure of their size). Since in the

72

MORPHOLOGIC VARIATION

later phases of growth, the increase in number of cells is much
greater than the increase in size of cells, such a curve of total cell
length does not differ much from the curve of the number of cells,
except during the early hours of growth. In Figure 15 the number

III/;

.no. of cells //
- total cell length / ,'

•
CO
M
<D

12;

A

J /

^^ 0

1 1 1

0 12 3^5

Hours.

Fig. is. Comparison of Growth Curves by Cell Counts and by Total

Cell Length in the Lag Phase. B. megatherium, Culture I.

of cells and the total cell length are plotted side by side. It will
be seen that the latter curve presents a somewhat shorter latent period
and a more gradual acceleration than the former.

A number of interesting facts are brought out by studying the
frequency distribution curves of the cells with regard to their size

CELLS OF BACILLUS MEGATHERIUM

73

5 4 ^S'
Abscissae: 1 unit^U*, Ordinates: 1 unit:=2%

Fig. 16. Frequency Distributions of Cells According to Length at
Different St.'^ges of Growth. B. megatherium, Culture I.

From "Morphologic Variations of Bacteria in the Lag Phase." Reproduced
through the courtesy of the Journal of Infectious Diseases (1926, 38, 54).

74 MORPHOLOGIC VARIATION

at various periods of growth. These frequency distributions for the
first twelve hours are presented in Figure 16. The material used for
seeding (zero hour) presented a practically symmetrical distribution,
with a relatively narrow range and a high mode. As growth pro-
gressed, the form of this distribution curve shows progressive changes.
At the end of two and one-half hours, the range is distinctly increased,
the mode is lower, and the curve has become somewhat skewed.
This increase in range, decrease in height of the mode, and increase
in skewness becomes continually greater in the succeeding curves.
At the end of four hours there is an apparent division, appearing as
though a new mode were being established. This is more evident
in the four and one-half hour curve, and in the following hour,
when the change in size was greatest of the entire period of observa-
tion, this new mode becomes higher; the original mode practically
disappears; and we now have established a somewhat irregular curve
with a long base line and a low mode. In the sixth hour curve the
last remnants of the original mode are seen for the last time. From
then on the curve, though irregular, draws in its base line, increases
its height, and at the end of eight hours is again approaching the
form of the original curve. This was finally reached at twelve
hours, and from then until the end of the period of observation
the frequency distribution curves show no significant variations.

It is clear, then, that the increase in size of cells is also accompan-
ied by an increased variation in size, which is quite evident by ob-
serving the standard deviations presented in Table VII. This seems
to be a general biological phenomenon which has been observed, for
instance, from the study of the growth of humans; the coefficients
of variation are higher in growing children than in adults.
The skewness and the apparent bimodality of the curves during the
period of increasing size, that is during the period of accelerating
growth, are of particular significance. They indicate that not all
of the inoculated cells commence to increase in size at once, for if
they did so, and all increased proportionately, the curve would sim-
ply move in the direction of greater size without changing in sym-
metry, but if at first a few, and then increasingly greater numbers
of cells commenced to grow and elongate, there would be produced
just such a series of changes in distribution as has been described,

CELLS OF BACILLUS MEGATHERIUM

75

namely increased range and dispersion and skewness. But as long
as this process is continuous, while there would be produced a grad-
ual upset of the whole curve with the final development of a new

O

L

6 S

Hours.

10

12

lU 16

Fig. 17. Influence of Size of Seeding on Growth Curves of B. mega-
therium: Cultures II and III.

76

MORPHOLOGIC VARIATION

mode, there would not occur distinct bimodality such as is apparent
here. If, however, out of the inoculated cells, one group started to
grow, while the others remained dormant, bimodality, with a new
mode growing at the expense of the old one, would result.

Further studies were made in broth cultures, and these were run
in parallel series with varying sizes of seeding, in order to deter-

iOn

1 T

Culture II

Culture III

J L

5
Hours.

10

Fig.

18. Influence of Size of Seeding on Length of Cells of B. mega-
therium: Cultures II and III.

mine more clearly the influence of the rate of growth upon the size
of the cells. The first series, marked Cultures II and III, were
inoculated simultaneously from a seven-hour agar slant culture, again
after transplanting every twelve hours for several days to get rid
of spores. Culture III being seeded with one-tenth as many cells as
Culture II. Since this culture started from a very young one, no
pronounced lag phase was noted. The experiment was therefore
repeated in Cultures IV, V, and VI, using a twenty-hour broth

CELLS OF BACILLUS MEGATHERIUM 77

culture for inoculation, hoping thus to get a lag period uncomplicated
by spore germination. This hope, however, was not fulfilled, lag
being absent. Culture V received one-tenth, and Culture VI one-
hundredth as many cells as Culture IV.

From these broth cultures it is quite apparent that the changes
in cell size are not nearly so extensive as those observed in the
preceding agar culture; but also the rate of growth is much lower.
In the agar culture (Culture I) the mean generation time for the
first twelve hours was sixty-eight minutes; in Culture II, it was
317 minutes; in Culture III, 153 minutes; Culture IV, 223 minutes;
Culture V, 111 minutes; Culture VI, 78 minutes. The maximum
size attained in these cultures was as follows: Culture I, 19.8[/.; Cul-
ture II, 8.5; Culture III, 9.7; Culture IV, 7.5; Culture V, 8.0;
Culture VI, 8.6. It is apparent, then, that the size attained depends
upon the actual rate of growth in the culture, the shorter the genera-
tion time the greater being the length of the cells; but since these
experiments were not all run under exactly the same conditions,
this correlation is of course not a perfect one. The relationship of
the size of the cells to the rate of growth is more clearly apparent by
consulting the graphs in Figures 17 and 18 (the data for which may
be found in Tables VIII, IX and X), which show the growth curves
and the curves for the size of cells in Cultures II and III. The
more heavily seeded culture (II) shows a lower growth rate, as
indicated by the slope of the curve during the logarithmic growth
phase, than Culture III, and correspondingly the curve for cell size
reaches a lower maximum and declines more rapidly than in the
culture (III), which was seeded with one-tenth as many cells. The
same thing was found to be true in Cultures IV, V and VI (data
for which will be found in Tables XI, XII, XIII, XIV), the curves
for cell size being somewhat irregular, but again the maximum was
higher in Culture VI, which was most lightly seeded, and this size
was maintained for a longer period of time.

Corresponding with the difference in actual size of the cells in
these broth cultures, there was found a distinct difference in the de-
gree of variation in size. For Cultures II and III the frequency
curves are presented in Figures 19 and 20. It is apparent at once
that in neither culture were the changes in the form of the distribu-

78

MORPHOLOGIC VARIATION

tion curve as pronounced as was found to be the case in the agar
culture, which had a much higher growth rate. In Culture II there

5 hrs.

1 1

1 M I 1 1

-

^^^

-

-

^^^k

.

J-L

j^^H

S hrs.
I M I LI I I I M

3 hrs.

1 1 1

1 1 1 1 1 1 1 1 1 1

_

^^^^^^

—

'\ \M

lidH^^ki

10 hrs.

LI M LI I I I I I I.

h hxs. 12 hrs.

Abscissae: 1 unit^lfx Ordinates: 1 unit=2%

Fig. 19. Frequency Distributions of Cells According to Length at
Different Stages of Growth, B. megatherium, Culture II.

is very little change other than skewness and a shifting of the mode.
In Culture III the skewness, shifting of the mode, and extension
of the base are still more apparent, and in the three hour sample,

CELLS OF BACILLUS MEGATHERIUM

79

approximately the period of maximum growth rate, there is some
tendency to bimodality. The same things were observed in Cultures
IV, V, and \T; with the decreasing sizes of seeding, that is with

I I I I M I I I I IJ M I I I I I I I I I ' I I I I 1 I ' I I I I I i_

I I I I

D hrs.
I I M I I I I I I I I I I I I

2 hxs.

T"« 1 1

1 1 1 1 1 1 1 1 1

—

_

,^^H^^,

-

—

,^^^^^^^^^^^

-

J J_-LJ

^^^^m^

^ • "

3 hrs.

j-r 1 1

T 1 M 1 1 1 M 1 1 1 1

1 1 1 1 1 1 l_

_

-

_

-

~M IJi

10 hrs.
I I I I I I I I ' I

k hrs. 12 hrs.

Abscissae: 1 unit^lM- Ordinates: 1 unit=2%

Fig. 20. Frequency Distributions of Cells According to Length at
Different Stages of Growth, B. megatherium: Culture III.

increasing growth rates, there was greater change in the form of the
distribution curves, and only in the last of these three was there any
evidence of bimodality in the curves.

From these quantitative studies in mass cultures, the same con-

80 MORPHOLOGIC VARIATION

elusions can be drawn as from the micro-colonies, though perhaps
the evidence is more clear, namely that the period of maximum
growth is accompanied by an increase in size of the cells, and that
this is correlated with the rate of growth, being greater with lightly
seeded cultures than with more heavily seeded ones. The maximum
size corresponds with the moment of maximum growth rate. There
is a lag period in the curve for cell size, and the lag period in the
curve for cell numbers is therefore not entirely occupied by the
transformation of the small cells to large, growing ones. But the
cells do increase in size before they begin to divide, and if we con-
sider growth both from the standpoint of cell size and cell numbers,
that is the total growth in the mass of proptoplasm, then the
lag phase is correspondingly shortened.

Not all of the inoculated cells grow at the same rate; first a few
and then more and more cells commence to grow, manifesting their
growth by increase in size; and by studying the frequency distribu-
tion of the cells according to their size, we can demonstrate this
variation in growth potentiality of the cells inoculated into the
medium. This degree of variation is greater with smaller seedings,
with very small seedings there being a pronounced tendency for bimo-
dality to develop in the distribution curves, indicating that under such
circumstances there does occur, as was postulated by Tenfold, a selec-
tion of a rapidly growing strain from the inoculum. That this de-
pends upon the size of the seeding will be apparent after a moment's
thought. If we inoculate a culture with two cells, there are two pos-
sibilities: First, that both of these will grow at the same rate; second,
that they will grow at different rates. Because of the tendency to
variation inherent in living organisms, the chances are in favor of the
second possibility. If we have three cells, then the chances are
that there will be three different growth rates, and by steadily in-
creasing the number of cells we increase the possibility of having
a perfectly graded series of cells with regard to their growth rate
when transferred to a new medium. In a culture thus inoculated
with a large number of cells those cells which have the maximum
growth potentiality will of course start to grow soonest, but they
will be followed quickly by the cells having the next highest growth
rate, and so on; and we will thus have a perfectly graded series

CELLS OF BACILLUS MEGATHERIUM

81

of cells starting to grow one after the other, and so producing a
skewed but continuous frequency distribution curve. If, however,
our seeding is small, the chances will be greater for there occurring
gaps between the cells having the maximum growth rate and those

yu

I r T T 1 1 T'

S

1 A \ '

A ^ \

7

/ \ \

—-

6

/'/ \ ^ \

5

k

>

/' w \

'■s,"

3

—

Quarter Strength Agar

n

1 1 1 1 1 1 1

0123 k 3 ^ 78

Hours.

Fig. 21. Influence of Concentration of Nutrients on Length of Cells
OF B. megatherium.
From "The Influence of the Concentration of Nutrients on the Size of the
Cells of B. megatherium." Through the courtesy of the Society for Experi-
mental Biology and Medicine (Proceedings, 1921, 19, 132).

cells which do not grow at all, and the smaller the size of the seed-
ing, therefore, the greater will be the tendency for the frequency
distribution curves to become broken into two or more modes.

The relationship of the size of the cells to the rate of growth
has been further studied in cultures inoculated with the same number

82 MORPHOLOGIC VARIATION

of cells but into media varying in the concentration of foodstuff
present. Ordinary nutrient broth was diluted to one-half, to one-
fourth, to one-eighth, and to one-sixteenth of its original concen-
tration, and these batches of broth were made up into agar slants
as described in Chapter III. All of the cultures were then simul-
taneously inoculated from the same suspension of cells of Bacillus
megatherium. Cell counts were not made from these cultures, so
that it is not possible to determine the actual rates of growth, but
from the data presented in Chapter II we know that the rate of
growth must have been proportionately higher with increasing con-
centrations of nutrient substances in the medium. The results
of the measurements of cell length are presented in Table XV, and
curves for the first three dilutions, full-strength, half-strength, and
quarter-strength medium, are presented in Figure 21. It will be
noted from Table XV that the results were not perfectly consistent,
the eighth-strength medium reaching a slightly higher maximum than
the quarter-strength; but that in general the maximum size attained
was directly proportional to the concentration of foodstuff in the
medium. Since observations were made only hourly, it is quite
evident that the peaks of the curve for full-strength and quarter-
strength agar would probably have been higher if observations had
been made more frequently, and this fact probably explains the
apparent inconsistency in the maximums which can be observed in
Table XV.

Frequency distribution curves for three of these cultures are
shown in Figure 22. They show in general the same characters
as have been previously described; with increase in the rate of
growth dependent upon the concentration of nutrient in the medium
there is a greater degree of variation in the size of the cells, greater
extension of the base line, and flattening of the curve. Only in
the full-strength medium is there a tendency to split into two groups
distributed about separate modes. This is somewhat difficult to
explain if the reasons given above for bimodality are accepted, but
it might be suggested that an increase in strength of the selecting
agent (nutrients) might as well increase the tendency to bimodality
as a decrease in the number of inoculated cells. At any rate it is
clear that increasing growth rate from whatever cause is accompanied

CELLS OF BACILLUS MEGATHERIUM 83

Full Quajter Sixteenth

Strength Strength Strength

.^L. 1 hr. 1 hr.

^^^ 2 hrs. 2 hrs.

L^^^J 3 hrs 3Jirs. ,

E E

5 hrs.

^ E E

^ hrs.

E

C E

g hrs. Shrs.

3 hrs.

Abscissa: 1 unit=l^i Ordinates: 1 unitr=:2%

Fig. 22. Influence of Concentration of Nutrients on Frequency Dis-
tribution OF Cells According to Length, B. megatherium.

84

MORPHOLOGIC VARIATION

0 2 4 6
Hours,

Parent
Culture.

2 H S
Hours,

Two hour
Subculture.

Four hour
Subculture.

Six hour
Subculture,

Hours.

Eight hour
Subculture

Fig. 23. Influence of Age of Parent Culture on Size of Cells,
B. megatherium.

CELLS OF BACILLUS MEGATHERIUM 85

by increased variation in size as well as by increased size of the
cells.

It will be remembered that Chesney demonstrated that the rate
of growth is also dependent upon the age of the cultures used for
inoculation, and this factor has also been studied with regard to
the size of the cells. A series of agar slant tubes prepared as de-
scribed in Chapter III were inoculated wdth a culture of Bacillus
megatherium; samples were removed every hour and measures of
cell length made. After tw^o, four, six, and eight hours, subcultures
were made onto other agar slant tubes prepared in the same way.
An attempt was made to keep the size of the seeding constant
by preparing suspensions of as nearly as possible the same degree
of turbidity, as determined by inspection. The results are presented
in Figure 23, which shows the mean length of the cells plotted
against the age of the culture. It will be seen that the parent culture
showed in general the same type of curve as has previously been
described. There was a slight latent period, after which the cells
increased in size, reached a maximum, decreased rapidly and then
more slowly until they reached the original size. In the culture
transplanted from the parent culture when it w'as tW'O hours old,
the cells continued to increase in size at the same rate as in the
parent culture, but reached a somewhat higher maximum and main-
tained their large size for a somewhat longer period of time. In
the culture transplanted from the parent culture when it was four
hours old, the cells showed a somewhat further increase in size,
but at a somewhat lower rate, and the increase in size of the four-
hour subculture over the two-hour subculture was less than the
increase of the two-hour subculture over the parent culture. By
continually subculturing during the maximum growth rate, then,
we can continue to increase the size of the cells, but this increase
is not proportional, the amount of increase being smaller with each
succeeding subculture, so that there is a definite limit to the size
attainable. In the culture transplanted from the parent culture at
the end of six hours, the cells immediately started to increase again,
and reached a somewhat higher maximum than in the parent culture;
but when transplanted at the end of eight hours, that is at a time
when the cells in the parent culture had returned to their minimum

86 MORPHOLOGIC VARIATION

once more, the cells in the new culture again showed a lag before
they commenced to increase, and reached practically the same maxi-
mum as in the parent culture, the curve having practically the same
form as that of the culture from which the seeding was taken.

Here, then, is further evidence that the size of the cells is cor-
related with the rate of growth. When a culture is transplanted
at a period during which the cells are increasing in size, they continue
to increase in size in the new medium; just as Chesney found when
a culture is transplanted during a period of maximum growth rate,
it continues to grow at a maximum growth rate in the new medium.
When the cells are transplanted during the period of decreasing
size, they immediately recover and increase again in the new medium,
but if transplanted at the time when they had reached their minimum
size, corresponding with the resting period or period of no growth,
they show a lag in the new medium, just as Chesney found there
was a lag in the curve for rate of growth.

Frequency distribution curves of these cultures also support the
previous observation that the degree of variation in cell size is cor-
related with the actual size attained, that is with the actual rate
of growth in the culture. The parent culture showed the extension
of the base line, and flattening of the peak during the period of
maximum cell size, as was observed in the previous cultures. In
the two-and four-hour subcultures, this extension and flattening was
increasingly greater, but in the eight-hour subculture the frequency
curves were practically identical with those in the parent culture.

CHAPTER V

THE SIZE AND FORM OF THE CELLS OF
THE COLON BACILLUS

In the studies of Bacillus megatherium immediately preceding,
the only measure of size considered was the length of the cells.
There were several reasons for this, the most important being that
the diameter of the cells is so small that by the technique used it
was impossible to make measurements of the thickness at all com-
parable in accuracy with the measurements of length. It was quite
apparent, however, that while there was a degree of correlation
between the length and diameter of the cells, the longer cells being
somewhat thicker, nevertheless the increase in thickness with in-
creasing length was by no means proportional; that is, the longer
cells were relatively much more slender, and it was assumed, there-
fore, that measurements of length alone would be sufficient to indi-
cate the character of the variations in size which were taking place.
It was found desirable, however, to make some observations on this
change in the relative proportions between length and thickness of
the cells, that is to obtain some sort of a quantitative expression
of the variations in the form of the cells which were taking place.

This was done with a culture of Bacillus coli, using photomicro-
graphs projected so as to give a final magnification of 30,000 X- From
the tracings of the projected cells the area was measured, as well as
the length along the major axis of the cell and from these
two measurements the quantitative expression of the form of
the cell was obtained by dividing the area of the projected image by
the length squared. The quantity so obtained, which I have desig-
nated the area-length index, is an abstract measure of form, regard-
less of the size of the cell. It will have its highest value if the cell
from which the tracing is made is a sphere, that is the image being
circular. As the length increases in proportion to the width, this value
becomes progressively smaller.

Such a measure suffices to express accurately the form of the

87

88 MORPHOLOGIC VARIATION

cell as long as this is symmetrical about both axes. It will be at
once apparent, however, that this does not serve if the cell is at
all asymmetrical. If, for instance, we have bulgings or constrictions
or lateral projections, such unusual cell forms might have the same
area-length index as a cell which was quite uniform throughout.
With the early stages of growth, however, such asymmetrical cells
are exceedingly rare, and therefore for this stage the area-length
index serves very well to indicate the variations in form of the
cells. By the use of this measure, together with the length of the

7

2X

193

100

50

l|lllM

Logajrithms of Numbers of Cells

— Index Kuiabers of Area-Length Indices.

Index Numbers of Average Lengths of pells.

Index Numbers or Average Lengths ot pe

0 3 w 9 12 18 30 ^ 72 9d

Hours.
Fig. 24. Growth Curve and Morphologic Variations of a Culture of

B. coli.
From "A Statistical Study of the Form and Growth of Bacterium coli." Re-
produced through the courtesy of the Society for Experimental Biology and
Medicine (Proceedings, 1923, 21, 215).

cells, we can therefore follow consecutive changes both in the size
and the form of the cells, and correlate these with each other, as
well as with the rate of growth.

The organism used was an old laboratory strain, a typical colon
bacillus. The culture was inoculated onto agar slants, following the
technique described in Chapter III. Samples were removed for
counting and measurement every three hours for the first twelve
hours, and again at eighteen, thirty, forty-eight, seventy-two, and
ninety-six hours.

:ells of the colon bacillus

89

Figure 24 shows in a general way the results of these observa-
tions. In this graph the logarithms of the number of cells indicate
the rate of growth ; the average length of the cells and the average area-
length index having been converted to index numbers, using as the
index the mean value of each for the entire period of observation.
The data are given in Tables XVII and XVIIL As with Bacillus
megatherium, there was noted a very marked increase in the length
of the cells during the period of active growth, reaching its maximum

Fig. 25.

0 3 b 9 12

Hours.

Representative Cells from a Culture of B. coli at Different
Stages of Growth.

probably again at the point where the growth curve changes from
positive to negative acceleration. It will be noted, however, that
the curve for area-length index during the early hours of growth
is practically the reverse of the curve for length of cells; that is,
as the cells increase in length they decrease in area-length index,
the longer cells being relatively more slender, even though they were
absolutely somewhat thicker. In this graph the average size and
form of the cells is indicated by tracings through the major image
of composite photographs of the cells prepared as described in

90

MORPHOLOGIC VARIATION

Chapter III. They illustrate well the increase in size during the
period of active growth. After growth has come to a standstill and
the cells have returned to their original size, subsequent variations
are slight and probably of no significance.

This negative correlation between the length and area-length
index of the cells will be perhaps more apparent by an inspection
of Figure 25, in which every tenth cell in the array of 200 from
each sample has been superimposed in a column upon the growth

3 hrs.

CECCJ

6 hrs. 30 hrs.

T — T — r

9 hrs. 96 hrs.

Frenuency Distributions: Length of Cells.
Abscissas: 1 unit=lu. Ordinates: 1 unlt=2^

_ 1 1 U 1 1 1 L

. 1 1 LI L

0 hr.

IS hrs.

ii

_l 111 1 _

3 hrs.

30 hrs.

||l 1 1 1 1 1 l_

rii 1 1 II 1 r

_l 1^ 1 1 1 _

6 hrs.

Frequency Dlst

Abscissae: 1

rll
w

Dutlon
iit=10

96 hrs.
s:Area-Len£th Indices
, Ordinates: 1 u.-ilt=2X

Fig.

26. Frequency Distributions of Cells According to Length and
Area-length Index. B. coli.

curve plotted logarithmically with regard to time, as well as to the
number of cells. (Since cell counts were made only every three
hours, this curve has been drawn in a rather free-hand fashion,
but is probably fairly close to the true growth curve.) Again the
marked increase in size during the period of active growth is appar-
ent. The columns of cells have been arranged in the order of their
area-length index, not in the order of their length, but it will be
seen that as a matter of fact this method does roughly classify
them according to their size. The longest cells in each column are

CELLS OF THE COLON BACILLUS

91

the cells with the lowest area-length index, that is the most slender,
and the cells of greatest area-length index, at the top of the columns
(i.e. those approaching spherical form), are also for the most part
the shortest.

I I I I I I I I I I I

I I I I I I

I I I I

I I I I

I I I I I ' I I 1

Fig. 27. Correlation Graph of Length of Cells Against Area-length

Index, B. coli.
Ordinates, 1 unit = 3\i (Length). Abscissae, 1 unit = 3 (Area-length Index).
Solid dots represent cells from eighteen-hour sample, hollow dots cells from
three-hour sample.

Frequency distribution curves for the length of the cells are
presented in Figure 26, together with curves for the distribution of

92 MORPHOLOGIC VARIATION

the cells according to their area-length indices. The curves for
length show the increased variation in length during the period of
active growth which was previously observed with Bacillus megath-
erium. There is, however, not nearly so pronounced a skewness
in the curves and no tendency toward bimodality. The culture
studied, however, w^as relatively heavily seeded, and it is probable that
if further studies were made with this organism, with lightly seeded
cultures, as has been done in the case of Bacillus megatherium, simi-
lar increased variation and a tendency toward bimodality in the
distribution curves would be noted. The frequency distribution
curves for area-length index of the cells show nothing very striking,
save that during the period of active growth (at three hours) the
range was somewhat restricted and the variation in form is some-
what less. The negative correlation between size and area-length
index, then, is also evident in the distribution curves.

The relation between size and form of the cells may be more
clearly seen in a correlation graph (Figure 27) in which the length
of the cells is plotted against the area-length index. For this graph
the data from the three-hour sample, as representative of the maxi-
mum growth phase; and from the eighteen-hour sample, as repre-
sentative of the resting phase, have been chosen. A curve has
been very roughly fitted to the data by simple inspection. This
curve is clearly hyperbolic, the ascending limb approaching as a
limit an area-length index of zero, i.e., a form in which the thick-
ness is infinitely small in proportion to the length; the descending
limb approaches as a limit a form in which the length equals the
thickness, i.e., a sphere. Such a hyperbolic curve would result if the
cells varied in length but remained constant in diameter. That
such is not the case is apparent from an inspection of Figure 24;
the three-hour cells are distinctly thicker than the others. But
proportionately to length, the increase in diameter is not great, and
consequently a hyperbolic curve results when length is plotted against
form. It is interesting in this connection to note that Minoux has
recently concluded from theoretical-mathematical considerations that
bacterial cells must grow much less in diameter than they do in
length.

CELLS OF THE COLON BACILLUS 93

Sufficient information has now been presented to warrant some
speculation regarding the significance of these variations in the size
and form of the cells of bacteria during the period of active growth.
While at first glance the simple explanation of Clark and Ruehl
that these variations are merely due to the growth of the cells pre-
paratory to cell division would seem sufficient, it should be clear
by now that such an explanation is quite inadequate. I have previ-
ously mentioned my belief that these changes in size, considered
together with certain other facts, are an indication that during the
period of active growth we have, with bacteria, to deal with a special
cell type characteristic of this growth phase, different in internal
constitution as well as in external form from the resting cells with
which we are mostly familiar. These young cells are therefore
in a sense analogous to the embryonic cells of a growing multicellular
organism. Now if we look upon them as having actually the signifi-
cance of such embryonic cells, there is opened up an attractive field
of speculation from the standpoint of phylogeny, for such embryonic
cells should tend to revert to an ancestral type. The fact that with
most species of bacteria these embryonic cells are long and slender,
tending towards a filamentous form, suggests the possibility that
these are attempting to form mycelium ; that the bacteria are actually
descendents of higher fungi. Such an hypothesis is, however, pure
speculation, hardly warranted by the data at hand.

More profit would seem to accrue from a consideration of the
data from a physical standpoint. Since the size of the cells con-
tinually increases as long as the growth rate is accelerating, the
synthesis of protoplasm during this phase proceeds more rapidly
than does cell division; and conversely, since the cells become in-
creasingly smaller when the growth rate becomes slower, cell division
must in this period proceed more rapidly than does the synthesis
of protoplasm. Since increase in size is accompanied by increased
relative slenderness, i.e., by a greater tendency towards the cylindrical
or filamentous form, and the subsequent decrease is accompanied
by a greater and greater tendency towards the spherical form, it
follows that during the period of accelerating growth there occurs
an increased amount of surface in proportion to the volume of
protoplasm for each cell, and that during the phase of negative

94 MORPHOLOGIC VARIATION

acceleration the surface-volume ratio is decreased. It should be
pointed out, however, that while this is true as long as individual
cells are concerned, the reverse condition obtains when we consider
the total amount of protoplasm, a greater surface being presented
by a mass divided into small particles even if these be spherical,
than by the same mass divided into larger particles, even though
these are filamentous. These considerations indicate that we may
with advantage consider the variations of the cells both in size and
form from the standpoint of surface phenomena.

"Among the forces which determine the forms of cells" says
D'Arcy Thompson, "whether they be solitary or arranged in con-
tact with one another, this force of surface tension is certainly of
great, and is probably of paramount importance." With organisms
as small as bacteria, where the surface in proportion to mass of
protoplasm is so greatly increased, this force must be, if possible, of
even greater importance than with larger cells; Thompson explains
in this way the relatively slight variation in form encountered with
bacteria as compared with higher organisms. While it would take
us too far afield (and require a much more extensive knowledge of
the subject than I possess), to discuss in all its details this relation-
ship of surface tension to the size and form of the cells of bacteria,
a word or two on the matter may well be interpolated here. For
a more detailed discussion of the matter, the reader is referred to
Thompson's work.

If bacterial cells were at all times homogeneous masses of proto-
plasm, their cells should at all times be spherical, as with the cocci
in the resting phase. But since with all forms, including the cocci
when actively growing, the cells are elongated, there must be present
some axially disposed force, some polar distribution of substances
(and consequently of energies) to give them this oval or cylindrical
or filamentous form. It is known with higher organisms, and gener-
ally assumed with bacteria, that forms other than the spherical
are maintained by a certain rigidity of the cell wall due to the depo-
sition of solid substances there. But, of course, the cell must have
assumed its form previous to the deposition of this substance; conse-
quently something must have been opposed to the rounding effect of
surface tension previous to the development of rigidity in the wall;

CELLS OF THE COLON BACILLUS 95

i.e., while a rigid cell wall may maintain a form other than the
spherical, it cannot give rise to it. Frobisher suggests that with
bacteria, since they elongate perceptibly without changing in diame-
ter, the rigid cell wall, if it exists, may be open at the ends. The
fact that the ends of cells where they are not in contact with
contiguous cells are always rounded supports this view.

If the elongated form of the cells is due to some internal, axially
disposed force, opposing the rounding effect of surface tension, then
when placed in a medium of lowered surface tension, this force
meeting less opposition should produce a more pronounced effect,
and the cells should become even longer and more slender. This
in fact is what occurs. Frobisher noted that the morphologic varia-
tions induced by adding surface tension depressants to the medium
were comparatively slight, but that with low tensions (35-40 dynes)
the cells became longer and more slender. He observed but slight
variation because he was dealing with cells which had passed the
maximum growth phase. In some as yet uncompleted observations*
on the effect of sodium ricinoleate upon the size and form of the
cells of the colon bacillus, the soap being added in excess to agar,
I have found that whereas after twelve hours or so the cells were
only somewhat longer and more slender than in control cultures,
during the active growth phase (three to five hours) they were
astonishingly elongated and filamentous in form, many cells extend-
ing over several oil immersion fields. On the other hand, when cal-
cium chloride (which tends to raise the surface tension somewhat)
is added to the medium, the cells are somewhat shorter and more
oval in form than in the normal medium.

If it be granted then that the elongation of the cells (with cor-
responding increasing slenderness) is due to the effect of some in-
ternal force opposing the surface tension of the medium, then we
must assume that this force is greater during the period of increasing

* These experiments are uncompleted and are not reported in detail here
because of the technical difficulties involved. It is impossible to determine the
surface tension of agar after it has jelled, and difficult to obtain clear morpho-
logical studies from liquid media. The colon bacillus becomes strongly clumped
in the soap medium so that it is impossible to obtain accurate measures of
growth by any method. The growth forms a slimy mass which does not
stain readily and which is difficult to spread in thin films.

96 MORPHOLOGIC VARIATION

growth rate than during the resting phase, and that it becomes
increasingly greater with increased acceleration of growth; that any
external or internal factor which tends to speed up growth also tends
to increase the magnitude of this force.

Clark and Ruehl noted that the large forms which they found
during the active growth phase were much more intensely stained
by the basic aniline dyes than were the "normal" forms found in
later stages of growth, and I have observed the same thing, the
intensity of staining being very striking. Eisenberg points out that
the Gram-staining of bacteria is variable with the age of the culture,
and that in very young cultures certain of the Gram-negative bac-
teria may retain the dye. Stearn and Stearn have shown that the
reaction of bacteria to acid and basic dyes depends upon the iso-
electric "zone" of the amphoteric protoplasm of which they are com-
posed, the difference between Gram-positive and Gram-negative bac-
teria being due to this difference in their iso-electric points. With
this information at hand we may generalize a bit and conclude that
in our "embryonic" bacteria the iso-electric "zone" is further on the
acid side of neutrality than is the case with the mature form, just
as in very old cultures it is further on the basic side. Gram-positive
organisms becoming Gram-negative, all organisms staining more
faintly with basic dyes, and some cells finally becoming stainable
with acid dyes like Congo red, as was pointed out in Chapter III.

If we accept the diffuse nucleus theory for bacteria, is it not
plausible that this increased affinity for basic dyes is due to an
increased proportion of nucleoproteins distributed through the proto-
plasm? And is this not analogous to the high nucleus-protoplasm
ratio characteristic of the embryonic cells of higher organisms?

Sherman and Albus (1923), as has been previously mentioned,
have found that young, actively growing cells of the colon bacillus
are much more susceptible to various injurious agents (heat, cold,
sodium chloride, phenol) than are older cultures. They also noted
that such young cultures were not susceptible to acid agglutination,
which may be correlated with the low iso-electric point of their
protoplasm as indicated above. They believe that these observa-
tions indicate a physiological difference between growing cells and
resting cells, that there is a "physiological youth" in bacteria analo-

CELLS OF THE COLON BACILLUS 97

gous to the physiological youth which has been demonstrated by
Child in multicellular tissues, as regenerating pieces of planarians.
It is noteworthy that Child also demonstrated this physiological
youth by increased susceptibility to injurious agents.

These three correlated properties: Increased length and slender-
ness of the cells, indicating a greater magnitude of some axially
disposed force opposing the surface tension of the medium; increased
intensity of staining with basic dyes and decreased susceptibility to
acid agglutination, indicating an iso-electric point of the protoplasm
more on the acid side; and increased susceptibility to injurious
agents, all serve to distinguish the young, actively growing cells from
the resting cells, and justify our recognizing these long cells as a
distinct morphologic type, as embryonic cells, characteristic of the
growth phase of the culture.

CHAPTER VI

SOME OBSERVATIONS OF A DIPHTHEROID
BACILLUS

It will be remembered that Clark and Ruehl observed that the
diphtheria and diphtheria-like bacilli showed a decrease in size dur-
ing the early stages of growth, followed by a gradual increase. Al-
bert, on the other hand, claims that the diphtheria bacillus increases
in size during the first four hours of growth, remains stationary
for about twenty hours, and then gradually decreases. This, how-
ever, I cannot confirm, as I find that during the actual growth
period they become smaller. This group therefore deserves some
particular study, but is not well suited for the type of work here
reported because most of the members of the group are very small
and are grown with some difficulty on artificial media. I was for-
tunate, however, in finding a saprophytic organism which seems
to possess the characters of the Corynebacteria, that is quite large
and could also be grown fairly readily. It is a red chromogen and
was isolated from lake water. Since this organism is apparently a
new species it will be described in some detail.

In morphology it resembles the Corynebacteria in the possession
of very prominent metachromatic granules, in pleomorphism which
is very prominent in old cultures, in a tendency toward a paHsade
arrangement of the cells at some stages of growth, and in the fact
that during the early stages of growth the cells are much smaller
than they are later. The length of the cells on ordinary agar may
be found in Table XX. In other media, particularly Loeffler's blood
serum and dextrose agar, they may be much larger. The diameter
does not vary so much as the length. The cells also vary consider-
ably in form. They are generally cylindrical with rounded ends;
when in pairs or chains the opposed ends are flattened. According
to the medium and age of the culture they vary from short oval
cells, sometimes almost spherical, to long filaments. In old cultures,
cells of irregular shape appear. In young cultures, particularly on

99

100 MORPHOLOGIC VARIATION

dextrose agar, the cells are arranged in short chains; later these
break up and the cells are singly distributed or clumped in the
parallel formation characteristic of diphtheria bacilli. It is non-motile.
With Gram's stain the reaction varies with the age of the culture. In
relatively young cultures, under forty-eight hours, the organism is
Gram-positive, but like the diphtheria bacillus not strongly so, being
readily decolorized if the action of the alcohol is somewhat pro-
longed.

The granules within the cells give all the microchemical reactions
of volutin; they disappear after treatment with boiling water, they
stain vitally with weak solutions of methylene blue and of neutral
red, they retain methylene blue after treatment with 1 per cent
sulphuric acid, Bismarck brown, and Gram's iodine, and they stain
violet or reddish with old" solutions of methylene blue and with
toluidine blue. They vary in size, number, shape, and position
within the cell with the age of the culture and the nature of the
medium. In agar slant cultures the granules are very small, but
because of their intense staining quite distinct. In a twenty-four-hour
culture the majority of the cells contain two granules, situated near
the extremities of the cell ; but some cells may contain a considerable
number scattered throughout. During the period of active growth,
however, the granules decrease in number, and after six hours on
agar most of the cells have no granules, and those that do seldom
have more than one, which is centrally located. In the involution
forms which appear in the old agar cultures the metachromatic ma-
terial is variously disposed, sometimes in large irregular masses,
frequently as a long, narrow, twisted filament in the middle of the
cell. In liquid media the granules appear much as on agar, save that
the variations occur much more slowly. On Loeffler's blood serum
the granules are very large, sometimes bulging the cell, and round.
When the granules are quite large, as on this medium, they appear
to stain more intensely at the periphery.

In media containing dextrose the granules more frequently oc-
cupy a position in the middle of the cell, and apparently divide
preceding cell division. This was most marked in a "synthetic"
medium composed of ammonium sulfate, dibasic potassium phos-
phate, magnesium sulphate, and dextrose. Here the cells were for

OBSERVATIONS OF DIPHTHEROID BACILLUS 101

the most part oval in shape, and in a resting state exhibit but one
large granule centrally located. From the various appearances pre-
sented by a series of cells from such a culture, it would appear that,
preceding cell division, the granule first elongates in a direction at
right angles to the axis of the cell, then splits in two along its own
axis, after which the cell divides on a plane between the two gran-
ules, and each of the new cells is thus provided with one granule, at
first situated at one pole, but later moving to the middle of the
cell and again becoming rounded. This apparent participation of the
metrachromatic granules in the process of cell division has been
noted by Ernst and by Williams, who have suggested that the gran-
ules must therefore be of the nature of nuclei. The fact, however,
that the division of the granules preceding cell division occurs only
at times or in certain media, and only in those cells where the gran-
ules happen to be located at the point of division, would argue against
their performing any essential function in the process; and it is not
necessary to assume that they possess either the structure or the
function of nuclei to explain the phenomenon, for it is clear that
if they happen to lie in the plane of fission they will be acted upon
by the forces which cause the cell to divide and must themselves be
divided.

The most striking of the cultural characters was the pigment
formation, which remained quite constant throughout the period
(several years) that the organism was under observation. It was
always a tint of grenadine red, usually grenadine pink (7d, Ridgway's
Color Standards and Nomenclature). The pigment is extracted
from the wet cells by ethyl and methyl alcohols, by acetone, and
by 20 per cent sodium hydroxide, but not by ether, petroleum
ether, toluene, xylene, or 10 per cent hydrochloric acid. From the
sodium hydroxide solution it can be shaken out with chloroform,
and acetone, only slightly with petroleum ether, and not at all
with ether, toluene, or xylene. From the methyl alcohol solution
somewhat diluted with water, it can be shaken out with petroleum
ether, more readily with carbon disulfide. The growth on agar
turns a bluish green when treated with sulphuric acid. The pigment
is a lipo chrome very similar to carotin.

On solid media the growth is smooth and glistening, somewhat

102 MORPHOLOGIC VARIATION

mucoid in consistency. In liquid media there is produced a diffuse
turbidity; after some time a slight pellicle appears. Gelatine is
rapidly liquefied, but egg, blood serum, and casein are not digested.
No indol is formed. None of the common sugars are fermented.
The organism grows best on slightly alkaline media and is sharply
inhibited at pH 6.4. It seemed to grow equally well at temperatures
between 30° and 37°.

Although various bacteriologists, notably Harris and Wade, have
noted the occurrence of saprophytic chromogenic diphtheriod bacilli,
which are common and widely distributed, these organisms have not
been thoroughly studied. I have found in the literature but one
producing a red pigment which has been adequately described and
named. This was reported by Hoag as "Bacillus X", and named
by Morse B. hoagii, which was changed to Corynebacterium hoagii
by Eberson. The organism here described differs markedly from
C. hoagii, however, in the unusually large size of the cells, the
deeper color, the liquefaction of gelatine, and the failure to ferment
dextrose and sucrose. While it grew very readily when first isolated
on all sorts of media, after a few years subcultivation it gradually
gave a scantier growth and finally died. Since the type culture
is therefore not available, I hesitate to name it.

The rate of growth, length of cells, and number of metachromatic
granules in the cells have been determined from a series of agar
slant cultures inoculated from a forty-eight-hour old agar culture,
following the technique described in Chapter III. Measurements
of cell length were made from Congo red slides projected at a mag-
nification of 3,000 X. The granules were counted from slides stained
by Ponder's method with toluidine blue; and for purposes of correla-
tion between number of granules and size of cells the length of
the cells was also determined from these latter slides by means
of a camera lucida and dividers, the magnification being 1,500 X.

The general character of the morphologic variations observed
can be seen in Figure 30, which presents a series of camera lucida
drawings from cultures of different ages. These were made from
cultures on dextrose agar, in which medium the changes in the
appearance of the cells are more pronounced than on ordinary agar
which was used for quantitative study; but the cell variations are

OBSERVATIONS OF DIPHTHEROID BACILLUS 103

of the same kind. The resting cells are large, with a faintly stained
protoplasm and deeply stained granules. When growth commences
they become much shorter, almost spherical in form, arranged in
chains; the granules disappear, and the entire protoplasm becomes
deeply stained.

The results of the various measurements are presented graphically
in Figure 28, and in tabular form in Tables XIX, XX, and XXL
In Figure 28 the logarithms of the numbers of cells are plotted as
a solid line, the average length of the cells as a dotted line, and
the average number of granules per cell as a dot-and-dash line.
The measurements of cell length show very pronounced minor fluc-
tuations which must be due to experimental error, but the major
tendency is clear. During the period of accelerating growth rate
the size shows very little change; if anything, the cells increase
somewhat. But as soon as the growth rate begins to slow up there
occurs a very pronounced decrease in the size of the cells, which
reach their minimum size at about the point of inflection between
the growth phase and the resting phase. The return to the original
size proceeds very slowly; the original size is attained only after
the culture had been in the resting phase for some twenty-four
hours. Since the growth curve was determined by cell counts rather
than by plate counts, it is quite possible that the culture had entered
the death phase before the original size was reached. While measure-
ments were not continued longer than forty-eight hours, simple in-
spection of the slides shows that they continue to elongate, becoming
long filaments after some days.

It will be seen at once that the variations in cell size with this
organism are of quite different nature from those found with B.
megatherium and the colon bacillus. It might be argued that the
slight initial increase in size apparent during the period of accelera-
ting growth is the same sort of a phenomenon as the pronounced in-
crease noted with the other two organisms, and that the pronounced
decrease during the period of negative acceleration is a peculiarity
of the diphtheroid group. In favor of this hypothesis is the fact that
the granules disappear and the protoplasm becomes rather deeply
stained before there occurs any decrease in the size of the cells.
But Clark and Ruehl failed to observe any initial increase in the size

104

MORPHOLOGIC VARIATION
1 \ 1 \

— 0.25

iG

2k
Hours.

32

1^0

Us

0.00

Fig. 28. Growth Curve (solid line), Length of Cells (dotted line), and

Metachromatic Granules per Cell (dot-and-dash line).

Diphtheroid Organism.

OBSERVATIONS OF DIPHTHEROID BACILLUS 105

of the cells of the Corynebacteria which they observed. It is note-
worthy that their curves also show a very slow increase to the original
size of the cells as compared with the very rapid decrease to the
original size observed in the other organisms studied. It is evident
that more work will have to be done with this group before the
changes in cell size can be clearly correlated with the phases of
growth. From the curve here submitted it would seem that the
small sized cells are more characteristic of the resting phase than
of the growth phase, but from evidence to be submitted later it is
suggested that this may be wrong.

That this diphtheriod organism behaves differently from the
other bacteria studied with regard to its variations in cell size is
still more apparent when studied from the standpoint of frequency
distributions, illustrated in Figure 29. No such pronounced varia-
tion is apparent as was the case with B. megatherium and the colon
bacillus; and the increased variation occurs not during the period
of most active growth, but during the resting period. The frequency
curves at six and twelve hours are practically symmetrical; all of
the others show some skewness. It will be noted from the coeffi-
cients of variation in Table XX that, while this value shows con-
siderable fluctuation, it is distinctly lower during the period from
nine to twelve hours, i.e., during the stage of decreasing size of
the cells, a period during which the growth rate is decreasing but
is still relatively high. While the cells are increasing in size again
during the resting phase, the coefficient of variation also increases.
It is clear, then, that variability in size is correlated with the actual
size of the cells, not necessarily with the rate of cell division. This
is notew^orthy because Taliaferro and Taliaferro have used the coeffi-
cient of variation as a measure of the rate of growth in trypanosomes,
on the supposition that during active cell division there will be found
more variation in size because of the presence in the population of
cells which are mature, cells. which have just divided, and all stages
between; whereas when growth ceases, there being no further cell
division,. all of the cells will be much of a size. This of course will
be true only if the cells show no other change in size than that
preparatory to division. However true it may be for trypanosomes,
it will not hold for bacteria. With the diphtheriod organism here

106

MORPHOLOGIC VARIATION

described the minimum coefficient of variation corresponds nearly
with the maximum rate of cell division.

6br5.

Ohr.

[K \A2

15 hrs.

\l hv5. 16 br5.

15 br5.

Abscissae, 1 unit =; 2^.

48 br^.
Ordinates, 1 unit = 5%.

Fig. 29. Frequency Distributions of Cells According to Length at
Different Stages of Growth, Diphtheroid Organism.

It has long been known that the metachromatic granules or
volutin of microorganisms tends to disappear during active growth,
and to reappear again after growth has ceased; in fact the generally
accepted view that these granules represent some sort of reserve

OBSERVATIONS OF DIPHTHEROID BACILLUS 107

food material is based upon this observation. Albert states that in
the case of the diphtheria bacillus:

Granules appear in cultures four to eight hours old. These are at first of
small size. Granules attain their largest average size in cultures twelve to
fifteen hours old. After this period the granules diminish in size and disappear.
The percentage of bacteria which contain granules increases rapidly from the
four-hour period when there are but a few granules to the twelve-hour
period when 91 per cent contain granules. At the end of two days most of
the bacilli have lost their granules.

The curve for the mean number of granules per cell in Figure 28
shows that the granules rapidly disappeared with active growth,
the minimum number being found at nine hours (when there was
only one granule to every ten cells) , which corresponds approximately
with the point of maximum growth. But the granules came back
very slowly, not nearly so rapidly as was noted by Albert with the
diphtheria bacillus. This is probably due to a difference in the
nature of the organisms, though perhaps in part to the fact that
he was using a much more favorable medium and probably much
more heavily seeded than my cultures.

The mere number of granules is not a very good measure of
the amount of volutin in the cells, since the granules also vary
greatly in size, in general being smallest when they are fewest, and
gradually increasing with age of the culture until they nearly, com-
pletely fill the cell. I have not noted their disappearance during
the late stages of growth as has been described by Albert; on the
contrary, they become more and more prominent, and very irregular
in form in the large filamentous and irregular cells found in old
cultures.

It will be noted in Figure 28 that the curves for cell length
and for number of granules per cell follow the same general course,
the approximation being greater during the later stages of growth
than at the beginning, since the granules decreased at once, whereas
the cells did not begin to shorten until after some hours. There
is apparent then a degree of correlation between the size of the
cells and the number of granules which they contain. I have cal-
culated the coefficients of this correlation for the various samples
in which the granules were counted (100 cells were counted and

108 MORPHOLOGIC VARIATION

measured), which are given in Table XXI. The coefficient is
high when the granules are numerous, low when they are scarce.
This is probably a spurious correlation; the two values happen to
move together during part of the growth cycle, but are not actually
related.

While one cannot clearly establish the relationship between the
morphologic variations of this organisms and the phases of growth
from the data at hand, it can be readily demonstrated that these
variations are dependent upon the rate of growth by observing the
effect of varying the size of the seeding and the concentration of
nutrients in the medium. This has not been done quantitatively,
but the drawings in Figure 30 show the effect. These were made
from cultures on dextrose agar (dextrose 5 per cent, peptone 1 per
cent) inoculated from a twenty-four-hour old culture on the same
medium, the cells of which are illustrated in the drawing marked
0 hour. Three series were run in parallel. The two marked "Heavy
inoculum" were seeded with the same number of cells; that marked
"Light inoculum" received one-tenth as many. The two marked
"Full strength medium" were inoculated onto agar slants having
the composition given above; the one marked "Dilute medium" was
planted onto agar slants having just one-fourth as much dextrose
and peptone. It will be seen that in the dilute medium the morpho-
logic variations were not so pronounced as in the full strength, and
that with a light seeding there was more variation than with the
larger inoculum. Those conditions which lead to a high growth
rate, viz., small seeding and large concentration of nutrients, cause
a development of more of the small, intensely stained, granule-free
cells and their persistence for a longer period of time. This being
the case, it is probable that these small cells are truly to be con-
sidered the form characteristic of the growth period rather than
the resting phase, that they are the embryonic forms of this type
of organism, and the failure of the curves to correlate in phase in
Figure 28 is really due to some experimental error.

It is rather striking in these drawings that the small forms be-
come almost spherical and are arranged in chains; this is especially
evident in the light inoculum, full strength medium. At ten and
fifteen hours the cells would hardly be recognized as being related

OBSERVATIONS OF DIPHTHEROID BACILLUS 109

0 hrs.

6 hrs

10 hrs,

1$ hrs.

24 hrs.

Heavy Inoculum,
Dilute Mcdixim.

\"n

\

c^e'^

Heavy Inoculum,
Fall StrtTigth
Mediun?.

/

**^

/^•'

.\

9§1#

q^^'^c^

Light li^oculuTn,
Full Strength
Medium

^

'--J

/

♦^

\^/-*
i

r-»

b^^

VSA

Fig. 30. Influence of Size of Seeding and Concentration of Nutrients on
Morphologic Variations of a Diphtheroid Organism.

110 MORPHOLOGIC VARIATION

to those used for seeding; they look like streptococci. This is rather
interesting because various authors, particularly Mellon, have sug-
gested that the diphtheroids and the streptococci are closely related.
If such a relationship exists, these observations might be construed
to indicate that the diphtheroids have developed from the strepto-
cocci and during their embryonic phase revert to the ancestral type.
At any rate it would seem likely that the so-called mutations of
diphtheroids to streptococci as described, among others, by Mellon,
may be the result of merely comparing different phases of growth,
a culture on one medium being in the resting phase when on another
it may still be actively growing.

CHAPTER VII

A NOTE ON SPORE FORMATION

It is frequently stated that bacteria form spores when condi-
tions become unfavorable for growth. This is, however, only true
in a narrow sense. Certain factors which tend to restrict the growth
rate also tend to delay rather than to accelerate spore formation.
Thus Migula has shown that spore formation is less rapid and
complete on either side of the optimum temperature for growth.
It is well known, however, that bacteria do not form spores until
the culture has practically stopped growing; that spores are struc-
tures peculiar to the resting cells rather than to the actively growing
cells. There is, however, as far as I can learn, no very definite
evidence in the literature concerning the actual relationships of
spore formation to phases of the growth curve.

This, however, is apparent in the data obtained on the growth
of Bacillus megatherium, presented in Chapter IV. During the
course of the investigation of the sizes of the cells in the first
culture reported in that chapter, the number of free spores was
incidentally determined, and is recorded in Table VI. The relation-
ship of spore formation to the rate of growth in the culture is indi-
cated in Figure 31, where the logarithms of the number of vegetative
cells and the number of free spores are plotted on the same scale
(sporangia being counted as vegetative cells, with only free spores
being considered as spores). It wilt be noted from this graph that
spore formation commenced practically at the point of inflection
between the logarithmic growth phase and the resting phase; that
after spore formation commenced it proceeded at a practically con-
stant rate for some time, the logarithms of the number of spores
falling on a straight line; but that spore formation also began to
slow up after some hours, the rate continually decreasing during
the period of observation. Not all of the cells, therefore, developed
into spores, at least within the period of time that the culture was
studied,

111

112

MORPHOLOGIC VARIATION

Since spore formation is a characteristic of the resting phase,
the rate of spore formation should be influenced by those factors

Log. of 1:0. of Cells __
Log. of Ko. of Spores

12

2k
Hours.

36

14S

Fig.

31. Relation of Rate of Spore Formation to Rate of Growth in
B. megatherium.

which also tend to influence the rate of growth and the form of
the growth curve. Migula has determined that spore formation

A NOTE ON SPORE FORMATION 113

seems to depend upon the density of the population in the culture.
He states that spore formation can be indefinitely postponed if the
culture is transferred to a new medium at a period preceding the
appearance of spores. It seems to begin after a definite number
of cell divisions have taken place, spore formation appearing at
a definite concentration of cells, regardless of the initial size of
the seeding for a given volume of medium, and begins always when
the medium is no longer suitable for vegetative multiplication, that
is when growth has ceased. Migula also found that spore formation
could not be delayed by the addition of nutrient to the medium in
concentrated form (dry peptone and meat extract), but, on the
other hand, by adding distilled water spore formation was postponed.
He concluded, therefore, that spore formation was not due to an
exhaustion of the medium; that it depended entirely upon the con-
centration of the cells, and was probably the result of an accumula-
tion of products of metabolism of the bacteria.

In order to determine more clearly this relationship between
rate of growth and rate of spore formation, the following experiment
was performed: A culture of Bacillus cohaerens some days old, in
which there were practically no vegetative cells, was inoculated on
to four sets of agar slants, two of normal agar, the other two con-
taining one- fourth as much peptone and meat extract. One set
of each kind was inoculated with a very heavy suspension, and
the other with the same suspension diluted 1-50. At regular inter-
vals samples were removed and the percentage of free spores among
200 cells was determined from each sample. The results are given
in Table XXII, and are presented graphically in Figure 32.

It will be noted from this graph that spore formation proceeded
more rapidly in the media of lower nutrient value than in the full
strength media, regardless of the size of seeding, and that it pro-
ceeded more rapidly in the heavily seeded cultures than in the lightly
seeded ones, though the difference here was not so pronounced as was
the effect produced by varying concentration of nutrients. In the
heavily seeded dilute medium, spores never entirely disappeared,
new spores starting to form before all of those introduced had germi-
nated. It would appear, therefore, that the rate of spore formation
is determined not by the concentration of cells alone, but rather

114

MORPHOLOGIC VARIATION

by the density of the population in relation to the concentration
of foodstuff in the medium. It appears earlier in the heavily seeded
cultures and in the more dilute media, because those cultures do
not have so long a period of vegetative reproduction; that is, because
the resting phase of the culture appears earlier. This is quite con-

100^

60^

kC^

20^

T f

1 I I r^

Heavy Inoculum

Full S trength Medium

Light Inoculum

Full Strength Medium

Heavy Inoculum
Dilute Medium

Light Inoculum
Dilute Medium

12 18

v^ 3t kk 56 9+ bO bb 72
Hours.

Fig. 32.

Influence of Size of Seeding and Concentration of Nutrients
ON Rate of Spore Formation in B. cohaerens.

trary to the results obtained by Migula, and I cannot help but
believe that his' observations were in error.

If we are to consider the large, deeply stained cells free of
granules as comparable with the embryonic cells of a growing plant
or animal, as has been proposed in Chapter V, then conversely we
must look upon those cells with which we are most familiar, the

A NOTE ON SPORE FORMATION 115

cells characteristic of the resting phase of the culture, as being
analogous to the differentiated cells which occur in the mature plant
or animal. Minot states that differentiation consists essentially in
this, that something new appears within the cell. I submit that
the metachromatic granules observed in diphtheroid bacteria, the
unstainable granules previously described in Bacillus megatherium,
the spores which are found in many species of bacteria, are some-
thing new which appears within the protoplasm analogous to the
various granules, secretory or reserve material, or of unknown func-
tion, which serve to distinguish the differentiated cells of a multi-
cellular organism from their embryonic prototypes.

It has been shown in Chapter VI that the metachromatic gran-
ules begin to reappear with a decrease in the growth rate, and are
most persistent in those cultures (heavily seeded, or dilute medium)
which have the lowest growth rate; which I have just shown is
also true of spores. Those factors then which tend to produce a
short growth phase of low rate tend to give a maximal development
of those characters which distinguish the mature or differentiated
cells, as conversely they tend to give a minimal development of
the embryonic types.

CHAPTER VIII

MORPHOLOGIC VARIATIONS OF THE
CHOLERA VIBRIO

The organism of cholera has long been a favorite for studies
of pleomorphism. Many bacteriologists, among them both pleomor-
phists and monomorphists, have found that this organism is more
prone to morphologic variation than many other forms; that cul-
tures of the same origin on different media, and cultures of different
origin on the same media, which may be identical in other characters,
may be widely different in the form of their cells. Short and long
rods, "comma" forms, spiral filaments, large and small spherical
cells, as well as budding and branching, bulging and irregular cell
forms, have all been described over and over again. Most of these
observations have been made on unusual media, though even in
cultures on standard beef extract agar considerable variation is en-
countered. These variants from the typical vibrio form have been
looked upon by some authors as involution forms, that is cells which
have suffered an injury or are dead; by others, as unusual repro-
ductive cells (for example, gametocytes by Almquist and Enderl.ein) ;
and by still others simply as the result of the action of physical or
chemical forces, particularly osmotic pressure, on the cells, implying
neither necessarily an injury nor a change in the mode of reproduc-
tion (for example, as plasmoptysis by A. Fischer). For complete
literature on the subject, the reader is referred to Lohnis' monograph
on "Life Cycles of Bacteria."

It was thought desirable, therefore, to extend these quantitative
studies to include this organism, and particularly to see if something
could be accomplished in the way of applying statistical methods
to variations in the form of the cells rather than their size. It will
be remembered that with the colon bacillus it was found that the
variations in form could be fairly readily expressed by the area-
length index (the area of the projected image divided by the length
squared), this measure serving to classify the cells into long fila-

117

118 MORPHOLOGIC VARIATION

mentous types on the one hand, and spherical cells on the other.
But it was pointed out that this measure only suffices as long as
the cells are symmetrical with regard to both axes. Cells which
show terminal or lateral or central bulgings, cells which are branched,
wedge-shaped cells, and such unusual types, are not classified by
this measure in regard to their asymmetry; they are merely placed
in the series according to whether their general form approaches
the filamentous or the spherical. Such asymmetrical cells were
found to be quite numerous with the cholera organism studied.
Nevertheless this method of studying variations in form was carried
through with the hope that, regardless of the asymmetry of some
of the cells, it would serve to classify them, although roughly, at
least to an extent which would give some idea as to the nature of
the variations which are occurring.

A five-day-old culture was inoculated onto agar slants, from
which samples were removed at regular intervals for observations.
Photomicrographs were prepared and reprojected to give a final
magnification of 21,000X, the cells being traced on to cards and
measured as described in Chapter III.

The general character of the morphologic variations of this or-
ganism are clearly indicated in Figure 33. This figure has been
prepared by plotting the logarithms of the number of cells against
the logarithms of the hours of growth. Since the growth curve has
been plotted from cell counts made by the direct microscopic method,
it does not show the true rate of death, as death of the cells is
not registered in these counts until they have autolysed and disap-
peared. I have indicated by a dotted line on this graph what I
believe was approximately the true death curve as it would have
been if plate counts had been made. This estimate may be justi-
fied by data to be presented later. Superimposed upon this curve
columns of cells have been erected in the order of their area-length
indices (those with the lowest index being at the bottom), choosing
each tenth cell from the samples of 200 which were measured.
The way in which the cells are classified by this value is apparent
from inspection of the figure. The longest and most slender cells
occur at the bottom, the cells becoming shorter and plmnper as
we approach the top of the column, those cells at the very top being

THE CHOLERA VIBRIO 119

practically spherical. It will be noted also from this figure that
the asymmetrical cells are not classified; that is, cells with terminal
bulgings, for instance, may be found at almost any position in the
column, their position being determined by their relative plumpness
or thinness, rather than by their asymmetry.

It will be noted that during the lag phase practically no varia-
tion in the cells occurred. The cells used for inoculation consisted
in part of typical vibrio forms, and in part of small oval and coccoid
cells with some irregular budding and branching cells. With the

L2 15 25 36 hS 72 qS 120 iGz 2Uc

Hours.

Fig. 33. Representative Cells from a Culture of the Cholera Vibrio at
Different Stages of Growth.

initiation of growth at six hours a definite alteration in form occurred,
the majority of the cells being larger and relatively straight. These
forms persisted during the period of active growth, but with the
cessation of growth at twelve hours the cells became more slender
and more curved, and during the resting period these typical vibrio
forms were predominant. Beginning at thirty-six hours there ap-
peared more and more irregular cell forms of the type which I have
designated asymmetrical. Without any measurements, then, one can
clearly distinguish in this graph the occurrence of three distinct types

120 MORPHOLOGIC VARIATION

of cells characteristic of the growth phase, the resting phase, and
the death phase, respectively; the first being large, relatively plump,
and straight, the second being more slender and markedly curved,
the third being characterized by asymmetry of the cells.

The variations in size of the cells were relatively slight as com-
pared with the variations noted previously in Bacillus megatherium
and the colon bacillus. The mean length of the cells is given in
Table XXV. There was a slight increase beginning at six hours,
that is at the end of the lag period, which persisted to practically
the end of the resting period (forty-eight hours), after which the
cells showed a considerable decrease in length. Clark and Ruehl,
however, found that the cholera vibrio showed an increase in length
of cells of practically the same degree as that noted with the other
organisms that they studied, and it is quite possible that in my
observations the change in the size of cells is not as great as is
usually the case. No studies of frequency distribution of the cells
with regard to their size have been made with this organism.

More significant results were obtained from the study of the
variations in form by means of the area-length index. These are
given in Table XXIV. No pronounced variation occurred until
after nine hours, corresponding approximately with the point of in-
flection between positive and negative acceleration in growth rate.
After this period the area-length index decreased considerably, reach-
ing its minimum at eighteen hours, which corresponds with the point
of inflection between the growth phase and the resting phase, after
which it again increased, remaining fairly constant at a high level
during the death phase. Slendemess of the cells is, then, a charac-
teristic of the resting phase, and plumpness a characteristic of the
death phase. When plump cells are used for seeding, the latter
half of the growth phase is occupied by a progressive increase in
their slenderness.

While the area-length index serves to measure the cells with re-
gard to their relative slenderness or plumpness, it does not take
care of another variation in form characteristic of the cholera vibrio,
namely the degree of curvature of the cells. This, however, can
be simply determined by measuring the length of the cells (by
means of dividers) along their major axis (L^) and the shortest

Plate II. Correlation of Morphologic Variations with the Rate of
Growth in a Culture of the Cholera Vibrio.

From "A Statistical Study of the Form and Growth of the Cholera Vibrio."
Through the courtesy of the Journal of Infectious Diseases. (1925, 37, 75).

THE CHOLERA VIBRIO 121

distance between their tips (LJ . The proportion between the difference

of these two values and the true length of the cells ^~ ^' serves

as an abstract measure of the variation in curvature. The mecin
values of this index of curvature are presented in Table XXV.
It will be seen that with the initiation of growth at six hours this
value began to increase, reaching its maximum at twelve hours, ap-
proximately at the end of growth, and then progressively decreasing
again to a minimum at the end of the period of observation, at
ten days. That the degree of curvature is correlated with the rate
of growth is quite apparent by an inspection of Plate 2. Particularly
during the first twenty-four hours the curve for index of curvature
is practically parallel with the growth curve. The maximum is main-
tained during the resting phase. Since the cells increase in curva-
ture as they decrease in area-length index, that is as they become
more slender, it seems probable that there would be found a cor-
relation between the degree of slenderness and the degree of curva-
ture, and this seems all the more apparent in the composite photo-
graphs given in Plate 2. When, however, the index of curvature
was plotted against the area-length index with cells from the nine
hour, eighteen hour, and thirty-six hour samples, no such correla-
tion could be established. That is, while the period of greatest
curvature corresponds with the period of greatest slenderness of
the cells, it is not necessarily the most slender cells which are the
most curved.

The frequency distribution of the cells with regard to their
area-length indices is presented in Plate 2, where an attempt has
been made to show in one illustration nearly all the morphologic
variations of this organism. The frequency distribution curves have
been erected in succession behind each other in isometric perspec-
tive. On the upper face of the parallelopiped containing them the
mean area-length index has _^ been plotted. On the right face the
rate of growth is shown by plotting the logarithms of the number
of cells. On this face there is also given the mean index of curva-
ture. The significance of the area-length index is indicated in the
small figure at the bottom of the plate, which presents composite
photographs of all of the cells from the five-hour sample arranged

122 MORPHOLOGIC VARIATION

in classes according to their area-length indices. The values given
are the maxima for the class, thirty for instance indicating all of
the cells with area-length indices from 25.1 to 30.0. This shows
well the variation from long, slender, curved cells, on the one hand,
to small, spherical cells on the other.

It will be noted from the frequency distribution curves that the
cells introduced into the medium apparently consisted of a large
group of relatively slender cells distributed about one mode, and
a small group of more oval cells distributed about another. As
growth proceeded the frequency distribution curves exhibit a regular
change in form. The minor mode is gradually absorbed and disap-
pears; the range becomes restricted, the mode higher; and at eighteen
hours we have a tall, symmetrical curve. With the beginning of
the death phase the frequency curves again return to the form ex-
hibited by the cells used for seeding; the base becomes extended,
the mode lower; and there appears a new minor mode of oval cells,
which reaches its maximum at four days, and then again disappears,
the final curve at ten days being simply a much skewed curve with
a broad base.

These variations in the frequency distribution are reflected in
the coefficients of variability of the area-length index, presented in
Table XXIV. This value showed considerable fluctuation, but was
distinctly lower during the period of active growth from nine to
eighteen hours, and increased again with the cessation of growth
to reach its maximum at four days. Since the area-length index
does not classify the asymmetrical cells, an attempt has been made
to follow these roughly by simple inspection of the drawings, and
the results are given in Table XXV. It will be seen that the per-
centage of asymmetrical cells steadfly decreased during growth, the
number remaining low until the beginning of the death phase at
forty-eight hours, when they again increased, reaching a maximum
at four days, after which there was again a slight decrease in num-
ber.

With all of these measurements at hand we are now in a posi-
tion to say something more precise concerning the three types of
cefls previously mentioned. During the initial period of dormancy
the cells showed no pronounced variation. With positive accelera-

THE CHOLERA VIBRIO 123

tion, in growth rate there was an increase in size, a decrease in index
of curvature, and only a very slight decrease in area-length index.
The embryonic cells are therefore large, plump, and straight. With
negative acceleration in growth there occurs a steady increase in
the index of curvature and a decrease in the area-length index, with
but little change in length. The maximum index of curvature and
the minimum area-length index are found during the resting period.
Therefore the mature cells are slender and curved, the typical vibrio
form. The death phase is characterized particularly by increased
variation in form, as indicated both by the coefficient of variability
of the area-length index, and by the percentage of asymmetrical
cells; further by the bimodality of the distribution curves indicating
a tendency to break up into two groups of cells, one elongated, the
other oval or spherical. These asymmetrical and coccoid cells are
therefore to be looked upon as the senescent forms.

The asymmetrical cells found in the death phase will be recog-
nized as those types which have so frequently been described as un-
usual reproductive cells, forming either buds or conidia or sexual
cells. It would seem at first glance that probably the small oval
or spherical cells have been derived from the longer, more slender
cells by such a process of budding, and this impression is strengthened
by the bimodality of the frequency distribution curves at this time,
for if the small, oval cells were produced simply by a contraction
of the longer cells, one would expect this process to be continuous,
that is to find all transitional stages between the long, slender and
the small, spherical types. If, however, a small portion of the cell
were constricted off, it would naturally be more oval or spherical
than the parent cell from which it was derived, and there would
be lacking such transition cells, and therefore a bimodal frequency
curve should result. It is curious that this bimodality in the fre-
quency curves disappears with continued observation, and that dur-
ing the latter stages of growth the curves tend again to become
more symmetrical. This might be explained on the basis that the
small, oval, or spherical types are dead, that they ultimately disinte-
grate, leaving behind only the more resistant slender types. Simple
observation, however, without any measurements, seems to indicate
that this is not the case; that with continued age of the culture the

124 MORPHOLOGIC VARIATION

tendency is for the cells to become smaller and smaller and more
and more spherical, cultures a month or more old, for instance,
appearing much like cultures of staphylococcus.

CHAPTER IX

THE SENESCENT FORMS OF THE COLON BACILLUS

The occurrence of irregular cell types which I have designated
as asymmetrical in the preceding chapter seems to be clearly cor-
related with death. Now these cells are just those forms which have
received the greatest attention from the exponents of life cycles in
bacteria, and just those types which have been most the subject
of controversy between the monomorphists and pleomorphists; the
former, by designating them involution forms, imply an abnormality
due to actual or approaching death; the latter believe them to be
reproductive cells of one type or another. I do not believe that any
data have as yet been brought forward to definitely settle this prob-
lem, for it may well be that those factors which lead to the death
of some of the cells in a culture may react upon others to cause
them to develop into one or another type of resting cell or reproduc-
tive body; but their occurrence only in the death phase is strongly
presumptive evidence of the orthodox view that they are involution
forms, and places squarely upon the pleomorphists the burden of
proving that they are otherwise. By naming these irregular types
"senescent" forms I express my opinion that the monomorphistic
teaching regarding them is correct. But I prefer this term to "in-
volution form" because it clearly expresses the analogy between
them and the cell changes which occur in old age in multicellular
organisms, and because it does not so much imply that they are
already actually dead.

The observations recorded in preceding chapters have for the
most part not been continued long enough to include more than the
very beginnings of the death phase. In none of them have plate
counts been made, so that it is not possible to determine the true
rate of death. Moreover, the measures of morphologic variation
used do not reveal except roughly the changes in the cells which
seem to be most characteristic of the death phase. It was
thought desirable therefore to continue these studies with some ob-

125

126 MORPHOLOGIC VARIATION

servations of the degree of variation in form of the cells during the
death phase; to devise a measure of this variation which would con-
sider all of the changes in form that occur; to include in the in-
vestigation the decrease in number of living cells by plate counts,
as well as the decrease in total number of cells by microscopic
counts; and to carry on the study in a number of different media
with the idea of thus varying the death rate and so determining more
conclusively whether or not the variations in cell form were directly
correlated with the rate of death. Very little work has been pub-
lished upon factors influencing the natural death rate in cultures,
and since it could not be predicted that this would vary with the
size of seeding and the concentration of nutrients, as does the growth
rate, it was thought that more would be accomplished by using
media of different composition, such as has been done by most of
the authors who have described unusual cell forms, in an attempt to
obtain the same types of morphologic variation as they have ob-
served.

With these aims in mind the following experiment was performed.
Five lots of agar were prepared, each having as nutrient 2 per
cent of peptone. These proved to be neutral (pH 7.4). One lot
was kept as such, another was made acid (pH 4.5) by the addition
of HCl, a third alkaline (pH 9.2) by the addition of NaOH. The
fourth had added to it 5 per cent of NaCl; the fifth, 3 per cent of
CaCP. These were inoculated with a stock culture of the colon
bacillus, the same strain as was used for the observations recorded
in Chapter V. It was not possible to make all of these observations
simultaneously, but conditions were kept as much alike as possible;
the media were all made from the same lot of peptone, the cultures
were inoculated from twenty-four-hour agar slants suspended in
water to about the same density. The tubes were incubated at
37° in a moist chamber. Samples were removed after one, two,
three, five, seven, ten, fourteen, nineteen, and twenty-five days (the
final sample of the NaCl series was removed on the twentieth day,
listed in Table XXVII as the nineteenth day). Five tubes were
removed for sampling, the growth of each suspended in 5 c.c. of
sterile water, and the contents of the five tubes were then pooled.
From this suspension a 1 c.c. sample was removed and serially

FORMS OF THE COLON BACILLUS 127

diluted in sterile tap water for plating. Plates were made for the
most part in triplicate, though on a few occasions five plates were
poured for each dilution. The counts recorded are the mean of all
plates of the same dilution, that dilution being chosen which gave the
nearest to one hundred colonies per plate. Another 1 c.c. sample
was mixed with Congo red solution, and slides were prepared for
counting and photographing preparatory to measuring their size
and form. The photographs were projected to give a final magnifi-
cation of 1 9,000 X; the image of the cells was traced on cards as
previously used for the cholera vibrio. In addition to the direct
microscopic counts the proportion of cells stained by Congo red
was also recorded, and this yielded information of considerable sig-
nificance.

The measurement of the rate of death in a culture, and the
determination of the limits of the death phase present some diffi-
culties. The death phase has generally been defined as beginning
at the point where the number of living cells determined by plate
counts begins to decrease, but undoubtedly the factor which causes
the cells to die, whatever it may be, actually begins to act upon them
earlier than this, at least at the point of inflection in the growth
curve from positive to negative acceleration in rate of growth, and
possibly earlier. This has been claimed by Wilson, who finds at
all stages of growth a discrepancy between the microscopic count
and the plate count, which can only be explained by the fact that
some of the cells are dying, while the majority of them are still
growing. By the time the culture has reached the beginning of the
death phase, as defined above, there must be already a considerable
number of dead cells present, and this is clearly indicated in my
own observations here reported by the appearance in every case
of a considerable number of cells stained by Congo red before there
had occurred any decrease in the plate counts. On the other hand,
when the number of living cells shows a decrease, and the culture
may definitely be said to have entered the death phase, some of the
cells are still multiplying, as indicated by the increase in microscopic
counts, which generally continues for some time longer.

In these experiments we have three sets of data from which to
compute the death rate; the plate counts which indicate the decrease

128 MORPHOLOGIC VARIATION

in actual number of living cells, the microscopic counts which indi-
cate not only the death of the cells but also their complete autolysis
and disappearance, and the proportion of cells stained by Congo
red. These values are given in Table XXVI. The property of
staining by Congo red, as was described in Chapter III, indicates
not only the death of the cells but also that they have undergone
a certain degree of autolysis. This being the case, as death proceeds
the plate counts should show a marked decrease before the micro-
scopic counts, and the proportion of stained cells should also increase
at a slower rate than the plate counts decrease. Such was found
to be the case. In order to show these three quantities graphically,
in a form in which they might be directly comparable, the curves
have been plotted in the following manner: The percentage of de-
crease from the maximum plate count was computed for the first
day in which a decrease was noted, and to this was added day by
day the further decreases in number of cells as a percentage of
the maximum number attained in the culture. Thus we obtain a
cumulative percentile curve which is directly comparable to the
curve for the percentage of stainable cells, which of course is also
cumulative. In a similar manner a curve for cumulative percentile
decrease in the number of cells in microscopic counts was plotted.
These three curves are given side by side for each of the five cul-
tures in Figure 36. It will be seen that in every case the curve
for percentage increase in stained cells (dotted line) occupies a
position between the curve for percentage decrease in viable cells
(solid line) and the percentage decrease in microscopic counts (dot
and dash line).

In most cases the decrease in microscopic counts was not marked
even at the end of twenty-five days. Their curves are therefore
not completed and not easily compared with the curves for plate
counts. The curves for stainable cells, however, are nearly complete
in most cases, and the distance between these curves and the curves
for plate counts then serves as a fairly good measure of the rate
of autolysis in the culture. This measure can be best taken on
the 50 percentile line, and is indicated in the graph, the figures
given being the number of days which have elapsed from the time
that half of the cells were dead, as determined by the plate counts.

FORMS OF THE COLON BACILLUS 129

and the time that half of the cells remaining in the culture have
become stained by Congo red.

The rate of autolysis so determined proved to be lowest in the
alkaline culture (17 days), the other cultures in order of increas-
ing autolysis rate being the neutral one (13.25 days), the salt medium
(6.25 days), the acid agar (4.25 days), and the medium containing
calcium chloride (4 days).

The highest rate of death was shown by the salt culture, the
mean daily decrease in the plate counts being 21.8 per cent; in
the alkaline medium it was 16.5 per cent, in the acid one 16.4
per cent, in the calcium chloride agar 11.9 per cent, and in the
neutral culture 9.7 per cent. The period of maximum death rate
occurred in the neutral agar between the second and third days, the
decrease per day being 60.2 per cent; in the acid medium between
the fifth and seventh days, 37.7 per cent; in the medium with sodium
chloride, fifth to seventh days, 32.2 per cent; in the calcium chloride
agar, second to third days, 26.8 per cent; and in the alkaline culture,
tenth to fourteenth day, 23.3 per cent. The order of the cultures
with regard to rate of autolysis does not agree with either of the
above, and the rate of autolysis is therefore independent of the
death rate.

The medium most favorable to growth, as measured by the
maximum yield (plate counts), was the acid agar; the others, in
descending order being the alkaline, the neutral, and those contain-
ing salt and calcium chloride. The death phase began earliest in
the neutral and alkaline media, both showing the maximum plate
counts on the second day; in the salt agar and calcium chloride
medium on the third day; and in the acid agar on the fifth day.

The microscopic counts in the death phase are determined not
only by the rates of death and autolysis, but also by the degree
to which those cells still viable are multiplying; for the marked in-
crease in the microscopic counts after the plate counts have reached
their maximum can only be explained by the fact that some of
the cells are still actively dividing and the progeny then soon dying.
This occurred in greatest degree in the neutral medium, the maxi-
mum microscopic count showing an increase of 176 per cent over
the maximum plate count. In the calcium chloride medium this

130 MORPHOLOGIC VARIATION

increase was 141 per cent, in the alkaline culture 58 per cent, in
the acid 41 per cent, and in the salt agar 12 per cent. It is interest-
ing in this connection that the first two cultures named also had a
high incidence of constricted (dividing) cells.

At first an attempt was made, as in the preceding study, to
follow variations in form by means of the area-length index. This,
however, failed to yield any significant results, the figures so ob-
tained showing no regular changes with increasing age of the culture,
nor noteworthy differences when the averages for the various cultures
were compared. It was quite apparent, however, by simply inspect-
ing the slides that the important changes in cell form were not so
much variations in relative slenderness or thickness as they were
departures from the normal symmetry, i.e., development of budding,
branching, bulging, and other unusual cell types. To measure the
degree of variation in form then some scheme had to be adopted
which would not only consider the relative slenderness and plump-
ness of the cells, but also these asymmetrical types, as well as curva-
ture in the cells. What was desired was a quantitative expression
of the degree of total variation encountered which could be com-
pared directly with rates of death and autolysis determined as above.

This was accomplished in the following way. The cells were
(more or less arbitrarily) divided into ten classes, illustrated in
Figure 34. This illustration has been made by tracing from the
cards, cells of the various classes from many different samples,
simply to show the types of cells included in each class. They
were chosen more or less at random, although an attempt was made
to include extreme cases. It should be borne in mind that these
composite fields do not indicate the frequency with which any of
the cell types occurred in any particular culture; they are intended
merely to indicate the nature of the classification followed. They
are arranged and numbered in the order of their mean frequency
in all of the cultures studied.

These ten classes include four classes of straight symmetrical
cells: viz., slender rods (Class IV), normal rods Class I), oval cells
(Class III), and spherical cells (Class VII). This is of course an
entirely arbitrary division, since there occurs every degree of gra-
dation from long filaments to perfectly round cells. But having

FORMS OF THE COLON BACILLUS

131

Fig. 34.

Representative Cells Illustrating the Ten Classes Used in
Studying Involution Forms of the Colon Bacillus.

From "The Involution Forms of Escherichia Coli." Through the courtesy of
the Journal of Infectious Diseases. (1926, 39, 429).

132 MORPHOLOGIC VARIATION

measured the area-length index, there was provided a means of
separating these into classes which would not depend at all upon
personal judgment. Area-length indices of 20, 40, and 60 were
chosen as dividing points for these classes. Long, slender filamentous
forms were considered as having an area-length index of 20 or
less; normal forms from 21 to 40; oval forms from 41 to 60; spheri-
cal cells, those with an index higher than 60.

Further classes were provided as follows: All curved cells (Class
II), regardless of their size or slenderness as long as there was
a definite curvature to their axis; all cells which were greater in
diameter at one end than the other (Class V), either provided with
a knob-like expansion or showing a uniform graduation in diameter
from the larger end to the smaller; all cells which showed a central
constriction (Class VI), i.e., cells probably in the act of dividing;
all cells which showed a lateral projection (Class VIII) whether
in the form of a small bud or definite side branch; all cells which
showed a sub-terminal constriction, i.e., which appeared to have
a terminal bud (Class X) ; and finally all cells which exhibited a
central bulging (Class IX).

Having decided upon this classification, all of the cards bearing
the tracings of cells were gone through, and the proportion of the cells
in each class determined for each sample. The average percentage
of each class for all the the cultures at all stages of growth was then
calculated, and the reciprocal of this percentage determined. The
product of the average percentage incidence of any class by its recipro-
cal will of course equal one (or for convenience 100), and therefore
the mean of these products for all ten classes will also equal 100. If
now in any given sample the percentage frequency of each class is
multiplied by the reciprocal of the percentage frequency of that
class for the entire series of samples, and the average of these prod-
ucts for the ten classes is taken, there is obtained a value which in-
dicates whether in that sample the degree of variation is greater or
less than the average of all of the samples. Where there occurs more
than the average number of the more frequent cell types (and there-
fore fewer of the less frequent types) this value will be less than 100.
while conversely if there occurs more than the average number of the
rarer cell forms the value will be greater than 100. This is really an

FORMS OF THE COLON BACILLUS 133

adaptation of the principle of index numbers to the problem at hand,
the index being the mean variation in form of all of the cultures being
compared. The value so computed may be called the index of vari-
ation of form. It is admittedly not an entirely satisfactory method,
being rather artificial and involving to some extent personal judgment
in classifying the cells, but in the absence of anything better may
serve to supply at least an approximate quantitative expression of
the degree of variation in form of the cells, which may be compared
from day to day and from culture to culture.

The incidence of the various cell classes in the five cultures at
various stages of growth is given in Table XXVII together with the
indices of variation calculated from them. The mean frequencies of
the ten cell types of the five cultures at all stages of growth are shown
graphically in Figure 35.

The fluctuations of the index of variation day by day are some-
what irregular and do not appear to be highly significant; this is
perhaps largely due to the small number of cells in each sample.
The alkaline culture showed very little variation; there was a slight
rise on the nineteenth day followed by a decline, corresponding with
the period of decrease in death rate. In the neutral medium there
occurred a well-defined increase reaching the maximum on the third
day, corresponding with the beginning of the death phase; this was
followed by a decrease and a second increase with its maximum on
the fourteenth day, at which point the death phase had nearly reached
its end, the number of viable cells remaining nearly the same during
the remaining period of observation. The medium containing NaCl
showed a very pronounced rise in the index of variation on the sec-
ond day, due largely to the remarkably high incidence of budding
and branching cells, followed by an irregular decline. The acid cul-
ture exhibited two maxima, one on the seventh day corresponding
with the beginning of the death phase, and one on the nineteenth
day when the death rate was decreasing. The medium containing
CaClg gave a slight initial rise followed by a decline, then a pro-
nounced increase which was sustained until the nineteenth day, there
being a marked reduction in the variation of the cells on the twenty-
fifth day.

134

MORPHOLOGIC VARIATION

These fluctuations with time, while somewhat irregular and vary-
ing considerably in the different cultures, show in general a tendency

60^1

Uo

20

Alkaline

I23U56759IO
Cell Classes.

Neutral

1

b

H

■

H

60^

IK)

20

123^1 5073910
Cell Classes.

60^,

ko

20

Sodium Chloride

Acid

L

■

l|

iU_

Pi

^m

I^H

gi

n

1^1

■

HH

H

■■

^H

^^

UO

20

I23U567S9IO
Cell Classes

123^^ 567s 9 la

Cell Classes
Calcium Chloride.

^f

ko

■

1

■

.h

■

■

^

^

■

1231156 789 10
Cell Classes.
Fig. 35. Frequency Distributions of the Ten Cell Classes in Various
Cultures of a Colon Bacillus.

to the formation of two maxima, one corresponding with the begin-
ning of the death phase and one with the end. By computing the

FORMS OF THE COLON BACILLUS 135

mean index of variation of the five cultures for each day we increase
the number of cases and obtain a correspondingly more accurate pic-
ture of the variations with time. There occurs a sharp rise at the
beginning of the death phase followed by a decrease, then a second
more sustained rise with another marked decrease on the last day.
This final drop in the degree of variation in the form of the cells,
occurring during the period of decreasing death rate as measured
by the plate counts and of increasing rate as measured by the micro-
scopic counts, can readly be explained by the final dissolution of
those cells which presented the greatest departure from the "normal"
form, and may be considered to support the argument that these ir-
regular cells are actually dead. I am at a loss,- however, to explain
the initial rise and fall in the index of variation.

While the daily fluctuations in the index of variation did not
prove very significant, and particularly there could not be established
any clear relationship between the period of greatest degree of vari-
ation and the period of most rapid death, important results were ob-
tained when the different cultures were compared with each other
with regard to their mean indices of variation. In order of magnitude
of this value they are arranged as follows: Alkaline (67.71), neutral
(87.56), salt (93.01), acid (93.42), and calcium chloride (157.25).
It will be noted that the values for the salt and acid cultures are
nearly identical. But the figures for the salt culture are not directly
comparable with the others, as no observation was made on the twen-
ty-fifth day. Since in all of the cultures there was observed a de-
crease in variation on the last day, it is probable that this would also
have occurred in the medium with sodium chloride, and that the mean
index of variation would have been somewhat reduced.

The rank order of the cultures with regard to their mean indices
of variation does not correspond with their rank order as regards
mean rate of death, maximum rate of death, or favorableness of the
medium for growth. In fact the cultures containing calcium chlo-
ride, which gave the highest^ degree of variation in cell form, had
nearly the lowest death rate. It is clear therefore that there cannot
be established a correlation between the index of variation and the
rate of death.

This rank order does, however, correspond definitely with the

136

MORPHOLOGIC VARIATION

rank order of the cultures according to their rates of autolysis, the
alkaline culture which required the longest time for half of the dead
cells to become stainable being the one with the lowest index of

Alkaline _ _

Days

1001,

■mill I \^^,

80

— ^^^"^^^^^^^^^^ ""

60

__ y / CaClp —

20
0

1 '^* —

h\\ 1 1 1 1/_J —

5

10

Pays

25

Fig. 36. Rates of Death and Autolysis in Various Cultures of a Colon

Bacillus.
Solid lines indicate percentage decrease in plate counts; dotted lines, per-
centage increase in stained cells; dot-and-dash lines, percentage decrease in
microscopic counts.

FORMS OF THE COLON BACILLUS 137

variation, and the calcium chloride culture with the highest autoly-
sis rate having also the highest index of variation. This is of course
only a rank order correlation, i.e., the figures for autolysis rate and
degree of variation in cell form vary in the same direction, but not
proportionately. But considering the imperfect nature of the measures
both of rate of autolysis and variation in form, I believe that this
correlation is significant, and offers a real clue as to the nature of the
unusual cell forms observed during the death phase.

It was pointed out in Chapter V that probably the form of bacte-
rial cells is maintained by a certain degree of rigidity of the cell walls.
If this membrane be softened or destroyed locally, it will lead to a
local bulging of the protoplasm, forming lateral or terminal buds or
branches or central bulgings depending upon the localization of the
area of destroyed or altered cell wall; while if the process involves
the entire cell it will give rise to oval and eventually spherical forms.
If autolysis be delayed, the dead cells will tend to maintain their
form even though the death rate be high. But if autolysis is rapid,
and alters the cell wall in the manner suggested, variation in form
will be great even though the death rate is relatively low.

While the differentiation between stained and unstained cells is
quite easily made by inspection through the microscope, there is
not sufficient contrast in the photographic negative when projected
at the magnification used to certainly separate the two classes when
making the tracings. Consequently no comparison can be made of
the variation in form between the stained and unstained cells. Micro-
scopic observation indicates, however, that the majority of the un-
usual forms, as budding, branching, and bulging cells, are not stained.
This, however, does not necessarily indicate that the variation in
form is not due to the same process which leads to the cells becom-
ing stainable; the changes in the cell membrane may occur in an
earlier stage of the process.

CHAPTER X

ON CYTOMORPHOSIS IN BACTERIA

The collection of these data has proven exceedingly tedious
and time-consuming. The observations recorded in the preceding
pages have required measurements of one character or another of
nearly 100,000 cells. For this reason the data are quite incomplete.
Much more work will have to be done, many more species of bac-
teria will have to be studied on a greater variety of media, before
the nature and significance of their morphologic variations can be
clearly understood in all details. But enough has been presented,
I believe, to indicate clearly that the problem includes factors not
recognized by either monomorphists or pleomorphists. The cells of
bacteria do vary in size and form and internal structure, but these
variations occur in a very regular and orderly fashion; they can be
measured, and the quantities obtained by such measurement when
plotted form regular curves. These variations are not confined to
the late stages of growth but occur continually through all stages.
Each character reaches its maximum development in some particular
phase or at some particular point of inflection of the growth curve.
Those factors which cause a variation in the growth rate equally
vary the rate and degree of change in morphologic characters. The
morphologic variations of bacteria are therefore an expression of
the variations in growth rate.

The character of these morphologic variations in the early stages
of growth depends largely upon the age of the cells used for seeding.
If these were embroyonic forms rapidly growing, they continue as
such in the new medium. If, however, they were of a later stage
of development, they may remain dormant for a time and even
continue to change in the same direction as in the old culture from
which they came. Thus in Cultures II and III of B. megatherium
described in Chapter IV, the cells decreased slightly in size during
the first hour of growth; with the diphtheriod organism recorded in
Chapter VI there was an initial increase in size; with the cholera

139

140 MORPHOLOGIC VARIATION

vibrio as described in Chapter VIII there was at the beginning a
slight decrease in size and increase in area-length index. While these
were not in all cases associated with dormancy of the cells as regards
division rate, it seems to me that we are dealing with the same sort
of phenomenon; the cells do not immediately recover from the aging
influence which was present in the old culture.

It would seem from my data that the division of the growth
curve into a lag phase (of accelerating growth), a logarithmic growth
phase, and a resting phase, is not so significant as a division into
a phase of accelerating growth and a phase of negative acceleration
in growth; the so-called logarithmic growth phase when present is
but a long drawn out point of inflection. For the morphologic
variations which occur during the early stages of growth progress
definitely to this "mid-point" (as Pearl designates it) of the growth
cycle, then turn sharply in the opposite direction. The embryonic
forms reach their maximum development just at the beginning of
negative acceleration in growth rate, the mature forms at the end
of growth. The transformation of aged cells to embryonic forms
precedes cell division; the initial period of dormancy then is occu-
pied in part (but only in part) by this transformation. The gradual
acceleration in growth rate is probably best explained by the nature
of the frequency distribution of the inoculated cells with regard to
their capacity for growth, as reflected in their frequency distribu-
tions with regard to those morphologic characters which distinguish
the embryonic forms.

These embryonic forms will vary in their characters with dif-
ferent species of bacteria, but it is apparent from what has been
presented here, as well as from the observations of Clark and Ruehl,
that with most forms, especially the rods, they differ from the
mature forms particularly in increased length and slenderness. The
diphtheroid group are apparently an exception, the embryonic forms
being shorter and more nearly approaching the spherical form. In
all cases these embryonic forms seem to possess a higher affinity
for the basic aniline dyes. In those cases where the young cells
show an increased size there is also apparent an increased varia-
bility in size. In those forms which develop intracellular granules
or other structures these are lacking in the embryonic cells. In

CYTOMORPHOSIS IN BACTERIA 141

the case of the cholera vibrio, which may perhaps be taken as a
type of the spiral organisms, the embryonic forms are characterized
particularly by straightness of their cells; they are bacillarly in form.

The mature or differentiated forms, beginning to develop with
negative acceleration in growth and reaching their maximum at the
end of growth, are just the reverse of the embryonic forms in the
characters enumerated above; in most species the cells are shorter
and more nearly approach the oval or spherical form; they are
less variable in size, more variable in form. In the diphtheriod
group they are longer and more slender. With the cholera vibrio
they are particularly characterized by curvature of the cells. In
all forms they are less deeply stained by the basic dyes. It is of
course obvious that these cells are not differentiated in the sense that
different cells show a great diversity of form and internal structure
as occurs in the differentiated cells of a multicellular organism;
such cannot be the case because the cells are all contained in the
same environment which must be nearly uniform throughout. But
the individual cells do show a differentiation in their internal struc-
ture, many forms developing within the protoplasm spores or granules
of one type or another, especially volutin. Now it is just this
development of internal "paraplasmatic" structures which charac-
terize the differentiated cells of a multicellular organism, and which
are either the result or the cause of their diversified function. In
this sense at least, then, there does occur differentiation in the
mature cells of bacteria.

The senescent forms begin to develop with the initiation of the
death phase. Perhaps because the data are too incomplete their
relationship to this phase of growth is not so clear as in the case
of the other two cell types; but it would seem that they are not so
definitely associated with the rate of death as with the rate of
autolysis during the death phase. They are characterized by in-
creased variation in form of cells and especially by an increasing
proportion of asymmetrical cells. I have already suggested that
this is due to the action of some agent upon the cell wall. These
senescent forms are also characterized by still less affinity for the
basic dyes and finally by the development of an affinity for the
acid dyes.

142 MORPHOLOGIC VARIATION

Throughout the preceding pages I have tried wherever possible
to call attention to the parallelism in the phenomena of growth and
the accompanying cell changes existing between cultures of bacteria
and multicellular organisms. While many biologists have studied
the cell changes of plants and animals which occur during growth,
senescence, and death, their significance has been especially empha-
sized by Minot, who named the process "Cytomorphosis."

In animals this cytomorphosis is characterized, according to
Minot, especially by variations in the ratio of size of the nucleus
to volume of protoplasm. This ratio is high in the embryonic
cells during the period of most active growth, and becomes pro-
gressively lower with increasing age of the animal, the cells be-
coming increasingly larger without any corresponding increase
in size of the nucleus, although, as he points out, these differen-
tiated cells are also characterized by the occurrence of various
structures within the protoplasm; the increase in size of the cell
being to a certain extent at least due to the accumulation of these
granules, vacuoles, and similar structures. Senescence is but a con-
tinuation of the same process, the nucleus continuing to shrink.
The protoplasm may either atrophy or hypertrophy, according to
the nature of the cell, and frequently also still further types of
granules or other structures may appear.

According to Minot, this process is due to intrinsic causes.
Once a cell has been started on the growth cycle, it must neces-
sarily pass through the remaining phases of this cycle. Growth,
senescence, and death are but outward manifestations of this cyto-
morphosis. As he states it, "the increase of protoplasm together
with its differentiation is to be regarded as the explanation (or should
we say cause) of senescence." Further, the process is to a degree
irreversible; "after a cell has progressed and is differentiated a cer-
tain distance, its fate is by so much determined." This he calls
the "law of genetic restriction."

In the light of later knowledge, however, these last statements
can hardly be accepted. The present day tendency is rather to con-
sider the changes which occur in the cells of a growing plant or
animal not so much a cause of growth as a result. They are mor-
phologic manifestations of physiological changes which are occurring

CYTOMORPHOSIS IN BACTERIA 143

in the organism. As an extreme example of this viewpoint, we may
quote Child, who maintains that what we call protoplasm of cells
is but a relatively inert framework about which the chemical pro-
cesses concerned with growth and age and death take place. He
states that, "According to this view the colloid substratum and the
morphologic structure of the organism represent, so to speak, the
sediment of the metabolic process."

Also the widespread occurrence of dedifferentiation in cells, not
generally admitted at the time of Minot's publications, now seems
to be generally accepted. This is especially evident in lower forms
of life, in the process of regeneration after excision of a part or in
the reconstitution of individuals from excised parts. It also seems
to have been established beyond question in the case of tissue cul-
tures. Here conditions are quite comparable with those which ob-
tain in cultures of bacteria. Cells left remaining in plasma tend to
differentiate and finally undergo senescence and death; while if
continually transferred to fresh plasma they may be continually
maintained in a relatively undifferentiated condition. Similarly, as
was shown in Chapter IV, bacteria left in the same medium tend
to revert to the mature form as indicated by shorter length of cells,
but if continually transferred to fresh agar remain in the embryonic
form characterized by long cells.

Child has studied the phenomena of growth and death from a
physiologic standpoint, and has discovered that the physiologic age
of an organism or tissue can be determined by its relative suscepti-
bility to certain poisons, there being a correlation between the degree
of susceptibility and the degree of physiologic youth of the cells
or organisms investigated. Not only is there a difference in the
susceptibility between young and old individuals, but it can be
shown that there is a difference in susceptibility between growing
and resting parts of an organism. Regenerating pieces of planarian
worms are physiologically ydunger than the same tissues when a
part of the whole animal; and similarly small pieces as measured
by their susceptibility have a higher metabolic rate than large ones.
This is very interesting from our standpoint, because, as has been
previously related, Sherman and Albus have discovered that cultures
of bacteria are more susceptible to various injurious agents, as heat

144 MORPHOLOGIC VARIATION

and cold and salt and phenol, than are resting cultures; that is,
there is with bacteria a physiologic youth that can be demonstrated
by the same means as physiologic youth in a growing animal or
plant.

I have also shown previously that cultures which are lightly
seeded develop a higher growth rate and a greater degree of mor-
phologic variation in the embryonic cells than cultures which are
heavily seeded, and have compared this with the higher growth
rate in regenerating tadpoles' tails when only a small stump is left.
It would appear, therefore, that whether measured by degree of
morphologic variation in the cells, by the actual rate of cell division,
or by physiological processes, such as the susceptibility to poisons,
the growth rate both of free populations of cells, and of multicellular
organisms is dependent upon the relative volume of protoplasm which
is in the process of growing.

While there have been made some quantitative studies of the
nucleo-cytoplasmic ratio of cells at different stages of growth in
developing animals, I am not aware of any quantitative investiga-
tions of cytomorphosis in multicellular organisms comparable with
those which I have presented for bacteria, by which the degree of
change in the cells can be directly correlated with the rate of growth.
There is, however, ample indirect evidence that such a correlation
exists. Minot noted that the "mitotic index," i.e., the percentage
of cells in mitosis (serving as a measure of rate of cell division),
steadily decreases with increasing differentiation, and that the rate
of differentiation becomes increasingly slower with age as does also
the growth rate. Child observed that the degree of differentiation
or dedifferentiation of the cells was correlated with the metabolic
rate as determined by susceptibility, which in turn was correlated
with the rate of growth. It would seem, therefore, not improbable
that the transition from embryonic to differentiated cells in multi-
cellular organisms is determined by a change from positive to nega-
tive acceleration in growth rate. A quantitative determination of
such a correlation, however, would present considerable technical
difficulties.

Minot predicted that with further study, cell cycles in one- celled
organisms, protozoa, would be found to be of the same nature as

CYTOMORPHOSIS IN BACTERIA 145

the cytomorphosis which occurs in multicellular organisms; that
with increasing age in protozoa a change in the nucleo-protoplasmic
ratio would be found comparable with that which occurs in differen-
tiating animal and plant cells. While I am not conversant with
the literature, I do not know that this has been done, that is that
the life cycles in protozoa have actually been studied as a cytomor-
phosis, and that any attempt has been made to trace in them a develop-
ment of the same laws which govern the cell changes in the multicellu-
lar organism ; but I believe that in the data which have been presented
in the preceding chapters of this book there is ample evidence that
the cell changes exhibited by bacteria are of this nature, and that
this term "cytomorphosis," with the present state of our knowl-
edge at least, more clearly explains the nature of the variations in
cell structure which are observed, than does the term "life cycle,"
since it does not necessarily involve any idea of multiplicity in forms
of reproduction, and particularly of alternating sexual and non-sexual
phases which generally is implied in the term "life cycle."

My conclusion that the cell changes occurring in cultures of
bacteria are a cytomorphosis of the same kind as that exhibited
by the cells of a multicellular organism is arrived at only by analogy.
No one can state definitely that the growth and cytomorphosis of
a population is governed by laws identical with those which govern
the growth and cytomorphosis of a multicellular plant or animal
until it has been discovered what those laws are. But the pheno-
mena so exactly parallel each other, no matter from what angle they
are viewed, the analogy is so perfect, that we are justified in
excepting this theory as at least a working hypothesis for further in-
vestigation.

The acceptance of this theory demands the acceptance of certain
corollaries. If it be granted that the cells of bacteria undergo a
metamorphosis of the same kind as that exhibited by multicellular
organisms, then it must be granted that to this degree a population
of free one-celled organisms, even though those cells have no connec-
tion other than the common nutrient fluid which bathes them, be-
haves like an individual. There has already been accumulated a
great deal of evidence of other kinds to support the idea that there
is no essential difference, that there can be drawn no hard and

146 MORPHOLOGIC VARIATION

fast line, between populations of one-celled organisms and multi-
cellular individuals; that a higher plant or animal is but a popula-
tion of more highly differentiated cells. But there has been, in the
past at least, a tendency to look upon cell differentiation in multi-
cellular organisms as being the result of some organizing agency
peculiar to such individuals. If, however, we find in cultures of
microorganisms where no such governing agency can be supposed
to exist, a differentiation of cells, even though very primitive, we
are forced to conclude that such is not the case; that the high degree
of organization of higher organisms is a result and not a cause of
the high degree of cell differentiation.

This theory of cytomorphosis in bacteria is not entirely in dis-
agreement with the current theory of monomorphism, which has
always recognized two of the three cell types; the so-called normal
forms which I designate mature or differentiated, and the involution
forms which I call senescent. It differs markedly, however, in the
recognition of a third type, the embryonic cells characteristic of
the stage of accelerating growth which will probably be found in
all species of microorganisms; but just this recognition of the em-
bryonic forms is sufficient to upset the entire theory of monomor-
phism, for it implies that morphologic variations are occurring con-
tinuously at all stages of growth, rather than only during the period
when cultures are becoming aged and dying. It would seem, there-
fore, that the monomorphistic theory without modification can no
longer be considered sound.

On the other hand, this theory of cytomorphosis is not necessarily
opposed to the theory of complex life cycles maintained by the
more recent pleomorphists. The two viewpoints are rather different
ways of looking at the same thing than are they mutually exclusive.
But with acceptance of the theory of cytomorphosis we are no longer
forced to explain every variation encountered as some sort of a mode
of reproduction other than simple cell division. Most of the cell
types described as extraordinary reproductive bodies by the pleomor-
phists, I have encountered only in the death phase, and their occur-
rence is apparently dependent on the rate of autolysis; it would
seem, therefore, more likely that they are cells which are undergoing
retrogressive changes than cells which are embarking upon some

CYTOMORPHOSIS IN BACTERIA 147

new mode of growth. But we have every evidence that in other
organisms modes of reproduction other than simple vegetative di-
vision or budding are also dependent upon the age of the culture.
As Child states, "there is good reason to believe that algae and
fungi may undergo senescence and rejuvenescence like the lower
animals, and that the different forms of reproduction are character-
istic of different stages of the life cycle" (i.e., phases of growth).
It is quite possible then that bacteria do develop during the death
phase some sorts of reproductive bodies that give rise to particular
forms of resting cells; that there does occur, as Mellon claims,
some sort of a reorganization process. But since we know that in
higher organisms along with these reproductive processes there occur
retrogressive changes in the vegetative cells, we are justified in de-
manding conclusive proof that these unusual cells found in the death
phase are really reproductive cells and not simply degenerating
cells. The burden of proof of complex life cycles in bacteria must
be placed upon the shoulders of those men who have declared that
such life cycles exist. So far such proof has not been brought for-
ward.

In connection with these theories of pleomorphism, it should
also be pointed out that every study of life cycles of bacteria should
be accompanied by observations of the rate of growth and a de-
termination of the phase of growth in which each type of cell
change occurs. The age of the culture is of paramount importance
in interpreting the nature of the cell variations which are observed,
and by this I mean physiologic age rather than age in hours. It is
not sufficient to state that a culture is so many hours or days old,
for the rate of growth may be varied by many factors. As has
been shown in the earlier part of this work, during the growth
phase cells may change in size within one or two hours to such a
degree that they would by ordinary standards not be recognized
as the same. One cannot, therefore, safely guess at the phase of
growth; it must be determined by measurement. Many of the
morphologic variations which have hitherto been described, par-
ticularly those indicating a possible mutation in the organisms studied,
have undoubtedly been not true permanent variations but the re-
sult of a comparison by the investigator of cultures which are in

148 MORPHOLOGIC VARIATION

different phases of growth. The diphtheriod organism which I have
studied, for instance, looks almost like a streptococcus in the period
of active growth, while it is a long, granular rod in the resting phase.
Now if this organism had been inoculated into different media, in
one of which it grew slowly and the other rapidly, and these were
compared, it would seem to the observer that the organism had
undergone a pronounced variation, and this would be apparent every
time the organism was subcultured on to these two media, other
conditions being the same. It is, therefore, necessary to determine
the phase of growth in every case where cultures are compared
with the view to establishing mutation.

It should also be pointed out here that those who may attempt
to repeat my experiments must inoculate their cultures Hghtly, if
they will see the degree of morphologic variation in the early stages
of growth which I have described. By the ordinary procedure of
transferring from one agar slant to another, generally a loopful of
the organism is carried over. A culture so inoculated will have an
extremely short period of accelerating growth, during which only
slight change in cell size will be observed. The difference between
the embryonic forms and mature forms is quantitative, and this
difference becomes greater the higher is the actual rate of growth
attained.

This chapter should not be closed without a word or two regard-
ing the relationship of these findings to problems of general bacteri-
ology. The morphologic variations described are but outward mani-
festations of equally profound variations in the physiology of the
organisms, concerning which we are practically in the dark. There
are scattered references in the literature to differences in degree of
virulence, of antigenic value, of susceptibility to antibody action,
or differences in fermentative power, between young and old cul-
tures; but in practically no cases have these variations in physiologic
properties been definitely correlated with phases of growth. It
would seem that this would offer an extremely attractive field for
investigation, especially from the standpoint of infection and im-
munity. Such investigations, however, are difficult to carry out,
because when the cells of bacteria are physiologically young, they
must necessarily be few in number in proportion to the volume of

CYTOMORPHOSIS IN BACTERIA 149

the medium, so few that they are not suitable for most types of ex-
periment. Every attempt to concentrate them, particularly to remove
them from their medium, immediately causes them to rapidly age,
and they are no longer physiologically young cells but mature cells.
For this reason I do not believe that very much experimental work
has so far been done with truly physiologically young bacteria. Some
way must be discovered for either working with small numbers of
cells in the medium in which they are growing, or else concentrating
them in some way that their properties are conserved, before much
can be accomplished in this direction.

APPENDIX

TABLES I TO XXVII

APPENDIX

Table I
Influence of Size of Seeding on Growth Curve of Yeast

Hours of

Number of Yeast Cells per

c.c. of Medium (10% dextrose, 2%

peptone)

mcubation

Flask No. 1

Flask No. 2

Flask No. 3

Flask No. 4

Flask No. 5

0

20,949,750

2,094,975

209,498

20,950

2,095

2

23,809,916

2,223,833

221,500

17,200

2,660

4

37,816,500

3,109,500

212,400

25,300

3,100

6

58,944,000

8,333,000

357,500

20,700

3,500

8

105,568,000

19,683,000

1,033,800

20,300

4,000

10

129,008,000

31,675,000

3,794,000

24,700

4,000

12

161,600,000

54,550,000

11,839,000

41,100

4,800

16

174,720,000

91,544,000

44,848,000

259,000

4,800

20

176,264,000

118,040,000

89,368,000

3,004,200

5,500

24

181,008,000

124,880,000

123,768,000

18,355,200

6,600

36

183,680,000

144,128,000

138,304,000

114,600,000

289,400

48

186,088,000

156,752,000

146,704,000

134,984,000

1,089,600

72

185,160,000

165,648,000

162,240,000

160,000,000

93,264,000

96

195,008,000

173,536,000

176,320,000

165,176,000

169,400,000

144

197,896,000

183,248,000

180,368,000

175,840,000

171,408,000

Table II
Influence of Concentration of Nutrients on Growth Curve of Yeast

Number of Yeast Cells per c.c. of Medium

Hours of
incubation

Flask No. 1

Flask No. 2

Flask No. 3

Flask No. 4

Flask No. 5

10% Dextrose,

5% Dextrose,

2.5% Dextrose,

1.2%, Dextrose,

0.6% Dextrose,

2% Peptone

0.4% Peptone

0.08% Peptone

0.016% Peptone

0.0032% Peptone

0

20,800

21,000

20,500

20,800

18,800

2

22,100

21,400

22,700

22,600

22,300

4

24,200

22.800

26,400

28,400

17,400

6

28,200

77,600

57,200

29,000

29,000

8

50,000

190,300

134,900

61,000

31,200

10

82,300

404,700

224,400

102,000

33,300

12

310,800

1,029,200

375,200

131,800

43,000

16

3,749,600

4,502,200

422,000

191,100

50,200

20

28,241,600

12,016,000

682,400

271.200

56,400

24

62,480,000

18,248,000

1,375,500

272,900

62,200

36

173,280.000

28,825,600

3,402,200

389,700

65,100

48

177,440,000

33,760,000

4,012,200

496,300

92,000

72

215,728.000

35,728,000

5,544,400

500,500

89.500

96

237,400,000

47,050,000

6,472,000

543,600

100,300

153

154

MORPHOLOGIC VARIATION

Table III
Micro-Colony I
Rate of Growth

Minutes

No. of cells

Average length

Sum of lengths

of growth

in colony

of cells-microns

of cells-microns

0

2

3.7

7.5

30

2

4.7

9.4

45

2

4.7

9.4

100

2

5.6

11.2

120

2

6.6

13.1

150

2

8.8

17.5

165

2

9.2

18.5

180

2

10.6

21.3

195

2

13.1

26.3

210

2

12.8

25.6

225

3

12.1

36.3

240

4

9.8

39.4

255

4

11.6

46.3

270

4

17.3

69.4

285

4

16.9

67.5

300

7

11.0

76.9

315

9

12.1

109.4

Table IV
Micro-Colony II
Rate of Growth

Minutes

No. of cells

Average length

Sum of lengths

of growth

in colony

of cells-microns

of cells-microns

0

2

6.6

13.1

120

3

5.6

16.9

180

4

5.5

21.9

210

4

5.1

20.6

225

5

5.2

26.2

240

5

6.1

30.6

270

5

7.5

37.5

280

5

10.0

50.0

295

6

8.9

53.7

310

7

8.9

62.5

325

8

8.0

63.7

APPENDIX

155

Table V
Micro-Colony III
Rate of Growth

Minutes

No. of cells

Average length

Sum of lengths

of growth

in colony

of cells-microns

of cells-microns

0

5

4.6

23.1

30

5

4.9

24.4

75

5

4.3

21.3

105

5

5.1

25.6

120

6

4.9

29.4

135

7

4.9

34.4

150

8

5.4

43.1

165

8

5.8

46.3

180

8

6.9

55.0

195

11

5.8

63.8

210

13

6.2

80.6

220

16

6.7

108.1

Table VI

B. Megatherium, Culture I

Rate of Growth

Hours

No. of cells

No. of spores

of growth

per cu. mm.

per cu. mm.

0

h

90

1

110

n

139

2

169

2^

178

3

196

Sh

209

4

244

^

416

5

1,370

5^

2,080

6

5,920

6i

10,380

7

15,040

156

MORPHOLOGIC VARIATION

Table VI {Continued)

Hours

No. of cells

No. of spores

of growth

per cu. mm.

per cu. mm.

71

18,540

8

21,080

Sh

38,409

9

57,240

9h

79,700

10

63,200

10^

77,900

11

76,800

lU

125,000

12

136,450

13

143,400

14

171,600

15

183,500 .

16

220,500

17

225,500

20

18

242,500

120

19

244,500

420

20

264,400

1,520

21

318,000

8,400

22

258,000

10,800

23

192,000

27,600

24

163,200

38,400

28

156,000

102,000

32

141,600

140,400

36

138,000

181,200

40

121,200

201,000

44

96,000

216,000

48

93,000

234,000

60

78,000

315,000

72

58,000

316,000

APPENDIX

157

Table VII
Megatherium, Culture I
Statistical Constants

Hours

Average

Mode

Median

Standard
deviation
(microns)

Coefficient

of growth

(microns)

(microns)

(microns)

of variation

0

3.4

3.0

3.1

0.80

23

2\

3.9

3.5

3.5

1.38

35

3

5.3

4.0

4.7

2.20

42

3i

6.5

3.5

5.5

3.58

55

4

9.1

3.5-4.0

10.2

5.55

61

4|

7.3

3.5

5.4

4.70

64

5

11.3

11.0-12.0

10.6

6.16

54

5^

19.8

14.5-15.0

16.8

8.67

43

6

13.0

10.5

12.8

4.42

34

6i

15.0

14.0

14.2

5.92

37

7

9.0

7.5

8.4

3.12

34

71

7.2

5.5

6.5

2.89

32

8

5.7

5.0

5.3

1.40

24

8i

5.0

4.0-4.5

4.4

1.87

36

9

4.3

3.5

3.7

1.98

46

9^

4.7

3.5

3.7

1.49

32

10

4.5

4.0

4.0

1.02

23

101

4.0

4.0

3.7

0.90

22

11

3.6

3.5

Z.Z

0.90

25

IH

3.4

3.5

3.1

0.69

20

12

3.4

3.0

3.1

0.80

23

13

3.9

4.0

3.6

0.86

22

14

4.1

3.5-4.0

3.7

0.90

21

15

3.8

4.0

3.6

0.86

22

16

4.3

4.0

4.0

1.06

24

17

4.1

4.0

3.7

0.93

23

18

4.1

4.0

3.8

0.85

21

19

4.0

4.0

3.7

0.80

20

20

3.8

3.5

3.5

0.95

25

21

4.5

4.0.

4.2

1.10

24

22

4.1

4.0

3.8

0.99

24

23

4.6

4.0

4.2

1.15

25

24

4.2

4.0

3.9

0.92

22

28

4.0

4.0

3.8

0.81

20

32

3.6

3.5

3.2

0.86

24

36

3.8

4.0

3.5

0.83

22

158

MORPHOLOGIC VARIATION

Table VII {Continued)

Hours

Average

Mode

Median

Standard
deviation
(microns)

Coefficient

of growth

(microns)

(microns)

(microns)

of variation

40

3.4

3.5

3.1

0.76

22

44

3.2

3.0

2.9

0.79

24

48

3.6

3.5

3.?,

0.80

22

60

3.9

3.5

3.6

1.15

29

72

2.8

2.5

2.4

0.82

29

Table VIII
B. Megatherium
Rate or Growth

Culture II

Culture III

Hours

No. of cells

Hours

No. of cells

of growth

per cu. mm.

of growth

per cu. mm.

1

2,340

1

240

2

3,120

3

2,520

3

6,910

4

3,670

4

8,340

5

4,110

6

10,620

6

5,100

8

11,800

7

5,740

10

12,280

9

6,240

12

15,900

10

6,920

14

18,630

13

8,150

16

21,120

15

10,400

18

30,810

16

13,740

20

43,830

18

16,110

22

52,830

20

28,230

24

54,760

22

36,200

24

43,820

APPENDIX

159

Table IX

B. Megatherium, Culture IT

Statistical Constants

Hours

Average

Mode

IVIedian

Standard
deviation
(microns)

Coefficient

of growth

(microns)

(microns)

(microns)

of variation

0

5.2

5.0

4.8

1.4

27

1

4.9

4.5

4.5

1.2

25

2

6.6

5.5

6.1

1.9

29

3

7.6

6.5

7.1

2.1

27

4

8.5

7.5

7.9

2.7

31

5

7.6

7.0

7.1

2.3

31

6

7.0

6.5

6.4

2.3

31

7

6.8

5.5

6.1

2.2

32

8

6.5

5.5

5.9

2.1

32

9

5.9

5.0-5.5

5.4

1.7

29

10

5.5

5.0

5.2

1.6

31

11

5.1

5.0

4.7

1.3

25

12

5.1

4.5

4.6

1.6

31

13

5.0

4.0

4.4

1.7

34

14

4.6

4.0

4.2

1.4

31

16

4.5

4.0

4.2

1.4

31

18

4.1

3.5

3.6

1.5

37

20

3.6

3.5

3.3

1.1

30

22

3.5

3.0

3.1

1.2

33

24

3.6

3.0

3.1

1.2

34

Table X

B. Megatherium, Culture III

Statistical Constants

Hours

Average

Mode

Median

Standard

Coefficient

of growth

(microns)

(microns)

-

(microns)

deviation
(microns)

of variation

0

5.2

5.0

4.8

1.4

27

1

5.0

4.5

4.6

1.2

24

2

6.9

6.0

6.3

2.0

30

3

9.3

7.5

8.8

2.7

28

4

9.7

8.5

9.0

2.8

29

160

MORPHOLOGIC VARIATION

Table X (Continued)

Hours
of growth

Average
(microns)

Mode
(microns)

Median
(microns)

Standard
deviation
(microns)

Coefficient
of variation

5
6

9.6
9.3

7.5-8.0
9.0

8.9

8.8

2.9

2.8

31
30

7

8.7

8.5

8.2

2.8

33

8
9

6.9

7.3

6.0
6.0

6.4
6.8

2.0
2.2

28
30

10
11

5.9
6.2

5.0
5.5-6.0

5.3

5.8

2.1
2.0

36

32

12

5.1

4.5

4.5

1.8

35

13

4.9

4.0

4.4

1.5

30

14
16

4.6
4.7

4.0
4.0

4.2
4.2

1.4
1.5

30
32

18

4.4

3.5

3.9

1.6

35

20
22

4.4
3.6

3.5
3.S

3.9
3.3

1.5
1.1

33
31

24

3.8

3.5

3.4

1.2

33

Table XI
B. Megatherixjm
Rate of Growth

Culture IV

Culture V

Culture VI

Hours of
growth

No of

Hours of

No. of

Hours of

No. of

cells per

growth

cells per

growth

cells per

cu. mm.

cu. mm.

cu. mm.

0

520

0

48

0

(5)*

1

662

1

59

3

33

2

990

3

225

5

340

4

1,980

4

405

6

1,040

5

2,450

5

1,620

7

1,560

6

3,170

7

2,290

9

2,250

8

4,030

8

2,610

10

2,680

10

4,480

9

3,120

12

2,920

12

4,800

12

4,180

* It was impossible to make accurate counts in this series with samples earUer
than three hours. Since this series was inoculated with 1/ 100th as much material

APPENDIX

161

as Culture IV, which contained 520 cells per cu. mm. at 0 hour, it is assumed that
5 cells per cu. mm. is a fair estimate of the number at the beginning of growth in
Culture VI.

Table XII

B. Megatherium, Culture IV

Statistical Constants

Hours

Average

Mode

Median

Standard

Coefficient

of growth

(microns)

(microns)

(microns)

deviation
(microns)

of variation

0

5.8

5.0

5.3

1.7

28

1

5.6

5.0

5.0

1.8

32

2

5.3

4.5

4.7

1.9

36

4

6.1

5.0

5.6

1.8

30

5

6.0

4.0

5.1

2.5

42

6

7.3

6.0

6.8

2.2

31

8

7.5

7.0

7.0

2.2

30

10

7.0

5.5

6.2

2.3

33

12

7.3

6.0

6.6

2.6

35

Table XIII

B. Megatherium, Culture V

Statistical Constants

Hours

Average

Mode

Median

Standard

Coefficient

of growth

(microns)

(microns)

(microns)

deviation
(microns)

of variation

0

5.8

5.0

5.3

1.7

28

1

4.3

4.0

3.9

1.2

28

3

5.3

4.5

4.7

1.8

34

4

6.3

5.5

5.8

1.7

28

5

8.0

7.0

7.5

2.1

26

7

7.2

6.5

6.6

2.0

28

8

7.1

6.0-6.5

6.6

2.0

27

9

8.3

8.5-9.0

8.0

2.0

24

12

9.0

7.0

8.5

2.6

28

162

MORPHOLOGIC VARIATION

Tablk XIV

B. Megatherium, Culture VI

Statistical Constants

Hours

Average

Mode

Median

Standard

Coefficient

of growth

(microns)

(microns)

(microns)

deviation
(microns)

of variation

0

5.8

5.0

5.3

1.7

28

2

4.5

4.0

4.1

1.0

23

3

5.1

4.5-5.0

4.8

1.2

23

5

7.3

6.0

6.8

2.2

30

6

7.5

8.0

7.1

2.2

29

7

8.6

5.0

7.8

3.4

39

9

8.5

8.5-9.0

8.1

2.3

26

10

7.0

6.0

6.3

2.2

31

12

6.9

6.0

6.3

1.8

26

Table XV
Effect of Concentration of Nutrients on Size of Cells (B. megatherium)

Hours of Growth

Average (microns)

Coefficient
of variation

Full Strength Agar

0

4.4

1.4

32

1

5.2

1.9

37

2

7.2

1.9

27

3

8.9

2.3

26

4

8.8

2.7

31

5

4.6

1.2

27

6

3.3

.8

23

7

3.2

.9

30

8

3.2

.8

25

Half Strength Agar

0

4.4

1.4

32

1

4.9

1.6

32

2

7.2

1.8

25

3

8.6

1.9

23

4

4.2

1.1

27

5

4.5

1.8

40

6

3.4

1.2

35

7

3.3

.7

21

8

2.8

.7

25

APPENDIX

163

Table XV (Continued)

Hours of Growth

Average (microns)

Standard Deviation
(microns)

Coefficient
of variation

Quarter Strength Agar

0

4.4

1.4

32

1

4.9

1.6

30

2

6.5

1.8

28

3

6.2

2.3

37

4

4.1

1.2

30

5

3.7

.9

27

6

3.1

.7

22

7

2.7

.6

24

8

2.5

.7

28

Eighth Strength Agar

0

4.4

1.4

32

1

4.9

2.1

46

2

7.2

2.2

30

3

6.5

2.0

30

4

3.9

1.0

27

5

2.9

.7

22

6

2.7

.6

22

7

2.6

.6

25

8

2.4

.6

27

Sixteenth Strength Agar

0

4.4

1.4

32

1

4.8

1.5

30

2

6.1

1.7

28

3

6.0

1.9

32

4

4.3

1.4

32

5

3.3

1.1

32

6

3.6

1.3

38

7

3.3

1.0

32

8

2.9 .

.8

29

164

MORPHOLOGIC VARIATION

Table XVI
Effect of Age of Parent Culture on Size of Cells (B. megatherium)

Hours of
Growth

Average
(microns)

Standard
Deviation
(microns)

Coefficient
of variation

Parent Culture

0

3.4

1.2

36

1

4.5

1.7

39

2

4.9

1.8

37

3

7.6

2.7

36

4

8.0

2.3

29

5

6.6

1.7

26

6

3.7

.9

25

7

3.6

1.0

28

8

3.1

.9

'28

Two-hour Subculture

0

4.9

1.8

37

3

8.4

2.6

31

4

7.4

1.9

26

5

6.9

2.2

32

6

4.0

1.2

31

8

3.7

1.4

37

Four-hour Subculture

APPENDIX

165

Table XVI (continued)

Hours of
Growth

Average
(microns)

Standard
Deviation
(microns)

Coefficient
of variation

Six-hour SubcuUure

0

3.7

.9

25

1

4.4

1.5

33

2

5.7

1.8

31

3

7.9

1.9

24

4

9.5

2.1

22

5

7.2

2.0

27

6

4.9

1.7

34

7

4.0

1.3

33

8

3.2

.7

23

Eight-hour Subculture

0

3.1

.9

28

1

3.5

1.0

29

2

4.1

1.1

27

3

6.0

1.8

30

4

8.3

2.3

27

5

7.7

2.0

26

6

4.2

1.2

28

7

3.5

1.1

31

Table XVIl
B. coLi, Rate of Growth

Hours

Cells per cu. mm.

0

40,400

3

371,600

6

1,592,000

9

2,466,800

12

3,060,400

18

3,190,000

30

3,577,200

72

3,792,400

96

4,594,400

166

MORPHOLOGIC VARIATION

Table XVIII
B. coLi, Statistical Constants

Standard

Average

Standard

Coefficient

Average

deviation of

Coefficient

Hours

length

deviation

of

area-length

area-length

of

(microns)

(microns)

variation

index

index

variation

0

1.48

0.39

26

33

9

28

3

4.06

2.13

53

13

8

34

6

1.83

0.61

33

11

8

30

9

2.05

0.55

25

30

9

29

12

1.84

0.55

30

29

8

28

18

1.90

0.74

39

31

9

29

30

2.13

0.79

37

29

7

24

48

2.10

0.71

34

30

9

31

96

1.51

0.61

40

26

10

37

Table XIX

Diphtheroid Bacillus

Rate of Growth

Hours of growth

No. of cells per ccm.

0

336,000

3

428,000

6

1,680,000

9

44,480,000

12

122,400,000

15

372,720,000

18

704,320,000

24

787,860,000

30

833,120,000

36

861,520,000

48

858,520,000

APPENDIX

167

Table XX
Diphtheroid Bacillus

Statistical Constants

Average

length

(microns)

Standard

Coefficient

Hours of

Mode

Median

deviation

of

growth

(microns)

(microns)

(microns)

variation

0

4.80

4.00

4.32

1.66

34.7

1

4.28

4.66

4.33

1.60

32.8

2

5.60

4.66

4.86

2.60

46.4

3

5.80

5.33

5.13

2.14

36.8

4

5.46

4.66

4.86

2.20

40.2

5

4.80

4.66

4.20

2.60

54.2

6

4.86

4.00

4.33

2.06

42.4

7

5.40

5.33

4.86

2.40

44.4

8

5.33

4.66

4.73

3.14

58.8

9

4.73

4.66

4.33

1.06

22.5

10

4.46

4.33

4.06

1.14

25.4

11

3.93

4.00

3.53

1.06

27.1

12

3.33

3.00

3.40

1.06

31.4

13

4.00

4.00

3.60

1.60

40.0

14

3.20

2.66

2.73

1.14

35.4

15

3.80

3.33

3.33

1.14

29.8

16

3.66

3.33

3.20

1.40

39.0

17

4.33

4.00

3.80

1.54

35.5

18

4.06

3.33

3.53

1.34

32.8

21

4.13

4.00

3.73

1.34

32.3

24

3.93

3.33

3.40

1.54

39.0

28

4.73

4.66

4.66

1.66

35.2

32

4.93

4.66

4.40

1.66

33.8

36

4.60

4.00

4.00

1.80

39.7

40

4.80

4.00

4.26

1.66

34.7

44

5.33

4.66

4.73

1.80

33.8

48

5.46

4.00

4.60

2.54

46.3

168

MORPHOLOGIC VARIATION

Table XXI
Diphtheroid Bacillus

Hours of growth

Granules per cell

Coeff. of Correlation
granules with length

0

1.09

83.3 + 4.6

3

0.71

78.4±1.8

6

0.16

46.1±3.6

9

0.10

17.3±6.5

12

0.15

42.6 + 3.9

15

0.21

54.7 + 4.7

18

0.21

38.4+5.7

24

0.29

64.6 + 3.9

36

0.42

56.0 + 3.3

48

0.98

61.8 + 2.9

Table XXII

b. cohaerens

Rate of Spore-formation

Hours of

Heavy inoculum

Light inoculum

Heavy inoculum Light inoculum

Full strength

Full strength

Dilute

Dilute

growth

medium

medium

medium

medium

2

98% spores

98% spores

98% spores

98% spores

3

92

82

79

72

4

15

13

27

22

5

5

1

19

7

6

1

0

3

2

8

0

0

1

0

11

0

0

4

0

15

1

0

5

0

21

3

1

22

7

28

7

5

42

31

36

20

14

50

43

48

48

46

75

66

72

65

56

92

84

APPENDIX

169

Table XXIII
Cholera Vibrio
Rate of Growth

Hour

Cells per cu. mm.

0

2,506

3

2,780

5

3,230

7

21,560

9

194,320

11

725,120

14

2,402,900

18

3,333,200

24

3,465,600

36

4,428,800

48

6,480,800

72

5,964,800

120

5,510,400

168

2,757,200

240

585,000

Table XXIV

Cholera Vibrio

Area-Length Index

Hour

Mean

Median

IVIode

Standard
deviations

Coefficient of
variability

0

35.70

28.67

22.5

19.73

55.30

3

40.40

36.58

22.5

20.92

51.78

6

31.45

29.85

27.5

20.15

64.09

9

33.46

31.51

32.5

9.30

27.79

12

28.50

29.20

27.5

9.49

33.31

18

21.20

21.06

22.5

5.99

28.25

24

24.96

22.40

22.5

11.09

44.43

36

24.20

19.25

17.5

16.21

66.98

48

37.70

27.99

17.5

19.74

52.36

72

36.60

31.40

17.5

20.62

56.34

96

34.16

34.88

30.0

21.13

61.86

120

39.20

33.93

27.5

20.49

52.27

168

34.31

35.39

27.5

16.03

46.72

240

34.84

30.48

22.5

18.89

54.23

170

MORPHOLOGIC VARIATION

Table XXV
Cholera Vibrio

Hours

Average length

Average index

Asymmetrical

of growth

of cells

of curvature

cells

0

2.22m

5.87

20.5%

3

2.12

3.84

10.5

6

2.57

5.82

7.5

9

2.85

8.22

6.0

12

2.75

8.96

1.0

18

2.86

8.88

2.5

24

2.22

7.82

5.5

36

2.61

5.58

1.5

48

2.78

6.77

26.5

72

2.12

6.46

28.5

96

2.23

6.58

45.5

120

2.03

6.87

31.5

168

1.86

6.46

37.5

240

1.99

4.10

37.0

Table XXVI
Colon Bacillus; Rate of Death

Days of growth

Plate counts
Cells per ccm.

Per cent
stained cells

Microscopic counts
Cells per ccm.

alkaline

1

2,790,000,000

0.0

2,555,000,000

2

3,650,000,000

3.7

3,820,000,000

3

3,200,000,000

5.1

4,760,000,000

5

2,265,000,000

8.8

5,350,000,000

7

1,700,000,000

11.7

5,500,000,000

10

760,000,000

12.8

5,760,000,000

14

54,000,000

20.1

5,670,000,000

19

8,700,000

23.6

5,600,000,000

25

1,610,000

56.7

5,380,000,000

1,090,000,000

2,240,000,000

890,000,000

322,500,000

0.0

4.0

6.5

11.2

2,900,000,000
4,640,000,000
5,590,000,000
6,050,000,000

APPENDIX

171

Table XXVI (continued)

Days of growth

Plate counts
Cells per ccm.

Per cent
stained cells

Microscopic counts
Cells per ccm.

7

226,000,000

27.1

6,190,000,000

10

175,000,000

36.7

5,610,000,000

14

112,200,000

45.2

5,510,000,000

19

106,000,000

57.0

4,600,000,000

25

100,400,000

59.6

3,420,000,000

SODIUM CHLORIDE

1

1,540,000,000

1.5

1,320,000,000

2

1,660,000,000

3.1

1,772,000,000

3

1,750,000,000

10.3

1,957,000,000

5

645,000,000

15.8

1,858,000,000

7

230,000,000

21.1

1,611,000,000

10

24,100,000

45.9

1,410,000,000

14

7,600,000

62.8

1,240,000,000

20

1,120,000

82.0

1,075,000,000

ACID

1

2,370.000.000

2.2

2,630,000,000

2

2,675,000,000

8.8

3,200,000,000

3

3,400,000,000

13.5

3,026,000,000

5

4,700,000,000

32.6

4,650,000,000

7

1,160,000,000

39.4

5,011,000,000

10

325,000,000

47.8

6,616,000,000

14

85,000,000

58.2

5,420,000,000

19

44,800,000

71.1

5,160,000,000

25

18,200,000

81.8

4,095,000,000

CALCIUM CHLORIDE

1

1,140,000,000

12.1

1,330,000,000

2

1,435,000,000

17.0

1,430,000,000

3

1,050,000,000

21.4

1,495,000,000

5

680,000,000

39.6

1,530,000,000

7

600,000,000

43.0

2,270,000,000

10

398,000,000

55.5

2,295,000,000

14

183,000,000

78.0

3,460,000,000

19

100,000,000

84.0

2,600,000,000

25

32,500,000

92.2

2,130,000,000

172

MORPHOLOGIC VARIATION

Table XXVII

Colon Bacillus; Incidence of Various Cell

Types in the Death Phase

Davs of Growth

1

2

3

5

7

10

14

19

25

Mean

Cell
class

.

ALKALINE

I

54.5%

51.0%

60.0%

45.0%

61.5%

60.0%

66.5%

53.5%

48.0%

55.55

II

15.5

8.5

8.5

5.0

9.0

7.5

3.0

5.5

8.0

7.83

III

2.0

23.0

20.5

39.5

11.5

15.0

17.5

22.0

29.0

20.00

IV

26.5

9.5

4.0

2.0

11,5

13.0

6.5

4.5

3.5

9.00

V

1.5

1.0

0.5

2.0

2.0

1.0

1.0

4.5

3.5

1.89

VI

0.0

5.0

3.0

2.0

2.0

2.5

2.0

6.5

4.0

3.00

VII

0.0

2.0

4.0

4.5

2.5

1.0

3.5

2.5

3.0

2.55

VIII

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.5

0.05

IX

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.5

0.05

X

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.11

Index

of vari-

ation

67.08

67.61

61.80

68.30

63.18

59.31

58.55

86.07

77.45

67.71

Cell

class

NEUTRAL

I

42.0

48.5

45.0

57.5

53.5

52.5

45.5

46.0

44.5

48.33

11

11.5

12.5

16.0

11.0

25.0

15.5

10.0

13.0

11.5

14.00

III

40.0

15.5

17.5

19.5

10.5

15.5

18.0

17.5

26.0

20.00

IV

0.5

7.5

4.0

6.0

2.5

5.5

7.5

10.0

5.0

5.39

V

0.5

1.0

3.5

2.0

4.0

0.5

3.5

3.0

4.0

2.47

VI

1.5

11.0

8.0

3.0

1.0

5.5

9.0

5.5

5.5

5.55

VII

3.5

1.0

3.5

1.0

1.0

3.5

3.5

2.5

2.0

2.39

VIII

0.5

0.5

0.0

0.0

2.0

0.5

1.0

2.0

1.5

0.89

IX

0.0

0.5

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.05

X

0.0

2.0

2.5

0.0

0.5

1.0

2.0

0.5

0.0

0.90

Index

of vari-

ation

64.37

87.82

115.27

55.98

73.17

86.03

119.25

95.41

80.70

87.56

Cell

class

SOD

JUM CH

LORIDE

I

62.5

28.5

44.5

46.0

57.0

44.5

60.0

42.0

48.12

II

16.0

30.0

28.5

19.5

25.5

31.5

19.0

37.5

25.93

III

14.5

12.0

10.5

15.0

4.5

8.5

9.0

4.5

9.81

IV

2.5

0.0

1.0

3.0

1.0

3.5

6.0

11.0

3.50

V

2.0

5.0

4.0

7.5

4.0

3.5

4.0

2.5

4.06

APPENDIX

173

Table XXVII (continued)

Days of Growth

1

2

3

5

7

10

14

19

25

Mean

VI

1.5

10.0

4.5

3.5

3.0

4.5

1.0

2.0

3.75

VII

0.0

1.0

0.5

2.0

0.5

0.0

0.5

0.0

0.56

VIII

1.0

11.0

4.0

3.0

4.5

3.0

0.5

0.0

3.37

IX

0.0

0.5

1.0

0.0

0.0

0.0

0.0

0.0

0.19

X

0.0

2.0

1.5

0.5

0.0

1.0

0.0

0.5

0.69

Index

of vari-

ation

50.48

181.55

117.79

96.34

80.24

92.94

55.58

69.18

93.01

Cell

class

ACID

I

51.0%

46.0%

36.0%

43.0%

45.0%

47.0%

47.5%

36.0%

35.0%

42.94%

II

21.0

23.0

25.5

20.0

15.0

15.0

13.5

11.5

11.5

17.33

III

5.5

10.5

18.5

12.5

13.5

18.0

11.5

22.5

32.0

16.05

IV

14.0

12.5

11.0

10.0

13.0

8.0

10.5

15.0

3.0

10.77

V

3.0

2.5

1.5

7.0

2.5

5.0

7.5

6.0

3.5

4.50

VI

0.5

4.5

2.5

3.5

4.5

2.5

5.5

2.0

3.0

3.17

VII

3.0

0.0

4.0

2.5

4.0

2.0

2.5

5.5

12.0

3.94

VIII

2.0

0.5

0.5

1.0

1.0

1.0

1.0

0.5

0.0

0.83

IX

0.0

0.5

0.0

0.5

1.0

0.5

0.0

0.5

0.0

0.33

X

0.0

0.0

0.5

0.0

0.5

1.0

0.5

0.5

0.0

0.33

Index

of vari-

ation

82.73

75.10

88.72

91.12

106.37

95.69

95.81

109.76

85.73 .

93.43

Cell

class

CALCIUM CHLORIDE

I

48.0

38.0

42.0

41.5

42.0

32.0

39.0

35.0

42.0

39.94

II

23.0

23.5

23.5

21.0

25.0

8.0

17.5

10.0

10.0

17.94

III

7.0

6.5

5.0

11.0

8.0

24.5

8.0

24.5

24.0

13.17

IV

1.0

5.0

7.0

1.5

1.5

2.0

4.0

3.0

2.0

3.00

V

10.5

14.5

9.5

13.5

11.5

11.0

9.0

8.5

6.5

10.50

VI

7.5

4.5

9.0

6.5

3.0

5.0

6.0

4.5

5.5

5.72

VII

2.5

5.0

0.5

1.5

1.0

3.5

3.0

3.5

2.0

2.50

VIII

0.0

2.0

0.5

1.5

1.0

4.5

4.0

2.5

3.5

2.17

IX

0.0

0.5

0.5

1.0

5.5

8.5

8.0

6.5

3.5

3.78

X

0.5

0.5

2.5

1.0 -

1.5

1.0

1.5

2.0

1.0

1.29

Index

of vari-

ation

88.06

123.78

127.97

118.30

161.35

226.02

220.21

200.83

148.63

157.25

APPENDIX

REFERENCES

APPENDIX 177

REFERENCES

Albert, H. 1921. Variations in the Morphology of the Diphtheria
Bacillus due to Age. Abstr. of Bact., vol. 5, p. 15.

Albert, R. & W. 1901. Chemische Vorgangen in der Abgetoteten
Hefezelle, Centralbl. f. Bakt., II, Orig., vol. 7. p. 737.

Almquist, E. 1908. Studien ueber das Verhalten einiger pathogenen
Mikroorganismen. Centralbl. f. Bakt., I. Orig., vol. 48, p. 175.
1911. Studien ueber filtrierbare Forman in Typhuskulturen. Cen-
tralbl. f. Bakt., I. Orig., vol. 60, p. 167.

1917. Wuchsformen, Fruktification, und Variationen der
Typhusbakterie. Zietschr. f. Hyg., vol. 83, p. 1.

1922. Variation and Life Cycles of Pathogenic Bacteria. Jour.
Inf. Dis., vol. 31, p. 483.

1924. Investigations on Bacterial Hybrids. Jour. Inf. Dis., vol.
35. p. 341.

1925. Biologische Forschungen ueber die Bakterien. Stockholm,
P. A. Norstedt & Soner. pp. 70.

Avery, O. T. and Morgan, M.J. 1924. The Occurrence of Peroxide

in Cultures of Pneumococcus. Jour. Exp. Med., vol. 39, p. 275.
Balls, A. K. and Brown, j. B. 1925. Studies in Yeast Metabolism.

Jour. Biol. Chem., vol. 62, p. 789.
Barber, M. A. 1908. The Rate of Multiplication of B. coli at

Different Temperatures. Jour. Inf. Dis., vol. 5, p. 379.
Benians, T. H. C. 1916. Relief Staining for Bacteria and Spiro-

choets. Brit. Med. Jour., vol. 2, p. 722.
Bergstrand, H. 1919. Ueber sogenannte Corynebakterien und ihre

Verwandten nebst Bemerkungen ueber Bakterien im Allgemein-

en. Acta Medica Scandinavica, vol. 52, p. m 1.

1921. On the Structure of Bacteria. Acta Medica Scandinavica.

vol. 55, p. 529.

1923. Further Studies on the Morphology of Bacteria. Jour.
Bact., vol. 8, p. 365.

Breed, R. S., and Brew, J. D. 1916. Counting Bacteria by Means
of the Microscope. N.Y. Agr. Exp. Sta., Tech. Bui. No. 49.

Brown, A. 1905. Influences Regulating the Reproductive Func-
tions of Saccharomyces cerevisiae. Jour. Chem. Soc. Lend.,
vol. 87, p. 1395.

178 MORPHOLOGIC VARIATION

Buchanan, R. E. 1918. Life Phases in a Bacterial Culture. Jour.
Inf. Dis., vol. 23, p. 109.

Burke, G. S. 1923. Studies on the Thermal Death Time of Spores
of Clostridium botulinum. II. Differential Staining of Living and
Dead Spores. Jour. Inf. Dis., vol. 32, p. 433.

Burke, V., Sprague A., and Barnes, L. 1925. Dormancy in Bac-
teria. Jour. Inf. Dis., vol. 36, p. 555.

Carlson, T. 1917. Ueber Geschwindigkeit und Grosse der Hefever-
mehrung, Biochem. Zietschr. vol. 57, p. 313.

Chesney, A. 1916. The Latent Period in the Growth of Bac-
teria Jour. Exp. Med., vol. 24, p. 387.

Chick, H. 1912. The Bactericidal properties of Blood Serum.
Jour. Hygiene, vol. 12, p. 414.

Child, C. M. 1915. Sensecence and Rejuvenescence. University
of Chicago Press, Chicago, 111.

Clark, N. A. 1922. The Rate of Formation and the Yield of Yeast
in Wort. Jour. Physical Chem., vol. 26, p. 42.

Clark, P. F., and Ruehl, W. H. 1919. Morphologic Changes Dur-
ing the Growth of Bacteria. Jour. Bact., vol. 4, p. 615.

Conradi, H. and Kurpjuweit, O. 1905. Ueber spontane Wachs-
tumshemmung der Bakterien infolge von Selbstvergiftung,
Mlinch. Med. Woch., 1905, p. 1761, p. 2164, p. 2228.

Coplans, M. 1909. Influences Affecting the Growth of Micro-
organisms— Latency; Inhibition; Mass Action. Jour. Path, and
Bact, vol. 14, p. 1.

Curran, H. R. 1925. The Growth of Bacteria. Thesis, Cornell
University, Ithaca, N.Y.

Cutler, D. W., and Crump, L. M. 1925. The Influence of Washing
upon the Reproductive Rate of Colpidium colpoda, Bio-
chem. Jour., vol. 19. p. 450.

Durbin, M. L. 1909. An Analysis of the Rate of Regeneration
throughout the Regenerative Process. Jour. Exp. Zool., vol. 7,
p. 397.

Eberson, F. 1918. A Bacteriological Study of the Diphtheroid Or-
ganisms with special reference to Hodgkin's Disease. Jour. Inf.
Dis., vol. 23, p. 1.

APPENDIX 179

Eijkman, C. 1904. Ueber Thermolabile Stoffwechselprodukte als
Ursache der naturlichen Wachstumshemmung der Mikroorganis-
men. Centralbl. f. Bakt., I. Orig., vol. 37, p. 436, p. 445.

1906. Ueber Natiirliche Wachstumshemmung der Bakter-
ien. Centralbl. f. Bakt., I., Orig., vol. 41, p. 367.

1907. Ueber die Ursache der Wachstumshemmung in Bakterien-
kulturen. Deutsch. Med. Woch., vol. 33, p. 265.

Eisenberg, P. 1923. Theorie der Bakterienfarbung. In Kraus and

Uhlenhuth, Handb. der Mikrobiol. Technik. Urban and Schwar-

zenberg, Berlin.
Ellis, M. M. 1909. The Relation of the Amount of Tail Regenerated

to the Amount Removed in Tadpoles of Rana claniitans. Jour.

Exp. Zool. vol. 7, p. 421.
Enderlein, G. 1925. Bakterien-Cyclogenie. Berlin und Leipzig.

W. de Gruyter & Co., pp. 390.
Ernst, P. 1889. Ueber Kern und Sporenbildung in Bakerien.

Zeitschr. f. Hyg., vol. 5, p. 428.
Ernst, P. 1921. Adsorptionserscheinungen an Blutkorperchen und

Bakterien. Ziegler's Beitrage zur Path. Anat. vol. 69, p. 152.
Eraser, O. G. 1920. The Action of Methylene Blue and Certain

Other Dyes on Living and Dead Yeast. Jour. Physical Chem.,

vol. 24, p. 741.
Frobisher, M. 1926. Relation of Surface Tension to Bacterial Phe-
nomena. Jour. Inf. Dis., vol. 38, p. 66.
Fuhrmann, F. 1908. Entwicklungszyklen bei Bakterien. Beihefte.

z. Botan. Centralbl. vol. 22>, (Abt. I), p. 1.

1926. Einfiihrung in die Grundlagen der Technischen Mykologie.

Zweite Auflage. Gustav Fischer, Jena. pp. 55-59.
Fulmer, E. I., and Buchanan, R. E. 1923. Studies on Toxicity. I.

Criteria of Toxicity Toward Microorganisms. Jour. Gen. Phys.

vol. 6, p. 75.
Gardner, A. D. 1925. The Growth of Branching Forms of Bacilli

("Three-point Multiplication"). Jour Path, and Bact., vol. 28.

p. 189.
Graham-Smith, G. S. 1920. The Behavior of Bacteria in Fluid Cul-
tures as Indicated by Daily Estimates of the Numbers of Living

Organisms. Jour. Hyg., vol. 19, p. 131.

180 MORPHOLOGIC VARIATION

Greenleaf, W. E. 1924. Influence of Volume of Culture Medium

and Cell Proximity on the Rate of Reproduction of Protozoa.

Proc. Soc. Exp. Biol, and Med., vol. 21, p. 405.
Hajos, K. 1922. Beitrage zur Frage der Wachstumhemmenden Wir-

kung von Bouillonkulturen. Centralbl. f. Bakt., I. Orig., vol. SS,

p. 583.
Harris, W. H., and Wade, H. W. 1915. The Widespread Distribu-
tion of Diphtheroids and their Occurrence in Various Lesions of

Human Tissues. Jour. Exp. Med., vol. 21, p. 493.
Henrici, A. T. 1923. Differential Counting of Living and Dead

Bacteria. Proc. Soc. Exp. Biol, and Med., vol. 20, p. 293.
Hoag. L. 1907. Organism ^'X," Probably of the Corynebacterium

Group, etc. Boston Med. and Surg. Jour., vol. 157, p. 10.
Hort, E. C. 1915. Epidemic Cerebrospinal Fever: The Place of the

Meningococcus in its Etiology. Brit. Med. Jour., vol. 1, p. 541,

p. 715; 1916, vol. 1, p. 156, vol. 2, p. 522.

1917. a. The Meningococcus of Weichselbaum. Brit. Med.

Jour., vol. 2, p. 373.

1917. b. Morphologic Studies in the Life Histories of Bacteria-

Proc. Roy. Soc. Lond., Ser. B., vol. 89, p. 468.

1920. The Reproduction of Aerobic Bacteria. Jour. Hyg., vol.

18, p. 369.
Jennings, H. S. 1920. Life and Death; Heredity and Evolution in

Unicellular Organisms. Boston, Gorham Press, pp. 233.
Kauffman, F. 1926. Beziehungen zwischen dem d'Herelleschen

Lysin, dem Antilysin, und den "Autotoxinen" (Conradi-Kurpju-

weit). Zeitschr. f. Hyg., vol. 106, p. 239.
Koser, S. A., and Mills, J. H. 1925. Differential Staining of Living

and Dead Bacterial Spores. Jour. Bact., vol. 10, p. 25.
Lane-Claypon, J. E. 1909. Multiplication of Bacteria and the In-
fluence of Temperature and some other Conditions thereon. Jour.

Hyg., vol. 9, p. 239.
Ledingham, J. C, and Penfold, W. J. 1914. Mathematical Analysis

of the Lag Phase in Bacterial Growth. Jour. Hyg., vol. 14, p.

242.
Lohnis, F., and Smith, N. R. 1916. Life Cycles of the Bacteria.

Jour. Agr. Res., vol. 6, p. 675.

APPENDIX 181

Lohnis, F. 1921. Studies upon the Life Cycles of Bacteria. Me-
moirs, Nat. Acad. Sci., vol. 16, 2nd Memoir, pp. 252.

Manteufel. 1906. Untersuchungen iiber die "Autotoxine" (Conradi)
und ihre Bedeutung als Ursache der Wachstumshemmung in Bak-
terienkulturen. Berl. Klin. Woch., p. 313.

McKendrick, A. G., and Pai, M. K. 1911. The Rate of Multiplica-
tion of Microorganisms; a Mathematical Study. Proc. Roy.
Soc. Edinburgh, vol. 31, 649.

M'Leod, J. W., and Govenlock, P. 1921. Production of Bactericidin
by Microorganisms. Lancet, vol. 1, p. 900.

M'Leod, J. W., and Gordon, J. 1922. a. On the Production of
Peroxides by Pneumococci and other Bacteria. Jour. Path, and
Bact., vol. 25, p. 139.

1922. b. The Production of Peroxide of Hydrogen by Bacteria.
Biochem. Jour., vol. 16, p. 499.

Mellon, R. R. 1917. A study of the Diphtheroid Group of Or-
ganisms, with Special Reference to their Relation to the Strepto-
cocci. Jour. Bact., vol. 2, p. 81, p. 269, p. 447.

1919. A Contribution to the Bacteriology of a Fusospirillary
Organism with Special Reference to its Life History. Jour. Bact.,
vol. 2, p. 505.

1920. The Life Cycle Changes of the so-called B. Hodgkini, and
their Relation to Mutation Changes in this Species. Jour. Med.
Res., vol. 42, p. 61.

1925. Studies in Microbic Heredity. I. Observations on a Primi-
tive Form of Sexuality (Zygospore Formation) in the Colon-
typhoid Group. Jour. Bact., vol. 10, p. 481.

1925. II. The Sexual Cycle of B. coli in Relation to the Origin
of Variants with Special Reference to Neisser and Massini's B.
coH mutabile. Jour. Bact., vol. 10, p. 579.

1926. HI. The Hereditary Origin of Group and Specific Ag-
glutinogens among the Colon-alkaligenes Organisms. Jour. Im-
munol, vol. 11, p. 139.

1926. V. The Biogenetic Law of Haeckel and the Origin of
Heterogeneity within Pure Lines of Bacteria. Jour. Bact., vol. 11,
p. 203.

182 MORPHOLOGIC VARIATION

1926. VI. The Infective and Taxonomic Significance of a Newly
Described Ascospore Stage for the Fungi of Blastomycosis. Jour.
Bact., vol. 11, p. 229.

1926. VII. Observations on the Genetic Origin of the Several
Types of Fungi Found in the Lesions of Blastomycosis Hominis.
Jour. Bact., vol. 11, p. 419.

1926. VIII. The Infectivity and Virulence of a Filterable Phase
in the Life History of B. fusiformis and Related Organisms. Jour.
Bact., vol. 12, p. 279.

1926. IX. Observations on the Biologic Origin and Physico-
chemical Nature of Inagglutinability with Freshly Isolated Ty-
phoid Bacilli. Jour. Immunol., vol. 12, p. 331.
1926. X. The Agglutinin-Absorption Reaction as Related to
the Newer Biology of Bacteria, with Special Reference to the
Nature of Spore Formation. Jour. Immunol., vol. 12, p. 355.

1926. XI. The Genetic Origin of Staphylococcus Albus and
Aureus from Common Ancestral Strains. Jour. Bact., vol. 12, p.
409.

1927. XII. Microbic Dissociation in Vivo as Illustrated by a
case of Subacute Septicopyaemia. General Summary of the Series.
Jour. Bact., vol. 13, p. 99.

Meyer, A. 1912. Die Zelle der Bakterien. Jena. Gustav Fischer,

pp. 285.
Migula, W. In Lafar F.: Handbuch der Technischen Mykologie,

Band I. Gustav Fischer, Jena, 1904-1907.
Minot, C. S. 1908. The Problem of Age, Growth and Death, G. P.

Putnam's Sons, New York, p. 280.

1913. Modern Problems of Biology. P. Blakiston's Sons, Phila-
delphia.
Minoux, M. 1927. Considerations generales sur la Forme chez les

Bacteries. Travail du Lab. de Microbiol, de la Faculte de Phar-

mecie de Nancy.
Morgan, M. J., and Avery, O. T. 1924. Inhibitory Substances in

Pneumococcus Cultures, Jour. Exp. Med., vol. 39, p. 335.
Morse, M. E. 1912. A Study of the Diphtheria Group of Organisms

by the Biometrical Method. Jour. Inf. Dis., vol. 11, p. 253.

APPENDIX 183

Miiller, M. 1895. Ueber den Einfluss von Fieber Temperaturcn aui
die Wachstumsgeschwindigkeit und die Virulenz des Typhus Ba-
cillus. Zeitschr. f. Hyg., vol. 20, p. 245.

Oebius, R. 1906. Ueber Spontane Wachstumshemmung der Bak-
terien auf kiinstlichen Nahrboden. Med. Klin. vol. 2, p. 598.

Otto, R., and Munter, H. 1923. Weitere Untersuchungcn zum
d'Herelleschen Phanomen. Zeitschr, f. Hyg., vol. 100, p. 402.

Paravincini, E. 1918. Zur Frage des Zellkernes der Bakterien. Cen-
tralbl. f. Bakt., II. Orig., vol. 49, p. 97.

Pearl, R. 1925. The Biology of Population Growth. Alfred A.
Knopf, New York, pp. 260.

Penfold, W. J. 1914. On the Nature of Bacterial Lag. Jour. Hyg.,
vol. 14, p. 215.

Penfold, W. J., and Norris, D. 1913. The Relation of the Concen-
tration of Food Supply to the Generation-Times of Bacteria.
Jour. Hyg., vol. 12, p. 527.

Peskett, G. L. 1924. Allelocatalysis and the Growth of Yeast. Bio-
chem. Jour., vol. 18, p. 866.

1925. Studies on the Growth of Yeasts. I. The Influence of
Volume of Culture Medium Employed. Biochem. Jour., vol. 19,
p. 464.

1925. II. A Further Note on Allelocatalysis. Biochem. Jour.,
vol. 19, p. 474.

Rahn, O. 1906. Ueber den Einfluss der Stoffwechselprodukte auf
das Wachstum der Bakterien. Centralbl. f. Bakt., II. Orig., vol.
16, p. 417, p. 609.

Robertson, T. B. 1923. The Chemical Basis of Growth and Senes-
cence. J. B. Lippincott, Philadelphia, pp. 389.

Roily, 1906. Experimentelle Untersuchungen uber das biologische
Verhalten der Bakterien im Dickdarm. Deutsch. Med. Woch. v.
32, p. 1733.

Rogers, L. A., and Whittier,^ E. O. 1927. Limiting Factors in the
Lactic Fermentation. Jour. Bact., vol. 13, p. 10.

Salter, R. C. 1919. Observations on the Rate of Growth of B. coli.
Jour. Inf. Dis., vol. 24, p. 260.

Scales, F. M. 1921. Induced Morphologic Variation in B. coli.
Jour. Inf. Dis., vol. 29, p. 591.

184 MORPHOLOGIC VARIATION

Seiffert, W. 1922. Vergleichende Farbeversuche an Lebenden und
Toten Bacterien. Centralbl. f. Bakt., I. Orig., vol. SS, p. 151.

Sherman, J. M., Holm, G. E., and Albus, W. R. 1922. Salt Effects
in Bacterial Growth. III. Salt Effects in Relation to the Lag
Period and Velocity of Growth. Jour. Bact., vol. 7, p. 583.

Sherman, J. M., and Albus, W. R. 1923. Physiological Youth in
Bacteria. Jour. Bact., vol. 8, p. 127.

1924. The Function of Lag in Bacterial Cultures. Jour. Bact.,
vol. 9, p. 303.

1924. Is there an Allelocatalytic Effect on the Growth of Bac-
terial Cultures? Abstr. Bact., vol. 8, p. 6.

Slator, A. 1913. The Rate of Fermentation by Growing Yeast
Cells. Biochem. Jour., vol. 7, p. 197.

1917. A Note on the Lag Phase in the Growth of Micro-
organisms. Jour. Hyg., vol. 16, p. 100.

1918. Some Observations on Yeast Growth. Biochem. Jour,
vol. 12, p. 248.

1919. Yeast Growth and Alcoholic Fermentation by Yeasts. Jour.
Soc. Chem. Ind., vol. 3S, p. 391.

Stearn, E. W., and Stearn, A. E. 1924. The Chemical Mechanism
of Bacterial Behavior. I. Behavior towards Dyes. Jour. Bact.,
vol. 9, p. 463.

Stewart, F. H. 1926. Mendelian Variation in the Paracolon-Muta-
bile-Colon Group, and the Application of Mendel's Principle to
the Theory of Acquired Virulence. Jour. Hyg., vol. 25, p. 237.

Taliaferro, W. H., and Taliaferro, L. G. 1922. The Resistance of
Different Hosts to Experimental Trypanosome Infections with
Special Reference to a New Method of Measuring this Resistance.
Am. Jour, of Hyg., vol. 2, p. 264.

Tanner, I. W., and Wallace, G. I. 1924. Relation of Temperature
to the Growth of Thermophilic Bacteria. Jour. Bact., vol. 10, p.
421.

Thompson, D. W. 1917. On Growth and Form. Cambridge Uni-
versity Press, pp. 793.

Topley, W. W. C, and Ayrton, J. 1923. The Segregation of Bio-
logical Factors in B. enteritidis. Jour. Hyg., vol. 22, p. 305.

APPENDIX 185

Von Niesson. 1926. Bakteriogenetisches. Centralbl. F., Bakt., II.

Orig., vol. 66, p. 321.
Williams, A. In Park and Williams. Pathogenic Microorganisms,

1924. Lea and Febiger, New York, p. 35.
Wilson, G. S. 1922. The Proportion of Viable Bacteria in Young

Cultures, with Especial Reference to the Technique Employed in

Counting. Jour. Bact., vol. 7, p. 405.
Zikes, H. 1919. a. Ueber den Einfluss der Konzentration der

WUrze auf die Biologic der Hefe. Centralbl. f. Bakt., II. Orig., vol.

49, p. 174.

1919. b. Ueber den Einfluss der Temperature auf verschiedene

Funktionen der Hefe. Centralbl. f. Bakt., vol. 49, p. 353; 1920,

vol. 50, p. 385.

INDEX

AUTHOR INDEX

Page

Albert H 99, 107

Albert, R 55

Albert, W 55

Albus, R 24, 34, 43, 44, 96, 143

Almquist, E

..2, 7, 10, 12, 13, 14, 15, 16, 117

Avery, O. T 39

Ayrton, J 15

Fraser, 0. G..
Frobisher, M..
Fuhrmann, F.
Fulmer, E. I..

Page

56

95

7

56

Balls, A. K.

42

Gardner, A. D 10, 12

Gordon, J 39

Govenlock, P 38

Graham-Smith, G. S

22, 24, 29, 32, 33, 40

Barber, M. A 21, 26 Greenleaf, W. E 43

Benians, T. H. C 53

Bergstrand, H 7

Breed, R. S 50

Brew, J. D 50

Brown, A 22, 24, 26, 29, 40, 42

Brown, J. B 42

Buchanan, R. E 20, 44, 56

Burke, G. S 56

Burke, V 44

Hajos, K 39, 40

Harris, W. H 102

Hoag, L 102

Holm, G 34

Hort, E. C 2, 3, 7, 10, 12, 13, 14

Jennings, H. S 14

Kauffman, F 37, 40

Koser, S. A 56

Kurpjuweit, 0 37, 40

Carlson, T 20, 29

Chesney, A 20, 26, 28, 38, 40, 42

Chick, H 21

Child, CM 97, 143, 144, 147 Lane-Claypon, J. E 20, 21

Clark, N. A 34 Ledingham, J. C : . . 20

Clark, P. F

59, 93, 96, 99, 103, 120

Conradi, H 37, 40, 42

Coplans, M 26

Crump, L. M 43

Curran, H. R 29, 32, 39, 41, 43

Cutler, D. W 43

Lohnis, F.

1, 2, 3, 4, 5, 6, 7, 11, 13, 14, 117

Manteufel 37

McKendrick, A. G 20, 24, 29

M'Leod, J. W 38, 39

Mellon, R. R

....2, 4, 7, 8, 12, 13, 14, 110, 147

Durbin, M. L 44 Meyer, A 12, 14

Migula, W Ill, 112, 113, 114

Eberson, F T 102 Mills, J. H 56

Eijkman, C 35, 36, 37, 38 Minot, C. S US, 142, 144

Eisenberg, P 53, 55, 96 Minoux, M 92

Ellis, M. M 26, 27 Morgan, M. J 39

Enderlein, G Morse, M. E 102

....2,5,6,10,11,12,13,14,117 Miiller, M 26

Ernst, P 55, 101 Munter, H 40

189

190

MORPHOLOGIC VARIATION

Page
Norris, D 2Q, 30, 40

Oebius, R 37

Otto, R 40

Pai, M. K 20, 24, 29

Paravicini, E 12

Pearl, R IS, 20, 25, 26

Penfold, W. J 20, 21

24, 26, 28, 20, 30. 40, 44. 45, 80
Peskett, G. L 43

Rahn, 0 24. 26, 37, 41, 42

Robertson. T. B 20, 42, 43

Roily 37

Rogers, L. A 35, 42

Ruehl, W. H.. .59, 93, 96, 99, 103, 120

Salter, R. C 29, 32, 40

Scales, F. M 5

Seiffert, W 56

Page

Sherman, J. M

24. 34, 43, -!4 06, 143

Slator, A 20

Stearn, A. E 96

Steam, E. W 96

Stewart, F. H 15

Taliaferro, L. G 105

Taliaferro, W. H 105

Tanner, I. W 24

Thompson, D. W 17, 94

Topley, W. W. C 15

Von Niesson 1

Wade, H. W 102

Wallace, G. 1 24

Whittier, E. 0 35,42

Williams, A 101

Wilson, G. S 50, 51, 127

Zikes, H 21, 29, 40

SUBJECT INDEX

Page

Acid

agglutination by 96, 97

influence of on cells 134, 135

influence of on rate of death.

129, 136

influence of on rate of auto-
lysis 129, 136

influence of on rate of growth. 34
Agar

preparation of slant cuHurrs. . 48

rate of growth on 21

Age, phys'ologic, of bacteria...

34, 44, 143, 144, 147

Alkali

influence of on cells 134, 135

influence of on rate of death

129, 136

influence of on rate of auto-
lysis 129, 136

influence of on rate of growth

34

Allelocatalytic effect 43

Antheridia 13

Antheridiospores 2

Antilysin 40

Area-length index

97, 89, 90, 117, 120, 130, 133

correlation with size 91, 92

with curvature. ... 121

Arthrogony 6

Arthrospores 3, 4

Asci 9, 11

Ascospores ^ , 10. 13, 14

Asymmetry of cells

88, 118, 122, 123, 125, 130

Autogamy " 5

Autolysis 55, 128

rate of 129, 137, 141

Autotoxines 34, 37, 39, 40

adsorption of 37, 38, 39

effect of heat on 36

effect of ether on 36, 38

Auxanogeny
Azotobacter

Page
6

3

Bacillus

anthracis 59

avisepticus 59

cohaerens 113

coli

32, 34, 35, 30, 41, 50, 87, 88, 125

coli mutabile IS

diphtherae 59, 99

fluorescens liqucfac'cns 38, 41

Hoffmann! 59

Hoagii 102

influenzae 59

lactis erythrogenes 38

megatherium 47, 49, 52, 53, 55,

60, 87, 103, 111, 112, 115, 139

paratyphosus 59

pertussis 59

pyocyaneus . 59

subtilis 59

typhosus 10, 59

vulgatus 59

xerosis 59

Bacteriophage 40

Bacterocidin 39

Bimodality 74, 80, 81, 82, 92, 123

Branching of bacteria

5, 9, 10, 118, 110, 130

Budding of bacteria 2, 3,

7, 9, 10, 14, 118, 110, 123, 130

Carotin 101

Calcium chloride

influence on forms of cells. .

95, 134, 135

influence on death rate.... 129, 136
influence on rate of auto-
lysis 129, 136

Centrosome 5

191

192

MORPHOLOGIC VARIATION

Page

Chlamydospores 7, 14

Cholera vibrio 6, 10, 59, 117, 120

Coefficient of variation 74

of area-length index 122

of length of cells 105

CoUargol, staining of granules

by 56

Composite photographs 58, 89

Congo red

negative staining with 53, 54

staining of dead cells by ... .

55, 56, 127

staining of granules by 55

Conidia 2, 10, 14

Counting chamber 49, 50

Counting of bacteria

by plating 127

microscopic 49, 50

Crowding of bacteria

effect of on growth 34

on spores 113

Curvature of cells 120

correlation with form 121

Cystit 6

Cystoid 6

Cytomorphosis 142, 144

"Dauerformen" 8

Death of bacteria, rate of

51, 127, 129, 141

Death phase. 20, 51, 120, 123, 125, 127

Dedifferentiation 143, 144

Differentiation. 115, 141, 142, 143, 144

Diphtheroid bacteria 7, 13, 60, 99

Diploid nuclei 2, 6, 13

Division of bacteria

6, 8, 11, 15, 62, 66, 93

equational 6

non-equational 6, 15

Dormancy 44, 69

Embryonic forms of bacteria..

93, 96, 108, 123, 140

Endospores 3, 4

Page

Ergastic substances 13

Exospores 3,4, 14

Filterable forms of bacteria. . . .2, 3, 5
Frequency distributions

according to area-length index

form of cells. . . . 134

length

73, 74, 78, 79, 82, 105, 106

Fruit flies, growth of 26

Fungi 5, 7, 8

Gametocytes of bacteria 6

Generation time 19, 59, 77

Glanders bacillus 59

Gonidia 2, 3, 4, 9, 11, 14

Gonit 6

Gram staining 55, 96, 100

Granules, intracellular

60, 61, 62, 66, 140, 141

Growth

cessation of 34

curves of 17, 18, 19

law of 18, 32

phases of 20

Haploid nuclei 2, 6, 13

Heredity in bacteria IS

Hybrids 2, 14, IS

Hydrogen peroxide 39

Impure cultures 9

Index of curvature 121

Index of variation of form. . .133, 135

correlation with autolysis 136

Inertia of bacteria 44

Inoculum, age of

influence on rate of growth. . . 26

size of cells 84

Inoculum, size of

influence on rate of growth. 24, 148

on size of cells. 76, 80, 108

on spore formation. 113

INDEX

Page

Involution forms

..7, 11, 12, 60, 61, 100, 117, 125
Isoelectric point of bacteria 96

Lag phase

20, 24, 26, 38, 42, 67, 69, 71, 119
Length of cells

59, 67, 71, 88, 103, 120

Life cycles of bacteria

3, 4, 5, 6, 8, 9, 11, 15

Lipochrome 101

Measurement of cells 52

Meningococcus 3, 10

Metachromatic granules

8, 100, 106, 107

(see also "volutin")
correlation with size of cells

107, 108

Microcolonies 61

Micrococcus grossus 38

Microcysts 3, 4

Mochlosis 7

Monogony 6

Monomorphism, monomorphists

10, 125, 139, 146

Mutations of bacteria

15, 110, 147, 148

MyceUum 7, 66, 93

Mych 6, 13

Mychit 6, 7

Neutral red, staining by 56

Nucleus-cytoplasm ratio

96, 142, 144, 145

Nuclei of bacteria

2, S, 6, 7, 8, 12, 96, :i01

Nutrients, concentration of
influence on rate of growth

29, 32, 35

on size of cells.. .82, 108
on spore formation. 113

Oit

Oospores

193

Page
.6, 10

2

Oidia

Paraplasmatic structures 141

Permeabihty of bacteria 56

Photomicrographs 52

Phylogeny 6, 93

Plasmodesmids 14

Plasmodium 14

Plasmolysis 55

Plasmoptysis 117

Pleomorphismf and pleomor-

phists..l, 2, 4, 10, 15, 125, 139, 146

Pliomychit 6

Pneumococci 38, 39

Ponder's stain 102

Probaeogeny 6

Probacterium 10

Protozoa 15, 42, 144, 145

Pseudomonas cerevisiae 8

Pyocyanase 38

Rate of growth of bacteria 17

factors influencing 21

relation to size of cells

59, 77, 86, 89, 103, 108

to form of cells. . . .89, 122
to spore formation. .

112, 113

Regeneration, rate of 26, 143, 144

Regenerative bodies 3, 4, 13

Resolving power of microscope. .5, 52

Resting cells 44

Resting phase

...20, 86, 103, 111, 115, 119, 123

* Schwarmer" 8

Segregation of characters 15

Selection 45, 80, 82

Senescent forms of bacteria. . . .

123, 125, 141

Sex in bacteria. . . .2, 4, 5, 7, 9, 14, 15

Skewness 74, 78, 81, 92

Soap, influence of on cells 95

194

MORPHOLOGIC VARIATION

Page

Spermit 6, 10

Sporangia 2, 61, 111

Spores ...3, 61, 111

Spore formation 12, 111

Staining, influence of on size of

bacteria 53

negative 53, 54

of dead cells 55, 56

Statistical methods 16, 57

Streptococcus lactis 35

pyogenes 59

Surface tension 94, 95

Surf ace- volume ratio 94

Page

Symplasm 4, 6, 11, 14

Syngamy 15

Temperature, effect on growth

21

Terminal growth 66

Thickness of bacteria cells 87

Viability of bacteria 11, 12

Volutin 12, 13, 55, 60, 100, 106

Zygospores 5, 11, 14

This book, MORPHOLOGIC VARIATION AND
THE RATE OF GROWTH OF BACTERIA by
Arthur T. Henrici, was designed, set, printed and
bound by the Collegiate Press of Menasha, Wis-
consin.

The type face is Linotype Old Style No. 7, set in 10
point on a 12 point body.

The paper is 60 pound Star English Finish, made
by the Ticonderoga Pulp & Paper Co., Ticonderoga
New York.

The binding is Munsell Cover.

With THOMAS BOOKS, careful attention is given
to all details of manufacturing and design. It is the
publisher's desire to present books that are satisfac-
tory as to their physical qualities and artistic possi-
bilities and appropriate for their particular use.

THOMAS BOOKS will be true to those laws of
quality that assure a good name and good will.

^,B,i

.^^

EOIE, MASS,