^

<^

PATTERNS AND PROBLEMS
OF DEVELOPMENT

THE UNIVERSITY OF CHICAGO PRESS . CHICAGO

THE BAKER & TAYLOR COMPANY, NEW YORK; THE CAMBRIDGE UNIVERSITY
PRESS, LONDON; THE M ARUZEN-K ABUSHIKI-K AISH A, TOKYO, OSAKA,
KYOTO, FUKUOKA, SENDAi; THE COMMERCIAL PRESS, LIMITED, SHANGHAI

Patterns and Problems
of Development

C. M. CHILD

The University of Chicago, Emeritus
and Stanford University

The University of Chicago Press
Chicago • Illinois

COPYRIGHT 1941 BY THE UNIVERSITY OF CHICAGO. ALL RIGHTS
RESERVED. PUBLISHED MAY 1941. COMPOSED AND PRINTED
BY THE UNIVERSITY OF CHICAGO PRESS, CHICAGO, ILLINOIS, U.S.A.

PREFACE

DEVELOPMENTAL physiology is often regarded as if it were
concerned only with embryonic development, and for some au-
thors of recent years it seems to be very largely a matter of verte-
brate or even of amphibian development. It cannot detract in any way
from the great interest and importance of the experimental work of Spe-
mann, his students, and many others on amphibian development to point
out that this and development of vertebrates generally constitute only a
small part and, so far as fundamental problems are concerned, probably
not the most important part of the field of developmental physiology.
Other patterns of embryonic development than those of amphibians or of
other vertebrates are of equal or greater significance in relation to the
physiological problems of development. Some of them appear to present
the problem of developmental pattern in simpler, more general form than
the vertebrates.

Moreover, the animal egg at the beginning of embryonic development
gives little information concerning the beginnings and origins of develop-
mental pattern. Pattern is already present in eggs — often advanced in
development — when what we commonly call "development" begins, and
the ovarian developmental period is at present almost inaccessible to
analytic experiment. If we are to be consistent, we must admit that em-
bryonic development in general is by no means the only material of de-
velopmental physiology. There are many other forms of development,
with other starting-points than eggs. Some of them, such as buds, recon-
stitutions of isolated pieces, and development of cell aggregates, bring us
much nearer the beginnings of developmental patterns and the factors
concerned in their origins and permit more extensive control and analysis
than do most eggs and embryos. In fact, embryonic development appears
generally to be the most highly speciahzed form of development. The
animal egg is one of the most highly differentiated cells of the body. Only
by comparison and analysis of all forms of development can we hope to
distinguish the physiologically fundamental factors in developmental pat-
tern from those which are incidental to a particular kind of development
and to arrive at an adequate concept of the physiological foundations of
development. The following pages, though very far from accomplishing

vi PREFACE

any such result, constitute in some degree an attempt at a comparative
consideration and analysis of data concerning some of the various sorts of
development and the characteristics of their patterns, with some discus-
sion of what is known or believed or remains to be discovered concerning
their origins. Embryonic development is considered not as the primary
and fundamental form of development to which all other patterns are to
be referred but rather as representing a relatively specialized type to be
interpreted in the light of what we learn concerning other more primitive
developmental patterns.

In certain species buds (often more than one kind of buds), isolated
pieces of various sizes from various parts of the body, and even aggregates
of dissociated cells may all give rise to individuals of the same sort as the
individual developing from an egg. These different starting-points ob-
viously do not all have the same pattern or organization as the egg, but we
can scarcely avoid the conclusion that the essential factors of pattern
must be similar in all. What are these factors, as distinguished from those
incidental to a particular sort of development? To the writer this ques-
tion appears to present a fundamental problem of developmental physiol-
ogy. With it this book is largely concerned.

Permission to reproduce copyrighted figures is gratefully acknowledged
as follows: to Verlag JuHus Springer, Berlin, to the Yale University
Press, and to Professor Spemann for Figure 156, A-H, a reproduction of
Figure i, a-h, from Spemann, Experimentelle Beitrage zu einer Theoric der
Entwicklung (1936; American ed.: Embryonic Development and Induction
[1938]) ; to Verlag Georg Thieme, Leipzig, for Figure 159, A-C, and Figure
160, reproductions of Figures 1-4 from Holtfreter, Biologisches Zentral-
hlatt (1933), Band 53, Hefte 7, 8; to the American Book Company for
Figure 206, B and C, after Figures 394 and 395 from Coulter, Barnes, and
Cowles, A Textbook of Botany (1910); to the University of Chicago Press
for Figure 208, A, after Figure 364 from Coulter and Chamberlain,
Morphology of Gymnos perms (1910); to D. Appleton-Century Company
for Figure 210, A-C, after Figures 600, 601, and 602 from Coulter and
Chamberlain, Morphology of Angiospcrms (1903); to Librairie-imprimerie
Gauthier-Villars, Paris, and to Professor Dalcq for Figure 219, A and B,
after Figure 23, a and b, from Dalcq, V organisation dc Va'uf chez les
C hordes (1935); to the Cambridge University Press and to Sir D'Arcy
Wentworth Thompson for Figures 222, 223, and 224, from Figures 369,
381, 382,404, 406, and 407 of Thompson, Grow//? aw^ Forw (1917). Special
acknowledgment is made with deep appreciation to Professor Dalcq for

PREFACE

Vll

his kindness in providing a photograph of an unpubHshed figure with per-
mission to reproduce and with full explanation, appearing as Figure 219,
C. For other figures from papers acknowledgment to the authors is made
in the legends with reference to the paper in each case. For redrawings of
certain figures and drawings or photographs of others the writer is deeply
indebted to Mr. Kenji Toda, artist of the Department of Zoology of the
University of Chicago, and to Mr. Dietrich Bodenstein, of the School of
Biological Sciences, Stanford University.

A few citations of papers too recent to be included in the Bibliography
and of some others which should have been included are given in footnotes
or in the text.

C. M. Child

Stanford University

TABLE OF CONTENTS

CHAPTER PAGE

I. Problems and Material i

II. Certain General Characteristics of Developmental Patterns . i6

III. Concerning Methods of Physiological Analysis 58

IV. Physiological Characteristics of Axiate Patterns 86

V. Differential Modification of Development: Coelenterates and

Flatworms 166

VI. Differential Modification of Development: Echinoderms . 197

VII. Differential Modification of Development: Other Animal

Groups 247

\'III. Gradients and Fields: Determination, Differentiation, and De-
differentiation 272

IX. Physiological Integration: Dominance and Physiological Isola-
tion 304

X. Physiological Dominance and Organization in Reconstitution . 332

XI. Reconstitutional Patterns in Relation to Experimental Condi-
tions 359

XII. Inductors and So-called "Organizers" in Embryonic Develop-
ment 435

XIII. Certain Embryonic Reconstitutions in Relation to Pre-existing
Pattern 5^4

XIV. Cleavage and Developmental Pattern 544

XV. Questions of Origin of Certain Agamic Patterns under Natural

Conditions 599

XVI. Origin AND Nature OF Embryonic Patterns: The Problems and the

Evidence 644

XVII. Physiological Integration, Differentiation, and Growth in the

Progress of Development 7°^

Appendixes 729

Bibliography 75°

Index 8°i

53668

1^1 LIBRARY

CHAPTER I

PROBLEMS AND MATERIAL

DEVELOPMENT of the individual organism is a continuous series
of events in space and time. At any period of development the
order is evident as a definite system of activities and structures
which constitute the pattern of the individual at that period. The indi-
vidual is a realization in these activities and structures of potentialities of
the protoplasm concerned in relation to conditions external to it, and these
external conditions are essential factors in determining what potentialities
are realized in any particular case. When we subject developing organ-
isms to experimental conditions, the course of development may be modi-
fied and the resulting organism may be widely different from that char-
acteristic of the species under natural conditions. These modifications
represent realization of other potentialities. Evidently the so-called "nor-
mal" individual is a realization of only a part of the potentialities of the
species-protoplasm. Since development of different organisms takes place
in protoplasms which differ specifically in constitution and possess differ-
ent potentialities, the individual structural and functional patterns which
develop also differ; and on the basis of these differences we distinguish
species, genera, and larger groups.

The individual organism is a physiologically integrated unit — a whole
with actual physiological relations between its parts, although at certain
developmental stages a greater or less degree of independence of some
parts as regards further differentiation may occur. In short, it seems nec-
essary to conceive development as essentially a series of events integrated
into a definite order, pattern, and unity which differs in character with the
constitution of the protoplasm in which development is taking place.
What problems does this phenomenon of development present? To ask
this question is to ask what the problems of biology are, for life is develop-
ment. And if we attempt to distinguish certain problems as fundamental,
we are in no better case, for, unless we are content with the mere accumu-
lation of facts, we cannot proceed far in the study of development from
any point of attack without coming face to face with the great problems
of biology.

2 PATTERNS AND PROBLEMS OF DEVELOPMENT

Workers in the ticld of developmental physiology have attempted by
many lines of experimental analysis to learn something more about devel-
opment than can be learned from observation alone. Their experiments
have not only given us a great body of data but have thrown much light
on various aspects of development concerning which we could have
learned little or nothing from mere observation, and have focused atten-
tion on various problems which promise to provide fruitful fields for fur-
ther investigation.

To many biologists, even many students of developmental physiology,
the word "development" means embryonic development. They recog-
nize, of course, that other forms of development occur, but consider them
of secondary importance. As a matter of fact, embryonic development is
only one among many forms or types of development, and probably the
most highly specialized of all. Many organisms, both plant and animal,
develop in various ways which result either in similar or in different sorts
of individuals. For example, hydroids may develop from buds, and the re-
sulting individuals may be like the parent or different from it; they may
also develop from isolated pieces of the mature body, or in some cases from
aggregates of dissociated cells, as well as from eggs or parts of eggs, em-
bryos, or larval forms. Again, the organization of the ascidian egg is evi-
dently not essential for development of an ascidian, for an individual of the
same sort may develop from an egg, from the tip of a stolon which itself
originates as a bud, directly from a bud without stolon development, from
cell aggregates formed from parts of an otherwise degenerating parent
(the so-called "winter buds" of various forms), from experimentally iso-
lated pieces of the mature individual, either by direct reconstitution or
from cells of the piece remaining after partial degeneration, and from
pieces of stolon. Unquestionably, something about the physiology of de-
velopment of a hydroid or an ascidian is to be learned from all these differ-
ent ways in which it occurs, perhaps even more from some of them than
can be learned from the highly differentiated egg and its development
alone. Plants develop from eggs and from buds of various sorts originating
from sporophy te or gametophyte tissues, many of the simpler forms from
naturally or experimentally isolated pieces or single cells of the plant body.

The occurrence of these many different forms or types of development
raises at once the question whether they are fundamentally different in
pattern or whether, in spite of the different starting-points, essential sim-
ilarity of pattern is to be found in some or all of them. Unless we are will-
ing to assume some metaphysical principle such as Driesch's entelechy

PROBLEMS AND MATERIAL 3

with power of free choice, underlying and guiding development, we should
expect to find fundamental similarities of pattern in all of them. In any
case, there is in all a definite developmental pattern involving spatial and
chronological order and physiological integration. There is reason to be-
lieve that the view so widely current, either implicitly or explicitly, that
embryonic development is of primary, and all other forms of development
of secondary, importance, has in some measure led us astray in concen-
trating much of our attention on problems concerned with this form of
development alone rather than on the problem of developmental pattern
in general and the question what, if any, features of developmental pat-
tern are common to all forms of development. By every criterion which we
can apply, the oocyte and the spermatozoon are highly specialized and
dift'erentiated cells. Evidently the differentiation of the egg cytoplasm is
a factor in determining the course of development of the embryo : ascidian
embryonic development is an excellent example. But, as noted above, an
ascidian can develop in various other ways from other starting-points —
from buds, pieces of body or stolon, etc., which certainly differ in organiza-
tion from the egg and from each other. Embryonic development in a par-
ticular species under natural conditions always has the same starting-
point, the egg, which always, except for genetic potentialities, has the
same organization in a particular species. It seems beyond question that
embryonic development is a relatively highly specialized form of develop-
ment. If this is true, there is undoubtedly much to be learned concerning
the fundamentals of developmental pattern from the bud, the piece, and
other forms of development, as well as from the egg. May we not hope to
attain a more adequate conception of development by including in our
analysis other forms of development as well as egg and embryo? May it
not even be possible to interpret certain features of embryogeny in the
light of what we learn from other forms of development?

Any consideration of the physiology of development, if it is to be more
than a mere compilation of data, must itself have an order and pattern
and be integr3,ted into a whole. The ordering and integrating factors in
the following pages are the problems of pattern, order, and integration,
particularly in the earlier stages of development.

THE PROBLEM OF ORGANIZATION
rROTOPLASM AND ORGANISM

A "protoplasm" is a complex physicochemical system differing in
constitution in different species and, to some extent, in different Individ-

4 PATTERNS AND PROBLEMS OF DEVELOPMENT

uals; an "organism" is a system of protoplasms, an order and pattern ap-
parently on a larger scale than that of a protoplasm. The most general
characteristic of organismic pattern is a surface-interior difference, and in
some simple organisms such difference apparently constitutes the only
persistent organismic pattern, though other patterns may be temporarily
superimposed on it. In Amoeba, for example, pattern is apparently pri-
marily surface-interior; but the formation of a pseudopod alters and com-
plicates the pattern temporarily, and in some cases the whole body may
become temporarily an anteroposterior pattern with activity of a single
pseudopodial region dominating it. Environmental factors are apparently
directly concerned in the origin of surface-interior pattern. The surface of
the protoplasm exposed to the external medium or in contact with other
systems like itself, as in multicellular tissues, becomes different from the
interior. The characteristics of the cell nucleus and its relations to the
cytoplasm suggest that it originated as a differentiation of the interior.
Nuclei or their parts may possess a very definite pattern, but this pat-
tern is not the pattern of the organism or the cell and apparently cannot
autonomously determine that pattern. At present there is no ground for
believing that even surface-interior pattern can arise independently of
environment.

AXIATE PATTERNS

In most organisms we find, in addition to surface-interior pattern, other
orders or patterns which biologists have been accustomed to distinguish as
polarity, radial and bilateral symmetry, ventrodorsality, dorsiventrality,
and various asymmetry patterns. A polarity is a single serial order in a
certain direction, referable to an imaginary polar axis. Symmetries and
asymmetries may be regarded as polarities in other directions, referable
to other axes — radial, ventrodorsal, dorsiventral, lateral — or in other di-
rections in organ systems. On the other hand, polarity is an asymmetry
in a certain direction. In other words, the different directions to which the
spatial orders of organismic pattern are referable do not necessarily repre-
sent fundamentally different features of pattern. The terms "polarity,"
"symmetry," and "asymmetry" merely provide convenient distinctions
for patterns in different directions in the developing organism,. the term
"polarity" usually being applied to the order which becomes evident first
in development or is most conspicuous, with "symmetry" and "asym-
metry" applied to secondary orders in other directions, usually becoming
evident under natural conditions later in development than polarity. It

PROBLEMS AND MATERIAL " 5

is also often convenient to conceive the direction in which these orders
appear as representing a physiological axis and to designate developmen-
tal pattern in which serial orders of this sort occur as axiate pattern. Not
only the whole organism but its various parts, appendages, many organs,
or even individual cells may develop axiate pattern, and the axes of such
patterns may be in all possible directions within the same organism.

The features of axiate pattern which first become evident in develop-
ment are polarity and symmetry or asymmetry of the whole. These seem
to constitute a sort of background in relation to which the further de-
velopment of pattern takes place. They may be compared to a system of
co-ordinates, a frame of reference, with respect to which each part has a
definite position. This analogy, however, must not be pushed too far, for
observation of the course of development indicates, and experimental
analysis demonstrates, that the factors which constitute this apparent
frame of reference are not merely formal in character but are physiologi-
cally operative in localization and determination of the course of develop-
ment and that their action can be altered or they can be obliterated experi-
mentally by change in physiological condition of the protoplasm con-
cerned. They apparently constitute wholeness in its simplest, most gen-
eral terms. They appear to be the primary organizing and also the pri-
mary integrating factors.

Polarities, symmetries, and asymmetries of one kind or another appear
in many nonliving systems. In crystals, for example, we find one sort of
polarity and symmetry; in flames, flowing streams, electric currents, etc.,
polarities of other kinds. Many biologists have attempted to interpret
axiate pattern in organisms in terms of protoplasmic molecular or micellar
orientation, symmetry, or asymmetry similar or analogous to that of the
crystal or other physical systems. While such molecular or micellar struc-
tures are unquestionably present in fibrillae, membranes, skeletal and
cuticular structures, shells, and various colloid particles, often constituting
local structural patterns, there is at present no evidence that they de-
termine the general order and pattern of the whole organism, and there is
considerable evidence that organismic pattern is quite different in char-
acter. These molecular and micellar patterns are apparently chiefly char-
acteristic of differentiated protoplasms of certain tissues and of nonliving
products of metabolism and of secondary and local significance in develop-
ment. They occur in all possible orientations with respect to the axes of
organisms and even of organs in which they develop. They appear to be
eft'ects or expressions of pattern rather than its primary framework.

6 PATTERNS AND PROBLEMS OF DEVELOPMENT

We may assume, as many biologists have, that the basis of developmen-
tal pattern is an inherent property of protoplasm and therefore continuous-
ly present and independent of external conditions; but the data of observa-
tion and experiment afford little support to this view, for we find no evi-
dence of the existence of such inherent pattern, and we can alter and ob-
literate patterns and determine new polarities and symmetries experi-
mentally in various ways. If an inherent ''intimate structure," such as is
often assumed, were the factor determining developmental pattern, we
should expect pattern to be less readily alterable by external conditions.
Unless we take refuge in so-called "vitalism," which amounts to giving
up the problem, we are forced, by the results of experiment, to the con-
clusion that the axiate individual is the expression of a pattern on a larger
scale than molecular or micellar pattern of a protoplasm and that factors
external to the individual play a part in determining it. Such a pattern
can be determined only by reaction of a protoplasmic system of specific
constitution to conditions in its environment, either within the parent
body or external. According to this view, molecular, colloidal, chromo-
somal, and other patterns may be present; but they do not constitute or-
ganismic developmental pattern, nor do they autonomously give rise
to it.

Organization undoubtedly involves chains of chemical reactions and
changes in character of reactions as development progresses, but any con-
ceivable number of chemical reactions cannot give rise to an organismic
pattern unless they are in some way definitely ordered in both space and
time. The chemical reactions in development are evidently so ordered in
a pattern which is spatially of molar order of magnitude and changes in
an orderly manner chronologically.

The Roux-Weismann hypothesis that organization results from the
separation of hereditary elements by qualitative nuclear divisions during
development has been abandoned because neither cytology nor experi-
mental investigation of development support it and because it is difficult
to conceive any mechanism or agent, short of a superintelligence, which
could fulfil all the requirements of the theory. It is now believed that each
nucleus contains all the genes. It is apparently maintained by some that
organization is determined by the genes; but if each cell contains all the
genes, the cells of a multicellular organism can become different only
through action of something external to the genes and determining differ-
ent gene effects in different cells. Obviously, regions, parts, and organs
are localized and become different in the course of development according

PROBLEMS AND MATERIAL 7

to an orderly pattern; that is, a progressive organization occurs. The
problem of organization is the problem of the origin and nature of the
pattern underlying and determining where, when, and how the differences
appear and in what they consist. It involves the physiology and behavior
of living protoplasms.

PHYSIOLOGICAL DIFFERENTIALS OR GRADIENTS

Cells commonly have rather sharply defined boundaries, but in the
earlier stages of developmental pattern sharply defined boundaries be-
tween the various regions or fields of the developing organism are usually
absent. Instead of such boundaries, we commonly find evidences of grad-
ed differences from one region to another. These differentials, gradations,
or gradients appear in various aspects. For example, we find gradations
in visible protoplasmic structure and in physical and chemical constitu-
tion; in rate of cell division; in cell size; in rate of growth, morphogenesis,
and reconstitution ; in degree of determination of organs; in respiration;
and in reaction to various external factors. Where it has been possible to
compare a number of these different gradients in the same individual, a
high degree of correspondence as regards region involved and direction has
usually been found. The most conspicuous of these gradients and the first
to become evident in development are those associated with the polarity
and symmetry of the whole organism. In fact, a gradient of some sort is
very commonly, if not always, the earliest distinguishable evidence of a
physiological axis. Similar gradients also appear in the early developmen-
tal stages of various organ systems and organs. Moreover, localized de-
velopmental activities very generally show evidence of a graded decrease
from a region of highest intensity in a certain or in all directions. It has
been possible, in many cases, to alter or obliterate such gradients and to
determine new ones experimentally and so to alter the axiate pattern of
development correspondingly in definite, controllable manner. The ex-
perimental data which will be considered in later chapters show, beyond
question, that various sorts of gradients which appear in development are
manifestations or expressions of underlying physiological differentials of
some sort, which are organismic in order of magnitude; that is, they coin-
cide with the scale of organization of the individual or part concerned.
In short, these physiological gradients are characteristic features of axiate
order and pattern : differentiation of regions and organs occurs in definite
relation to them. Obviously, they are associated with development in
some way, cither as factors which arc operative and determinative or as

8 PATTERNS AND PROBLEMS OF DEVELOPMENT

incidents or effects of more fundamental factors. Different gradients may
differ as regards the activities occurring in them, and the evidence indi-
cates that the presence of a gradient may be as significant for development
as its nature. It is evident, then, that consideration of developmental pat-
tern involves the problems of the nature and origin of these gradients and
of the parts which they play in development.

THE PROBLEM OF PHYSIOLOGICAL INTEGRATION

The "organism as a whole" is the sum not merely of its parts but also
of the relations between them, their actions, and effects on one another.
These relations, commonly called "physiological correlation," represent
in each particular case a relation of control or dominance, on the one hand,
and of. being controlled or subordination, on the other. The dominance
may range from slight and momentary to complete and permanent. Since
these integrating factors are associated with the pattern of organization,
they too constitute an orderly spatial and chronological pattern. That
physiological dominance exists, even early in development, that it is an
essential factor in development, and that in many of the simpler animals
it is necessary throughout life for persistence of the individual as a whole
have been demonstrated by many lines of experiment.

Present knowledge leads us to conclude that physiological dominance
of one region or part over another may be effected in two different ways — ■
by initiation and transmission of energy changes and by production and
transport in mass of substances. Dominance of the transmissive type may
be regarded as including production and transmission of mechanical,
thermal, and electrical changes; but the most important factor is excita-
tion and its transmission. Protoplasms in general are capable of excitation
and some degree of transmission of excitatory changes; but the highest
development of this type of dominance appears in the nervous system and
its receptors, which have become in the higher animals highly differenti-
ated organs of excitation and its transmission. This type of dominance is
possible without any pre-existing differentiation from each other of the
parts concerned. The region primarily excited is dominant, at least tem-
porarily, over regions to which the excitation is transmitted. The pri-
mary excitation establishes the difference between these regions and so de-
termines the dominance. This type of dominance is therefore to be re-
garded as the primary integrating factor in organismic pattern.

Transportative or chemical dominance presupposes some degree of dif-
ferentiation of the parts concerned: if differentiation is absent, all parts

PROBLEMS AND MATERIAL 9

will produce the same substances. Dominance of this kind unquestionably
occurs widely in organisms but evidently increases in complexity and im-
portance with increasing differentiation of organs, both in individual de-
velopment and in evolution. The hormone interrelations in the higher
vertebrates represent its highest development. Transmissive dominance
is a control based primarily on intensity factors; transportative domi-
nance, on specific substances. The one may be regarded as essentially
quantitative; the other, as qualitative. Both transmission and transport
may be concerned in some types of control, as, for example, in the media-
tion of nerve impulses by chemical substances. According to current the-
ory, movement of ions occurs in transmission of excitation, but only over
short distances; and transmission depends on the electric charges, not on
the chemical nature of the ions. The problem of physiological integration
of the organism as a whole is the problem of the origin and nature of the
factors concerned in dominance and of their effects on subordinate parts.
Self-differentiation or independent difTerentiation of parts, correlative or
dependent differentiation, induction, organizers — all raise questions con-
cerning the part which physiological dominance plays in development and
the nature of the factors concerned in each case. From another viewpoint,
the nature of excitation and transmission, the methods of transport of
substances, and the chemical constitution of hormones and other trans-
ported substances and of the parts affected are all involved in the problem
of physiological integration.

In consequence of the researches of Spemann and his co-workers on
amphibian development the terms "inductor" and "organizer" have found
wide acceptance among students of developmental physiology. The con-
cepts of inductor and organizer have undergone changes as amphibian ex-
perimentation has advanced. The two terms have often been used indis-
criminately, but with progress of experiment distinction has become in-
creasingly desirable. At present an "inductor" may be defined as an agent
which brings about a definite developmental effect — a determination or
differentiation of a particular tissue, for example, or a new axiate pattern.
In earlier amphibian experiment the inductors were parts of a developing
embryo, either of the same species as that in which the induced effect
occurred or of another species. An organizer is an inductor which deter-
mines a definite, orderly developmental pattern in another part. The in-
ductors from the region constituting the dorsal lip of the amphibian blas-
topore are, in general, also organizers, since they determine a new axiate
pattern, an organization. As experiment progressed, it was discovered

lo PATTERNS AND PROBLEMS OF DEVELOPMENT

that more or less differentiation of neural tissue in amphibian embryos
could be induced not only by the living tissue of the dorsal lip region but
also by various other amphibian tissues of the same or other species, gen-
era, or orders; and not only that, but dead tissues, as well as living, proved
to be inductors. With further experiment many living and dead tissues
from many different animals, representing most invertebrate and verte-
brate groups, tissue extracts, and various synthetic chemical substances,
proved to have more or less inducing action. Some of these foreign in-
ductors are apparently organizers: certain plant tissues, for example, have
been reported to determine neural plate. Induction is not restricted to
amphibian development but occurs in other vertebrate embryos; and in
hydroids and planarians small pieces (in planarians chiefly pieces from
the region of the cephalic ganglia) may determine new axiate patterns (see
pp. 378-87). In the hydroid, Corymorpha, a lacerated incision may serve
as an organizer.^ Moreover, photic, electric, thermal, and chemical dif-
ferentials and gravity can determine new physiological axes and develop-
mental patterns in at least some plant or animal species. These agents are
as truly inductors and organizers as living or dead tissues or organic chem-
icals.' Inductors and organizers are, then, nothing new but are simply
cases of physiological dominance established in one way or another and
are characteristic of development in general. The character of the domi-
nance remains to be determined in each case. It is probable that it will
be found impossible to make any sharp distinction between inductors and
organizers, for, as will appear in later chapters, it may be questioned wheth-
er the organizer does anything more than determine or play a part in de-
termining a relation of dominance and subordination which becomes the
real organizing factor. In view of this possibility, which is supported by
experimental evidence, it is perhaps desirable to drop entirely the term
"organizer."

It has usually been assumed that chemical substances are the actual
inducing agents in vertebrate development. Whether or to what extent
this is the case is perhaps still uncertain as regards some inductions. There
is, however, considerable evidence which makes it probable that develop-

■ For earlier experiments showing or indicating dominance and induction in reconstitution
see: Browne, 1909; Rand, 1911, 1912; Child, 1911U, c. Many of the experiments of Driesch,
Morgan, Child, and others on the hydroid, Tiibitluria, suggested dominance of the hydranth
region and of distal over proximal regions. Dominance of the vegetative tip in plants has long
been familiar to botanists.

' Experimental determination of axes, inductors, and organizers will be more fully discussed
in chaps, xi and xii.

PROBLEMS AND MATERIAL ii

mental induction and axiate organization may be initiated by transmis-
sion of excitation in some of the invertebrates. In its more primitive ex-
pressions dominance is intimately associated with physiological gradients.
The high region of a gradient is primarily the chief dominant region; but
a given gradient level, when isolated from more anterior levels, may, to
some degree, dominate lower gradient levels, at least in the polar gradient
of hydroids and planarians. Experimental localization of a new dominant
region generally results in induction of a gradient or gradient system, with
the dominant region as its high end or as its center or otherwise located,
according to the form and pre-existing pattern of the material. Trans-
portative dominance in its more highly specific forms is not necessarily
associated with a gradient, though there may be an indirect relation if the
parts concerned were determined in relation to a gradient.

In many organisms, particularly in the simpler animals, dominance ap-
parently decreases in effectiveness with increase in distance from the dom-
inant region or is limited in range of effectiveness, and this range varies
with physiological condition and changes during the course of develop-
ment. In consequence of this hmited range, more or less physiological
isolation of parts of the organism which come to lie beyond this range is
possible. "Physiological isolation" may be defined as isolation in greater
or less degree of a part of an organism from control by a dominant region,
without physical discontinuity. Physiological isolation of a part may be
brought about in various ways, to be considered in chapter ix. In the
simpler organisms the result of physiological isolation is usually alteration
of the part of the axiate pattern present or development of a new inde-
pendent axiate pattern in the isolated part.^

Agamic reproduction in axiate organisms is intimately associated with
physiological isolation. In many cases either the degree of isolation is in-
suflicient or the potentialities of the isolated part are too limited to per-
mit development of a complete individual, and a part — for example, a
segment — may develop. Local dominances and physiological isolations
maybe concerned in determining the developmental order of various repet-
itive parts, such as tentacles. In fact, development apparently involves
progressively increasing complexity of relations of dominance and sub-
ordination, of both the transmissive and the transportative type. Ques-
tions of the origin and nature of physiological dominance in any particular
case; of its changes in character during the course of development; of its

J For earlier discussions of dominance and physiological isolation see Child, 1911a; 1915^,
chaps, iv and v; 19246, chaps, x-xii.

12 PATTERNS AND PROBLEMS OF DEVELOPMENT

role in determining order, scale of organization, and growth forms of mul-
tiaxiate organisms; of its relation to induction; of the significance of physi-
ological isolation in development; and the factors involved in different
organisms and under different conditions — these are some of the questions
involved in the problem of physiological integration.

THE MATERIAL OF DEVELOPMENTAL PHYSIOLOGY

Comparative investigation of embryonic development in difTerent or-
ganisms and groups has given us a broader conception of this form of de-
velopment than could have been obtained from a single species or group.
Comparative study and experimental analysis of different forms of devel-
opment must provide a basis for a still broader conception and assist us
in distinguishing generally essential factors from those concerned with a
particular form of development. In all forms of development physiological
order, pattern, and integration appear, and evidence is accumulating to
show that the essential differences in the different forms are not as great
as has often been believed. Some of them present possibilities of experi-
mental analysis which are present to a lesser degree in embryonic develop-
ment. Sometimes it is even desirable to interpret embryonic development
in the light of what we learn from other forms of development.

Embryonic development, development from spores, gemmules, stato-
blasts, and other special reproductive bodies formed under natural condi-
tions occur in cells or cell masses which have, in each case, a particular
past history which is always essentially the same in each species. Some
of these reproductive bodies appear to be highly differentiated cells or cell
groups, the egg and spermatozoon (the most highly differentiated of all)
having undergone greater change from the undifferentiated or embryonic
cell than most other cells of the individual. There can be no doubt that
the past history of the egg plays some part in determining the course of its
development, at least in the earlier stages. In many species there is a con-
siderable degree of regional differentiation in the egg cytoplasm preceding
or following fertilization; when present, this may become an important
factor in determining the manner in which development takes place, even
though the resulting individual may be similar to an individual which de-
velops from a physiologically or physically isolated part of the adult ani-
mal or from a bud.

The reproductive cells known as "spores" arise in some unicellular or-
ganisms by fragmentation of the individual, but in multicellular organisms
they usually originate only under certain physiological conditions and

PROBLEMS AND MATERIAL 13

from certain parts of the body. In some forms the spore is apparently a
more highly differentiated cell than other body cells, and in many plants
it is formed only in highly differentiated organs; yet it has the potentiality
of developing into a new multicellular individual. Multicellular spore-
like bodies — sponge gemmules, bryozoan statoblasts, and some ascidian
winter buds — have a different past history from that of the egg but can
develop into the same sort of individual.

In cases of fission which give rise to new individuals, the reproductive
body has previously been merely a part of the parent body. As such it has
had a past history of development and more or less differentiation; more-
over, it is not always the same part of the parent body which is separated
by fission, but the kind of individual which develops from it does not de-
pend on this differentiation. In the fiatworms and annelids which under-
go fission the course of developmental reorganization may differ according
to the body region separated by fission, but similar individuals result.
Origin of segments seems to be essentially a repeated fission, primarily
mesodermal, with limited development of each. Segments are formed in
embryonic development, in the development of zooids and of isolated
pieces of segmented animals, but the manner in which they arise differs
somewhat in the different forms of development, though the final result
may be the same.

Experimental isolation of pieces of the individual by section, with fol-
lowing reconstitutional development, is purely accidental, as far as the
original individual is concerned. The different parts have had different
past histories and undergo different courses of reorganization in giving
rise to similar individuals. Also, they may be of widely different size and
develop into individuals of different size ; and the animals from which they
are taken may be of different age, nutritive condition, etc. Such forms of
development provide much more favorable material for various lines of
physiological analysis than does embryonic development.

Various multicellular axiate algae separate into individual cells under
certain unfavorable conditions, and these cells are capable of developing
into new axiate multicellular individuals.'' Certain planarian species, if
kept above a certain temperature, separate into fragments when they at-
tain a certain size, the size attained and the size of the fragments varying
with the character of nutrition (Child," 1913c, igi^d). At low temperature
they become sexually mature instead of fragmenting (Castle, 1928). Cer-
tain nemertean species fragment on stimulation. In the planarian frag-

^ Tobler, 1902, 1904, 1906; Child, igijb.

14 PATTERNS AND PROBLEMS OF DEVELOPMENT

ments there is complete degeneration of internal organs during an en-
cysted stage; nevertheless, development of planarians finally occurs. De-
velopment of the nemertean fragments is a reconstitution like that follow-
ing isolation of pieces by section.

In development of new axiate individuals from buds there is much of
interest for developmental physiology. The bud makes its appearance as
a localized region of cell activation and growth, which in certain plants
may begin in a single cell of a leaf or other part of the vegetative body,
and in various unicellular organisms, in some portion of a cell. Buds of
many multicellular forms, both plant and animal, develop into new axiate
multicellular individuals or members of an individuation of higher order,
as in multiaxiate plants, hydroids, etc. Moreover, the individual which
develops from the bud may be like or unlike the individual from which the
bud arises or the individual which develops from the egg. Buds of some
forms from different regions of the parent body or from parents in different
physiological condition develop into different kinds of individuals; hy-
dranth buds and medusa buds of certain hydroids, and vegetative and
flower buds of plants, are examples. Many organs, such as tentacles, ap-
pendages, etc., originate as localized budlike regions of growth and differ-
entiation, apparently similar to buds which give rise to complete individ-
uals, but with limited developmental potentialities and often, like the am-
phibian limb bud (pp. 390-95), deriving a part of their axiate pattern
from the parent body. Fusion of eggs, embryos, or other parts may result
in the development of single individuals; and transplantation and im-
plantation of parts in embryonic and later stages make possible extensive
experimental analysis of determination in parts, of dominance, of induc-
tion, and of effects of different organismic environments, different gradi-
ent levels, etc., on the course of development.

Since different levels of a physiological gradient differ in susceptibility
to many, if not all, external chemical and physical agents which are toxic
to living protoplasms, it is possible to alter and control the course of de-
velopment differentially by subjecting the whole organism to the action
of such agents. That the various forms of development provide practically
unlimited material for biochemical and biophysical investigation is ob-
vious. Needham's Chemical Embryology (1931), which is concerned with
the chemistry of only a single form of development, is sufficient evidence
on this point. And finally, it is evident that both genetic and environmen-
tal factors are concerned in all forms of development. All of them are to
be regarded as reactions of protoplasms of specific genetic constitutions to

PROBLEMS AND MATERIAL 15

environmental factors within the parent organism and in the external
world. In this reaction certain potentialities of the protoplasm are real-
ized as developmental order and pattern. All forms of development pro-
vide material for investigation of the roles of heredity and environment in
determining patterns.

In short, embryonic development is not the only, and perhaps in many
respects not the most interesting and significant, material for study of
various developmental problems. An adequate general theory of develop-
ment must be based on the less specialized forms of development; it must
recognize and distinguish the factors common to different forms of devel-
opment from those characteristic of only a particular form; also, it must
attempt to interpret the more highly specialized embryonic form of de-
velopment as far as possible in terms of the common factors.

In the following pages a consideration of the physiology of development
is undertaken with particular reference to the problems of the origin and
nature of developmental pattern. Data of observation and experiment
concerning different forms of development, the conclusions which have
been or may be drawn from them, and their bearing on the general prob-
lem of pattern and particularly on the question whether or to what extent
different forms of development may be fundamentally different or funda-
mentally similar in the essential features of pattern are discussed. Per-
haps this survey of developmental physiology on a somewhat broader
basis than has usually served— that is, in terms of a comparative analysis
of different forms of development with an attempt to discover common
factors of pattern — may have some value for the future because it endeav-
ors to show that the patterns of other forms of development are no less, or
even more, important than embryonic pattern in relation to the general
problem. Perhaps it may also serve as a stimulus or an irritant to further
investigation.

CHAPTER II

CERTAIN GENERAL CHARACTERISTICS OF
DEVELOPMENTAL PATTERNS

THE various forms of agamic development, reconstitution from
experimentally isolated pieces of individuals, and development
from the egg have different starting-points and follow more or less
different courses. The question of their physiological resemblances and
differences is of great interest and importance, since it is a question wheth-
er the same organism may arise in fundamentally different ways or wheth-
er the different forms of development are physiologically more or less sim-
ilar in pattern. Observation of different forms of development under nat-
ural conditions and of experimentally induced development shows certain
features which indicate or suggest that underlying physiological factors
may be more or less similar. Certain of these characteristics of patterns
which are directly evident are briefly pointed out in the present chapter.
Attention has been called to some of them in earlier publications.'

BUD PATTERNS

In both plants and animals new axes which become new individuals,
branches, zooids, organ systems, or organs such as leaves, roots, tentacles,

A B C

Fig. I. — Diagrammatic sectional outlines of stages of bud growth; arrows indicate direc-
tions of gradients.

appendages of various sorts, develop from what we call "buds." In its
early stages a bud is a locus, usually at or near an external or internal sur-
face of the parent body, in which the conditions which determine its de-
velopment operate in decreasing degree from a physiological center which
may or may not be the geometrical center. These conditions apparently

' Child, 1915c, pp. 65-87; 19246, pp. 74-100; i928</.

16

CERTAIN GENERAL CIL\RACTERISTICS

17

consist primarily in an activation of some sort which decreases in intensity
from the center. As the bud develops, the more rapid growth of the physi-
ologically central region brings about elongation, usually more or less ver-
tically to the surface on which the bud appears, and the radial differential
becomes longitudinal (Fig. i). Usually the central region becomes the
apical end of the new axis; but in some buds the center may change its
position, the axis may become asymmetrical, or the original center may
divide into two or more, as in dichotomous branching and various sorts of
appendages.

Figs. 2 and 3. — Development of an adventitious bud from epidermal cells of a Begonia
leaf. Fig. 2, surface view; Fig. 3, longitudinal section. The old cellulose membranes are indi-
cated in Fig. 2 by double lines; the membranes of new cells, by single lines. In Fig. 3 the heavier
line indicates an old cellulose membrane undergoing resorption (from Regel, 1876).

Adventitious buds which develop from epidermal cells of parts of plants '
which are physically or more or less physiologically isolated from active
vegetative tips provide almost diagrammatic examples of the gradient
system and its transformation. Buds of this sort developing from epider-
mal cells of the leaf blade or petiole in certain species of Begonia, after iso-
lation of the leaf, were described and figured by Regel (1876). Before the
bud appears, the epidermal cells of the leaf possess cellulose membranes
and large vacuoles, have ceased to divide and grow, and under the usual
conditions would never divide or grow again. In short, they have all the
characteristics of differentiated cells. The locus of a bud may be a single
one of these cells, or it may involve several of them. The beginning of
bud formation is indicated by disappearance of the differentiated vacuo-
lated condition, accumulation of cytoplasm, and rapid division within the

i8

PATTERNS AND PROBLEMS OF DEVELOPMENT

old cellulose membrane. Figure 2 shows a case in which the locus of bud
formation involves most intensely the adjoining parts of four cells. The
more deeply shaded cells in the central region are filled with finely granu-
lar "embryonic" cytoplasm, and the optical density of this decreases from
the center peripherally, as indicated by the decreasing shading. The

Fig. 4. — Five stages in development of an adventitious bud from an epidermal cell or
cells of flax hj'pocotyl: a, four derivatives of single epidermal cell; e, epidermis of hypocotyl;
Ip, leaf primordium of adventitious bud (from Crooks, 1933).

peripheral unshaded cells are still more or less vacuolated, although they
have undergone division. The figure shows very clearly a radial gradient
system in rate of division indicated by cell size and a gradient in the same
direction of cytoplasmic character and content of the cells. Evidently the
intensity of the activation initiating formation of the bud decreases from
the central region peripherally. As development proceeds, the old cellu-

CERTAIN GENERAL CHARACTERISTICS 19

lose membranes are resorbed, and the whole area consists of embryonic
growing and dividing cells; but the gradient system persists. Since divi-
sion and growth are more rapid in the center of the locus than peripheral-
ly, the developing bud grows out from the leaf surface, and the radial
gradient system becomes an apicobasal axial gradient system. Figure 3
shows a vertical section of a bud after outgrowth has begun. It is evident
that the gradient in cell size and cytoplasmic content extends from the
free surface proximally as well as from the center peripherally. In the
stages shown in Figures 2 and 3 the bud represents the apical region of the
new axis. With further elongation the free end becomes the vegetative
tip, leaf buds appear in definite relation to it, and differentiation of vascu-
lar bundles begins at a certain distance from it.

In the seedlings of flax {Linum usitatissimum L.) development of ad-
ventitious buds can be induced from epidermal cells of the hypocotyl by
removal of all buds distal to this region. Five stages in development of
these buds are shown in Figure 4. In the earliest stage (upper left) only a
single cell has begun to divide; the later stages figured show a general
radial and basipetal gradation of increasing cell size, indicating decrease in
rate of division from the region of primary activation. In most cases
among the seed plants adventitious buds do not arise from epidermal cells
but from tissues below the surface, but the visible pattern does not differ
fundamentally from that described here. Adventitious buds may arise in
callus tissue which develops on wounded surfaces in many woody plants.
Here, also, bud pattern is essentially similar to that of the epidermal bud.
Roots also develop from buds which originate beneath the surface of the
plant body. As in other buds, the apical region develops first, and other
parts arise from it as it elongates.

So far as gradient pattern is concerned, early stages of animal buds are
very similar to those of plants. Figure 5, an early stage of amphibian
limb buds, is of interest in comparison with the figures of plant buds. The
decrease in number of nuclei per unit area from the mesodermal region in
contact with the ectoderm suggests an activity gradient of some sort. In
the relation between the primary gradient pattern of the bud and morpho-
genesis the differences between plant and animal bud development are
much like those between embryonic plant and animal development. In
the plant bud the apical region usually remains for a time or indefinitely
an embryonic growing region, and morphogenesis begins at a certain dis-
tance from it. As new material is added to the axis from the tip, each level
begins morphogenesis when the critical distance between it and the tip is

20

PATTERNS AND PROBLEMS OF DEVELOPMENT

attained. In the animal bud morphogenesis usually occurs first at the
apex and progresses basipetally, and the region of most rapid growth may
also shift basipetally and perhaps becomes basal or posterior at some later
stage. In certain syllid annelids the posterior segment-forming region
buds, giving rise to new regions like itself, which become the apical regions
of the bud axes (except for the anal segment) but actually represent poste-
rior body regions and produce new segments anteriorly, the first formed
becoming the head (H. P. Johnson, 1901, 1902). In these buds the new
segment-forming regions remain embryonic for a time and give rise suc-
cessively to new segments; they and the posterior segment-forming regions

Fig. 5.

1933)-

-Early stage in development of amphibian limb bud in section (from Filatow,

of annelids in general, as persistent growing regions, are somewhat similar
to the vegetative tips of plant axes.

In its most primitive form bud pattern becomes evident as primarily a
radial gradient system; but many buds, both plant and animal, show earli-
er or later a bilaterality or an asymmetry, usually definitely related to the
axes of the parent body. The bilaterality of many leaves and flowers, the
dorsivcntrality of lateral branches of many conifers and other plants,
and the anteroposterior and dorsiventral asymmetry of the amphibian
limb indicate that the axiate pattern of the region from which the bud
arises persists in, or is impressed on, the bud.

In general, budlike developmental loci, whether plant or animal, wheth-
er developing into new individuals or into organ systems or organs, show
very similar gradient patterns in early stages. Unquestionably, the nature

CERTAIN GENERAL CHARACTERISTICS 21

of the physiological differential from center to periphery differs in different
Kinds of buds, but the relation of the differential to center and to periph-
ery is very similar; and in the adventitious bud of the plant or the bud
which becomes a new zooid in the hydroid or a new appendage in the am-
phibian— the pattern — originally more or less radial, becomes longitudinal
in consequence of differential growth.

Buds which develop into complete individuals must finally undergo
complete separation from the parent body. Usually the free end of the
bud axis becomes apical or anterior; the attached end, basal or posterior.
The fission which separates bud individual from parent occurs when the
attached region has attained a certain stage of development. A familiar
example is the hydra bud. In the buds of the segment-forming region of
certain syllids mentioned above, the head develops from the attached end
and its developmental stage determines separation, but the primary bud
pattern is apparently similar to that of other buds. So-called "buds" of
some forms — for example, the "winter buds" of some ascidian species —
are really fission pieces or cell aggregates rather than buds and separate
from the parent body before any development occurs. It seems preferable
to restrict the term "bud" to that form of development in which a new
axis originates in a local activation and differential growth.

In some axes developing from buds or budlike outgrowths the region
of most active growth remains at the base instead of becoming apical.
The hairs of certain algae, the long bladelike leaves of various plants, and
probably some animal appendages are examples. These axes represent
progressive differentiation from base to tip during growth or development;
but to what extent this type of axis is an integrated whole, not merely a
differentiation gradient of a certain kind, is not known. This kind of bud
axis is much less common than that with the high end of the gradient at
the tip. Some of these "inverse" bud axes — perhaps all of them — are ap-
parently similar to the usual type of bud in their early stages, and in some
the region of most active cell formation and growth may shift from base to
tip in the course of development. Most of the bud axes mentioned in the
following pages are of the usual type, with the region of most intense ac-
tivation becoming the tip, at least primarily. The inverse bud axes usual-
ly, if not always, give rise to differentiated parts, not to entire individuals.

Certain questions concerning bud pattern, conditions which determine
the origin of a bud, and other lines of evidence concerning its pattern, or-
ganism bud, and organ buds are taken up in later chapters.

2 2 PATTERNS AND PROBLEMS OF DEVELOPMENT

MULTIAXIATE GROWTH FORMS

The branching or multiaxiate plant, whether alga or tree, originates as a
series of buds from the primary axis. Each branch of the common type
of multiaxiate plant consists, at least during its development, of a vegeta-
tive tip, which is the first part formed, and other axial levels formed by the
tip. Protoplasmic growth decreases basipetally in the tip, except as new
bud loci are activated in certain spatial relations to the tip; and other
forms of growth, such as increase of cell size by increase in size of vacuole,
usually also decrease basipetally from a certain level below the tip where
they begin. Cells in certain regions of these lower levels, however, may
still continue division and protoplasmic growth and bring about increase
in transverse diameter along the axis. It is evident that the growth form
of the whole multiaxiate plant must be determined by relations between
the different axes. The tapering growth form of the fir and many other
conifers, of various other trees and herbaceous plants, and of many algae
with branches radially arranged or in a single plane must result from rela-
tions of some sort between the main axis and lateral branches. Moreover,
the variations in this growth form, that is, from a short cone with large
base to a long cone with small base, and the almost cylindrical, ovoid, or
nearly spherical forms of some other trees can result only from orderly,
graded differences in growth of the different axes, differing in different
species. Many multiaxiate plants — many trees, for example — show, when
young, a more or less conical form with marked difference in growth and
form of the main axis and lateral branches; but later these differences be-
tween main and lateral axes may decrease, and all may become more or
less completely equivalent and a form with spreading crown of more or
less similar branches results. The quantitative study of growth of lemon
shoots by Reed (1928) shows differences in growth of apical and lateral
shoots of definite and orderly character. Plants with a multiaxiate in-
florescence usually show a gradient, either acropetal or basipetal, in de-
velopment of flowers. In the cotton plant, for example, blooming begins
at the bases of the lowest fruiting branches and progresses acropetally
along each branch and along the main axis (McClelland, 191 6). Many
other examples of both acropetal and basipetal progress of blooming
might be given. Which of the two orders occurs apparently depends on
the condition of the tip at the time of flowering. If it is still growing, the
order of flowering is, in general, acropetal; if early transformation of the
tips into flower axes occurs, the order is basipetal.

The growth form of the multiaxiate ciliate protozoan Zoothamnium is

CERTAIN GENERAL CHARACTERISTICS 23

much like that of a conifer with single main axis and lateral branches, ex-
cept that the branches lie in one plane. Faure-Fremiet (1930) has shown
that in Z. alternans the colony, as a whole, shows a certain degree of in-
dividuality in the distribution of growth and multiplication of its cells in a
graded order. The zooid axes of certain other branching ciliates — Carche-
simn, Epistylis, etc. — ^are apparently equivalent, and the growth form is
similar to that of a plant with a spreading crown of equivalent branches.
The branching hydroids and bryozoa show growth forms very similar
to those of multiaxiate plants and apparently resulting from similar rela-
tions between different axes. In certain species of Pennaria and Obelia,
for example, the growth form, when not altered by injury, crowding, or
other conditions, is in general conical like that of conifers. In many other
hydroids the branches are more or less equivalent as regards growth; and
the spreading, less regular growth form results. An acropetal gradient in
time of appearance of medusa buds, whether on the hydranth bodies or in
special zooids, occurs in many species of branching hydroids. Apparently,
lower levels of the multiaxiate whole arrive at the physiological condition
which determines formation of medusa buds earlier than higher levels.
Buds which develop into new individuals arise in definite graded orders on
stolons in certain medusae, siphonophores, and ascidians. This order re-
sults in gradients of size, of stage of development of the buds, and often of
arrangement of buds; in some forms these gradients are repeated in order-
ly sequence, indicating periodic or cyclical changes in condition.^ In gen-
eral, the character of multiaxiate growth forms is determined by the spa-
tial and chronological order of development of the buds which constitute
the axes and by the rates of growth of the dilTerent axes. All these indicate
the existence of graded differentials of some sort.

PATTERN IN RELATION TO FISSION

Fission in many of the simpler unicellular organisms is apparently
nothing or little more than cell division; but in many forms with definite
axiate pattern which reproduce by fission, a definite developmental pat-
tern becomes evident in the reconstitution of the parts resulting from fis-
sion. Brief mention of a few cases will serve to illustrate the point. Many
of the flagellate protozoa divide longitudinally, with cytoplasmic division
beginning at the anterior end and progressing posteriorly. Figures 6-9

2 For interesting examples of these orders see Chun, 1896; Braem, 1908, for budding of
medusae; W. K. Brooks, 1893; M. Johnson, 1910; Ritter, 191 1, for bud orders in salpid as
cidians.

8

Figs. 6-9. — Four stages of cell division in Euglena spirogyra; chromosomes and nuclear
division not shown. Fig. 6, intranuclear body which forms the blepharoplasts has moved to
anterior end of nucleus and divided. Fig. 7, two blepharoplasts, resulting from a second division
of the intranuclear bodies, have reached the base of the reservoir, where the flagellum arises;
new axial fibers have developed from them and united with the old, and the flagellum has
divided. Figs. 8 and 9, later stages, showing progress of division of cell body from anterior
end (after Ratcliffe, 1927).

CERTAIN GENERAL CHARACTERISTICS 25

show four stages in the cell division of Euglena, as described by Ratcliffe
(1927). The nucleus approaches the anterior end before division, and the
behavior of the blepharoplasts and the progress of cytoplasmic division
from the anterior end posteriorly indicate that physiological conditions in
the anterior region of the cell differ in some way from those in more poste-
rior parts. In certain choanoflagellates division progresses posteriorly in
the cell body and anteriorly in the collar. Division in the cystoflagellate
Nodiluca progresses over the body from the mouth region, which is un-
doubtedly to be regarded as apical. Division of certain trichomonad
flagellates progresses from the anterior end posteriorly (Kofoid and Swezy,
191 5). In Noctiliica and the trichomonads the nucleus is situated in the
anterior end, but in others which show the same course of the division fur-
row from anterior to posterior this is not the case. In the two species of
Spirotrichonympha recently described by Cleveland (1938) the reconstitu-
tion of the highly differentiated morphological structure in connection
with fission progresses basipetally from the apical end of the new individ-
ual, which in one species is half the apical end of the original animal and
in the other is a new apical region at the original posterior end. Longitu-
dinal division also occurs in the stalked ciliate protozoa such as Vorticella,
Carchesium, Zoothamnium, etc., the division furrow appearing first at the
distal end of the zooid and progressing proximally. It is well known that
in Stentor and various other forms the peristome region develops first in
fission; development of the new gullet in fission of the heterotrichous cili-
ate Bursaria progresses from the anterior end (Lund, 1917).

Indications of a longitudinal differential also appear in free-swimming
ciliates which divide transversely, though these are usually less definite
than in flagellates and sessile ciliates. When the temperature of a Parame-
cium culture is raised from i8°-2 2° C. to 26°-30° C, the anterior member
of a pair resulting from fission undergoes the following division earlier
than the posterior member; but when the temperature is lowered from
18° 22° C. to 13°-! 7° C, the posterior member divides first. Even in the
third generation at high and low temperatures some indications of this
difference persist (W. Petersen, 1927). The author interprets these results
as due to a difference in physiological condition in anterior and posterior
regions which results in a differential susceptibility to high and low tem-
peratures, the high temperature accelerating, the low temperature de-
pressing, the anterior more than the posterior member. De Garis (1928)
followed division rates in five lines of Paramecium with selection for ante-
rior and posterior origin. In three lines the anterior selections showed a

26 PATTERNS AND PROBLEMS OF DEVELOPMENT

higher mean fission rate; in one rates were equal, and in another the rate
was higher in posterior selections. These data, particularly those of Pe-
tersen, suggest an anteroposterior differential in physiological condition
but indicate that, although the differences in the products of division may
be inherited, they do not necessarily persist indefinitely.

A more or less complete replacement of peristome and its membranelles
and of the cirri occurs in both members in fission of hypotrichous ciliates.
The new cirri appear in certain areas or fields before the old cirri disap-
pear. In Euplotes the primordium of the peristome is the first indication
of reorganization of the posterior product of fission, the primordia of the
cirri appearing somewhat later (Wallengren, 1901). This suggests a gradi-
ent in the developmental field of the posterior member. From certain of
his figures {H and J) of Stylonychia it appears that the new cirri of the
anterior member are slightly more advanced in development than those
of the posterior member. If this is true, it suggests an anteroposterior
gradient in the parent individual.

The gradient of the parent body is clearly evident in the successive
fissions constituting strobilation of the scyphistoma of Amelia and other
discomedusae. The longitudinal fissions which occur occasionally in hydra
and actinians, either as the result of injury (Roudabush, 1934) or other
conditions, begin apically and progress basipetally. Fission also occurs in
many flatworms, nemerteans, and annelids. In some cases it is essentially
fragmentation without definite fission zones, as in some nemerteans and
annelids. In others fission occurs at definite body-levels, and visible re-
constitution may precede or follow separation. Reconstitution of a whole
from the part in connection with fission in these forms is essentially simi-
lar to reconstitution following experimental isolation of pieces by section,
which is discussed below. Usually a part of the parent pattern persists in
the fission piece and becomes the starting-point for orderly reconstitution.
Fragmentation of nemerteans, Lumbriculus, and other forms apparently
results from strong muscular contraction without predetermination of
body-level. Separation at a definite body-level precedes morphological
development in Dugesia ( = Euplanana) dorotocephala and some other pla-
narians; but the new developmental pattern is physiologically distinguish-
able (pp. 108-17), and separation occurs when its physiological isolation
becomes sufficient to permit an independent motor reaction. In many
forms development precedes separation so far that chains consisting of
several zooids may result, separation occurring when head and posterior
end of adjoining zooids attain a certain degree of development. Figure 10

CERTAIN GENERAL CHARACTERISTICS

27

shows four stages in development of a chain in the rhabdocoel, Steno-
stomuvi. The second and following fissions occur in order from anterior to
posterior zooid of the chain; that is, increase in length of a zooid sufficient
to permit determination of a new head region, which can occur only at a
certain distance from the head of the zooid, is most rapid in the anterior

A

nil

^iii

V-^II2

Fig. 10. — Four stages in development of zooid chain of Stowstomum grande; fission zones of
first, second, and third order are indicated by Roman numerals; the order of their appearance
in second and third series, by Arabic numerals.

zooid and less rapid from zooid to zooid posteriorly, as the figure indicates.
Consequently, the chain shows a gradient pattern as regards order of ap-
pearance of new zooids, and each fission plane apparently results from de-
termination of a new head region in consequence of a certain degree of
physiological isolation of the body-level concerned from the dominance of
the head region of that zooid. Among microdrilous oligochetes similar pat-

28 PATTERNS AND PROBLEMS OF DEVELOPMENT

terns occur, but the chains usually consist of fewer zooids because growth
in length and physiological isolation of the posterior regions of existing
zooids are relatively less rapid than development of zooids. In some poly-
chete annelids chains of zooids occur, but the posterior members usually
separate successively before they undergo fission. Fissions provide ex-
amples of reconstitution occurring under natural conditions, in conse-
quence either of physiological isolation or of physical isolation. The polar-
ity and ventrodorsality of the new individual are usually in the same di-
rection as in the parent : the new head develops at the most anterior re-
gion of the part, which becomes the zooid, and its ventrodorsal pattern is
apparently determined by the parent ventrodorsality. Except as regards
effects of physiological isolation of parts in initiating development, the
natural fissions give us little information concerning developmental pat-
tern that is not obtainable from the reconstitutions resulting from experi-
mental section, and they have the disadvantage for experimental analysis
that they occur only under certain physiological conditions, although
these conditions can be induced experimentally in some forms, and are
usually limited to certain regions of the body. The wider range of possi-
bilities of analysis in the reconstitutions following experimental section
will become evident in the following section.

POSTEMBRYONIC RECONSTITUTIONS

Reconstitutional development may be defined as the alteration of pat-
tern and course of development in a part of a pre-existing individual fol-
lowing its physical or physiological isolation from its normal organismic
environment. In other words, it is the realization, in the isolated part, of
other developmental potentialities than those which have been, or would
be, more or less fully realized without isolation. This form of develop-
ment has been variously called "regeneration," "restitution," "repara-
tion," "reproduction," and, in general, "form regulation." For Driesch
and many others it has involved the assumption that pattern and de-
velopment of the isolated part are so altered that it approaches in some
degree or becomes a whole, a normal individual; that is, the isolated part
gives rise to more of the individual than if it had remained an integrated
part of the original whole. This conception of reconstitution is implied in
the terms "restitution," "reparation," "regeneration," and "form regula-
tion." While it is, of course, true that very often an isolated part does be-
come more nearly a whole, it is also true that very often it does not. Its
pattern may be so altered that it gives rise to some other part or parts of

CERTAIN GENERAL CKL'VRACTERISTICS 29

the individual than those which it formed or would form without isola-
tion; but whether these represent more, or possibly in some cases less, of
the whole depends entirely on what we mean by "more" or ''less." For
example, when a short piece of a Tuhularia or Corymorpha stem develops
into only the extreme oral end of a hydranth or to two opposed oral ends
(pp. 334, 346), has it become more nearly a whole than it was originally?
Unquestionably, its developmental pattern has undergone alteration; but
what has occurred can scarcely be called a restitution, a reparation, a form
regulation, or even a regeneration in any strict sense. It is, however, a
reconstitution, but one which makes long-continued existence impossible.
It is perhaps even pertinent to raise the question whether continued dif-
ferentiation of an isolated part in the same way as without isolation may
not, at least in some cases, involve some reconstitution. The fact that
the isolated part is capable of more or less independent or self-differentia-
tion does not prove that it differentiated independently when not isolated.
In view of the experimental data the term "reconstitution" seems prefer-
able to any of the others used for this form of development, because it in-
volves no implications of approach to wholeness, regulation, or replace-
ment of missing parts. It implies nothing beyond alteration of develop-
mental pattern in the isolated part. As already noted, physiological or
physical isolation and reconstitution of parts occur in the various forms
of budding and fission. Either a new axiate pattern develops in a former
part of an individual, as in buds, or the partial pattern present undergoes
reconstitution, usually into the pattern of a whole. It is perhaps worth
pointing out that even embryonic development may be regarded as re-
constitutional in character. The egg and the spermatozoon are morpholog-
ically, and apparently also physiologically, highly differentiated and spe-
cialized cells, and in embryonic development the differentiation and pat-
tern undergo great alterations. For parthenogenetic eggs isolation from
the earlier organismic environment is sufficient to initiate development,
but in other eggs activation by the spermatozoon or by chemical agents
is also necessary. In any case the egg does not develop so long as it re-
mains as a part of the parent body in the ovarian environment. It is diffi-
cult to understand how anyone can observe the changes which egg and
spermatozoon undergo in fertilization and development and doubt or deny
the occurrence of dedifferentiation. If the gametes are, in any sense, parts
of the parent body during their development as gametes, embryonic de-
velopment is a more extreme reconstitution than any other form of de-
velopment.

30 PATTERNS AND PROBLEMS OF DEVELOPMENT

The chief purpose of the present section, however, is to call attention
to some of the graded differences in reconstitutional patterns of experi-
mentally isolated parts of more or less mature individuals and particularly
to their relations to body-level of origin of the part. The reconstitutions of
mature organisms, particularly of those with elongated polar axes, usually
show such differences more clearly than those of isolated parts of eggs and
early embryos. Experimental analysis of differentials along an axis in the
egg and in early embryonic stages, by isolation of parts, is to some extent
limited by the presence of the whole axiate pattern in one or a few cells and
often also by the regional cytoplasmic differentiations which occur in
many eggs. These apparently represent developmental expressions of pat-
tern in relation to ovarian environment, maturation, and fertilization,
rather than the fundamental pattern itself. For these reasons considera-
tion of egg and embryonic reconstitutions is postponed to a later chapter,
following presentation of other lines of experimental analysis of patterns
of various forms of development.

For convenience three types of reconstitution may be distinguished:
substitution, redifferentiation or reorganization, and regeneration. Sub-
stitution is reconstitution in multiaxiate forms by substitution of another
axis for one removed. Substitution occurs very commonly in multiaxiate
plants in consequence of isolation from a dominant region. Removal of a
dominant axis or of its dominant region results in reconstitution of an-
other, or of more than one, axis (previously subordinate) into dominant
axes; or in initiation and development of one or more new axes — adventi-
tious axes — from cells which were previously parts of an axiate pattern
already present. In certain conifers, for example, removal of the domi-
nant region of the main axis, the vegetative tip, is followed by a turning-
upward of one or more of the uppermost lateral branches and its trans-
formation from a bilateral to a radial pattern of branching. Bud primordia
inhibited by a dominant region of another axis may be activated and may
develop following removal or inhibition of that region; or, if no bud pri-
mordia are present , adventitious buds may develop following isolation from
a dominant region. Substitution may occur in multiaxiate animals. A
hydra bud may remain attached and become the apical region if the par-
ent body distal to the bud-level is removed.'

Redifferentiation or reorganization is reconstitution of an isolated part
into something else without outgrowth of new tissue from surfaces of sec-
tion. A piece of Tuhularia stem, for example, may undergo reconstitution

3 Weimer, 1928; Rulon and Child, i937«.

CERTAIN GENERAL CHARACTERISTICS 31

into a hydranth or the distal part of a hydranth by reorganization or re-
differentiation without any ''replacement of missing parts" by outgrowth
of new tissue.

Regeneration is reconstitution by outgrowth of new tissue from the cut
surface and its differentiation into a part of an individual, often, but by no
means always, more or less similar to the part removed. Regeneration, as
defined, occurs widely in animals but is not common among plants,
though some algae and fungi regenerate and parts of vegetative tips re-
moved may be regenerated in some of the higher plants.

Redifferentiation and regeneration, as here defined, are probably not
very different physiologically. Reconstitution in stem pieces of the hy-
droids Tiihularia and Corymorpha is wholly redifferentiation; there is no
replacement of parts removed, but transformation of the piece. In most
planarians both regeneration and redifferentiation occur in reconstitution,
and the proportion of each differs in pieces from different body-levels of
the same individual and can also be altered experimentally. In general,
redifferentiation occurs to a greater extent in the less, regeneration in the
more, stably differentiated forms, even within the same phylum or class.
In Dugcsia { = Euplanaria), for example, considerable redifferentiation
occurs; in Procotyla, another triclad of different family, almost none.
Reconstitution in later developmental stages of arthropods and verte-
brates is limited to regeneration of certain parts, such as appendages and
various tissues.

The experimental reconstitutions resulting from isolation of parts by
section provide material for experimental analysis which cannot be ob-
tained in other ways. In various animals they can be initiated at will; and
the region of the body, size, and form of the isolated part can be varied
and controlled within wide limits, the effects of nutritive condition, physi-
ological age, and other factors can be investigated, and much concerning
physiological dominance can be learned.

PLANT RECONSTITUTIONS

A few examples of reconstitution in plants are briefly described. Re-
constitution in certain algae shows graded differences along an axis which
are very similar to those in hydroids. The thallus of the dXgd, Acetabular ia
mediterranea consists of a long stem from which whorls of hairs develop,
an umbrella-like apical region, and rhizoids at the basal end, all formed
from a single cell with the nucleus in the rhizoid region. Nucleated pieces
reconstitute completely, and even nonnucleated pieces of the stem are

32 PATTERNS AND PROBLEMS OF DEVELOPMENT

capable of considerable reconstitution. The original polarity is usually
retained, but short pieces may give rise to bipolar forms. Distal nonnu-
cleated pieces reconstitute apical parts; proximal pieces, rhizoids more
readily or more completely (Hammerling, 1934^, b, c; 1936). To account
for the observed results Hammerling postulates two opposed gradients of
different formative substances produced by the nucleus and accumulating
in concentrations decreasing from each end of the axis. The differences in
reconstitution in pieces from different body-levels are not very different
from those in certain hydroids. For the hydroids, however, the hypothesis
of two opposed gradients of different formative substances advanced by
Morgan was later abandoned as inadequate. Since these substances are
assumed to accumulate toward opposite poles of the axis, their distribu-
tion must be determined by a polarity of some sort to which they react.
If this be granted, the possibility that that polarity, rather than the hypo-
thetical formative substances, is the essential factor in determining the
differences in reconstitution at different body-levels may perhaps be ad-
mitted. Undoubtedly, the nucleus is essential for the continued life and
activity of the cytoplasm, but whether it is responsible for the axiate pat-
tern appears open to question. It does not appear that the nucleus has
any mechanism for determining the distribution of the two formative sub-
stances in two opposed gradients, even if the hypothesis of the substances
and their distribution be accepted. But whatever the interpretation, the
experiments show the existence in this single cell body of a gradient pat-
tern very similar, as regards its effect on reconstitution, to that in hy-
droids and various other animals.

The alga Caulerpa prolifera, a single multinucleate cell with differentia-
tion into rhizomes, rhizoids, and leaflike parts, also exhibits interesting
evidences of a gradient pattern in reconstitution." In certain fungi pres-
ence of a gradient pattern and dominance of the apical region are indi-
cated by the data of reconstitution.

An interesting example of substitution by formation of adventitious
axes appears in the liverwort Marchantia. The thallus pattern resembles
that of a bilateral animal in that it consists of a longitudinal and dorsiven-
tral axis with the dominant region at one pole (Fig. 11, ^). The notched
end may be regarded as anterior, for the apical cell, which constitutes the
growing tip and gives rise to other cells, lies in the notch, and this is the
dominant pole. The apical cell at times undergoes dichotomous division,
giving rise to two apical cells, each of which becomes the anterior pole of

■< Zimmcrmann, 1929, and references to earlier papers given by him.

CERTAIN GENERAL CHARACTERISTICS

33

a new thallus. Isolated pieces of the thallus reconstitute with interesting
evidences of gradient pattern. Any cell of the thallus is capable of recon-
stituting a new thallus, but in multicellular pieces preferential localization
of reconstitution occurs (Vochting, 1885). When an anterior region is re-
moved by transverse section at any level, a new apical cell and a new thal-
lus develop from the ventral side of the midrib just posterior to the cut
surface (Fig. 11, B). In general, in pieces containing any part of the mid-
rib the new thallus develops from its anterior ventral region (Fig. 11, C,
D). In pieces entirely lateral to the midrib with transverse anterior cut

C E

Fig. II, A-E. Reconstitution in the liverwort Marchantia. A, unbranched thallus with
midrib indicated; B, reconstitution from ventral side of midrib after transverse section; C, D,
reconstitution from half-midrib in pieces of half-thallus width; E, reconstitution from most
anterior level of piece without midrib and with anterior oblique cut surface (from Vochting,
188s).

end, the new thallus develops from the most nearly median part of the
anterior region; but in pieces lateral to the midrib with oblique anterior
section, development is from the most anterior region of the piece, which
is at the lateral border (Fig. 11, E). These and various other experiments
indicate the existence in the thallus of a longitudinal and mediolateral
gradient pattern of some sort. Any level of either can be experimentally
determined as the locus of development by altering the shape of the piece
and the region included in it. Planarian pieces of dilTerent shape and in-
cluding different regions give results which are similar in certain respects
as regards the gradient pattern which they indicate (pp. 52, 364-65).
In multiaxiate seed plants the chief growing tip may completely in-

34 PATTERNS AND PROBLEMS OF DEVELOPMENT

hibit or greatly retard development beyond a very early stage of bud pri-
mordia below it or may determine their development as lateral branches.
When this tip is removed or inhibited, the usual result is that the upper-
most buds develop, or one or more of the uppermost branches reconstitute
a main stem and development or transformation of axes below is in-
hibited. On removal of the reconstituted axis or axes the next uppermost
buds or branches react first, and so on. Even in the potato tuber, which
is a modified subterranean stem, this relation appears to some extent
(Appleman, 1918, 1924). In certain other plants only a difference in rate
or time of development of buds along the stem results from removal of the
tip, the uppermost buds developing first, and others in basipetal order.
The relation of development of lateral roots to a main root tip is very sim-
ilar.

Adventitious buds which develop from differentiated cells in some
plants after removal of all growing tips and preformed buds may also
show preferential localization or more rapid development at more apical
levels. When they develop on pieces of roots, as in the dandelion (Nemec,
1908), they tend to be localized at the basal end of the piece, that is, on the
end toward the stem; but in very short pieces they may appear anywhere
without preferential localization. The data indicate a gradient pattern
more or less effective in determining localization of the buds in longer
pieces but ineffective in very short pieces because in them there is so little
regional difference. Rhizoids and roots usually show an opposite localiza-
tion differential from shoot buds, that is, they tend to appear basally on
isolated stem pieces; they can, however, be determined at any level.

The art of pruning, trimming, and "forcing" development of certain
axes by removal of others in plants depends on reconstitution by substitu-
tion and on presence of a gradient pattern of some kind along the axes con-
cerned.

RECONSTITUTION IN ANIMALS IN RELATION TO BODY-LEVEL
CILIATE PROTOZOA

Some indications of a longitudinal gradient appear in connection with
fission in ciliates (p. 24); and, although experiments on reconstitution
have usually been chiefly concerned with other problems, they add some
evidence. Reconstitution of the gullet of Bursaria after section progresses
from the anterior end, as in fission (Lund, 191 7). The multiple monsters
which sometimes result from incisions in the cell body of Paramecium may
show a considerable number of partial axes, each of which represents at

CERTAIN GENERAL CHARACTERISTICS

35

first an anterior region, the posterior part of the body being undeveloped
(Fig. 12). Development of some of these may continue posteriorly until a
complete individual separates from the mass (Calkins, 191 1). More re-
cently Sonneborn (1932) has described monsters of the ciliate Colpidium
campylum which appear very similar to the Paramecium monsters but re-
sult from culture with a strain of Micrococcus as food. Stylonychia mytilus
has a micronucleus and meganucleus in the anterior half of the body and

Fig. 12. — Multiple Paramecium monsters (from Calkins, 191 1)

another similar pair in the posterior half. In fissio. the two fields from
which the cirri develop are localized near each pair of nuclei. When recon-
stitution occurs after any injury which leaves the two pairs of nuclei in-
tact, the field from which the new cirri develop appears in relation to the
anterior pair of nuclei. Only when these nuclei are removed does a cirrus
field develop in relation to the posterior nuclei. The field of the marginal
cirri always develops in the anterior half of the body, and development
progresses from anterior to posterior (Dembowska, 1925).

36

PATTERNS AND PROBLEMS OF DEVELOPMENT

■ i t

COELENTERATE RECONSTITUTIONS

Decrease in rate of tentacle development and in number of tentacles
appearing in a given time from distal to proximal levels of section has been
observed in pieces of hydra {P elmatohydra oligactis) .^ Reconstitution oc-
curs at all levels of the stem or hydrocaulus of the hydroid Tubularia,
which in some species may attain a length of lo cm. or more without
^ — -^ branching. The hydranth primordium develops

within the perisarc adjoining a cut end by rediffer-
entiation of the coenosarc, the tentacles arising as
longitudinal ridges and separating from the hy-
dranth body progressively from tip to base (Fig. 13).
Emergence of the hydranth is brought about later
by elongation of the coenosarc proximal to it. Evi-
dently hydranth development progresses from the
tip of the piece proximally. Further evidence of this
course of development is given by the reconstitu-
tion of very short pieces into apical parts of hy-
dranths. In these as much of the hydranth is
formed basipetally as length of piece and scale of
organization permit (pp. 344-49). Pieces a centi-
meter or more in length usually give rise to a hy*
dranth at each end (oral and aboral hydranth) ; but,
in general, results of reconstitution vary with length
of piece, body-level, and condition of animal.''

Rate of hydranth development, as measured by
time from section to emergence from the perisarc,
is higher at the distal than at the proximal end of
the piece, except when very short; the rates at the
two ends may then be equal, because the gradient is
practically absent. Sometimes, also, the rates maybe
almost equal at the two ends of extremely long stem
pieces, apparently because the proximal end has become physiologically
isolated from the dominant hydranth and has undergone change in physi-
ological condition as a result. Some species of Tubularia form buds and
new axes under these conditions, that is, when proximal regions come to lie
beyond the range of dominance of the apical hydranth or when its domi-

5 Peebles, 1897; Koelitz, 1910; Weimer, 1928; Rulon and Child, 1937.
^ For experimental data see Driesch, 1897, 18996; Morgan, 19016, 1902a, 190317, 1905,
19066, 1908; Morgan and Stevens, 1904; Child, 1907a-/.

a. i

mn

Fig. 13. — Hydranth
reconstitution in Tubu-
laria; a long hydranth
primordium at oral end
of a distal stem-level and
a short aboral primor-
dium from a proximal
level.

CERTAIN GENERAL CHARACTERISTICS

37

nance decreases. In pieces of equal length from different stem-levels rate
of hydranth development decreases at both ends of the piece with increas-
ing distance of the level of reconstitution from the oral end of the stem,
but in extreme proximal regions the differences may be slight or absent.
Although all earlier investigators who concerned themselves with the
question of rate at different levels of the Tubularia stem were essentially
agreed as to the presence of these differences, Garcia-Banus (1918) pre-
sented experimental data to show that they do not exist. Hyman (19206),
however, showed conclusively that they are present in animals in good
condition and pointed out that the rates recorded by Garcia-Banus were
so much lower than others had found them that his material could not
have been in a condition comparable to that used by others. Hyman's

TABLE

1

Number
OF Pieces

Hours

Distal
Halves

Proximal
Halves

June (room temperature)

December (12°+ 2° C.)

/35

1I8

\36

42

53

71
69

52
73
83

77

data for oral hydranths on distal and proximal halves of the same stems
are given in Table i.

In a more recent study Barth (1938a), taking volume of hydranth
primordium divided by time from section to emergence of hydranth
{irr^L/t) as the measure of rate of reconstitution, obtains much greater
difference in rate at different stem-levels, since stem diameter and length
of hydranth primordium decrease and length of time from section to emer-
gence increases from distal to proximal levels. It seems open to question
whether volume of hydranth primordium should be employed in a measure
of rate in this case. Diameter of the primordium depends on diameter of
the stem, which has nothing to do with reconstitution; and length of the
primordium is not the result of growth but depends on the gradient length
and scale of organization determined a few hours after section, before the
primordium becomes visible (see pp. 101-2). Neither diameter nor length
has anything to do with cell formation or growth; consequently their rela-
tion to rate of reconstitution appears at least doubtful. Driesch (18996),

38 PATTERNS AND PROBLEMS OF DEVELOPMENT

Morgan (igoib, 1903c), and Child (igo'jb-e) have shown that length of
hydranth primordium differs in definite ways at different levels of section
on oral and aboral ends of the same piece and in short pieces with differ-
ence in length of piece. These differences involve differences in localiza-
tion and length of parts of the hydranth, that is, differences in scale of or-
ganization. Some of the measurements are given in order to show what
actual lengths are found and how they differ. Driesch (18996), using two
successive pieces of T. mesemhryanthemiim, each 7-10 mm. in length, gave
the following measurements: oral primordia of distal pieces, 1.67 mm.,
of proximal pieces, 1.21 mm.; aboral primordia of distal pieces, 0.96 mm.,
of proximal pieces, 0.9 mm. Measurements by Child (19076), using two
successive pieces 20-25 mm. in length, were: oral primordia of distal
pieces, 2.26 mm., of proximal pieces, 1.33 mm. With three successive
pieces, each 10 mm. in length, primordium lengths were as follows: oral
primordia of distal pieces, 2.27 mm., of middle, 1.5 mm., of proximal, 1.3
mm. Aboral primordia of the same pieces were, respectively, 0.96, 0.85,
and 0.91 mm. in length. The slight increase in length at the aboral end of
the proximal piece probably results from some degree of physiological iso-
lation of this region. With decrease in length of piece, length of oral and
aboral primordia becomes, in general, more nearly equal; but even in
pieces little longer than the two primordia the aboral may be shorter or
less complete proximally than the oral. The proportions of the primordi-
um, that is, the relative lengths of the four distinguishable zones (Fig. 13),
also show a graded change from distal to proximal stem-levels, and the re-
lation of primordium length to length of hydranth after emergence dif-
fers in the same way (Child, 19076, c).

Differences in rate of development and length of primordium similar
to those in Tuhularia occur in Corymorpha, another gymnoblast hydroid,
with difference in level of reconstitution (Child and Hyman, 1926). The
hydranth of this species, however, develops gradually on the naked stem;
and even during reconstitution contraction and extension occur; conse-
quently, rate of reconstitution and length of hydranth primordium cannot
be determined as exactly as in Tubularia.

The gradient pattern in a branching hydroid is very beautifully shown
by Pennaria cavolinii, not only in the growth form but in the order of de-
velopment of hydranths after their removal and the order of subsequent
degeneration (Cast and Godlewski, 1903). Lateral branches in this species
are oblique to the main axis and show a regular gradation in length basip-
etally. When all hydranths arc removed, new hydranths develop basip-

CERTAIN GENERAL CHARACTERISTICS

39

etally in order from the apical region of the multiaxiate complex and from
the apical end of each branch (Fig. 14, A). Later these hydranths degen-

FiG. 14, A~C. — Reconstitution and degeneration of hydranths in Pomaria cavolinii. A,
reconstitution of hydranths after removal progresses basipetally in the whole and in each
axis; B, degeneration of hydranths and retraction of coenosarc progresses acropetally; C, the
reconstituted axial system at the proximal end persists and grows longer than other parts
(from Gast and Godlewski, 1903).

erate, retraction of the coenosarc in the branches occurs, and the empty
parts of the perisarc drop off. These changes progress from the most basal

40

PATTERNS AND PROBLEMS OF DEVELOPMENT

branches acropetally (Fig. 14, B). Meanwhile the original apical region
of the main axis and the new multiaxiate individual which has developed
from the proximal cut end may grow and give rise to new hydranths; evi-
dently they are able to live at the expense of lower axial levels (Fig. 14, C).
A gradient pattern is evident in this whole series of events. The rate of re-
constitution decreases basipetally in the whole and in each axis; as the
animals starve in the laboratory, lower axial levels are used in maintaining
higher levels. Possibly differential tolerance to laboratory conditions is
also involved (see chap. v). The reconstituted multiaxiate individual at
the proximal end undoubtedly represents younger, more intensely active
tissues than the old coenosarc, perhaps even more active than the original
apical region, and so is able to persist and grow for a longer time than any
other part. A few unpublished experiments with P. tiarella indicate that

TABLE 2

Levels

Days

3

8

14

26

1-2

0-5
0
0
0

7-10

6- 8

5- 7
0
0

20-25

lo-is

S-io

2

0

12-20

b

12-15

C

d

e

12-18

10
0-5

a similar gradient exists in that form, though less regular, as the less regu-
lar growth form of the species would suggest.

A gradient in rate of reconstitution in Obelia has been noted by Billard
(1904) and Lund (1923a) and in Eudendrium by Goldforb (1907). In the
sessile scyphozoan Haliclystus auricula (Child, 1933^) and in the actinians
Cerianthus solitarius (Child, 19030), C. aestuarii (Child, 1908), and Haren-
actis attenuata (Child, 1909a) a similar gradient in rate occurs. Table 2
gives tentacle length for Harenactis pieces at different times after section
at the five body-levels indicated in Figure 15, ^. The tentacles from the
more distal levels appear earlier, and in general grow more rapidly and
attain greater length, than those of levels farther proximal. In the last
column, 26 days, the tentacles of the a-pieces have begun to decrease in
length, as all do sooner or later in animals kept without food and not per-
mitted to burrow in sand. In these pieces the aboral ends were intact, and
consequently the pieces differed greatly in length; but other experiments

A

Fig. i6

Fig. 15, A,B. — Harenactis attenuata. A, outline indicating levels of section, a-e of Table 2;
B, bipolar reconstitution from esophageal region; tentacles of disk at oral end develop more
rapidly and become longer than those of aboral disk (from Child, igoga, h).

Fig. 16. — Outline of planarian indicating approximate levels at which fission can be in-
duced in long animals.

42 PATTERNS AND PROBLEMS OF DEVELOPMENT

showed that length of piece became significant only in shorter pieces than
those used here. Short pieces from the esophageal region often show bi-
polar reconstitution with development at the original oral end more rapid
than at the aboral (Fig. 15, -B).

Some studies of coelenterate reconstitution have been chiefly concerned
with other points and give no information concerning differences in rela-
tion to body-level. Moreover, the original gradient differences in some of
the more sensitive branching hydroids apparently decrease or even disap-
pear under laboratory conditions, particularly if the animals are kept in
standing water; hydranths degenerate and stolons develop in place of hy-
dranths over the whole stock.

Although ctenophores are not now regarded as coelenterates, it may be
noted here that no evidence of gradient difference has been observed in the
reconstitution of Mnemiopsis (Coonfield and Goldin, 1937), although
other evidences of an apical-oral gradient have been found (pp. 106, 327).

BODY-LEVEL AND RECONSTITUTION IN TURBELLARIA

Among the turbellaria the triclads are the most interesting forms in this
connection because of the great capacity for reconstitution of certain
species of the group and various limitations of this capacity in other spe-
cies, because reconstitution in the group has been studied by many in-
vestigators, and because the gradient pattern is clearly evident in it."

The described American species of Dugesia undergo fission at definite
body-levels posterior to the mouth. The act of fission, which consists of an
independent motor reaction of the posterior zooid region, is not usually
preceded by visible morphogenesis but is a transverse rupture followed by
reconstitution. This rupture, however, occurs at a definite body-level, and
various lines of physiological evidence indicate that in animals above a

' Recent taxonomic studies of American planarians by Hyman (1931) and Kenk (193511)
have resulted in extensive change in names of genera and species. According to Hyman,
the genus Planaria does not occur in the United States, and the genus name Euplanaria re-
placed it for American species. However, it has recently been discovered by Hyman (1939)
that the genus name Dugesia has priority over Euplanaria. This is an unfortunate discovery,
for the names Planaria and Euplanaria are good descriptive designations and the anglicized
"planarian" has definite meaning. Now American forms of the genus become Dugesia, while
the European Planaria remains unchanged. The species distinguished by Hyman (1931) as
Euplanaria maculata and E. novangliae, are, according to Kenk, a single species; and for this
species the name Dugesia tigrina has priority, according to Hyman and Kenk. This species is
regarded as including the form named P. lata by Sivickis (1923); see also Watanabe, igssb-
Planaria simplicissima (Curtis, 1900) becomes Curtisiaforemanii; P. velaia becomes Fonticola
velala; and the American dendrocoelid, often called Dendrocoelum lacteum, is Procotyla fluviatilis
(Hyman, 1931). The revised nomenclature is used in the following pages.

CERTAIN GENERAL CHARACTERISTICS 43

certain length the posterior part of the body consists of one or more zooids,
the number of zooids increasing with body length.^ Occurrence of a second
set of sex organs in the posterior zooid is certainly rare but has been ob-
served by Kenk (1935a) and is regarded by him as morphological evidence
of the presence of a posterior zooid. Fission can be prevented or induced
experim.entally in D. dorotocephala. When it is prevented, animals 30-35
mm. in length, sometimes even longer, result. These give evidence of the
presence of two, three, or more zooids, and fission can be induced at sev-
eral levels (Fig. 16, F, F' , F") of the postpharyngeal region (Child, 1910a).
In the species which do not undergo fission, no evidence of any kind indi-
cating presence of a posterior zooid has been found. Reconstitution in

17 18

Figs. 17-20. — Planarian regeneration. Fig. 17, difference in rate at anterior and posterior
surfaces of section and between median and lateral regions of a single surface. Fig. 18, more
rapid regeneration ventrally than dorsally. Figs. 19, 20, regeneration in certain concentrations
of Ringer solution without contraction of surface of section; cut ends of gut unite with ecto-
derm forming an opening into gut anteriorly and posteriorly; greater tolerance to Ringer an-
teriorly than posteriorly.

most planarians consists of more or less regeneration at surfaces of section
with redifferentiation elsewhere, but in some it consists chiefly or wholly of
regeneration.

As regards rate of reconstitution, it may be noted first that in isolated
pieces, except when very short, regeneration in the earlier stages is more
rapid at the anterior, than at the posterior, end of the piece (Fig. 17).''
Since this dift'erence in early stages has no constant relation to the final
amount of regeneration, it suggests the existence of a longitudinal differ-
ential or gradient in the planarian body. Early growth of new tissue is

* Child, 1910a, 191 1&, d, 1913&, 1930; Hyman, 19236.

9 This difference has been observed in D. ligrina and D. agilis (Child, unpublished), in the
form called Planaria lata by Sivickis (1923), in a Japanese species resembling D. dorotocephala
(Child, 19326), in Phagocata gracilis (Buchanan, 1933), and repeatedly by groups of students in
the study of reconstitution of D. dorotocephala.

44 PATTERNS AND PROBLEMS OF DEVELOPMENT

also more rapid in the median ventral region than laterally and dorsally
(Figs. 17 and 18), suggesting a differential from median ventral laterally
and dorsally. These relations may be altered by incidental or experimen-
tal conditions; anterior or posterior outgrowth may be inhibited differen-
tially (pp. 175-76). One example of some interest may be mentioned here.
When pieces develop in approximately isotonic or hypertonic concentra-
tions of Ringer solution, the cut ends do not contract, as they do in water;
the cut end of the gut remains open and usually unites with the ectoderm
at both ends of the piece, so that openings into the gut persist at both ends
and the regenerating tissue is more or less divided (Figs. 19, 20). When
level of section is through the pharyngeal pouch, its wall may unite with
the ectoderm and so give rise to a terminal "mouth" with growth of new
tissue on each side of it. The inhibition of median regeneration in these
cases is evidently an incidental result of failure of the cut surfaces to con-
tract in the salt solution and does not represent alteration in gradient re-
lations. If the divided new tissue is removed and the piece returned to
water, regeneration occurs as usual. Moreover, the wound does not re-
main open, as was maintained by J. W. Wilson (1926).

Rates of development of the head at different body-levels may be com-
pared by determining the time between section and appearance of the
black pigment of the eyespots. According to a statistical study (Wata-
nabe, 1935a) rate of head development, as determined in this way, de-
creases from anterior to posterior levels as far as the fission zone and in-
creases again in the posterior zooid region of D. dorotocephala. Other
planarian species investigated show a similar relation between rate and
body-level, except that in those that do not undergo fission the decrease
extends to the posterior end of the body or as far posteriorly as head regen-
eration occurs." Rate of development of posterior new tissue shows a sim-
ilar differential, slight in some species, strongly marked in others, except
at extreme anterior levels; heads isolated by section at levels only slightly
posterior to the eyespots heal posteriorly without further regeneration;
sectioned at levels slightly farther posterior, they regenerate posterior ends
incompletely or slowly. Posterior to these levels the minimum length of
piece which reconstitutes a complete, normal individual increases posteri-
orly to the fission zone in species which undergo fission and is again less

"Evidence of similar difference of rate has been found in the following: a European
form called P. dorotocephala but probably another species (Pourbaix, 193 1) ; a Chinese planarian
(Li and Shen, 1934); P. gonocephala (Abeloos, 1930); Dendrocoehim lacteitm (Sivickis, 1931a,
1933); Procotylajhiviaiilis (Child, unpublished); Phagocata gracilis (Buchanan, 1933).

CERTAIN GENERAL CHARACTERISTICS

45

in the posterior zooid; in most other species it increases to the posterior
end or as far posteriorly as head regeneration occurs, but even very small
fragments of Fonticola velata from any postcephalic level reconstitute com-
plete individuals, though only after encystment, practically complete de-
generation of internal organs, and reorganization of body wall and paren-
chyma."

Marked differences in scale of organization of the preoral region occur in
planarian reconstitution in relation to level of origin of the piece. These
consist in a gradual decrease in relative length of the preoral region and
more anterior localization of the pharynx in successive pieces of equal
length from the most anterior level posteriorly to the fission zone, except
in pieces from pharyngeal levels; in those the pharyngeal region or some

Fig. 21, A-G. — Differences in length of prepharyngeal region and in position of pharynx in
pieces from different body-levels of Dugesia tigrina. A-E, decrease in prepharyngeal length in
successive pieces from anterior to posterior end of anterior zooid; F, G, increase in prepharyn-
geal length in pieces from posterior zooid region.

part of it is already present, but its level may undergo gradual change in
consequence of differential growth in relation to scale of organization
(Fig. 21, A-E). In the region of the posterior zooid length of preoral re-
gion is again greater (Fig. 21, F, G). In species without a posterior zooid
region decrease in preoral length continues to the posterior end. Pieces
from anterior and posterior levels of Curtisia foremanii (Fig. 22, A, B) and
similar pieces from young individuals of Fonticola velata reconstituting
without encystment show this difference in scale clearly."

Scale of organization of the new head in earlier stages shows a similar
relation to level of origin of the piece. It decreases from the anterior level
to the fission zone and is again larger in the posterior zooid region in the
species with posterior zooids (Fig. 21). In those without such zooids the

" For fragmentation, encystment, and reorganization in F. velata see Child, 1913c, igi^d;
W. A. Castle, 1928.

'^ For further data see the following: Morgan, 1898, 1900a, 1901c; Child, 1906, 1911&,
1913c, igi^d; Sivickis, 1923; Buchanan, 1927.

46

PATTERNS AND PROBLEMS OF DEVELOPMENT

decrease continues to the posterior end (Fig. 22). As these figures also
show, the length of regenerated tissue posterior to the eyespots increases
from anterior to posterior levels of the anterior zooid and is again less in
the posterior zooid region (Fig. 21), or in species without posterior zooid it
increases to the posterior end (Fig. 22). In other words, with increasing
distance of level from the anterior end, an increasing proportion of the
anterior region of the new individual is formed by regeneration rather than
by redif^erentiation, with decrease in regeneration again posterior to the
fission zone when it is present. These differences in relative size and pro-
portion of parts, as primarily localized and determined in pieces from dif-
ferent body-levels, gradually approach the normal in consequence of dif-

P^iG. 22, A-D. — Differences in prepharyngeal length and position of pharynx in pieces from
anterior and posterior levels. A, B, Curtisia foremanii; C, D, Fonticola vclata.

ferential growth. Parts that are "too small" grow more rapidly than other
parts in later stages, or in animals kept without food, at the expense of
other parts with reduction in size of the whole. Changes in proportion are
much more rapid in fed than in starved animals.

Certain planarian species show increasing inhibition of head develop-
ment with decrease in length of piece below a certain fraction of body or
zooid length. The length of piece at which inhibition becomes evident in-
creases from anterior levels posteriorly or to the fission zone, but in the
posterior zooid region it becomes very small. In pieces of equal length be-
low a certain fraction of body or zooid length the degrees of inhibition in-
crease from anterior to posterior levels, or to the fission zone, and are much
less in the posterior zooid.'-' The inhibited head forms and conditions
which determine them are discussed in other connections (pp. 177-96).

The polyclad turbellarian Lcptoplana docs not reconstitute a head

'3 For data see Child, 191 ib; Child and Watanabe, 1935a; Watanabe, 19356; Sivickis, 1923,
i93ici, b, 1933; Abcloos, 1930; Buchanan, 1933.

CERTAIN GENERAL CHARACTERISTICS 47

anteriorly from levels posterior to the cephalic ganglia but does regenerate
posterior parts at all levels from a short distance posterior to the ganglia
to the posterior end of the body (Child, igo^b). Both rate of growth and
total length of posterior new tissue produced in pieces kept without food
decrease from anterior to posterior levels. Rates and amounts of regen-
eration are considerably greater at any level when the cephalic ganglia
are present than when they are absent. Pieces without ganglia are in-
capable of normal co-ordinated locomotion and show little motor activ-
ity. Apparently growth of the posterior regenerating tissue is stimulated
in the normally active animal. Various rhabdocoel species show essen-
tially similar conditions.

DATA FROM OTHER GROUPS

Some species of nemerteans show even greater capacity for reconstitu-
tion than planarians. According to Nusbaum and Oxner (191 1), the
slender form of Linens ruber reconstitutes complete individuals at all
levels; but pieces from the middle region have the highest rate. More
recently Coe (193 1) has found in L. vegetus and L. socialis in general a
decrease in rate from anterior to posterior levels, but with considerable
variation.

The many studies of regeneration in the annelids show very different
degrees of regenerative capacity in different species. Briefly stated, the
relation between regeneration and body-level in polychetes and oligo-
chetes is, in general, as follows: In many forms studied the rate of
regeneration at different levels has not been determined; but, according
to the data at hand, rate of head regeneration, and in many species ability
to regenerate a head, decrease from anterior to posterior levels. In some
species a head is regenerated only at extreme anterior levels. Ability to
regenerate a posterior end usually increases from extreme anterior levels
over a certain portion of body length, which is very different in different
species; and at more posterior levels rate of posterior regeneration de-
creases. In short, the differences in rate of regeneration at different body-
levels in those forms which are capable of regeneration over most of the
body length do not differ very greatly from those in planarians; but
capacity for, and rate of, posterior regeneration usually increase over a
greater distance from the anterior end than in planarians, though not
necessarily greater in proportion to body length. ^^

'■•See Hescheler, 1896; Morgan, 1897; Morgulis, 1907, 1909; H. R. Hunt, 1915; Hyman,
1916a; Korschelt, 1919; Sayles, 1934; and references cited by these authors. See also Sayles,
1940, Biol. Bull., 78, 3.

48 PATTERNS AND PROBLEMS OF DEVELOPMENT

Many annelid species regenerate no more than a certain number of seg-
ments anteriorly, even though more than that number have been re-
moved. This number is characteristic for the species and may be re-
garded as representing a more or less specialized head region.'^ Recon-
stitution by redifferentiation of some of the old segments posterior to the
regenerated anterior end into more anterior segments replacing those
which were removed above the number regenerated is known to occur in
some species; but for others no data are at hand, though such reorganiza-
tion of old segments is probably of very general occurrence. Here, then,
as in planarians, reconstitution — at least in certain species — is in part re-
generation, in part redifferentiation or reorganization of old parts.
In some annehds, however, number of segments regenerated anteriorly is
apparently not definitely limited but is more directly dependent on the
number removed. In these cases there is probably less or no reorganiza-
tion of old segments. Whether a difference in scale of organization in rela-
tion to body-level occurs in annelids does not appear from the data avail-
able, but an inhibition of head regeneration in relation to body-level and
length of piece, much as in planarians, appears in Lumbriculus (Hyman,
1916a). As regards existence of a ventrodorsal or a transverse differential,
it is found that outgrowth of new tissue from a cut end is, in general, more
rapid in the median ventral region than elsewhere.

Relations between the different axes in certain of the branching
bryozoan species resemble, more or less, those in the hydroids. When tips
are removed from several branches at the same time, reconstitution begins
first on the chief branch, also the ability to regenerate decreases basip-
etally in each axis (Otto, 1921).

In most other animal phyla the capacity for reconstitution is usually
more narrowly Hmited. Asteroid starfishes regenerate arms, and a few
species are able to regenerate disk and other arms from a single arm with-
out any portion of the original disk. Each arm develops like a new polar
axis from the tip basipetally, and outgrowth of new tissue is more rapid
in the median ventral region than elsewhere; that is, a longitudinal and
ventrodorsal gradient are indicated in development of the arms. Rate of
regeneration of larval legs of the insect Cloe diptera decreases from anterior
to posterior levels (von Ubisch, 191 5). Attention directed to this point
would perhaps show similar differences in other arthropods. Some of the
ascidians are capable of extensive reconstitution. Pieces of Clavelina from

■s Morgan, 1897; Hyman, 1916a; Gross and Huxley, 1935; Berrill, 1936". b- Sec also Hyman,
1940, Amer. Nat., 74, No. 755.

CERTAIN GENERAL CHARACTERISTICS 49

the "esophageal" region proximal to the branchial chamber develop into
bipolar forms with a branchial chamber at each end, but distal develop-
ment is more rapid than proximal (Brien, 1930; Pasquini, 1933).

THE QUESTION OF REGENERATION IN RELATION TO DEGREE OF INJURY

Increase in rate of regeneration with increase in size or number of parts
removed has been reported for certain forms and has raised the question
whether increase in degree of injury increases the rate.'^ In this connec-
tion it is of interest to inquire whether the differences in rate of recon-
stitution at different levels of the longitudinal axis noted earlier in this
chapter are similarly related to degree of injury. Rates of hydranth re-
constitution in hydroids and of apical reconstitution in other coelenterates
and rate of head regeneration in planarians and annelids decrease from
anterior to posterior levels, that is, with increase in size of the part re-
moved anteriorly, obviously an increase in degree of injury. Evidently
this increase is not concerned in these cases in determining rate of regen-
eration, for the rate is highest when the injury is least. Posterior regen-
eration in planarians and annelids is usually absent, incomplete, or slow at
levels close to the head, that is, when degree of injury is greatest; at levels
a little farther posterior, with a lesser degree of injury the rate is highest
and decreases from these levels posteriorly. This decrease might perhaps
be regarded as related to decrease in degree of injury if head regeneration
did not decrease in rate in the same direction with increasing degree of in-
jury anteriorly. If rate at either end is affected by degree of injury at both
ends, little difference in rate in pieces of equal length from different levels
should occur, for degree of injury is much the same in such pieces; if the
part removed anteriorly is small, the posterior part removed is large, and
vice versa. In short, these differences in rate at different levels of the
longitudinal axis appear to be related to inherent graded differences in
physiological pattern at different levels rather than to degree of injury.
Moreover, the scale of organization of the oral hydranth in Tuhularia and
Corymorpha is largest when the degree of injury at the oral end of the piece
is least, and that of the aboral hydranth is largest when degree of aboral
injury is greatest. It appears probable that in cases of apparent relation
between rate of regeneration and degree of injury, for example, the in-
crease in rate in the starfish with removal of an increasing number of
arms, the increasing disturbance of function, is a stimulus to regenerative
growth.

'^Zeleny, 10050, b, 1909; Ellis, 1907, 1909.

50

PATTERNS AND PROBLEMS OF DEVELOPMENT

RECONSTITUTION AFTER SECTION OBLIQUE TO THE LONGITUDINAL AXIS

Differences in rate of reconstitution in relation to different levels of a
cut end oblique to the longitudinal axis have been observed in a number
of coelenterates and planarians. When stem pieces of Tubularia or
Corymorpha are cut with plane of section strongly oblique to the longi-
tudinal axis, the distal tentacles of the hydranth often develop first at the

Fig. 23. — Haliclystiis auricula. Lines AA-GG, XX, IT, indicate levels and directions of
section (from Child, i933&).

most oral region and progressively to lower levels, forming a ring of distal
tentacles oblique to the polar axis of the stem. In these forms, however,
closure of cut ends is rapid, and the oblique surface may become rounded
before tentacles develop. When this occurs, distal tentacles develop near
the oral end as usual, but the plane of the tentacle ring is transverse, not
oblique.

Difference in rate in relation to an oblique cut surface appears much
more clearly in the sessile scyphozoan HaUdystus. After transverse sec-
tion at various levels of umbrella or stalk (Fig. 23, AA-EE) reconstitution

CERTAIN GENERAL CHARACTERISTICS

SI

of the distal part removed occurs as indicated in Figure 24, A and B. If
the plane of section is strongly oblique to the longitudinal axis (Fig. 23,
FF), reconstitution occurs most rapidly on the most distal part, and the
rate decreases proximally until at the most proximal levels nothing more
than wound-healing may occur (Fig. 24, C). If the oblique section is made
so that one or two of the original marginal organs remain (Fig. 23, GG),
reconstitution of marginal structures may be largely or even com-

LA

A

D

Fig. 24, A-D. — Reconstitution in Haliclystus. A, early stage after transverse section at
levels BB or CC of Fig. 23; B, advanced stage after section at level EE of Fig. 23; C, early
stage after obliciue section with removal of all marginal organs; D, reconstitution after oblique
section leaving small part of original margin (from Child, 1933&).

pletely inhibited (Fig. 24, D). Later removal of the original marginal
parts, even after a week or two, is followed by development of at least
some of the marginal structures previously inhibited. When only a small
part of the marginal region is removed, as indicated by A'A'^ and YY in
Figure 23, little or no development of the parts removed occurs (Child,

1933^)-

After oblique section of the scyphistoma of Aurelia the rate of recon-
stitution decreases from distal to proximal levels of the oblique surface, as
in Haliclystus. Tentacles from the most distal level attain full length in a

52

PATTERNS AND PROBLEMS OF DEVELOPMENT

few days; but, although tentacle development is not completely inhibited
at the most proximal levels, the full normal number, sixteen, is not usually
attained, even after 3 or 4 weeks, and the tentacles farthest proximal ap-
pear last and develop very slowly. Experiment on another species of
stalked scyphozoan {Thaumatoscyphus distinctus) shows a delay of recon-
stitution at proximal levels of an obhque cut surface more nearly like
that in the scyphistoma than in Haliclystus (Hanaoka, 1935). In these
forms there is not only a difference in time of appearance of marginal
structures at the different levels of the oblique section but a continuing
dominance of the distal region which greatly retards reconstitution at the
more proximal levels or, in Haliclystus, completely inhibits it. Similar

_ CD

Figs. 25, yl-D.— Reconstitution after oblique section. Fig. 25, A, early, and B, late, stages
of distal reconstitution in Cerianthiis solitaries; C, anterior and posterior planarian regenera-
tion after oblique section; D, development of both head and posterior end on oblique posterior
cut end of short piece.

differences in rate on an oblique cut end appear in the actinians Ceriantkus
solitarius (Child, 1904a) and C. aestuarii (Child, 1908), as indicated in
Figure 25, ^ and B. In both species development at the most distal level
is more rapid, and that at the proximal level less rapid, than on transverse
cut surfaces at corresponding levels.

On strongly oblique anterior cut ends of planarian pieces head regen-
eration is localized on the more anterior part; and the more anterior side
of the head, including the eyespot, usually develops more rapidly than
the other side (Fig. 25, C, /)). This asymmetry constitutes further evi-
dence of difference in rate of head development at different body-levels. A
regenerating posterior end is localized on the more posterior part of an
oblique posterior cut surface (Fig. 25, C). The posterior end is a subordi-
nate part and is determined by levels anterior to it (pp. 339-40). Physio-
logical dominance of more anterior levels is undoubtedly concerned in

CERTAIN GENERAL CHARACTERISTICS 53

determining its position on the oblique surface. That this is the case is
further indicated by occasional development of a head on the more an-
terior part of the posterior, as well as of the anterior, oblique end in short
pieces and more rarely of both head and posterior end on the posterior
oblique surface (Fig. 25, D).^''

In the reconstitution of short pieces half, or less than half, the width of
the planarian body the head may develop on the angle between anterior
and median cut surfaces or entirely on the median, instead of on the
anterior, surface. The longitudinal axis thus determined is oblique or
vertical to the original anteroposterior axis. In these heads the eyespot
of the side nearest the original anterior end often appears first, and the
whole head may be asymmetrical in earlier stages.'*

VISIBLE EVIDENCES OF PATTERN IN EARLY EMBRYONIC DEVELOPMENT

The earliest indications of embryonic order and pattern are usually
regional differentials or gradients of one kind or another — of amount of
yolk, visible character and staining of the cytoplasm of the egg or cleav-
age stages, rate of cell division, size of cells, rate of growth, rate of dif-
ferentiation, etc. Some of these evidences of pattern — ^yolk gradients, for
example — are present in undivided eggs of some species; others appear
during cleavage or later; still others, only as particular organ systems
begin visible development. Very commonly pattern is first evident along
what becomes the polar axis and becomes evident in other directions
corresponding to other axes as development proceeds. However, some
eggs exhibit, at or before the beginning of development, more or less
definite regional differentiations in the cytoplasm which are, or become,
features of embryonic pattern. So-called "mosaic development" furnishes
numerous examples of this type of pattern (see chap. xiv). Such eggs may
be, and now usually are, regarded as having already undergone a greater
or less degree of regional cytoplasmic differentiation which becomes the
basic of embryonic differentiations. Since these differentiations are, or be-
come, definitely related to axiate embryonic pattern and since they occur
only in certain species, they probably represent derivatives, effects, ex-
pressions of pattern, rather than the primary pattern itself. But whatever
the condition of the egg cytoplasm at the beginning of development,

" These oblique reconstitutions in planarians have been repeatedly described. See, e.g.,
Morgan, 1900, 1901c, 19046,- Bardeen, 1902; Rand and Boyden, 1913. They have been used
by the writer as class experiments for many years.

'* Morgan, 1900, 1901c, 1904&; Child, 1915c, p. 164; Olmsted, 1918; Beyer and Child, 1930.
See also pp. 365-66.

54 PATTERNS AND PROBLEMS OF DEVELOPMENT

gradients in rate of cell division, in size of cells, and in rate of growth and
of differentiation appear in relation to the axes of the whole and of parts
as development progresses.

The very general occurrence of an anteroposterior developmental
gradient has been recognized in the so-called "law of anteroposterior,
cephalocaudal, or craniocaudal development.'"^ But developmental gra-
dients equally definite in character and equally constant in occurrence
appear in other directions — for example, from the ventral region laterally
and dorsally in turbellarians, annelids, and arthropods and from the mid-
dorsal region laterally and ventrally in vertebrates. In certain other
groups — coelenterates, trematodes, cestodes, nemerteans, nemathel-
minths, mollusks, and echinoderms — an apicobasal or longitudinal de-
velopmental gradient appears more or less clearly, at least in the egg, in
cleavage or in larval stages; but other gradients differ according to de-
velopmental pattern in different members of these groups. As develop-
ment proceeds, new gradients may arise at various levels and in various
directions with localization of organ primordia. Every such localization
alters, and may even obHterate or reverse, the pre-existing gradient or
gradients. Development of the first hydranth from the originally basal
end of the planula of calyptoblast hydroids with reversal of the original
polarity is an example (pp. 96, 97). Many axiate organs and organ sys-
tems make their appearance as localized areas or fields of developmental
activity in which the activity decreases from a region more or less cen-
trally or otherwise localized in the area. A tentacle, for example, usually
begins its development as a region in which developmental activity de-
creases more or less radially from a center and becomes axiate in conse-
quence of greater growth of the central region ; its developmental pattern
is essentially like that of other buds. Conditions in early developmental
stages of many appendages are apparently similar, but the most active
region is not necessarily at the geometrical center of the field. In short,
observation alone gives evidence of gradients and gradient systems ap-
pearing and undergoing change in definite order and relation as char-
acteristic features of earlier embryonic developmental stages of many
organisms and organs. Gradients appear to precede sharply defined and
bounded morphogenetic differentiation.

Even in advanced stages of development, in which the changes consist

'9 Morphologically, this gradient is merely a timetable of events, but there must be a
physiological basis for this timetable. In this connection see Kingsbury, IQ24, 1926, 1932;
Child, 1925c.

CERTAIN GENERAL CHARACTERISTICS 55

largely of growth in size, growth gradients appear in many parts as either
absolute or relative differences in growth rate. Huxley's recent quantita-
tive study of relative growth is of interest in this connection (Huxley,
1932).^° Growth gradients and allometric growth occur very generally in
the course of development. Allometric growth, that is, growth of a part at
a different rate from growth of the whole, appears almost universally in
the apicobasal or longitudinal axis at some stage of development. In the
hydroids Tuhularia and Corymorpha length of hydranth in early stages
may be half or more than half the total body length, but both absolute
and relative lengths of stem increase until in large individuals stem length
is many times hydranth length. Evidently the stem grows in length more
rapidly than the whole. In planarians also, relative growth in length with
increase in total length increases from anterior to posterior levels (see,
e.g., Abeloos, 1928). Embryologically the pharynx arises near the pos-
terior end, but in long individuals it lies anterior to the middle. This
higher growth rate of posterior regions is most conspicuous in those
species which develop a posterior zooid or zooids and undergo fission; but
in these forms it is not, properly speaking, growth of a single individual.
Such growth gradients do not necessarily correspond in direction with the
primary gradient of the longitudinal axis, and certain characteristics of the
primary gradient may persist independently of growth gradients of later
stages. In Tuhularia and Corymorpha the high end of the gradient, indi-
cated by respiration, differential susceptibility, and reduction of vital
dyes, is the hydranth; but it is the low end of the gradient of longitudinal
growth in postembryonic stages. In fact, growth is very often greater at
the lower than at the higher levels of gradients, indicated by differences
in respiration and various other physiological factors.^^ In general, the
correspondence between these gradients indicated by different methods
appears to be closer in earlier than in later stages of development. Un-
doubtedly, gradients characterized by other activities than growth occur.
In fact, higher growth rate at lower levels of such gradients may occur
because other activities are less intense there.

Graded morphological differences corresponding to the physiological
gradients present often appear, particularly in the longitudinal axes, in
adult animals. Distribution of setae, pigmentation, and nephridia of cer-

^^ Huxley and Teissier (1936) have suggested that the term "heterogony," used by Huxley
(1932), be replaced by the term "allometry."

^' In his discussion of gradients Needham (1931, pp. 582-606) has apparently failed to
grasp this point, that a gradient indicated by differences in respiration, susceptibility, etc.,
does not necessarily correspond to a growth gradient in the same axis.

56 PATTERNS AND PROBLEMS OF DEVELOPMENT

tain oligochetes show such graded differences along the anteroposterior
axes of certain ohgochetes coinciding with the physiological gradients
present (Pickford, 1930; Sivickis, 1930). The functional activities of
various organs also give evidence of the presence of physiological gradients
of some sort. Transmission of impulses may occur more rapidly, or with
greater range, in one direction than in the reverse, as in the ctenophore
plate row (Child, 1917c, 1933a). Isolated pieces of the mammalian intes-
tine show a gradient decreasing posteriorly in rate of rhythmical contrac-
tion and a gradient in the same direction of excitability and various other
physiological characteristics (Alvarez, 1928).

CONCLUSION

It has already been noted that buds provide material much more
accessible than the ovarian oocyte for obtaining information concerning
the beginnings of developmental pattern. Development of an individual
from a bud is physiologically one of the most interesting and important
forms of development occurring under natural conditions. Bud pattern
may originate in a single cell or in a cell group which has previously con-
stituted a part of an individual, and cells from different parts of the indi-
vidual may give rise to buds. They have not had the same past history,
nor have they developed an organization, a pattern, like that of the egg.
As far as they are concerned, the new pattern may be entirely accidental:
they merely happen to be in the locus of the activation. The transforma-
tion of the radial gradient system primarily characteristic of buds into an
axial system by differential growth is highly suggestive for the problem of
axiate pattern. Is the physiological polarity which develops from the pri-
marily radial pattern anything more at its inception than an altered
spatial order of the radial gradient system?

Fissions involve reconstitution with either modification of a pre-existing
partial pattern or development of new pattern; but they have the dis-
advantage, for certain experimental purposes, that they usually occur
only in certain regions of the parent body and under certain conditions,
though it is sometimes possible to induce them experimentally. The re-
constitutions following section can be initiated at will with variation of
form, size, region or origin of the isolated part, and physiological condition
of the parent individual, as affected by age or nutrition or as altered ex-
perimentally. Also, the effect of experimental conditions on developmen-
tal pattern of isolated pieces of different origin and in different physio-
logical condition can be investigated. Only certain directly distinguish-

CERTAIN GENERAL CHARACTERISTICS 57

able characteristics of reconstitutional patterns have been considered in
this chapter, but these have brought to hght certain points of interest.

It is evident that, even though they finally become similar individuals,
the patterns of reconstitution in pieces from different body-levels are pri-
marily different and that these differences are, in general, graded in char-
acter and indicate graded differences of some sort in the parent body.
Except for failure of extreme apical or anterior regions to reconstitute
more basal or posterior parts and absence of head regeneration at the
more posterior body-levels in certain species, the differences in pattern
are quantitative, rather than qualitative or specific, differences in rate,
size, length, and proportion rather than in the kind of organs which de-
velop. These facts suggest that the differences are expressions of es-
sentially quantitative differences along the parent axes, whatever the
specific organ differences which may be present at different levels. If we
admit this, it follows that these quantitative differences in the parent
body are essential factors in determining developmental pattern in the
isolated pieces.

Gradations or gradients of some sort, whose presence is indicated in one
way or another in development under natural conditions, are character-
istic of embryonic, as well as of other developmental, patterns. Considera-
tion of the question of their significance in embryonic pattern and of their
possible relations to the apparently specific regional differences in egg and
embryo, which are revealed by experimental procedures, is postponed un-
til other experimental data bearing upon the problem of pattern have
been presented.

CHAPTER III

CONCERNING METHODS OF PHYSIOLOGICAL ANALYSIS

IN THE attempt to throw light on the problems of development many
methods have been developed and employed and have given a large
body of evidence. Since the mature organism is the result of develop-
ment, it, as well as earlier stages, has been used as material in attacks on
problems of development, but the data obtained are of interest chiefly in
the light of what is learned from earlier stages. Some consideration of
these methods, their objects, their limitations, and particularly some of the
difficulties involved and the precautions necessary appears desirable as a
preliminary to discussion of results obtained with them.

METHODS FOR DETERMINING RESPIRATION

Since the axial developmental gradients appear to involve protoplasmic
activity of some kind and since respiration, as determined by oxygen con-
sumption and carbon dioxide production, is a measure of certain proto-
plasmic activities and, to greater or less degree, an indicator of others, the
question at once arises whether differences in rate of respiration bear any
relation to the gradients or to other features of developmental pattern.
Unfortunately, most of the methods available at present for direct deter-
mination of respiratory rate are rather unsatisfactory for the purpose of
determining whether such differences exist, because they make it necessary
to separate different regions of the body and to determine respiration in
the isolated pieces. This procedure introduces various difficulties and com-
plicating factors, some of which cannot be entirely overcome or avoided ;
but if they are not considered and ehminated or controlled as far as possi-
ble, the results obtained are without definite significance. When an or-
ganism, particularly an animal, is separated into pieces for determinations
of respiratory rate of different regions along an axis, several questions
arise: First, is respiratory rate altered in consequence of section? Second,
if alteration occurs, is the rate increased or decreased? Third, are pieces
from different levels of the axis affected to the same degree by section?
Fourth, if alteration of rate occurs after section, is it persistent or tem-
porary? Fifth, if the alteration is temporary, does the rate sooner or later

58

METHODS OF PHYSIOLOGICAL ANALYSIS 59

become the same as that of the region of the intact body which the piece
represents, or does it remain above or fall below that? Sixth, does loss of
body fluids or blood in consequence of section affect the respiratory rate
of pieces? Seventh, in pieces which undergo reconstitution is respiratory
rate altered by the reconstitutional activities, and, if so, how soon after
section does this alteration occur? Eighth, if pieces are motile, is motor
activity of pieces from different body-levels different in degree, and, if it
is, to what extent does this affect respiration? In certain cases the ques-
tion of occurrence or of intensity of anaerobic respiration must also be
considered.

At present it is not possible to answer all these questions with certainty.
For example, we cannot be entirely certain whether the respiratory rate
of an isolated piece from a certain body-level is the same as when the piece
was a part of the intact body; nor can we determine with certainty when
reconstitutional activity begins to affect the respiratory rate, although it
appears beyond question that it does increase the rate. As a matter of
fact, in most animal species thus far investigated the respiratory rate is un-
doubtedly increased, sometimes very greatly, immediately following sec-
tion and gradually decreases during one to several hours following; the
rate of all the pieces together may finally become about the same as that
of the animal before section, but we have no means of knowing that the
rate of each piece is the same as when it was part of the intact body. The
total rate of pieces of some other animals may fall below that of the intact
animal; consequently, the significance of differences in pieces from differ-
ent levels becomes still more uncertain. Motor activity presents another
difficulty in motile animals. In planarians, for example, pieces above a
certain fraction of body length from anterior regions of the body show, in
general, more motor activity for some time after section than pieces from
more posterior levels. These differences, however, can be practically elimi-
nated by decreasing length of pieces, controlling illumination, and provid-
ing conditions favorable to aggregation of the pieces. In naked-bodied
aquatic forms which can withstand only brief exposure to air, weighing of
pieces presents certain difficulties. Some forms secrete considerable slime
when exposed to air, and transfer connected with weighing may also in-
crease slime production. If perisarc or other skeletal substances are pres-
ent, they usually differ in amount at different body-levels and become a
factor in weight.

Elongated organisms of considerable size, at least several millimeters
in length, and without great regional differences are the most favorable

6o PATTERNS AND PROBLEMS OF DEVELOPMENT

material for determinations of difference in respiratory rate along the
polar axis, but difficulties and more or less uncertainty are involved in all
such determinations. Such organisms are either fully developed or in an
advanced stage, and data from them can be no more than suggestive or
indicative of conditions in earlier developmental stages. They give no in-
formation concerning axial characteristics at the beginning of develop-
ment or of changes which may have occurred during its course. More-
over, some of them, such as planarians and annelids, have certain organs
at particular body-levels which may have respiratory rates different from
those of other parts of the body at the same level; also, differences in rate
along the gut may not be in the same direction as differences in the body
wall. In planarians differential susceptibility and rate of dye reduction
in the body wall decrease from the head posteriorly to the fission zone in
forms with a posterior zooid, but there is some evidence that the gradient
in the gut decreases in both directions from the pharyngeal region. If
respiratory rates parallel these differences, the respiratory rates of pieces
from different levels of the planarian body may not show the differences
which are actually present, because differences in the gut may more or less
balance differences in the body wall. In practice we may apparently de-
crease differences in the gut by starving the animals for a week or more,
but even with this precaution we cannot be certain that we obtain a true
picture of the respiratory differentials in the body; nevertheless, the fact
that differences in rate are usually absent in pieces of animals with food
in the gut and become definite and consistent in sufficiently starved ani-
mals appears in some degree significant (pp. io8, 737). The unbranched
stem (hydrocaulus) of the hydroid Tubularia, or Corymorpha, meets the
desired conditions, as far as differentiation of regions is concerned, as
nearly as any animal species ; but in the case of Tubularia the presence and
different thickness of the perisarc at different levels and in Corymorpha the
increase in volume of the entodermal parenchyma basipetally and the
presence of perisarc on the basal part of the stem show that even these
forms are not without possible complicating factors for respiratory deter-
minations (Child and Hyman, 1926; Hyman, 19260).

Determinations or estimations of CO2 production also give some infor-
mation concerning regional differences in respiratory rate, but essentially
the same difficulties are involved as with determinations of oxygen con-
sumption.

The chief methods available for determinations of oxygen consumption
bearing upon the question of a respiratory pattern in the individual are the

METHODS OF PHYSIOLOGICAL ANALYSIS 6i

Winkler method and the respirometer and microrespirometer methods.
The Winkler method, as used for aquatic respiration, consists in deter-
mination of the amount of oxygen dissolved in water, before and after a
period of respiration, by the material concerned in a closed container with-
out air. The procedure consists in the addition of the Winkler reagent to
a known volume of water and titration. With the respirometer methods
the quantity of oxygen consumed is determined from decrease in gas vol-
ume measured in the manometer connected with the respiratory changer,
the carbon dioxide produced being removed. When the material is
aquatic, it is usually placed in a small volume of water in the respiratory
chamber, and provision is made for maintenance of equilibrium of oxygen
in air and water in the chamber, usually by some type of shaker. Constant
temperature is maintained.

The Winkler method, as used by Hyman and others with various in-
vertebrates,' has been criticized on the ground that absorption of iodine
by substances in the water, particularly organic substances, such as slime
or tissue fluids, may result in incorrect oxygen values.^

Carbon dioxide production can be determined or estimated in various
ways. The first data on differences along the polar axis were obtained
with planarian material by Tashiro with the "biometer" (Child, 1913a),
but most of the later work along this line has been done with colorimetric
methods. A method of comparative estimation used with Dugesia
dorotocephala (Robbins and Child, 1920) and with Corymorpha (Child and
Hyman, 1926) is based on color change of a nontoxic acid-alkali indicator
(phenolsulphonephthalein) in solution in the water containing the pieces,
equal weights of animal material and equal volumes of solution being used
for comparison. The rate of change in hydrogen-ion concentration, as in-
dicated by change in color of the indicator solution or the time required to
attain a certain color as compared with standard colors of known pH,
serves for comparison of different lots of material. In another colorimetric
method^ the material, in a small volume of water, if aquatic and unable to
stand exposure to saturated air, gives off CO^, into the water; it passes
from the water into the air of the respiratory chamber and from the air
into the indicator solution. Equilibrium between air and indicator solu-
tion is maintained by continued to-and-fro movement less rapid than the

' Hyman, 19166, 1919a, b, c, d, e, 1920a, c, 19236, 1925, 1926a, 1932a; Hyman and Galigher,
1921; Child and Hyman, 1926; Lund, 1928.

^ See Appendix I, pp. 729-31.

3 Parker, 1925, 1929; Watanabe, 1931; Watanabe and Child, 1933.

62 PATTERNS AND PROBLEMS OF DEVELOPMENT

usual shaking.'* Neither of these two methods provides for the rather re-
mote possibihty of production of other gaseous or volatile acid-forming
substances or for production of ammonia.

As regards the problem of developmental pattern, it is much more im-
portant to know whether a respiratory pattern is present in eggs and early
developmental stages, and what changes it may undergo during develop-
ment, than to determine whether such pattern is present in adult indi-
viduals; but the difficulties are even greater. Although the data thus far
obtained with adult forms give no information concerning earlier stages,
they are of some interest to developmental physiology because, in those
cases in which evidence of a respiratory gradient has been obtained, this
gradient parallels gradients indicated by other methods, and in some forms
these other gradient expressions persist from early development through-
out life, at least in certain parts of the body. Gradient patterns indicated
by differential susceptibility, differential dye reduction, and in other ways
are characteristic of early developmental stages; and their coincidence
with respiratory gradients in fully developed individuals suggests the
possibility that a respiratory gradient pattern may also be present in
earlier stages of development. Granting the possibility of sufficiently ac-
curate determination, however, separation of the unfertilized or fertilized
egg into pieces leaves the nucleus in one piece if it is not destroyed or in-
jured, although in many eggs fertihzation of the nonnucleated piece is
possible. Separation of blastomeres and of parts of embryos and larval
stages is possible, but the same questions as to effects of separation arise
as with pieces of adult animals, and the possibihty of periodic change in
respiratory rate in connection with the cell-division cycle must also be
considered. As development progresses, localization of organs and appear-
ance of new axes may introduce further complication. Respiration of parts
of amphibian embryos isolated by section has been determined by J.
Brachet (1934a, b, c, 1936), and more recently determinations have been
made on different regions of intact unfertilized eggs and embryos by intro-
duction of a single egg or embryo into a capillary tube with diameter of
lumen equal to, or slightly smaller than, that of the egg and connection
of both ends of the tube with a microrespirometer (J. Brachet and Sha-
piro, 1937). With this method the organism is not injured in any way, and
oxygen uptake of two opposed sides can be determined separately. Re-
sults obtained in this way are discussed in the following chapter. Still

■• See Appendix I, p. 731.

METHODS OF PHYSIOLOGICAL ANALYSIS 63

more recently studies on respiration and glycolysis have been made on
small fragments of amphibian embryos with the Cartesian diver "ultra-
microrespirometer" (Boell et at., 1938, 1939).

In so far as they involve separation of the material into pieces, the
methods for determining regional differences in glycolysis present the
same difffculties and raise the same questions as those for oxygen consump-
tion and CO2 production.

DIFFERENTIAL REDUCTION OF POTASSIUM PERMANGANATE

Potassium permanganate is a powerful oxidizing agent and is readily
reduced by living protoplasms to oxides of manganese, which are precipi-
tated and stain the protoplasm brown or, in high concentration, opaque
black. As might be expected, this substance is highly toxic and may be
used as a lethal agent to show death gradients. In very low concentrations,
however, intracellular reduction and coloration of the protoplasm take
place slowly; and axial differentials in rate and, in some cases, in depth or
density of coloration can be observed. Concentrations used in this way
range from m/2,000 to m/ioo,ooo, according to material and rate of colora-
tion desired. The method is chiefly of value for small more or less trans-
parent forms, such as protozoa, some eggs, blastulae, planulae, and gas-
trulae, but may also be used to show differential reduction in the ecto-
derm of hydroids, medusae, and various larval forms and on cut surfaces
at different levels of an axis. Even small embryonic or larval stages of
some species become opaque black if the reaction is allowed to continue to
completion and final death of all parts. In some of these, however, the
organism becomes more or less translucent after dehydration and clearing,
and a gradient in depth of color is visible ; but in the clearing oil or in bal-
sam mounts the color gradually disappears. It is perhaps scarcely neces-
sary to point out that observation of the color through different thick-
nesses of tissue in the same organism may lead to incorrect conclusions.
In organisms killed by boiling water or by various fixing agents the amount
of reduction, as indicated by depth of color, and the rate of reduction are
greatly decreased; and either no gradient appears or a slight gradient is
present immediately after killing but soon disappears.

The gradient in rate of reduction might conceivably result from a gra-
dient in rate of penetration of KMn04, but evidence indicating appreciable
differential in penetration has not been obtained. The reduction color is
at first uniform, as reduction occurs superficially in the cell or cells, and

64 PATTERNS AND PROBLEMS OF DEVELOPMENT

the color gradient appears only gradually. If any permeability gradient
were originally present, it is probably destroyed at once.^

DIFFERENTIAL INTRACELLULAR INDOPHENOL REACTION

The indophenol blue reaction between a-naphthol and dimethyl-para-
phenylenediamine, which is catalyzed by certain oxidases, has been used
with various ciliate protozoa, embryonic and larval stages, and small ani-
mals, also in certain experiments on reconstitution. A very distinct axi-
al gradient in appearance within the cell or cells of the indophenol blue
color is characteristic of the living animal, but none or only traces for a
short time if the organisms are first killed. The reagents must be used
in extremely dilute solutions, to avoid killing. The extensive litera-
ture on the indophenol reaction, the indophenol oxidase or oxidases, and
their relation to tissue oxidation and respiration need not concern us here.
It has been shown that dimethyl-paraphenylenediamine increases oxygen
consumption in certain tissues and that a-naphthol inhibits oxidations.
Moreover, not all the oxidation products of paraphenylencdiamine give
indophenol blue with a-naphthol, and it may happen that the more vigor-
ous the oxidation, the less blue is formed. Also, indophenol blue may be
decolorized by the reducing action of certain tissues. The greater part of
the work with this reaction has been done with vertebrate tissues, and it
is by no means certain that the conclusions and hypotheses based on it
hold without modification for the lower invertebrates, embryonic tissues,
and unicellular organisms.

Attempts have been made with hydroid planulae and starfish blastulae
and gastrulae to determine whether or to what extent differential perme-
ability to the agents might be concerned. The material was placed first
in one of the agents for one to several minutes, then in the other. It was
found that with this procedure both agents, in the very low concentrations
used, were present in all parts of the material within a short time in suffi-
cient quantity to give the same reaction as with much longer exposure.
Five minutes or less was sufficient for complete penetration of either re-
agent, but difference in time of the reaction along a polar axis may be an
hour or more. That the indophenol gradient indicates an axial physiologi-
cal differential of some sort is evident; that it indicates a differential as-
sociated with indophenol oxidase appears highly probable; and that it

5 For use of KMn04 in this way and results obtained see Child, 1919a,/, 1921a, 1925a,
1926a; Hyman, ig2ob; Galighcr, 1921a.

METHODS OF PHYSIOLOGICAL ANALYSIS 65

represents one aspect of an axial differential which expresses itself in
various ways is indicated by the fact that it parallels gradients shown by
other methods in the same material.

Benzidin has been used as a test for peroxidase activity in sea-urchin
eggs and embryos;'' but since the method involves killing the material, the
significance for the living organism of observed differences or absence of
difference remains a question.

DIFFERENTIAL STAINING BY VITAL DYES

The so-called "vital dyes" which stain intracellular constituents in the
living organism are more or less toxic and, with sufficient staining, kill.
Certain of these dyes undergo reduction in living cells under anaerobic
conditions or in low oxygen, becoming colorless (methylene blue, etc.) or
changing color in the course of reduction (Janus green). Use of dye-re-
duction methods is considered in the following section. For the present,
staining with oxidized basic dyes is the chief concern. In water of suffi-
ciently high oxygen content to prevent intracellular reduction of the dye,
organisms or parts with little regional differentiation along the polar axis,
coelenterate blastulae and planulae, echinoderm blastulae and gastrulae,
the body of hydra, the naked stem of Corymorpha, hydroid tentacles, etc.,
stain uniformly or almost uniformly at first; that is, there is little or no
evidence of differential penetration of dye in the ectoderm of such forms.
With continued staining, however, an axial differential often appears. The
depth of staining becomes greatest apically or anteriorly, decreasing basip-
etally or posteriorly, and cytolysis and death produced by the dye follow
the same course. A study of staining gradients in relation to susceptibility
gradients and to the gradient problem in general has been made by J. W.
MacArthur (1921) with protozoa, hydra, planarians, other fiatworms, and
annelids as material. Like others before him, he finds the basic dyes much
more effective, both as dyes and as toxic agents, than the acid dyes. Acid
dyes in neutral solution do not stain and are rarely sufficiently toxic to
kill. Increased acidity increases effectiveness of acid dyes. In a planarian
which shows certain minor differences in relative susceptibility to basic
and acid agents (pp. 11 2-13), MacArthur finds that basic dyes act like
acid agents, and acid dyes, when effective at all, like basic agents. He also
finds that the staining gradient shows the same axial relations as the later

' Ranzi e Falkenheim, 1937; Pitotti, 1939.

66 PATTERNS AND PROBLEMS OF DEVELOPMENT

death gradient in the dye and in other agents: his results are confirmed
by many other observations on staining and toxic effects of various dyes.^
In his discussion of the Uterature of vital staining MacArthur calls at-
tention to the apparent relation between rate of staining and physiological
condition of different regions and concludes from the evidence of his own
and other work that the axial staining gradient cannot be due to a differ-
ential permeability in any merely physical sense. The cell surface is living
and concerned in the activity of the cell. The staining and toxic effect of
the dye depend on its adsorption, precipitation, flocculation, or chemical
combination on or with cell constituents. Its continued entrance and ac-
cumulation within the cell depend on its removal from the solution in the

Evidence from other work also indicates that graded differences in vital
staining, particularly in the simpler animals and earlier developmental
stages, are associated not merely with differences in permeability but with
graded differences in physiological condition of the cells. Numerous ob-
servations on vital staining of protozoa, blastulae, planulae, young hy-
droids, stages of reconstitution of Corymorpha and other hydroids, turbel-
laria, and microdrilous oligochetes agree with MacArthur's data in show-
ing that with sufficient staining axial staining gradients usually appear
which parallel closely the death gradients observed with other physical
and chemical agents. The gradients of injury, cytolysis, and death, with
further continued staining, coincide with the staining gradient.

A point of some interest in connection with this differential staining is
that, when staining has progressed to a certain stage, the gradient of sus-
ceptibility to cyanide disappears, all levels being equally susceptible.
With further staining the susceptibility gradient undergoes reversal, that
is, the regions which were originally most susceptible to cyanide become
least susceptible after injury by the dye. This reversal of differential sus-
ceptibility to cyanide along an axis by deep staining has been observed

" Protozoa, Child and Deviney, 1926; Hydra, Child and Hyman, 1919; embryonic, larval,
and adult hydroids. Child, i()i()d, 1926a; larval stages of the polyclad, Stylochus, Watanabe
and Child, 1933, and other turbellaria.

* Hypotheses concerning the manner in which vital staining occurs are not lacking. Since
MacArthur's work is under discussion, it may be noted that he suggests an electrochemical
hypothesis. The cation of the dye, which is the color ion, unites with an intracellular anion,
which is a product of catabolism; if such anions are formed by dissociation of amphoteric pro-
teins as acids, their production should be decreased by acidity and increased by alkalinity and
by abundant o.xygen, as seems to be true. According to McCutcheon andLucke (1924), how-
ever, e.xternal alkalinity favors, internal alkalinity decreases, staining with basic dyes. They
suggest that these dyes combine with some acid cell constituent.

METHODS OF PHYSIOLOGICAL ANALYSIS 67

with methylene blue and some other basic dyes in several algae, hydroid
planulae and later stages, and other invertebrates (Child).

Methylene blue is known to increase respiration in tissues of higher ani-
mals and to protect against the action of cyanide. '^ Whether the reversal
of the susceptibility gradient to cyanide by dyes is due to increased respi-
ration of the more deeply stained regions, which antagonizes the action of
cyanide, or to decreased respiration resulting from injury by the dye, and
consequently decreased susceptibility to cyanide, has not been deter-
mined.

Recently differences in vital staining of regions and blastomeres in cer-
tain eggs and embryos have been interpreted as evidences of differentia-
tion. In the absence of information concerning technique, concentrations
used, periods of staining, and toxic effects of the dyes used, it is perhaps
questionable whether some of the differences described— for example, in
cleavage stages — are not due to other factors than differentiation." Deep
staining of one cell earlier than others about it may in some cases indicate
only greater susceptibility and earlier injury by the dye, possibly in con-
nection with certain stages of the division cycle. Some basic dyes are
much more toxic than others; but, in general, intracellular accumulation
of dye beyond a certain point results in injury to the protoplasm, and with
further staining cytolysis and death occur. The graded differences in
these effects along physiological axes of many organisms suggest quanti-
tative, rather than specific, regional differences, but they do not exclude
the possibility that specific differences are present without effect on dye
action or that the staining characteristics of certain cells or organs may be
associated with their differentiation. It appears beyond question, how-
ever, that use of a wide range of dye concentrations and staining periods is
necessary as a basis for conclusions concerning the significance of differ-
ences in staining.

DIFFERENTIAL REDUCTION AND OXIDATION
OF VITAL DYES

Intracellular reduction with change or loss of color of certain vital dyes
occurs under anaerobic conditions or when oxygen tension of the medium
falls below a certain point. The oxygen-level at which reduction of a par-

»See, e.g., Eddy, 1931; Gerard, 1932, pp. 504-7; M. M. Brooks, 1932, 1935; Chen, Rose,
and Clowes, 1933; Marsh, 1934; Solandt ei al., 1934; Chrisler, 1935; Bodine and Boell, 1937,
and citations by these authors.

" Ries, 1936, 1937; Ries and Scholzel, 1934; Ries and Gersch, 1936; Gersch and Ries, 1937.

68 PATTERNS AND PROBLEMS OF DEVELOPMENT

ticular dye occurs differs in different organisms, and regional differences in
time and rate of reduction in the individual are associated with differences
in rate of oxidation. With gradual decrease of oxygen after staining, such
differences are most clearly seen. In embryonic stages and even in adult
individuals of many animal species these differences appear not as sharply
defined and bounded areas but as reduction gradients, that is, gradients of
change or loss of color in definite and characteristic relation to physiologi-
cal axes of the whole, of organ systems, and of loci of developmental activ-
ity, such as buds. If injury by the dye has not occurred, return of color
after reduction may be brought about by readmission of oxygen. With
small organisms in a small amount of water reoxidation may occur almost
instantaneously or within a few seconds, but a gradient of return of color
is usually temporarily visible. If not injured by lack of oxygen or dye, the
same material may serve for repeated reduction and reoxidation.

A number of the basic dyes which are available as oxidation-reduction
indicators can be prepared as reduced, colorless leucobases with either a
reducing agent, sodium hyposulphite, or its formaldehyde compound, ron-
galite, or by catalytic reduction with molecular hydrogen by a platinum or
palladium catalyzer." The leucobases penetrate more readily than the
oxidized dyes, and the color appears in the material when it is exposed to
oxygen. Leucobases prepared with hyposulphite, however, are much
more toxic than the oxidized dyes, even those which stain readily in oxi-
dized form; consequently, great care is necessary in their use. In regions
of the living material injured by them reduction is retarded or does not
occur at all. Leucobases catalytically prepared are less toxic, but oxidize
rapidly on exposure to oxygen and enter living material without oxidation
only under anaerobic conditions.

With many organisms it is possible to distinguish definite and constant
regional differences in rate of appearance of color after penetration of the
leucobase and exposure to oxygen or with hyposulphite-leucobase, as it
enters the cells in an oxygen-containing medium. In other words, the leu-
cobase is more rapidly oxidized in some regions than in others; an oxida-
tion gradient may appear temporarily. Regional differences in rate of in-
tracellular oxidation of leucobase have been regarded as indicating differ-
ence in oxidase activity; but since the leucobases oxidize rapidly on expo-
sure to oxygen without a catalyst, it may be questioned whether this reac-
tion is always an adequate indicator of oxidase activity alone. But wheth-
er the dye enters the cell or organism as leucobase or in its usual form as

" See Fischer und Hartwig, 1937, and their citations.

METHODS OF PHYSIOLOGICAL ANALYSIS 69

oxidized dye, the most readily observed and most conclusive evidence of
oxidation-reduction pattern is obtained by intracellular reduction of the
oxidized dye. This can be made to occur rapidly or slowly as desired. The
material can be brought at once, after staining, into anaerobic conditions,
where reduction will take place rapidly or decrease in oxygen may be
brought about gradually by the oxygen consumption of the material itself
in a sealed container of small volume. It is necessary to make certain that
injury from lack of oxygen or accumulation of carbon dioxide or other
products of metabolism does not occur before reduction.

Suitable concentrations and periods of exposure to the dye must be de-
termined by trial for the organism concerned, with both oxidized dye and
leucobase. Too high concentration or too long exposure results in differen-
tial injury and retardation or absence of reduction of dye in the injured
region; consequently, the reduction pattern becomes different from that of
the uninjured individual. An axial reduction gradient may even be the re-
verse of the normal, because the region which normally reduces most rap-
idly is also most susceptible to injury by the dye and, when injured to a
certain degree, though still living, reduces least rapidly. Examples are
given in the following chapter.

If a true picture of oxidation-reduction pattern is to be obtained, it is
also highly important, with gradual oxygen decrease by the living mate-
rial, that equal oxygen distribution in the medium be maintained; other-
wise differences in reduction which have no relation to physiological pat-
tern may occur. For example, when two or more eggs or embryos lie close
together without change of position after staining, reduction usually oc-
curs first in the adjoining regions of each, because oxygen decreases more
rapidly between them than elsewhere. A region of an embryo remaining
continuously in contact with the bottom of the container is likely to re-
duce before others. Free-swimming organisms in a sealed container pro-
vide more or less completely for equal oxygen distribution; but if they
tend to aggregate, unequal distribution usually results. An aggregation of
Paramecium reduces much more rapidly after staining with methylene
blue than isolated individuals swimming free in the same preparation, and
at a certain oxygen-level an individual which has become colorless in the
aggregation becomes blue at once on leaving it. With nonmotile forms —
cleavage stages, for example — in a small sealed container equal distribu-
tion of oxygen may be maintained by frequent change in position of the
preparation with respect to gravity.

It is not to be assumed that a dye-reduction pattern — for example, an

70 PATTERNS AND PROBLEMS OF DEVELOPMENT

axial dye-reduction gradient — must always, and of necessity, coincide
with the respiratory gradient of the same axis or that a respiratory gradi-
ent must always be distinguishable in an axis which shows a dye-reduction
gradient. Dye reduction is regarded as chiefly dependent on intermediate
oxidation-reduction reactions rather than on total oxygen use. On the
other hand, when respiratory gradient and dye-reduction gradient do co-
incide, each aids in confirming the significance of the other."

DIFFERENTIAL SUSCEPTIBILITY

Living protoplasms may be stimulated or depressed, accelerated or re-
tarded, in activity or specifically altered in condition by external physical
and chemical agents; and any of these effects, beyond a certain degree or
threshold, becomes injurious or lethal. Agents acting within the organism
may also affect a particular cell or cell group, or an organ or organ system,
in all these ways. These effects are indicative of the sensitiveness or sus-
ceptibility of Hving protoplasm to factors in its environment. If life is a
continuous dynamic equilibration, protoplasmic susceptibility signifies
that this equilibration process can be altered in various ways or arrested
by external factors. We do not doubt that, as far as the organism is con-
cerned, differences in susceptibiHty of regions and organs, whether specific
for certain agents or not, depend upon differences in physicochemical con-
stitution and physiological condition or activity of the protoplasms, re-
gions, or organs concerned. It is evident, however, that differences in sus-
ceptibiHty to a particular concentration or intensity of a particular agent
tell us little or nothing more than that differences of some sort in constitu-
tion or condition exist. In order to obtain all the information possible con-
cerning susceptibility to any agent, it is usually necessary to determine
the susceptibilities of individuals, in different physiological condition and
of different developmental stages, to a wide range of concentrations or in-
tensities of the agent and often of other agents as well, to compare the
susceptibilities of different species, and to check the data on susceptibility
with those obtained by other methods. Conclusions concerning the nature
of regional differences in protoplasmic constitution and condition have
often been drawn from results obtained with a single agent in a single con-
centration or intensity. Such conclusions are often highly uncertain.
Many external agents act more or less specifically on certain organs or or-
gan systems of adult individuals, particularly of forms in which a high de-
gree of histological differentiation is evident. A certain agent may act
" For further details concerning use of the method and for references, see Appendix II.

METHODS OF PHYSIOLOGICAL ANALYSIS 71

chiefly or perhaps entirely, so far as we can determine, on a certain organ;
another on another organ. In case of some agents, action is more or less
specitic in a certain range of concentration or dosage ; but above this range
the action becomes more general, and in sufficiently high concentration or
intensity such agents may affect the whole organism. For pharmacologi-
cal purposes agents which produce physiological effects have sometimes
been separated into those with specific and those with general action, but
this distinction cannot usually be sharply drawn.

Another aspect of protoplasmic susceptibility appears in the relation
between the susceptibility of a particular species-protoplasm, individual,
or organ to a particular agent or to many agents and its physiological con-
dition. For example, the physiologically young individual is usually more
susceptible than the old to various agents within a certain range of con-
centration or intensity. Increase in motor activity increases the suscepti-
bility of the planarian to cyanide in a certain range of concentration
(Child, 1913a), but increased motor activity may decrease the suscepti-
bility of many animals to alcohol and certain other agents within certain
limits of concentration. Susceptibility of an organ system may change
greatly during development. Susceptibility of the nervous system to anes-
thetics or to strychnine changes relatively to that of other organs from the
medullary plate stage to the adult in vertebrates; but susceptibility of the
hydroid nerve net to these agents undergoes no such relative change, and
strychnine, like anesthetics, depresses nervous activity in the adult.

From the viewpoint of developmental physiology the questions of chief
interest as regards susceptibility concern axial and regional differences in
susceptibility, susceptibihty patterns at the beginning and in early stages
of development, and the changes which occur as development progresses.
So far as data are available, axial or other regional differences in suscepti-
bility which are specific for particular agents are usually less evident in
earlier, than in later, developmental stages or absent, though regions of
high yolk content in eggs may be highly susceptible to fat-soluble agents,
and certain other differences which have been regarded as specific appear
in some forms. Graded differences in susceptibility (susceptibility gradi-
ents) appear as characteristic features of early developmental pattern.
As development progresses, these gradients may undergo alteration or
even obliteration, and new gradients may appear; but in many of the sim-
pler animals gradients of earlier stages persist throughout life. These
gradients constitute what has been called "differential susceptibility."
They are, in general, nonspecific to a high degree; that is, the gradient is

72 PATTERNS AND PROBLEMS OF DEVELOPMENT

the same in direction with respect to developmental pattern, physiological
axes, or radii for certain ranges of concentration or intensity of many dif-
ferent agents which act on living protoplasms in different ways.

Differential susceptibility to external agents appears in various aspects:
differential death, differential inhibition, differential tolerance, differen-
tial acclimation, differential recovery, differential acceleration, and, in
some cases in later stages, differential irritability and other functional dif-
ferentials.

Using death or the cellular changes — cytolysis, coagulation, disintegra-
tion, etc. --associated with, or indicating, the approach of death, as a cri-
terion, we find that death gradients are characteristic features of develop-
mental pattern. For example, when the whole organism is equally ex-
posed to action of an external agent in concentration or intensity above
the limit or tolerance, but only slowly lethal, death occurs first in a certain
region of a physiological axis and progresses along the axis from that re-
gion. In the earlier stages of buds and of organs which develop as budlike
outgrowths from localized fields death progresses radially, though not al-
ways equally in all directions, from a region of highest susceptibility,
which may be central or otherwise situated in the field. As such an axis
develops, the radial susceptibility gradient becomes a longitudinal gradi-
ent. The differences in time of death at different levels of an axis may dif-
fer widely with different agents and with concentration or intensity of a
single agent; but with certain ranges of concentration or intensity of many
different agents, both chemical and physical, the progress of death is in
the same direction.

Differential inhibition of embryonic or reconstitutional development or
of other activities may occur with concentrations or intensities of many
agents which are more or less toxic or inhibitory but not directly lethal.
In consequence of differential inhibition developmental form and propor-
tion can be modified in definite and controllable directions.

With a certain range of lower concentration or intensity, which, of
course, must be determined experimentally for each species and each
agent, and usually for different developmental stages and for physiologi-
cally young and old individuals, evidences of differential tolerance and of
ability to acquire increased tolerance have been found in cases examined.
In earlier publications this differential in degree or limit of tolerance or
threshold of toxic or lethal action of the agent and the differential acquire-
ment of increased tolerance have been grouped together and called "dif-
ferential acclimation." Now, however, it seems desirable to distinguish, at

METHODS OF PHYSIOLOGICAL ANALYSIS 73

least formally, two aspects of this differential: first, the differential in
the threshold of irreversible toxic or lethal effect of the agent or the upper
limit of tolerance to it; second, a differential in the rise of this threshold,
that is, in acquirement of increased tolerance during continuous exposure
to the agent. This is differential acclimation or conditioning in the strict
sense. Apparently, there is in such cases a differential in the ability of the
protoplasm to dispose of the agent in some way or to render it nontoxic or
less toxic. The tolerance differential may be so great that a part of the in-
dividual finally dies after days or weeks of exposure to the agent, while the
other part continues to live indefinitely; or, with effect of agent near the
upper limit of tolerance, the whole individual may finally die, but the
progress of death along the axis is in the reverse direction from that with
action of the same agent above the limit of tolerance. Dugesia {^Eupla-
naria) serves as an excellent example. In concentrations of ethyl alcohol
which kill within a few hours, physiologically young animals are more
susceptible than old, and the head and posterior zooid region are most sus-
ceptible, death progressing posteriorly from the head in the anterior
zooid. In a certain range of lower concentration young animals are more
tolerant than old; and, even though both young and old are at first anes-
thetized more or less completely, the young recover activity before the
old with continuous exposure to the alcohol, apparently because of ac-
quirement of increased tolerance. Also, if death occurs anywhere, it be-
gins at the posterior end of the anterior zooid and progresses anteriorly; it
may stop at a certain level, and anterior region and posterior zooid region
may continue to live indefinitely in the same concentration of alcohol ; the
anterior part may even develop a posterior end, and the posterior zooid a
head, in the same concentration which was earlier partially lethal. Differ-
ential tolerance and apparently also some degree of differential acclima-
tion or conditioning appear in early developmental stages with some
agents in secondary differential modifications of form and proportion in
directions opposite to those of the primary differential inhibition. In
echinoderm development regions most inhibited or killed by more extreme
toxic action show relatively more rapid development and larger size than
others (chap. vi). These modifications occur even though all parts are
more or less inhibited; and they follow a differential inhibition in earlier
stages, indicating some degree of differential conditioning. Since the re-
gions which are primarily most tolerant show the greatest capacity to ac-
quire increased tolerance, the modifications of form and proportion result-
ing from differential conditioning are, in general, similar to those of differ-

74 PATTERNS AND PROBLEMS OF DEVELOPMENT

ential tolerance. It is often difficult to determine whether the modifica-
tions observed are due merely to the primary differences in tolerance or to
actual differential conditioning. Thus far. death gradients associated with
differential tolerance and differential conditioning have been investigated
in only a few forms and with a few agents, but difference in ability to toler-
ate and to acquire tolerance to different agents is evident. The differential
in tolerance to ethyl alcohol along the polar axis of forms tested is con-
siderable, and differential conditioning occurs relatively rapidly; but dif-
ferential tolerance to cyanide is shght, and differential conditioning occurs
ver>^ slowly if at all (Child, 1932&). It is possible that the mechanism of
tolerance to certain agents may gradually fail with time and that a pro-
gressive differential sensitization to certain agents may conceivably occur,
but evidence is lacking on these points.

Living organisms are able to recover more or less completely from less
extreme toxic action of external agents after return to the normal medium.
Axial differentials in recovery after slight differential inhibition are, in gen-
eral, parallel to those of differential tolerance and differential acclimation
but are often more strongly marked, since the inhibiting agent is no longer
present in the recovery period. With more extreme action of the agent the
more susceptible regions may be injured to such an extent that they can-
not recover, even after return to normal environment, and only less sus-
ceptible regions recover. This is not, properly speaking, a differential re-
covery but rather a partial recovery, that is, a recover}- of the least sus-
ceptible and least inhibited part; usually it is by no means complete and
may be Httle more than a continuation of life after return to the natural
medium. The parts that do not recover usually die, though in some cases
they may remain alive in highly inhibited condition.

Differential tolerance, conditioning, and recover}- are all secondary
modifications, as far as distinguishable alterations of pattern are con-
cerned. They depend on the physiological condition and metabolism of
the part concerned, rather than directly on the action of the external
agent. As regards both differential death and modification of form and
proportion in development, their effects are the reverse of the primary in-
hibitions in their relations to physiological pattern.

As regards the physiological significance of differential susceptibility,
the fact that the lethal gradient is the same in direction for many different
agents, both chemical and physical, indicates that it does not depend on
the constitution or nature of the agent. It is certain that lethal effects of
different agents are not all brought about in the same way. The only con-

METHODS OF PHYSIOLOGICAL ANALYSIS 75

elusion possible seems to be that the susceptibility gradient represents a
quantitative differential or gradient of some sort. In other words, differen-
tial susceptibility indicates quantitative features of physiological axes but
gives no information concerning qualitative differences of substance which
may or may not be present at different levels. Differential permeability
may be a factor in determining differential susceptibility to some agents,
but it is not the only or the chief factor, for the same gradients appear with
agents that penetrate readily at all levels, with agents that penetrate only
as they injure cell surfaces, and with physical agents, such as ultra-violet,
X-rays and radium, temperature extremes, and lack of oxygen. Moreover,
in all cases involving the same material, susceptibility gradients parallel
very closely the respiratory gradients, gradients of dye reduction, and in-
dophenol gradients. Differential tolerance and differential conditioning
must involve a differential activity of the organism in relation to many
agents. The anterior region of Dugesia is more tolerant and acclimates
more rapidly or more completely to low concentrations of alcohol, and to
some extent to low concentrations of cyanide, but is more susceptible to
lethal concentrations than more posterior regions (Child, 191 le, 1913^,
1914&, 19326). Also, physiologically young planarians are more tolerant
to low concentrations, and more susceptible to high concentrations, than
old. It has been suggested in earlier papers that a general parallelism ex-
ists between differential susceptibility and metabolic, respiratory, or oxi-
dative rates at different levels of the individual organism. That this paral-
lelism is absolutely complete or universal has not been maintained, nor
need it be assumed that all the agents used in demonstrating differential
susceptibility act directly on oxidation, reduction, or other metabolic reac-
tions. However, it does appear highly probable that any sort of disturb-
ance of the protoplasmic system by an external agent, if sufficient to bring
about death or inhibition as direct effect, will be likely to kill or inhibit
earlier in regions in which change is going on more rapidly than in those of
less activity and that regions of more intense metabolism will have, in
general, a higher tolerance and a greater ability to acclimate to, or to re-
cover from, effects which are not too extreme. If a living protoplasm is a
system undergoing continuous dynamic equilibration, any action of an
external agent, if sufficient in amount or intensity, must sooner or later,
directly or indirectly, alter some essential factor or factors of the system so
that alteration of the whole system results, or its continued existence be-
comes impossible. The higher the rate of equilibration in a particular re-
gion of a physiological axis, the earlier will such alteration or destruction

76 PATTERNS AND PROBLEMS OF DEVELOPMENT

occur, as in differential death and differential inhibition. On the other
hand, if the external action is slight in amount or intensity, the more rap-
idly equilibrating region will eliminate in some way the disturbing factor
or equilibrate to it more rapidly than a region which is equilibrating more
slowly. This is apparently what happens in differential tolerance, differ-
ential conditioning, and differential recovery. For susceptibilities which
are specific for particular agents, these relations do not hold; but they may
hold for different regions of a part or organ which, as a whole, is specifically
susceptible. Moreover, it is not necessary to assume that respiration or
oxidation is always the primary factor in the various expressions of differ-
ential susceptibility, but it appears impossible to account for the facts in-
dependently of metabolic activity; and in most organisms respiration or
oxidation-reduction appears to be a more or less trustworthy indicator of
such activity, or at least of some of its fundamental factors. The relation
between respiration or oxidation and susceptibility may not be the same in
all organisms or in all organs of adult individuals; when glycolysis is the
chief source of energy, it is obviously different from that in aerobic respira-
tion. All that is maintained here is that there appears to be a general
parallelism between the phenomena of differential susceptibility, particu-
larly in early developmental stages and in the simpler organisms, and the
basal metabolic activity of the protoplasmic system concerned. Differen-
tial susceptibility is at best merely one method of indicating differences in
physiological condition which appear to be quantitative and which, as will
appear in following chapters, are essential factors in development ; without
the aid of other methods it gives no direct information as to the nature of
the differences indicated.'^

THE QUESTION OF DIFFERENTIAL PERMEABILITY

It was pointed out above that gradients indicated by external chemical
agents are not dependent merely on a permeability gradient. Neverthe-
less, the question whether a differential permeability is present along phys-
iological axes as one feature of the axial gradient is of interest. It might
perhaps be expected that rate of staining by vital dyes would throw some
fight on this question. MacArthur (1921) maintained that differential
permeability alone could not account for the gradients of staining and
susceptibility to the dyes. The writer's observations bear out his conclu-

'3 For references to the literature on differential susceptibility and further discussion of the
question of its relation to metabolism see Appendix III.

METHODS OF PHYSIOLOGICAL ANALYSIS 77

sions. With low concentrations of methylene blue the entoplasm of Para-
mecium stains before the ectoplasm, and the posterior before the anterior
ectoplasm. With a certain higher range of concentrations ectoplasm and
entoplasm stain about equally. With still higher concentrations ecto-
plasm stains before entoplasm, and anterior ectoplasm finally becomes
more deeply stained than posterior (Child, 19346). Other protozoa show
similar relations, and in metazoa staining gradients in opposite directions
along an axis have been observed with low and high concentrations of dye.
Even when earlier stages of staining are quite uniform along an axis, as far
as can be determined, a very distinct gradient in depth of staining and in
toxic or lethal effect appears later. This susceptibility gradient evidently
results from difference in condition inside the cells and a consequent differ-
ence in adsorption or combination of the dye, rather than from any differ-
ence in permeability. Experiments with Paramecium on penetration of
ammonia and acetic acid and with hydroids, using neutral red as indicator,
have shown some evidences of an axial differential permeability to ex-
tremely low concentrations. '■• These concentrations, however, are only
shghtly or not appreciably toxic; and with the higher concentrations, in
which a lethal gradient appears, penetration seems to occur equally at all
levels. Strong bases and strong acids do not alter the color of the intracel-
lular neutral red until cytolysis begins, but the death gradient is the same
as with weak bases and acids. A very distinct susceptibiUty gradient ap-
pears in single elongated cells of monosiphonous algae, even though no
difference in penetration or depth of staining can be observed (Child,
1916c, e, 1917a, 1919/).

ELECTRIC POTENTIAL DIFFERENCES

For many years electric potential differences and their changes with
activity have been investigated in nerve and muscle, and numerous ob-
servations have been made on other organs of plants and animals. Most
of these investigations, however, have been on organs of fully developed
individuals, and the work on nerve and muscle has been concerned very
largely with vertebrate material. These investigations, important as they
are, are only remotely connected with problems of developmental pattern.
Potential difference in relation to developmental pattern and physiologi-
cal axes has received less attention, but data on axial potentials in certain

'I Child and Devincy, 1926; Child, ig2ba, and unpublished data.

78 PATTERNS AND PROBLEMS OF DEVELOPMENT

plants and animals and their changes under experimental conditions have
been obtained, though largely with growing plants and adult animals.'^

It is evident from the data that potential gradients are characteristic
features of physiological axes, but different authors are not in complete
agreement as regards direction of change in sign. Mathews; Morgan and
Dimon; Hyman; Hyman and Bellamy; and Watanabe found in various in-
vertebrates a decrease galvanometrically in negativity from the apical or
anterior region basipetally or posteriorly and in annelids a second gradient
of negativity decreasing from the posterior end anteriorly to a certain
level. According to Lund and his co-workers, galvanometric positivity de-
creases from the apical region basipetally and increases again toward the
basal region of the hydroid Obelia. Also, the tips of stem axes of the
Douglas fir are galvanometrically positive (internally negative) to lower
levels, and in the onion root there is a decrease in galvanometric positivity
from the tip basipetally to a certain level with some increase farther basal-
ly. Barth, working with hydroids, does not find any constancy in direction
of potential difference. In Tuhularia the hydranth is negative to the mid-
dle stem region, and a cut or reconstituting end is usually negative to a
hydranth. In Eudendrium reconstituting regions are usually positive to
other regions. In this connection it may be noted that hydranth reconsti-
tution in Tuhularia is a redifferentiation of a portion of the stem without
outgrowth, while in Eudendrium and Ohelia outgrowth of tissue from the
cut end occurs before the hydranth develops.

As regards hypotheses concerning the origin of bioelectric potentials,
Du Bois Reymond held that they originated in purely physical factors;
among the earlier physiologists, Hermann, Hering, Biedermann, and
Waller did not accept this view but maintained that they were associated
in some way with reactions of metabolism, though Hermann apparently
discarded this hypothesis later. In general, the hypothesis of a relation to
metabolism seems to have been regarded more favorably than others.

'5 Numerous determinations have been made on plant axes, usually showing the tip elec-
tronegative to lower levels, a few with the reverse potential difference. The following refer-
ences are more or less directly concerned with axial potentials in animals: Mathews, 1903,
hydroids; Hyde, 1904, eggs of fish and turtle; Morgan and Dimon, 1904, earthworm; Hyman,
1918, 19326; Hyman and Bellamy, 1922, sponges, hydroids, hydromedusae, planarian, anne-
lids; Child and Hyman, 1926, hydroid; Lund, 1921c, 1922, 1923a, 1924a, b, 1930, 1931a, h, c,
1932a, b; Lund and Kenyon, 1927; Lund and Bush, 1930; Lund and Hanszen, 1931; Lund and
Moorman, 1931; Marsh, 1928, 1930, 1932; Rosene, 1930; Rosene and Lund, 1934, hydroids,
plant stems, and roots; Watanabe, 1928, earthworm; Burr, 1932; Burr and Lane, 1935;
Burr and Hovland, igsja, b, developmental stages of amphibia and mouse; Barth, 1934^,
hydroids; Hasama, 1935, artificially fertilized amphibian egg.

METHODS OF PHYSIOLOGICAL ANALYSIS 79

Hyman (1918), Hyman and Bellamy (1922), and Child and Hyman (1926)
suggested that the axial potential difference observed in many animals
originates in metabolic differences. Lund and his co-workers, in various
papers from 1926 on, have maintained, as the result of determinations of
oxygen consumption and comparative estimations of CO2 production, de-
crease of potential difference and of oxygen consumption by cyanide, dif-
ferential reduction of methylene blue, and other experiments, that electric
polarity is quantitatively correlated with oxidation — in other words, that
the axial potential differences are expressions of an axial metabolic gradi-
ent. It was further suggested by Child (1924&, chap, xi) that, if the poten-
tial differences give rise to currents under natural conditions, such currents
may be important factors in physiological dominance and correlation. In
his later papers Lund assumes that currents do flow under natural, as well
as under experimental, conditions, and emphasizes their importance as
factors in correlation. Barth, on the other hand, finds "no coincidence of
electric polarity and organic polarity in different hydroids"; that is, in
some forms the apical region is galvanometrically positive, in others nega-
tive, to lower levels, and in isolated pieces cut ends may be at first nega-
tive, later positive, or continuously positive to other levels, according to
the species. In view of the fact that potential difference appears to be a
characteristic property of physiological axes, it seem rather improbable
that no definite relation between it and other axial characteristics exists.
Moreover, comparison of the apical regions and of the manner in which
hydranths are reconstituted in the hydroid species in which potential
differences have been determined suggests that apical galvonometric nega-
tivity may be associated with activity of differentiated or differentiating
hydranths or with the stimulation following section and apical positivity
with predominance of growth activity. Again, assuming that current re-
sults from the potential differences, the current may have different effects
on regions which it reaches, according to the sign of the apical region. A
current resulting from galvanometric negativity of the apical region or
cut end may conceivably result in activation and dedifferentiation of cells
which it reaches and organization of a hydranth primordium from a part
of the stem without outgrowth, as in Tuhularia (see Fig. 13) ; and current
from an electropositive end may establish a growth gradient, such as ap-
pears in Ohelia. According to Barth (1934a), inhibition of reconstitution
by an applied current may occur at either cathode or anode according to
species; and in Tuhularia, according to current density. It is to be ex-
pected that reconstitution will be inhibited, either by inhibition of the ac-

8o PATTERNS AND PROBLEMS OF DEVELOPMENT

tivation following section, which is apparently the primary factor in de-
termining the development which follows, or by inhibition of the resulting
growth or differentiation, or by both. The results as regards cathodal and
anodal inhibition in different species and with different current densities
may perhaps mean that, in the one case, one of these activities, in the
other case, the other, is inhibited. But aside from the question of inter-
pretation, an electric-potential gradient is certainly characteristic of the
polar axis of plants and animals, as far as investigated, parallels closely
other gradient expressions, and appears to be correlated with metabolic
activity; but many questions and problems await further investigation.

In many animals galvanotactic reactions show a definite relation to the
polar gradient. The reaction, however, cannot be regarded as a simple at-
traction of unlike poles resulting from an electric gradient in the organism
but is a physiological reaction to an external differential and can be al-
tered or reversed by change in environment or physiological condition."^

OTHER METHODS CONCERNED WITH PATTERN

Other biophysical, biochemical, and histochemical methods give evi-
dence of differential distribution or regional localization of substances in
relation to developmental pattern. As differentiation progresses, many
such localized differences of course appear as consequences of the earlier
pattern; but some quantitative differentials, in content of water, fat, car-
bohydrate, etc., may persist or appear in relation to axiate pattern in adult
individuals. The nitroprusside reaction indicates a polar glutathione gra-
dient in some forms; in others no differential has been found. A modifica-
tion of the nitroprusside reaction has been used by J. Brachet (1938) as an
indicator of differential distribution of SH-proteins in amphibian develop-
ment.

Evidence concerning early developmental stages bears more directly
on the problem of pattern. A differential localization of sulphydryl pro-
teins in very definite relation to developmental pattern in the amphibian
egg and embryo has recently been reported (J. Brachet, 1938). By stain-
ing and injecting eggs with acid-alkali indicators the appearance of a polar
differential in hydrogen-ion concentration at the time of polar-body for-
mation has been observed in eggs of certain annelids, mollusks, and tele-
osts, resulting finally in some eggs in sharply defined acid and alkaline
zones.'^ The question of distribution and disappearance of glycogen has

•^Hyman and Bellamy, 1922; A. R. Moore, 1923; Hyman, 1932/;.
'7 Spek, 1930, 1933, 19341', ^' ^-

METHODS OF PHYSIOLOGICAL ANALYSIS 8i

arisen in connection with amphibian gastrulation and induction (see pp.
154, 477). Presence of axial functional differentials has been demonstrated
or indicated in many organisms and organs in many ways. Functional
dominance of apical or anterior regions is a familiar fact. In planarians
and hydroids any body -level dominates, to some degree, more basal or
more posterior levels, provided more apical or anterior regions are absent.
The central nervous system in general, the ctenophore plate row and its
conducting path, the heart, the mammalian alimentary tract, and particu-
larly the small intestine, all present somewhat similar relations of domi-
nance.

Study of the physical condition of protoplasm may show regional dif-
ferences or chronological changes in viscosity, gelation, etc., which are re-
lated in some way to developmental pattern. Certainly the colloidal dif-
ferences between the cell surface and interior are fundamental factors in
the pattern of the cell as an organism. With the aid of polarized light and
X-rays much has been learned concerning the ultra-structure of many
morphological differentiations and products of living protoplasms, muscu-
lar and other fibers, both permanent and temporary, membranes, chromo-
somes, cellulose, hair, etc. The crystalHne or other orientations of mole-
cules or micellae discovered or inferred appear, in general, to be character-
istic of rather highly differentiated protoplasms or products or related to
local conditions, mechanical tensions or pressures, surfaces, or interfaces.
Up to the present no fundamental ultra-structure of a protoplasm that
might be regarded as the foundation of organismic and developmental pat-
tern has come to light. Many of these ultrastructural patterns are more
or less continuously forming and disappearing in protoplasms, and those
that are persistent and associated with morphological differentiation are
apparently derivatives rather than fundamental factors of pattern.

METHODS OF ANALYSIS OF DEVELOPMENTAL POTENTIALITIES
AND POTENCIES

As the words are used here and in following pages, "potentiality"
means possibility and "potency" means power or ability. Not all develop-
mental potentialities are realized in development under natural condi-
tions, for realization of all potentialities development in all possible en-
vironments is essential; it is impossible to determine what the potential-
ities are in any other way. Developmental potency, the ability to develop
in a certain way, represents the potentiality which is realized in a particu-
lar environment, intraorganismic or external. In the course of develop-

82 PATTERNS AND PROBLEMS OF DEVELOPMENT

ment potentialities may become progressively more limited, at least as
far as can be inferred from behavior of the material in certain environ-
ments; but we cannot be certain that in some other environment the po-
tentialities supposedly lost might not be realized. New potencies, how-
ever, may appear as development progresses: the ability to develop in a
certain manner is attained only when a certain stage is reached, that is,
when the developing system has attained a certain physiological condition
in consequence of changes occurring within itself, or in its intra- or extra-
organismic environment, or both. In case of a developing part of the intact
organism, both the part and its intraorganismic environment are appar-
ently factors in determining its developmental potency or potencies (cf.
Gilchrist, 1937a, ^).

The methods most widely used in attempts to analyze potentialities and
potencies are those that are not infrequently regarded as the methods of
developmental physiology — isolation, explantation and transplantation or
grafting of parts of organisms, union of individuals, and isolation and ag-
gregation of cells. The method of differentially modifying development by
exposure of the whole developing organism or system to experimental en-
vironments, which was mentioned above under differential susceptibility,
also serves for the realization of other developmental potentialities than
those realized under natural conditions.

ISOLATION OF PARTS

Isolation of parts in experiment has usually meant physical isolation-
The part is separated from other parts and remains in the normal external
medium or may be brought into an experimental environment. An isola-
tion experiment involves at least two, often more than two, parts. A hy-
droid or a planarian may be separated into two or many parts; the blas-
tomeres of a two-cell or later stage may be isolated; a small part, an organ
primordium, a limb bud, the optic primordium, etc., may be isolated from
other parts, but those parts are also isolated from it, and experiment may
be concerned with the effect on either or both.

Physiological isolation of parts without physical separation occurs in
many organisms in connection with agamic reproduction and with func-
tion and probably to a greater or less extent in development of the in-
dividual (see chap. ix). Results of physiological isolation are, in general,
essentially similar to those of physical isolation as regards further develop-
ment, though they may be less extreme.

METHODS OF PHYSIOLOGICAL ANALYSIS 83

EXPLANTATION

As the term is usually employed, explantation is merely isolation of a
part in a special extraorganismic environment. This environment may be
supposedly indifferent : a balanced salt solution such as the various con-
centrations and modifications of Ringer solution adjusted to the organisms
concerned. If properly adjusted, such a medium prevents the loss of elec-
trolytes which may occur in water, particularly if isolation involves a
wound. Cultures in vitro of living cells, cell masses, organ primordia or or-
gans in special media, plasmas, and gels (nutritive, growth-stimulating,
etc.) are explantations. These experiments involve not only the effect of
isolation but also that of the medium on the isolated part. Explantation
methods have been extensively used in the culture of particular cells or
tissues of embryonic or later stages, but recently culture in vitro of larger
parts of embryos, chiefly of vertebrates, has been undertaken in the at-
tempt to provide environments favorable to further development.

TRANSPLANTATION OR GRAFTING

Transplantation consists essentially in altering the organismic environ-
ment of the part concerned. This can be accomplished in many ways:
transplantation may be autoplastic, to the same individual, homoplastic,
to another individual of the same species, heteroplastic, to another species
of the same genus, xenoplastic, to an individual of another genus, family,
order, etc. In any of these relations it may be orthotopic, to the position
originally occupied by the part concerned, or heterotopic, to some other
position. The part may be transplanted in normal orientation to the axi-
ate pattern of the host or in some other orientation. Donor, transplant,
or host may be subjected to experimental conditions before or after trans-
plantation. Size of transplant and developmental stage of donor or host
can be varied. Transplantation to the extraembryonic membranes of
birds is possible; this environment is still physiological but more or less
isolated from the axiate pattern of the host. Up to the present the most
extensive transplantation experiments with embryonic stages have been
made with amphibian and avian material, but transplantations of parts of
fully differentiated hydroids, planarians, and annelids have given results
of great value for analysis of developmental pattern.

FUSION OF INDIVIDUALS

With some organisms it has been possible to unite developmental stages
of two intact individuals. Here the questions are whether they can recon-

84 PATTERNS AND PROBLEMS OF DEVELOPMENT

stitute to a single individual or retain their individuality and whether one
may dominate the other and alter its development. The question whether
fusion of individuals of different races or species is possible also arises.

DISSOCIATION AND AGGREGATION OF CELLS

Cells of certain sponges and hydroids can be dissociated and, in contact
with each other, may aggregate into masses, which may be varied in size
as desired. Under certain conditions these masses develop into complete
individuals. These cases raise questions of the origin of developmental
pattern and of the possibility of cell dedifferentiation and redifferentiation
in relation to the new pattern.

GENERAL PURPOSES OF THESE METHODS

These methods of experiment have been developed in the attempt to
obtain information concerning developmental potentialities and potencies.
By means of them we endeavor to discover whether, or to what extent,
realization of potentialities or developmental expression of potencies is in-
trinsic in the part concerned or is dependent either on a particular regional
relation to other parts of the organism or on general factors of intraorgan-
ismic environment, and whether or how it is affected by different nonor-
ganismic environments. Analysis of physiological dominance, both in
preventing physiological isolation and in determining the course of de-
velopment of subordinate parts (induction), has been greatly advanced by
these methods.

When isolated or transplanted parts undergo reconstitution, that is, al-
teration of pattern, we infer that they or some of their cells are not so
stably differentiated or their course of development so unalterably estab-
lished that they cannot react to the altered conditions, and that their de-
velopmental or other behavior as parts of the intact individual must be de-
pendent, at least in part, on factors in their intraorganismic environ-
ment. If they continue to develop or otherwise behave as in the intact in-
dividual, intrinsic factors independent of other parts are supposedly con-
cerned. As regards development, these differences are commonly distin-
guished, following Roux (1885), as dependent or correlative differentiation
and independent or self -differentiation. Actually, however, the dift'erence
involves not only differentiation but the whole pattern of developmental
and other behavior. A part may be independent in certain respects, de-
pendent in others. Moreover, as Roux pointed out, a development which
is independent as regards relation to other parts is dependent on relations

METHODS OF PHYSIOLOGICAL ANALYSIS 85

within its developing system. As far as relations to other parts are con-
cerned, a reconstituting hydroid or planarian piece is undergoing self-de-
velopment; but within the piece reconstitution depends on an orderly and
definite pattern of relations, on dominance of certain parts and subordina-
tion of others. Self-differentiation or self-development occurs only when a
pattern already present persists in the part after its isolation or transplan-
tation. Such a part is said to be "determined," that is, its pattern is re-
garded as fixed. Labile and definitive determination are often distin-
guished. It cannot be too strongly emphasized, however, that determina-
tion is always relative to a particular environment or environments. To
conclude that a pattern is unalterable because it is not altered by isolation
or by transplantation is entirely unjustified. There is always the possibil-
ity that in some other environment it may undergo alteration. In fact, this
occurs in some transplantation experiments; the pattern of a part trans-
planted to a certain region of a host may persist, but with transplantation
to another region it may undergo alteration. Undoubtedly, progressive
stabilization or fixation of pattern or the basis of differentiation does occur
in the course of development, even before pattern or differentiation be-
come visible. Determination, in the sense of a more or less stable disposi-
tion or tendency to develop in a certain way before that development be-
comes evident, obviously represents a change from an earher undeter-
mined condition. Isolation or transplantation may give evidence of it but
tells us nothing concerning its nature. To say that a part is determined
means only that it possesses a certain disposition as regards further activ-
ities and involves no implications concerning the nature of that disposi-
tion. If we keep in mind the possibility that even the most obstinate dis-
positions may not be unalterable and that the word "determination" is an
expression of our ignorance, it serves a useful purpose.

CHAPTER IV

PHYSIOLOGICAL CHARACTERISTICS OF
AXIATE PATTERNS

BEFORE turning to the experimental analysis of morphogenesis,
, "determination," and differentiation, it seems desirable to devote
some attention to certain physiological characteristics of axiate
pattern. Since agamic development and reconstitution of isolated cells or
pieces occur in many organisms in a definite relation to the pattern of the
parent, some of the physiological features of adult pattern, as well as
those of early developmental stages, are of interest in this connection.

PLANTS

A few points concerning plants are noted because they suggest certain
similarities of plants and animals as regards general pattern. In some
fifty species of axiate algae examined with various agents, differential sus-
ceptibility decreases basipetally in axes with apical growing tips, at least
in the younger parts of the axis (Child, 1916c, e; 1917a, h; 1919/). In the
older parts of the thallus irregularities often appear. Multiaxiate forms
with regularly arranged bipinnate or radial branches show the basipetal
gradient in each branch and in the system as a whole; that is, the suscepti-
bility of the growing tips of branches decreases from branch to branch
basipetally. In axes consisting of a single series of elongated cells the
progress of cytolysis can be observed, not only from cell to cell, but along
each cell from the distal to the proximal end. The whole bipinnate thallus
of the alga, Bryopsis, is a single cell; but the gradient is the same as in sim-
ilar multicellular axes (Child, 1916c). In axes with a basal growing region
susceptibility decreases from the base acropetally; the frond of the kelp
Nereocystis (Child, 1919/) and hairs of the Fiicus thallus (Child, 1917a) are
examples. Unpubhshed data on the differential susceptibility of Vohox to
a large number of agents (see Appendix III, p. 735) show decrease from the
pole of the growing region, the posterior pole in locomotion.

Polarity of the egg of the alga Fucus can be determined by various ex-
ternal differentials (see pp. 423-25). Light is probably the usual deter-
mining factor under natural conditions, but there is evidence that the egg

86

PHYSIOLOGICAL CIL^RACTERISTICS OF AXIATE PATTERNS 87

has a polarity at the time of discharge from the thalkis that may be effec-
tive in the absence of sufficiently intense external differential after dis-
charge. The rhizoid outgrowth appears on one side of the cell before the
first division, and the plane of the first division is normally at right angles

B

Fig. 26, A-C. — Three developmental stages of the alga, Fiicus. A, growth of rhizoid pre-
ceding first division; B, two-cell stage, rhizoid gradient; C, later stage, thallus and rhizoid
gradients.

to the direction of rhizoid outgrowth (see Fig. 26, A, B). The pole oppo-
site the primary rhizoid becomes the apical pole of the thallus. No definite
death gradient has been observed in the eggs before determination of po-
larity by light or some other external factor; but when the rhizoid begins
to grow out, it becomes a region of high susceptibility, and death pro-

88 PATTERNS AND PROBLEMS OF DEVELOPMENT

gresses basipetally from its tip (Fig. 26, A, B). Only in somewhat later
stages, when the developing thallus consists of several or many cells, does
its apical region become more susceptible than other parts. From these
stages on, cytolysis progresses basipetally from, the tip of the thallus and
from the tip of each rhizoid present (Fig. 26, C). Reduction of potassium
permanganate shows the same gradient pattern (Child, 1919a). An elec-
tric polarity is present in single cells of Pithophora and Nitella (Lund,
1938). Mold hyphae developing from spores show a basipetal suscepti-
bility gradient.

Axial differences in rate of respiration in the higher plants are often of
uncertain significance because the proportions of active and relatively
inactive or dead cells and of protoplasm and water, cellulose or other non-
living substances differ at different levels. Structural differences along the
axis of the potato tuber are not great before sprouting occurs: basal
halves produce from 6.5 to 6.8 per cent less carbon dioxide per unit of
weight than apical halves in a variety tested; after apical sprouts begin
to grow, the difference is 41.8 per cent, and after removal of sprouts, 10.7
to 19.8 per cent (Appleman, 191 5). The fact that apical "eyes" tend to
develop first and inhibit others more or less completely suggests that there
is an axiate pattern in the quiescent tuber and that the difference in CO2
production is a feature of it.

Many observations by many investigators have shown that electric po-
larity is a characteristic feature of plant axes, and various sources of origin
of the potential differences observed have been suggested. According to
Lund and his students, the growing tip of the onion root and of other
roots and the tips of the main stem and branches of the Douglas fir are
externally electropositive to lower levels of the same axis, and the tip of
the main axis of the fir is positive to the tips of lateral branches of the first
whorl.' Rate of methylene blue reduction and CO2 production in the root
are highest in the most positive region and decrease basipetally with it to
a level some distance from the tip, where there is a region showing some
increase, followed again by decrease. In both root and fir stimulation of a
tip decreases its positivity and may reverse the potential gradient. The
electric polarity of the fir shows a differential susceptibility to tempera-
ture. Decrease in temperature decreases electric polarity because decrease
of positivity is greater in regions of high than of low positivity; in other
words, a differential inhibition apparently occurs. With return to the

■ Lund, 1928a, b, c, 1929a, b, 1930, 1931a, b, c, 1932a, b; Lund and Kenyon, 1927; Marsh,
1928.

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS 89

original temperature the most highly positive region shows a greater "re-
bound" than lower levels (differential recovery).

The root tip shows an essentially similar relation between potential
difference and local deprivation of oxygen. Removal of oxygen from about
regions of the root which have oppositely oriented electric polarities pro-
duces opposite effects on the potential of the whole. Removal of oxygen
from the region of active cell division at the root tip greatly decreases and
often reverses the potential difference between this and a lower root-level,
that is, the effect is differential (Rosene, 1934; Rosene and Lund, 1935).
Presence of a glutathione or SH-gradient in the growing tips of certain
plants has been reported (Camp, 1929); this perhaps indicates a gradient
in oxidative metabolism. Lund holds that the axial electric-potential
gradient results from a gradient in rate of oxidation.

Determination of oxygen uptake and of flocculability of colloids in
parts of flowers show parallel respiratory and flocculation gradients cor-
responding to the morphological symmetry of the flower concerned. Re-
sults with bilateral flowers are particularly interesting: the gradients are
in the direction of the plane of symmetry. In leguminous flowers tested,
rate of respiration decreases and flocculability of colloids increases from
the superior to the inferior parts of the corolla; in some other bilateral
flowers the gradient is in the reverse direction (Zanoni, 1934a, b).

Analysis shows gradients of various substances along the axes of the
higher plants. These, or some of them, are merely indicative of progres-
sive changes in physiological condition and progressive differentiation
with increasing distance from the growing tip. They are consequences of
the type of pattern characteristic of most plant axes, a growing tip re-
maining embryonic and continuously giving rise to new cells, all or a part
of which gradually differentiate. In general, it seems evident that physio-
logical gradients are characteristic and essential features of axiate pattern
of plants. Their presence can even be shown in single elongated cells of
some multiceUular axes, and they are present in multiaxiate unicellular
plants, such as Bryopsis. The axes of the higher plants are obviously
gradients, at least in the younger regions near the growing tips. When iso-
lated parts of a plant axis undergo reconstitution, these gradients play a
part in determining localization of shoots and roots. Doubtless, gradients
in roots and shoots and in other plant organs differ in character, but per-
haps the presence of a gradient is just as significant in relation to the prob-
lem of pattern as the character of the reactions which occur in it and the
chemical constitution of the substances involved.

go PATTERNS AND PROBLEMS OF DEVELOPMENT

PROTOZOA

An ectoplasmic gradient has been demonstrated in many protozoa,
chiefly cihates, by differential reduction of methylene blue and differential
susceptibility to many chemical and physical agents, and the indophenol
blue reaction and reduction of permanganate have been used with a few
species. Results obtained with the different methods show a rather re-
markable agreement.

In Amoeba susceptibility to cyanide decreases from tip to base of the
pseudopod (Hyman, 1917); and, according to Lynch (1919), animals in
the limax condition, that is, with a temporary anteroposterior axis, show
an anteroposterior gradient in the whole body. Bovie and Barr (1924)
have observed a similar gradient in susceptibility to radiation. Suscepti-
bility and permanganate reduction decrease from the mouth region basip-
etally in Noctiluca. Evidence of a longitudinal ectoplasmic gradient has
been obtained from all ciliates examined for the purpose — some thirty
species; but Paramecium has been more extensively studied than other
forms. With staining which is not appreciably toxic rate of reduction of
methylene blue in Paramecium ectoplasm decreases from the anterior end
posteriorly (Child, 1934&). In very low concentrations of oxidized dye
the ectoplasm of P. caudatum does not stain appreciably, but the ento-
plasm stains slowly, at least in the posterior region (Fig. 27, ^). In
slightly higher concentrations anterior ectoplasm does not stain, while the
posterior ectoplasm stains gradually and sooner or later shows toxic effect
with loss of structure, beginning at the posterior end and progressing ante-
riorly, and finally cytolysis progressing in the same direction (Fig. 27, 5).
These differential stainings occur when the dye solution is exposed to air.
They are not due to deeper staining of the posterior region in consequence
of entrance of dye through the mouth into the posterior entoplasm and
passage into the ectoplasm; the entoplasm may be uniformly stained
throughout, with ectoplasmic staining only in the posterior region. With
further increase of concentration, ectoplasm stains equally throughout;
and with decrease of oxygen, rate of dye reduction decreases from anterior
to posterior end (Fig. 27, C). Ectoplasmic reduction seems to occur a
little more rapidly about the oral region than in the general ectoplasm of
that body-level; this, however, is not entirely certain and is not indicated
in the figures.

The facts suggest that some dye reduction occurs in the ectoplasm at
oxygen-levels in media exposed to air. If this is the case, it appears that

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS 91

with extremely low concentrations all levels of the ectoplasm may reduce
the dye as rapidly as it enters; with slightly higher concentrations, the re-
duction gradient appears with rate of reduction decreasing from anterior
to posterior levels. The only alternative to this suggestion is apparently
a differential ability of the ectoplasm to decolorize the dye in some other
way than by reduction.

In still higher dye concentrations the anterior ectoplasm becomes more
deeply stained than other parts, and reduction is retarded in it or does not

Fig. 27, A-D. — Differential staining and differential reduction in Paramecium (methylene
blue). A, very low concentration, staining deepest parts of posterior entoplasm first; B, higher
concentration, stains posterior ectoplasm and with continued staining, injury, loss of struc-
ture and ability to reduce occur, progressing anteriorly while anterior region remains un-
stained; C, early stage of differential reduction after staining of whole ectoplasm; reduction
progresses from anterior end posteriorly; D, high concentration or long staining in high
oxygen; anterior ectoplasm stains more deeply than rest, and its reducing power is decreased
or destroyed, while posterior region still reduces (from Child, 19346).

occur, while more posterior regions are still uninjured and able to reduce
(Fig. 27, Z)). When the leucobase of methylene blue prepared with hypo-
sulphite (see p. 68) is added to culture fluid containing Paramecium and
exposed to air, it penetrates at once and is almost immediately oxidized
in the animals. For a few seconds, however, an oxidation gradient appears,
rate of dye oxidation decreasing in the ectoplasm from anterior to poste-
rior levels; but staining very soon becomes uniform. Extremely high dye
concentrations, particularly of the leucobase with hyposulphite, are highly
irritating and induce long-continued backward locomotion. Under these
conditions the posterior ectoplasm oxidizes the leucobase and stains more

92 PATTERNS AND PROBLEMS OF DEVELOPMENT

rapidly and more deeply than other regions, and cytolysis begins poste-
riorly, often in I or 2 minutes.^

Most of the observations on oxidation and reduction of dyes have been
made with P. caudatum. The dye-reduction gradient is even more distinct
in P. miiltimicronucleatum, decreasing posteriorly in uninjured animals
and reversed in direction when differential injury retards or abolishes re-
duction anteriorly. As might be expected, the continuously circulating
entoplasm of Paramecium shows no evidence of an intrinsic gradient, the
course of reduction in it being apparently determined by the ectoplasmic
gradient. The indophenol blue reaction (see p. 64) with highly dilute
reagents, so that the reaction occurs before the animals are killed, is most
sharply localized in the inner portion of the ectoplasm or on the boundary
between ectoplasm and entoplasm and decreases in rate from anterior to
posterior end (Child and Deviney, 1926).

Differential susceptibility of P. caudatum to various gradually lethal
chemical and physical agents, as indicated by cytolysis or other changes
in the ectoplasm, decreases from the anterior end posteriorly, with a sec-

=> Only the reversed reduction gradient resulting from differential injury of the ectoplasm
by the leucobase, with retardation or absence of reduction anteriorly, was observed by Roskin
and Semenoff (1933) with the use of a leucobase which was toxic, according to their own state-
ment. Their conclusion that reduction occurs more rapidly in the posterior region is therefore
mistaken. Using thionine reduced by rongalite (see p. 68) and oxidized Janus green, Gersch
(1937) concludes that oxidation of the leucobase occurs more rapidly anteriorly than pos-
teriorly in Paramecium and that reduction of the oxidized dye is more rapid posteriorly than
anteriorly. No data concerning concentrations used or staining periods are given, but the
rongalite-leucobase solution is stated to be somewhat toxic. The author concludes that oxida-
tion is more rapid anteriorly; reduction more rapid posteriorly. The question how oxidation
occurs in the anterior region without reduction or more rapid oxidation with less rapid reduc-
tion, and in the posterior region more rapid reduction with less rapid oxidation, is of interest in
connection with this conclusion. Apparently the dye- reduction gradient decreasing anteriorly,
as observed by Gersch, is, like that found by Roskin and Semenoff, the result of a differential
toxic effect decreasing from the anterior end posteriorly, with reduction most retarded or
absent anteriorly. (See also Kalmus, 1928.) Both the rongalite-leucobase and Janus green
are toxic, the latter extremely so. That it is essential to use a wide range of concentrations
of dye and, with oxidized dyes, different exposure periods, in order to avoid, as far as possible,
misleading results, is sufficiently evident from the results obtained with Paramecium. It is
possible to obtain both an oxidation and a reduction gradient in either direction, according
to concentration of leucobase or oxidized dye used and according to staining period with oxi-
dized dye. Only by varying the procedure sufficiently to determine which gradients most
nearly represent physiological condition and which are results of differential toxic effect is
it possible to attain an adequate basis for any conclusions. To assume that the animals are
uninjured because they continue to swim is not justified. The anterior ectoplasm may be so
injured that it is entirely unable to reduce the dye at all, while the animal is still able to swim.
These reversed reduction gradients in Paramecium resulting from differential toxic effect of
the dye are excellent examples of the differential susceptibility along the anteroposterior axis.

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS 93

ond region of higher susceptibility at the extreme posterior end in some,
not in all, individuals, perhaps associated with reconstitution after fission
or possibly with backward locomotion.' A characteristic effect of exposure
to ultra-violet radiation is shortening of the anterior region (Fig. 28, A-C) ;
with sufficient exposure this may be followed by complete cytolysis of the

G

H

I

Fig. 28, A-L. — Differential susceptibility of Paramecium. A-D, ultra-violet; E, F, HCl;
G, H, methylene blue, high concentration; /, /, neutral KCN; A', L, lack of oxygen (from Child
and Deviney, 1926).

ectoplasm, progressing from the anterior end posteriorly (Fig. 28, D), or
with less intense effect by a similar gradient of loss of structure in the ecto-
plasm and gradual approach to spherical form without destruction of the
pellicle. In a certain range of concentration of HCl a wave of complete
cytolysis of ectoplasm runs from anterior to posterior end (Fig. 28, E, F).
In methylene blue and other basic dyes in toxic concentrations the anterior
region of the ectoplasm gradually becomes more deeply stained than other

3 Child, 1914a; Bills, 1924; Child and Deviney, 1926.

94 PATTERNS AND PROBLEMS OF DEVELOPMENT

parts, then shortens, and clear vesicles appear on its surface, first ante-
riorly, later elsewhere (Fig. 28, G, H). In neutral cyanide anterior shorten-
ing is followed by complete ectoplasmic cytolysis, progressing from the
anterior end posteriorly (Fig. 28, /), or in somewhat lower concentrations
bv ectoplasmic disintegration within the pellicle (Fig. 28, J). In death
from lack of oxygen anterior shortening is followed by appearance of clear
areas within, or clear vesicles on the surface, these changes also progressing
from the anterior end posteriorly (Fig. 28, K, L). In view of the findings
that cyanide does not greatly decrease oxygen uptake in Paramecium,
ectoplasmic differential susceptibility to this agent and to lack of oxygen
is of particular interest (see Appendix III, p. 736). The fact that the same
susceptibility gradient appears with both strong and weak acids and bases,
ultra-violet radiation, and lack of oxygen is also of interest, as indicating
that differential permeability is not the chief factor in determining the
gradient. An ectoplasmic gradient of osmophilic substance, decreasing
from the posterior region anteriorly, was found by Parke (1929) in P.
caudatum. The anterior vacuole of this species has a more rapid rhythm
than the posterior (Child and Deviney, 1926; Unger, 1926); but in P.
aurelia the posterior vacuole is more rapid, and in P. calkinsi the rate of
both is the same. These species differences in relative rate of vacuoles are
perhaps associated with difference in the gradient of the different species
and differences in form of body. Susceptibility and reduction gradients ap-
pear to be more strongly marked in P. caudatum than in P. aurelia, but in
general the posterior vacuole apparently receives fluid from a larger part of
the body than the anterior; this may be a factor in determining its higher
rate of contraction in some species.

Other ciliates — Frontonia, Spirostomtim, Dileptus, and Woodruffia"^ —
show an anteroposterior ectoplasmic gradient of dye reduction. The gradi-
ent of Dileptus is particularly interesting; the whiplike, highly motile ecto-
plasmic organ at the anterior end reduces methylene blue much more
rapidly than the rest of the body, if not too deeply stained, and is also
much more susceptible than other parts to toxic action of dyes (Child,
1934&). Spirostomum is highly susceptible to dyes; but when slightly
stained with methylene blue, a slight reduction gradient appears; with
deeper staining this is reversed in direction in consequence of differential
injury by the dye. The susceptibility gradient in these forms parallels the
reduction gradient, except that in some individuals of Spirostomum a sec-
ond posteroanterior gradient appears in the posterior region. Spirostomum

* Child, 19346, except unpublished data on Woodruffia.

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS 95

moves backward almost as frequently as forward; consequently, the possi-
bility suggests itself that this secondary gradient, evident only in some in-
dividuals of a lot, may, like the reversed oxidation gradient of Paramecium
(p. 91), be associated with frequent backward locomotion. Cytolysis of
Stent or coeruleiis in cyanide progresses over the body in a wave from the
peristome, and in VorticeUa and Carchesium the differential in the body is
similar. The contractile stalks of these animals, however, show evidence
of specific susceptibility to certain agents. According to Merton (1929),
their susceptibility to formol is less than that of the peristomial cilia, but
greater to lactic acid, pilocarpin, and alcohol; indications of specific sus-
ceptibility of the contractile fiber of the stalk in these forms have also been
observed with other agents (Child) . This fiber, capable of extremely rapid
contraction, undoubtedly is a highly differentiated part of the cell; it is
not surprising, therefore, to find it specifically susceptible to particular
agents. All hypotrichous ciliates examined for differential susceptibility,
with one exception,-^ showed anteroposterior gradient. Anteroposterior
susceptibility and indophenol gradients were observed in a Monocystis
from a marine poly clad.

The susceptibility gradient of holotrichous, heterotrichous, and hypo-
trichous ciliates is evident, not only in change of form and cytolysis but
also in retardation and cessation of ciliary movement, which progresses
from the anterior end posteriorly, except in those individuals which show
a posteroanterior gradient in the posterior region ; in these this gradient
appeared in the cilia.

The evidence for the existence of a gradient pattern in axiate protozoa
has been presented in considerable detail because presence in the morpho-
logically differentiated part of a single cell of a spatial physiological pat-
tern showing essentially the same characteristics as are found along the
axes of multicellular animals, at least during earlier developmental stages
and often throughout life, is significant as indicating that organismic pat-
tern is independent of cell boundaries. The same physiological features of
pattern appear within the single cell and in the individual consisting of
millions of cells. That an axiate, or, more strictly speaking, polar, proto-
zoan pattern consists only of these gradients in the fully developed animal
is certainly far from true, but that they are essential factors in the develop-

5 Siylonychia, Oxytricha, Onychodromits, Kerona, and Euplotes all show an anteroposterior
ectoplasmic gradient. In a Japanese species somewhat similar to Uronychia a posteroanterior
gradient was observed. Whether this is associated with the very large and highly developed
posterior cirri, as seems probable, or with some other factor could not be determined, since
only a few individuals were found and since attempts at cultivation were not successful.

96 PATTERNS AND PROBLEMS OF DEVELOPMENT

ment of the morphological pattern appears beyond question. They are
evidently concerned in reconstitution of individuals in fission or after ex-
perimental isolation of parts. Moreover, it may at least be questioned
whether axial co-ordination of ciliary beat and definitely directed locomo-
tion are possible without a gradient pattern.

SPONGES, COELENTERATES, AND CTENOPHORES

SPONGES

Oxygen uptake and CO2 production are, in general, greater in apical
than in basal pieces of the sponge commonly known as Grantia.^ Accord-
ing to Hyman and Bellamy (1922), the oscular region is galvanometrically
negative to more basal levels. In longitudinally split individuals of Gran-
tia and Leucoselenia and several other elongated species with single oscu-
lum, the rate of reduction of permanganate and the resulting depth of
color decrease from the osculum basipetally.

HYDROZOA

Gradients of early embryonic stages and their changes in the course of
development, as indicated by differential susceptibility and reduction of
permanganate and in some cases by the indophenol reaction, are known
for several species (Child, 1925a). The oocyte of the calyptoblast hydro-
zoan Phialidium gregarium is attached in the gonad by one pole, the future
vegetal pole, the free pole becoming the animal or apical pole (Fig. 2g,A).
Polar bodies form at the apical pole, the first cleavage furrow progresses
from it, and it becomes the apical pole of the early blastula and planula
and the high end of the primary gradient (Fig. 29, B, C, D). In the later
planula a second gradient appears at the original basal end (Fig. 29, D);
the primary gradient becomes less distinct; the planula, if in good condi-
tion, attaches itself by the original apical end; and the secondary basal
gradient becomes the hydranth-stem gradient (Fig. 29, E). In the later
development of the branching hydroid the growth form is sympodial, that
is, each new hydranth bud is only temporarily the apical member and be-
comes a lateral branch when the next bud appears (Child, i9i9c^). The
reversal of polarity and gradient in development of the hydranth from the
planula may be regarded as the first step in sympodial development ; the
first hydranth bud arises as a new axis from the basal regions of the plan-

<> Oxygen uptake higher in eleven of thirteen determinations on different lots; CO2 produc-
tion higher in ten of eleven determinations (Hyman, 1925).

PHYSIOLOGICAL CIL^RACTERISTICS OF AXIATE PATTERNS 97

ula. The same gradient relations have been found during development of
two other calyptoblast species. The planula of a Japanese species, Ser-
tularella minurensis , develops the secondary gradient in the original basal
region at a somewhat earlier stage; in the fully developed planula the
end originally basal is already more susceptible than the apical end, and
the secondary gradient has extended over most of the body length.

/-: .-*^

D

Fig. 29, A~E. — Susceptibility gradients in development of the hydrozoan Pkialidinm. A,
oocyte isolated by teasing; B, first cleavage (in A and B disintegrated portion indicated in
broken line); C, early planula with gradient from apical end; D, later planula with second
gradient at original basal end; E, after attachment by original apical end; secondary gradient
of basal end becomes hydranth-stem axis. Arrows show direction of progress of disintegration
(from Child, 1925a, with modifications).

The egg of the gymnoblast hydroid Corymorpha palm a, when shed,
sinks and adheres firmly to any solid substratum. A gradient is present
in early stages with high end at the free pole, but it has not been possible
to determine whether the egg orients itself in sinking or whether the gradi-
ent results from the differential in conditions from the free pole to the pole
in contact. The planula never swims but may progress a short distance in
contact with the substrate with apical end in advance and secreting peri-
sarc as it goes. In the early planula differential susceptibility, permanga-
nate reduction, methylene blue reduction, and the indophenol reaction

98

PATTERNS AND PROBLEMS OF DEVELOPMENT

agree in indicating a single gradient with high end apical (Fig. 30, A).
Somewhat later a secondary gradient appears at the basal end but re-
mains short (Fig. 30, B) and becomes the region of holdfast origin (Fig.
30, C), while the hydranth develops at the primary apical end.

Determinations of oxygen uptake, comparative estimations of CO^ pro-
duction, differential reduction of methylene blue and permanganate, and
electric-potential differences all give evidence of gradient pattern in adult
hydrozoa. Corymorpha palma is exceptionally favorable material for re-

\^

Fig. 30, A-C. — Development of the hydrozoan Corymorpha palma from early planula. .4,
early planula with single basipetal gradient; B, later planula with secondary acropetal gradient
in basal region; C, young hydroid, primary gradient becomes hydranth-stem axis (from Child,
1925a).

spiratory determinations because of large size — it may attain a length of
10-12 cm. — and because of absence of perisarc over most of the stem
length and complete absence of branches.^ Determinations of oxygen up-
take at different levels of the naked stem region are given in Table 3, and
comparative data on CO^ production in Table 4.

Other CO2 estimations showed a much higher rate in hydranth than in
stem, and in short distal than in short middle pieces. The question of sig-
nificance of these data and criticisms advanced are discussed in Appendix I
(p. 729). The respiratory gradient parallels the gradient of methylene

7 Hyman, i923(j; Child and Hyman, 1926.

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS 99

blue reduction and the susceptibility gradient, which concern the ecto-
derm only. The respiration of the large supporting cells which till the in-
terior of the Corymorpha stem is probably extremely low. Determinations
of oxygen uptake in stems of Tuhularla (Hyman, 1926a) and Ohelia (Lund,

TABLE 3

Oxygen Uptake of Distal and Proximal Halves and Distal

Middle, and Proximal Thirds of the Naked Stem

of Corymorpha palma

(From Child and Hyman, 1926)

HALVES

Determination
Number

Hours from

Section to

Determination

Respiration

Period in

Hours

Oxygen Uptake per Gram

PER 24 Hours in

Cubic Centimeters

Distal

Proximal

I

2

20

20

18

18

16

7

7

7

21
21
22

22

24

22
22
22

0.97
0.99
1.89
2. 19

3 69

1.42
2.08
1.72

0. 70

0-93
1.29
1.89
2 .69

7

A

c

6

I. 17
1.48

7

8

1-53

THIRDS

Determination
Number

Hours from

Section to

Determination

Respiration

Period in

Hours

Oxygen Uptake per Gram

PER 24 Hours in

Cubic Centimeters

Distal

Middle

Proximal

I

16
16

20
20
20
20

23
23
23
'-i
24
24

1.63
1-65
2.66
2.78
4.86
7.01

1.28
I .46
2.06
2.18
3 04
3.82

IIS
1 .46

1-75
1.98

2

3

4

5

6

2.62
2. IS

1928c, 1931a) also show a decrease basipetally; and, according to Lund,
oxygen uptake shows a differential susceptibility to cyanide; that is, the
decrease in oxygen uptake by a given concentration of KCN is greater in
apical than in basal pieces. Barth {Biol. Bull., 78, 1940) also finds a
basipetal decrease in oxygen uptake in Tubularia.

lOO

PATTERNS AND PROBLEMS OF DEVELOPMENT

TABLE 4

Comparative CO^ Production of Different Stem Regions of

Corymorpha palma COLORIMETRICALLY ESTIMATED WITH

Phenolsulphonephthalein as Indicator
(From Child and Hyman, 1926)

Estimation
Number

Body
Length in
Centi-
meters

Region

No. of
Pieces

Weight
in Grams

Hours
from
Section
to Esti-
mation

Hours pH

Change

from 8.1

to 7.3

I

10-12

dist. 1/2
prox. 1/2

6
6

0.729
0.7285

18

4:4s
5:20

2

10

dist. 1/2
prox. 1/2

5
5 .

0.2525
0.2845

43

7:30

9: 10

7.

8-10

dist. 1/2
prox. 1/2

7
7

0.2398
0.2522

48

8:30

9:30

A

6-7.5

dist. 1/2
prox. 1/2

12
9

0 . 3005
03057

17:30

4:00

6:05

e

5

dist. 1/2
prox. 1/2

9

8

O.I557
0. 1614

18

6:15

8:00

dist. 1/3

18

0. 2042

6:00

6

6-7.5

mid. 1/3

II

0.2073

20

8:15

prox. 1/3

12

0. 2212

8:35

dist. 1/3

20

0.2333

5:1s

7

6

mid. 1/3

17

0. 2397

18

T- 15

prox. 1/3

15

0.2512

8:30

dist. 1/3

22

0.1696

6:30

8

6-7

mid. 1/3

19

0.1743

18

7:50

prox. 1/3

14

0.1717

II :oo

9

6-7 ■ 5

mid. 1/3
prox. 1/3

II
6

0.1369
0.1425

22

9:00
10:45

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS loi

In regard to data on differential dye reduction in adult hydrozoa, the
use of methylene blue with Pelmato hydra oligactis, the common brown
hydra, shows reduction progressing basipetally in each tentacle and basip-
etally from the hypostome to, or almost to, the more slender stalk. In
unattached animals the foot usually reduces earlier than the stalk, and re-
duction progresses acropetally over more or less of the stalk. In attached
animals or advanced buds attached to the parent the stalk usually reduces
later than the body proper. Motor activity, local contractions of body or
stalk, may alter the reduction gradients to a considerable degree. In un-
attached animals the stalk is usually more or less motile, contracting and
extending frequently, and reduces more rapidly than when inactive. When
an animal bends strongly to one side while reduction is proceeding, the
color may disappear completely from the contracted side and reappear
when relaxation occurs. This local reduction accompanying or following
contraction, with reoxidation of the dye on relaxation, has been observed
repeatedly.

Methylene blue reduction progresses basipetally in each tentacle of
Corymorpha and in the hydranth body if not altered by contraction. That
oxygen uptake of the hydranth is relatively great is indicated by occur-
rence of dye reduction in open solutions of dye in the hydranth body and
the proximal regions of the crowded tentacles if the hydranth is left with-
out change of position for even a few moments. Frequent change of posi-
tion or increase of oxygen about the hydranth by agitation of the medium
brings about reoxidation. The dependence of hydranth reconstitution on
oxygen tension in the related genus, Tuhularia, indicates similar condi-
tions as regards oxygen uptake of the hydranth (Barth, 1938^). Reduc-
tion also progresses basipetally in the ectoderm of the naked Corymorpha
stem, except for a slight increase in rate of reduction in the extreme prox-
imal region. From the basal region of the stem, inclosed in perisarc, nu-
merous holdfast stolons arise as buds and elongate with growing region at
the free end; in each of these rate of reduction decreases from the tip basip-
etally (Child and Watanabe, 19356). In isolated pieces of the Cory-
morpha stem, rate of reduction adjoining a cut end is increased following
section, and a gradient of decreasing rate extends from the cut end for a
greater or less distance. The region from which the hydranth will develop
later is the high region of this gradient, and within a few hours after sec-
tion the length of the hydranth primordium is indicated by a well-defined
region at one or both ends of the piece, where reduction is more rapid than

I02

PATTERNS AND PROBLEMS OF DEVELOPMENT

elsewhere. Even at this stage, before any morphological evidence of hy-
dranth development appears, it can be determined, from the presence of a
region of rapid reduction at one or at both ends of the piece, whether it
will develop as a unipolar or a bipolar form (Fig. 31).^

Electric-potential gradients have been observed in a considerable num-
ber of hydroids.'' Mathews, Hyman, and Bellamy and Child and Hyman
found apical levels galvanometrically electronegative to more basal levels.

ABC

Fig. 31, A-C. — Differential reduction of methylene blue in isolated pieces of Corymorpha
stem. A, bipolar form; B, unipolar with distal hydranth; C, unipolar with proximal hydranth.
Arrows indicate direction of progress of reduction (from Watanabe, 1935c).

According to Lund, however, galvanometric positivity decreases from the
apical region basipetally and increases again toward the basal end of the
main axis of Obclia. Lund's data concern cut pieces without hydranths.
Barth docs not find uniformity of potential difference in different species
or in the same species under different conditions. The hydranth of Tubu-
laria is galvanometrically negative to middle levels, both distal and prox-

* Child and Watanabe, 1935&; Watanabe, 193SC.

9 Mathews, 1903, Parypha {Tuhularia), Pennaria, Canipamdaria; Hyman, 19206,- Hyman
and Bellamy, 1922, Tubiilaria crocea, Obelia geniculata, 0. borealis, Schizotricha lenella, Eiiden-
drium ramosum, Pennaria tiarella; Lund, 1922, 1923a, 19246, 1925, 1926; Rosene and Lund,
1930, 0. commisuralis; Child and Hyman, 1926, Corymorpha palma; Barth, 19346, Tuhularia,
Pennaria, Eudendrium.

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS 103

imal cut ends of a piece are negative to the middle region, a cut end is
negative to the hydranth, and a distal cut end is negative to a proximal
but may become positive during reconstitution. In Pennaria pieces an
end undergoing reconstitution is usually externally negative to other re-
gions; in Eudendrium it is usually positive. It does not appear from
Earth's data whether the potential of a developing hydranth differs from
that of a developing stolon or whether reconstitution of the Tuhularia hy-
dranth by reorganization of cells already present and reconstitution by re-
generation, that is, by outgrowth of new tissue from the cut end before
hydranth development, may account for some of these species differences.
At present the data appear to indicate that, in general, the apical region
is externally electronegative in intact hydroids with fully developed hy-
dranths; in forms like Tuhularia and Corymorpha, in which hydranth re-
constitution occurs by redifferentiation of a part of the stem without out-
growth of new tissue, the reconstituting hydranth is usually externally
negative to other regions; in forms like Obelia, in which outgrowth of new
tissue resembling a stolon precedes hydranth development at its free end,
this growing region is externally positive. Moreover, it is possible that the
sign of potential difference between the outgrowth and other levels of the
piece may differ according as the outgrowth is stolonic in character or a
hydranth-stem axis, and the potential of outgrowing tissue may differ
from that of the differentiating or differentiated hydranth. Further inves-
tigation is necessary to determine whether these suggestions have any
value; but concerning the presence of electric-potential differences as
characteristics of axiate pattern in hydroids there can be no doubt. Lund
regards the potentials as associated with, and dependent on, oxidation-
reduction. He has shown that in Obelia the axial potential difference can
be reversibly decreased or reversed in direction by cyanide, ether, and
chloroform — in other words, the axial potential exhibits a differential
susceptibility to these agents.

In several species of hydromedusae Hyman and Bellamy found the
distal end of the manubrium most strongly negative externally, the mar-
gin of the umbrella next, then the subumbrellar surface, and the exum-
brellar surface positive to other parts.

The galvanotactic reaction of hydra and of various hydroids and medu-
sae shows definite relation to the gradient pattern of the body or part con-
cerned."

The gradients indicated by differential susceptibility and differential

" Hyman, 19326; Bancroft, 1904; Hyman and Bellamy, 1922.

I04 PATTERNS AND PROBLEMS OF DEVELOPMENT

reduction of permanganate in the body of hydra, in tentacles, hydranths,
stems, and stolons of hydroids, in medusa buds and fully developed medu-
sae parallel closely those observed by other methods. They parallel the
electric-potential gradients, but apical regions which have been found to
be externally electropositive and those electronegative to other regions ap-
pear as regions of high susceptibility and high rate of permanganate re-
duction. Assuming that the data on potential differences are correct and
that the direction of sign change may result from different kinds of physi-
ological activity in different species, particularly in reconstitution, this
lack of correspondence as regards direction of gradient is to be expected,
since differential susceptibiUty and differential reduction apparently indi-
cate merely differences in rate or intensity."

It is a point of particular interest that early stages in the form of buds
of hydranths, tentacles, medusae, and stolons appear as local gradient sys-
tems, at first more or less radial from a center and becoming longitudinal
as outgrowth proceeds. In the naked contractile hydra, however, the high
end of the bud gradient is not always continuously apical. Both the reduc-
tion and the susceptibility gradient of the bud may be reversed by con-
tractile activity of the parent body, which often involves the basal region
of the bud. Contractile activity of the stalk in detached animals increases
its susceptibility, and bending of the hydra body may increase suscepti-
bility of the contracted side ; these effects parallel those observed in dye
reduction (p. loi).

In branching hydroids with more or less definite growth form of the
whole {Pennaria, Obelia, Gonothyraea) the multiaxiate system shows a
basipetal decrease in susceptibility similar to that in multiaxiate algae
(p. 86). In these cases susceptibility of the terminal hydranths decreases
from the apical hydranth of the system basipetally and also basipetally
in each branch, but with occasional irregularities. This gradient is in the
same direction as the gradient in rate of hydranth reconstitution observed
in another species of Pennaria (pp. 38, 39).

The gradients in a monophyid siphonophore are of interest. The coeno-
some or stolon arises from one side in the median plane of the bilaterally
symmetrical nectomedusoid. Groups of zooids, each consisting of a nu-
tritive zooid or manubrium with tentacles developing near its base, a me-
dusoid, and a bract, develop successively from the region of origin of the

" For data on differential susceptibility and permanganate reduction see Child and Hyman,
1919; J. W. MacArthur, 192 1 ; Wcimer, 1928, hydra, several species; Child, 1926a, Corymorpha;
Child, 1919a, d, i()2id; Uyman, 19206, various hydroids and medusae.

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS 105

coenosome; and successive separation of the groups occurs at the tip, the
medusoid of the group becoming the nectomedusoid of a new system and
giving rise to a new coenosome. The developing zooids on the coenosome
show the usual radial gradient system of buds which becomes longitudinal
with outgrowth. Susceptibility and reduction gradients in the subum-
brellar ectoderm of the fully developed nectomedusoid are bilateral, that
is, susceptibility and rate of reduction decrease from the side opposite the
coenosome to the region where it develops. This bilaterality appears as a
gradual change from the basipetal gradient of the earlier medusoid bud
and is evident before morphological bilaterality is distinguishable in the
developing medusoid attached to the coenosome. That side of the bud
toward the attached end of the coenosome becomes the high end of the bi-
lateral gradient; and the highest region on that side is next to the velum,
that is, the polar gradient is not entirely obliterated. Susceptibility and
rate of reduction increase with development of motility in the zooids.
The definite and constant relation of the bilaterality of the nectomedusoid
to the axis of the coenosome or stolon to which it is attached suggests that
it originates in a gradient along the coenosome axis. The medusoid origi-
nates as a bud at right angles to the coenosome axis; and the side of the
bud toward the attached, physiologically younger end of the coenosome
becomes the high side of the bilateral gradient. The polar gradient of the
medusoid originates from the radial gradient system of the early bud,
which becomes longitudinal with outgrowth.

OTHER COELENTERATES

Susceptibility and rate of reduction decrease basipetally in the ectoderm
of the Aurelia scyphistoma and the sessile scyphozoan, Haliclystus
(Child). A gradient in the indophenol blue reaction has also been shown
to be characteristic of polarities originating in reconstitution of the stalk
of Haliclystus (Watanabe, 1937). Oxygen uptake of exumbrella and
mesogloea of the scyphomedusa Cassiopea xamachana is only about 25 per
cent of that of the intact animal at rest; since respiration of the meso-
gloea is extremely low, the exumbrellar epithelium evidently has a much
lower respiration than that of the subumbrella. How much of the differ-
ence is due to absence of entoderm in the exumbrella is not known (Mc-
Clendon, 1917). Developmental stages of an alcyonarian "sea pen" from
early cleavage to the first polyp show basipetal decrease in rate of dye re-
duction.

Few data concerning actinians are available. Young individuals of

io6 PATTERNS AND PROBLEMS OF DEVELOPMENT

Epiadis and Peachia show basipetal decrease in susceptibility and rate of
reduction in each tentacle and in the body. In colorimetric determinations
of CO2 production of distal and proximal pieces of the actinians Sagartia
luciae and Metridium marginatum no definite difference was found (Park-
er, 1929), but the data have httle or no significance.'^

Entodermal susceptibility and reduction gradients in these coelenter-
ates appear, in general, to be the same as the ectodermal. In well-fed
hydras and hydroids the entoderm of the apical region may cytolyze be-
fore the ectoderm; but in other regions entodermal susceptibility is ap-
parently about the same as that of the ectoderm, except locally in the re-
gion of recently ingested food undergoing digestion. In starved animals
entodermal susceptibility may be distinctly less than that of the ectoderm.
Entodermal dye reduction usually follows ectodermal reduction and pro-
gresses in the same direction. In the solid tentacles of hydroids and in the
Corymorpha stem with its core of supporting cells of entodermal origin it
can usually be seen that dye reduction is most rapid in the outer parts of
the entoderm cells, those parts next to the ectoderm. Respiration is prob-
ably very low in these cells, and dye reduction in them may be largely an
induced result of the much higher oxygen uptake of the ectoderm. The
question whether there is an intrinsic gradient in the entoderm of the
hydrozoa remains open.

CTENOPHORES

Differential susceptibility of Mnemiopsis, BoUnopsis, Pleurobrachia,
and Beroe decreases, in general, from the apical region in the oral direction
with increase in the oral region of Pleurobrachia and in the lobes of the lo-
bate forms, Mnemiopsis and Bolinopsis. The swimming-plate rows also ex-
hibit a very marked gradient in the same direction, but this may be altered
by differences in motor activity of the plates at different levels of a row.
In the higher, more rapidly lethal concentrations of various agents in
which movement of all plates is completely stopped within 1-2 hours, ces-
sation of movement and death of the plates and their bases with change
from the transparency of the living plates to opaque white begin at the
aboral (apical) end of the row and progress orally from plate to plate. In

" The possibility of stimulation of the pieces following section or of laceration of the foot
in detaching the animals is not considered; time between section and determination is not
given; and only one set of determinations on each of four lots is presented. Moreover, since
the foot is a secretory organ, it may have a high CO, production, and this may be increased
by detachment, even without laceration. Determinations on at least three pieces— apical,
middle, and basal— must be made and repeated at different periods after section before any
conclusions can be drawn. See Child and Watanabe, 1933.

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS 107

somewhat lower, but still finally lethal concentrations, the impulses from
the apical nervous organ decrease in frequency so long before the more
oral regions of the row are appreciably injured by the agent that these
latter regions become physiologically isolated (pp. 327-28) and develop
a vigorous beat entirely independent of the apical region, while the aboral
part of the row is much less active or its plates have ceased to beat. If
these conditions persist for an hour or more before death, the more active
plates toward the oral end of the row may die before the less active or in-
active aboral plates. The change of the plates to opaque white at death is
apparently due to coagulation of their colloids. When movement is in-
hibited soon after exposure to the agent and death occurs in an hour or
two, the base of each plate becomes white earlier than the plate itself, and
the change in appearance progresses from base to tip of the cilia compos-
ing the plate. When active movement of plates continues for several
hours during exposure to the agent, with final death, coagulation begins
at the tips of the cilia and progresses basipetally in each plate, the plate
base being the last to undergo the change.

The lobate genus BoUnopsis is extremely sensitive and often undergoes
complete and practically instantaneous disintegration when subjected to
sudden shock. Disintegration may be induced by sudden exposure to cer-
tain chemical agents, but with different concentrations or rates of addition
of the agent to sea water the degree of disintegration may be varied and
controlled. In concentrations near the lower limit of shock effect disin-
tegration may occur only after several seconds or even i or 2 minutes. In
such cases disintegration begins at the apical (aboral) pole and at about
the same time at the tips of the oral lobes and progresses orally over the
body and aborally in the lobes. It may progress from the apical pole over
a fourth or a third of the body and stop, leaving the rest of the body, or all
except the tips of the lobes, intact. In short, the disintegration gradient
following shock is the same as the death gradient in slowly lethal agents
which do not produce shock effect (Child, 1917c, 1933a). The ctenophore
plate row is a particularly good example of a gradient with functional ex-
pression showing relations of dominance, subordination, and physiological
isolation and under some conditions even reversal in direction, essen-
tially similar to relations and alterations in developmental gradients. The
reversal in direction of the death gradient in the cilia of the plate in rela-
tion to quiescence and motor activity is another feature of interest. The
changes in the ctenophore plate row do not result in altered morphological
development, but they do bring about altered functional development.

io8 PATTERNS AND PROBLEMS OF DEVELOPMENT

As regards gradient pattern and its changes, it will perhaps appear more
clearly in following chapters that there is considerable similarity between
this functional pattern and physiological patterns of development.

Transplantation and reconstitution of pieces of Mnemiopsis give evi-
dence of a persistent polarity but not of a gradient in time of reconstitu-
tion in pieces from different body-levels/-^

TURBELLARIA
PLANARIANS

Because of their capacity for reconstitutional development, planarians
have been extensively used by many investigators as material for various
lines of experiment. The relation of reconstitutional pattern to body-level
of origin of the piece, as described in chapter ii, indicates that axiate pat-
tern of the parent body plays an essential part in determining the pattern
of reconstitution in the isolated piece. How far this is true will appear
more clearly in following chapters. Assuming that it is true, it is obvious
that information concerning the physiological characteristics of planarian
axiate pattern may be expected to aid in analysis of reconstitution and
perhaps of other forms of development.

Determinations of oxygen uptake in pieces from different body-levels
of Dugesia dorotocephala by the Winkler method, with due care to elimi-
nate, as far as possible, sources of error and complicating factors, give
definite and consistent results.^^ Determinations on pieces of different
lengths at different periods after section showed that, following section,
rate of respiration of planarian bodies cut into short pieces (i/8) is much
higher than when cut into longer pieces (1/3) and that the longer pieces
show only a slightly higher rate than bodies with only head and posterior
end removed. Averages from the determinations in Hyman's Table i are:
for bodies with head and posterior end removed, 1.40 cc. of oxygen per

'3 Coonfield, 1936a, b, 1937a, b; Coonfield and Goldin, 1937.

'^ Hyman, 1923&. Since an earlier study of the effect of feeding and starvation on the
respiration of this species (Hyman, 19196) had shown that a marked increase in oxygen uptake
occurs on feeding, and that, in absence of food, oxygen uptake decreases in about a week to a
level which remains almost constant for some weeks following, and since size of the digestive
tract differs at different levels, the animals for determination were kept without food 6-18
days in order to eliminate, as far as possible, effects on total respiration of differences in
activity in different parts of the digestive tract. Animals of the same length, from one stock and
without sex organs, and pieces as nearly as possible the same length were used in all lots to
be compared. In all determinations head and posterior end were removed so that a cut sur-
face was present at each end of each piece. Each determination was repeated on from seven
to eleven different lots of material. See Appendix I (p. 729).

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS 109

gram per hour; for thirds, 1.5 1, and for eighths, 2.06. It may be noted
here that differences in susceptibility to cyanide show similar differences
in long and short pieces following section. It was also determined that in
1/8 pieces from anterior and posterior regions of the anterior zooid, that is,
just posterior to the head and in the mouth region, the rate of oxygen up-
take is about the same a short time after section, while 23-26 hours after
section rate of the posterior pieces is distinctly lower than that of the ante-
rior pieces. Here, again, susceptibility to cyanide shows the same like-
ness immediately after section and lower susceptibility of posterior pieces

TABLE 5

Oxygen Uptake of Anterior and Posterior Halves of
THE Anterior Zooid and a Piece of the Same Length
from the Posterior Zooid Region of Dugesia doroto-

cephala; TEMPERATURE 20° ± 1° C.

(From Hyman, 19236)

Number of
Experiment

3

4
5
6

7
8

9

10

Days since
Feeding

14
15
16

14
IS
13

14

Cubic Centimeters of Oxygen
Consumed per Gram per 24^ Hours

Anterior

Half of

First Zooid

Posterior

Half of

First Zooid

Piece from

Posterior

Zooid

later. These experiments show, first, that respiratory rate is increased in
short pieces and little or not at all in long pieces following section ; second,
that in pieces from the posterior region of the anterior zooid oxygen up-
take immediately following section is about equal to, but after 24 hours
lower than, that in anterior pieces of the same length.

Table 5 gives data on oxygen uptake of pieces of equal length from
three regions of the planarian body, anterior and posterior halves of the
anterior zooid, and pieces of equal length from the posterior zooid re-
gion. Pieces of this length show no marked change in respiration during
the first 24 hours following section; consequently, the experiments were

no PATTERNS AND PROBLEMS OF DEVELOPMENT

allowed to run 15-24 hours, and the oxygen uptake was calculated per
gram per 24 hours. In all ten lots anterior halves of the anterior zooid
consumed more oxygen than posterior halves; and posterior zooid pieces,
about the same as the anterior pieces. Susceptibility to cyanide gives sim-
ilar results.^^

Colorimetric estimations of CO, production on pieces of D. doroto-
cephala gave results essentially similar to those of Hyman as regards in-
crease of respiration following section in shorter pieces and presence of a
gradient in the anterior zooid (Robbins and Child, 1920). From colori-
metric CO2 determinations on pieces of D. tigrina Parker (1929) concluded
that this species shows no evidence of a gradient; but, as a matter of fact,
his data agree with and confirm those of Hyman and Robbins and Child."'

Because of their pigmentation and susceptibility to basic dyes, most
planarian species are not favorable material for observation of dye reduc-
tion; but reduction can be followed readily on the unpigmented, or only
slightly pigmented, ventral surface of a number of pigmented species. In
D. dorotocephala dye reduction (Janus green) occurs most rapidly, and at
about the same time, anteriorly and in the posterior zooid region and pro-
gresses posteriorly in the anterior zooid, the last region to reduce being
postoral, just anterior to the fission zone. In animals long enough to have
a second posterior zooid the posterior part of the posterior zooid region
may show slightly more rapid reduction than its anterior part. Differen-

'5 Dr. R. M. Fraps permits mention of unpublished data obtained with a respirometer de
vised by him (Fraps, 1930) possessing certain advantages for this type of experiment. These
data confirm Hyman's determinations as regards difference of respiratory rate at different
body- levels in Dugesia.

'^ Parker finds no definite difference in anterior and posterior halves. None is to be expected,
because this species, like D. dorotocephala , has a posterior zooid region with a higher respiration
than the posterior part of the anterior zooid. Table 5 shows that a piece from the anterior re-
gion has about the same oxygen uptake as one of equal length from the posterior zooid. In
another series of experiments, using the second to the fifth i /6 pieces, anterior and posterior
sixths being discarded, the average rates of CO^ production in milligrams per gram per
minute are: second 1/6, 0.00721; third, 0.00694; fourth, 0.00702; fifth, 0.0071 1. Second and
third pieces are in the anterior zooid, and COj production is less in the third than in the second;
the fourth sixth includes a part of the second zooid, and its CO, production is consequently
higher than that of the third. The fifth piece is wholly in the posterior zooid region, perhaps
even in a third zooid (see pp. 41, 321) and has a still higher CO2 production. These results are
exactly according to expectation on the basis of the other data on respiration and data on
head frequency and susceptibility to cyanide. The only question is whether the differences of
Parker's data are large enough to be significant. Whether the animals used in these experi-
ments were sexually mature is not known. The presence of the genital cloacal complex posterior
to the mouth may increase the respiration of this region over that in asexual animals. Deter-
minations of oxygen uptake of planarians by Shearer are discussed in .\ppendix 1 (p. 730).

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS iii

tial reduction in the unpigmented Procotyla, a form without posterior
zooids, progresses from head to posterior end, except that the extreme pos-
terior tip may reduce somewhat more rapidly than levels slightly anterior
to it. The curves of Figure 32 indicate diagrammatically the differences in
the reduction gradients of Dugesia and Procotyla. They are intended
merely to show roughly the differences in rate of dye reduction, as indi-
cated by decoloration of methylene blue or change to red of Janus green

Fig. 32. — Diagrammatic curves without quantitative significance, indicating differential
dye reduction in Dugesia dorotocephala {upper) and Procotyla flitviatilis Qoicer). A, anterior
end; F, level of fission;ordinates indicate degrees of decoloration or color change; abscissae,
body-levels.

at a stage when reduction is about completed in the head region. In the
upper curve for D. dorotocephala the rate decreases from the head to the
fission zone, where it increases rather abruptly and remains high in the
posterior zooid region. The curve for Procotyla (below) indicates a con-
tinuous decrease in rate from anterior end posteriorly, except at the
posterior tip.^'^

'7 With slow staining, planarian ectoderm is rather susceptible to basic dyes; cytolysis
usually begins in the head region, while the body is only slightly or not visibly stained. With
relatively high concentrations (e.g., 1/20,000 Janus green, 1/5,000 or higher methylene blue)
and with rapid oxygen decrease during staining, it is possible to demonstrate the reduction
gradients on the ventral surface before cytolysis begins (Child, unpublished).

112 PATTERNS AND PROBLEMS OF DEVELOPMENT

Differential susceptibility of planarians to many agents has been deter-
mined (see Appendix III, p. 734). In species without posterior zooids and
fission susceptibility of ectoderm and body wall, as indicated by disinte-
gration or other death changes, progresses from the head region posterior-
ly over the whole body length, except that the extreme posterior tip is
usually somewhat more susceptible than regions immediately anterior to
it, agreeing in this respect with results of dye reduction. Species develop-
ing posterior zooids and undergoing fission show a less simple longitudinal
susceptibility gradient. ^^ In these forms one or more posterior zooids are
more or less clearly distinguishable by their susceptibility, though not
morphologically. Two, three, or even more posterior zooids are often
present in long individuals of D. dorotocephala, and under certain condi-
tions fission of any one can be induced (Child, 1910a, 191 1 J).

In gradually lethal concentrations or intensities above the limit of tol-
erance death of the anterior zooid progresses from the head posteriorly,
but lateral margins are more susceptible than the median region to alka-
line and irritating agents (Fig. t,^,, A-D) and less susceptible to agents
which kill without stimulation or irritation (Fig. ^^, E-G). The differ-
ences between margin and median region suggest a certain degree of spe-
cific susceptibility of the margins, perhaps associated with the presence
there of numerous gland cells; these are apparently stimulated to secre-
tion by alkaline and irritating agents. Dorsal and ventral epithelium and
pharynx also show indications of specific susceptibility. To some agents
the dorsal, to others the ventral, epithelium is more susceptible, and the
pharynx is highly susceptible to certain agents but no more susceptible
than other parts at the pharyngeal level to others. In general, the poste-
rior zooid region is distinctly more susceptible than the posterior part of
the anterior zooid (Fig. ^iZ^ -^^D); but under certain conditions — for ex-
ample, in neutral or slightly acid cyanide — its relative susceptibility is
decreased (Fig. ^t,, E-G), and it may be less susceptible than the anterior
zooid. In these respects the cUfferences between anterior and posterior
zooids are very similar to those between old and young individuals.

Susceptibilities of isolated pieces to cyanide differ with length of piece
and time after section. Pieces one-third to one-fourth the total body
length of animals 1 5 mm. or more in length show little or no increase in sus-
ceptibility after section, except immediately adjoining cut surfaces. Simi-

'* Species used for observation of differential susceptibility are as follows: without pos-
terior zooids and fission: Procoiyla flimatilis, Fonticola velata, Phagocata gracilis, Ciirtisia
foremami (see p. 41, footnote 7); with posterior zooids and undergoing fission at a definite zone:
D. dorotocephala, D. agilis, D. tigrina,a.r\d a Japanese species, externally very similar to Diigesia.

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS 113

lar fractions of body length from small, young animals usually show in-
crease. With decrease in relative length of piece, increase in susceptibility
after section becomes greater. Moreover, in pieces of equal length from
the anterior zooid the increase is greater with increasing distance of the
piece from the anterior end; susceptibility of pieces from the posterior
part of the anterior zooid becomes as high as, or even higher than, that of
anterior pieces. Pieces from the posterior zooid region resemble anterior
pieces in showing less increase in susceptibility than posterior pieces of

/X

T'.

Ir-

J

\r-l^

B

YT

l-J

H

r'i i^

TT

D G J

Fig. 2,2,, A-J . — Differential susceptibility of Dngesia dorotocephala in KCN. A-D, m/i,ooo,
alkaline; E-G, ni/i,ooo, neutral or slightly acid; H-J, m/io,ooo, alkaline. Arrows indicate
direction of progress of disintegration.

the anterior zooid. In general, the effect of section on these shorter pieces
is obliteration or, in shorter pieces, inversion of the susceptibility gradient
of intact animals. The posterior region of the anterior zooid becomes
equally susceptible to, or more susceptible than, other regions. These al-
terations of susceptibility are temporary and apparently represent a stim-
ulation of the pieces following section. They gradually disappear, the
gradient in pieces becoming similar to that of intact animals after 12 hours
or more (Child, 1914c). That they indicate real, though temporary,
changes in physiological condition at different body-levels, in relation to

114 PATTERNS AND PROBLEMS OF DEVELOPMENT

the general gradient pattern, seems evident. They are also of interest be-
cause degree of inhibition of head regeneration in pieces parallels the
increases in susceptibility following section ; and if the increases are pre-
vented, head development is not inhibited (see pp. 177, 406). The changes
in rate of oxygen uptake found by Hyman are, so far as determined, es-
sentially parallel to the changes in susceptibility.

In low concentrations which kill very slowly but are still above the lim-
it of tolerance (e.g., KCN m/ 10,000), more than one posterior zooid, if
present, may become distinguishable by slight differences in susceptibility
(Fig. ^T), H, I, J). As regards the death gradient of the digestive tract,
it is difficult to attain certainty, for direct exposure, at least to chemical
agents, depends more or less on disintegration of the body wall. There is,
however, some evidence that in well-fed animals death progresses from the
pharyngeal region anteriorly and posteriorly. In agents which penetrate
readily, such as cyanide, the gut of well-fed animals disintegrates as early
as, or earlier than, the body wall; but in starved animals it is much less
susceptible and may still be intact after the body wall has disintegrated.

In the earlier stages of reconstitution the new tissue at both ends is
more susceptible than the old, and the developing head more susceptible
than the posterior tissue; in the old tissue susceptibility decreases from
the head posteriorly. When a head develops at the posterior end of a
piece or in any other than the usual position, a longer or shorter suscepti-
bility gradient arises in relation to it.

Attention must be called to the fact that the anteroposterior gradient
pattern in postembryonic stages of the planarian body does not represent
a gradient of growth in body length. The rate of growth increases poste-
riorly, as is at once evident from comparison of individuals of different
length; Abeloos (1928) has given statistical data concerning this point.
The physiological gradients are not necessarily growth gradients and may
or may not parallel such gradients (see Needham, 1931, p. 584).

Evidence of differential tolerance and differential acclimation or con-
ditioning has been obtained with D. dorotocephala . In low concentra-
tions of ethyl alcohol (1-1.5 P^'" cent) most individuals remain alive and
intact for at least a week or two and show more or less evidence of in-
creased tolerance or acclimation. During the first few days they are more
or less anesthetized, but motor activity increases after several days of ex-
posure. Small young individuals, whether from eggs or reconstituted
pieces, while more susceptible than large individuals to higher concentra-
tions of alcohol and at first less active than large in these low concentra-

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS 115

tions, acclimate more rapidly and more completely than the large ani-
mals. They become active earlier and show less differential death than
the older individuals. After a week or two disintegration usually begins
to appear in the posterior region of the anterior zooid in some of the larger
animals, that is, just anterior to the level of fission, and progresses ante-
riorly from this region (Fig. 34, A). The body gradually becomes sepa-
rated into two independent pieces, the posterior zooid region and the an-
terior part of the anterior zooid (Fig. 34, B). After progressing a greater
or less distance anteriorly in the anterior zooid, the disintegration may
cease, and healing of the wound and regenera-
tion of a new posterior end may occur slowly
in the same concentration that brought about
the disintegration (Fig. 34, C). The posterior
zooid may also remain active and slowly re-
generate a new head (Fig. 34, C). In a num-
ber of individuals of the same length in the
same container, differences in the time of disin-
tegration and the rate and amount of its prog-
ress anteriorly occur. Some individuals may
disintegrate completely (the head region and
the posterior zooid last of all) , while in others
disintegration is limited to the posterior part
of the anterior zooid, and both anterior piece
and posterior zooid or zooids regenerate and re-
,main alive indefinitely. Small, physiologically
young animals usually remain alive and intact

in concentrations in which partial death of the large, old individuals oc-
curs; in slightly higher concentrations they may show a similar partial
death. The range of concentrations giving these results differs with tem-
perature, nutritive condition, and size of animals and must be determined
experimentally for a particular stock. Susceptibility to toxic and lethal
effects of these low concentrations increases with decrease and decreases
with increase of temperature, the reverse of susceptibiHty to concentra-
tions above the limit of tolerance.

Occasionally heads only, or heads and a short anterior portion of a few
of the larger animals, may disintegrate early, and a new, more or less in-
hibited head develops slowly. In somewhat higher concentrations this
may sometimes occur in smaller animals. These cases represent the begin-
ning of the death gradient characteristic of higher concentrations. The

Fig. 34, A~C. — Dugesia doro-
tocephala, differential tolerance
and differential acclimation.

ii6 PATTERNS AND PROBLEMS OF DEVELOPMENT

physiologically old heads have a lower range of tolerance than those of
young animals and are sometimes killed by the primary toxic action of the
agent before they are able to acquire tolerance. If the heads do not die
during the first few days of exposure to the agent, they usually remain
alive indefinitely, or until starvation is far advanced, when the whole ani-
mal may die.^^

In KCN m/ioo,ooo most individuals live indefinitely; in KCN m
4/100,000 most or all die. Between these hmits, however, some degree of
differential tolerance appears in many individuals, with reversal of the
death gradient in the anterior zooid and acchmation of more or less of the
anterior region and of the posterior zooid region. Anterior ends of pieces
show greater tolerance or acchmation in regeneration than posterior ends
(Child, 1933c). In animals kept at 3°-5° C. a similar differential tolerance
appears in many individuals, with separation of the body into an anterior
piece and the posterior zooid by disintegration of the posterior part of the
anterior zooid, and later acclimation with regeneration of a posterior end in
the anterior piece and a head in the posterior piece at the same tempera-
ture.

The point of chief interest in these experiments is the reversal in direc-
tion of the death gradient and the greater tolerance to low concentrations
of those regions which are most susceptible to high concentrations of the
same agents. They also constitute further evidence concerning the char-
acter of the longitudinal physiological pattern of the planarian and its
relation to reconstitution in pieces from different levels. The evidence at
hand from oxygen uptake, CO2 production, differential dye reduction, and
differential susceptibility to lethal effects of agents suggests that the more
tolerant regions, also more capable of acclimation, are more active regions,
regions of higher metabolism. In the case of alcohol, tolerance apparently
depends on abihty to oxidize it. In animals living in low concentrations of
alcohol for several weeks without feeding, the rate of oxygen uptake in-
creases greatly, sometimes 500 per cent, though the animals are sluggish
and do not lose weight more rapidly than controls in water (Buchanan,
1922, pp. 28, 29). In differential tolerance and acclimation the regions of
greater tolerance and acclimation are apparently regions of higher rate of
oxidative metabolism. As will appear in later chapters, evidence of differ-
ential tolerance and differential acchmation or conditioning within the in-

'9 Child, 191 le, 1913a, 19146. It is perhaps unnecessary to note that in these experiments
on tolerance to alcohol the animals were kept in closed containers of large volume with small
air space, and solutions were renewed daily or every 2 days.

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS 117

dividual is to be obtained not only from postembryonic stages of planari-
ans but from early embryonic stages of various animals and in the same
relation to gradient pattern.

OTHER TURBELLARIA

Observations on the zooid chains of the rhabdocoel Stenostomum, in
which morphological development of zooids precedes separation (Fig. 10),
show that rate of methylene blue reduction and susceptibility decrease
from the head region of each zooid after an early stage of zooid develop-
ment.^''

Susceptibility, as indicated by cytolysis, and the indophenol blue reac-
tion in eggs removed from the uteri, cleavage stages, and early larvae of

A B C D

Fig. 35, A~D. — Susceptibility gradients in eggs and developmental stages of Stylockiis
ijimai. A, egg removed from uterus; B, early larva; C, D, later larval stages. The indophenol
blue gradient is similar (after Watanabe and Child, 1933).

the Japanese polyclad Stylochus ijimai show a gradient decreasing from
apical to basal pole (Fig. 35, A, B). In later larvae the apicobasal
gradient persists in the general ectoderm, but the cells of the two ciliated
bands also show a higher susceptibility and rate of indophenol reaction
than adjoining regions (Fig. 35, C), and in the most advanced larvae ex-
amined the whole posterior region shows increased susceptibility and rate
of indophenol reaction (Fig. 35, Z)). Conditions during metamorphosis are
not known. Determinations of CO2 production of pieces from different
body-levels of adults show a U-shaped gradient in most individuals (Fig.
36) . Whether the high CO2 production of the posterior region is associated
with the presence in it of the terminal genital complex is not certain. As
noted above, the posterior region shows increase in susceptibility and rate
of indophenol reaction in later larval stages, before there is any trace of the

" Child, 1 9 246, susceptibility; t 934(7, dye reduction.

ii8

PATTERNS AND PROBLEMS OF DEVELOPMENT

genital organs. Ectodermal susceptibility of adults decreases from the
posterior end anteriorly, with a slight increase at the anterior end in most
individuals (Watanabe and Child, 1933). The CO2 gradient represents, of

Fig. 36. — CO2 production of Stylochus ijimai divided into three pieces, A, B, CD; extreme
anterior and posterior ends discarded; five series of determinations, each lot consisting of three
pieces from the same body-level; ordinate, thousandths of a milligram of CO2 per gram per
minute; abscissae, body-levels. Horizontal broken lines indicate CO^ production of intact ani-
mals, upper curves, CO, production immediately after section, lower curves, COj production 2,
4, and 6 hr. after section (from Watanabe and Child, 1933).

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS 119

course, the total CO2 production of all organs at the different levels; and
the data on dye reduction in a related species, presented below, indicate
that gradients in dijfferent organs differ in direction. The ectodermal sus-
ceptibility gradient gives no information concerning internal organs.
However, the very slight variations in successive CO2 determinations on
the same lots 2, 4, and 6 hours after section indicate that they are phys-
iologically significant as indicating a definite respiratory pattern of the
whole.

The gradient of the earlier larval stages of a Califomian Leptoplana is
apicobasal, like that of corresponding stages of Stylochus. Observations on
differential reduction of methylene blue in adults of this species and of a
planocerid closely resembling Stylochus appear to throw some further light
on the question of gradient pattern in these forms. When only the body
wall is stained, rate of dye reduction decreases from the posterior end ante-
riorly and from median to lateral regions. When animals are stained
throughout, internal reduction is most rapid in the cephalic ganglia and
decreases in all directions from them, more rapidly posteriorly than anteri-
orly. In both species susceptibihty of the ectoderm and body wall, like
dye reduction, shows a gradient decreasing from the posterior end anteri-
orly. Apparently, then, there are in these forms at least two gradients — •
an ectodermal and an internal — in opposite directions. It is a point of some
interest that the cephalic ganglia reduce more rapidly than other internal
organs. When the genital complex is present, it reduces more rapidly than
adjoining regions, but not as rapidly as the ganglia. The uncertainty as
regards interpretation of respiratory gradients of adult animals with dif-
ferent, more or less localized organ systems is well illustrated by these
polyclads. The different lines of evidence suggest that the primary gradi-
ent persists internally, apparently in relation to the nervous system,
while secondary gradient changes occur in ectoderm and body wall. In
cleavage and gastrula stages of the planocerid species rate of reduction of
methylene blue in low oxygen decreases basipetally from the apical region.

ANNELIDS

The bipolar migration of alkaline and acid substances in the egg of
Nereis at the time of polar-body formation and their accumulation, re-
spectively, in apical and basal regions indicate the existence of an axial
differential of some sort (Spek, 1930, 1934&). Spek suggests that the mi-
gration is due to an electrical factor, but the existence of such a factor im-
plies a physiological differential.

I20

PATTERNS AND PROBLEMS OF DEVELOPMENT

A few data on differential susceptibility of developmental stages are
available (Child, igijd; Hyman, 1916a). Susceptibility of the egg of
Chaetopterus pergamentaceus before and after fertilization decreases from
apical to basal pole (Fig. 37, A). During maturation susceptibility of
the apical region is relatively high. The basipctal gradient persists dur-
ing cleavage and after gastrulation in the ectoderm, but in early motile
stages susceptibility increases in the dorsiposterior region, slightly anterior
to the extreme posterior end, and in the early trochophore this region be-
comes more susceptible than the apical region and extends ventrally
around the larva (Fig. 37, 5, C). This is the region of the so-called "so-
matic plate," which gives rise to the trunk segments. The region of the

A

B

Fig. 37, A~C. — Differential susceptibility of Chaetopieriis pergamentaceus egg and larvae.
A, basipetal gradient in undivided egg; B, C, stages of cytolysis in trochophore (from Child,
1917J).

developing mouth also shows increased susceptibility at this stage (Fig.
37, C). Nereis limbata and Arenicola cristata, also polychetes, show a simi-
lar simple gradient from apical pole in early stages; later, as the somatic
plate becomes active, its susceptibility increases. The cells of the Nereis
prototroch also show increased susceptibility as their jciiiary activity de-
velops. What may be called a U-shaped gradient is apparently character-
istic of annelids from, or somewhat before, the stage when development of
trunk segments begins, throughout life, or as long as segment formation
continues. In these polychetes it originates in the trochophore. The term
"U-shaped" is merely a convenient designation for a gradient pattern in
the longitudinal axis with two high ends and a low region between them;
strictly speaking, such a gradient is not U-shaped but consists of two arms
with differential in opposite directions. It is present at least temporarily in

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS 121

other segmented animals, as well as in annelids, and also in the polyclad
larva (see Fig. 35, Z)) preceding posterior elongation, though segmentation
does not occur. It is apparently associated with posterior elongation of
the body, whether by formation of segments or not. In early stages of
embryonic elongation of Tubifex tubifex, an oligochete, susceptibility
shows only the posterior arm, that is, it decreases from the posterior seg-
ment-forming region anteriorly without evidence of higher susceptibility
in the head region (Fig. 38, A, B), according to Hyman (1916a). The ac-
tivity of the rapidly developing segments appears to dominate the whole
embryo. Somewhat later, however, when segments between the head and
the posterior growing region have undergone further development, the
U-shaped gradient characteristic of later, and probably also of earlier,
stages appears (Fig. 38, C).

A B

Fig. 38, A-C. — Differential susceptibility of embryonic stages of Tubifex tubifex. A, B,
earlier, and C, later, stages (from Hyman, 1916a).

Respiration in relation to body-level has been determined in several
annelid species, both polychetes and oligochetes. Considering the data
with some departures from order of appearance, Hyman and Galigher
(192 1), using the Winkler method, found a U-shaped gradient of oxygen
uptake, with the posterior end usually higher than the anterior, in Nereis
virens, N. vexillosa, and Lumbriculus inconstans. Shearer (1924), using the
Haldane respirometer, found the oxygen uptake of anterior pieces of an
earthworm (species not given) at least twice as high as that of posterior
pieces and a similar difference in the oxygen uptake of acetone powders of
similar pieces. Since he used only anterior and posterior pieces, his data
do not show whether a U-shaped gradient was present in the species used.
In a later paper (Shearer, 1930) he concludes that his earlier data were in-
correct but presents no new determinations on the earthworm. Perkins
(1929) found a slight U-shaped gradient of oxygen uptake in earthworms
{Lumhrlcus sp.; Allolobophora sp.) but noted that it did not correspond
with the gradients of total iodine equivalence, extractable sulphydryl, and

cu

<5^

122

PATTERNS AND PROBLEMS OF DEVELOPMENT

total sulphur. He suggested, however, that the oxygen-uptake gradient
doubtless includes other oxidation systems than those associated with
growth. Okada (1929) and Kawaguti (1934) found a U-shaped gradient
of oxygen uptake and CO2 production in the microdrilous oligochete
Branchiura. Determinations of CO2 production, of oxidizable substance
by a modification of the Manoilov method, and of the temperature of on-
set of heat-shortening in the earthworms Pheretima communissima and
Allolohophora (Watanabe, 1931) give very similar U-shaped gradients for
both species after increase of CO, production following section has disap-

- -0.005

Fjq ^g — Gradient of oxidizable substance (heavy line), of CO2 production (medium line),
and of heat-shortening temperature (light line) in Allolohophora foetida. CC, cubic centimeters
of KMnO,, required for oxidation; C, centigrade temperature; Mgm, milligrams of CO2 per
gram per minute; abscissae, body-levels; A, anterior (from Watanabe, 1928).

peared (Fig. 39). Except for Shearer's conclusion, not based on evidence,
that his experimental data were incorrect, there is essential agreement
among all these investigators.

Determinations of CO2 production on N. virens (Parker, 1929) give very
different results. He finds rate of CO2 production of anterior and posterior
pieces lower than that of middle pieces. Examination of his data, how-
ever, shows that rates of the pieces are from 1 2 to 30 per cent lower than
that of intact animals and that this difference is due entirely to decrease
in the rates of anterior and posterior pieces after section. He finds ante-
rior and posterior regions more susceptible than the middle region to cy-
anide, and his data on CO2 production suggest that they are also more sus-

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS 123

ceptible to loss of blood and experimental conditions. Since his determina-
tions were made "a few hours" after section and were not repeated on the
same lots at different periods, they are without definite significance in rela-
tion to the gradient problem, except in so far as they indicate that anterior
and posterior regions have undergone depression since section.

These results with Nereis and the views advanced in connection with
them led Hyman (1932a) to undertake a re-examination of oxygen up-
take in N. virens with the Winkler method. She finds that during the first
3 hours after section the rates are irregular but that from 3 to 9 hours after
section the U-shaped gradient is characteristic. Pieces from the posterior
region undergo rapid depression in physiological condition after section;
consequently, she regards her data as showing a much lower rate in these
pieces than the normal. Electric-potential differences and galvanotactic
reaction in Nereis are also dependent on freshness of the material (Hyman
and Bellamy, 1922).

Using extremely short pieces in order to eliminate motor activity, Mal-
oeuf (1936) maintains that there are no significant diiTerences in oxygen
uptake at different levels of an earthworm, the differences observed by
others being attributed to differences in motor activity, but presents no
actual evidence in support of his opinion. This work, like that of Parker,
provides another illustration of the difficulties involved in determining the
significance of respiratory determinations in isolated pieces of animals.
The effect on respiration of loss of blood and of the large area of cut sur-
face in relation to volume is entirely unknown. Maloeuf also finds that
total solids, total organic material minus fat, and fat of the body wall and
of the whole starved animal, in percentages of wet weight, increase poste-
riorly in the preclitellar region and decrease from the clitellum posteriorly,
except for an increase in fat at the posterior end. It is of some interest to
note that these gradients are essentially the inverse of susceptibility and
dye-reduction gradients and of the respiratory gradients found by most
investigators. That such gradients are present without any corresponding
differences in respiratory rate seems rather remarkable. The anterior re-
gions contain less reduced glutathione than other regions, in which the
differences are negligible.

It is sufficiently evident from these data on annelid respiration that a
single set of determinations on pieces after section does not afford a basis
for any definite conclusions, except that further investigation is necessary.
Repeated determinations on the same lots at different times after section
and determinations on pieces representing different fractions of the body

124

PATTERNS AND PROBLEMS OF DEVELOPMENT

length must be made; effects of conditions favoring or stimulating motor
activity and those favoring quiescence must be evaluated as far as possi-
ble before reasonable certainty can be attained as to the presence or ab-
sence and significance of a respiratory gradient independent of motor ac-
tivity and other incidental factors. At present most of the evidence indi-
cates presence of a U-shaped

_\ L respiratory gradient and is in

agreement with results of dye
reduction, susceptibiHty, elec-
tric-potential difference, and
galvanotactic reactions, some
of which are obtained in com-
plete absence of motor ac-
tivity.

\

\

1

Observations on differential
dye reduction in several micro-
drilous oligochete species and
several earthworms show a
U-shaped reduction gradient
with posterior arm much long-
er than anterior, except in cer-
tain forms after segment de-
velopment is completed and
the posterior growing region
disappears. In zooid chains of
Nais and Aeolosoma each new
zooid develops a reduction
gradient with rate decreasing
posteriorly from the head re-
gion and anteriorly from the
posterior-segment-forming re-
gion, when that is present and
active. In Aeolosoma the posterior arm of the gradient is present only
temporarily, since the posterior growing region disappears after a certain
number of segments is formed, and the gradient in the fully developed in-
dividual becomes a simple gradient, decreasing from the head region to
the posterior end (Child, 1934a; Child and Rulon, 1936).

Stages of differential dye reduction in Tuhifex are indicated diagram-
matically in Figure 40. Reduction is most rapid anteriorly and posterior-

FiG. 40. — Diagrammatic, indicating successive
stages of differential reduction of dyes in body wall
of Tubifex tuhifex when not appreciably injured by
the dye. Ordinates indicate degree of reduction ; hor-
izontal base line, no visible reduction; anterior end
at left; diagram at bottom indicating reduction in
animal with well-developed clitellum, C (from Child
and Rulon, 1936).

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS 125

Fig. 41. — Stages of differential dye reduction in two-
zooid chains of A eolosoma, Nais, and Stylaria, indicated
as in Fig. 41 ; fission zone at F (from Child and Rulon,
1936).

ly, except that the anal segment, indicated by the short heavy hne on the
base line at the right, reduces slowly (Fig. 40, top). The anal segment, of
course, is not a part of the posterior growing region but is the original pos-
terior end of the body. It is interesting to find it so sharply marked off
from the growing region immediately anterior to it.

A stage of reduction in animals with well-developed clitellum (C) is in-
dicated in the diagram at the bottom of Figure 40. The clitellum is sharp-
ly distinguishable from ad-
joining regions by its much
higher rate of reduction.
With overstaining, differen-
tial susceptibility to the dye
results in earlier injury to
anterior and posterior re-
gions than to other parts,
with U-shaped gradient. Re-
duction is differentially re-
tarded in the injured regions, with the result that the normal reduction
gradient undergoes complete inversion, the middle region reducing most
rapidly, though no more rapidly than in normal animals, and rate of re-
duction decreasing both anteriorly and posteriorly.

The normal U-shaped gradient appears in actively motile animals and
also in animals with motility practically completely eliminated by anes-
thesia, though possibly slightly less steep in the anesthetized animals.
Motor activity of animals isolated for a week or two in clear water in day-
light decreases greatly. These animals often show no visible movement
when observed continuously during the period of dye reduction, but the
reduction gradient is essentially the same as in actively motile animals.
Evidently differential dye reduction does not result from differential
motor activity. Differential dye reduction in two-zooid chains of A eolo-
soma, Nais, and Stylaria is indicated in Figure 41, the fission plane being
at F. This gradient may also be completely inverted by overstaining and
differential injury of anterior and posterior regions of each zooid. Motor
activity is certainly not greater in the region of fission than in adjoining
regions; it is probably less, for the posterior region of the anterior zooid
and the head of the second zooid are not fully developed.

In several unidentified species of earthworms a similar two-armed re-
duction gradient is present. In one form, however, without indication of
loss of posterior segments, rate of reduction decreased from anterior to

126

PATTERNS AND PROBLEMS OF DEVELOPMENT

posterior end. This was perhaps a species in which new segments are not
formed after hatching (Sun and Pratt, 1931). In all forms tested for dye
reduction a distinct ventrodorsal reduction gradient appears, rate of re-
duction decreasing from the mid- ventral region laterally and dorsally. In
the more or less transparent microdrilous oligochetes the segmental gan-
glia are visible through the body wall in the living individual. With sufh-

If^"

70 90 110 150

B

A

Fig. 42, v4-C.— Susceptibility gradients of microdrilous oligochetes. A, Litmbriculiis in-
constans; B, Tubifex rivitlorum; C, Dero limosa. In B and C ordinates indicate minutes; ab-
scissae, body-levels as segment number from anterior end at left (from Hyman, 1916a).

cient exposure to the dye they become stained, and repeated observations
seem to show a higher rate of reduction in them than in the ventral body
wall. However, since they are seen through the stained body wall, the
possibility of error exists.

Susceptibility gradients observed by Hyman (1916a) in microdrilous
oligochetes parallel the reduction gradients. Except in the fully developed
Aeolosoma individual, the forms examined show a U-shaped death gradi-
ent. Different species differ as regards relative length and height of an-
terior and posterior arms of this gradient (Fig. ^2,A,B,C). The anal seg-

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS 127

ment of later developmental stages is sharply marked off by a much lower
susceptibility from the segment-forming region immediately anterior to
it, except in individuals in which a new posterior end has recently de-
veloped in connection with fission (Fig. 42, C). In this respect, also, dif-
ferential susceptibility parallels differential dye reduction. Inversion of
susceptibility gradients has been observed in Dero and some other micro-
drilous species kept in clear standing water in diffuse light for a week or
two. Susceptibility to cyanide is decreased but increases from the two
ends toward a region anterior to the middle. Evidently the animals are
differentially susceptible to the unfavorable environment, and the differ-
ential effect on physiological condition inverts differential susceptibility to
cyanide.

The electric-potential gradient of earthworms tested also shows two
opposed arms, at least on the ventral side, both ends being galvanometri-
cally negative to an intermediate region (Morgan and Dimon, 1904;
Watanabe, 1928). The galvanotactic reaction of the earthworm and of
several species of polychetes is also significant in this connection. When
exposed to the current, the animals bend into a U-shape with the two ends
toward the cathode. ^^ If the current is strong or if the animals are exposed
for a long time, a rather sudden reversal has been observed in a number of
species, both ends turning toward the anode. If these animals are removed
from the current for 15-30 minutes, they recover, and both ends are again
cathodic on renewed exposure.

Water content, distribution of setae, and pigmentation show a U-shaped
distribution in various species of earthworms, according to several au-
thors; but an increase in water content from anterior to posterior levels
has been reported in two species." The difference in temperature at which
heat-shortening occurs at different body-levels has been determined for
several species of annelids. The differences are characteristic for the spe-
cies, showing in some a simple, in others a U-shaped, in still others an in-
verted U-shaped, gradient and in some a more varied differential. In the
earthworms tested, these temperature gradients are U-shaped and corre-
spond closely to the other gradient expressions. Their relations to other
conditions at different body-levels in polychetes are not known ; they may
perhaps correspond to regional differentiations (Hatai, 1924a; Watanabe,
1928).

Few observations have been made on susceptibility gradients in adult

" Moore and Kellogg, 1918; Hyman and Bellamy, 1922; Moore, 1923.

"Hatai, 1924a; Kopenhaver, 1937, water content. Hatai, 19246; Sivickis, 1930, setae.
Pickford, 1930, pigmentation.

128 PATTERNS AND PROBLEMS OF DEVELOPMENT

polychetes. In unidentified syllids individuals developing from buds
showed the U-shaped gradient after the posterior segment-forming region
was established. The two ends of N. virens are more susceptible to cyanide
than the middle.^'^

If it be granted that a U-shaped gradient involving metabolism is pres-
ent in the body walls of adults of a considerable number of annelid species
and that the posterior arm of the gradient is related to the progressive
formation and development of new segments, there is every reason to be-
lieve, as was noted years ago by Hyman (1916a), that the two arms of the
gradient are not identical in kind of metabolism. The anterior arm appar-
ently represents the final expression of the primary embryonic polar gradi-
ent in this fully differentiated region; the posterior arm represents differ-
ent stages of growth and differentiation of segments, the earhest stages
being anterior to the anal segment, with stages of development and physi-
ological age progressively more advanced from segment to segment ante-
riorly as long as segment formation and development continue. It would
be most surprising if there were not a metabolic gradient in this region
during the period of segment formation and development. But in those
species in which segment formation ceases, either at hatching or at some
later stage, and all segments sooner or later attain full development, this
posterior arm of the gradient may disappear completely, and only a sim-
ple gradient from anterior to posterior end may be present ; or, with special
differentiations and functions of particular regions other gradient modifi-
cations may appear. In Tubifex, for example, development of the clitellum
alters the gradient of the body wall in the region where it appears. Deter-
minations of oxygen uptake, CO, production, differential dye reduction,
and differential susceptibility show only certain quantitative factors and
give no information concerning other axial differences, which may be and
doubtless are, present in the adult. That the metabolism indicated by
oxygen consumed or CO2 produced is the same in character in anterior and
posterior regions of an annelid body does not follow from the data and is
not assumed; but that there are differences in basal rate, that the rate de-

'i Parker, 1929. This author concluded that the higher susceptibility of anterior and pos-
terior regions resulted from entrance of cyanide through the mouth and anus. This is
certainly not the case in other forms: death and disintegration begin on the outer surface of
the body; the posterior arm of the susceptibility gradient of Tubifex and other microdrilous
oligochetes extends over more than half the body length; that cyanide or other agents should
be carried so far anteriorly in the intestine is extremely improbable. Moreover, differential
death of the body wall may occur in vital dyes without staining of the intestine. Also, differ-
ential susceptibility to lack of oxygen is the same as to cyanide.

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS 129

creases from the head posteriorly for a certain distance, and that the earh-
er stages of segment development in the posterior region have a higher
rate than later stages at more anterior levels is indicated by a large body of
evidence, even though there is not complete agreement.

EMBRYONIC AND LARVAL DEVELOPMENT OF ECHINODERMS

Embryonic development of echinoderms is one of the classic materials
of developmental physiology. It has been the object of a great number of
experimental studies dealing with many special problems, but all more or
less directly concerned in one way or another with the problem of develop-
mental pattern. Among the many investigations concerned with this
material, the concept of a gradient or gradients of some kind as essential
factors of axiate pattern has appeared in various forms. Probably the
first suggestion that the polar axis is represented by a gradient was ad-
vanced by Boveri (1901). The earlier experimental studies of Runn-
strom (1914, 191 5) on echinoderm development led him to the view that
the axiate pattern of sea-urchin development is based on systems of layers
or concentration gradients of substances. Further development of this
hypothesis by Runnstrom and his co-workers, particularly Horstadius
and Lindahl, will be discussed as occasion arises. A hypothesis of gradient
pattern in sea-urchin development in terms of the colloid substrate has
been advanced by von Ubisch (1936, 1938). Observations on differential
susceptibility, as indicated both by differential death and by differential
modification of development, on differential reduction of potassium per-
manganate and later of dyes and in the indophenol blue reaction, in other
forms as well as in echinoderms, led Child to a hypothesis of axiate pattern
in terms of gradients in which differences in rate of metabolism were re-
garded as essential factors and probably the most important in early de-
velopment.^^

In this chapter only some of the more direct evidence on which these
hypotheses are based will be dealt with.-^

'•t Child, 1913a, d; 1914/; 1915a; 1916a, d; i928(f; 1936a, b.

^5 In order to avoid possible confusion, certain axial relations in echinoderm development
and the terms used in describing them are noted. The terms "animal" and "vegetal" or "vege-
tative," dating from the earlier days of embryology but still widely used for eggs and early
embryonic stages generally, have reference to the more or less marked segregation of proto-
plasm and yolk and the difference in developmental activity in opposite polar regions of the
eggs of many animals. The more active protoplasmic pole where polar bodies are usually
formed is the animal pole. The application of these terms to echinoderm development in
which polar differences are slight in early stages is largely conventional. In the following pages

I30 PATTERNS AND PROBLEMS OF DEVELOPMENT

OOCYTE AND UNDIVIDED EGG

Differential dye reduction in early, growing oocytes of echinoids"^ seems
to indicate a very slight decrease in rate of reduction from the region where
the nucleus is nearest the cell surface. This is usually, but by no means al-
ways, nearer the free than the attached pole. In full-grown oocytes ap-
proaching maturation, with nucleus no longer visible and retaining evi-
dence of position of stalk of attachment, a slight reduction gradient with
highest rate at the pole opposite the region of attachment seems to be
present: maturation stages have not been obtained. In a few unfertihzed
eggs after maturation a slight gradient, with most rapid reduction at the
egg surface in the region opposite the pole of attachment, or in some eggs
in the region where the nucleus is near the surface, has been observed; but
whether the nucleus is still at the apical pole or elsewhere is not known.
In all stages preceding fertilization these reduction gradients are so slight
that their presence becomes reasonably certain only with repeated ob-
servations. Only uncertain indications of a possible differential suscepti-
biUty have been observed in echinoid stages preceding fertilization (Child,
1916a). A higher susceptibility to cyanide of the ventral side in the undi-
vided egg of the sea urchin Paracentrotus lividus has been reported (Foer-
ster and Orstrom, 1933).

The nucleus of the full-grown asteroid oocyte usually lies near the cell
surface at some point, more often nearer the free than the attached pole.
With slight staining, reduction apparently decreases in rate from that
part of the cortical region where the nucleus is nearest the surface; but
here, as in the echinoids, the gradient is at best very sUght. According to

the terms "apical" and "basal" are often used in place of them, as being more convenient.
Gradients observed are often designated as "basipetal" or "acropetal," meaning that the
progress of the change indicating the gradient is basipetal, toward the vegetal or basal pole, or
acropetal, toward the animal or apical pole. In echinoids the apical polar region becomes the
oral lobe of the pluteus larva (see Fig. 73), the basal region, mesenchyme, and entoderm; and
the most basal region of the ectoderm becomes the anal side of the pluteus. The apical polar
region of the asteroid starfish egg becomes the apical and functionally anterior end of the
larva (see Fig. 83). The so-called "ventrodorsal axis" of the echinoid and asteroid larva is
more or less nearly at right angles to the apicobasal or polar axis; that is, one side of the em-
bryo becomes conventionally ventral, the opposite side dorsal. The ventral side of the echinoid
pluteus, the side including the oral lobe and the anal arms (see Fig. 73), is functionally anterior
and has sometimes been called "anterior" (Child, 1915a, 1916a, d), sometimes the "oral side"
or "field"; the opposite pointed or rounded end of the pluteus, "posterior" or "aboral." In the
asteroid larva, however, the ventral side is oral, and the anus comes to lie more or less ven-
trally.

^f- Material, Strongylocentrolus purpuraliis, Deudraster excentricus; dyes, methylene blue and
Janus green; Child, 1936a. In this paper dye concentrations and staining periods are given
for all figures.

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS 131

E. B. Wilson and Mathews (1895), that region of the oocyte where the
nucleus is nearest the surface becomes the apical or animal pole in Asterias
forbesii, and the polar bodies form there. Polar bodies in this species are
said by Yatsu (1910a) always to appear halfway between equator and
free pole. In Patiria miniata, another asteroid genus, their position, like
the position of the oocyte nucleus, is highly variable, ranging fron the free
pole to a position between equator and attached pole (Child, 1936a, Figs.
12-14). According to Schaxel (1914; 1915, pp- 35, 36) the attached pole
becomes the apical pole in Asterias glacialis; this is not the usual relation
of egg polarity to ovarian attachment in echinoderms or other inverte-
brates.

When maturation of the Patiria egg begins, a very distinct cortical dye-
reduction gradient appears, with most rapid reduction in the region of
polar-body formation (Fig. 43, A). When the cell is stained throughout,
reduction progresses basipetally in the interior as well as in the cortex
(Fig. 43, B). The region about the vegetal pole reduces last, but little or
no gradient is visible there. The susceptibility gradient with cyanide in
the ovarian oocyte oi A. forbesii (Fig. 43, C, D) progresses from the region
where the nucleus is nearest the surface through the whole cell (Child,
1915a). The oocyte nucleus, whether in the cell or isolated, also shows a
susceptibility gradient the same as that of the cytoplasm in nuclei not
isolated (Fig. 43, £). Whether this actually represents a nuclear gradient
or results from differential exposure of the nucleus as cytoplasmic cytol-
ysis occurs is not certain, but the differential susceptibility of isolated
nuclei suggests a real nuclear gradient. Conceivably a nuclear gradient
might be determined by the differential of its position near the surface of
the cell.

From experiments with hypertonic and hypotonic solutions Dalcq
(1925) infers presence of a gradient of electrical charge on the plasma
membrane of the egg of A. glacialis with basipetal increase of positivity.
Under natural conditions the egg of A. forbesii does not show the migration
of alkaline and acid substances to opposite polar regions observed in some
other eggs, probably because of high viscosity of the cytoplasm, but under
conditions which decrease viscosity the bipolar migration occurs (Spek,
1934c). The migration is apparently in relation to the axis of attachment
of the egg in the ovary; but, if earlier investigators are correct, this is not
the final polarity of the egg and embryo. Spek's description and figures
indicate accumulation of acid substances at the pole of attachment, alka-
line substances at the free pole.

A

D

Fig. 43, ^-£.— Differential dye reduction and differential susceptibility to cyanide in aster-
oid oocytes; arrows indicate direction of progress of reduction and cytolysis. A, B, maturation
stages, reduction progressing from apical pole {Paliria miniata; from Child, 1936a); C, /?,
differential cytolysis in ovarian oocyte of Asterias forbesii; E, differential cytolysis of both
nucleus and cytoplasm in Asterias oocyte (from Child, 19130).

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS 133

CLEAVAGE STAGES AND EARLY BLASTULA

In early cleavage stages of Strongylocentrotus purpuratus and S.francis-
canns rate of reduction decreases basipetally, but the differential is slight;

A

B

C D

Fig. 44, A-D. — Dye-reduction gradient in early cleavage stages of Strongylocentrotus
franciscanus; arrows indicate direction of progress of reduction (from Child, 1936(7).

A B

Fig. 45, A, B. — Dye-reduction gradient in early echinoid blastulae; optical sections, A,
Strongylocentrotus purpitratiis; B, Dendraster excentriciis (from Child, 1936a).

no evidence of more rapid reduction in the micromeres at the basal pole,-
either during their formation or in following cleavage stages, has been ob-
tained (Fig. 44). Results of a few observations on susceptibility of these

134 PATTERNS AND PROBLEMS OF DEVELOPMENT

stages are uncertain as regards presence of a differential. In the early
echinoid blastula the basipetal reduction gradient becomes more distinct
(Fig. 45). At all levels reduction is most rapid in the inner ends of the
cells and progresses outward; this perhaps indicates merely that oxygen
tension in the blastocoel falls below that in the external medium.

Early cleavage and early blastula stages of the asteroid Patiria show
a more strongly marked basipetal reduction gradient than the echinoids,
that is, the difference in rate of reduction at the two poles is greater. The
susceptibility gradient is also distinct in the early blastula and basipetal,
like the reduction gradient (Child, igi$a, 1916a).

THE LATER BLASTULA

In later blastula stages a change in the reduction picture occurs. A sec-
ond reduction gradient appears in the basal region, extends acropetally,
and partly obliterates the primary gradient. In the echinoids the center
of this secondary gradient system is the basal pole where the mesenchyme
cells, products of the micromeres, lie; but the adjoining entoderm reduces
almost as rapidly (Fig. 46). As immigration of the mesenchyme cells be-
gins, they become the most rapidly reducing cells of the whole individual,
and the prospective entoderm is only slightly less rapid (Fig. 46, B, E).

Formation of primary mesenchyme preceding gastrulation does not
normally occur in asteroids, but a secondary gradient develops in the
basal region of the later Patiria blastula, though usually not as strongly
marked and not extending as far acropetally as in the echinoids; earlier
reduction on the presumably ventral side is also evident (Fig. 47).

The earlier observations on differential susceptibility of the later blas-
tula stages in the sea urchin Arhacia punctulata and the starfish Asterias
forbesii agree with the data on dye reduction in that they show a decrease
in susceptibility from the apical pole basipetally and from one side, pre-
sumably ventral, according to evidence from gastrula stages; but they do
not show increase in susceptibility of mesenchyme and prospective ento-
derm (Child, 1915a, 1916a). More recent data on differential susceptibil-
ity to cyanide, lack of oxygen, and the dyes methylene blue and Janus
green do show this increase in Strongylocentrotiis; but the basal gradient is
more narrowly limited to the mid-basal region than is the dye-reduction
gradient, as comparison of Figure 48 with Figure 46 will show (see Appen-
dix IV, p. 739). The presumably ventrodorsal differential is evident in
Figure 48. In the asteroid basal increase in susceptibility does not appear
until later.

B

Fig. 46, ^-£.— Differential dye reduction in late echinoid blastulae. A, B, Strongylo-
centrotus franciscanus, with numerals, 1-4, indicating time-order of reduction of different re-
gions in B; C-E, S. purpuratus, preceding and during immigration of mesenchyme. Arrows
indicate direction of progress of reduction, and an attempt is made to indicate by their lengths
the relative rate of spread of reduction from the center of earUest appearance (from Child,
1936a).

136

PATTERNS AND PROBLEMS OF DEVELOPMENT

The change in condition of the basal region preceding gastrulation, as
indicated by dye reduction, suggests that it is undergoing activation; the
changes in its developmental behavior — immigration of the mesenchyme,
invagination of the entoderm — also suggest that a change of some sort has

Fig. 47. — Differential dye reduction in late blastulae of Patiria; reduction more rapid on
one side, the left in the figure, presumably the ventral side (from Child, 19360).

A

B

Fig. 48, .4, B. — Differential susceptibility of late Strongylocentrotus blastulae to lack of
oxygen and to cyanide. A, before immigration of mesenchyme only mesenchyme cytolyzed
in basal region; B, after immigration mesenchyme and middle entodermal region cytolyzed;
ventral(?) side more susceptible than dorsal (from Child, 1936a).

occurred. If the change is an activation, as dye reduction suggests, it is
not necessary to assume that the two opposed gradient systems represent
identical activities. Gradients involving different metabolic reactions and
different protoplasmic substances may be similarly indicated by reduction
or susceptibility differentials.

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS 137

THE GASTRULA AND LATER STAGES

The dye-reduction gradients in early echinoid gastrulae are essentially
similar to those of late blastulae. The mesenchyme in the blastocoel re-
duces earlier than any other cells. The entoderm follows closely; and as it
elongates inward, it develops a distinct basipetal gradient, that is, the
radial differential present before invagination becomes longitudinal with
invagination, as in a bud. In the ectoderm rate of reduction decreases
from both apical and basal poles and from the ventral side dorsally (Fig.
49, .4). Later the beginning of development of the stomodeum and of the
oral lobe from the apical region is accompanied by increased rate of reduc-
tion in those regions, and rate of reduction ventrally is distinctly higher
than dorsally (Fig. 49, B).

As the pluteus form begins to develop, the oral lobe and the region of
each anal arm show reduction, decreasing in rate from the tips. Elonga-
tion of the dorsal region is also accompanied by a slight reduction gradi-
ent, with high end at the tip. With completion of pluteus development
the reduction gradients become less distinct : changes associated with met-
amorphosis have not been observed.

In the early gastrula of Patina the invaginating entoderm does not
show a high rate of reduction at its tip, but entodermal rate is highest at
the blastopore, decreasing in both directions; in the ectoderm, rate de-
creases from apical and basal regions (Fig. 49, C). As enlargement and
decrease in thickness of the inner free end of the entoderm begins, a new
reduction gradient with high end at the tip appears, that is, the basipetal
gradient appears here, but later than in the echinoid (Fig. 49, D). That
the difference is in some way associated with formation and isolation from
the epithelial layer of mesenchyme preceding gastrulation in the echinoid
and at a considerably later stage (after invagination in the starfish) seems
probable.

As decrease in thickness of the apical end of the entoderm progresses,
with separation of mesenchyme cells into the blastocoel and development
of the esophagus, the basipetal gradient becomes still more distinct (Fig.
49, E). The region about the blastopore, however, still shows early reduc-
tion, perhaps associated with contractile activity there, which is evident
at this stage in closure and opening of the blastopore, as if a sphincter
were present. The beginning stomodeal invagination is represented by a
locaHzed radial gradient system or held of rapid reduction, and the whole
ventral side of the larva reduces more rapidly than the dorsal (Fig. 49, E).
As the ciliated bands develop, they reduce somewhat earlier than the

Fig. 49, A-E. — Differential dye reduction in echinoderm gastrulae. Arrows show direction
of progress of reduction, their relative lengths, relative rates of progress; numerals i-j, adjoin-
ing certain regions, indicate the general time-order of reduction, with / as most rapid; two
numerals (/, 2; j, 4) adjoining the same region indicate individual variations in order. A, B,
Strongylocentrolus pjtrpuralus {S. franciscanus and Dendraster excentriciis similar); C-E,
Paliria {A, C-E, from Child, 1936a).

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS 139

general ectoderm about them. As the two coelom sacs develop from the
thin-walled apical archenteron, they seem to reduce slightly more rapidly
than the adjoining entoderm; but this is somewhat uncertain because the
whole apical archenteron at these stages reduces very rapidly, often show-
ing reduction while staining in open solution, probably because oxygen
content in that part of the enteric cavity is below the critical level. Here,
as in the echinoids, the gradients associated with development of organs
become less distinct or disappear with termination of larval development.
Apical and ventral regions usually still reduce somewhat more rapidly
than the dorsal ectoderm. Some changes in entodermal reduction occur,
probably associated with regional differentiation. Reduction during met-
amorphosis has not been observed.

It seems evident that these reduction gradients and their changes indi-
cate some of the quantitative features of echinoderm developmental pat-
tern. The changes in the gradient systems in the course of development
and their close association with development of morphological form are
of particular interest. However, they give no direct evidence concerning
absence or presence of regional specificities or qualities, though it seems
highly probable that there is a relation of some sort between them and
regional differentiation. In general, the gradients appear to precede any-
thing clearly distinguishable as differentiation.

The earlier observations on susceptibility in A rhacia and A sterias gas-
trulae agree with these more recent data on dye reduction, so far as ecto-
derm is concerned. Susceptibility was found to decrease from the apical
region of the A sterias entoderm of middle and later gastrula stages; and
later, unpublished observations showed high susceptibility about the
A sterias blastopore. Developing oral lobe and anal arms of Arbacia
showed basipetal susceptibility gradients.

According to a recent study of dye reduction in development of two
other species of sea urchins, dye reduction progresses from the apical pole
basipetally throughout larval development, with no indication of a sec-
ondary acropetal gradient in late blastula and gastrula. It is also described
as progressing from the outer cell surfaces inward, instead of from within,
outward, as observed by Child (Ranzi e Falkenheim, 1937). These au-
thors find no evidence of an indophenol blue gradient, as might have been
expected, for concentration of the reagent used was so high that it cer-
tainly killed at once, and dead embryos and larvae show no indophenol
gradients or only the merest traces of them. Only use of the reagents in
very high dilutions will show the gradients in living animals. Also, no

I40 PATTERNS AND PROBLEMS OF DEVELOPMENT

gradient was observed with the nitroprusside reaction for gkitathione.
Although a basipetal reduction gradient appears in their material, their
conclusion is that there is no actual gradient but that the basal region
produces a substance inhibiting metabolism and raising redox potential;
this decreases in concentration acropetally and so gives the appearance
of a basipetal gradient. This substance and its effects are entirely hypo-
thetical, no evidence of its existence being given. They, of course, do not
accept the interpretations of the Runnstrom school or the conclusions of
Spek.

OTHER DATA AND HYPOTHESES

The indophenol blue reaction with extremely dilute reagents, which
permit appearance of intracellular indophenol blue in the living animals,
gives a very distinct basipetal color gradient in blastulae and gastrulae of
Asterias, supposedly indicating a gradient of active indophenol oxidase.
Unfortunately, attention was not given to the question of presence or
absence of a ventrodorsal gradient (Child, 1915a). Susceptibility of the
sand dollar Echinarachnius parma to a number of agents is like that of
Arbacia (J. W. MacArthur, 192 1), and dye-reduction gradients of another
sand dollar, Dendraster excentriciis , are like those of Strongylocentrotus but
more distinct (Child, 1936a). The question whether the change in condi-
tion indicated by dye reduction in Strongylocentrotus and Dendraster oc-
curs later in Arbacia and Echinarachnius remains open for further investi-
gation.

In a long series of papers from 1914 on, Runnstrom has advanced a
hypothesis of gradient pattern of sea-urchin egg and embryo in terms of
two opposed overlapping gradients in the polar axis. These are regarded
as concentration gradients of specifically different substances, the "ani-
mal" gradient decreasing in concentration basipetally, the "vegetal" gra-
dient acropetally. Ventrodorsality is considered to be represented by an-
other concentration gradient approximately at right angles to the polar
gradients; and the lateral asymmetry of the larva, by still another. Since
these are regarded as coexisting and interpenetrating, each must differ
specifically from the others. His co-workers, particularly Horstadius and
Lindahl, have applied this hypothesis in interpretation of their experi-
ments on echinoderm development.^^

Using potassium-free sea water with several species of sea urchins,
Runnstrom (1925a) finds a differential susceptibihty to lack of potassium

2' See Runnstrom, Horstadius, and Lindahl in Bibliography.

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS 141

in the apicobasal and ventrodorsal axes, basal and ventral regions being
most susceptible. According to his conclusions, water content and per-
meability to salts are highest in basal and ventral regions, and absence of
potassium brings about loss of water, first from apical and dorsal regions
of lower water content to the basal region and then from the whole egg
or embryo; decrease in dispersion of colloids is also believed to occur. The
differential susceptibility is regarded as resulting from differential per-
meability to potassium ions, decreasing from basal and ventral regions
acropetally and dorsally (see Appendix V, p. 742). According to Runn-
strom, the basal region is also more susceptible than the apical to high
concentrations of sodium thiocyanate, nile blue sulphate, and lithium
salts; development in a CO-O2 atmosphere in light shows injury in both
apical and basal regions, and evidence of a ventrodorsal and left-right
gradient is also found.^^

Most of Runnstrom's data on susceptibility are concerned with modi-
fications of development rather than with cytolytic or death gradients.
Differential modifications of development are considered in the following
chapter, but his views require some further consideration here in connec-
tion with the more direct evidence concerning gradient pattern. He ap-
parently regards the effects of external agents as more or less regionally
specific, some acting on the animal, others on the vegetal gradient, and
so on. It does not appear, however, that he has considered the possibility
of differential tolerance to external agents and of differential recovery
after temporary exposure. As will appear more clearly in the next chap-
ter, these reactions of the organism may bring about differential modifi-
cations of form and proportion opposite in direction, as regards axial rela-
tions, to those resulting from direct action of the agent. Nor does the
possibility that the gradient pattern or parts of it may undergo change
during development seem to have been considered. There is often little
evidence in his data of use of a wide range of concentrations and exposure
periods. Whether, or to what extent, some of the modifications described
represent differential inhibitions, differential tolerance, differential recov-
ery, or regionally specific effects seems, at present, uncertain.

Differential cytolysis and death and differential dye reduction give no
evidence of overlapping, specifically different gradients in the polar axis.
They indicate a single polar gradient, the "animal" gradient in early
stages, and, according to the recent studies of Child (1936a), a second
"vegetal" gradient appearing later and partially obliterating or replacing

■'^ Runnstrom, 1928a, b; 1929^; 1933; 1935a-

142 PATTERNS AND PROBLEMS OF DEVELOPMENT

the original gradient. But, as noted in these studies, differential dye re-
duction gives no information as to presence or absence of regional specific
differences. Oxidation-reduction reactions may differ as regards reacting
substances, the reactions themselves, and the products formed but may
still show the same rates or similar gradients of dye reduction. So far as
the results of differential dye reduction and differential susceptibility are
concerned, the specific overlapping gradients postulated by Runnstrom
may or may not be present. On the other hand, concentration gradients
of substance can accomplish nothing in the way of development without
metabolism. Moreover, two opposed and overlapping substance gradients
may be associated with a single gradient of rate or intensity. The over-
lapping polar gradients of yolk and protoplasm in many eggs and the
difference in rate of cleavage and other developmental activities at differ-
ent levels showing a single gradient decreasing basipetally will serve as an
example. Specifically different substances may be present in apical and
basal regions of the echinoderm egg at the beginning of development; if
not present then, they doubtless are later. When present, they are un-
doubtedly concerned in differentiation of cells; but form and proportion —
morphogenesis in its larger, more general aspects — appears, at least during
earlier developmental stages, to depend to a much greater extent on quan-
titative factors of pattern, factors of rate or intensity, than on regional
distribution of specific substances. Substance is significant in develop-
ment only as it takes part in or affects activities; the activities, not the
substances, are formative. Differential dye reduction indicates some of
the activities and their quantitative differences and changes of rate but
tells us nothing about the substances concerned. These differences and
changes in rate appear to be essential factors in development of form and
proportion, and they seem to the writer to offer less difficulty in inter-
pretation of the more general features of early echinoderm morphogenesis
than the specific concentration gradients of Runnstrom. Even if this is
true, however, the specific concentration gradients may also be present.
Nevertheless, the data at hand seem to favor the view that factors of rate
or intensity are more important in determining form and proportions in
the earlier stages of echinoderm development than specific or qualitative
differences of substance. Definite patterns of form and proportion are
evident before cellular and regional differentiations are distinguishable
with certainty. Further investigation is, of course, necessary; but in spite
of the fact that data and conclusions concerning different species are not
in complete agreement as regards presence or appearance before gastrula-

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS 143

tion of a secondary acropetal gradient, and in spite of the hypothesis of
Ranzi and Falkenheim (p. 139) that there is no gradient, although their
data indicate one, it appears evident that a gradient pattern of some sort
is a fundamental factor in early echinoderm development. This will per-
haps become even more evident in the following chapter. It remains for
the future to determine to what extent use of different methods and pro-
cedures, of different species as material, and possibly preconceived views
may be concerned in the lack of agreement and lack of information on
certain points. In later chapters some analysis of other lines of recent ex-
periment on echinoderm development will be attempted.

OTHER INVERTEBRATES

A few data are at hand concerning mollusks. Early developmental
stages of several gasteropods show basipetal decrease of susceptibility, as
indicated by cytolysis and death and later a local increase in the region
which becomes the shell gland (Child, unpublished). Differential suscepti-
bility to various agents has been observed in early cephalopod develop-
ment (Ranzi, 1927; 1928; 1929a, b). In gasteropods and a cephalopod the
apical region of the egg becomes more alkaline than the basal (Spek,
1934a; Ries und Gersch, 1936). Methylene blue leucobase (rongalite
white) oxidizes most rapidly in the apical region of the egg of Aplysia,
an opisthobranch gasteropod. The more rapid oxidation of the leucobase
by single blastomeres, even by single or certain micromeres, observed by
these authors, raises certain questions. Oxidation is more rapid in one
cell of the two- and four-cell stages and in some of the micromeres of later
stages. Since the prototroch of the early larva also oxidizes leucobase
more rapidly than other cells, they regard the differential oxidation in
earlier stages as indicating cell lineage of the prototroch and velum from
one cell of the two-cell stage and a single quadrant of the four-cell stage
and from certain micromeres of this quadrant. The cleavage is not fol-
lowed in detail ; and in the light of earlier studies on cell lineage in mollusks
and annelids, such origin of the prototroch appears improbable. The pos-
sibility suggests itself that difference in condition associated with stage
of the division cycle may be responsible for some of the differences in
rate of leucobase oxidation in individual cells; such difference may also
determine a difference in susceptibility to the toxic rongalite white, and
the deep staining of certain blastomeres may result from injury. These
authors also find that reduction of Janus green decreases in rate from
the basal pole acropetally in maturation and cleavage stages of Aplysia.

144 PATTERNS AND PROBLEMS OF DEVELOPMENT

Since the basal region of this egg consists largely of yolk, the apical region
of cytoplasm, and the rate of cleavage and morphogenesis is much higher
apically, this result is rather surprising. There can be little doubt that it,
like the reduction differential observed by Gersch in Paramecium (p. 92,
footnote 2), results from injury by the very toxic Janus green to the
apical region, and consequent retardation or absence of reduction there.
There is no evidence in this work of any attempt to determine whether
different concentrations of dye and dififerent staining periods would give
different results. ^^

Direct evidence of physiological differentials or regional specificities in
early stages of arthropod development appears to be lacking, except for
demonstration of differential suceptibility to cyanide in the insect Bruchus
quadrimacidatus (Coleoptera). In stages from egg-laying to 6| hours sus-
ceptibility becomes highest on the ventral surface, and the most suscepti-
ble area gradually becomes localized in the median ventral, presumptive
prothoracic-maxillary region. From 6^ to 12 hours ectodermal suscepti-
bility decreases from this region anteriorly through the head region and
posteriorly through the embryonic plate; from 12 to 16 hours it remains
essentially the same. The locus of highest susceptibility in the ventral
prothoracic region is the position of the "differentiation center" of Seidel
(pp. 515-21). The aggregations of nuclei at certain regions of the cortical
layer of the insect egg, their changes of position in some eggs, and the
progress of formation of cell boundaries about them from certain egg
regions indicate, of course, presence of physiological pattern of some
sort; but of what sort, these observations do not show.

During the summer and autumn, 1940, suggestive data on differential
reduction of methylene blue and Janus green in the ovaries of the insect
Drosophila hydei have been obtained (Child, unpublished). The ovaries
consist of parallel series or strings of follicles in each of which the basal
cell becomes the egg, the other cells becoming accessory or nurse cells.
The apical pole of the egg is toward the apical ovarian pole, that is, the
pole where follicle development begins; and the ventral side of the egg is
toward the outer ovarian surface. If overstaining and differential injury
are avoided, dye reduction in low oxygen progresses basipetally in the

^' These authors have investigated various aspects of cytoplasmic differentials and differ-
entiations in eggs and early stages of a considerable number of animals, with certain results of
interest and undoubted value; but in some respects the data presented appear inadequate as a
basis for the conclusions drawn and suggest the need of a considerably wider range of experi-
ments. See Ries und Gersch, 1Q37; Ries, 1937; Gersch und Ries, 1937; and their citations of
earlier papers.

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS 145

apical ovarian regions consisting of the younger follicles and from the
outer ovarian surface inward. In older single follicles in which the grow-
ing oocyte is distinguishable from nurse cells in living material, the nurse
cells reduce before the oocyte, and reduction in the oocyte and in the
follicle wall progresses basipetally. In the full-grown egg cell a basipetal
reduction gradient usually appears, and some evidence of a ventrodorsal
gradient has been obtained; but this gradient, if present, is slight. Since
the data show a uniform and constant relation between morphological
and physiological pattern of ovary, folhcle, and egg, they suggest that
the ovarian differential determines which cell in the follicle becomes
egg and the axiate pattern of that egg.

EARLY CHORDATE DEVELOPMENT
ASCIDIANS

A polar-dorsiventral pattern is directly visible in some ascidians in
differences in color and appearance of certain cytoplasmic regions, but
this pattern appears only after fertilization. Moreover, the visible pat-
tern is of no significance for ascidian development, for in other species no
such pattern is visible, but development follows the same course, and a
polar-dorsiventral pattern is inferred.'" But it is perhaps worth repeating
that, whatever the egg pattern is, it is not necessary for ascidian develop-
ment. Ascidians can develop from pieces of the body, from stolon tips
and isolated pieces of stolon, from buds, and from aggregations of cells.
Evidently, the regional cytoplasmic pattern of the ascidian egg is a prod-
uct or expression of a more fundamental pattern of some sort; that is,
more or less regional differentiation occurs in this egg before cleavage be-
gins. Apparently, the only direct evidence of gradient pattern in ascidian
development is the differential reduction of potassium permanganate in
Corella willmeriana (Child, 1927(f). In early cleavage the reduction gra-
dient is distinctly basipetal (Fig. 50, A). In the gastrula a region corre-
sponding to the dorsal region of the amphibian embryo (see Fig. 156)
in its relation to the blastopore, though not in relation to the primary
polarity of the egg, reduces most rapidly, with lateral margins of the
blastopore following (Fig. 50, B, C). Even before outgrowth of the tail,
increase in rate of reduction appears in the region of the prospective tail
(Fig. 50, D), and the longitudinal gradient becomes two-armed or U-
shaped; that is, a new gradient, opposed in direction to the primary

^oConklin, 19050, 6, c, 1906, 1931; Dalcq, 1935, 1938; and various citations by these
authors.

146

PATTERNS AND PROBLEMS OF DEVELOPMENT

gradient, has appeared, as apparently in annelids and also in echinoderms.
With outgrowth of the tail a very distinct reduction gradient appears in
it, the tip being the most rapidly reducing region of the larva; the neural

Fig. 50, A-H. — Differential KMn04 reduction in larval development of the ascidian Corella
willmeriana, arrows indicating direction of progress. A, four-cell stage with basipetal decrease
in rate; B, C, gastrula in lateral and dorsal view; D, increase in rate of reduction in region of
future tail; E, caudal gradient with higher rate of reduction at tip than in any other part of
body and early reduction of scattered tunic cells; F, G, later stages of larval development; //,
caudal resorption about to begin (after Child, ig2yd).

side of the larval body and the end reduce more rapidly than the other
side (Fig. 50, E, F, G). During its development the tail is also the most
susceptible region of the larva. In the full-grown larva the caudal reduc-
tion gradient is still distinct; but, as compared with the body, rate of

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS 147

caudal reduction is relatively less than earlier. Before resorption of the
tail begins, its reduction gradient becomes very slight; during resorption
traces of it may persist, or only irregular areas of more rapid reduction
scattered along the tail appear (Fig. 50, H). As this figure indicates, the
more rapid reduction on the neural side of the larval body persists.

The appearance at the gastrula stage of a region of high rate of reduc-
tion corresponding to the dorsal inductor region of amphibia, the dorsal
lip of the blastopore, is of interest, as compared with the higher rate of
respiration and oxidation in this region of the amphibian gastrula ob-
served or indicated by most experimental data (pp. 153-58) and in rela-
tion to the view, now widely held, that there are fundamental similarities
in spatial embryonic pattern throughout the chordates. Differential per-
manganate reduction does not, of course, give any information as legards
presence or absence of regional cytoplasmic specificities. It does suggest,
however, that a quantitative gradient pattern perhaps underlies and pre-
cedes the regionally specific pattern.

CYCLOSTOMES

Differential susceptibility to ammonia and acetic acid of the lamprey
(Entosphenus appendix), as determined by Hyman (19266), is shown in
Figure 51. The susceptibility gradient of the unfertilized egg (A) and
early cleavage stages (B, C) is basipetal. In more advanced cleavage it
is in general basipetal; but some irregularities as regards cells or cell
groups appear, perhaps because susceptibility differs at different stages
of the division cycle (D). In the early blastula the basipetal gradient is
distinct and symmetrical (E); but in the later blastula one side, pre-
sumably dorsal, becomes more susceptible than the other {F, G). The
region of invagination is more susceptible than any other part in the early
gastrula {H), the dorsal side is more susceptible than the ventral (/, /),
and this dorsiventral difference increases as gastrulation proceeds {K, L).
As the neural groove develops, its anterior region becomes the most sus-
ceptible part of the embryo {M), but the original basipetal (anteropos-
terior) gradient and the higher susceptibility of the blastopore region are
still evident in the later stages of disintegration (A^) . With elongation of
the head region death progresses from the anterior end of the brain pos-
teriorly in the dorsal region and laterally and ventrally from this region,
and a region extending dorsally and anteriorly from the blastopore also
shows high susceptibility (0, P) ; this, according to Hyman, is the segmen-
tal plate or somite-forming region. At the time of hatching, the gradient

148

PATTERNS AND PROBLEMS OF DEVELOPMENT

is predominantly anteroposterior (Q, R) ; 2-3 days later susceptibility of
the posterior region has increased so that anterior and posterior arms of
the gradient are almost equal (S, T). No essential difference in the re-

FiG. 51, A-T. — Differential susceptibility in development of the lamprey Entosphemis
appendix; disintegrated regions indicated by dotted bounding-lines; and the chief directions of
progress of disintegration, by arrows. A, unfertilized egg; B, eight-cell stage; C, about thirty-
two cells; D, later cleavage; E, late cleavage or early blastula; F, G, two stages of disintegration
in late blastula; H-J, three stages of disintegration at beginning of gastrulation; A', L, later
gastrula; M, N, neural groove stage, disintegration progressing posteriorly, but original
apicobasal gradient evident in later disintegration (N); 0, P, susceptibility decreasing posteri-
orly from the elongating head along the nervous system and an increased susceptibility along
the segmental plate; Q, R, at hatching, susceptibility decreasing posteriorly along the nervous
system, with secondary region of high susceptibility posteriorly; S, T, 2-3 days after hatching,
anterior and posterior arms of the gradient almost equal (after Hyman, 19266).

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS 149

gional susceptibilities to the two agents was observed; the same pattern
appears with both. The relation between susceptibility and developmen-
tal activity seems evident. The higher susceptibility of the presumptive
dorsal side, and later of the actual dorsal side and the region about the
blastopore, indicates a change in condition of this region as gastrulation
approaches. This is of interest in comparison with the increase in rate
of reduction in the region corresponding to the dorsal lip of the blastopore
in the ascidian (Fig. 50, B, C) and the increase in susceptibility of the
amphibian dorsal region at gastrulation; it is also in line with the evidence
of higher rate of respiration and of dye reduction in this region (see
pp. 153-58).

TELEOSTS

Differential susceptibility of developmental stages of three teleost
fishes, Tautogolahrus adspersus, Fundulus heteroclitus , and Gadus morrhua,
has been determined by Hyman (192 1). Figure 52 shows some of the
stages. During early cleavage of Tautogolahrus susceptibility decreases
from center to periphery of the blastoderm {A). In the later blastoderm
the center is no longer the most susceptible region; disintegration begins
at the posterior border of the blastoderm, that is, in the region which is to
become the embryonic area, and proceeds anteriorly {B, C). This region
represents the dorsal lip of the blastopore, and its high susceptibility at
this stage corresponds with the findings in the ascidian, cyclostome, and
amphibian. After the embryo appears, the gradient in it is from anterior
to posterior and from dorsal to ventral {D, E). In still later stages the
posterior end of the embryo becomes a secondary region of high suscepti-
bility, and death progresses anteriorly from it, as well as posteriorly from
the head region and laterally and ventrally from the dorsal region {F).
In its earlier stages Fundulus resembles Tautogolahrus, but in the embryo
the posterior region of high susceptibility appears earlier and, up to ad-
vanced embryonic stages, is more susceptible than the anterior end (G-/).

The development of Gadus shows somewhat different susceptibihty re-
lations in earlier stages (Fig. 52, J-P). In this species the germ ring de-
velops early, and it and the embryonic shield are much more distinct and
more sharply marked off from other parts than in the other two forms. In
the earliest stages observed, the periphery of the blastoderm where the
germ ring is forming is more susceptible than the center (/). In early
germ-ring stages one side of the ring is more susceptible than the other;
and when the embryonic shield forms, it is the most susceptible region
{K, L). In more advanced stages of the shield and in early stages of the

ISO

PATTERNS AND PROBLEMS OF DEVELOPMENT

embryo an anteroposterior gradient appears, corresponding to the polar
axis of the embryo (M, N). At a somewhat later stage, before closure of
the germ ring, a second region of high susceptibility appears at the pos-
terior end of the embryonic shield and persists in the posterior end of the
embryo {O, P). As particular organs — the eyes, the heart, etc. — are local-

FiG. 52, A-P. — Differential susceptibility in early teleost development; regions and prog-
ress of disintegration indicated as in Fig. 51. A-F, Taulogohibrns adspersus; G-I, Fundidus
heteroclitus; J~P, Gadus morrhua (after Hyman, 192 1).

ized and develop, they appear as regions of high susceptibility; in the em-
bryonic heart susceptibility decreases from the venous end.

It is suggested by Hyman that the decrease of susceptibility from the
center of the early blastoderm peripherally in Tautogolahrns (Fig. 52, ^)
indicates a more primitive type of development with developmental ac-
tivity centered about the animal pole; the early appearance of higher
susceptibility peripherally in the blastoderm of Gadus is regarded as as-
sociated with the earlier appearance and more definite development of
the germ ring in this species. In any case, the relation of differential sus-

PHYSIOLOGICAL CIL\RACTERISTICS OF AXIATE PATTERNS 151

ceptibility to developmental activity is evident in these teleosts, as else-
where. In Figure $2, B, C, K, and L, the region representing the dorsal
lip of the blastopore shows high susceptibility; in B and C stages are
figured showing the highest susceptibility in the region of early invagina-
tion. During elongation of the embryonic shield its anterior end becomes
the most susceptible region; but sooner or later, at somewhat different
stages in the different species, a second region of high susceptibility appears
posteriorly, as in other segmented animals.

An electric-potential difference exists between apical and basal poles
in early stages of Fundulus development and undergoes cyclical reversal,
apparently in relation to cyclical changes in the cells. A potential differ-
ence also occurs along one axis of the blastoderm (Hyde, 1904). At the
time of maturation alkaline colloids accumulate apically, the blastodisc
becoming distinctly marked off from the acid remainder (Spek, 1933).

AMPHIBIA

The great volume of investigation concerned with amphibian develop-
ment in recent years has revealed many facts and perhaps presented even
more problems. We have learned something about the pattern of am-
phibian development, but the problem of pattern remains and continually
presents new aspects. Some of these are touched upon in later chapters;
the present concern is chiefly with some of the more direct evidence bear-
ing on the general characteristics of amphibian spatial or regional pattern
in earlier stages of development.

The protoplasm-yolk gradient of the amphibian egg, the superficial
pigmentation present in most species except in the basal region of high
yolk content, and the gradient, decreasing basipetally, in rate of cleavage
indicate an apicobasal pattern. Dorsiventrality is clearly indicated in the
fertilized eggs of some anurans by the gray crescent, a less deeply pig-
mented, subequatorial region between the deeply pigmented apical, and
the light basal region, decreasing in width laterally and ventrally from
the mid-dorsal region. It is indicated in Figure 53, A-C. In eggs of other
amphibia a corresponding region is more or less clearly indicated by a
difference in appearance from other parts. Observations on differential
susceptibility, as indicated by cytolysis, give evidence of a definite pattern
of physiological condition undergoing progressive change with the prog-
ress of development. ■''

^' Bellamy, 1919; Bellamy and Child, 1924; material, Rana pipiens, Chorophihis nigritus
Bufo americaniis, and some data on Amblystoma tigrinum.

152

PATTERNS AND PROBLEMS OF DEVELOPMENT

In the unsegmcnted egg disintegration progresses basipetally from the
apical pole but more rapidly in a broad band over the dorsal surface to
the region of the gray crescent (Fig. 53, yl). In early cleavage essentially

Fig. 53, A-D. — Differential susceptibility in early development of Rana pipiens. A, four
stages of disintegration in undivided egg; B, four stages in early cleavage; C, in late blastula;
D, in early gastrula (from Bellamy, 1919).

the same susceptibility pattern appears (Fig. 53, B). In late cleavage a
region near the middle of the gray crescent begins to disintegrate at about
the same time as the apical pole, and the two areas spread and meet along
the dorsal side of the embryo (Fig. 53, C) and gradually extend basipetally

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS 153

and ventrally over other parts. The region of the dorsal Hp of the blasto-
pore is more susceptible than the apical pole in the early gastrula; from
this region disintegration extends anteriorly to a greater extent than in
earher stages, but also extends basipetally from the apical pole (Fig. 53, /)).

The anterior end of the embryo is localized near, or more or less ventral
to, the apical or animal pole of the egg, according to the species; the apical
region, then, becomes anterior. After gastrulation is completed, two areas
of high susceptibility still persist, one anterior, the other dorsal; but the
dorsal area is less susceptible than the anterior. In the earlier neurula
stages disintegration begins in the median anterior region of the neural
plate; and in slightly later stages, usually at two points where the optic
primordia are developing. From this anterior region it spreads posteriorly
along the neural plate and joins the area of disintegration at the posterior
end of the embryo. Susceptibility of neural-plate stage to radium shows
essentially the same differentials (Stachowitz, 1914). In early stages of
organ development optic vesicles, nasal pits, ventral suckers, and tail bud
all appear as local areas of high susceptibility with local radial or axial
gradients.

The data on susceptibility indicate that the dorsal region is different
in some way from lateral and ventral regions, but the apical region is
primarily the most susceptible. As development progresses, the relative
susceptibility of the region which is to become the dorsal lip of the
blastopore increases, until at the beginning of gastrulation it is the most
susceptible part of the embryo, and in it susceptibiHty decreases anteriorly
and laterally from the blastopore. This is the region which, after invagina-
tion, becomes the so-called ''organizer" or "organization center" (see
pp. 454-70). In a study of differential susceptibility in other amphibian
species Cannon (1923) found practically no uniformity in the suscepti-
bility of different regions at various stages of development (see Appendix

VI, p. 743)-

Recent investigations on the respiration of the dorsal lip region, as
compared with other parts, are of much interest in relation to the role
of this region as inductor of the neural plate and to the data on suscepti-
biHty. In thirty-six out of forty-four cases (82 per cent) in which the
dorsal lip region was destroyed, oxygen uptake was found to be lowered
by 30 per cent or more, as compared with controls in which an equivalent
destruction of an indifferent region had been accomplished. Gastrulae
from which the dorsal lip region has been removed have a lower CO^
production than those from which an equal volume of cells has been re-

154 PATTERNS AND PROBLEMS OF DEVELOPMENT

moved from the ventral marginal zone of the blastopore. Also, the CO^
production of the isolated dorsal lip region is distinctly higher than that
of an equal cell volume from the ventral marginal zone (J. Brachet, 1934a,
b). Later determinations by Waddington, Needham, and Brachet (1936)
on fragments taken immediately above the dorsal lip and fragments of
presumptive ventral ectoderm showed oxygen uptake of the same order
of magnitude in both. According to determinations by J. Brachet (1936),
however, CO^ production is about 85 per cent higher in pieces from the
dorsal lip than in ventral pieces; oxygen uptake per milligram of nitro-
gen is 35 to 40 per cent higher in the dorsal than in the ventral pieces;
and the respiratory quotient is 1.06 in the dorsal, as compared with 0.76
in ventral pieces. The higher CO2 production in the dorsal lip results in
part from a higher absolute rate of metabolism and in part from the
higher respiratory quotient.

Simultaneous determination of oxygen uptake of the two sides of single
intact eggs and embryos showed the dorsal side averaging 47 per cent
higher than the ventral in gastrulae of the frog {Rana sylvatica) .^^ With
orientation of the embryos in the capillary so that the dorsal lip was
symmetrically placed in relation to the two sides of the apparatus, no
significant difference in oxygen uptake appeared. The authors suggest
that the earlier failure by Waddington, Needham, and Brachet (1936) to
discover a difference in respiration of dorsal and ventral piece probably re-
sulted from use of pieces taken from above the dorsal lip rather than from
the dorsal lip itself. They also point out that alteration of respiratory
rate after isolation may be concerned in determinations on isolated pieces.
In unfertilized eggs of R. pipiens they found no consistent difference in
oxygen uptake between apical and basal hemispheres. If there is any
actual difference, it is probably very slight until activation occurs on fer-
tilization or by other means, and it may decrease when the egg remains
for some time without fertilization.

Determinations of oxygen uptake, respiratory quotient, ammonia pro-
duction, anaerobic and aerobic glycolysis in very small fragments of dorsal
and ventral regions of amphibian embryos of various species with the
Cartesian diver used as an ultramicrorespirometer show no consistent dif-
ferences in oxygen uptake, a higher respiratory quotient dorsally (about
i), as found by Brachet, anaerobic glycolysis of dorsal pieces about three

3^ J. Brachet and Shapiro, 1938. Eggs and embryos were introduced singly into capillary
tubes with lumen about equal to, or slightly larger than, their diameter, and each end of the
tube was connected with a manometer.

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS 155

times that of ventral, ammonia production higher dorsally, and Httle or no
aerobic glycolysis in any part.-^' It is not evident that the very small frag-
ments used in these determinations contribute in any way to greater
accuracy or certainty of results. If metabolism is altered by isolation,
greater alteration may be expected with smaller than with larger pieces;
in isolated pieces of planarians and hydroids such differences appear. The
data do not indicate whether isolation alters metabolism in these embry-
onic pieces or whether size of piece may be a factor in rate after isolation.

According to another series of determinations of oxygen uptake, ^^ that
of the apical half of the blastula is 3.1 times that of the basal half, that
of the dorsal inductor region twice that of the intact gastrula, that of the
neural plate of the neural stage 2.61 times that of the whole neurula, and
that of the anterior region of the later embryo higher than that of the
posterior part.

In a later paper J. Brachet (1939) finds little difference in oxygen up-
take between dorsal and ventral sides in R. fusca at the beginning of
gastrulation and agrees with Boell, Needham, et al. that the great dififer-
ence found earlier probably resulted from the presence of two cell layers
in the dorsal region; that is, the determinations on single embryos required
so much time and temperature was so high that invagination was ad-
vanced before they were completed. However, it is perhaps still a ques-
tion whether the second cell layer beneath some part of the dorsal ecto-
derm would increase oxygen uptake 47 per cent. There is also the possi-
bility the oxygen uptake increases in the invaginated portion of the dorsal
region. Brachet's later determinations of CO, production by a colori-
metric method show the dorsal side 29 per cent higher than the ventral;
but this difference is, at least in part, due to the difference in respiratory
quotient. He also finds dorsal pieces more susceptible than ventral to
agents which inhibit glycolysis (monoiodoacetate and NaF). Carbohy-
drate metabolism, indicated for the dorsal region by the respiratory quo-
tient about unity, he regards as not essential to induction, and he shows
that induction is not necessarily prevented by inhibition of glycolysis.

^3 Needham and Boell, 1938; Boell, 1938; Boell, Needham, and Rogers, 1939; Boell and
Needham, 1939; Boell, Koch, and Needham, 1939; Needham, Rogers, and Shih-Chang Shen,
1939. More recently Boell and Nicholas (1940, Absir. Amer. Soc. ZooL, Anal. Bee, 78, 4,
Suppl.) have found a basipetal decrease in respiration dorsally and a steeper decrease ven-
trally, but they maintain that when corrections are made for the different amounts of yolk
in the different regions actual quantitative respiratory differences do not appear. See also
Needham, 1939, "Biochemical aspects of organizer's phenomena," Groidh, Suppl.

" Fischer und Hartwig, 1938; oxygen as mm^O, per 10 mg. dry weight per hour.

156 PATTERNS AND PROBLEMS OF DEVELOPMENT

Still another series of determinations of oxygen uptake in pieces from
Amhly stoma and Rana embryos" shows the dorsal lip region much higher
than the ventral lip and slightly higher than the presumptive epidermis,
but Barth points out that the cells of the dorsal lip contain more yolk
than those of the epidermis; consequently, intensity of respiration in the
metabolizing protoplasm of the dorsal lip must be considerably higher
than appears from the data. On the ventral side oxygen uptake decreases
basipetally from the apical region.

Somewhat earlier, glycogen distribution in amphibian embryos was in-
investigated by Woerdeman (1933a, h, d). His findings show more or less
uniform distribution in the apical hemisphere with decrease in the dorsal
lip on invagination. Similar conclusions were reached by Tanaka (1934),
and a decrease was also observed on invagination of transplanted pieces
of the dorsal lip region (Raven, 1933a, 1935a). Using a different method
for demonstrating glycogen, Pasteels (1936c) maintains that the supposed
decrease does not occur. However, work with a quantitative microchem-
ical method indicates that before gastrulation glycogen decreases basip-
etally from the apical pole and that during gastrulation there is decrease
in amount in all regions, but that decrease is greatest in the invaginating
material (Heatley and Lindahl, 1937). These observations are interesting
in relation to the data of Boell, Needham, et al. on glycolysis.

Most amphibian eggs and embryos are unsatisfactory material for dye
reduction because of their pigmentation. Recently, however, some obser-
vations have been made on unpigmented or shghtly pigmented forms.
Staining the unpigmented neurula of the urodele Triton cristatus by re-
oxidation of the leucobase of brilliant cresyl blue and observing reduction
under strictly anaerobic conditions, Fischer und Hartwig (1936) found
reduction most rapid in the floor of the open neural plate. Using the same
material and the same procedure, Piepho (1938) followed the course of
reduction in blastulae and gastrulae. In blastulae reduction occurred most
rapidly in a region extending from the apical pole almost to the equator
and from there progressively to the basal pole. Gastrulae, so far as they
gave definite results (60 per cent of the total), showed, like blastulae,
more rapid reduction in the region apical to the equator and also in the
region of the dorsal lip of the blastopore, the inductor region. Whether
one of these regions reduced more rapidly than the other could not be
determined. In these experiments reduction occurred under nitrogen, that

35 Barth, 1939a, h; oxygen uptake in cubic millimeters of oxygen per 100 mg. dry weight
per hour.

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS 157

is, oxygen was completely excluded from the beginning. The lesser differ-
ences in rate are undoubtedly more clearly distinguishable with gradual
oxygen decrease than with complete exclusion at once. Probably with
gradual decrease of oxygen a reduction gradient progressing basipetally
from the apical pole would have appeared instead of reduction of the
apical hemisphere without differential.

In eggs and embryos of Triturus rividaris there is considerable variation
in depth of pigmentation. Some lots of eggs were found by Child to be
light enough to permit observation of reduction. -^^ With gradual decrease
of oxygen dye reduction could be observed to progress basipetally in
blastulae, but whether a dorsiventral differential was present could not
be determined; if present, it was very slight. In the naked early gastrula,
however, reduction was most rapid at the dorsal border of the blastopore
and progressed laterally and anteriorly from this region. Apical and dorsal
reduction could not be observed in the same individual because of the
flattening, but embryos in proper orientation showed reduction decreasing
in rate from the apical region basipetally. In stages with open neural
plate rate of reduction appears highest in the floor of the plate and de-
creases from the anterior region posteriorly. It occurs over the whole
plate, including a narrow region just outside the neural folds, before re-
duction is clearly evident in the general ectoderm." A few observations
on elongated embryos with closed neural folds showed rate of reduction
decreasing from the head posteriorly along the dorsal region and laterally
and ventrally from that, with a posterior region of higher rate as develop-
ment of the tail began.

Using the method of Giroud and Bulliard (1933) for indicating occur-
rence and distribution of sulphydryl compounds by a color reaction,
J. Brachet (1938) has shown a definite and changing distribution during
development of several amphibians of what he believes to be sulphydryl-
proteins. In the unfertilized egg and soon after fertilization the reaction
occurs in a region about the apical pole. Some 2 hours later it is localized
more or less closely in a crescent, corresponding, in forms in which dorsal
and ventral sides are distinguishable, to the position of the gray crescent
of the frog's egg. In later cleavage and blastula stages the reaction occurs

^^ Slight staining with Janus green of blastulae within the membrane was found possible
after removal of the jelly. With care the membrane can be removed from gastrulae; but, when
naked, the gastrulae become flattened on the surface in contact.

37 The Triturus material was provided through the kindness of Dr. V. C. Twitty, and his
skill in removing the membranes of early gastrulae was of great assistance.

158 PATTERNS AND PROBLEMS OF DEVELOPMENT

in the apical half, sometimes extending farther basipetally on one side.
At the beginning of gastrulation the apical region still reacts strongly, the
reaction extends farther basipetally on the dorsal than on the ventral side,
and the dorsal lip of the blastopore shows in most cases an intensification
of the reaction. In the neurula the reaction is most intense anteriorly on
the floor of the neural plate and about the blastopore, less intense along
the whole dorsal region. The head in later stages reacts most intensely,
but the reaction occurs along the whole dorsal region. These data are of
special interest in relation to the widely accepted view that sulphydryl
compounds are essential factors in oxidation-reduction. It is perhaps also
of some significance that, except for stages just before and during early
cleavage, the regions of most intense reaction are practically identical with
those of highest susceptibility and most rapid dye reduction.

Although the different lines of evidence are not in complete agreement
on certain points, they all show the presence of a definite spatial pattern
of physiological condition and activity. This pattern is on a molar or
regional scale, with gradation in condition from the apical pole basipetally
and from the dorsal region ventrally. A progressive change in physio-
logical condition and apparently, according to most of the data, a greater
increase in rate of metabolism of some kind occurs in the dorsal region
than elsewhere in the course of pregastrular development. Most of the
available evidence indicates that at the time of gastrulation the region
which becomes the dorsal lip of the blastopore, and after invagination the
inductor also, becomes the most intensely active region of the embryo.
The higher respiratory quotient of this region, indicating that carbohy-
drate metabolism is predominant, has already been noted. Since the re-
spiratory quotient of the whole embryo later approaches or becomes unity,
it is conceivable that the dorsal region, in consequence of its higher level
of activity, exercises a certain degree of dominance over other parts in
raising their metabolic levels; and with the rise in level, change in the
character of metabolism may perhaps occur. In fact, it is suggested by
Waddington, Needham, and Brachet (1936) that the change in metabo-
lism indicated by the increase in respiratory quotient may spread over
the embryo from the dorsal lip. The question whether the high metabolic
level of the dorsal region is concerned in its action in inducing develop-
ment of the neural plate after its invagination is also of fundamental im-
portance, but its consideration is postponed to a later chapter, in which
the problem of induction in general is taken up.

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS 159

THE CHICK EMBRYO

Determinations of oxygen uptake by Shearer (1924) showed a higher
rate in anterior than in posterior parts; but in a later paper, without fur-
ther experiment, the earher data are discarded as without significance

Fig. 54, A~G. — Differential reduction of Janus green in earlier stages of the chick embryo;
reduction indicated by stippling. A-C, progress of reduction in primitive-streaiv stage; D, E,
head-process stage; F, G, stage of first intersomitic groove; injury of neural folds and retarda-
tion of reduction indicated by small crosses (from Rulon, 1935).

(Shearer, 1930). A recent study of dye reduction has given results of
value (Rulon, 1935). Reduction of Janus green to the red form by the
embryos occurs readily in low oxygen. In primitive-streak and head-proc-
ess stages rate of reduction decreases posteriorly from the region of the
node (Fig. 54, A-C) and the head process (Fig. 54, D) and laterally from

i6o PATTERNS AND PROBLEMS OF DEVELOPMENT

the median region. The margin of the area pellucida where yolk is being
digested also reduces rapidly (Fig. 54, A-E). At the stage of the first
somitic groove the border of the anterior end of the neural plate is just
beginning to fold dorsally and is so highly susceptible to the dye that,
when the rest of the embryo is appreciably stained, it is irreversibly in-
jured, and reduction in it is greatly retarded or does not occur at all; with
slight staining, however, it shows a very high rate of reduction. In Figure
54, F and G, this region is indicated by small crosses as injured. At this
stage the region of the node reduces somewhat less rapidly than the an-
terior region, and reduction progresses anteriorly and posteriorly from
it (Fig. 54, F, G). At stages of three to four somites the region of the
neural folds most susceptible to the dye is shghtly posterior to the anterior
end (Fig. 55, yl) ; as in the earlier stage, it is a region of very rapid reduc-
tion after slight staining. Except for the neural folds, the region of the
node reduces most rapidly with rate decreasing anteriorly and posteriorly
from it (Fig. 55, A~C). Somewhat more slowly reduction occurs at the
extreme anterior end of the head and progresses posteriorly, except for the
injured neural folds, which reduce slowly or not at all. At the stage of
eleven to twelve somites reduction is first evident anteriorly in the open
neural folds of the anterior head region, the region of the neuropore, and
the lateral regions, which are to form the optic vesicles, reducing most
rapidly in this anterior region (Fig. 55, D, E). These regions also stain
somewhat more deeply than other parts of the neural plate. Rate of re-
duction decreases posteriorly from the anterior end over most of the
embryo; but the primitive streak, still present in the posterior region,
appears as a second region of rapid reduction, with rate decreasing an-
teriorly and posteriorly from it, and the developing heart now becomes a
region of rapid reduction (Fig. 55, Z), £). Preceding torsion of the embryo,
the level at which torsion occurs is temporarily a region of rapid reduc-
tion. At the twenty-somite stage torsion is completed, and flexion of the
head is taking place. The regions of most rapid reduction at this stage
are the primitive streak, the auditory vesicles, the future gill slits, and
the region of the hindbrain, where flexion is greatest. The reduction gradi-
ent is now predominantly from the anterior end (Fig. 55, G, //, /) ; reduc-
tion in the primitive streak is relatively less rapid than earlier, and the
gradient does not extend very far anteriorly from it. The region of most
rapid reduction posteriorly is the future tail bud (Fig. 55, G, H). Figure
55, J, shows reduction in an early stage of the limb buds. Even before

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS i6i

any marked outgrowth the hmb buds are regions of rapid reduction with
decrease in rate more or less radially from a center, as in buds generally.
In these embryos the rapid reduction in the region of the heart and gills

Fig. 55, A~J. — Differential reduction of Janus green in later stages of chick embryo.
A-C, four-somite stage with injury and retardation or absence of reduction in neural folds;
D-F, twelve-somite stage; G-I, twenty-somite stage; /, early limb-bud stage(from Rulon, 1935).

may be in part due to more rapid decrease of oxygen there in consequence
of the crowding of parts. In still later stages each feather germ stands
out as a region of more rapid reduction than the skin about it.

l62

PATTERNS AND PROBLEMS OF DEVELOPMENT

The reduction gradient in three stages of the developing heart is shown
in Figure 56, A-C. The posterior region, where the omphalomesenteric
veins fuse, shows the highest rate with progressive decrease anteriorly,
but with higher rate on the right side (left in the figures).

Studies of differential susceptibility to cyanide, ammonia, and sodium
hydroxide by Hyman (1927a), to ultra-violet radiation by Hinrichs (1927),

B

c

Fig. 56, .4-C.— Differential dye reduction in three developmental stages of chick heart
(fromRulon, 1935).

and to hydrocyanic acid by Buchanan (1926^) are in essential agreement,
not only with each other but with the results of differential dye reduction,
so far as the same stages are concerned. In primitive-streak and head-
process stages a simple anteroposterior gradient appears. From early neu-
ral fold stages on, the anterior region and the region of the node are, in
general, most susceptible; but other regions of high susceptibility appear,
some temporarily. In early somite stages the region of closure of the

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS 163

neural folds is more susceptible than other parts of the head; and as
closure progresses, the region of high susceptibility moves posteriorly and
then disappears. As the optic lobes develop, they become temporarily
more susceptible than other parts of the head, and later the otic primordia
also appear as regions of high susceptibility. Preceding torsion, the hind-
brain region again becomes highly susceptible, a change probably associ-
ated with torsion and disappearing later. Susceptibility becomes high in
the primordium of the heart, decreasing anteriorly from the posterior end,
where the sinus, the dominant region of the heart, finally develops (Hy-
man, 1927&). As the tail bud develops, high susceptibility appears in it,
and the limb buds become highly susceptible as they appear. In general,
except where lateral organs become localized regions of high susceptibility,
the embryo shows decrease in susceptibility from the median region lat-
erally. The remarkable correspondence of regions of high susceptibility
to agents as different in their action as potassium cyanide, ammonia,
sodium hydroxide, and ultra-violet light with regions of rapid dye reduc-
tion and, so far as observations go, with regions of high susceptibility to
Janus green is highly significant not only as evidence of the changes in
physiological pattern as development proceeds but also as indicating that
differential susceptibility shows certain real physiological differentials and
their changes. The results suggest that activation is an essential factor,
in the early stages of development, not only of the whole embryo but of
particular regions and organs.

MAMMALS

There is, at present, no direct evidence bearing on the question of
physiological pattern in early mammalian development, but much indirect
evidence from the occurrence of teratological forms, many of which are
similar to the differential modifications of development under experimen-
tal conditions in other vertebrates described in the following chapter.
Moreover, there is adequate ground for the belief, very generally held,
that there are fundamental similarities in developmental pattern among
vertebrates and that the differences in course of development are, in large
part, dependent on presence and amount of yolk in fishes, amphibians,
reptiles, and birds and its absence, together with uterine development, in
mammals.

Functional patterns of various organs of adult mammals indicate a
gradient pattern of some kind as an underlying factor. The heart with
its dominant region at the sinus end, the high end of a gradient in early

1 64 PATTERNS AND PROBLEMS OF DEVELOPMENT

cardiac development of birds (Fig. 56), is an example. The mammalian
alimentary tract, particularly the small intestine, is another. There are
several regions of relative dominance along the tract, beginning with the
pharynx. These may function separately, or temporary subordination of
one or more may occur. In the small intestine gradients of irritability,
latent period, muscle tone, rate of rhythmic contraction of isolated pieces,
susceptibility, and respiration have been demonstrated, all corresponding
in direction, but some temporarily reversible in direction by local stimu-
lation at some level below the dominant region at the pylorus There is
a remarkable similarity in functional pattern of the mammalian intestine
and the ctenophore plate row (see pp. 106, 327).-'^ The ureter is another
organ in which there is evidently a very similar pattern of function. The
central nervous system shows various evidences of a gradient pattern,
as might be expected from its close developmental association with the
gradient pattern of early stages. ^^

CONCLUSION

The data of this chapter, though incomplete or fragmentary and often
only indicative or suggestive rather than conclusive, constitute, on the
whole, a remarkably consistent body of evidence. While complete agree-
ment is lacking on various points, the high degree of parallehsm of results
obtained on the same material with different methods is sufhciently ob-
vious. There is little room for doubt that the methods are not all con-
cerned with the same aspects of protoplasmic differentials or gradients,
but it does seem evident that they are concerned with different aspects
of the same pattern. The spatial patterns of early development are not
merely patterns of physical condition, of structure, of concentration of
substances, of metabolism, but of all of these — in short, they are spatial
patterns of living. If this is true, we shall learn all that they are only as
we learn all that living protoplasms are. Whatever they are, metabolism
appears to be the effective factor in the progress of development. Without
it physical structure and chemical substances would not bring about de-
velopment. Moreover, not only metabolism, but metabolism in a spatial
pattern, is essential to orderly axiate development.

Most of the evidence available indicates or suggests that this pattern
is primarily quantitative, or at least that differences in rate of living
play primarily a more important part in development than differences in

J* See Alvarez, 1928, and his citations of special papers,
•i"* See, e.g., Herrick, 1924; Coghill, 1929.

PHYSIOLOGICAL CHARACTERISTICS OF AXIATE PATTERNS 165

kind of living. They appear to be the primary factors in bringing about
axial and regional differences in rate of development and in form and
proportions.

Eggs of different animal species differ greatly at the beginning of em-
bryonic development : in some there is apparently little or nothing more
in the way of developmental pattern than differentials in rate ; others give
direct or indirect evidence of more or less specific regional differentiation
in the cytoplasm. Even in these, however, the cytoplasmic differentia-
tions apparently are, or become, localized in definite relation to a general
gradient pattern and are probably to be regarded, like other features of
development, as expressions or consequences of that pattern, appearing
in the egg rather than in later stages (see chap. xiv). Gradient pattern
along any physiological axis is itself a three-dimensional pattern, a gradi-
ent system; it may perhaps be regarded as a general background on which
details are gradually filled in, or as a frame of reference within which,
and in relation to which, developmental events occur. But unless the
following pages are wholly in error, it is more than a background or a
frame of reference; it is a physiologically active and effective factor in
initiating the order and unity of the individual.

CHAPTER V

DIFFERENTIAL MODIFICATION OF DEVELOPMENT:
COELENTERATES AND FLATWORMS

IN SO-CALLED ''normal" development only certain developmental
potentialities are realized; it is possible to alter the course of realiza-
tion experimentally by altering environmental conditions and so to
alter form and proportions of parts, localization of particular differentia-
tions, and even the general patterns of symmetry, asymmetry, and polar-
ity. We call the individuals thus produced "abnormal forms," "mon-
sters," "terata"; but it is scarcely necessary to point out that normal de-
velopment represents the reactions of the protoplasm concerned to a par-
ticular environment and abnormal development its reactions to other en-
vironments.

Differential susceptibility was discussed in chapter iii, but certain
points concerning its relation to development are briefly recalled to atten-
tion here. Experiment has shown for many animal species that with ex-
posure in early developmental stages to an external inhibiting factor in
sufhciently high concentration or dosage to bring about more or less in-
hibition of all parts, but not rapidly lethal, the inhibition of development
is differential — ^in other words, a gradation in degree of inhibition, de-
creasing from the most to the least susceptible regions is evident. So far
as adequate data are available, the gradient pattern indicated by differ-
ential inhibition of development is, in general, essentially the same as that
indicated by other methods and is definitely related to the axiate pattern
of the organism (see chap. iv).

With a certain lower range of concentration or dosage differential tol-
erance or acquirement of increased tolerance (a differential conditioning)
to many agents appears sooner or later following a primary inhibition.
Under these conditions differential modification of form and proportions
is opposite in direction to that in differential inhibition. The regions most
inhibited primarily show secondarily the greatest tolerance or condition-
ing. After temporary exposure to the inhibiting agent, provided its action
has not produced persistent injury, a differential recovery may occur,
with modification of development similar, in general, to that in differen-
tial tolerance and conditioning but often more extreme. If the more sus-

i66

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. I 167

ceptible regions are so injured that they cannot recover, the less suscep-
tible may still be able to recover to some extent; it seems desirable to dis-
tinguish this, as partial recovery, from the differential recovery of the
whole, in which the regions primarily the most susceptible show the great-
est recovery. Differential acceleration of development has also been ob-
served in some forms with certain agents. As pointed out in chapters iii
and iv, the appearance of the same differentials or gradient patterns with
many different agents, both physical and chemical, suggests that differ-
ential susceptibility depends primarily on factors of rate rather than of
kind of activity, the most active regions being primarily most susceptible
to the more extreme effects of external factors and most capable of toler-
ance or conditioning to, and recovery from, less extreme effects and in
some cases most susceptible to accelerating effects. In other words, dif-
ferential susceptibility appears to be, to a high degree, nonspecific for
different agents and apparently depends on quantitative factors of gradi-
ent pattern rather than on specific regional differences in the organism,
if such are present, or on the nature of the agent and the particular man-
ner in which it acts on living protoplasms.

The data on experimental modifications of early development resulting
from exposure of the entire intact organism to altered environmental
conditions indicate that most, if not all, of the modifications thus pro-
duced depend on nonspecific differential susceptibility and fall under one
of the above heads (Child, 1924a, igiSd). Moreover, it appears that phys-
iological inhibiting factors of various sorts act differentially on develop-
ment in the same way as external factors. The present chapter and the
two following are chiefly concerned with some of the differential modifica-
tions of development and their bearing on the problem of pattern.

DIFFERENTIAL MODIFICATION OF HYDROID DEVELOPMENT

Differential inhibition of embryonic development of the hydromedusa
Phialidium gregarium has been effected by KCN, LiCl, phenyl urethane,
dilute neutral red, and CO^ with essentially similar modification by all
agents (Child, 1925&). In normal development cells from the basal re-
gion of the blastula immigrate singly into the blastocoel to form ento-
derm (Fig. 57, A-C). Under the inhibiting conditions the number of im-
migrating cells increases, the region of immigration extends farther api-
cally (Fig. 57, D), and, instead of forming a single cell layer of entoderm,
the immigrating cells may form a solid mass, more or less completely
obliterating the blastocoel (Fig. 57, E-G). Cells may also be given off

i68

PATTERNS AND PROBLEMS OF DEVELOPMENT

externally as well as internally, usually from the basal region (Fig. 57,
H, I), occasionally over most of the surface. This emigration occurs
more frequently and to a greater extent in LiCl than in other agents used.
The cells given off are normal in appearance, with no evidence of cytoly-

B

D

Fig. 57, A-K. — Phialidiiim gregariiini. Normal and differentially inhibited development.
A-C, normal development of planula; D, E, differential inhibition in early stage with increase in
number of immigrating cells and extension apically of region of immigration; F, G, differen-
tially inhibited early planulae, excessive immigration; H, I, immigration and emigration;
/, K, obliteration of polarity, equal immigration from all regions resulting in solid spherical
mass (from Child, 19256).

sis; the solid planulae remain alive and, as will appear, may develop fur-
ther. With more extreme inhibition in early blastula stages all evidences
of polarity are obliterated, immigration occurs equally from all parts of
the wall, giving rise to completely solid spherical larvae without visible
structural differences of any kind (Fig. 57, J, K). Ciliary co-ordination

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. I 169

or control is evidently lost in these forms; locomotion is limited to an in-
definite rolling-about instead of definitely directed with apical end in
advance, as in normal planulae. If returned to water, these forms may
live 2 weeks or more without further development and without any indi-
cation of a developmental pattern. Apparently they are unable to develop
unless a new pattern is determined in them (see pp. 425-26). Except for
the more frequent emigration of cells in LiCl the same modifications ap-
pear as with other agents used. All of them "entodermize"; that is, the
higher gradient levels are more or less completely inhibited down to the

Fig. 58, A-D. — Normal and inhibited development of axes in Phialidium. A, normal at-
tachment of planula by end originally apical, and development of hydranth-stem axis from end
originally basal; B, C, inhibited planulae with some degree of differential tolerance or differ-
ential recovery, giving rise to stolon axes; D, hydranth-stem axis developing from planula
stolon after recovery (from Child, 19256).

level characteristic of the basal region in normal individuals, so that im-
migration of cells is not limited to the basal region but occurs farther api-
cally or, in the spherical forms, equally from all parts of the wall, and
some cells may emigrate instead of immigrating. The entodermization re-
sembles that occurring in echinoderms, and the emigration of cells is prob-
ably comparable to echinoderm exogastrulation (see chap. vi).

With lower concentrations of inhibiting agents which do not completely
obliterate polarity, retarded elongation of the planula occurs. The normal
planula in good condition swims actively and attaches by the apical end
(Child, 1925a), and the first hydranth develops from the original basal
end (Fig. 58, A). The slightly inhibited planulae do not swim free but

I70 PATTERNS AND PROBLEMS OF DEVELOPMENT

move along the bottom of the container and finally come to rest with
one side, instead of the apical end, in contact. After several days indica-
tions of differential conditioning, or of differential recovery if they are
returned to water, appear; these consist in outgrowth of stolons from the
apical end or from both ends (Fig. 58, B, C), representing either the
primary gradient of the planula or both gradients (pp. 96-97). They
may continue to grow as stolons at the expense of other parts of the
planula body or, if returned to water, may recover suf]ficiently to give
rise to a hydranth-stem axis (Fig. 58, D). Failure of the planulae to
attach by the apical end and development of stolons instead of hydranth-
stem axes from one or both ends occur with very slight degrees of inhibi-
tion— for example, in slightly crowded cultures or with insufficient aera-
tion, as well as with low concentrations of various chemical agents.

The normal development of the gymnoblast hydroid Corymorpha after
attachment of the planula is shown in Figure 59, A-C. As indicated in
Figure 30 (p. 98), the high end of the primary gradient is apical, and a
secondary gradient appears basally in the course of development. The
egg of Corymorpha is so opaque that effects of differential inhibition on
early development are obscured; but polarity can apparently be obliter-
ated, as in Phialidium, by cyanide, lithium, alcohol, ether, and probably
by methylene blue and neutral red, the embryo remaining spherical within
the membrane, though alive. With less extreme inhibition, retarding de-
velopment but not preventing hatching and not obliterating polarity, the
planulae become much longer than normal because the less susceptible
basal and middle regions are less inhibited than the apical end (Fig. 59,
D, E). Further development of these forms in the inhibiting agent is
indicated in Figure 59, F-H. If tentacles appear at all, they do not de-
velop beyond early stages; distal tentacles are more inhibited than proxi-
mal and often do not form (Fig. 59, F) ; often development does not go
beyond the stages of Figure 59, D and E. In all cases the hydranth region
is small and the stem large and elongated. Evidently the less susceptible
stem has been able to grow to a greater extent, as compared with the
hydranth, than in normal development. If the forms of Figure 59, F-H,
are returned to water, hydranth development proceeds, the hydranth
growing at the expense of the stem, so that the usual proportions are
approached, that is, differential recovery occurs. With slightly greater
inhibition development rarely goes beyond the stages of Figure 59, / and
J , except after return to water.

Differential regression of development may also be brought about by

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. I 171

inhibiting agents. Stages like Figure 59, B and C, may undergo regression
to forms essentially like F, G, and H and develop again if returned to

Fig. 59, /1-L.— Normal and differentially inhibited development of hydroid of Corytnorpha
palma from planula. A-C, stages of normal development; D, E, differentially inhibited de-
velopment in planula stage; F, G, H, later stages of differentially inhibited development; /, /,
more extreme inhibition; K, L, regression from stage of B to planula-like stage.

water. Regression of stages like Figure 59, B, to forms resembling early
planulae completely inclosed in perisarc is possible with inhibiting agents,

172

PATTERNS AND PROBLEMS OF DEVELOPMENT

but renewed development of these forms following return to water has
not been observed. These regressions take place without visible disinte-
gration, but the tentacle cells are apparently resorbed and may serve as
nutrition. No evidence of specificity of the various modifications for any
agent used has been found, and comparison with the lethal susceptibilities
(pp. 10 1-4) and with results of differential dye reduction shows that
the same gradient pattern is indicated by the different methods.'

Fig. 60, A-C. — Boiigainvillia mertensii. Stolon development from hydranth-stem axes
under inhibiting conditions. A, 48 hr. in ethyl urethane m/200; B, KCN m/50,000; C, KCN
m/50,000 for 6 days, then KCN m/ioo,ooo with further decrease in concentration for 8 days
(from Child, 1923a).

The transformation of hydranth-stem axes into stolon axes in hydroids
is of interest as a differential inhibition in later stages of these forms
(Child, 1923a). From the study of many hydroid species it appears that
hydranths are much more susceptible than stolons and that the motile
hydranth is more susceptible than the nonmotile bud. In some hydroid
species stolons are not infrequently found developing from apical ends of

' The data on Corymorplui have not been previously published.

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. I 173

branches, particularly in the more basal parts of the system, under natural
conditions. Experiment throws light on the character of these transforma-
tions. Certain highly susceptible hydroids, e.g., BougainvilliasindPlumu-
laria, when kept in standing water in the laboratory, lose their hydranths
in a few days, sometimes in a few hours— the fully developed hydranths
usually by disintegration, the buds apparently by regression and resorp-
tion. A few days later stolons develop in place of hydranths, even from
the apical ends of axes, the whole system often showing nothing but

Fig. 61. — Gonothyraea clarkii. Stolon development from hydranth-stem axes in standing
water (from Child, 1923a).

stolons. Still later, under the same conditions, appearance of hydranth
buds suggests some degree of conditioning. With low concentrations of
inhibiting' agents these transformations may be even more rapid. In
ethyl urethane m/200 Bougaimillia may undergo complete transforma-
tion in 48 hours (Fig. 60, A). In KCN m/50,000 the stolons are often
subapical, and after a week or more the appearance of hydrant hbuds in-
dicates some degree of conditioning, but the buds do not develop beyond
an early stage (Fig. 60, B) . A stolon system developing in cyanide from
a single apical end is shown in Figure 60, C. Transformation in Gono-
thyraea is slower but may take place in standing water (Fig. 61) ; in Obelia

174 PATTERNS AND PROBLEMS OF DEVELOPMENT

and several other species similar transformations occur. Ethyl urethane,
cyanide, MgS02, LiCl, neutral red, and CO2 all give essentially similar
results. Pieces of Plumularia setacea (California) in standing water lose
the original hydranths and develop numerous stolons from lateral
branches and from both cut ends of the chief axis, and no, or very few,
hydranths. In flowing water hydranths remain and hydranth-stem axes
develop from one or both cut ends of the chief axis, no stolons or very few
appearing. Sertularella miurensis (Japan) in standing water loses hy-
dranths and develops lateral and terminal stolons. In hypotonic standing
sea water, 75 and 50 per cent, and in flowing normal sea water hydranths
remain alive, and few or no stolons develop until inanition is far advanced.
These transformations of hydranth-stem axes into stolon axes obviously
result from differential inhibition. The most susceptible parts, the hy-
dranths, die or are resorbed, and the inhibited axes develop stolons. With
differential conditioning or recovery the stolons may give rise to hy-
dranths. Each stolon represents a growth gradient, the tip growing and
remaining in good condition, at the expense of other parts, in the absence
of food or after separation from the stock. As the stolon increases in
length, separation from the parent stock occurs because the cells at the
low end of the stolon gradient serve as food for cells nearer the tip. The
tip continues to grow at the expense of more proximal levels until reduc-
tion of the coenosarc to very small size. In Figures 60 {C) and 61 the
graded shading of the stolons indicates the gradation from the tips, where
cells are in good condition and fill the perisarc completely, to levels where
only a slender coenosarcal strand of atrophying cells persists. Unshaded
parts of these figures indicate empty perisarc. The stolon axis is a growth
gradient, the hydranth-stem axis is a differentiation gradient; the stolon
axis develops under somewhat depressing or inhibiting conditions, the
hydranth-stem axis under conditions more nearly optimal. The inhibiting
factor may be external, as it undoubtedly is in the transformations of
apical regions into stolons in standing water and in slightly toxic solutions;
or it may be physiological, as in determination of stolon development at
proximal levels of an axis by dominance of the hydranth-region (see
pp. 314-15). In any case, stolon development evidently indicates a rela-
tively low level of certain metabolic reactions. Oxygen is probably often
insufficient to support hydranth metabohsm at the more basal levels of
hydroid systems as their own stems and branches become more numerous
or in consequence of overgrowth by other forms. Many facts indicate that
the oxygen content of sea water under natural conditions is often not far

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. I 175

above the critical level for hydranth development in various hydroid
species.' When it falls below that level, stolons develop in some forms;
in others there is no development, at least under laboratory conditions,
but insufficient food or other incidental conditions may be responsible for
its absence. Doubtless stolon metaboHsm is different from hydranth me-
tabolism, but apparently a low level of the kind of metabolism which
brings about hydranth development makes stolon development possible.

DIFFERENTIAL MODIFICATION OF PLANARIAN RECONSTITUTION

Isolated planarian pieces provide interesting and valuable material for
the study of differential developmental modification in its relation to de-
velopmental pattern. In these pieces it is possible to determine the differ-
ential effects not only of external factors but also of certain physiological
conditions. Pieces of different lengths from different body-levels, from
individuals of different physiological age, nutritive condition, previous
conditioning, etc., provide somewhat different starting-points for physio-
logical analysis. Moreover, the differential modifications are not limited
to the longitudinal axis.

LONGITUDINAL MODIFICATIONS

Under natural conditions regeneration of new tissue is more rapid at
the anterior than at the posterior end of the piece in Dugesia (see Figs.
17, 18), and the amount of new tissue formed in head-development is ap-
parently greater than in development of the posterior end. When pieces
are subjected immediately after section to concentrations of agents which
inhibit reconstitution but are not lethal, development of new tissue is
inhibited at both ends, apparently with little, if any, difference. At this
time, when activation of the cells near the cut ends is occurring, there is
apparently little difference in condition at the two ends. However, if the
concentrations of inhibiting agents do not completely inhibit regenera-
tion, the anterior new tissue gradually begins to grow and slowly develops
into a head, which may be normal or more or less differentially inhibited,
while posterior regeneration is almost or quite inhibited. Under these
conditions pieces which in normal environment give rise to complete indi-
viduals develop into tailless forms, often without pharynx (Fig. 62, A, B).
In these more new tissue appears anteriorly than posteriorly, but less
than in normal head development; and the head is formed in greater or
less part from the old tissue of the piece (Fig. 62, B-D) instead of entirely

= H. B. Torrey, 1912; Child and Watanabe, 1935^' Barth, 1937^/ J- A. Miller, 1937.

176 PATTERNS AND PROBLEMS OF DEVELOPMENT

from the regenerated tissue, as normally. In these cases the anterior re-
generating tissue of the piece apparently undergoes conditioning more
rapidly, or to a greater degree, than the posterior end. These secondary
modifications have been obtained with KCN, ethyl alcohol, chloretone
and other anesthetics, CO^, hydrogen ion, and low temperature. With
cyanide they develop very slowly, usually becoming evident only after sev-
eral weeks' exposure; in alcohol they begin to appear in a few days; with
other agents used, they develop at different rates between these extremes.
The history of these pieces suggests that immediately after section
there is little difference in condition at the two cut ends, but later there
is evidently a difference. If the pieces remain in water 48 hours after
section and are then subjected to the inhibiting agent, anterior regenera-

A B C

Fig. 62, A-D. — Anteroposterior differential tolerance and differential conditioning to in-
hibiting agents in pieces of Dugesia dorotocephala. A and B, early, C and D, later, stages of
differential tolerance or conditioning, showing greater inhibition of posterior reconstitution.

tion is more inhibited than posterior; that is, a direct differential inhibi-
tion occurs. This is more clearly evident with agents such as cyanide, in
which the secondary modification occurs very slowly. Measurements of
length and width of anterior and posterior new tissue make it possible to
compare areas of anterior and posterior regenerating tissue in controls
and experimental pieces; since thickness of the two regions is much the
same, the measurements serve as a rough ratio of amounts of anterior
and posterior regenerated tissue. The following data serve as examples:

Anterior regeneration
Posterior regeneration

Controls, 8 days in water i -43*

. f 2 days in water, 6 days in KCN m/ 100,000. . i.ii

\ Same lot after 14 days in KCN m/ioo,ooo . . i . 66

• The figures are the sums of products of lengths and widths of anterior, divided
by the sums of products of lengths and widths of posterior regenerated tissue, as
determined in each piece of the experimental lot.

In the experimental lot the ratio of anterior to posterior regeneration
after 6 days in cyanide is far lower than in the control, that is, inhibition

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. I 177

is greater anteriorly than posteriorly; after 14 days in cyanide the an-
terior-posterior ratio in the same pieces is considerably higher than in the
control, the conditioning of the anterior region to cyanide being appar-
ently greater than that of the posterior. These differences are directly
visible and often become greater in later stages, but these have not been
measured. With some agents — for example, ethyl alcohol — the secondary
modification occurs so rapidly that the differential inhibition preceding it
is slight or scarcely appreciable. In general, the greater susceptibility of
the regenerating head region than of the posterior regenerate to lethal
concentrations or dosages and its more rapid or greater conditioning in
a lower range of concentrations are conspicuous characteristics of planar-
ian pieces undergoing reconstitution. That they are expressions of a longi-
tudinal pattern appears beyond question.

HEAD FORMS AND HEAD FREQUENCIES IN PLANARIAN RECONSTITUTION

In the reconstitution of pieces of a number of planarian species graded
series of head forms develop, ranging from the normal fully developed
head to the completely acephalic condition. These head forms are evi-
dently expressions under different conditions of a differential or gradient
pattern in the parent body and persisting in the regenerating head with
both anteroposterior and mediolateral components. On the one hand,
the head forms constitute a continuous series of differential inhibitions of
head development; on the other, a series of secondary modifications rep-
resenting differential conditioning and recovery. Chief attention is given
to the inhibition series because of its relation to physiological, as well as
external, factors, because the degree of differential inhibition and conse-
quently the head form can be experimentally altered and controlled in
many ways, and because this series has thus far provided a basis for a
wide range of experimental analysis. The secondary modifications, though
physiologically equally significant, appear under rather narrowly limited
experimental conditions; they will be considered in a later section.

Head forms. — The head forms of the inhibition series have been classed
in five groups for convenience, but it must be remembered that the limits
of these groups are arbitrary and that the series is actually continuous.
The groups are as follows:

The normal head: The head form which is typical of the species in
natural environment is triangular in general outline, with two separate,
bilaterally localized "eyes" (photoreceptors) and two cephalic lobes (chem-

178

PATTERNS AND PROBLEMS OF DEVELOPMENT

oreceptors) on lateral margins of the head at a level slightly posterior
to that of the eyes. The head form is shown in Figure 63, ^, a transverse
section of head and ganglia at eye-level in Figure 64, A.

The teratophthalmic head : In animals developing from pieces of equal
length from the same body-level of the parent teratophthalmic heads are
usually slightly smaller than normal heads and often less sharply pointed

>1) CM) (H) CM) (X4)

(M) iT^) (m (M) ;*; (M)

Fig. 63, A-L.^The inhibition series of head forms of Diigesia dorotocephala. A, normal
head; B, teratophthalmic head; C, eyes of teratophthalmic heads; D-G, teratomorphic heads;
H-J, anophthalmic heads; K, L, acephalic forms.

anteriorly and shorter, at least in earlier stages. The eyes are localized
nearer the median line than normally, and the pigment cups are more or
less connected; all degrees of approximation, even to complete cyclopia,
occur in these forms (Fig. 63, B, C). The two eyes are sometimes unequal
and localized at somewhat different levels, but these irregularities ap-
parently result from incidental differences on the two sides of the piece ;
for example, the presence of an intestinal branch at the level of section

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. I 179

on one side may determine slightly slower growth on that side because
parenchyma cells are less numerous. Figure 64, B, shows that the median
ganglionic region is apparently absent and that the more lateral parts of
the ganglia are approximated to the median line and partially fused.

The teratomorphic head: This head is rounded anteriorly with median
eye, apparently single (cyclopia),^ and with cephalic lobes localized more
or less anteriorly instead of laterally and showing all degrees of approxi-
mation to the median line, with a single median lobe as the extreme (Fig.
63, D-G). The ganglionic modifications correspond in degree to those of
other parts of the head (Fig. 64, C, D). In the extreme teratomorphic

Fig. 64, A-D. — Transverse sections of normal and differentially inhibited heads. A,
normal head of large animal; B, teratophthalmic; C, D, teratomorphic, reconstituted heads
(from Child and McKie, 191 1).

forms (Fig. 63, G) the region corresponding to the normal head is only a
narrow median band, the median parts of the normal head not being
represented at all; the originally lateral regions of the normal head are
now anterior and near, or in the median plane.

The anophthalmic form: In these forms the anterior regenerate may
resemble the extreme teratomorphic head with a single median cephalic
lobe but is without an eye (Fig. 6^, H), or the head may be represented
only by an outgrowth of new tissue without distinguishable external dif-
ferentiation (Fig. 63, /, /). The ganglionic mass shows various degrees
of more extreme reduction than in the teratomorphic head, and in some
cases only traces of gangHa appear at best; and it might even be ques-

^ In sections two pigment cups forming a rounded mass and two nerves, side by side or
practically unite'd, or two pigment cups, one ventral to the other, are sometimes distinguish-
able; but usually there is only one (see Fig. 64, C, D).

i8o

PATTERNS AND PROBLEMS OF DEVELOPMENT

tioned whether the outgrowth should properly be called a head. However,
even forms like Figure 63, / and /, behave much more nearly like normal
animals than do the completely acephalic forms.

Acephalic forms: In these the anterior cut surface becomes strongly
contracted and a small amount of new tissue fills in the concavity and
differentiates into a continuation of the lateral margins which bound the
contracted cut (Fig. 63, K, L). There is no trace of cephalic ganglia or
of eyes, and locomotion usually continues only a short time when induced
by strong stimulation. These forms can be distinguished, even from
anophthalmic forms, by their locomotor behavior.
A pharynx and mouth develop in acephalic forms
from prepharyngeal levels (Fig. 63, K) but, so far
as observed, never in those from postpharyngeal
levels (Fig. 63, L).

In Figure 65, ^, the parts of the head which are
absent in different degrees of teratophthalmiaare ap-
proximately those lying between the corresponding
broken lines on the two sides of the median plane.
Figure 65, B, indicates in the same way the regions
absent in teratomorphic and anophthalmic forms.
The mediolateral differential susceptibility deter-
mines this differential inhibition, progressing from
the median region laterally. Particular attention is
called to the fact that the inhibition is not specific
for particular organs of the head but is regional, in-
volving those parts of cephalic organs which happen
to lie in the region concerned. Actually, of course,
the inhibition occurs before the organs have differen-
tiated; but when differentiation occurs, the relation of particular organs
to certain regions is clear. When the median head region is inhibited, the
median region of the cephalic nervous system does not develop elsewhere
but is absent; with inhibition extending farther laterally the ganglionic
defect extends farther laterally, the eyes are approximated, or cyclopia
or anophthalmia results, and the lateral cephalic lobes show all degrees
of approximation to the median line, finally a single median lobe; in
anophthalmic forms there are only the rudiments of head development.
In short, this differential inhibition of the planarian head shows no rela-
tion to particular organs or their parts but simply eliminates regions pro-
gressively from the median plane laterally. Other differential inhibitions
to be described later show similar characteristics.

B

Fig. 65, A, ZJ.— Cor-
responding broken lines
on the two sides indicate
approximately the re-
gions most reduced or
absent in teratophthal-
mic {A), teratomorphic
and anophthalmic {B),
forms.

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. I i8i

Head frequencies. — ^The frequency of occurrence of the various degrees
of differential inhibition of the anterior end in a particular lot of pieces
of the same length from the same body-level of animals of a certain
length and as nearly as possible in similar physiological condition has been
called, for convenience, "head frequency." In such lots head frequency
shows a high degree of constancy, but it is a rather sensitive indicator of
differences in physiological condition and in external environment. With
animals of a certain length from a particular locality, which probably
signifies similar nutritive condition, possibly in some cases conditioning
to some external factor, and with similar laboratory environment, head-
frequency values differ in definite directions with length of piece, level of
body, and conditions under which reconstitution occurs. In pieces from
animals of different length the values also differ in definite directions, with
fraction of body length represented. Moreover, they differ in character-
istic ways with differences in nutritive condition of the animals and with
differences in temperature and other factors of the environment of the
stocks preceding section. Animals from different localities or from the
same locality at different seasons of the year may give characteristically
different frequencies. Acclimating or conditioning animals to certain en-
vironmental factors preceding section alters the frequencies. And finally,
so far as comparison is possible, different species of the American genus
Dugesia { = Eu plana ria), and at least some species of the genus Planaria,
show characteristic species differences in head frequency. ">

HEAD FREQUENCY IN RELATION TO CERTAIN PHYSIOLOGICAL FACTORS

The relation between head frequency, length of piece, and body-level
is shown graphically for Dugesia dorotocephala in Figure 66, in which the
head-frequency indices for pieces 1/3, 1/4, 1/6, and 1/8 of the post-
cephalic body length are plotted as ordinates against body-levels as ab-
scissae. As the graph shows, head frequency decreases from the anterior
level to the region of the fission zone, that is, in the anterior zooid, scarcely
at all in 1/3 pieces, more in 1/4, still more in 1/6, and most in 1/8 pieces.
At the most anterior level the frequency is the same in pieces of all
lengths, but decreases more steeply as length of piece decreases. From the
region of the fission zone it increases posteriorly. With still shorter pieces
the decrease and increase are still steeper, but with sufficiently short
pieces the most anterior and posterior levels show a decrease in frequency,
as compared with longer pieces (Child and Watanabe, 1935a). In other

1 For the literature chiefly concerned with planarian head frequency, the method used for
obtaining a "head-frequency index," and the appHcation of statistical methods to evaluation
of head-frequency data, see Appendix VII (p. 745).

182

PATTERNS AND PROBLEMS OF DEVELOPMENT

words, the head frequency at any level decreases with decrease in length
of piece below a certain limit, but this limit differs widely at different
body-levels. The length of piece at which differential inhibition of head
development begins to appear increases posteriorly in the anterior zooid

V8-8

Fig. 66. — Head frequencies of 1/3, 1/4, 1/6, and 1/8 pieces of postcephalic regions of
Dugesia dorotocephala. Head-frequency indices plotted as ordinates against body-levels as
abscissae; unbroken and broken lines represent frequencies obtained by different groups of
students in different years.

and decreases again in the posterior-zooid region. In these lots of pieces
only head forms of the inhibition series appear, except for occasional sec-
ondary modifications indicating differential tolerance, conditioning to, or
recovery from, the inhibiting factor, and slight diiTerences in size, position,
and time of appearance of the two eyes, evidently due to incidental differ-
ences on the two sides of the cut end.

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. I 183

The physiological factor determining the graded series of head forms
in pieces below certain lengths is an effect resulting from section at the
posterior end of the piece and the following stimulation of the nervous
system and activation of cells there; it is apparently, at least in large
part, nervous in character, for section of the nerve cords posterior to the
level of head development inhibits the head to almost the same degree as
section of the whole body. It is also very clearly shown that the inhibition
of the head results from posterior section by delaying either posterior or
anterior section for different periods after the other section has been
made. Delay of posterior section even for i or 2 hours increases head
frequency in short pieces; but with delay of anterior section head develop-
ment is inhibited over a much longer period, complete normal head de-
velopment being attained only with 3 or 4 days delay. ^

Head frequencies are usually lower in pieces from animals kept without
food for several weeks before section than in similar pieces from well-fed
animals. Probably in the starved animals the cells from which the head
develops are less intensely activated and therefore more inhibited by the
stimulation resulting from posterior section. In pieces representing equal
fractions of body length head frequency is less in those from small, young
animals, doubtless because the pieces are shorter and the inhibiting factor
from the posterior section therefore more effective than in the longer
pieces from large individuals (Child, 191 1/). In general, animals in good
condition give a steeper head-frequency gradient than those in poor condi-
tion; this gradient is almost invariably steeper in newly collected animals
than after long periods in the laboratory (Rulon, 1937). These data sug-
gest that head frequency may be altered in two different ways : by alter-
ing the condition of the cells concerned in head development cHrectly, and
by altering the intensity or effectiveness of the inhibiting factor resulting
from posterior section and so indirectly altering the condition of the
head-forming cells. The data presented in the following section indicate
that both these possibilities are realized with external factors.

HEAD FREQUENCY IN RELATION TO EXTERNAL FACTORS

Head frequency may be increased or decreased by environmental fac-
tors. With low concentrations of KCN it may be decreased or increased,
according to length of piece, body-level of origin, period of exposure, and
concentration of cyanide. With the same concentration of cyanide and

5 For further discussion of the experiments on section of nerve cords and on delay of
anterior and posterior section see pp. 406-1 1 ; also Child and Watanabe, 1935a, and Watanabe,
1935^-

PATTERNS AND PROBLEMS OF DEVELOPMENT

b

the same exposure period it can be decreased in anterior, and increased
in posterior, pieces of the same length from the anterior zooid and again
decreased in similar pieces from the posterior-zooid region. In general, in
pieces with high head frequency in water these concentrations of cyanide
decrease it; in pieces with low frequency they increase it (Child, 1916^).
A few data are presented by way of illustration. Table 6 gives head fre-
quencies in pieces of different lengths and from different body-levels for
controls and different exposure periods with several concentrations of
cyanide. In the pieces x of Table 6 and Figure 67, which are so long that
the physiological inhibiting factor does not greatly
affect head frequency, cyanide decreases it in all
concentrations used, and the decrease is greater
with continuous exposure to a lower concentration
(Table 6, Ix) than after temporary exposure to a
higher (Table 6, II:r). In shorter pieces, approxi-
mately thirds of the anterior zooid (Fig. 6'j,a,b,c),
a-pieces show a decrease in head frequency with all
concentrations and exposure periods that are at all
effective. Three examples are given in Table 6, la,
Ila, and Ilia. In the c-pieces, on the other hand,
continuous exposure to low concentrations of cya-
nide increases head frequency from the more inhibi-
ted forms to teratophthalmic heads (Table 6, Ic)
and with shorter exposures may increase frequency
of normal heads (Child, 1916^). With higher concen-
tration and continuous exposure increase becomes
less, is not altered, or decrease occurs. In lie of Table 6 there is an increase
from anophthalmic and acephalic forms to teratophthalmic heads, but
there is also a decrease from normal to teratophthalmic; consequently, the
index shows no significant change. Shorter exposures (48-72 hours) to the
same concentration as lie may, however, produce marked increase. But
even brief exposures to still higher concentrations decrease frequency in
c-pieces, and deaths begin to occur (Table 6, IIIc). The ^-pieces (Fig. 67)
which are not included in Table 6 give results intermediate between a
and c. Often they show no significant change in frequency under condi-
tions in which decrease occurs in a and increase in c.

Within certain limits high temperature increases, low temperature de-
creases, head frequency; pieces from animals conditioned to a certain
temperature give different frequencies from those of animals conditioned

Fig. 67. — Outline in-
dicating lengths and lev-
els of pieces x, a, and c
of Tables 6 and 7

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. I 185

to a different temperature when both lots reconstitute at the same tem-
perature (Behre, 1918). Some effects on head frequency of temperature

TABLE 6
Alteration of Head Frequency by KCN (Dngesia dorotocephala)
Fifty pieces in each lot; lots a and c with same Roman numeral are from the same an-
imals; frequencies in percentages.

Pieces

Concentration
and Exposure

Normal

Teratoph-
thalmic

Terato-
morphic

Anoph-
thalmic

Acephalic

Dead

Index

control

m/ 200,000, con-
tinuous

96
32

4
68

99.2
86.4

I X

control

m/ioo,ooo, 96
hours

92

70

8

30

98.4
94

II .V

control

m/ 200,000, con-
tinuous

84
4

16
80

96.8
74-4

I a

6

6

2

control

m/ioo,ooo, con-
tinuous

86

14

76

97.2
71.6

II a

10

10

4

control

m/50,000, 24
hours

74
5^

26

48

94

88.4

III a

I c

control

m/200,000, con-
tinuous

10
8

28
78

2

18
8

42
6

49.2

74.8

II c

control

m/ioo,ooo, con-
tinuous

46
24

26

74

6

2

12

10

77.2
73-2

III c

control

m/50,000, 24
hours

12

8

50

44

8
6

16

20

14
18

4

66

58.4

and temperature-conditioning are shown in Table 7. Low temperature
before and after section determines a much lower frequency in a-pieces
than medium temperature before and after section (lai, Ilai), but only
slightly lower in c-pieces (Ici, IIci). Pieces from animals conditioned to

1 86

PATTERNS AND PROBLEMS OF DEVELOPMENT

low temperature but reconstituting at medium temperature (Ia2, Ici)
show a much higher frequency than those conditioned to, and reconstitut-
ing at, low temperature (lai, Ici) and also higher than those conditioned
to, and reconstituting at, medium temperature (Ilai, IIci). Conditioning
to medium, with reconstitution at low temperature, results in great de-
crease in a-pieces with some deaths and decrease with many deaths in

TABLE 7
Head Frequency and Temperature
Pieces a and c (Fig. 67) of each series from the same animals (50 pieces in each lot); fre-
quencies in percentages. "Low" temperature is 8°-io° C; "medium," i8°-2o°C.; "high,"
2 7°-3o°C.

Series
and Lot

Temperature before
and after Section

Normal

Teratoph-
thalmic

Terato-
morphic

Anoph-
thalmic

Acephalic

Dead

Index

I

low-low
low-medium

24

86

76
14

84.8

a

2

97.2

I

I
c
2

low-low
low-medium

10

14
62

14

30

8

50

4

4

2

34-4

72

medium-medium
medium-low

78

44

22
50

95-6
84

a

6

II

I
c
2

medium-medium
medium-low

20

6
6

24
28

50
40

24

39-2

24.4
(32.1)*

medium-medium
medium-high

76
92

24

8

95-2
98.4

a

III

I
c
2

medium-medium
medium-high

2
28

22
60

2

28
6

46

2

42.2
80

* Second index for living pieces only.

c-pieces (Ser. II). Series III shows the great increase resulting from re-
constitution at high temperature of pieces conditioned to medium tempera-
ture. Comparison of Ilai with Illai and of IIci with IIIci also shows
how slight the variation is in similar pieces from similar lots under the
same external conditions.

A number of anesthetics have been found to act essentially like cyanide,
the same concentration and exposure period decreasing or increasing

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. I 187

head frequency according to level of origin of the piece (Buchanan, 1922).
Long periods of exposure to low concentrations of caffein also decrease
head frequency in anterior, and increase it in posterior, pieces, an effect
like that of cyanide; but brief exposures to high concentrations mav in-
crease the frequency in ante-
rior, and decrease it in poste-
rior, pieces (Hinrichs, 1924a).''
Induced increase of motor ac-
tivity increases head frequen-
cy, whether by providing for
more adequate respiratory ex-
change by change in position
or by otherwise stimulating the
head-forming cells is uncertain
(Child, 1911/).

Carbon dioxide, hydrogen
ion, and certain organic acids
are highly effective in increas-
ing head frequency in posterior
pieces of the anterior zooid but
have very little effect in de-
creasing it in anterior pieces
(Rulon, 1936a, 1937). The ef-
fect of CO2 is shown in Figure
68. Hydrogen ion in the con-
centration found effective in-
creases the frequency in the
more posterior pieces, where it
is low; the increase is much

100

.

80

" \\

60

V\

\ ■ — ■m^f

40

\

20

1 1 1

B

C

D

E

Fig. 68. — Increase in head frequencies of 1/8
pieces from anterior zooids of 14-16-mm. animals
by CO2 {M) over control {L) ; indices as ordinates
against body-levels as abscissae (from Rulon,

1936a).

greater in pieces from animals previously conditioned to CO, and calci-
um antagonizes the effect of hydrogen ion (Fig. 69).

The action of strychnine (sulphate) is very similar to that of anesthetics
and CO2. It increases the frequency greatly in posterior pieces but alters
it little or not at all in anterior pieces (Figs. 70, 71) ; it apparently retards
development slightly, but the pieces are highly sensitive to stimulation.

^ In order to show increase in frequency in anterior pieces, it is desirable to use pieces which
give relatively low frequencies under standard conditions— for example, pieces from small
young animals, from starved animals, or pieces from large animals, but short enough to show
a low frequency.

1 88

PATTERNS AND PROBLEMS OF DEVELOPMENT

Evidently it acts chiefly on the physiological, supposedly nervous, factor
which inhibits head development; but, instead of making this factor more
effective, it decreases its effectiveness in some way, perhaps by general
stimulation of the nervous system with resulting functional disorganiza-
tion (F. S. Miller, 1937). With
certain concentrations and ex-
posure periods strychnine is so
highly effective that pieces
from the more posterior levels
of the anterior zooid, largely
acephalic in the controls, de-
velop a high percentage of nor-
mal heads.

In pieces so long that there
is no physiological inhibition of
head development in controls,
cyanide, various anesthetics,
caffein in certain concentra-
tions, CO2, hydrogen ion, and
low temperature all decrease
head frequency more or less;
that is, inhibited head forms
appear in significant percent-
ages, instead of high normal
frequency.

The head-frequency gradi-
ent of Dugesia a gills and D.
tigrina, both of which have a
posterior-zooid region, is es-
sentially like that of D. doroto-
cephala, except that in D. ti-
grina the physiological inhib-
iting factor is apparently less
effective and the gradient is less steep.'' Certain other triclad species with-
out posterior zooids show a decrease in frequency of head development
from anterior to posterior levels over the whole body length. In some

7 Head frequencies under natural conditions have been determined in the forms earher
known as Planaria lata (Sivickis, 1923), Euplanaria jnacidata (Watanabe, 1935/'), and E.
novangliae (Child) , all of which are now regarded by Hyman as D. tigrina (seep. 41, footnote 7).
A Japanese planarian from one locality was found to give differentially inhibited heads only in

100

— »#/

•M

80

" %.-

60

''\\a\">

40

'^s \\ ^^

\^

20

1 1

1

A B C D E

Fig. 69. — Head frequencies of 1/8 pieces from an-
terior zooids of animals conditioned to CO^ and non-
conditioned animals under different experimental
conditions. L, nonconditioned, reconstituted in water;
L', conditioned, reconstituted in water; M, noncon-
ditioned; M', conditioned, reconstituted in carbonate-
free water, pH 4.15; N, nonconditioned; N', condi-
tioned, reconstituted in carbonate-free water, pH 4. 5,
plus 0.5 gm. of CaCU per liter. Indices as ordinates
against body-levels as abscissae (from Rulon, 19360).

Fig. 70. — Head frequencies of 1/8 pieces of i6-cm. animals reconstituting in strychnine
sulphate m/450,000 (continuous lines) and controls (broken lines) ; corresponding experimental
and control lots are indicated by numbers 1-4. Values obtained from combined counts in heavy
lines; indices as ordinates, body-levels as abscissae (from F. S. Miller, 1937).

Fig. 71. — Head frequencies of 1/8 pieces of i6-mm. animals reconstituting in water after
3 days' exposure to m/300,000 strychnine sulphate; graphed like Fig. 70 (from F. S. Miller,
1937).

I90 PATTERNS AND PROBLEMS OF DEVELOPMENT

of these, differentially inhibited head forms appear (Abeloos, 1930);
in others, an essentially normal head form or no head at all devel-
ops on pieces of any length posterior to a certain level. ^ Apparently,
if the initial activation of the head-forming cells occurs in these latter
forms it is not inhibited physiologically, even in short pieces. It is per-
haps of interest to note that in these species there is more regeneration
of new tissue at the anterior ends of pieces than in Dugesia; consequently,
the developing head and ganglia are isolated to a greater degree from the
old nerve cords and so from any nervous factor originating at the posterior
cut surface and inhibiting head development. In the dendrocoelids Proco-
tyla and Dendrocoelum rate of head development decreases from anterior
levels of section to a region near the middle of the body; posterior to this
level head development has not been observed, irrespective of length of
piece or presence of a posterior cut surface, although some new tissue is
formed; but posterior ends regenerate at all except extreme anterior
levels. Apparently, activation of the cells at the anterior end of the pieces
following section decreases more steeply from anterior to posterior levels
in these forms than in Dugesia and at a certain level becomes insuffi-
cient to initiate development of ganglia and head, though still sufficient
under the dominance of more anterior levels to develop a posterior end.

SECONDARY DIFFERENTIAL MODIFICATIONS OF THE PLANARIAN HEAD

Not infrequently pieces undergoing reconstitution in water, but with
heads inhibited by the physiological factor, show secondary modifications
at the anterior end, suggesting a conditioning to, or a recovery from, the
earlier inhibition, more probably the latter, since the physiological in-
hibiting factor is present only temporarily following section. For example,
in anophthalmic forms like Figure 63, / or /, further development often
occurs after a week or more, resulting in forms like Figure 72, ^ and B.
In the case of Figure 72, ^, the median region of the anterior new tissue
has undergone elongation; in Figure 72, 5, not only elongation but de-
velopment of a median cephalic lobe and eye has occurred. In this case
the outgrowth finally attains a condition permitting development of a
teratomorphic head. Development of a normal or teratophthalmic head
from these primarily anophthalmic pieces has never been observed.

extremely short pieces, 1/16 or less of the body length, under natural conditions. Animals, ap-
parently of the same species, from another locality several hundred miles from the first showed
a head-frequency gradient essentially like that of D. dorotocephala, according to data obtained
by Professor T. Minoura.

* Sivickis, 1931a, 1933; Buchanan, 1933.

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. I 191

Heads inhibited by the physiological factor to the less extreme terato-
morphic forms such as Figure 63, E, often undergo a secondary transfor-
mation into forms like Figure 72, C, with resorption of the original
cephahc lobes and development of new lobes in the normal position and
of two eyes, also normally placed. Here the median head region, not rep-
resented in the primary development, appears secondarily, and a normal
head, except for the primary median eye, results. In other cases all three

Fig. 72, A-I. — Secondary modifications of head form, differential conditioning, and
differential recovery in Dugesia. A, B, differential recovery of anophthalmic heads; C, differ-
ential recovery of teratomorphic head; D, reduction of head and loss of preocular region in
low concentrations of alcohol and ether; E, F, secondary development of median region under
the same conditions that produced the primary effect, D (differential conditioning); G, ap-
proach to normal form after return to water; H, reconstitution of head in alcohol, ether, and
low temperature; /, secondary development of head in alcohol and ether (differential condi-
tioning) and with rise in temperature (differential acceleration) (in part from Child, 1921c).

eyes may be more or less fused, that is, the head does not develop beyond
the teratophthalmic condition.

Still more extreme secondary modifications of head form appear with
differential conditioning to, and recovery from, effects of external inhibit-
ing agents. When intact animals are placed in low concentrations of ethyl
alcohol or ether, ^ the head decreases in size relatively to the body, and
the preocular region often undergoes more or less complete reduction by
gradual resorption or by disintegration in the course of 2 or 3 weeks
(Fig. 72, D). At about this time or somewhat later new growth of the
median region may begin in the same concentration in which resorption

' .Alcohol 1-1.5 per cent; ether 0.2-0.4 per cent.

192 PATTERNS AND PROBLEMS OF DEVELOPMENT

or disintegration occurred earlier, and after about a month in the solution
head forms like Figure 72, £ and F, have developed. Apparently the
median region gradually acquires a greater ability to grow and develop
than lateral regions. After such forms are returned to water, the head
may gradually approach normal form, but even after two months or more
the preocular region is still larger than normal (Fig. 72, G).

Very similar modifications of head form occur in reconstitution of heads
on pieces under external inhibiting conditions. In ether, alcohol, chlore-
tone, and other anesthetics heads like Figure 72, H, appear frequently:
in these the growth of new tissue is largely inhibited; but in the course of
8-10 days a single median eye appears, often in the pigmented old tissue.
With some further development such heads would doubtless be terato-
morphic. After 2 weeks or more in the solution further growth may begin,
resulting in heads with elongated median region, and often two new eyes
and cephalic lobes appear in normal position (Fig. 72, /). Here the me-
dian regions not represented in earlier stages gradually develop as condi-
tioning proceeds and become overdeveloped relative to other parts of the
head. Essentially the same head forms appear in differential recovery
after exposure for a week or two to the same agents. Secondary modifica-
tions of the same type have also been obtained in the differential accelera-
tion of development resulting from change to high temperature after
conditioning to low. Pieces from a low-temperature stock (conditioned
to 3°-5° C.) reconstituting at 4°-8° C. produce little new tissue, and most
of them remain acephalic or like Figure 72, H, even after 3 months.
Brought to room temperature (2o°-24° C.) many pieces develop heads
with median elongations Hke Figure 72, /, or without median eye. These
cases may involve differential conditioning to the higher temperature, as
well as differential acceleration of development. The occurrence of these
forms with various anesthetics and with change from low to high tem-
perature makes it probable that they, like other modifications of planarian
head form, are expressions of differential susceptibility and not specific
for any particular agent. Doubtless they can be produced with many
other agents which permit relatively rapid differential conditioning or re-
covery.

THE PLANARIAN HEAD: INTERPRETATIONS AND SUGGESTIONS

It is a point of considerable importance that both the physiological in-
hibiting factor and the external inhibiting agents give the same inhibition
scries of head forms in Dugesia, a continuous graded series of medio-

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. I 193

lateral differential inhibitions of development. The facts suggest a grada-
tion decreasing from the median region laterally as regards the "physio-
logical level" necessary for initiation of development of the head. That
this level involves metabolism seems evident. The action of both physio-
logical and external factors in inhibiting head development apparently
consists in preventing activation of the cells concerned to the level re-
quired. The inhibition of development involves progressively regions far-
ther lateral as the action of the inhibiting factor increases.

This interpretation will serve for the longer pieces on which no in-
hibited heads appear under natural conditions. As regards the shorter
pieces, however, it appears, at first glance, to be in contradiction to the
results of experiments, for agents which inhibit head development differ-
entially in longer pieces with high head frequency usually have the op-
posite efifect on shorter pieces with a naturally low frequency. These
agents decrease the mediolateral differential inhibition even though they
retard development of the head. These effects of external agents indicate,
and experiments on delay of posterior and anterior section, to be pre-
sented later (pp. 406-11), provide conclusive evidence, that in the shorter
planarian pieces two antagonistic factors are concerned in determining
the head form on a particular piece : the one the activation of the cells
concerned in head formation; the other the effect, nervous stimulation,
or whatever it may be, resulting from section of the nerve cords at the
posterior end of the piece. This effect is probably not essentially different
from a functional stimulus which tends to maintain the cells as cells of a
particular body-level. In order to give rise to a head, they must become
free of relations to other parts. Removal of more anterior regions has
freed them from relations to those parts; but if nervous stimuli from more
posterior regions are sufficient to keep them, to some extent, functional
parts of a particular body-level, they do not attain the level of activation
necessary for development of a normal head, and the inhibition series
of head forms results, according to the effectiveness of the physiological
factor. Transverse section of the longitudinal nerve cords with minimum
injury of other parts, if within a certain distance of the anterior end of
the piece, decreases head frequency; but transverse section of other re-
gions at the same level has little or no effect on frequency (Watanabe,
19356). Also, removal of a short piece half the body width, involving
section of one nerve cord a short distance posterior to a level of head
regeneration, usually results in asymmetry of the head, the side anterior
to the half-section being more or less inhibited (Rulon, 19366).

194 PATTERNS AND PROBLEMS OF DEVELOPMENT

The shorter the piece at a given body-level and the farther posterior
in the anterior zooid the level of origin, the more effective is the physio-
logical inhibiting factor. The length of piece at which inhibition of the
head begins to appear increases from anterior to posterior levels of the
anterior zooid. At the anterior level the physiological factor is effective
only in very short pieces. The intensity or rate of activation of the head-
forming cells and the rate of head development decrease from anterior
to posterior levels (p. 43), and it may be that the nervous effect of
posterior section is better transmitted in the anterior direction at more
posterior levels. Either or both of these conditions may determine the
greater effectiveness at a given distance and effectiveness at a greater
distance of the physiological inhibiting factor at more posterior levels.

In any lot of similar pieces head frequency depends on the relation be-
tween the condition of the cells directly concerned in head formation and
the physiological factor preventing those cells from undergoing the change
in condition necessary for complete head development. The physiological
factor has been called an "inhibiting factor" because it inhibits head de-
velopment differentially, but it appears actually to be a factor which tends
to maintain the cells as functional parts of a certain region of the body.
In order to give rise to a head, the cells must become free from functional
relation to other parts of the body; and the physiological differential in-
hibition of head development represents all degrees of inability to attain
that freedom, because the nervous stimulation resulting from posterior
section tends to maintain another sort of cellular behavior. The parts of
the head which require the highest level of activation for initiation of their
development are first and most inhibited; these are the median head
regions. With lesser degrees of freedom more lateral regions are reduced
or prevented from developing, until, when the cells are completely domi-
nated by the physiological effect of posterior section, they do not react
at all to the absence of regions anterior to their level, and acephalic forms
result. According to length of piece, level of origin, physiological condition
of parent animal, and nature and concentration of external agent and
period of exposure, the degree of differential inhibition of head develop-
ment varies in definite, orderly ways. Experimental control of head fre-
quency is possible by controlling length of piece and level of origin and
condition of parent animal and also by external inhibiting and accelerat-
ing agents. All these methods of control are merely ways of controlling
and altering the balance between the two antagonistic factors concerned
in determining the degree of differential inhibition of head development.

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. I 195

Differences in effect of the various external agents on head frequency
depend on their relative effect on the two factors. These which inhibit
directly the activation of the head-forming cells decrease head frequency
even in long pieces in which the heads are not physiologically inhibited.
Those which inhibit nervous activity are effective in increasing head fre-
quency in short and posterior pieces in which it is physiologically in-
hibited, but they may also inhibit the head-forming cells directly and so
decrease head frequency in long and anterior pieces in which it is not
physiologically inhibited. Cyanide, for example, inhibits both factors and
may decrease or increase head frequency according to length of piece,
level of body, period of exposure, and concentration; but results can be
controlled and, with sufficient experimental background, predicted. Since
the physiological factor inhibiting head development is effective only tem-
porarily following section, temporary exposure to cyanide for a day or
two after section is most effective in increasing head frequency because,
after return to water, the head-forming cells recover more or less com-
pletely and are not longer inhibited by the nervous stimulation resulting
from posterior section. Exposure to cyanide during the whole period of
reconstitution is most effective in decreasing head frequency, because
under these conditions its direct action on the head-forming cells is con-
tinuous. The anesthetics used in controlling head frequency act more or
less like cyanide but are more effective in increasing frequency by in-
hibiting the nervous stimulation than in decreasing it by direct action on
the head-forming cells. Carbon dioxide, hydrogen ion, organic acids, and
strychnine inhibit the nervous factor but have relatively little effect, in
the concentrations used, on the head-forming cells; consequently, they
are much more effective in increasing head frequency in short and pos-
terior pieces than in decreasing it in long and anterior pieces. Tempera-
ture, on the other hand, apparently affects chiefly the head-forming cells:
a rise in temperature increases, a fall decreases, head frequency. Caffein
apparently may serve as an accelerating or a depressing agent according
to concentration and exposure period; consequently, it may either increase
or decrease head frequency in pieces of the same sort.

Differential conditioning to low concentrations of external agents and
differential recovery may also alter head frequency secondarily : terato-
morphic heads may become normal, except for the original median eye,
and anophthalmic forms may become teratomorphic. Also, a series of
differentially modified head forms, with overdevelopment instead of inhi-
bition of the median region, occurs. These modifications are, of course,

196 PATTERNS AND PROBLEMS OF DEVELOPMENT

not direct effects of physiological or external factors inhibiting head de-
velopment but represent secondary differential reactions of the develop-
ing head region following a primary inhibition.

There is, at present, no evidence of specificity in the differential modifi-
cations of head form. It was noted above that they do not concern any
particular organ or region of the head but involve any and all parts. The
differential inhibitions of head form constitute a continuous graded series
ranging from slight inhibition of the median region to complete inhibition
of the whole head. The inhibiting factor, whether physiological or an ex-
ternal agent, acts, of course, on the earlier stages of head regeneration be-
fore the various organs are present as localized differentiations. The sec-
ondary modifications, although appearing under a rather narrowly limited
range of conditions, evidently also constitute a continuous graded series.
The appearance of the same series of forms with physiological factors and
with many external agents suggests that the effects of all these different
factors on the primary pattern of the planarian head must be essentially
similar. In other words, the only possible conclusion, in view of all the
evidence, seems to be that the mediolateral pattern which is differentially
modified must consist primarily in a quantitative gradient or differential.
If specific differences are present in the cells from which the head develops,
they are not concerned in the differential modifications of form and pro-
portion. However, the differentiations which appear in the modified heads
suggest that the quantitative gradient provides the physiological basis
for differentiation, and the differential susceptibility of different levels
of this gradient to physiological and external factors determines not only
form and proportions of the modified heads but the differentiations which
appear in them.

Absence of physiological differential inhibition of head development
ii;i certain triclads may be due to one or more of several factors : activation
of the head-forming cells may be so intense that the nervous factor cannot
prevent it, or the nervous factor may be slight or less readily transmitted
anteriorly in some species than in others; in some species head regenera-
tion is more completely isolated from the old nerve cords by regenerating
tissue posterior to it; in some the head develops so slowly that the transi-
tory nervous effect of posterior section may disappear before it can be
effective in inhibiting head development. Only further experiment with
the various forms can throw light on these points.

CHAPTER VI

DIFFERENTIAL MODIFICATION OF DEVELOPMENT:
ECHINODERMS

BECAUSE of its plasticity under experimental conditions, echino-
derm development in its early stages is exceptionally interesting
material for investigation of differential modification of develop-
ment. The alterations of form and proportions and the changes in locali-
zation of parts resulting from exposure of the whole developing organism
to experimental environments are, to a high degree, definite in character
and experimentally reproducible. They are not simply "abnormal" or
teratological forms to be described as curious or mysterious anomalies of
development, but alterations in definite ways under controllable condi-
tions; as such, they constitute a highly significant body of evidence bear-
ing on problems of the physiology of developmental pattern. At present
little more than a beginning has been made in this field of physiological
analysis, but results obtained speak for themselves and promise much of
interest to further investigation. Thus far, only echinoids and asteroids
have been used as material.'

ECHINOIDS
FORM AND PROPORTIONS UNDER NATURAL CONDITIONS

Echinoid development to the pluteus larva follows, in general, the same
course in different groups and species with only minor differences in form
and proportions, skeletal pattern, etc. Cleavage, blastula, and gastrula
stages are figured in chapter iv (Figs. 44-46, and 49, A, B). In Figure
73 the gastrula {A), the prepluteus (B, C), the pluteus {D, E) of the
sand dollar Dendraster excentricus, and the pluteus of the sea urchin
Arhacia punctulata (F, G) are shown.

■ In the following account figures are used extensively as more effective than any amount
of description. Most of the figures are based on ocular micrometer measurements of the di-
mensions of living individuals, supplemented by sketches of the same individuals. Form and
proportions are shown as exactly as possible, but various details of structure are often omitted
or diagrammatically indicated, particularly mesenchyme and skeleton. All figures of a group
are on the same scale. For data concerning concentrations of agents and exposure periods,
supplementing those given in the legends, see Appendix VIII (p. 747).

197

KiG 73 A-I -Echinoid development under natural and slightly inhibiting conditions.
A-E development of Dendraster excentricus from gastrula to pluteus under natural conditions;
F, G, pluteus of Arbacia punctulata; H, I, slight differential inhibition of Arbacta pluteus by de-
velopment in KCN m/ 100,000 {F-I from Child, igibb).

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. II 199

DIFFERENTIAL INHIBITION

In slight degrees of differential inhibition at early stages the oral lobe
is most inhibited, the angle between the anal arms (brachial angle) is
slightly decreased, and the foregut or esophagus is usually small (Fig. 73,
H, I); that is, apical and mid-ventral ectodermal regions are most in-

D

Fig. 74, ^-/^.—Differential inhibition in Arbacia. A-E, KCN; F-H, CUSO4 and LiCl
(A-E from Child, i9i6<^).

hibited. With greater inhibition there is further reduction of the oral
lobe, often to complete absence, and further decrease of brachial angle to
parallel, fused, or single median arms with skeletal rods parallel or even
converging toward the tips (Fig. 74). In Figure 74, A-E are cyanide
material, F-H, CUSO4 and LiCl, with return to water after exposure long
enough so that the differential inhibition persists, although development
may continue. With such procedure the differential inhibition has become

200

PATTERNS AND PROBLEMS OF DEVELOPMENT

irreversible, at least up to the most advanced stage attained. Similar
forms appear in acid sea water, sea water plus ammonia, ethyl alcohol,
hypotonic sea water, MgClz, etc., with certain concentrations and exposure
periods. Unfertilized eggs in KCN, m/ioo up to a few hours, then fer-
tilized and developing in water, give similar forms. The forms of Figure 74
show not only an apicobasal but also a ventrodorsal and a mediolateral

Fig. 75, A-I. — Differential inhibition in Arbacia, showing various degrees of obliteration
of the original ventrodorsaHty and the resulting bilaterality. A-D, ventrodorsality still
evident; E-H, bilaterality perhaps secondary and determined by transverse orientation and
elongation of skeletal rod; /, completely radial form without skeleton. Ciliated band shaded.

differential inhibition. It seems evident that decrease of brachial angle,
approximation and fusion of arms, and development of a single median
arm result from different degrees of ventral inhibition, decreasing from
the median region laterally and quite similar in principle to the mediolat-
eral differential inhibition of planarian head development (pp. 177 81).
A series of forms with further degrees of differential inhibition and de-
crease of axiate pattern is shown in Figure 75. These forms occur most
frequently with long-time exposures (e.g., 36 hours or more) from early

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. II 201

stages to LiCl in concentrations slightly above the range in which differ-
ential tolerance or conditioning is possible. The long period of exposure
makes the inhibition practically irreversible within the developmental
period, though there is some further development after return to water
and perhaps slight differential recovery, but differential inhibition is evi-
dently predominant. These forms, however, are not specific for lithium
but have been seen with other agents having a high differential action —
for example, CUSO4, HgCl2, and very similar forms, though with differen-
tiation more inhibited, are produced by KCN. In the individual of Fig-
ure 75, A and B, there is evidently extreme apical inhibition, the oral lobe
being completely absent; the skeletal rods converge toward what is ap-
parently the ventral side, but arms do not develop, and the ciliated band
is apparently more or less transverse. Figure 75, C and D, shows a more
extreme case of this type, with further approach to radial form but still
with some evidence of ventrodorsality in position of the skeletal rods. In
lots subjected to the same conditions forms like EI of Figure 75, also
occur; in these forms apicobasal pattern is almost completely obliterated
and only a single skeletal rod is present, apparently transverse and in the
same plane as the ciliated band, suggesting a possible relation. But
whether, or to what extent, position of arms 180° apart, when they de-
velop, is determined by prospective arm areas in the ectoderm or merely
by elongation of the skeletal rod is uncertain. Elongation of a rod can in-
duce development of an arm or armlike outgrowth in other than the posi-
tion of the original arm area. In forms with excess of skeleton supernu-
merary arms are often formed, three-, four-, and five-armed forms result-
ing; and in the apparently anaxiate ectoderm of extreme exogastrulae a
skeletal rod may induce a short armlike outgrowth in various positions,
even at what was originally the apical pole. It seems possible, therefore,
that the arms of forms like E and F of Figure 75 result from skeletal
elongation. If this is the case, they may be without definite relation to the
original ventrodorsality and constitute a bilaterality independent of it.
With less elongation of the skeletal rod bilaterality is less evident (Fig. 75,
G, H). That the original ventrodorsality and the resulting bilaterality
may be completely obliterated is suggested by absence of stomodeum and
by completely radial form of entoderm. These forms are very different
from the wide-angled forms resulting from secondary modifications de-
scribed in the following section and appear only under rather extreme
inhibiting conditions, but they probably do involve some slight recovery,
for development of ciliated band, skeleton, and arms occurs only after

202 PATTERNS AND PROBLEMS OF DEVELOPMENT

return to water. In the absence of skeletal development completely radial
individuals with apparently transverse ciliated band are frequent in the
same experimental lots (Fig. 75, /). This series of forms has been obtained
with Arbacia, Strongylocentrotus , and Dendraster.

As degree of inhibition increases, there is progressively less evidence of
axiate pattern, the skeleton is represented by a few spicules or does not
develop, and mesenchyme cells are scattered instead of being localized
bilaterally in the regions of arm development. The particular form of
larva resulting depends largely on the stage at which exposure begins and
the rate at which inhibition occurs in relation to rate of development.
With relatively rapid inhibition, beginning at the two-cell stage and re-
turn to water before the increase in susceptibility of entoderm associated
with gastrulation, ectodermal development may attain sKght ventrodor-
sality (Fig. 76, yl) or be completely radial, and more or less regional ento-
dermal differentiation may occur. Although these forms probably repre-
sent a slight degree of differential recovery, they are mentioned here be-
cause they represent chiefly considerable degrees of differential inhibition.
Other somewhat similar forms are described below under differential re-
covery. With sufficient inhibition involving the early gastrula entoderm
is inhibited, develops incompletely, or may become spherical, separate
completely from the closed blastopore, and lie free in the blastocoel.
Ectoderm may show slight polarity or be completely anaxiate (Fig. 76,
B, C). Forms of this sort, returned to water, may live for 2 weeks or more.
Completely anaxiate forms like Figure 76, C, do not show definitely di-
rected locomotion but roll about in all directions. With still more extreme
inhibition in early blastula stages more or less cytolysis and disintegration
progress basipetally from the apical pole. With return to water at the
proper time the disintegration is arrested, the cells come together apically,
and invagination of entoderm may take place (Fig. 76, D, E). Some of
these forms remain completely anaxiate; others show slight differential
recovery (pp. 208-10). With these relatively extreme inhibiting condi-
tions continuing from early stages to the stage when entodermal suscepti-
bility increases, the entoderm tends to lose its epithelial character and be-
come a solid cell mass, from the surface of which cells dissociate. This
dissociation may begin either without invagination or during its early
stages (Fig. 76, F-I). Cells may also dissociate from regions apical to
the prospective entoderm, in some cases from all levels, as in Pkialidium
(pp. 167-69). In many of these cases, particularly with LiCl but also
with other agents, there is entodermization of prospective ectoderm; that

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. II 203

is, a part, or in extreme cases all, of the prospective ectoderm becomes
entoderm. If such forms are able to develop further after return to water,
they become exogastrulae with small or no ectoderm and large external
entoderm. With exposure to LiCl beginning in the later blastula stages.

Fig. 76, A-K. — Differential inhibition. A-E, Arbacia, KCN, ethyl alcohol, etc. A, slight
axiation remaining; B, ventrodorsality not evident, apparently slight polarity; C, completely
anaxiate; D, E, partial differential death, apical region of early blastula killed, development of
basal region, but E anaxiate; F-I, Dendraster, extreme differential inhibition in high concentra-
tions of LiCl (e.g., m/20) continuing from two-cell stage to stage when susceptibility of prospec-
tive entoderm becomes higher than that of ectoderm; prospective entoderm and more or less of
the entodermized ectoderm becoming a solid cell mass from which cells dissociate externally or
internally or both; J, K, Strongylocentrotus franciscanus, showing inhibition of entoderm with
exposure to LiCl (m/40) beginning in later blastula stages {A-D, after Child, igiCxi; H, I,
from Child, 1940; /, K, from Child, 19366).

about the time of increase in susceptibility of the entoderm, not only may
further entodermal development be inhibited (Fig. 76, J) but entoderm
may decrease in extent and thickness so that more or less of it becomes
indistinguishable from ectoderm (Fig. 76, K). Forms of this type suggest
that some of the prospective entoderm has been ectodermized. If this is

204 PATTERNS AND PROBLEMS OF DEVELOPMENT

the case, it appears that the same inhibiting agent may entodermize pro-
spective ectoderm in early stages, when it represents higher gradient-levels
than entoderm, and ectodermize prospective entoderm in later stages,
when it has attained a higher gradient-level than ectoderm. Further
examples of this differential inhibition will be given in the section on
exogastrulation.

The so-called "sterroblastula," appearing so frequently under widely
varied inhibiting conditions and often in pathological lots of eggs without
external inhibition, is a blastula in form with blastocoel more or less com-
pletely filled with cells. It represents a differential inhibition with dis-
sociation of cells from the entodermal region and perhaps also from other
parts of the wall. Apparently, all that is necessary to produce sterroblas-
tulae is a sufficient degree of inhibition. In general, the developmental
modifications resulting from differential inhibition differ with concentra-
tion or intensity of agent, with exposure period, and with stage when ex-
posure begins; but there is no conclusive evidence of specific regional
effects of particular agents: the differences appear rather to be indicative
of quantitative factors in physiological condition and their changes in the
course of development than of specific factors.

SECONDARY MODIFICATIONS: DIFFERENTIAL TOLERANCE,
CONDITIONING, AND RECOVERY

These secondary modifications of development follow an initial inhibi-
tion; they are much greater and occur much more rapidly with some
agents than with others. With cyanide, for example, they occur slowly
and are not extreme; with ethyl alcohol, acidified sea water, hypotonic
sea water, and various other agents they appear earlier and may be ex-
treme; with CUSO4 and MgCl^ they appear somewhat more slowly but
may still be great. ^

With slight inhibition in early stages the first evidence of differential
tolerance or conditioning is relative elongation and enlargement of the
apical region; this may appear in the gastrula with slight inhibition by
agents such as ethyl alcohol, acidified or hypotonic sea water, and various
others (Fig. 77, A,B), but with many agents it becomes evident somewhat
later. In the prepluteus the oral lobe becomes relatively large and long
(Fig. 77, C), and the resulting pluteus has a large oral lobe and a wide
brachial angle up to 180°, a relatively large ventral, and a small dorsal

" Arhacia has been the chief material in these experiments, but data on Strougylocentrotiis
and Dendraster are sufticient to show that the secondary modifications in these species are
essentially similar.

Fig. 77, ^-/.—Secondary modifications resulting from differential tolerance or condition-
ing and differential recovery. A-G, Arbacia; H-1, Dendraster. A, B, secondary modification
of apical region of gastrula in ethyl alcohol 2 per cent; C, prepluteus with elongated apical
region (oral lobe), a secondary modification characteristic with various agents; D-C, two
plutei in anal {D, F) and lateral (£, G) outline with oral lobe and ventral regions relatively
enlarged (characteristic with various agents); H, differential tolerance or conditioning with
LiCl m/90; /, J, differential recovery after 4^ hr. in LiCl m/30 from early blastula.

206

PATTERNS AND PROBLEMS OF DEVELOPMENT

region. Comparison of the Arhacia plutei of Figure 77, D-G, with those
developing under natural conditions (Fig. 73, B, E) shows the alteration
in proportions. Similar modifications with low concentrations of LiCl ap-
pear in Dendraster: Figure 77, H, is an example of differential tolerance or
conditioning to lithium, and / and / of Figure 77 are cases of differential
recovery from lithium: pluteus form under natural conditions is shown
in Figure 73, F. In all these wide-angled plutei the ciliated band has the
usual relation to oral lobe and arms, and the characteristic skeletal pat-
tern develops but with alterations in proportions corresponding to those
of other parts.

Fig. 78, A-E. — More extreme secondary modifications in Arhacia with conditioning to,
and recovery from, alcohol, acid sea water, CuSO^, HgCL, etc. A-C, ventral, anal, and lateral
outlines of an individual; D, E, ventral and lateral outlines of another individual without
skeleton (after Child, I9i6fc).

With somewhat greater degrees of inhibition the secondary modifica-
tions are greater. Characteristic forms of Arhacia are outlined in Fig-
ure 78. These larvae may be said to consist largely of oral lobe; this is
much enlarged and elongated relatively to other parts but still shows
inhibition in later differentiation. The larvae are flattened ventrodorsally,
and arms are more or less inhibited or absent. The skeleton may be lim-
ited to a single rod, as in Figure 78, A-C, or the brachial rods of the two
sides may be separate; and in some animals there are other spicules or
the skeleton may be entirely absent (Fig. 78, D, E). Here the secondary
modifications of ectoderm are evidently apicoventral and decrease basip-
etally. The usually large foregut suggests secondary modification in the
apical entodermal region, the high end of the entodermal gradient. The
skeleton is more or less inhibited or absent, probably because activation

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. II 207

of mesenchyme at the time of immigration is more or less inhibited; and
before tolerance, conditioning, or recovery occurs, it is too late for normal
skeletal development. These forms also are not specific for any particular
agent, but their appearance in a given species depends on concentration
or intensity, developmental stage, and period of exposure. Although the
skeleton may consist of a single rod transversely oriented, these forms are
of quite different origin from those of Figure 75, E-H, which are results

Fig. 79, A-G. — Secondary modifications in Arbacia. A, B, lateral and basal outlines, con-
tinuous in alcohol, 1.5 per cent, with apical region secondarily modified, basal region differ-
entially inhibited; C, D, lateral and basal outlines, CUSO4 m/2, 500,000, continuous; E-G,
secondary modifications in apical region only, ventrodorsality obliterated, CuSO^ m/2, 500,000,
continuous.

of relatively extreme inhibition with almost complete or complete oblitera-
tion of ventrodorsality and decrease of polarity.

The most extreme secondary modifications resulting from differential
conditioning with continuous exposure to low concentrations of agent have
been obtained with ethyl alcohol and CUSO4. In Figure 79, A and B
(alcohol, 1.5 per cent) and C and D (CUSO4 m/2, 500,000) are examples.
In these the elongation and enlargement of the apical region is secondary,
early development being inhibited. Inhibition persists in the basal region,
arms being completely absent and the skeletal rods oriented at less than
the usual angle {B), or arms are present and greatly inhibited, but the
angle of the skeletal rods is slightly wider than normal, suggesting slight

2o8 PATTERNS AND PROBLEMS OF DEVELOPMENT

secondary modification ventrally {D). In somewhat more inhibited indi-
viduals secondary modification may be limited to the apical region. In
Figure 79, £, F, and G, relatively extreme examples are shown of a type
very common with certain ranges of concentration and exposure for all
agents used. Ventrodorsality is obliterated in the body, and a transverse
ciliated band surrounds the basal region. •^ This altered position of the
band is characteristic of these forms and also appears in many exogastrulae
(pp. 224-25). Apparently the more apical levels of the ventral side are
so much inhibited that the band cannot develop there; but with secondary
modification in the less inhibited basal region and the altered relations of
regions it differentiates there, in part with localization different from the
normal. However, the tip of the secondary apical outgrowth often de-
velops the specialized ciliated epithelium characteristic of the band.

Inhibition beginning just before, or at the earliest stages of, entodermal
invagination and within the range permitting some degree of secondary
modification gives results of interest. At these stages entoderm and mes-
enchyme have undergone or are undergoing increase in susceptibility and
rate of dye reduction. Conditioning or recovery occurs in the apical re-
gion, and entoderm and mesenchyme are more or less inhibited. These
modifications appear with the various agents used, even in a certain range
of concentrations of LiCl (Fig. 80). In earlier stages lithium may ento-
dermize prospective ectoderm, as already noted, and so increase entoderm
at the expense of ectoderm; but at this stage entoderm has become so
highly susceptible that it is inhibited, while the apical ectoderm, at first
inhibited, gradually becomes thicker and elongates. The lithium forms
of Figure 80 do not usually develop much farther with continued exposure.
No skeleton is formed, the mesenchyme, not shown in Figure 80, remain-
ing scattered or irregularly massed in the blastocoel. Incidentally it may
be noted that these forms with relative enlargement of the apical region
suggest an approach to asteroid larval form with its large preoral region
(Child, 1938). The apicobasal dye-reduction gradient in early stages of
the starfish appears to be steeper than that in the sea urchin (Child,
1936a), and these differential modifications involve an increase in steep-
ness of the gradient, at least at the more apical levels.

Lesser degrees of differential tolerance or conditioning are outlined in
Figure 81, and of differential recovery after temporary exposure in

3 In some individuals the band develops as two separated parts symmetrically localized on
the two sides, indicating that the bilateral features of ventrodorsality have not been completely
obliterated.

Fig. 8o. — Secondary modifications of gastrula of Dendraster with 40 hr. of exposure to
Li CI m/60 from beginning of gastrulation; apical region secondarily elongated, entoderm
inhibited (from Child, 1938).

Fig. 81. — Lesser degrees of secondary modification in Arbacia limited to apical region with
continuous exposure to various agents (from Child, igi()d).

2IO

PATTERNS AND PROBLEMS OF DEVELOPMENT

Figure 82. The two series are very similar in that the secondary modifica-
tions are limited to the apical region. Position of the mouth at one side

Fig. 82. — Slight differential recoveries in Arabacia after relatively extreme inhibition by
various agents (from Child, i9i6(/).

in some, presumably the ventral side, of the apical outgrowths suggests
that a trace of ventrodorsality is still present apically; but there is no
evidence of it elsewhere. In others, however, the mouth is apparently
apical, suggesting possible induction of a mouth by the entoderm. Degree

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. II 211

of development of entoderm is highly variable: in some, all three regions
are present; in others, two; while in still others it is a spherical vesicle.
Very often it separates from the region of the blastopore, of which no
trace remains, and usually it comes to lie near the secondarily modified
apical region. The secondary apical modification consists in outgrowth,
often with development of ciliated band epithelium, in development of a
mouth, or occasionally only in apparent attachment of entoderm to the
apical ectoderm. If a ciliated band develops, it is basal. The blastocoel
contains cells in varying number (not figured), the mesenchyme, and in
many individuals cells dissociated from prospective entoderm, but no
skeleton develops. Figures 81 and 82 represent the most advanced stages
of these forms, but they often live longer without further development
than fully developed plutei from the same lots of eggs."*

ASTEROIDS

Differential modifications of asteroid development show the same rela-
tions as those of echinoids to the gradient pattern indicated by other
methods. In differential inhibition apical and ventral regions of ectoderm
and apical region of entoderm are most inhibited, and in the secondary
modifications these regions become relatively large.

EARLY LARVAL FORM UNDER NATURAL CONDITIONS

The gastrula of the starfish Patiria miniata was outlined in Figure 49,
C-E, of chapter iv; a later stage is shown in Figure 83; development of
Asterias is similar. As a basis for consideration of the differential modifi-
cations it is important to note the following points of difference between
asteroid and echinoid development: first, in the asteroid the mesenchyme
does not immigrate before gastrulation but is first formed from the apical
entoderm after invagination is completed and increase in rate of dye re-
duction in the enlarged apical entoderm has occurred (Child, 19366);
second, the invaginated asteroid archenteron extends only about halfway
from the basal to the apical pole; third, instead of becoming a flattened
oral lobe, as in echinoids, the apical region of the starfish develops into a

4 These types of secondary modification have been obtained with alcohol, acid and hypo-
tonic sea water, CuSO^, HgCL, ultra-violet radiation, and visible hght with sensitization by
eosin. Differential tolerance or conditioning to KCN is shght, but differential recovery gives
the whole series of forms after certain ranges of concentration. Similar secondary ectodermal
modifications occur with LiCl, but very commonly in association with some degree of exogas-
trulation.

212

PATTERNS AND PROBLEMS OF DEVELOPMENT

rounded preoral region, empty except for mesenchyme; fourth, two ciH-
ated bands extend obhquely around the body, as indicated in Figure 83.

DIFFERENTIAL INHIBITION

Experiment on asteroids is less extensive than on echinoids, but a num-
ber of agents have been used to produce modifications.^ Not the sHghtest
evidence has appeared in more than a hundred experimental lots of spe-
cific relation between any of these agents and any particular modification,
not even between lithium, exogastrulation, and entodermization of pro-

FiG. 83.— Lateral and ventral outlines of larva of Paliria miniata; ciliated bands indicated
by broken lines (from Child, 1938).

spective ectoderm. Eggs from animals kept in laboratory aquaria sup-
plied with flowing sea water for 9 or 10 days before use develop differen-
tially inhibited forms identical with those obtained with cyanide and
lithium, etc. With exposure to the agent in inhibiting concentrations at
beginning or early cleavage inhibition at first decreases basipetally, as in
echinoids. In differentially inhibited blastulae the polar axis is short and
the apical ectoderm thicker than normal. If the inhibition is not too ex-
treme to permit gastrulation, these blastulae, returned to water shortly
before gastrulation, or gastrulating before return, form gastrulae like
those of Figure 84. In all of these ectoderm is more inhibited than ento-

5 The following agents have been used: LiCl and KCN in many concentrations; hypotonic
sea water; the dyes neutral red, Janus green, and Nile blue sulphate; crowded conditions; and
sea water at pH 7, with CO, probably the cflective agent.

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. II 213

derm, which is relatively or absolutely "too large." In the first gastrula
of Figure 84, 24 hours in water after LiCl, the thin apical ectoderm and
the extremely large apical entoderm suggest some degree of differential
recovery.

As invagination progresses, entoderm becomes more susceptible than
ectoderm, at least in its more apical levels, and when sufficiently in-
hibited it begins to lose its epithelial character and to undergo dissocia-
tion. The lesser degrees of this change resemble immigration of mesen-
chyme in echinoids, but dissociation may involve the whole entoderm

Fig. 84, A-E. — Differentially inhibited gastrulae of Patiria. A, LiCl m/25, 11 hr. from
sixty-four cells, 24 hr. water; B, LiCl m/30, 14 hr. from sixty-four cells, 26 hr. water; C,
LiCl m/2S, 195 hr. from thirty-two cells, 4 days water; D, LiCl m/25, n hr. from sixty-four
cells, 24 hr. water; E, KCN m/200,000 from two cells; perhaps some differential recovery in
apical archenteron of A and B.

(Fig. 85). In the case of Figure 85, A, exposure to a very high concentra-
tion of LiCl (m/io) began just before gastrulation ; some invagination
occurred before the agent became fully effective, but gradually the apical
entoderm began to lose epithelial order and to dissociate, and the change
progressed basipetally but the ectoderm remained intact. This was the
characteristic effect in the experimental lot. In Figure 85, B, from another
lithium lot, the invaginated entoderm has become a solid cell mass and is
dissociating. These results, characteristic of lithium in sufficient concen-
tration, with exposure beginning just before gastrulation (A) or continu-
ing into that stage, indicate the change in condition of the entoderm
associated with gastrulation. As in echinoids, lithium in early pregastru-

214

PATTERNS AND PROBLEMS OF DEVELOPMENT

lar stages may entodermize prospective ectoderm, but later it inhibits
and dissociates entoderm; the increase in entodermal susceptibility ap-
parently occurs at a somewhat later stage in asteroids than in echinoids,
as does the increase in rate of dye reduction (pp. 134-37)- Inhibition be-
ginning in early stages may prevent gastrulation, but with long-continued
exposure dissociation into the blastocoel of entodermal cells may take
place without invagination; apparently, increase of entodermal suscepti-
bility is possible without invagination.

Fig. 85, .4-F.— Loss of epithelial character and dissociation of the entodermal region in
Patiria with relatively extreme differential inhibition continuing to gastrulation or later.

Different degrees of entodermal dissociation with more extreme inhibi-
tions are shown in Figure 85, C~F; in C there is some invagination, and
dissociation is beginning apically; in D invagination apparently began,
but the whole prospective entoderm, except for a thin external layer, has
lost its coherence; in E and F gastrulation is completely inhibited, and the
cell immigration from the entodermal region resembles immigration of
mesenchyme in echinoids; there is apparently some entodermization in
E and in F. In Figure 85, A and B are effects of LiCl, C is from a lot in
KCN, D from sea water at pH 7, £ from another lot in KCN, and F from

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. II 215

a crowded lot; loss of epithelial character and dissociation of entoderm
are evidently not specific effects of a particular agent : they depend on con-
centration and exposure period. The apparent asymmetry of C and D
results from ventrodorsal differential inhibition. The inhibiting conditions
may completely prevent development of ventrodorsality (Fig. 85, E, F);
but if it develops, inhibition decreases from the ventral region dorsally.
That it is actually the ventral region that is more inhibited cannot be de-
termined from forms like C and D, but many individuals undergo suffi-
cient development to permit certain identification of ventral and dorsal
regions.** Gastrulae such as B and D of Figure 85 usually show little or
no further development; but forms like A, C\ E, and F may continue

Fig. 86, A-D. — Slight secondary modifications in Patiria after relatively extreme inhibi-
tion. A, KCN m/50,000, 18 hr. from two-cell stage, 11 days water; B, LiCl m/20, 16 hr.
from two-cell stage, 8 days water; C, LiCl m/30, 38J hr. from early cleavage, 8 days water;
D, LiCl, m/25, 19J hr. from early cleavage, 5 days water.

development as differentially inhibited forms or show some secondary
modification, according to experimental conditions.

The question arises whether the cells dissociating from the entodermal
region represent premature mesenchyme, and, if so, how it is possible
that inhibiting factors determine their appearance. The data on dye re-
duction indicate low oxygen in the blastocoel ; in the starfish it is so low
that the apical region of the archenteron, from which mesenchyme nor-
mally dissociates, reduces both methylene blue and Janus green in water
of the usual oxygen content and open to the air (Child, 1936a). It is sug-
gested that under inhibiting conditions which decrease oxygen or inhibit
its intracellular utilization, dissociation may occur without activation of
the region concerned or with slight activation. But whether dissociated

^ Forms like those of Fig. 85 also appear in low oxygen, in hypotonic sea water, with Janus
green, and with ultra-violet radiation.

2l6

PATTERNS AND PROBLEMS OF DEVELOPMENT

cells function as mesenchyme or remain as inhibited cells is not known,
except for the fact that some of them may form epithelial vesicles in
echinoids (pp. 238-39).

SECONDARY AND COELOMIC MODIFICATIONS

Forms resulting from temporary exposure in early stages to severe in-
hibiting conditions, but with slight secondary modifications, are shown
in Figure 86. In A the only indication of possible differential recovery is

Fig. 87, A-E. — Differential recovery in water after differential inhibition. A, B, Patiria,
LiCl m/20, 10^ hr. from early blastula; C-E, Asterias; C, D, KCN m/200,000, continuous from
one-cell stage but with gradually decreasing concentration after 2 days; E, Patiria, LiCl m/30,
21 hr. from two-cell stage.

the elongation, the earlier gastrula being flattened apically; B shows sim-
ilar elongation and an apical mouth; slight apical thickening and out-
growth appear in C, but there is complete breakdown of entoderm, except
a thin outer layer; apical recovery is evident in D. With somewhat less
extreme inhibition further development and more or less differential re-
covery follow return to water, but ventrodorsality may be completely or
almost completely obliterated. In Figure 87, A and B, lithium forms of
Patiria, have apparently completely radial ectoderms, but the form of
the enteron in A suggests that some slight ventrodorsality is still pres-
ent or was present at a stage affecting entodermal development. In B
there is no evidence of ventrodorsahty ; and C, D, and E, modifica-

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. II 217

tions of Asterias, are apparently completely radial. In E differential re-
covery involves only the extreme apical region and perhaps the foregut.
Similar forms are produced by other agents with certain concentrations
and exposure periods; they also appear in eggs from animals kept too
long under laboratory conditions. All of them represent some degree of
secondary modification, differential tolerance, conditioning, or recovery
of a more or less extensive apical region, for they appear only in the lower
inhibiting concentrations or after return to water, and they develop from
gastrulae with inhibited apical ectoderm.

The development of one or two ciliated bands in some of these forms
is of interest (Fig. 87, B, C, D). As in many echinoid ectoderms inhibited
to radial form, the bands have become transverse circles, that is, their
localization is different from the normal. A few individuals with three
bands, at least ventrally, have been observed (see Fig. 89, H). This locali-
zation of ciliated bands suggests approach to the larval pattern of the
crinoid with five bands and little development of ventrodorsahty in larval
form.

Final stages of differentially inhibited forms with ventrodorsahty and
some degree of apical recovery are shown in Figure 88, A-D. In A, B,
and C the whole preoral region is represented only by an ectodermal
thickening, in D by secondary apical outgrowth; entoderm shows little
or no recovery, the foregut being relatively small {A, B, C) or not differ-
entiated (D). When these forms develop a ciliated band, it is transverse
(A, B). Special interest attaches to these modifications because they ap-
proach in form and proportions the echinoid prepluteus (see Fig. 73, B, C).
Since the apicobasal gradient of the starfish is apparently steeper than
that of the echinoid in earlier stages (Child, 1936a), it is perhaps signifi-
cant that with decrease of its steepness by differential inhibition there is
approach to echinoid form and proportions. These types evidently result
from a relatively high degree of differential inhibiting action between
apico ventral and other regions. They have been seen most frequently
with LiCl, which evidently has a high differential action, but they also
appear with other agents.

With somewhat greater degree of differential recovery forms like E and
F of Figure 88 appear, often in the same lots as those of Figure 88, A-D.
In a given lot animals swimming free or at the surface show, in general,
greater recovery than those at the bottom or in aggregations. When the
primary inhibition is slight, forms resulting from differential recovery are
of the general type of Figure 88, G. The preoral region is enlarged; mouth

2l8

PATTERNS AND PROBLEMS OF DEVELOPMENT

and foregut are usually relatively large ; and the ventral surface is almost
or quite flat, rather than concave. All degrees of this modification occur

Fig. 88, A~I. — Secondary modifications of Patiria following differential inhibition. A-D.
forms with slight apical recovery resembling echinoid preplutei in form, developing in water
after LiCl m/20, m/25, m/30, 10-21 hr. from early cleavage; £, differential recovery after
21 hr. LiCl m/30 from two-cell stage; F, differential recovery after 145 hr. LiCl m/30 from
si.xty-four cell stage; G, differential recovery after 21 hr. in 80 per cent sea water from two-cell
stage; //, differential conditioning in 90 per cent sea water 13 days; /, secondary modification
in Nile blue sulphate, initial concentration 1/1,000,000.

in recovery after exposure to all inhibiting agents used, when inhibiting
action is not too great. The modifications resulting from differential tol-
erance or conditioning with continuous exposure to low ranges of inhibit-

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. II 219

ing action are, in general, similar in type to the differential recoveries but
are usually, as might be expected, less extreme and with a degree of sec-
ondary modification differing with different agents. For example, second-
ary modifications are slight in cyanide; but in various other agents, even
in LiCl, they may be considerable or even extreme. A characteristic form
developing in 90 per cent sea water is shown in Figure 88, H; the preoral
region is relatively very large and almost spherical, the postoral region
and entoderm are still small. Forms of this type develop from gastrulae
with slight apical inhibition. In low concentrations of neutral red and
Nile blue sulphate similar forms develop. Figure 88, /, is an example from
a Nile blue sulphate lot in which the dye gradually disappeared, both
from the solution (open to the air) and from the animals. Evidently there
is a differential in ability to dispose of the dye in some way. Entoderm
stains more deeply than ectoderm and may remain differentially inhibited,
as in the figure, or show some secondary modification apically.

Of the two starfish coeloms, originating as localized budlike outgrowths
on right and left sides of the foregut and separating from it as small
vesicles, the left is usually the larger, or becomes larger, gives rise to the
madreporic canal, and becomes the hydrocoel. Reversal or obliteration
of this asymmetry, as an occasional occurrence under natural conditions,
has been noted repeatedly and induced experimentally; and single coe-
loms, either dorsal or ventral, have also been observed in sea urchins
under experimental conditions.'^ These modifications of coelom develop-
ment occur in Patiria under all inhibiting conditions thus far used by the
writer. They are undoubtedly to be regarded as cases of differential in-
hibition. When the enlarged apical region of the archenteron extends ven-
trally toward the stomodeal region, growth in length is much greater on
the apicodorsal side than on the other. This is evident in the form of the
foregut in later stages (Fig. 83, A). With differential inhibition of this
development the coeloms tend to appear nearer the median line on the
dorsal side. All degrees of approximation to the median region appear
from a slender connection between the two (Fig. 89, A) to development
of a single median sac, bilaterally symmetrical with two madreporic canals
(Fig. 89, B) or with a single median dorsal canal (Fig. 89, C). With con-
tinuous exposure to inhibiting conditions the differential elongation of
the foregut may be completely inhibited so that it remains a blind,
rounded, or somewhat elongated region at the apical end of the archen-
teron. If coelomic development occurs in such cases^ it is usually apical

"> Newman, 1925, Patiria, low temperature; see also Newman, 1921a, b, 1922; Runnstrom,
19256, c, Psammechinus, hypotonic sea water.

220

PATTERNS AND PROBLEMS OF DEVELOPMENT

and may show some bilaterality (Fig. 89, E) or be single and median
(Fig. 89, F). In some cases when the foregut elongates, a single coelom
develops on the ventral side (Fig. 89, G, H), between it and the stomodeal
region. The various degrees of differential modification from bilateral to
median origin of the coeloms with approximation, partial fusion, or single
median coelom with bilaterality obliterated present a series of differential

Fig. 89, A-H. — Modifications of coelom development by differential inhibition in Patiria;
outline of foregut and part of midgut with coelom: A, B, D-F, from dorsal side, C, G, H, from
left side. A, right and left connected, asymmetry evident; B, single symmetrical and median
with two canals, both 1 1 hr. LiCl m/25 from thirty- two cells; C, single median dorsal with medi-
an dorsal canal, LiCl m/40, continuous from two-cell stage; D, single symmetrical without
canal, LiCl m/30, 2ii hr. from two-cell stage; E, F, apical (anterior), LiCl m/40, continuous
from two-cell stage; G, single median ventral, between foregut and stomodeum, eggs from
animals long time in laboratory; H, single median ventral, LiCl m/30, 38I hr. from si.xty-four
cells, undoubtedly some recovery but occurring late, after ventrodorsality of foregut appar-
ently reversed.

inhibitions similar to those of the planarian head, the anal arms of the
sea-urchin larva, and the approximation to the median line of eyes and
other bilateral organs of the vertebrate head. In all these cases the modi-
fications evidently depend on a mediolateral physiological differential, a
median region being more inhibited than lateral; and as inhibition in-
creases, lateral organs appear successively nearer the median line and
become median. Coelom development on the ventral side of the foregut
probably results from reversal of the ventrodorsal differential by inhibi-

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. II 221

tion, that is, the apicodorsal region is so strongly inhibited that only on
the less inhibited ventral side is coelom development possible. Oblitera-
tion and reversal of the normal coelomic asymmetry also represent dif-
ferential modifications and are characteristic results of certain degrees of
inhibition by all agents used.

It is possible that with sufficient inhibition of the more susceptible
region the normally less susceptible and presumably less active region is,
to some extent, physiologically isolated and undergoes activation, par-
ticularly after return to water. If the left coelom normally represents a
higher level of activity and is in some degree dominant, differential in-
hibition may permit sufficient activation of the right side to reverse asym-
metry. Dorsiventral differential inhibition of the foregut may bring about
physiological isolation and permit activation of the ventral side and de-
velopment of a coelom there after return to water. But whether physio-
logical isolation is or is not involved, it seems evident that localization of
coelom development on the foregut is determined by nonspecific factors,
quantitative differences in physiological condition, rather than by defi-
nitely localized, "coelom-forming substances." With recovery after rela-
tively extreme inhibition of the apical entodermal region three or even
four coelum buds may develop, usually equidistant from each other,
about the circumference of the radial or almost radial foregut. In these
cases asymmetry has been obliterated, there is little or no evidence of
ventrodorsahty, and the whole circumference of the foregut is appar-
ently equipotent for coelom. The question how the coelom buds are
localized about the circumference of the foregut in a particular case is
the same as that concerning locahzation of individual tentacles about the
circumference of the body in various hydroids. The small coelomic ves-
icles developing from the posterior region of the midgut also show change
in position, number, and asymmetry under inhibiting conditions and with
secondary modifications. In recoveries as many as six of them may de-
velop in a circle about the midgut.

ECHINODERM EXOGASTRULATION AND ASSOCIATED MODIFICATIONS

It was discovered by Herbst (1892, 1895, 1896a) that when early stages
of sea-urchin development were treated for a time, beginning soon after
fertilization or in early cleavage, with lithium (usually LiCl) in certain
concentrations the entodermal region evaginated at the stage of gastrula-
tion instead of invaginating. Another effect of lithium was differential
entoclermization of prospective ectoderm, progressing from basal ecto-

222 PATTERNS AND PROBLEMS OF DEVELOPMENT

dermal levels acropetally with increase in lithium effect, until in extreme
cases all the ectoderm had become entoderm and an organism consisting
only of entoderm and mesenchyme resulted. All gradations from the nor-
mal gastrula to these completely entodermized forms were found, accord-
ing to conditions; forms with partly evaginated, partly invaginated ento-
derm, entexogastrulae or exentogastrulae, were also described. Herbst was
at first inclined to regard exogastrulation as a more or less specific effect
of lithium; but he soon found that sodium butyrate and lack of mag-
nesium would also produce exogastrulation (Herbst, 1895, 1897). Since
Herbst's work exogastrulation has been produced by many investigators
in various species of sea urchins and several species of starfish not only
by lithium but by a great number of other agents, both chemical and
physical.^ Probably further experiment will add other agents to this list.
In the light of the data at hand it is evident that exogastrulation is not a
developmental modification determined by specific action dependent on
the chemical composition or specific physical effect of a particular agent
or agents and specific regional differences in egg or embryo. It is ex-
tremely improbable that all agents listed in the footnote act in the same
manner on particular regions of sea urchin, sand dollar, and starfish
embryos; but if they do not, the question at once arises whether exogastru-
lation may not be, like other modifications described in preceding sections,
a differential modification, depending on nonspecific alterations of a pri-
marily quantitative gradient pattern. « A recent study of differential dye

s E.g., exogastrulation has been produced by NaCl, HgCU, CUSO4, KCN, NiCU, and by
various other salts and salt mixtures (MacArthur, 1924; Waterman, 1932, 1934, 1938), by
crowding and acidified sea water (MacArthur, 1924; Child), by high and low temperatures
(Driesch, 1893; Waterman, 1934), by neutral red (MacArthur, 1924), by methylene blue and
Janus green (Child), by auxin, glycogen, and HCIO3 (Motomura, 1934), by tobacco smoke
(Child), by lack of oxygen (Child), by X-rays, and occasionally a few exogastrulae by respira-
tory stimulants (Waterman, 1934, 1938). It has been observed by Runnstrom (1933) that
lithium and CO are additive in effect. In crowding lack of oxygen is probably the effective
factor rather than CO,, for the modifications occur when pH has not decreased below 7.3 or
7.4; but it is possible that an excreted toxic metabolite may be concerned. In acidified sea
water CO^ is probably the inhibiting factor.

9 The writer's interest in the problem of exogastrulation, extending over many years, has
led, as opportunity offered, to a wide range of experiment, particularly with LiCl in concen-
trations ranging from the lowest effective to those completely inhibiting development and
killing rapidly and with exposure periods beginning at various developmental stages and rang-
ing from an hour or two to continuous, also some exposures to inhibiting agents before fertiliza-
tion. The following species have been used as material: echinoids: Arbacia piinctiilata,
Strongylocentrotiis franciscanns, S. purpuratus, Echinarachnius parma, and Detidraster excen-
tricus; asteroids: Asterias forbesii and Patiria niiniata. For the most recent general discussion
of echinoderm exogastrulation, a paper not included in the Bibliography, see Child, 1940,
"Lithium and echinoderm exogastrulation," Physiol. Zool., 13.

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. II 223

reduction in relation to exogastrulation suggested an interpretation of this
series of developmental modifications, differing somewhat from hypothe-
ses advanced by others (Child, 19366). Further experiment, chiefly with
Dendr aster as material, appears to throw some additional light on condi-
tions concerned in development of these extremely interesting forms.

ECHINOID EXOGASTRULAE: A FEW EXAMPLES OF FINAL STAGES

Developmental modifications of exogastrulae are essentially similar in
the echinoid species thus far used as material. The following account is
chiefly concerned with lithium exogastrulae of Dendr aster excentricus. Fig-
ure 73, A-E, showing forms of gastrula and later stages under natural con-
ditions, will serve for comparison with exogastrulae. A few of the almost
innumerable modifications of form and proportion in exogastrulae are out-
lined in Figures 90 and 91," most of them lithium forms developing with
different concentrations and exposure periods, beginning at, or soon after,
the first cleavage. Attention is called first to the differential inhibitions
of the ectoderm in these forms. With the lower concentrations or rela-
tively short exposures ectoderm may attain more or less completely
pluteus form (Fig. 90, A, B). Figure 90, C, a result of crowding, shows
somewhat less advanced ectodermal development. With greater inhibi-
tion ectoderm is smaller and oral lobe is almost completely inhibited, but
arms may develop at varying wide angles up to 180° (Fig. 90, D-F). As
will appear more clearly below, the basal part of the prospective ectoderm
of these forms has been entodermized by lithium, that is, has become
entoderm; consequently, the arms develop from ectoderm much farther
apical than the original arm area. These forms are similar as regards
ectodermal development to the inhibited, wide-angled modifications dis-
cussed above (pp. 200-201) and undoubtedly result from essentially similar
conditions. They appear in a wide range of concentrations according to
exposure period. More inhibited ectoderms develop parallel, fused, or
single arms with double- or single-arm skeleton, and in some of them a
stomodeum develops in a position relative to the arm, which indicates
that the arm is ventral and median (Fig. 91, yl), as in the one-armed forms

" These and following figures concerned with exogastrulation in Dendraster are drawn to
the same scale from ocular-micrometer measurements of living individuals and sketches of
the same individuals made at the same time. Levels of localization of mesenchyme are indi-
cated; but, as regards actual numbers of cells of mesenchyme or dissociated entoderm in the
blastocoel, the figures are diagrammatic, though they attempt to show whether free cells are
numerous or few and to represent loss of epithelial character and dissociation in entoderm.
Pigment cells are indicated, chiefly along the ciliated band, where they are most numerous
when it develops. Patiria exogastrulae in Fig. 95 are on a smaller scale.

224

PATTERNS AND PROBLEMS OF DEVELOPMENT

of Figure 74. With further increase of inhibiting action the ectoderm is
radial; the ciliated band, if it develops, is localized around the basal region
(Fig. 91, B-D); and if skeleton develops at all, it forms an irregular ring
of spicules, often subapical {B) instead of basal. In absence of skeletal
development a ring of mesenchyme may occupy the same position (C, D).

Fig. 90, A-F. — Less extreme types of Dendraster exogastrulae. A, LiCl m/90, continuous
from early blastula; B, LiCl m/40, 14 hr. from early blastula; C, crowded 10 hr. from early
blastula; D, LiCl m/25, 5 hr. from first cleavage; E, m/70, continuous from first cleavage; F,
LiCl m/60, continuous from first cleavage.

Further reduction of ectoderm is shown in Figure 91, E-G. Position of
the skeletal rods in E is probably not indicative of the original ventro-
dorsality, and form of the ectoderm appears to be determined by their
elongation. In F and G their direction is apicobasal; occasionally in such
forms a short armlikc outgrowth develops at the apical pole, apparently
induced by the skeleton. In Figure 91, H, the only part of the original

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. II 225

ectoderm remaining is the tiny button at the apical pole. A thin-walled
"neck" connecting external entoderm and ectoderm develops in many
exogastrulae, apparently always as a secondary modification, not as a
direct effect of lithium (Figs. 90, A-E, and 91, A-C, E-G), and apparently
of ectodermal origin. Its possible significance is discussed later.

Fig. 91, A-I. — More extreme types of Dendraster exogastrulae: final stages attained in
water after various exposures to LiCl. A, w/25, 7 hr. from first cleavage; B, m/30, i8§ hr.
from eight cells; C, m/25, 7 hr. from first cleavage; D, m/50, continuous from first cleavage;
E, m/30, 18^ hr. from first cleavage; F, m/25, 7 hr. from first cleavage; G, m/20, 9 hr. from
early blastula; H, I, m/20, 95 hr. from four cells.

Exogastrular entoderm also shows a wide range of modifications. The
original entoderm may invaginate more or less but is inhibited in further
development (Figs. 90, D, and 91, D); or, with more extreme inhibition,
it may lose epithelial character and dissociate more or less completely
and the exogastrular entoderm develops in large part, perhaps in some
cases entirely, from entodermized ectoderm (Figs. 90, F, and 91, E-H).
It is often impossible to determine how large a part of the total entoderm
is original prospective entoderm. On the other hand, with relatively slight

226

PATTERNS AND PROBLEMS OF DEVELOPMENT

inhibitions, all or almost all of the exogastrular entoderm may develop
from the evaginated original entoderm (Fig. 90, A-C). With the less ex-
treme inhibitions three entodermal regions may be more or less distinctly
marked off by constrictions (Fig. 90, J5, C). In recovery after greater in-

FiG. 92, A-I. — Echinoid exogastrulae with small entoderms, resulting from exposure to
LiCl beginning in later pregastrular stages or at beginning gastrulation. A-C, Detidraster;
D-G, Strongylocentrotus purpuratus; H, I, S. Jranciscanus. A, B, m/45, 39 hr. from 6-hr.
blastula; C, m/25, 24 hr. from beginning gastrulation; D-G, m/40, 20 hr. from late blastula;
H, I, m/40, 24 hr. or more from late blastula. Arrows in G and // indicate directions of prog-
ress of dye reduction in low oxygen and numerals i and /, that dye reduction becomes evident
at about the same time in the regions indicated {D-I from Child, 19366).

hibition development of two entodermal regions of large size is common,
often with a third region of small size or barely indicated (Fig. 91, A-C,
E-G). After extreme inhibition there is little or no regional differentiation

(Fig. 91,^)-

When exposure to lithium begins in late pregastrular stages or, in mod-
erate concentration, continues up to, or through, gastrulation, entoderm

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. II 227

Is generally more inhibited than ectoderm ; and if exogastrulation occurs,
the entoderm is usually relatively small (Fig. 92, A-C). Figure 92, C,
is an entexogastrula in which partial invagination preceded e vagina tion,
but entoderm is evidently inhibited in development. Forms resulting from
exposure of Strongylocentrotus to lithium in late pregastrular stages are
shown in Figure 92, D-I . It was suggested in an earlier paper (Child,
19366) that in cases like D-F of Figure 92 more or less ectodermization of
prospective entoderm has taken place. Cells in the blastocoel do not in-
dicate extensive dissociation of entoderm.

DEVELOPMENT OF ECHINOn) EXOGASTRULAE

The final forms of many exogastrulae give little information concerning
their developmental physiology: for such information it is necessary to
look to the earlier stages, as modified by different concentrations and ex-
posure periods. Since the pioneer work of Herbst it has been known that
lithium retards and, in sufficient concentration, completely stops develop-
ment and kills. That entodermization of prospective ectoderm may occur
in connection with exogastrulation was also observed by Herbst. Lithium
is believed by Runnstrom and his co-workers to have a specific action
on the postulated vegetative concentration gradient. In recent papers
it is maintained that lithium inhibits the animal kind of metabolism and
increases the vegetative kind." We have to inquire whether, or to what
extent, the development of exogastrulae supports this hypothesis.

Normally the primary mesenchyme after immigration and before bi-
lateral aggregation is localized in a more or less definite ring or circle on
the inner surface of the ectoderm near its junction with the prospective
entoderm. As is well known, this ring may be localized much farther api-
cally in developing exogastrulae; that is, the ectodermal level which local-
izes it is displaced apically. This shift apically of certain unknown con-
ditions in the ectoderm is usually associated with entodermization of the
more basal levels of prospective ectoderm, though slight apical shift is
apparently possible without appreciable entodermization. Lithium has
still another effect: it inhibits development of entoderm. That it ento-
dermizes ectoderm and also inhibits entoderm is not the paradox that it
may appear to be, but is very simply interpreted. And, finally, there is
evidence indicating that some ectodermization of prospective entoderm
may occur under certain conditions, particularly in recovery after return

" See, e.g., Lindahl, 1936, p. 339, etc., and references to earlier papers by Runnstrom;
also Runnstrom, 1928Z), i93Sa.

228

PATTERNS AND PROBLEMS OF DEVELOPMENT

to water. The maximum inhibiting and modifying effect of a given con-
centration of LiCl is attained only gradually, presumably according to
rate of penetration. High concentrations are more rapidly effective than
low ; and the effects of the higher concentrations, even with short exposure,
are more persistent than those of low, perhaps because intracellular con-
centration remains above the threshold of effectiveness for a longer time.
Within certain hmits longer exposure to the lower effective concentrations

Fig. 93, A-I. — Developmental stages of Dendr aster exogastrulae showing displacement
apically of primary mesenchyme {A-G), entodermization of prospective ectoderm {A-I), and
loss of epithelial character and dissociation of prospective entoderm (J5, D-F) and of basal
parts of entodermized ectoderm (G-I). A, LiCl m/50, 145 hr. from first cleavage, identical
forms in crowded lots; B, crowded 10 hr. from early blastula, 6 hr. water; C, LiCl m/30, 5 hr.
from first cleavage, 18 hr. water; D, E, LiCl m/60, 25 hr. from first cleavage; F, LiCl m/2S,
7 hr. from first cleavage, 14 hr. water; G, LiCl m/20, 9 hr. from first cleavage, 18 hr. water;
H, LiCl m/30, 18^ hr. from first cleavage; /, crowded 27 hr. from first cleavage.

gives modifications essentially similar to those resulting from shorter ex-
posure to higher concentrations.

Early stages of exogastrulae with different degrees of entodermization
and displacement apically of mesenchyme are shown in Figure 93. In ^
of this figure, a rather long exposure to a relatively low concentration, the
apical region is obviously inhibited, and almost the whole basal half of
the ectoderm has become entoderm. With continued exposure the original
entoderm may invaginate more or less but is inhibited in further develop-
ment, the entodermized ectoderm becomes the external entoderm of the
exogastrula, and ventrodorsality is usually obliterated in the remaining

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. II 229

ectoderm (Fig. 91, D). This figure represents the most advanced stage
attained with continuous exposure to LiCl m/50. Evidently both invagi-
nated entoderm and entodermized ectoderm are inhibited in development,
and some dissociation of entoderm has occurred. After return to water
at the stage of Figure 93, ^, ectoderm may develop more or less ventro-
dorsality (Fig. 90, D, E), and entoderm may show some further develop-
ment. Figure 93, B, shows entodermization resulting from crowding in
early stages; after return to water, forms of the general type of Figures 90,
C, result. Evidently, entodermization is not a specific effect of lithium.
Figure 93, C, a short exposure to rather high lithium concentration, shows
extreme entodermization and greatly inhibited invagination, even after
18 hours in water. With further recovery, however, ectodermal ventro-
dorsality appears in many individuals, the most advanced attaining ap-
proximately the condition of Figure 90, D, with all gradations from this
form to completely radial forms like Figure gi,D. In all, the later develop-
ment of original entoderm and entodermized ectoderm remains inhibited.

With higher concentrations or longer exposures the original entoderm
may lose, in part or wholly, its epithelial character and become a solid
cell mass from which cells dissociate internally or externally or both,
though in some individuals a thin external layer remains epithelial or re-
gains epithelial character after return to water (Fig. 93, D-G). Crowding
has a similar effect (Fig. 93, 5). With still more extreme inhibiting condi-
tions, either by long exposure or by high concentrations, loss of epithelial
character and dissociation progress acropetally in the entodermized ecto-
derm (Fig. 93, H, I). In the most advanced stages attained in recovery
from these effects more or less of the original entoderm may remain a
cell mass, internal or external, at the tip of the entodermized ectoderm,
or an external layer may constitute epithelium at the free end of the
exogastrular entoderm."

In these forms it is impossible to determine how much of the exo-
gastrular entoderm develops from the original prospective entoderm;
but the earlier stages, such as E-G of Figure 93, indicate that the ex-
ternal entoderm may consist mostly, perhaps in some cases wholly, of
entodermized ectoderm. Dissociated entodermal cells do not usually ap-
pear to take any part in further development, though it is uncertain
whether they may function as mesenchyme in some cases of recovery.
Usually, however, with inhibition sufficient to produce entodermal dis-
sociation, little or no skeleton develops.

'^ See Fig. 91, B, C, E, G, H , /, and probably also F; also Fig. 93, D, E, G.

230 PATTERNS AND PROBLEMS OF DEVELOPMENT

With these more extreme lithium ejffects mesenchyme is locahzed near,
or even at, the apical pole, indicating entodermization of most of the
ectoderm (Fig. 93, E-G)\ with complete entodermization it may not be
definitely localized (Fig. 93, H, I). Occasionally two circles of mesen-
chyme appear at different levels (Fig. 93, £), and in a crowded lot indi-
viduals with three circles have been observed

With concentrations of LiCl high enough to stop development in rather
early stages, particularly with sea-water solutions, which are hypertonic,
entodermization of ectoderm is apparently complete, and the basal region
becomes extremely thick, with partial or progressive obliteration of the
blastocoel (Fig. 94). Under these conditions many individuals become
spherical solid cell masses. With return to water before lethal effect, forms

Fig. 94, /I -Z).— Entodermization with continuous exposure to high concentrations of LiCl
from early cleavage. A, B, m/25, 26 hr.; C, D, m/20 in sea water (hypertonic).

like A and B of Figure 94 may become exogastrulae like Figure 91, E-H,
even though they appear earlier to be completely entodermized.'^

Exogastrulae of Strongylocentrotus, Arbacia, and Echinarachniiis, and,
so far as can be determined from the work of others, those of other
echinoid species, do not differ essentially from Dendraster exogastrulae.
Differences in proportions of certain parts, particularly in the secondary
modifications, apparently result from differences in steepness of the gradi-
ents and in degree of activation of entoderm in different species; but as
regards the general types of modification, there is a high degree of simi-
larity.

ASTEROID EXOGASTRULATION

General characteristics of asteroid and echinoid exogastrulation are
similar, but there are minor differences resulting from absence in the star-
fish of mesenchyme formation preceding gastrulation and from the later

'3 In the hypertonic solutions (Fig. 94, C, D) there is little or no dissociation, probably be-
cause of hypcrtonicity; but after return to water it may be e.xtensive or more or less complete.

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. II 231

increase in susceptibility and apparent activation of entoderm (pp.
134-39). Exogastrulae of the starfish Patiria (Fig. 95) show various de-

I

Fig. 95, A-I. — Lithium exogastrulae of Patiria. A-F, continuous exposure from two-
cell stage: A, m/30, 60 hr.; B, m/40, 77 hr.; C, D, m/40, 95 hr.; E, F, m/40, 94 hr., slight
secondary modification apically. G-I, differential recoveries: G, H, m/30, 18 hr. from first
cleavage, 2 days water; /, m/20, 19 hr. from first cleavage, 3 days water; H and /, e.xentogas-
trulae. Arrows indicate directions of progress of dye reduction in low oxygen, and numerals
I and 2 the time-order in which reduction becomes evident in different regions (from Child,
19366).

grees of entodermization of prospective ectoderm; cells often dissociate
from entoderm into the blastocoel, resembling the primary mesenchyme of
echinoids (Fig. 95, A-C, E, G); and the apical ectoderm, when strongly

232 PATTERNS AND PROBLEMS OF DEVELOPMENT

inhibited, also gives off cells (Fig. gs, A, C, D, I). Thus far, little differ-
entiation has been seen in the remaining ectoderm of starfish exogastrulae,
even in recovery. In forms with considerable entodermization the remain-
ing ectoderm represents more or less of the prospective preoral region,
but there is no evidence that it reconstitutes a more or less complete
larval ectoderm. It may become rounded and in some individuals an
apical thickening appears as a secondary modification (Fig. 95, E, F).
Extensive, perhaps complete, entodermization of ectoderm and loss of
epithelial character in the prospective entoderm results in forms like Fig-
ure 93, G, of Dendr aster, an effect of long-continued exposure in both;
and more or less entodermal dissociation may occur in later gastrula
stages without entodermization or exogastrulation (see Fig. 85). Whether
dissociated cells of the prospective entoderm may function as mesenchyme
is not known; but here, as in echinoids, they appear to be merely in-
hibited cells without particular function.

In the asteroid, as in echinoids, lithium entodermizes prospective ecto-
derm and, after the increase in entodermal susceptibility occurs, also in-
hibits entoderm. Exogastrulation has also been produced in Patiria by
crowding and by acidified sea water, by KCN, and by Janus green; in
Asterias by lack of oxygen. With all these agents there is apparently more
or less entodermization of prospective ectoderm. Probably still other
agents will produce exogastrulation and entodermization in starfish, as in
echinoids.

EXOGASTRULAE AND PSEUDO-EXOGASTRULAE

In this discussion of exogastrulation all forms with external entoderm
have been called "exogastrulae," but they are not all alike in origin.
Strictly speaking, exogastrulae are modifications in which the entoderm
evaginates instead of invaginating. In many so-called "exogastrulae,"
however, the original prospective entoderm invaginates, and the external
entoderm is entodermized ectoderm (Figs. 90, D, and gi, D). These have
been called "entexogastrulae," or, when the invagination is secondary,
"exentogastrulae" (Fig. 95, H, I). In many cases, also, the prospective
entoderm takes little, perhaps no part, in formation of the external ento-
derm but loses epithelial character and remains a cell mass, usually with
more or less complete dissociation, while the external entodermal epithe-
lium is, in large part or wholly, entodermized ectoderm, as in Figures 91,
B, C, E, G, and I, and probably also in F and H. The entodermized ecto-
derm of these forms has not evaginated instead of invaginating, and it
shows no tendency to invaginate after return to water. If evagination re-

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. II 233

suits from a reversal of cell polarity in the prospective entoderm, there is
no evidence of such reversal in entodermized ectoderm. It remains in
essentially the same physical relations to other parts as when it was
prospective ectoderm of the blastula. So-called "exogastrulae" of this
type are actually not exogastrulae at all but partly entodermized blastulae
with more or less regional differentiation. For convenience and in a purely
descriptive sense they may be called "exogastrulae" because their ento-
derms are external instead of internal, but it should not be forgotten that,
as regards origin and development, they are very different from true
exogastrulae; in so far as the entodermized ectoderm is concerned, they
are pseudo-exogastrulae. Entodermization of ectoderm apparently has no
relation to gastrulation.

Since the physical relations of entodermized ectoderm to other parts
are not altered by its entodermization and there is no evidence of reversal
of polarity in its cells, it remains a question whether the external entoderm
formed by it is inside out and which end is physiologically its apical end.
The end attached to the remaining ectoderm has developed from the
higher, more nearly apical, level of the prospective ectoderm. Moreover,
differential dye reduction indicates that this end usually shows higher
rate of dye reduction.'^

EXOGASTRULATION AS A DIFFERENTIAL MODIFICATION OF DEVELOPMENT

In general, the modifications of development by lithium do not differ
essentially from those produced by other inhibiting agents. All degrees
of inhibition of ectodermal development appear with lithium, as with
other agents; and the secondary modifications resulting from differential
tolerance, conditioning, or recovery are the same in character with lithium
and other agents. Even entodermization of prospective ectoderm is not a
specific effect of lithium but is produced by various other inhibiting
agents — -crowding and Janus green in Dendraster, several agents in Patiria;
further investigation will probably show that many other agents have a
similar effect. In fact, the evidence suggests that entodermization may
be a nonspecific differential inhibition. Differential susceptibility to grad-
ually lethal action of many agents and to inhibiting action on develop-
ment and differences in rate of dye reduction in low oxygen's indicate a
single gradient of physiological condition in earlier pregastrular stages of
echinoids and asteroids, with decrease in susceptibility and rate of dye

'^ Fig. 95, A,B,F, G, H, and Child, 19366.

'5 Chap, iv and preceding sections of present chapter.

234 PATTERNS AND PROBLEMS OF DEVELOPMENT

reduction basipetally from the apical pole and a change in condition, ap-
parently an activation, in the basal region- — in echinoids as gastrulation
approaches, in asteroids somewhat later. Prospective entoderm consti-
tutes the lower levels of the primary gradient. Exposure to lithium and
other agents entodermizes only when it begins in the earlier pregastrular
stages. In view of all the facts, it seems probable that the primary action
in entodermization is a depression or inhibition, rather than a direct in-
crease, in concentration and extent of a specific vegetal substance gradient
or stimulation of a specific vegetal metabolism, as Runnstrom maintains.
Lower levels of prospective ectoderm are entodermized by lesser degrees
of inhibition than higher levels because they are nearer the critical level.
This depression must result in alteration of concentration of many react-
ing substances in the cells concerned and in their relations to other parts,
and so in progressive increase of their specificity as entoderm. It has been
pointed out that lithium not only entodermizes but inhibits development
of original entoderm and of entodermized ectoderm. It is a point of some
importance that the exogastrulae with extremely large entoderms (Fig. 91,
B, C, E-G) are not direct effects of lithium or other agents but secondary
modifications, appearing in most extreme form in recovery but also to
some extent with differential tolerance or conditioning in low concentra-
tions. Comparison of Figure 90, A, E, and F, continuous exposures, with
Figure 90, B and C, recoveries after return to water, and of Figure 91, D,
continuous, with Figure 91, A-C and E-G, recoveries, will show the dif-
ference in entodermal development.

In recovery after high concentrations of lithium, which practically stop
development in earlier pregastrular stages, secondary modifications are
often greater, if exposure is not too long, than after somewhat lower con-
centrations, apparently because with the higher concentrations, the stage
of activation of entoderm is reached only after return to water, while
with lower concentrations it may be reached and activation inhibited
during exposure. A comparison of Figure 91, A, C, and F, 7 hours in
LiCl m/25 from two-cell stage, with Figure 91, D, in m/50 from the same
stage, illustrates the point.

Lithium, like other agents, inhibits more or less completely skeletal
development; but in secondary modifications, particularly in recovery
from the less extreme inhibitions, excess of skeleton with supernumerary
arms and other skeletal structures not normally present often develops.
Whether mesenchyme cells increase in number secondarily or dissociated
entodermal cells function as mesenchyme is not known.

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. II 235

The possibility that specific vegetal substance is already present in
early stages is, of course, not excluded; but the evidence seems to support
the view that the primary effect of lithium in entodermization, as in other
modifications, is depression or inhibition, an action on quantitative dif-
ferences in physiological condition, with origin or increase of specific dif-
ferences as a result. In short, according to this view, lithium and many
other agents are primarily inhibitors of early echinoderm development,
with regional differentials in effect depending on nonspecific differences
in physiological condition and their changes in the course of development.
Secondary differential modifications of tolerance, conditioning, or recov-
ery are not direct effects of the inhibiting agents but results of action of
the gradient factors of physiological pattern.

There is considerable evidence that ectodermization of prospective en-
toderm and re-ectodermization of entodermized ectoderm are possible.
The forms of Strongylocentrotiis in Figure 92, D, E, and F, with exposure
to lithium beginning at late blastula stage, suggest partial ectodermiza-
tion, as does Figure 92, ^, of Dendraster. At the late blastula stage, or a
little later, entoderm becomes more susceptible than ectoderm; and it
seems not impossible that with depression or inhibition it may still be ecto-
dermized. The ectodermized and "animalized" forms discussed below (pp.
243-45) have a different origin.

In exogastrulae with ectoderm decreased by entodermization the locali-
zation of mesenchyme near the apical ectodermal pole suggests that ento-
dermization extended earlier to, or almost to, that level but that with
secondary modification more or less re-ectodermization has left the mes-
enchyme where it was localized by the entodermization (Fig. 91, A-D).
Localization of mesenchyme at two or more levels may represent stages
in entodermization or in re-ectodermization. It was noted above that in-
dividuals with apparently completely entodermized ectoderm (Fig. 94,
A, B) may, in recovery, become exogastrulae with some ectoderm, like
Figure 91, E-H. These cases also indicate re-ectodermization. In a dif-
ferent line of experiment re-ectodermization has been observed by von
Ubisch (1925a, 1929).

At present there seems to be no adequate reason for regarding either
true or pseudo-exogastrulae as primarily anything but results of differen-
tial inhibition, often with various degrees of secondary modification due
to differential tolerance, conditioning, or recovery. The question how any
agent brings about true exogastrulation remains. What determines evagi-
nation instead of invagination? If we knew how invagination is deter-

236 PATTERNS AND PROBLEMS OF DEVELOPMENT

mined, we might hope to answer this question; but, although various
hypotheses have been advanced in terms of mechanical factors, differen-
tial growth, specific constitution of the basal region, differential colloidal
swelling inside and outside, etc., they remain hypotheses until we know
more about living protoplasms and how external agents act on them.
The hypothesis of differences in colloidal swelling as the determining fac-
tor not only in gastrulation but in other invaginations, and of its reversal
by agents producing exogastrulation, was advanced by Spek (1918).
Evagination of prospective entoderm suggests reversal of polarity of some
sort in the cells of this region. Differential dye reduction indicates lower
oxygen content in the blastocoel than outside (see pp. 133-40). Conceiv-
ably, this differential or some other between blastocoel and exterior may
induce gastrulation, either by colloidal changes or otherwise, and the
activation of the prospective entoderm preceding gastrulation in echinoids
may be expected to make it more susceptible to such a differential; but
in the starfish the activation is apparently less marked before gastrulation
than in echinoids. Perhaps the agents which bring about evagination ob-
literate or reverse a polarity in the entodermal cells. According to Lindahl
(1936), lithium inhibits one component of respiration in sea-urchin em-
bryos. A point of some interest is that the stomodeum, instead of invagi-
nating, often evaginates and is considerably enlarged in exogastrulae, sug-
gesting that exogastrulation is not primarily concerned with a specific
vegetal gradient.

There is, however, no evidence of such reversal in cells of entodermized
ectoderm. They retain their original relations to other parts and to the
blastocoel and show no tendency to invaginate, even in the most com-
plete recoveries. Conceivably, either their failure to invaginate may be
due to insufficient activation and susceptibility to the surface-interior
differential, or in inhibited forms, as all these are, the differential between
blastocoel and exterior may be insufficient to bring about reaction or in
some cases may be altered in character by the dissociated cells in it. But
whatever the physiology of exogastrulation, it is sufficiently evident that
neither true exogastrulation, that is, evagination of prospective entoderm,
nor pseudoexogastrulation, development of an external entoderm from
entodermized ectoderm, are specific effects of lithium.

RECONSTITUTION IN EXOGASTRULATION

Development of exogastrulae with more or less entodermization of pro-
spective ectoderm may involve extensive reconstitution. All ectodermal

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. II 237

differentiations except those of the extreme apical region and perhaps the
skeleton develop in other than their original prospective regions or cells.
Anal arms, if they appear, develop from various levels, all apical to the
prospective levels. If ventrodorsahty is not obliterated and a ciliated
band develops around the ventral side, all except its apical levels develop
from other cells than normally. When ventrodorsal ity is obliterated and
the band develops around the basal ectoderm (Fig. 91, B, C, D), its locali-
zation is, in large part, different from normal. The scale of ectodermal
organization may be greatly decreased, so that the apical half or even less
of the prospective ectoderm may approach pluteus form. This decrease
in scale is similar to that occurring in hydroid reconstitution : a piece of
Tubularia or Corymorpha stem so short that it gives rise only to a hy-
dranth or the apical part of a hydranth under natural conditions may,
under inhibiting conditions, develop a complete hydranth on a smaller
scale and stem, or in Corymorpha, hydranth, stem, and basal holdfast
region (pp. 344-49). A similar decrease in scale of organization results
from inhibiting conditions in planarian reconstitution (pp. 349-54). The
evidence indicates that in hydroids and planarians the decrease in scale
results from depression and decrease in length of a quantitative physio-
logical gradient, and at present convincing grounds for regarding the de-
crease in echinoderms as fundamentally different do not appear. The
small, thin-walled "neck" connecting ectoderm and entoderm in many
exogastrulae is evidently a reconstitution, and its possible significance is
an interesting question.^^ It is apparently ectodermal in origin; and after
return to water the entoderm often separates from it, leaving it attached
to the ectoderm. Usually, if not always, it is a secondary modification, not
a direct effect of the primary inhibition. Since it develops from what
would have been the anal region in an entogastrula, it might be regarded
as a proctodeum, developing under the experimental conditions though
absent or not developing appreciably in the entogastrula. On the other
hand, if the attached end of the entodermized ectoderm is physiologically
apical, it might perhaps be regarded as representing more nearly a stomo-
deum induced by the entoderm, particularly in the forms with radial
ectoderm and basal ciliated band, in which the flattened basal ectoderm
seems to be more or less Hke the normal ventral region in certain respects
(Fig. gi, B,C). It shows the highest rate of dye reduction after return to
water (Child, 19366), and the presence of the ciliated band about it sug-
gests that it approaches the ventral region in physiological condition.
'" Fig. 90, B, C, D; Fig. 91, .4, B, C, E, G.

238 PATTERNS AND PROBLEMS OF DEVELOPMENT

However, decision between proctodeum and stomodeum is neither pos-
sible nor necessary. The neck is apparently a secondary ectodermal modi-
fication resembling an everted proctodeum or stomodeum.

In partly entodermized forms there is also entodermal reconstitution;
entodermization is itself a reconstitution. The original prospective ento-
derm may form only a small part of the total entoderm, much less than
under natural conditions, and probably in some cases it is completely dis-
sociated; but development, either of entodermized region together with
original entoderm, or of the former alone, is definite and orderly and ap-
parently entodermal in character. Three entodermal regions, separated
by constrictions, often develop secondarily; or after greater inhibition,
only two, sometimes with a very small third region at the tip; or with
still more inhibited development there may be no regional differentiation.
Supposedly, the three or two regions correspond to entodermal regions
differentiating in normal development; but if this is so, how is the regional
differentiation determined in an external entoderm consisting in part of
original entoderm, in part of entodermized ectoderm, or entirely of the
latter? From the data available, it appears that regional differentiation
of entodermized ectoderm does not occur unless some ectoderm remains,
and it may be absent or almost absent if the ectoderm is very small
(Fig. gi, H); but whether absence of entodermal differentiation indicates
absence of some determining or inducing action of ectoderm, or is merely
the result of extreme inhibition, is at present not known. In any case,
differentiation of entodermized ectoderm with or without original ento-
derm and without invagination suggests orderly and definite interrela-
tions between parts, even in the differentially modified forms.

Entoderm which has lost its epithehal character during exposure to
lithium or other agents may, if exposure is not too long, regain epithelial
order to som^e extent during recovery in water. For example, as noted
above, forms like A and B of Figure 94 may become exogastrulae like
E-H of Figure 91. That there is recovery of epithelial character in these
cases seems evident. Moreover, after extreme inhibition spherical ento-
dermal vesicles often appear free in the blastocoel during recovery; ex-
amples are shown in Figure 96. Individuals in which the whole basal
region has become merely a cell mass with more or less dissociation (Fig.
96, A) may, after return to water, attain the condition of Figure 96, B,
with large external epithelial sac, many dissociated cells in the blastocoel,
and one or more epithelial vesicles. Many individuals like Figure 96, C,
do not recover, but 20-25 per cent may attain the condition of Figure 96,

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. II 239

D or E. In D there is apparently re-ectodermization of the apical region,
but E appears to be completely anaxiate; and it may perhaps be ques-
tioned whether the external epithelium retains entodermal character or
has become ectoderm; it has undergone the decrease in thickness char-
acteristic of ectoderm.

When entoderms of individuals in the condition of Figure 96, A and C,
come into contact, they tend to stick together ;'7 and large masses, con-
sisting of many individuals, often result. In these the entodermal regions
in contact gradually dissociate into the interior of the mass, and the
parts directly exposed to water become a continuous epithelium with the

Fig. 96, A-E. — Dendraster, recovery after extreme inhibition. A, B, LiCl m/20, 9 hr. from
two-cell stage: A, 17 hr. water, B, 40 hr. water. C, LiCl m/20, 26 hr. from first cleavage.
D, E, LiCl m/20, 26 hr., 24 hr. water.

apical regions at first protruding from the surface but gradually becoming
incorporated into it, so that finally a large vesicle full of cells results, with
complete loss of individuality of the component members. Within these
masses, however, numerous epithelial vesicles may develop after return
to water; they often persist free in the water after disintegration of indi-
viduals or multiple masses. The multiple forms are interesting cases of a
sort of reconstitution with more or less complete obliteration of axiate
pattern of the original individuals composing it and development of a
new individuality in which there is no evidence of anything but surface-
interior pattern.

Secondary reconstitutional modifications of skeletal development in

'" Less extremely inhibited exogastrulae often stick together by their entoderms if there is
loss of epithelial character at the tip, but they do not lose their individuality so completely as
the forms in which axiate pattern is almost obliterated.

240 PATTERNS AND PROBLEMS OF DEVELOPMENT

echinoids present a wide range of variation, with aberrant localizations,
deficiencies, excess of skeleton, and development of structures widely dif-
ferent from those characteristic of the species under natural conditions.
Only a few of these are indicated in the figures.

The definite and orderly character of the reconstitutional changes in
localization and differentiation of parts in exogastrulae represents the
realization of potentialities in relation to an axiate pattern, modified pri-
marily by differential inhibition and secondarily by differential tolerance,
conditioning, or recovery. It is also evident that differential inhibition may
completely obliterate axiate pattern, ventrodorsality being obHterated
with less extreme inhibition than polarity. With increasing degree of ob-
literation there is progressively less evidence of localization and differen-
tiation of particular parts, and with complete obliteration the individual
remains completely anaxiate in development and without any regional
differentiation. Particular features of axiate pattern are not fixedly as-
sociated with particular regions of egg or embry^o but may be shifted in
position with the experimental alterations of the pattern. At present it
seems difficult to account for all the results of experiment otherwise than
in terms of a primarily quantitative gradient or differential pattern in-
volving metabolism, within which new gradients and specific differences
gradually arise. When this pattern is altered differentially by external
factors, the region of the embryo or larva in which a particular differentia-
tion takes place may be shifted in one direction or another; or if the altera-
tion is sufficient, the differentiation does not appear.

OTHER EXPERIMENTS AND INTERPRETATIONS

The early experiments of Herbst on the effects of artificial sea water,
with certain salts or ions increased in amount, lacking, or replaced by
others, were primarily attempts to determine what substances were neces-
sary for sea-urchin development.'^ They produced various modifications
of development, including exogastrulation, but these were described as
effects of particular experimental environments. Many of them are simi-
lar to the differential modifications discussed in preceding sections of this
chapter, but the possibiHty that development might be altered differen-
tially in similar ways by many environmental factors or that secondary^
modifications, opposite in direction to the direct effects of external agents,
might occur after a primary inhibition, seems not to have been recognized.
The significance of many of the modifications described and figured is,

'* Herbst, 1892, 1895, 1896^, 1897, 1901a, 1904.

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. II 241

therefore, uncertain; but it is evident that many of them are differential
inhibitions. It is possible, however, that in some cases a modification re-
sulting from absence of a particular constituent of sea water may be
more or less specific in character, but to prove this requires extensive ex-
periment— it must not be taken for granted. Herbst was concerned with
other problems than those of developmental pattern, but many of his
data point to the same conclusions as those of this chapter.

The attempts to interpret differential modifications of sea-urchin de-
velopment, including exogastrulation, in terms of a single apicobasal
gradient are, in the light of more recent work, applicable only to the
earlier stages preceding gastrulation and to the ectoderm of later stages.^''
The change in physiological condition of prospective entoderm and mes-
enchyme was not known at the time of these papers; and, though evi-
dences of it were observed in some cases, their significance was not recog-
nized, nor was dissociation of prospective entoderm actually observed.
As regards the ectoderm, however, the interpretation suggested is essen-
tially the same as that in this chapter.

Runnstrom's concept of two opposed and overlapping gradients in the
apicobasal axis of the sea-urchin egg was originally stated in terms of con-
centration gradients of "animal" and "vegetal" substances, but more re-
cently it has been maintained that specifically different animal and vegetal
metabolisms characterize the axis. The hypothesis of two overlapping
gradients was stated as if established fact in the introduction to the first
of Runnstrom's experimental studies on sea-urchin development (1914)
and has appeared in many papers since."" According to Runnstrom and
Lindahl, lithium inhibits the kind of metaboHsm characterizing the ani-
mal gradient and stimulates or in some way favors that of the vegetal
gradient. As already pointed out, differential susceptibiUty to gradually
lethal agents, differential modification of development, and differential
dye reduction give no information concerning presence or absence of such
overlapping gradients. Moreover, it may be noted that two overlapping
concentration gradients of different substances may be associated with
a single gradient of metabolic rate; and metabolism, rather than concen-
trations of substances, is the effective factor in development of organismic
pattern. Under natural conditions concentration gradients do not appear
to be the primary factors determining metaboHsm; metabolism is unques-

•9 Child, 19166; MacArthur, 1924, accepting Child's views.

2° Runnstrom, 1914, 1915, 1925a, 1928a, b,c, 19296, 1931, 1933, ig^S^, b, etc.; Horstadius,
1928&, 1931, 1935, 1936a, b, 1937a, 1938; Lindahl, 1932a, b, 1933, 1935, 1936.

242 PATTERNS AND PROBLEMS OF DEVELOPMENT

tionably an essential factor in establishing them. Differential modifica-
tions of development indicate a single primary activity gradient in the
apicobasal axis and a secondary gradient arising later in the basal region
and partly obliterating the primary gradient. Analysis of developmental
stages of exogastrulae gives further evidence in support of this conclusion.
Prospective entoderm is less susceptible than ectoderm in early stages,
and more susceptible in later stages, to lithium and other agents. Lithium
and various other agents entodermize prospective ectoderm in early
stages, apparently by depressing or inhibiting it, but, after a certain de-
velopmental stage, inhibit entoderm more than ectoderm; that is, their
action is inhibitory throughout. These experiments do not prove the ab-
sence of the gradients postulated by Runnstrom but indicate a different
sort of gradient pattern as the primary effective factor in determining
form and proportions, and alterations of form and proportion are often
associated with changes in localization and differentiation of parts, indi-
cating that these factors of development originate in the same pattern.
Evidences of specific regional and local differences certainly become in-
creasingly evident as development progresses, but it does not necessarily
follow that they are the primary features of pattern.

Runnstrom has used potassium-free sea water, ZnS04, lithium, CO, and
other agents in modifying echinoid development; but determination of the
effects of a wide range of concentrations of the chemical agents and of
different exposure periods at different stages of development does not
appear to have been undertaken to any great extent. Apparently, neither
the possibility of change in susceptibility of prospective entoderm in the
course of development nor that of secondary developmental modifications,
opposite in character to the primary, in low concentrations of external
agents or in recovery after return to water, has been recognized. Conse-
quently, it is often difficult to determine whether forms described are direct
effects of the agents used or secondary modifications. The experiments
with potassium-free sea water suggest several possibilities: distribution of
potassium may differ regionally in the egg or embryo; with lack of ex-
ternal potassium internal distribution may be altered; and the effect on
development of absence of an essential ion, such as potassium, may be
different from that of a chemical or physical agent acting in addition to
the natural environment. Experiment with different amounts of potas-
sium, both below and above the content of natural sea water, appear
desirable as an adequate basis for analytic interpretation.

In several papers Lindahl and others have presented voluminous experi-

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. II 243

mental data on developmental modification and metabolism in the sea
urchin. They have described effects of various agents and of absence of
the sulphate ion and from these experiments have drawn certain conclu-
sions concerning metabolism.^' As regards some of this work, it is diffi-
cult to resist the impression that the experimental data do not always
provide an entirely adequate basis for the hypotheses advanced. Only a
few points of these investigations can be touched on here. Accepting
Runnstrom's hypothesis of two opposed, overlapping gradients, Lindahl
and his co-workers present various lines of evidence which they regard
as justifying the conclusion that the
metabolisms of animal and vegetal
regions differ specifically in character.
Certain agents are believed to "ani-
mahze," that is, to increase and ex-
tend the animal kind of metabolism;
certain others, to "vegetativize" (veg-
etalize). Here, also, the possibility
of differential tolerance, conditioning,
or recovery is apparently not recog-
nized, although it is noted incidental-
ly that further modification of form
may occur after return to water.

Perhaps the most interesting of
the modifications described are the
animalized larvae obtained by ex-
posure of eggs before fertilization to
NaSCN in calcium-free sea water,
also to Nal and sometimes by expo-
sure to calcium-free sea water alone, with fertilization and develop-
ment in natural sea water. The animalized larvae are without entoderm
or mesenchyme, that is, entirely ectodermal; and in the more extreme
types the long cilia normally appearing as a group or tuft about the apical
pole in certain earlier stages develop over much or all of the surface of
the blastula-like forms (Fig. 97). The cells bearing the long cilia also be-
come different from the general ectodermal epithelium and similar to cells
of the apical tuft. These forms do not develop beyond the stage of ciliated
blastula-like larvae, though they lose their long cilia in later stages. This
modification is prevented by exposure to cyanide or to lithium after fer-

^' Lindahl, 1933, 1935, 1936; Lindahl and Stordal, 1937; Lindahl and Ohman, 1938.

Fig. 97. Faracentrotus lividus. — In iso-
tonic NaSCN (5 cc. NaSCN, 100 cc. sea
water) for 18 hr. before fertilization; fer-
tilization and 32 hr. development in sea
water (from Lindahl, 1936).

244 PATTERNS AND PROBLEMS OF DEVELOPMENT

tilization; and with lithium, exogastrulae may develop from eggs treated
with thiocyanate before fertilization. According to Lindahl, some of
these exogastrulae are bipolar forms united by the distal ends of their ex-
ternal entoderms." The possibility that the animalized forms may not be
direct effects of the thiocyanate or iodide but secondary modifications re-
sulting from recovery after return to water, together with the activation
associated with fertilization, is not considered by Lindahl. Some degree
of ectodermization or animalization may apparently occur following ex-
posure to lithium after fertilization (pp. 227, 235), and the possibility that
Lindahl's animalized forms are cases of recovery cannot, at present, be
excluded. The fact that cyanide and lithium after fertilization prevent the
animalization of eggs treated with thiocyanate before fertilization sup-
ports, rather than conflicts with, the view that the animalization repre-
sents a recovery, not a direct effect of thiocyanate or iodide. Lindahl
finds, also, that temporary exposure to thiocyanate or iodide after fertiliza-
tion animahzes in earlier, vegetalizes in later, stages. This is what might
be expected with differential recovery following inhibition.

Following Lindahl's procedure, Rulon (1938, 1940) has obtained ani-
malized forms of Dendraster, though less extreme than some of those de-
scribed by Lindahl for other species, and suggests that they represent
recovery from a primary inhibition. According to this suggestion, thio-
cyanate inhibits differentially the slight apicobasal gradient present in the
unfertilized egg, so that any dominance of higher gradient-levels or any
definite relation between parts is almost or quite abolished; that is, in
extreme cases axiate pattern is virtuall}' obliterated. After return to water
and fertilization, with the accompanying activation, the regions which
would have been the lower gradient-levels and would, therefore, have de-
veloped as entoderm are more or less physiologically isolated, since they
are in the same, or almost the same, physiological condition as more apical
levels. Under these conditions they develop as higher gradient-levels.
This may be regarded as essentially a reconstitution similar in principle
to reconstitution of an apical region by a physically isolated basal half of
a sea-urchin embryo and to reconstitution of apical regions by physically
isolated pieces of hydroids and planarians.

" In Lindahl's figures the supposedly bipolar exogastrulae are identical with cases of union
of two exogastrulae by the tips of their entoderms. These appear in large numbers with the
more extreme lithium inhibitions, and, as described above (p. 239), many individuals may
stick together in this way with gradual obliteration of their individuality. At present it seems
not impossible that the "bipolar" forms figured by Lindahl are not actually bipolar, in the sense
that they have developed from a single egg, but are two exogastrulae united by the tips of
their entoderms. In these unions the blastocoels become continuous by dissociation of the
cells in contact, exactly as described by Lindahl.

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. II 245

Moreover, Rulon finds that, even with concentrations of thiocyanate
and exposure periods preceding fertihzation which produce the highest
percentages (30-40) of ectodermized or animahzed forms, the forms in
a given lot range from these through normal individuals to differentially
inhibited modifications. Also, with increase in concentration or exposure
period preceding fertilization, frequency of animalized forms decreases and
that of differential inhibitions increases. If animalization were a direct
effect of thiocyanate, increase in frequency of animalized forms up to
100 per cent, or nearly, would be expected with increase in concentration
and exposure, at least up to a certain limit. That is what happens as re-
gards frequency of exogastrulation produced by lithium. But after thio-
cyanate, even in the most favorable cases, less than half the individuals
exposed are animalized; and with increase in thiocyanate action beyond
this point animalization is progressively decreased, and modification in
the opposite direction, differential inhibition, increases. This is an effect
exactly similar to the relations between differential recovery and differ-
ential inhibition with various other agents and stages after fertilization
as concentration or exposure increases. Another point suggesting that the
direct effect of thiocyanate is differential inhibition before fertilization as
well as after is that in the more extreme animalizations ventrodorsality is
apparently completely obliterated, though more or less polarity may still
be present. Obliteration of ventrodorsality with polarity still an effective
factor in development is a characteristic result of certain degrees of differ-
ential inhibition with various agents, as many of the forms in Figures
91-96 indicate. In the light of all the data available Lindahl's conclusion
that animalization is a direct effect of thiocyanate treatment before fertili-
zation seems not entirely satisfactory.

Experiments with sulphate-free sea water (Lindahl and Stordal, 1937),
together with other earlier experiments, have led to the hypothesis that
carbohydrate metabolism is characteristic of the animal region, protein
metabolism of the vegetal. Only one line of evidence regarded as support-
ing this hypothesis is mentioned here. Development is found to be more
modified in absence of sulphate than when it is present in the artificial sea
water used; from this it is argued that in sulphate-free water the develop-
ing embryos give off more substance that inhibits development than when
sulphate is present. Also, developing vegetal halves are injured by lack
of sulphate, animal halves are not: on this basis the hypothesis is ad-
vanced that the vegetal halves produce poisons which are rendered non-
toxic by sulphate. The possibility that absence of sulphate may itself
inhibit development seems not to have been considered. Moreover, if the

246 PATTERNS AND PROBLEMS OF DEVELOPMENT

vegetal region, that is, prospective entoderm and perhaps lower levels of
ectoderm, undergo activation to a level above that of the apical or animal
half at the time of gastrulation, as various lines of evidence considered in
this and earlier chapters indicate, its greater injury by absence of sulphate
may be due to increased susceptibility, as its dissociation by lithium ap-
parently is; or after its activation it may produce certain metabolites in
larger quantity than the animal half. The modifications figured and de-
scribed in connection with these experiments apparently do not differ
essentially from those produced by other agents. The hypotheses ad-
vanced may be entirely correct, but the evidence on which they are based
does not appear adequate to exclude other possibilities.

Summing up, it appears that much of the evidence regarded as indicat-
ing presence of regional specificities in the sea-urchin egg and early embryo
is open to other interpretations and that the data of differential suscepti-
bility, both of differential death and differential developmental modifica-
tion, and the data of differential dye reduction give no evidence of re-
gional specificities. In fact, they indicate that such specificities are either
absent in early stages or not sufficient to give rise to distinct, specifically
different regional effects on development with different agents. Doubtless
the metabolisms of ectoderm, entoderm, and mesenchyme do sooner or
later become specifically different. Perhaps they begin to do so from the
earliest stages of development or earlier; but, if so, the differences do not
become sufficient to fix definitively their characteristics, that is, they do
not become definitively "determined" until a later stage. Probably no
one would maintain that all the local specific differences of later stages
are present in the egg as actual, localized differences at the beginning of
development: this, of course, would mean complete predetermination.
But if new differences can originate and be localized during the course of
development, the possibility cannot be excluded that specific differences
of ectoderm, entoderm, and mesenchyme may not be present primarily
or may be so slight at the beginning of development that they have little
effect but increase gradually during development. In short, it appears
possible, and much of the evidence supports the view, that specific re-
gional differences arise secondarily from a primarily quantitative pattern.
In early echinoid and asteroid development graded differences in rate of
metabolism or of certain metabolic reactions appear to be much more
important in determining the course of development and its modifications
under experimental conditions than any regional specificities that may be
present.

CHAPTER VII

DIFFERENTIAL MODIFICATION OF DEVELOPMENT:
OTHER ANIMAL GROUPS

INVERTEBRATES

SLIGHT differential inhibition of larval development has been pro-
duced with a number of chemical agents in several species of poly-
chete annelids (Child, igi'jd). In these modifications apical and
segment-forming regions are most inhibited, as might be expected from
their greater susceptibility to lethal concentrations (pp. 120-29). How-
ever, the regions of the trochophore body are differentiated so early and
the period from fertilization to swimming trochophore is so short in the
species used that the alterations in form and proportions by external
agents are not great, though they are clearly evident. Some modifications
of later larval stages are apparently secondary, but they are so slight
that their significance remains uncertain.

In reconstitution of pieces of oligochete annelids, particularly the micro-
drilous forms, physiological factors determining head frequency appear
to be very similar to those in planarians. As Hyman (1916a) has shown,
some microdrilous species regenerate a head at any body-level; in others
head regeneration occurs only at levels near the anterior end of the body,
irrespective of length of piece; and in still others, notably Lumhriculus,
head frequency in pieces depends on level of body and length of piece,
much as in Diigesia { = Euplanaria). According to Hyman, the physio-
logical factor inhibiting head development in the shorter pieces of Lum-
hriculus results from the posterior section, as it does in Dugesia, and is a
stimulation of the piece. Apparently, also, head development may be dif-
ferentially inhibited in Lumhriculus. Hyman distinguishes microprosto-
mial and aprostomial anterior ends and certain outgrowths apparently
intermediate between head and tail or beginning development as one and
undergoing partial transformation into the other. In pieces which give
low head frequency in natural environment head frequency is increased
by temporary exposure to an inhibiting agent (e.g., KCN), exactly as in
Dugesia. This rather close parallelism in certain aspects of reconstitu-

247

248

PATTERNS AND PROBLEMS OF DEVELOPMENT

tional development in planarians and annelids suggests existence of some-
what similar physiological patterns in the longitudinal axis. The problem
of bipolar forms will be considered in a later chapter.

In several gasteropod species with free-swimming veliger larvae de-
velopment of velum and shell gland have been inhibited by external chem-
ical agents. In the inhibited forms the apical region remains smaller than
in normal larvae, and the prototroch persists as a simple circular girdle
of ciliated epithelium ; even though the larvae remain alive for much longer
periods than necessary for development of shell gland and spiral coiling

Fig. g8, A-E. — Differential inhibition in a cephalopod (after Ranzi, 1928, 1929(7)

and velum, these may all be completely inhibited, the larva remaining in
the primary trochophore stage (Child, unpublished).

Differential inhibition of development in the cephalopod Loligo vulgaris
and other species has been described by Ranzi.' The modifications were
produced chiefly by LiCl, although the author states that MgCU and
ultra-violet radiation produced similar effects. As with other forms, the
individual modifications differ according to stage of development at
which embryos are subjected to the agent, length of exposure period, con-
centration, and susceptibility of the individual. The most conspicuous
feature of the inhibited forms is the differential inhibition of the head re-

' Ranzi, 1926; 1927; 1928; 19290, b; 1931; 1932.

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. Ill 249

gion, resulting in all degrees of approximation of eyes to complete cyclopia,
reduction in size of eyes, anophthalmia, and acephaly (Fig. 98, A-E).
Evagination, reduction, or absence of the stomodeum also results. Other
organs are also inhibited in many cases.

Ranzi interprets cyclopia in the squid as indicating a primary median
position of the optic primordium, and the graded series of forms between
cyclopia and normal position of eyes as different degrees of inhibition of
its duplication and change of position, but there is no evidence in normal
development of a primarily median and single optic primordium. It ap-
pears more probable that here, as in the planarian head, susceptibility
decreases from the median region laterally, so that the median region is
most inhibited; and with increasing degree of inhibition eyes develop
nearer the median line or a single median eye forms. Moreover, the arms
of the sea-urchin pluteus show a graded series of approximations to the
median plane and a single median arm, and finally ventrodorsality is
obliterated. It is certain that the arms do not develop from a primordium
primarily median, but they show the same series of approximations to
the median plane as the eyes of the squid and the planarian.

The hypothesis is advanced by Ranzi that differential susceptibility
has nothing to do with quantitative physiological or metabolic gradients
but that the most susceptible regions are those in which more complex
embryogenic processes are going on, and that these regions are more
susceptible only while these processes are occurring. What he means by
"more and less complex processes" is not at all clear. If morphological
complexity is meant, there is abundant evidence that his hypothesis is
incorrect. The apical region of an alga axis is certainly, if anything, less
complex morphologically than other parts, but it is more susceptible: the
apical region of the sea-urchin embryo does not appear to be any more
complex morphologically than other parts but is more susceptible. The
regenerating planarian head is more susceptible than parts posterior to it
and reduces dyes more rapidly in low oxygen from the beginning of its
development on, as long as the animal is in good condition. If Ranzi^

^ Ranzi, 1926, 1927, 192S, 1929a, b, 1931, 1932, 1938; Ranzi e Falkenheim, 1937, 1938. In
several of these papers it is maintained that physiological gradient pattern either does not
exist or is of no fundamental importance in development. Since the experimental evidence
presented in these papers indicates presence of gradients, the views presented are personal
opinions rather than conclusions from the data of experiment, and it need only be said that
they are not in accord with a great volume of experimental evidence.

2 so PATTERNS AND PROBLEMS OF DEVELOPMENT

means physiological or chemical complexity, the hypothesis is merely an
opinion without basis of evidence.

The only data on experimental differential modification of development
in arthropods are apparently those obtained by Brauer (1938) on em-
bryonic stages of a brucid beetle. In early stages (oviposition to 6^ hours)
axiate pattern can be completely obliterated by cyanide. In following
stages (6|-i2 hours) susceptibility decreases anteriorly, posteriorly, and
laterally from the presumptive prothoracic-maxillary region of the em-
bryonic plate. In consequence of complete inhibition of the median re-
gion, more or less complete duplication may result by formation of new
embryonic plates in the less susceptible and less inhibited lateral regions,
which would give rise to lateral parts under natural conditions. Partial
duplications of embryos show complete doubling in the prothoracic-maxil-
lary region with incomplete duplication of heads and more posterior parts;
complete dupUcations of all parts also occur (pp. 518-19).

ASCIDIANS

Very considerable differential modifications of development result from
action of external inhibiting factors on early developmental stages of the
ascidian Corella willmeriana (Child, 1927(f). Developmental stages under
natural conditions are outlined in Figure 99: the larva before hatching
{A), the fully developed swimming larva {B), a stage of tail resorption
(C), tail completely resorbed, with chorda cells aggregated in a mass {D).
Development to the swimming larva takes place in the atrial cavity of the
parent at a pH below that of normal sea water, and removal of eggs or
early embryonic stages from the atrial chamber to sea water inhibits de-
velopment differentially. Experiment indicates, however, that CO^ con-
tent of water, rather than hydrogen ion concentration, is the important
factor in the atrial chamber, the eggs apparently being conditioned to a
higher CO2 content than that of normal sea water. Sea water with alkahn-
ity increased by NaOH is more strongly inhibiting than natural sea water.

The chief modifications thus far observed are indicated in Figures 99
(£-G), 100, and loi. In uninhibited larval development the tail always
coils ventrally around the larval body (Fig. 99, ^). With slight inhibition
it may be bent in various directions (Fig. 99, E), but with greater inhibi-
tion it extends dorsally instead of ventrally and is shortened and folded
or is represented by a rounded mass in which cells of the notochord are
visible but without definite order (Fig. 99, F, G, and Fig. 100). The dorsal

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. Ill 251

region of the body is also decreased in size and flat instead of rounded; the
tail appears to arise farther dorsally, the sensory vesicle and tail are closer
together, and the papillae at the "anterior" end are more dorsal than in

Fig. 99, A-G. — Uninhibited and differentially inhibited larval development of Corella
willmeriana. A, uninhibited before hatching; B, fully developed larva; C, tail partially re-
sorbed; D, tail completely resorbed; E, F, G, lesser degrees of differential inhibition (from
Child, ig2jd).

uninhibited individuals. Comparison of Figure 99, A and B, with Fig-
ure 99, E-G, less extremely inhibited forms, and with Figure 100, more
extreme inhibitions, will show the differences. Differential reduction of
KMnO^ in developing larvae indicates the tail as the most intensely active

252 PATTERNS AND PROBLEMS OF DEVELOPMENT

region, the dorsal region next; the developmental modifications show
them to be the most susceptible regions.

In uninhibited individuals the sensory vesicle contining the two pig-
mented sense organs lies somewhat to the right of the median dorsal
region. Among the more inhibited forms, individuals apparently com-
pletely bilateral, with pigment spots bilaterally localized at some distance

Fig. ioo, .4-G.— More extreme differential inhibitions of Corella; in all the most advanced
stage of tail development is figured; all show dorsiventral differential inhibition (from Child,
1927 d).

from the median plane, appear rather frequently. The pigment spots are
sometimes connected by a band of pigment like the eyespots of tera-
tophthalmic planarian heads (Fig. loi). It would be of interest to know
what sort of ascidian would develop from these apparently completely
bilateral forms, but in all observed thus far there has been little or no
development beyond the stages figured.

In uninhibited development the larval tail shows a very distinct gradi-
ent with high end at the tip. At or before the beginning of tail resorption
this gradient disappears (pp. 145-47)- With differential inhibition the
caudal gradient is decreased or quite obliterated; consequently, the tail
is smaller, shorter, or a mere cell mass containing chorda cells but without

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. Ill 253

definite pattern. The apparent association of orderly arrangement of
chorda cells in a definite notochord with presence of a gradient in the

Fig. ioi, A-E. — Bilaterally symmetrical forms of Corella resulting from differential in-
hibition (from Child, ig2-]d).

Fig. 102, A-E. — Stages of metamorphosis of Corella. A, uninhibited; B-D, differentially in-
hibited individuals metamorphosing without hatching; £, metamorphosis of somewhat inhib-
ited individual after hatching but without development of tail at any stage (from Child, ig2'jd).

developing tail and loss of definite arrangement when the caudal gradient
disappears is perhaps of interest.

The more inhibited individuals very commonly attain more or less ad-
vanced stages of metamorphosis within the egg membrane (Fig. 102) and,

254 PATTERNS AND PROBLEMS OF DEVELOPMENT

if not too much inhibited, may become young ascidians before hatching.
Developmental stages after resorption of the tail are much less susceptible
than those of larval development; and differentially inhibited larvae in
which a tail never develops may become apparently completely normal
ascidians, though many of them fail to develop beyond early stages of
metamorphosis. Much inhibited forms sometimes hatch but are incapable
of movement, or the aborted tail may show only slight muscular tremors
(Fig. IOC, F,G).

VERTEBRATES

Extended consideration of the enormous literature of vertebrate tera-
togeny and teratology is quite beyond the present purpose. We are pri-
marily concerned here with developmental modifications, apparently dif-
ferential in character, that is, not specific for particular agents, and re-
sulting from controlled exposure of the entire organism in early stages
of development, or of egg or spermatozoon, to the action of physical and
chemical agents or conditions outside the range of so-called "normal en-
vironment." These experimental modifications throw some light on the
problem of origin of certain of the "accidental" teratological forms result-
ing from uncontrolled and unknown conditions. One of the most interest-
ing results in this field of experimental vertebrate teratogeny is the high
degree of similarity of teratological forms produced by different external
factors. Although various authors have, from time to time, regarded
terata produced by certain agents as specific for these agents, further ex-
periment has shown that the supposed specificity did not exist. In the
more advanced developmental stages after a considerable degree of dif-
ferentiation particular agents may act more or less specifically on par-
ticular organs; but in the earlier stages less, or no, evidence of such local-
ized specific action appears. If the yolk content differs greatly in different
regions of the egg or embryo, as in amphibians, the yolk-laden parts may
be more susceptible to certain agents and less susceptible to others than
those with little yolk. This, of course, represents a specific difference re-
sulting from regional differentiation already present in the egg. However,
aside from such differences as these, the general similarity of the modifica-
tions produced by many different agents is evident, even in the work of
the earlier investigators in this field, ^ and some of them called attention
to it. The evidence from later work supports the view that the modifica-
tions of early stages of vertebrate development by external agents depend

3 See, e.g., Dareste, 1891; O. Hertwig, 1892, 1895, 1898; Gurwitsch, 1895; Bataillon, 1901,
1904; Rabaud, 1901-2.

DIFFERENTL\L DEVELOPMENTAL MODIFICATION. Ill 255

primarily on a differential susceptibility of different embryonic region,
which is, to a high degree, similar for different agents though the differen-
tial in effect may be greater with some than with others. In many cases,
however, analytic interpretation of the experimental data is difficult or
impossible because of experimental procedure. The range of concentration
or intensity used is often narrow; exposure to the agent is sometimes con-
tinuous, sometimes temporary, and, in certain experiments, repeated at
intervals, and effects of exposures beginning at different developmental
stages are not always clearly distinguished. The possibility of differential
tolerance or conditioning to low concentrations or intensities and of dif-
ferential recovery after temporary exposure has received no attention from
most authors. In most of the experiments with earlier stages susceptibility
decreases from the prospective anterior region posteriorly, but sooner or
later a second region of high susceptibility appears at a more posterior
level. In the fishes this occurs earlier in some species used in experiment,
later in others (see Fig. 52, p. 149). In the amphibians the dorsal lip region
has become highly susceptible at the beginning of gastrulation (Fig. 53,
p. 152), and in the chick the region of the primitive streak from early
stages to its disappearance is highly susceptible (Hyman, 1927a). Differ-
ential inhibition of the earlier stages always involves the head region to a
greater or less degree ; but modifications resulting from inhibition of the
region of gastrulation may be absent, differ widely in degree, or appear at
different body-levels, according to stage at which exposure begins, period
of exposure, and rapidity and degree of action of agent, and possible indi-
vidual differences in susceptibility. A variety of modifications may result
from differential inhibition in a single experimental lot; and, if secondary
modifications of conditioning or recovery are possible, the range of forms
becomes still greater. Some have maintained that there is no discernible
law and order in the teratogenic action of external agents (e.g., Kellicott,
1916; Cannon, 1923). As a matter of fact, however, law and order are
becoming increasingly evident as experiment and analysis continue; but
it is only by working with a considerable range of concentrations or in-
tensities, exposure periods, and developmental stages that we can hope
to attain definite results.

Numerous agents have been used in experimental vertebrate teratog-
eny : for example, magnesium salts, lithium salts, cyanides, inorganic and
organic acids, bases, various other electrolytes, anesthetics, alkaloids,
acetone, formaldehyde, low and high temperature, ultra-violet radiation,
X-rays, and radium. In most experiments the agent used has been ap-

2S6

PATTERNS AND PROBLEMS OF DEVELOPMENT

plied directly to developmental stages, but in some cases the effects on
development of treatment of ova or spermatozoa preceding fertilization
have been determined. In certain cases genetic factors have been found
to be concerned in producing developmental modifications similar to those
resulting from experimental conditions.

Fig. 103, A-K. — Differential inhibition of head in Fiindulus heterocUtus. A, uninhibited;
B, cyclopia; C, cyclopia with reduction in size of eye; D, anophthalmia; E-K, anterior views of
uninhibited head and various degrees of differential inhibition {A, B, E-K, after Stockard,
1909; C, D, after Werber, 1916a).

EXPERIMENTS ON DEVELOPMENTAL STAGES OF FISHES

The most conspicuous and most discussed feature of differential inhibi-
tion in fishes is the mediolateral differential inhibition of the head."* In
the inhibited heads eyes develop in all degrees of approximation to the
median line to complete cyclopia, reduction in size of the median eye,
and anophthalmia (Fig. 103). In the more extreme cases there is usually
reduction of the anterior brain region; but, according to Stockard, cyclopia

'• See, e.g., Stockard, 1907a, b, 1909, 1910a, 1921, and other papers; Lewis, 1909; McClen-
don, 1912a, b; Werber, 1915, 1916a, b; Kellicott, 1916; Gianferrari, 1921; Hinrichs, 1925.

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. Ill 257

may occur without appreciable brain reduction. Stockard's interpretation
of these modifications as indicating a median origin of the optic primordia
and prevention of their later separation or of attainment of the usual
position does not necessarily follow from the experimental data (see
pp. 282-85).

In many of the modifications in fishes the head region is strongly in-
hibited; posterior regions, little or not at all, because exposure to the
inhibiting agent was temporary and ended before the posterior region
developed high susceptibihty. In various other cases further experiment
is necessary to determine whether differential conditioning or recovery
may be concerned in the result. In the more extreme degrees of inhibition
posterior regions are usually more or less inhibited or absent (Fig. 103,
C, D). Development of the heart may also be differentially inhibited by
external agents, so that it remains tubular; and a periodic reversal in direc-
tion of beat has been observed in certain cases (Gowanlock, 1923).

Asymmetric forms with a single eye on one side of the head have been
described frequently. Probably many of these result from unequal expo-
sure to the inhibiting agent or from unequal oxygen supply or unequal dif-
fusion of CO,. If the developing embryos lie undisturbed on the bottom
of a container or if floating eggs are in contact or aggregated in groups,
such differences may arise. When exposure to the agent begins in early
stages, difference in stage of the division cycle on the two sides of the
blastoderm, with resulting difference in susceptibility at the time when
the agent becomes effective, may determine later asymmetry. But what-
ever the factor concerned in a particular case, it seems to be incidental.
Evidences of differential acceleration of development with production of
megacephalic forms in Macropodus have been obtained by Gowanlock.^

EXPERIMENTAL MODIFICATIONS OF AMPHIBIAN DEVELOPMENT

In the many studies of modification of amphibian development differ-
ent agents, concentrations, and intensities have been used, different de-
velopmental stages have been exposed and for different periods, and a
great variety of forms has resulted. Most of these show the characteris-
tics of differential inhibition and indicate the presence of a definite differ-
ential susceptibility pattern, that is, the modifications with different
agents give little or no evidence of regional specificity.'*

5 Unpublished. See Child, 19246, p. 85, for figure.

^ The following references illustrate, to some extent, the development of investigation in
this field: O. Hertwig, 1892, 1895, 1898; Gurwitsch, 1895; C. B. Wilson, 1897; Bataillon, 1901,
1904; Bohn, 1903; Morgan, 1903; Schaper, 1904; Jenkinson, 19066, 1911a; Levy, 1906; Bardeen,

258

PATTERNS AND PROBLEMS OF DEVELOPMENT

According to Bellamy (1919), who has given particular attention to the
modifications of early stages, differential inhibition appears first in altera-
tion of the "cleavage ratio," that is, a graded decrease in rate of cleavage,

d.L

B

C D

Fig. 104, A-D. — Differential inhibition in early development of frog {Rana pipiens). A,
differential inhibition of cleavage, KCN m/ 1,000, 24 hr. from two-cell stage, then in KCN
m/5,000, 24 hr.; B, differential inhibition of dorsal lip in gastrulation, median region most
inhibited, LiCl m/10.62, 38 hr. from advanced cleavage; C, more extreme inhibition of gas-
trulation, median dorsal lip almost completely inhibited, lateral lips also retarded, LiCl
m/10.62, 28^ hr. from one-cell stage; D, equatorial gastrulation resulting from differential
inhibition, LiCl m/10.62, 76 hr. from one-cell stage (from Bellamy, 1919).

greatest apically and decreasing basipetally. With certain degrees of this
inhibition blastomeres may be approximately the same size throughout
(Fig. 104, A). Under inhibiting conditions so severe that development

1909; Doms, 1915; Baldwin, 1915, 1919; Bellamy, 1919, 1922; Leplat, 1920; Cotronei, 1921,
1922; Higgins and Sheard, 1926; Motomura, 1931; Holtfreter, 19346; Lehmann, 1933, 1936a,
b, 1937a, b, c, 1938.

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. Ill 259

stops in early cleavage, a few cleavage furrows may appear about the
apical pole, evidently before the agent is fully effective ; but their progress
basipetally is soon stopped, and the basal region does not undergo cleavage
at all. With certain fat-soluble agents, such as alcohol, this type of modi-
fication may perhaps result from a higher concentration in regions of
higher yolk content or from a more or less specific action of the fat-soluble
agent on this region.

Under inhibiting conditions permitting continuation of development to
the blastula stage or later, the less susceptible yolk cells may proHferate
into the blastocoel and in some cases completely obliterate it, the degree
of such modification depending on concentration of agent and stage when
exposure begins. It has been generally observed that inhibiting agents
determine disturbances in gastrulation. According to Bellamy, the dis-
tance between apical pole and level of earliest stages of the blastopore
decreases with inhibition. The median dorsal lip of the blastopore is most
inhibited with the lesser degrees of inhibiting action; and, as gastrulation
proceeds, the blastopore takes the form of an inverted U or V and later
becomes ovoid in outline with its long axis in the median plane (Fig. 104,
B). Under conditions severe enough to inhibit lateral, as well as dorsal,
lips, the developing blastopore takes the form of a transversely flattened
crescent (Fig. 104, C) and progresses subequatorially or even equatorially
around the embryo (Fig. 104, D). In these individuals the blastopore
makes its appearance near or at the equator, that is, much nearer the
apical pole than normally; and the progress over the yolk of the blasto-
pore lips may be completely inhibited, so that much or all of the basal
hemisphere remains a permanent yolk plug. In some gastrulae of this
type more or less elongation of the pigmented region in the apicobasal
axis may take place in recovery, but dorsiventrality and bilaterality are
apparently completely obHterated (Fig. 105, yl). These equatorial gastru-
lae have not been observed to develop appreciably farther, even after re-
turn to water; probably their failure to do so is associated with the more
or less complete obliteration through differential susceptibility of dorsi-
ventral and mediolateral gradient differences. However, a secondary in-
vagination often develops between the original blastopore and the apical
pole and may extend partly or completely around the embryo. Its char-
acteristics are similar to those of the primary invagination ; it begins dor-
sally and progresses ventrally (Fig. 105, B-D). The possibiHty that the
beginning of normal gastrulation involves some degree of physiological

26o

PATTERNS AND PROBLEMS OF DEVELOPMENT

isolation from apical dominance and that under inhibiting conditions a
secondary isolation may occur nearer the apical pole than the primary
and determine a secondary invagination is suggested by Bellamy. These
modifications have been obtained with LiCl m/io, HgCU m/500,000,

inv.2

B

inv.2

znv.

C D

Fig. 105, .4 -D.— Differential inhibition in frog; gastrula stages. .1, equatorial gastrulation
and radially symmetrical embryo, LiCl m/ 10.6 2, 76 hr. from one-cell stage, water 20 hr.;
B, equatorial gastrulation and secondary invagination nearer apical pole, HgCL m/500,000, 24
hr. from one-cell stage, water 24 hr.; C, D, secondary invaginations, LiCl m/io, continuous
from one-cell stage, d.l., dorsal lip; hp., blastopore; y.p., yolk plug; inv. 2, secondary invagina-
tion (from Bellamy, 1919).

o.ooi per cent formaldehyde, and KCN m/i,ooo with successive decreases
in concentration after 24 hours.

Under experimental conditions permitting further progress of develop-
ment not only do the resulting forms represent various degrees of dif-
ferential inhibitions but with low concentrations or intensities unques-
tionable secondary modifications appear in later developmental stages.

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. Ill 261

Neurula stages with relatively slight inhibition show reduction of the
anterior part of the neural plate and often a permanently open blastopore

B

D

vs

Fig. 106, A-F. — More advanced stages of differential modifications of frog development.
A-E, differential inhibition. A, KCN, 4 days from two-cell stage, water 2 days; B, formalde-
hyde 0.0075 per cent, 2 days from four- to eight-cell stage; C, microcephalic with shallow neural
groove and persistent yolk plug, KCN m/5,000, 24 hr., water 5 days; D, KCN m/io,ooo, 4
days, water 8 days; forms like A-D also obtained with LiCl; E, fused suckers and nasal pits,
tail dorsiventrally differentially inhibited, LiCl m/s, 3 hr. at late cleavage, water 4 days; F,
differential recovery, LiCl m/io, 2 days, water 5 days, t, tail; 11 s, ventral sucker; y.p., yolk
plug (A-D from Bellamy, 1922; E, F, from Bellamy, 1919).

varying in size. With increasing inhibition these characteristics become
increasingly conspicuous (Fig. 106, A-C). With further development neu-
ral folds may fail to close posteriorly and be separated by the open

262 PATTERNS AND PROBLEMS OF DEVELOPMENT

blastopore, giving various degrees of spina bifida, usually with micro-
cephaly, sometimes almost complete acephaly. In cases of spina bifida two
tails often develop, usually more or less bent dorsally, indicating greater
inhibition on the more susceptible dorsal side (Fig. io6, D). This dorsal
curvature of the tail or tails is highly characteristic of differential inhibi-
tion, even in absence of spina bifida (Fig. io6, E). All degrees of medio-
lateral differential inhibition appear in the head region. Ventral suckers,
olfactory pits, and eyes develop with increasing inhibition progressively
nearer the median plane, may become single median organs, often with
decrease in size, or may be entirely absent; and inhibited forms almost or
quite acephalic occur. Since different regions of the embryo attain high
susceptibility at different developmental stages, the forms resulting from
differential inhibition differ according to stage and period of exposure to
the agent. For example, with temporary exposure in earlier stages the
head region may be greatly inhibited, the tail almost or quite normal,
because the tail region was not highly susceptible at the stage of exposure.
With temporary exposure at certain later stages the head may be little,
the tail greatly, inhibited; but these differences apparently do not repre-
sent specific differentiations — at least they are not specific for particular
agents.

The most conspicuous secondary modification with the less extreme
degrees of inhibition in early stages is development of a relatively large
tail bud with ventral curvature of the developing tail (Fig. io6, F), that
is, greater elongation of the dorsal side, a modification opposite in
character to that of direct differential inhibition (Fig. io6, D, E). These
forms are frequent with return to water after temporary exposure and
appear, beyond question, to be cases of dorsiventral differential recovery.
They are less frequent with continuous exposure, but sometimes appear
apparently with differential conditioning, though less extreme than after
return to water. Animals inhibited to the spina bifida condition (Fig. io6,
A-D) usually show little or no secondary modification; they may undergo
some further development after return to water, but differential inhibi-
tion remains predominant.

With low concentrations of NaOH and of alkaline KCN and by addi-
tion of HCl to a well water with high carbonate content Bellamy (1922)
obtained acceleration of development with indications of differential ac-
celeration in precocious head development and megacephalic tadpoles
(also Child, unpublished).

Forms described by Higgins and Sheard (1926), resulting from daily

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. Ill 263

I -minute exposures to ultra-violet radiation of a certain wave length
range, show strong ventral curvature and are apparently cases of differen-
tial tolerance or conditioning, while those resulting from daily 2-minute
exposures are, at least in part, difTerentially inhibited forms with persist-
ent open blastopore.

Bellamy's experiments show that the dorsal lip of the blastopore at
the beginning of gastrulation is highly susceptible to many agents and
that its median region is more susceptible than lateral regions, but they
do not indicate specific susceptibility to any of the agents used. Moreover,
they do not show whether, or to what extent, ectodermal differential
inhibitions of the head and dorsal region are direct effects of the agent on
the ectoderm or results of differential inhibition of the chorda-mesoderm,
the dorsal inductor (see chap. xii). The high susceptibiHty of the dorsal
lip, decreasing anteriorly and laterally, suggests that differential inhibition
of this region may be of primary importance in various modifications of
the ectoderm, but doubtless direct differential inhibition of the ectoderm
after induction has taken place plays a part in the modifications — for
example, in those of the head.

By removal of membranes of axolotl and anuran blastulae and culture
in a modified Ringer solution partial or complete exogastrulation has been
produced. '^ In total exogastrulation no neural plate develops in the ecto-
derm, apparently because of absence beneath it of the entomesodermal
inductor, but the evaginated entomesoderm differentiates without evi-
dence of retardation or differential inhibition (see pp. 463-65) . Exogastru-
lation also occurs in some cases in water after removal of membranes, ap-
parently in consequence of collapse of the blastocoel; but in salt solution
all axolotl embryos become exogastrulae. From the data it appears prob-
able that a differential susceptibility is not primarily concerned in this
type of exogastrulation. The collapse of the blastocoel in the soft, highly
fluid axolotl embryo without membranes is probably the chief factor.

The view that lithium has specific regional effects, differing at differ-
ent developmental stages and dependent on critical stages or phases in
development of the dorsal inductor region, and that action of external
agents in general on embryonic development shows many evidences of
specific regional effects differing at different stages has been advanced
recently in a number of papers by Lehmann.^ The evidence presented in
these papers, however, does not at present seem to provide adequate basis

7 Holtfreter, 1931a, igs^d, e; see also Goertller, 1926.

* F. E. Lehmann, 1933; 1934; 1936a, b; 1937a, b, c; 1938.

264 PATTERNS AND PROBLEMS OF DEVELOPMENT

for the conclusion that the local differences in susceptibility to lithium or
other agents constitute evidence for the presence of local specific differ-
ences in the embryo. Considering one case, for example, Lehmann finds
that treatment with lithium at the beginning of gastrulation produces
otocephaly with anterior head region normal, while lithium treatment at
the midgastrula stage produces cyclopia, with otic region normal. The
possibility that in the first case early invagination of the part of the dorsal
inductor which attains later the most anterior position may protect it
to some extent from lithium action does not seem to be considered. If in-
vagination does protect in this way, Lehmann 's results may be a matter
of differential exposure to lithium rather than of specific regional differ-
ence. The possibility that different degrees of recovery may occur in dif-
ferent regions is also not considered. In the second case, lithium treat-
ment at midgastrula stage, anterior head region, and otic region may be
more or less equally exposed, and the more susceptible anterior region
more inhibited. However, cyclopia may also result from exposure to in-
hibiting agents beginning in early cleavage stages.

Formation of notochord may also be inhibited by lithium treatment,
according to Lehmann, the prospective notochord being "mesodermized"
and incorporated into the somites, and these become continuous across
the median plane. Continuity of somites across the median plane has
been obtained by lithium treatment of late blastulae or early gastrulae of
Ambly stoma with retarded development, not absence, of notochord
(Cohen, 1938a). In these individuals the notochord has evidently not
been mesodermized, and transverse continuity of somites dorsal to it is
apparently a matter of physical conditions resulting from retardation of
its development. It is suggested by Cohen that in Lehmann 's material
the notochord is not mesodermized but is never segregated from the
archenteric roof, that is, under the inhibiting conditions it remains ento-
derm. The modifications of anuran development, evidently differential
inhibitions, obtained by Hoadley (1938) with high temperature, show
similar inhibitions of notochord and median continuity of somites. Evi-
dently this modification is not specific for lithium, and the fact that it is
more conspicuous anteriorly suggests differential susceptibility in relation
to an anteroposterior gradient in the invaginated chordamesoderm ; the
presence of such a gradient is shown by differential lethal susceptibility
and by differential dye reduction and is also suggested by the different
degrees of inductor action by different portions of the dorsal inductor re-
gion when implanted. Once more it may be emphasized that what may

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. Ill 265

appear, with a few experiments, to be regionally specific effects, may
prove, on more extensive investigation, to be nothing of the sort. Leh-
mann's conclusions may be entirely correct, but the experimental evi-
dence does not, at present, exclude other possibilities, and the develop-
mental modifications concerned do not appear to be essentially different
from others, apparently not specific. Further experiment with different
developmental stages, different agents, and a wide range of concentra-
tions and exposure periods is still highly desirable.

HIGHER VERTEBRATES

Experimental modification of development in the chick by subjection
of the embryo to particular external agents and altered environmental
conditions presents certain difiiculties. When a chemical agent is applied
inside the shell, there is no certainty that all embryonic regions are sub-
jected to the same concentration at the same time ; and when it is allowed
to diffuse through the shell, the possibihty of control of distribution is
even less. Uniform decrease of oxygen supply to different parts of the
embryo cannot be certainly attained by decreasing the surface through
which oxygen can enter or by decreasing oxygen in the surrounding at-
mosphere, because differences in oxygen consumption of different embry-
onic parts may decrease the supply at very different rates about different
embryonic regions. The only external factors which meet the require-
ments seem to be change of temperature and perhaps X-rays and radium.
In view of this situation and of the possibility of differential tolerance,
conditioning, and recovery it is often impossible to determine the signifi-
cance of a particular teratological embryo. Nevertheless, the general sim-
ilarity of many modifications of chick development to those of fishes and
amphibia, which apparently depend on differential susceptibihty, leaves
little doubt that essentially similar physiological factors are involved in
all three groups. More than forty years ago Dareste (1891) concluded
that similar modifications are produced by different external conditions,
that the same external factors do not always produce the same modifica-
tions, and that the teratological forms depend on time, intensity, and
duration of action rather than on the nature of the external agent. Later
work, in large part, confirms these conclusions. Hyman (1927a, b), dis-
cussing differential susceptibility in the chick, points out that the regions
most modified are those most susceptible to lethal agents.

Inhibition of head development, approximation of eyes and other lateral
cephalic organs to the median plane, their development as single median

266 PATTERNS AND PROBLEMS OF DEVELOPMENT

organs, their absence, anencephaly, and acephaly all occur in the chick.
The developing heart is, in its earlier stages, a region of high susceptibility
and frequent, apparently differential inhibition. Decreased elongation,
spina bifida of various degrees at various levels, and inhibition of posterior
regions are, as Hyman points out, associated with the high susceptibility
of the region of posterior elongation which becomes separated from the
highly susceptible head regions with the progress posteriorly of the node
and primitive streak. With sufficient inhibition the whole embryo may
be absent.^

The early developmental stages of mammals are, as yet, scarcely ac-
cessible to long-continued experiment with altered environment. The ef-
fects on the nervous system of irradiation with X-rays and radium (Bagg,
1922) and of administration of alcohol to parents suggest embryonic dif-
ferential susceptibility. In some of these experiments, however, the direct
effect of the experimental conditions is on the germ cell rather than on
the embryo; in all cases further experiment is necessary with data con-
cerning early stages.

OTHER FACTORS IN MODIFICATION OF VERTEBRATE DEVELOPMENT

In his experiments on development of ligatured Triton eggs Spemann
(1904) found that, when the ligature was slightly oblique to the median
plane, more or less complete duplication of the anterior end resulted, as
with median ligatures, but that the head developing on the side with
less of the median anterior region may show all gradations of inhibition
to complete cyclopia, the modifications being the same as the differential
inhibitions produced by chemical and physical agents. In the light of
more recent work of Spemann and his school it seems evident that the
inhibition of head development is associated with inadequacy of induc-
tion, perhaps also with absence of the apical ectoderm.

Inhibition of development, apparently differential in character, follow-
ing injury to the spermatozoon, is of particular interest because the in-
jury, doubtless chiefly of the sperm nucleus, appears in development as
inhibition with regional differential along one or more axes, as in cases of
direct exposure of the ovum or developing embryo to an inhibiting agent.
Eggs fertilized by spermatozoa irradiated by X-rays or radium, or ir-

9 For references to, and discussion of, the work of earlier experimenters — Fere, Windle,
Kaestner, Mitrophanow, and others — see Hyman, 1926a. See also the following: Rabaud,
1901-2; Oilman and Baetjer, 1904; Tur, 1904; Reese, 1912; Stockard, 1914; Alsop, 1919;
Riddle, 1923; Byerly, 1926; Buchanan, 1926c; Hinrichs, 1927; Gradzinski, 1933, 1934; also
citations by these authors.

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. Ill 267

radiated ova fertilized by normal spermatozoa, have been found to give
developmental modifications essentially similar to those resulting from
direct action of inhibiting agents on developmental stages; treatment of
ova or spermatozoa with chemical agents has given similar results.'" In
the case of injured spermatozoa the effect of the injury on development
is, in general, greatest in the regions most susceptible to inhibition by
direct action of external agents, the regions of greatest developmental ac-
tivity; that is, the cells of these regions with injured nuclei do not attain
the physiological levels essential to full development, but development
at lower gradient levels is less, or not at all, affected.

It has been shown that certain heterogenic hybrids, notably among the
fishes, often show inhibitions of development similar in character to the
differential inhibitions by external agents. In a series of papers concerned
with fish hybrids particular attention has been called to this point by
Newman." He has noted in some detail the similarities of the hybrid
terata to those determined by direct action of chemical and physical
agents. In these hybrids development is retarded, and the monsters usu-
ally show what appear to be various degrees of differential inhibition, the
heads being small, eyes approximated or cyclopic, the heart, and in some
the posterior part of the body, inhibited as with external agents. More-
over, forms similar to cases of differential tolerance, conditioning, and
recovery also appear. The similarity of hybrid monsters to modifications
produced by external inhibiting agents was also noted by Loeb (191 2,
191 5). More recently similar terata have been reported by Montalenti
(1933) as characteristic of the amphibian cross Bujo viridis 9 X B. vul-
garis & but infrequent in the reciprocal cross.

Various suggestions have been advanced in attempting to account for
the factors concerned in these modifications of hybrid development: toxic
action of the sperm, in general more severe the less closely related the
species crossed; nuclear ''incompatibility" resulting in aberrant distribu-
tion of chromosomes or perhaps in inactivation or loss of chromosomes of
parts of chromosomes; alteration of the nucleoplasmic ratio; asynchrony
of developmental rate, particularly in the case of sperm from a species
with slower development than the maternal species.

" See, e.g., Opperman, 1913, fish spermatozoa, radium; Gee, 1916, fisli spermatozoa and
ova, alcoiiol, NaOH; Bardeen, 1907, amphibian spermatozoa. X-rays; O. Hertwig, 191 1, 1913,
amphibian spermatozoa, radium; G. Hertwig, 191 1, amphibian ova, radium; G. and P.
Hertwig, 1913, amphibian spermatozoa, nicotine, strychnine, chloral hydrate, various dyes;
Stockard, 191 2, 1913, guinea-pig spermatozoa and ova, alcohol.

" Newman, 1908, 1914, 19150, 1917&, 1918.

268 PATTERNS AND PROBLEMS OF DEVELOPMENT

Whatever the particular mechanism involved, it appears that in the
hybrid monsters, as in those resulting from injury to the germ cells, the
regions more susceptible to external inhibiting factors are, in general,
more inhibited by these physiological factors which interfere in some un-
known manner with the activities of life. The effect of the interference
is primarily greatest in the regions most intensely active, but in some cases
these regions apparently become more or less completely adjusted to the
physiological interference or "outgrow" it. One is tempted to suggest
that development of parts originating at high gradient-levels is more
"dilhcult" than that of other parts; that is, for such development the
mechanism must be in the best working order, and relatively slight inter-
ferences decrease its effectiveness more than that at lower levels, but the
"best working order" also involves, within limits, the possibility of more
rapid or more complete equilibration or adjustment to, or recovery from,
the lesser interferences.

The enormous literature of descriptive teratology records numerous
monsters, chiefly in man and other mammals, apparently representing all
possible degrees of differential inhibition and perhaps in some cases sec-
ondary modifications. These have been variously classified and inter-
preted.^^ Little or nothing is known concerning the conditions determin-
ing these terata, but in the light of experimental teratogeny it is highly
probable that many physiological or pathological factors may give rise
to essentially similar modifications.

A large and extremely interesting series of monsters occurring in certain
branches of an inbred strain of guinea pigs with relatively high frequency
has been described and its genetic basis discussed by Wright and Wagner
(1934; Wright, 1934). In this series a wide range of inhibited head forms
appears with resemblances to experimental differential inhibitions, but
most of them fall into two, rather than into a single, series. The chief
forms are designated and defined by Wright and Wagner as follows:

Practically it turns out that a two-dimensional scheme includes nearly all of the
combinations described. The mandible is regularly more susceptible than the maxillary
process. This gives a basis for one series of grades.

Brachygnathus: Mandible short but of normal width.

Micrognathus: Mandible diminutive. Ventral approach of ears.

P'ypoagnathus: Mandible absent. Ear ossicles united.

Synzygo-agnathus: As above with more or less fusion of zygomatic arches.

" See, e.g., St. Hilaire, 1832-37; Ahlfeld, 1882; Ballantyne, 1904; Schwalbe, 1906-13, par-
ticularly Teil III, Abt. i, Kap. 5 and 6, "Cyclopie, Otocephalie, Triocephalie, etc."; Abt. 2,
Kap. I, "DieMissbildungendes Auges," Kap. 2, "Missbildungen des Nervensystems," Kap. 5,
"Missbildungen des Gebisses," Kap. 6, "Die Missbildungen des Ohres."

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. Ill 269

Azygo-agnathus: Zygomatic arch absent or vestigial. No mandible.

Premaxillary defect has no regular relation to this series, and the same is true of
defects of the nose, brain and eyes but premaxillary defects always seem to occur with
the latter group and these usually occur in a regular sequence of increasing susceptibil-
ity from anterior to posterior. This gives the basis for another series of grades. Most
of the terms are familiar in the literature of the subject.

Brachyrhynchus: Premaxillary reduced or absent. Brain normal.

Arrhinencephalus: As above except defect of olfactory lobes and more or less union
of cerebral hemispheres. Usually one nostril (monorhinus).

Rhinocyclops: Forebrain vesicular. Eyes more or less united. Proboscis above the
single orbit.

Cyclops arhinus: As above except for absence of proboscis.

Anops: As above except for absence of eye.

Aprosopus: Cerebellum and anterior parts of brain absent.

Monoto-aprosopus: As above, ear vesicles united.

According to the interpretation offered by Wright and Wagner, differ-
ent regions of the developing head attain their maximum susceptibiHty
at different stages, and the character of the individual modification de-
pends on the stage at which the inhibiting or depressing factor becomes
effective. That genetic factors are concerned in these modifications has
been demonstrated by Wright (1934). He suggests that they act by bring-
ing about, directly or indirectly, a general depression of vital activity at
a particular critical moment with permanent effects on the development
at the anterior end of the polar embryonic axis as the most active, and
hence the most susceptible, region at the time when the action takes
place. The depressing action at an earlier stage produces different modi-
fications from action at a later stage. The same is true for action of ex-
ternal inhibiting agents. But chance conditions also play a part in deter-
mining the individual forms; chance differences in implantation or in
blood supply or chance mutations are suggested by Wright and Wagner
as possible factors. "The randomness of occurrence within litters of each
size indicates that each monster is due to a highly localized chance event,
supplementing a genetic tendency common to all members of the group"
(Wright, 1934, p. 502).

In an inbred strain of mice Little and Bagg (1924) obtained forms with
reduced lower jaw and frequent microphthalmia, apparently representing
somewhat less extreme degrees of differential inhibition than the guinea
pigs. Occurrence occasionally in mammals of a single individual or a
single litter showing characteristics of differential inhibition has been at-
tributed to various factors — for example, to maternal toxemia associated
with disease or other conditions — but practically nothing is known con-
cerning the particular factors involved in many cases. Results of experi-

270 PATTERNS AND PROBLEMS OF DEVELOPMENT

mental teratogeny lead us to believe that the determining factors may
differ widely even though the modifications are similar.

Apparently inhibition in earlier stages may produce cyclopia in verte-
brates; in later stages after normal localization of optic primordia, only
microphthalmia. The occurrence of cyclopia has suggested that the optic
primordium may be primarily median and become bilateral later and that
the inhibiting conditions merely prevent the secondary change in locali-
zation. This question is discussed later in connection with certain trans-
plantation experiments which bear upon it (pp. 282-85).

CONCLUSION

Differential modifications of development indicate presence of physio-
logical factors of developmental pattern which are very similar as regards
regional differences and relations in various animal groups, from coelen-
terates and flatworms to the higher vertebrates. The fact that a large
number of external agents, both chemical and physical, within certain
ranges of concentration or intensity, produce similar modifications of early
development with axial gradations in degree of modification indicates
that the physiological factors on which the modifications depend consti-
tute a quantitative gradient pattern, rather than a regionally specific
pattern of different substances and metabolisms. The possibihty that a
regionally specific pattern may also be present is, of course, not excluded
by the characteristics of the modifications; but if it is present, it is evi-
dently not the chief factor in determining the similar graded characteris-
tics of the differential modifications. Moreover, the alterations in locali-
zation of particular differentiations and the complete obliteration of axiate
pattern and all localized differentiation by differential inhibition and ob-
Hteration of the gradient pattern suggests that differentiations are de-
pendent on, and results of, this pattern.

The various lines of evidence considered in this and preceding chap-
ters indicate that axiate pattern is primarily a quantitative gradient pat-
tern, involving the essential activities of living. Within this general pat-
tern specific differences of substance and reaction arise and become the
basis of differentiations. With the changes in activity in the course of
development new centers of activity and new gradient systems may arise,
some of them with metabolisms different in character from that of the
primary pattern. As these differences increase, they affect activities in
other parts, and physiological interrelations and integrating factors be-
come increasingly effective and varied in character. Apparently, how-
ever, quantitative gradient factors may persist and be physiologically ef-

DIFFERENTIAL DEVELOPMENTAL MODIFICATION. Ill 271

fective from the beginning of development to maturity in many of the
simpler organisms and in some of the organ systems, even of vertebrates
and man. Such factors become evident, for example, in the reconstitution
of isolated pieces of hydroids, planarians, and many other forms; the
localization of the reconstituting hydranth or head evidently depends on
these gradient factors rather than on the specifically differentiated organs
in the piece. Gradient factors are also evident in the functional gradients
of many axiate organs — for example, the mammalian intestine (see
p. 164).

In many animal eggs specific regional differentiations and metaboHsms
are apparently already present in the cytoplasm at the beginning of em-
bryonic development, but quantitative gradient factors may also be pres-
ent and effective. All that we know concerning developmental physiology
indicates progressive increase of specificity of parts during the earUer
stages: the progress of so-called "determination of parts" and of visible
differentiation certainly constitute abundant evidence of this increase.
By a sort of physiological extrapolation to successively earlier stages of
development, to the ovarian development of the oocyte, to the beginnings
of reconstitution in the isolated piece, to the bud in its initial stages, and
to the aggregate of dissociated cells we seem to find it necessary to postu-
late a quantitative gradient pattern as the primary axiate pattern, and
many fines of experiment show the presence of such pattern. It is quite
unnecessary, however, to assume that every individual organism begins
its development with nothing but this primary gradient pattern. Obvi-
ously, this is not the case. The reconstituting planarian piece may con-
tain various organs; its polarity usually determines on which end the
head shah develop, and the ventrodorsality of the regenerated head is evi-
dently derived from that of the piece, but the new gradient pattern re-
sulting from activation and formation of new growing tissue at the an-
terior end makes over — -reorganizes — the piece. AH that we can learn con-
cerning the organization of eggs at the beginning of embryonic develop-
ment and the changes during the course of development is, of course, of
value in the analytic investigation of that kind of development; but it
should not be forgotten that this organization may be far from the pri-
mary developmental pattern and significant rather as a feature of a par-
ticular kind and stage of development than as representing fundamental
factors of developmental pattern. We have already learned that various
features of egg organization are results — effects — rather than essential
factors of the real pattern.

CHAPTER VIII

GRADIENTS AND FIELDS: DETERMINATION, DIFFER-
ENTIATION, AND DEDIFFERENTIATION

GRADIENTS, GRADIENT SYSTEMS, AND DEVELOPMENTAL FIELDS

THE term "physiological gradient" has been applied to spatial
patterns in living organisms characterized by a gradual progres-
sive differential in certain expressions of physiological condition.
That quantitative metabolic differentials seem to be the most conspicuous
features in early stages of many forms and that they are essential factors
in development is a justifiable conclusion from the data of chapters ii-vii ;
but concerning the particular chemical reactions and the substrate and
how they differ at different levels, we know little. However, if decrease
in concentration or amount of a certain substance or substance-complex
occurs in one direction along a gradient, there must be increase in con-
centration or amount of some other substance or substances unless there
is decrease in volume. A substance gradient decreasing from apical to
basal pole in an egg, for example, must be complemented by another sub-
stance gradient, different in character in the opposite direction. In many
eggs we find two such opposed substance gradients: the active metabohz-
ing cytoplasm decreases basipetally, yolk acropetally; but the resultant
in such cases, at least in early development, may be a single activity
gradient, decreasing basipetally. In some other eggs we find no direct
evidence of substance gradient; but a single activity gradient, also de-
creasing basipetally, is nevertheless present and unquestionably associated
with graded difference in the protoplasmic substrate. The presence of an
activity gradient, as determined or indicated by methods at present avail-
able, gives no definite information concerning quantitative or qualitative
character of differences in the substrate. Development as a continuing
series of changes is an expression of the dynamics of living protoplasms,
and very commonly its pattern in early stages appears to be represented
wholly or largely by the gradient system present. Specific or qualitative
material regional differences are apparently not necessarily concerned in
the earliest stages of developmental patterns, though there is no question

272

GRADIENTS, FIELDS, AND DETERMINATION 273

as to their significance in later stages. It is by no means necessary to
assume, nor is it probable, that the component reactions are the same in
different gradients even in the same individual. On the other hand, it is
probable that specific differences begin to appear almost at once in a
gradient primarily quantitative; but the important point, so far as early
developmental pattern and order are concerned, is that quantitative
gradient factors appear to be primary and specific regional differences to
develop gradually. The graded differentials in rate of development, oxi-
dation-reduction, electric potential, susceptibility to toxic action, etc.,
characteristic of these gradients give, of course, only general and partial
information concerning them; but since developmental order and pattern
show very definite relations to the gradient pattern of which these graded
differentials are partial expressions, it appears, beyond question, that the
physiological gradients are operative and effective factors in development
of axiate order and pattern. The metabolisms of different gradients in
the same individual may differ widely, but the gradient pattern is appar-
ently as essential to axiate development as the kind of reaction in it.

As will appear in following chapters, gradients in living protoplasms
can be determined experimentally by various environmental factors and
under natural conditions are often determined by local activation in rela-
tion to factors in the organismic environment of the part concerned.
Presence of differentiation is not necessary for their initiation. All that
is necessary is localization of an activated region in some way ; from this
the activation spreads, irradiates, and is transmitted with a decrement
in intensity. The activated region corresponds to a region of excitation,
and the resulting gradient to excitation transmitted with a decrement.
If the activated region persists long enough, it may determine a persistent
gradient, which in turn determines axiate development. Whatever the
factor or factors initiating activation or a gradient, specific constitution
of the protoplasm in which it appears and the physiological condition, as
determined by various factors, are, of course, the chief factors in deter-
mining the character of the gradient, steepness of decrement and effective
length, as well as the kind of physiological activity characterizing it and
the kind of development resulting from it. As in nervous excitation and
transmission, the activating factor initiates and the protoplasmic mecha-
nism determines character of effect.

A gradient in a certain direction appears to constitute a physiological
basis for definite order and sequence in development in that direction, and
a gradient system seems to serve as a sort of a physiological co-ordinate

2 74 PATTERNS AND PROBLEMS OF DEVELOPMENT

system with reference to which axiate pattern develops. The question
whether and how specific differences can arise at different levels of a gradi-
ent is considered in another connection (p. 297) ; but, assuming for the
moment that this is possible, the conception of a gradient system as a
physiological co-ordinate system requires some further consideration, A
radial gradient system, such as that of many buds, results in a radial de-
velopmental pattern which becomes polar in consequence of differential
growth (p. 16); and a longitudinal system, like that of a Corymorpha
piece, results in a longitudinal developmental pattern, within which par-
ticular organs or parts arise at particular levels and where scale of organi-
zation can be increased or decreased by altering height and length of the
gradient (pp. 38, 344-57). But a question at once arises: Can two or
more gradients or gradient systems in two or more different directions in
the same protoplasm constitute a co-ordinate system which specifically
determines each region? In other words, can such a system determine the
developmental pattern of organisms with ventrodorsality or dorsiventral-
ity and often asymmetry? If the two or more gradients are of the same
kind and intensity, the chief axis of the pattern determined will be in the
direction of the resultant of the two. In the accompanying diagram (Fig.
107, A) anterior, posterior, dorsal, and ventral are indicated hy A,P, D,
and V. The anteroposterior gradient is indicated by the numerals 4-1
above; the ventrodorsal gradient, by the numerals on the right. The nu-
merals of the different areas are the sums of those indicating respective
levels of the two gradients; for the sake of simplicity the two gradients
are assumed to be additive in effect. The diagram shows that the chief
pattern differences are in the direction of the heavy lines oblique to both
gradients. The broken obUque lines indicate directions along which pat-
tern is the same. This is not the sort of pattern usually found in organisms
with a longitudinal and ventrodorsal axis.

In Figure 107, B, the anteroposterior gradient is assumed to be opera-
tive and to have induced some degree of specific differences, a-d, at dif-
ferent levels before the ventrodorsal gradient becomes markedly effec-
tive, its effectiveness being assumed to increase gradually. In this case
each area is in different physiological condition and is defined by its rela-
tion to the two gradients, and the pattern has an anteroposterior and
ventrodorsal axis. If the ventrodorsal gradient is primarily quantitative,
its effect on the specifically different regions a-d may differ in character
because of their differences and may determine further alteration. Fig-
ure 107, B, serves merely to suggest in a general wav how a developmen-

2 ^ 3

Xw^ ^

X. **

X. •

X^^ 1

Xv •

Xv •
X^x

• X.
• \

y 4 \

X. •
X^ •

X^^^X

x'x. \

x^ ^

X. ^

\

X. ^

\ •

X^ x'^
xy*

X. •

V>/'»

F

A

c^ ,

/d\

c\

b\

/ d2

C2

hi

a2 \

\ d^

C3

hz

a3 /

\dA

CA

bA

aAyA

A

V

B

Fig. 107, .1, B. — Diagrams of effects of two gradients in different directions; explanation in

text.

276 PATTERNS AND PROBLEMS OF DEVELOPMENT

tal pattern with anteroposterior and ventrodorsal axes — or, if the second-
ary gradient were in the opposite direction, a dorsi ventral axis — may arise
from a pattern consisting of two gradients at right angles, one becoming
operative or being much more effective before the other. As a matter of
fact, the ventrodorsal or dorsi ventral gradient usually becomes directly
evident by the methods at present available, at a later stage of embryonic
development than the anteroposterior gradient, although certain experi-
mental data indicate that something constituting a basis for it may be
present in the unfertilized egg but with little or no visible effect on early
developmental stages. As the diagram is drawn, with the high end of the
secondary gradient becoming the ventral region, it might be regarded as
representing a plane projection of early planarian or annelid pattern.
With the high region dorsal, it might perhaps be regarded as resembling
the pattern determined in the amphibian Hmb bud by its relations with
the pattern of the body.

In recent years the field concept has been applied to various develop-
mental phenomena; but before the word "field" came into use in develop-
mental physiology, we find what is essentially the field concept stated
in other terms. For example, Spemann (1912a) suggested that early em-
bryonic potency to develop an organ such as the lens of the amphibian
eye involves an area analogous to a diffraction circle, the degree of deter-
mination being highest in the center and decreasing peripherally. Harri-
son (19 1 8) says with reference to amphibian limb development:

The limb rudiment may be thus regarded, not as a definite circumscribed area like
a stone in a mosaic, but as a center of differentiation in which the intensity of the
process diminishes as the distance from the center increases until it passes away into
an indifferent region. Many other systems, such as the nose, ear, hypophysis, gills,
seem to have the same indefinite boundaries which may even overlap each other.

Harrison is speaking here not merely of the limb bud itself, the locus of
actual development of the limb or other organ, but of the whole area
about this locus, which is found by experiment to be more or less capable
of giving rise to the organ. The concept of the developmental or morpho-
genetic field has been further developed and applied by various authors
independently of, or in relation to, the gradient concept; and, as is usual
in such cases, reference of certain developmental activities or results to
a field seems sometimes to be regarded as advancing our knowledge of
developmental physiology.' Without further analysis the field concept,

•Spemann, 1921; Gurwitsch, 1922, 1923, 1927; Weiss, 1924, 1926, 1928, 1939; Guyenot et
Schotte, 1926a; Guyenot, 1927a, b; De Beer, 1928; von Bertalanffy, 1928; Guyenot et Ponse,
1930; Huxley, 1932; Waddington and Schmidt, 1933; Huxley and De Beer, 1934; Dalcq, 1938.

GRADIENTS, FIELDS, AND DETERMINATION 277

as applied to development, has only a formal value. Identification of a
hydranth field, a limb field, an eye field, or, in general, reference of certain
developmental phenomena to a field indicates merely presence and a cer-
tain order of capacities or potencies for these phenomena in a certain
region but gives us no information concerning the physiological character
of the order or the potencies, or the conditions determining realization of
potencies in a particular part of the field. In other words, reference to a
field merely states experimental data in terms of an unknown, of a con-
cept without definite content, and the field often becomes little more
than a verbalistic refuge.

The developmental field concept implies an ordering or controlling fac-
tor or factors of some sort. Evidently, however, many developmental
fields include differences in actual orders and patterns. For example, in
the case of the amphibian limb the area capable, or becoming capable
under experimental conditions, of developing a limb is considerably more
extensive than the field of actual limb development in any particular case,
and the same is true for various other organs and organ systems. Obvi-
ously, the potency field and the field of actual differentiation differ in
some way and must be distinguished. The individuation field of Wad-
dington and Schmidt (1933) does not throw any light on individuation;
and, as the authors apply it to vertebrate development, it seems to sug-
gest that individuation results from action of the chorda-mesoderm as
inductor, but actually individuation is present before induction and the
chorda-mesoderm is itself a part of the individual.

If the physiological gradients are operative factors in development, the
question of the relation of gradients and dominance to developmental
fields is important. According to Huxley and De Beer (1934, P- 274),
it is found that the original control of differentiation in all cases appears to be exerted
in relation to what may be called a biological or morphogenetic field. Within these
fields various processes concerned with morphogenesis appear to be quantitatively
graded so that the most suitable name for them is field-gradient systems or simply
gradient-fields.

This statement raises several questions, for it seems to imply that the
field is the factor determining differentiation; yet the field is regarded as
a gradient field — in other words, as a gradient system. What is the field,
as distinguished from the gradient or gradient system in it? Does the
field determine the gradients, or do gradients constitute the field? How
does the field determine or control differentiation? What part does domi-
nance play in the field?

278 PATTERNS AND PROBLEMS OF DEVELOPMENT

The experimental evidence concerning physiological gradients seems to
indicate that developmental fields in their simplest, most general forms
are gradient systems, that is, the field is constituted by the gradient or
gradients present ; the gradients are the vectors of the field and determine
its extent and the orderly relations within it. Such a field may originate
in a region of localized activity which determines a gradient or gradients.
In the adventitious bud field of a plant (see Figs. 1-4) the primary region
of activity is the physiological center of the field and approximates, in
this case, the geometric center; and the gradient system or field is at
first radial and becomes axiate and polarized in consequence of differen-
tial growth. Such a field may be larger than the actual bud which develops
from its higher gradient levels. The earlier stages of these adventitious
buds of plants and many buds in animals, whether they develop into com-
plete individuals or parts, such as tentacles, appendages, etc., may be re-
garded as examples of simple developmental fields. The extent of the re-
gion of activation, that is, the extent of the radial gradient system, may
or may not be greater than the field of actual differentiation. These are
evidently gradient fields, and the field is the gradient system.

Realization of field potencies in actual development is localized in the
high region of the gradient or gradient system constituting the field simply
because this region is the most intensely active. Experimental activation
of some other region of the field may bring about development there, pro-
vided it is not inhibited by the dominance of the physiological center. In
pieces of Corymorpha stem activation of the cells adjoining the level of
section determines a gradient or a system of parallel gradients extending
for a certain distance (see Fig. 31). This is a reconstitution field; and
within it parts of the new hydranth are localized, and alteration and
elongation of the stem occurs, so far as length of piece and scale of or-
ganization, as determined by intensity of activation and length and slope
of the resulting gradient, permit. A similar reconstitution field arises
following section in pieces of planarians, nemerteans, annelids, etc., in
relation to the activated region which becomes a head. These fields are
axiate and polar from the beginning. The dominant region is the high
end of the gradient system and is an inductor in the sense that it is the
primary factor in determining the gradient system, and consequently the
levels at which particular parts shall develop. Such fields as these are
fields of actual development, differentiation fields, and represent acti-
vated regions of the potency field, which in Corymorpha is the whole

GRADIENTS, FIELDS, AND DETERMINATION 279

stem. During the course of reconstitution in Corymorpha new fields — for
example, the tentacle fields — arise at certain levels of the gradient system.
These also appear to be primarily gradient systems like those of buds;
but since they determine particular organs rather than complete individ-
uals, the gradient activities in them doubtless differ in some way from
those in the original reconstitution field in which they arise. The appear-
ance of such secondary fields within the primary field indicates that the
latter becomes, sooner or later, something more than a simple quantita-
tive gradient system. Apparently specific differences of some sort appear
at certain levels, and some of these determine local activation with forma-
tion of organ fields at those levels. At present it is difficult to describe the
reconstitution of Corymorpha in any other terms than these. The gradi-
ents arise as the earliest distinguishable evidences of development and
can be made visible in the Hving animals. They may be determined at
any level and at either end of a stem piece by section, by a graft from
any stem-level, or by a lacerated wound; and their length may be altered
and controlled experimentally by environmental factors with correspond-
ing alteration of scale of organization (see chap. x). As already pointed
out, such forms of development come nearer providing a starting-point
for physiological analysis of development in general than does embryonic
development from an egg already more or less highly differentiated be-
yond the earlier stages of pattern.

Symmetry or asymmetry of bud fields and organ fields arising in rela-
tion to pre-existing gradients in the body from which they develop may
be determined or influenced by these gradients. The fields of origin of
new cilia and cirri in the division and reconstitution of various ciliate
protozoa are evidently related in a definite way to the pattern of the
original individual;^ and some evidence concerning that pattern is given
by the ectoplasmic gradient in rate of dye reduction and susceptibility
in these forms. The "dorsiventral" or bilateral pattern of tentacle devel-
opment in the bud of Pelmatohydra is apparently determined by the gra-
dient and dominance in the parent body (Rulon and Child, 1937). The
anteroposterior and dorsiventral pattern of the amphibian limb arises
in definite relation to anteroposterior and dorsiventral pattern of the body.

The question of the relation of metabolism to developmental pattern,
and particularly to the developmental field as this is conceived by some.

- See, e.g., Dembowska, 1925, 1926, and citations.

28o PATTERNS AND PROBLEMS OF DEVELOPMENT

is of fundamental importance. The following discussion of this question
is quoted from a recent paper.

Some biologists apparently believe that metabolism is not a fundamental factor
in development. For example, Shearer [1930, p. 266] says: "morphological organiza-
tion has nothing to do with metabolism." Parker [1929, p. 424], in criticizing the
gradient concept, makes the statement: "the metabolic activity of the organism is not
a true formative process, but the result of such a process." Spemann [1939, pp. 321 ff.]
seems to hold somewhat the same opinion; his discussion of gradients involves the mis-
taken assumption that, according to the gradient theory, there must be only quantita-
tive metabolic differences in the amphibian egg at the beginning of embryonic develop-
ment. In his recent book, Weiss [1939, pp. 373-83] seems to hold that specifically
different capacities for organization exist in different regions independently of metabo-
lism. Granting that these capacities are present in many eggs, how did they arise
except through metabolism? And even if they are present in eggs, are they present in
the early stages of buds, of pieces of Tubularia or Corymorpha stem or in aggregates of
dissociated Corymorpha cells? In isolated pieces of the planarian body, of various
annelids and of some ascidians development of a particular part has no definite or
constant relation to a particular region of the individual or to a pre-existing organiza-
tion, but a new organization originates. Does it originate independently of metabo-
lism? If we inhibit metabolism it does not appear and in many cases it is possible to
initiate its development by purely quantitative external differentials which in one way
or another determine gradients involving metabolism. It has been pointed out re-
peatedly that, even if these gradients are primarily quantitative, they probably do not
remain so for any great length of time, though quantitative factors may continue to
exist and be effective, even in many adult animals. How do the regional organizing
capacities and the presumably specifically different metabolisms of hydranth, stem
and holdfast region in a reconstituting piece of Corymorpha originate if not through
metabolism? Since any level of a new individual may develop from any level of the

stem these capacities are certainly not localized preceding isolation of a piece

The arguments of Spemann and Weiss are based on the egg and embryonic develop-
ment rather than on development in general Like many other embryologists

they maintain that organization or specific capacity for it is present, but do not tell us
how it originates. If organization consists in localized presence of specific substances
how can these substances originate and be localized except in the final analysis through
metabolism? Probably no one now believes that they are aU present in the primitive
germ cell. It is difiicult to believe, for example, that localized formation of specific
substances in the dorsal inductor region of the amphibian egg and embryo can take
place independently of metabolism. Without metabolism ovarian development of the
oocyte and embryonic development cease. Even the differentiation of various organs
and tissues does not persist if their metabohsm is decreased below certain levels. How
are electric and other physiological regional differences established and maintained
except through metabolism? What is the possible nature of formative processes as-
sumed to be independent of metabolism? In short, is there any more fundamental
characteristic of hving protoplasms than metabolism? At present evidence of any such
characteristic seems to be whoU}- lacking. It may be argued that structure of some
sort is more fundamental, but structure without metabolism is not living protoplasm

GRADIENTS, FIELDS, AND DETERMINATION 281

and accomplishes nothing that can be regarded as organismic development. Even if
the organization of the egg originates in the orientation of elongate dipolar and sym-
metrical or asymmetrical protein molecules, as Harrison [1937] and others have sug-
gested, the question at once arises: to what do the molecules orient? Metabohsm is
going on continuously in the ovarian oocyte and in the parent organism to which it
is attached. Can the orientation possibly occur without relation to this metabolism?
In reconstitution of a piece of Corymorpha or of a planarian the new organization fol-
lows certain changes in metabolism in the region concerned and without them does not
develop. Doubtless changes in structure are also involved and affect metabolism, but
metabolism is the active and effective factor. At present there is no evidence that the
changes in metabolism characteristic of the new organization result from orientation
of protein molecules in all the thousands of cells concerned. The polarities of the indi-
vidual cells of the hydroid and planarian are apparently determined by local surface-
interior differences without relation to the axiate pattern of the whole, but these cells
become parts of the axiate pattern of the new individual and still retain their original
polarities. If their molecules are, or become oriented, are they oriented with respect
to the local surface-interior factors or to the axiate pattern of the whole individual?
If they are locally oriented then the new pattern is independent of them. If the new
pattern is determined by their reorientation, how are their original polarities main-
tained?

Weiss, in his book [1939], has much to say of the field concept, of field energy, of
strong and weak fields and of decrease in field energy from a center, but he does not
tell us what a field is or may be as an active and effective factor in development, nor
does he say what makes a field strong or weak or what the source and character of its
energy may be. What is the basis for decrease of field energy from a center to periphery
of the field? New fields originate in the course of development: how do they arise?
Is their origin independent of metabolism? Is there some other source of energy in a
field than metabolism? Without information, or at least hypothesis, concerning these
points the field concept remains almost mystical in character.

That organization with localized specific differences of substance and metabolism
does originate in some way and that dynamic, not merely structural factors, are
essential to its origin appears beyond question; that regional specificities originate
and increase during development is indicated by many lines of evidence. If metabolism
is merely incidental to, or a result of these changes, it would seem that we must
again postulate a specific vital energy: Driesch's entelechy wiU scarcely serve our pur-
pose, for that was conceived, not as a source of energy, but rather as controlling energy
transformation, and that in living protoplasm is in the final analysis, controlling
metabolism.

The concept of physiological gradients in terms of dynamic factors effective in
bringing about development, rather than of purely structural factors, does not in any
way conflict with the concept of organization in terms of specific, regionally localized
material differences. It merely maintains on the basis of many lines of evidence
that such organization is not the primary pattern of development, but the result of
metabolic activity in a primarily quantitative pattern from which the regional specifi-
cities gradually developed. In short, this concept is an attempt to look beyond organi-
zation already present and to throw some light on the problem of its origin and develop-
ment [Child, 1940, "Lithium and echinoderm exogastrulation," Physiol. ZooL, 13].

282 PATTERNS AND PROBLEMS OF DEVELOPMENT

The many examples of embryonic developmental fields given by Huxley
and De Beer (1934) and by Weiss (1939) make it unnecessary to do more
than call attention to a few points of interest here.

THE EYE FIELD OF AMPHIBIAN AND CHICK

His earlier experiments led Spemann to believe that the optic primordia,
even those of cyclopian eyes, are determined and localized in the neural
plate as a mosaic of independently developing parts.' Later experiment
showed the incorrectness of this view (Spemann und Bautzmann, 1927).
The optic primordium was regarded by Stockard (19 13 and earlier papers)
as first median, spreading laterally and giving rise to two growth centers.
According to this view, cyclopia in the amphibian and probably in other
vertebrates is fundamentally different from cyclopia in a planarian and
from the approximation to, and development in, the median plane of the
anal arms of the sea-urchin larva with differential inhibition. There are
no grounds for believing that the primordium of the planarian eyespots
is originally median, and the primordium of the sea-urchin arms certainly
is not. The starfish coeloms become median and single with differential
inhibition (p. 219), but no reason appears for postulating an originally
median primordium. Moreover, the suckers of the anuran head and the
olfactory pits show the same graded approximation to the median plane,
single median development, and, with extreme inhibition, absence. Here,
also, there is no ground for postulating an originally median primordium.

More recent investigation shows the existence in Amhlystoma embryos
of an eye field including anterior, median, and lateral regions of the neural
plate back to a certain level.'' Pieces from different parts of this region
transplanted to the belly wall give rise to eyes. According to Adelmann
in transplantation of pieces without mesentodermal substrate eye develop-
ment is much more frequent in pieces from the median region (70.8 per
cent of transplants) than in lateral pieces (ii.i per cent), that is, eye
potency is apparently much higher in the median than in the lateral
regions of the field. In transplants with the underlying tissue which in-
duced development of the neural plate 72.6 per cent of the median and
54.4 per cent of lateral pieces develop eyes; that is, the underlying
mesentoderm increases eye frequency in lateral but not in median pieces.
Moreover, of the median transplants with substrate, 70.8 per cent give

3 Spemann, iqokz, 1903a, 1904, 1912a, h; see also Fischel, 1921, and H. Petersen, 1924.
'• Adelmann, 1929a, b, 1930, 1934, with references to earlier work; also Leplat, 1919; Man-
chot, 1929.

GRADIENTS, FIELDS, AND DETERMINATION 283

rise to two eyes. Also, removal of the mesoderm underlying the anterior
part of the neural plate soon after it attains that position results in a high
frequency of cyclopia or median approximation of eyes (Mangold, 1931a,
p. 365). Adelmann concludes that eye potency is higher in the median
region than laterally and that the underlying inductor determines bilateral
eye development. The following suggestion perhaps serves to bring verte-
brate cyclopia and related modilications more nearly into line with similar
modifications in other forms. The inducing action of the underlying mes-
entoderm unquestionably brings about a change in physiological condition
in the ectoderm involving activation, whatever its other effects may be.
The mediolateral decrease in inducing capacity of the inductor tissue
(p. 459) makes it probable that the change in condition of ectoderm re-
sulting from induction is more rapid and greater in the median region
than laterally. In the course of this change the median region passes
through a stage of physiological level which represents capacity for eye
formation; and since it attains this condition earlier than lateral regions,
it may, when physiologically isolated by transplantation at certain stages,
give a higher frequency of eye development than lateral ectoderm. Under
natural conditions, however, the median region is prevented, by its rela-
tions with other parts, from developing an eye or eyes at the stage when
capacity for such development is present; instead it undergoes further
change in condition, probably with further activation, and finally becomes
part of the brain floor, while lateral regions finally attain the condition
representing full capacity for eye development. In other words, what de-
velops under natural conditions represents, in general, the full or final
effect of induction; the eye potency of the median region represents
merely a temporary condition intermediate between the condition pre-
ceding induction and the final condition. Eyes are normally bilateral be-
cause the final conchtion constitutes the action system initiating eye de-
velopment in lateral regions at a certain distance from the median plane
and at a certain level of the anteroposterior axis. When the underlying
tissue is present in the transplant, the median region may be activated
above the level determining eye development; consequently, two eyes,
bilaterally localized, develop, the distance from the median plane varying
with conditions in the individual transplant. In similar transplants with-
out underlying inductor tissue the level determining eye development is
usually present only in the median region, and a single eye is formed.
With differential inhibition by toxic agents optic primordia are locahzed
nearer or in the median plane because the more lateral regions never at-

284 PATTERNS AND PROBLEMS OF DEVELOPMENT

tain the eye-level; with increasing inhibition this level is localized pro-
gressively nearer the median plane or becomes median because all lateral
regions fail to attain the eye-level. Apparently, however, approximation
of eyes and cyclopia cannot be regarded as resulting entirely from differ-
ential inhibition of the inductor; inhibition of eqtoderm seems also to be
concerned. Even the maximum inducing or activating action may not
bring any but the median part of the inhibited ectoderm up to the eye-
level. The occurrence in fishes of cyclopia with normal brain indicates
inhibition largely or wholly ectodermal.

This suggested interpretation not only avoids the assumption of specifi-
cally different actions of closely adjoining median and lateral regions of
the mesentoderm but also accounts for the change in localization of eye
capacity or potency in the course of development and in differential inhi-
bition. According to it, eye potency and optic primordia are not the
same: potency may be present in an extensive area, though not neces-
sarily at the same developmental stage in all parts of the area, but optic
primordia originate where eyes actually develop. The eye field appears
to be primarily a gradient system resulting in large part or wholly from
induction in the ectoderm and subject to experimental alteration with
altered localization of optic primordia.

The eye field of the chick embryo at certain stages, as indicated by
chorio-allantoic grafts of different regions of the blastoderm, is also a bi-
lateral field including median and lateral regions.^ At these stages eye
development or differentiation of eye tissue occurs in both median and
lateral pieces from a certain blastodermal level, but more frequently and
with more advanced differentiation in median pieces.'' A slight asym-
metry of the field is indicated by a larger amount and more advanced dif-
ferentiation of eye tissue from left than from right lateral grafts. The
presence of a mediolateral gradient of susceptibility and rate of dye re-
duction in the chick blastoderm has been shown (pp 159-62). Except as
regards the asymmetry, the interpretation suggested above applies here
also. If it is correct, the mediolateral extent of the eye field in the am-
phibian and bird, the higher eye potency in transplants of median regions
at certain stages, and the approximation of eyes and cyclopia under in-
hibiting conditions are all expressions of a mediolateral pattern, primarily
a quantitative gradient pattern rather than regionally specific; and ap-
proximation of eyes and cyclopia in vertebrates do not differ in principle

5 Clarke, 1936; Rawles, 1936. See also Figs. 167, 170, pp. 531, 534.

GRADIENTS, FIELDS, AND DETERMINATION 285

from approximation to the median plane of other bilateral organs in ver-
tebrates and invertebrates under inhibiting conditions.

One other point requires notice in connection with these transplanta-
tion experiments on parts of the eye field. It is by no means certain that
isolation and transplantation of a piece of the embryo is possible without
any alteration of physiological condition of the piece. In the light of ex-
periments with hydroids and planarians it appears possible that a tem-
porary stimulation may follow isolation and perhaps be followed in turn
by a depression. If the pattern is a gradient pattern, such changes will
certainly play a part in determining what develops from the transplant.
There may be differences in susceptibility to the altered environment in
pieces from different regions of the field, and these may influence the de-
velopmental result. In short, potency for eye development or for any
other development may depend as much on the environment of an isolated
or transplanted piece and the operative eft'ects on it as on the condition
of the region concerned before operative procedure.

THE AMPHIBIAN LIMB FIELD

In consequence of the great amount and variety of experiment more is
perhaps known concerning fields of organ systems and organs in amphib-
ian development than in other forms. Among these the amphibian limb
field has received much attention and serves as an excellent example.
The limb arises laterally on the body; and the potency field, the region
in which limb development can occur, is in earlier stages more extensive
than the region of actual development in any particular case. Its physio-
logical center, where the potency is highest, and where the limb normally
develops, is apparently nearer the anterior than the posterior border, that
is, in the higher levels of the portion of the anteroposterior gradient in-
cluded within the field. In the primitive limb disk the anterodorsal quad-
rant is apparently to be regarded as the dominant region (Swett, 1923),
suggesting that perhaps the dorsiventral gradient of the body may also
be concerned in its localization. The frequent duplication or triplication
of limbs in connection with transplantation of limb buds and regeneration
of limbs, evidently a result of the appearance of more than one physiologi-
cal center, is discussed in another connection (pp. 390-95). As others have
pointed out, the limb field is not established all at once; but certain of its
characteristics appear at one stage, others at another. It is not a static
pattern but a product of continuous progressive change. It is at least a
pertinent question whether anything other than a metabolic pattern with

286 PATTERNS AND PROBLEMS OF DEVELOPMENT

progressively increasing regional specificity can account for the establish-
ment and development of this field. In Amblystoma the pattern of the
limb bud is determined in the anteroposterior direction of the body earlier
than the pattern in the dorsivejitral direction; and both of these before
the mediolateral, that is, the longitudinal or polar axis of the limb/'
Anteroposterior and dorsiventral axes are apparently determined about
the same time in anura and in Triton, but experiment on these forms is
less extensive than on Amblystoma.'^ Determination evidently follows
much the same course in the regenerating as in the original limb bud.
Apparently, the anteroposterior and the dorsiventral gradient pattern of
the body are imposed on the hmb primordium. In Amblystoma the dorsi-
ventral pattern is less effective and becomes fixed in the Hmb primordium
later than the anteroposterior. The longitudinal axis of the limb, how-
ever, is apparently determined, as in other buds that become axiate, by
the gradient system in the bud itself. The radial decrease in developmen-
tal activity about the physiological center becomes a longitudinal gradi-
ent as outgrowth occurs.

That there is a real effective dominance in the limb field is suggested
by the fact that a limb region removed from its original position and im-
planted on the flank of the same or another animal is usually inhibited
or resorbed if within a certain distance of the limb regenerating from the
original site or of a normal developing limb, but develops if at a greater
distance (Detwiler, 1918; Hellmich, 1930). It might perhaps be expected
that the transplanted limb region, since it represents the region of highest
potency in the field, would dominate and inhibit regeneration from the
original site. Its failure to do this may result from depression brought
about by removal and implantation in a new environment, or the activa-
tion in the original site following its removal may be sufficient to inhibit
it. Even though polarity of the hmb in regeneration is not determined by
the proximal stump or by the potency field of the limb, the limb field
plays a part in determining the character of distal regenerating or trans-
planted parts. The regenerating outgrowth on the stump of an amputated
fore leg transplanted in an early stage to the stump of an amputated hind
leg may develop as hind leg, or that from a hind leg as fore leg when
transplanted to a fore-leg stump (Milojevic, 1924). It may be suggested
that the determining factor in this case is the difference in the asymmetry

'' Harrison, 1921(7, b, 1925a; Svvett, 1926, 1927, 1928a, b, c, 1932, 1937a, b, 1938a, b, c, 1939;
Swett, 1936, experiments on hind limb. See also pp. 390-95.

7 Graper, 1922a, b, 1923, 1924, 1925, 1926a, b; Brandt, 1924a, b, 1925; Milojevic, 1924.

GRADIENTS, FIELDS, AND DETERMINATION 287

patterns of fore-leg and hind-leg stumps, for this asymmetry pattern pre-
sents a differential or gradient pattern at the cut surface of the stump
which may be imposed on the regenerating tissue and perhaps also deter-
mines the asymmetry pattern of the transplant. But the regenerating
outgrowth on a limb stump is not limited or definitively determined in its
earlier stages as regards development into a limb, for it may develop into
a tail-like structure if transplanted to the tail region (Guyenot, 1927;
Guyenot et Ponse, 1930). According to Weiss (1927^), the regenerating
outgrowth of an amputated tail transplanted to a Hmb stump may de-
velop as hmb. In these cases, also, the symmetry pattern of the tail or the
asymmetry pattern of the limb is probably the factor determining the
character of development. However, Liosner and Woronzowa^ find a con-
siderable specificity in transplants of muscular tissue from tail to limb, etc.

Implantation of the limb region with more or less rotation in the dorsi-
ventral plane from normal orientation is sometimes followed by rotation
to normal orientation.'^ This reorientation apparently occurs in relation
to the region from which the shoulder girdle develops, not to the body as
a whole, and has been regarded as in some way associated with the shoul-
der girdle. A small implant {ih somites) which develops no shoulder girdle
and a large implant (5 somites) which develops a complete girdle do not
rotate, but an implant of intermediate size (3I somites) which develops
part of a girdle may rotate. If a 3^-somite region and a ring of tissue
representing the region around it are separated and implanted with inde-
pendent and different rotation, the limb region rotates to normal orienta-
tion with the ring. The question arises whether the girdle as such or the
gradient pattern is the factor determining rotation. Undoubtedly, both
the asymmetry gradient pattern of the part undergoing rotation and that
of the part on which it rotates must have attained a certain degree of de-
velopment in order that rotation may occur. The fraction of the pattern
in the i^-somite piece may be insufficient to bring about reaction, and the
5-somite piece may fail to rotate because of physical conditions associated
with its size or because it possesses sufficient pattern to alter adjoining
regions to some extent instead of rotating in reaction to them.

It appears highly probable, in view of the data, that the gradient sys-
tem or field about the transplanted or regenerating limb primordium does,
or may, influence it and impose pattern on it, or determine postural rota-
tion. The fact that an anteroposterior asymmetry of the limb coincides

* Liosner and Woronzowa, 1937; Liosner, 1938; Woronzowa, 1938.
'Harrison, 1921a; Nicholas, 1924, 1925, 1926.

288 PATTERNS AND PROBLEMS OF DEVELOPMENT

approximately in direction with the anteroposterior axis of the body and
is apparently the expression in the localized Hmb region of that pattern
is perhaps significant as indicating that polarity and asymmetry are not
fundamentally different in character; something constituting polarity in
the body as a whole determines an anteroposterior asymmetry in the
limb; similarly, something constituting dorsiventrality and bilaterahty in
the body determines a dorsiventral asymmetry in the limb. It seems diffi-
cult to account for facts such as these except in terms of gradient pattern.
Apparently, work is done in the postural rotation of an implant, and this
impHes a dynamic factor of some sort.

Within the potency field of the limb development of a limb can be in-
duced not only by implantation of part of a limb bud but of an otic vesi-
cle, an olfactory placode, brain tissue, eye, or even a piece of celloidin
and also by a nerve deflected to a region of the field. The induced limb
appears later the farther posterior the position of the inductor and reac-
tivity to an inductor disappears from anterior to posterior regions of the
field progressively with advance of development.'"

The inductor, whether implant or nerve, apparently serves in these
cases merely as a nonspecific activator, and its action seems to be pri-
marily on the mesenchyme. The developmental result is determined not
by the inductor but by the polar gradient or gradients resulting from the
locaHzed activation and outgrowth, the character of the field in which the
outgrowth takes place, and the asymmetry pattern representing a part
of the general body pattern. Either ectoderm or mesenchyme of the
axolotl limb bud at certain stages may determine limb development with
foreign mesenchyme or ectoderm, according to Filatow (1930a). This may
mean that both are activated sufficiently so that cither can act as a domi-
nant region.

SOME OTHER FIELDS

Experiment has shown the existence of various other fields in amphib-
ian development— ear, gills, urodele balancer, etc. — and in many of them
gradient characteristics appear. Like the limb and eye fields, the potency
field of the ear is more extensive than the differentiation field in earlier
stages. Otic development induced in other parts of the field than the
physiological center is more complete near the normal differentiation field
than at greater distances, and the anteroposterior axis is determined ear-
lier than the dorsiventral." Implantation of the presumptive ear region

■" Detwiler, 1918; Balinsky, 1925, 1926, 1927, a, b, 1933, 19370; ^" Filatow, 1927.
" Yntema, 1933; Harrison, 1936a; Albaum and Nestler, 1937; Hall, 1937, 1939.

GRADIENTS, FIELDS, AND DETERMINATION 289

of Amhly stoma in the normal differentiation field of the ear (orthotopic
implantation) with various degrees of rotation may result, at certain
stages, in reversal of both anteroposterior and dorsiventral axes, or one
may become the other. It seems evident that these changes in pattern
are imposed on the implant by the pattern about it.

The field of the larval urodele balancer in the ventral head ectoderm
is apparently not axiate, and the balancer itself appears to be radially
symmetrical. A radial potency gradient decreasing from a physiological
center has been demonstrated; ectoderm from this region, but not other
ectoderm, transplanted to other head regions at certain stages, will de-
termine development of a balancer with mesoderm from the region of im-
plantation. In later stages just preceding appearance of the balancer bud
transplanted ectoderm of this region will determine balancer develop-
ment in trunk as well as in head regions. Harrison (19256) regards this
difference in inducing capacity as indicating increase in specificity, but it
seems entirely possible that it may indicate merely a more intense activa-
tion of the balancer ectoderm at the stage when outgrowth is about to
begin. Development of supernumerary balancers in the field can be in-
duced by implants of various tissues, neural plate cells, foregut cells, even
tissue from urodeles which have no balancer (Mangold, 193 16) and cells
of the anuran neural crest (Raven, 193 1, 1933&). That the inducing agent
is anything more than a nonspecific activator here, as in limb induction,
seems improbable.

These examples are sufficient to indicate the character of field phenom-
ena in amphibian development where they have been most studied. That
the potency field is primarily a gradient system of a certain kind and that
the normal differentiation field represents its high region is indicated by
various lines of experimental evidence. The limb field, the eye field, the
ear field, appear in the course of development, but their patterns may
not be established all at once. The hmb field is a region from which limb
develops before the pattern of a particular limb is finally fixed, that is,
the pattern of the limb may be altered by transplantation to an altered
physiological environment. According to data at hand, amphibian po-
tency fields, as areas more extensive than the fields of actual differentia-
tion, undergo progressive restriction in most cases with progress of de-
velopment and finally disappear completely. The possibility remains,
however, that under other experimental conditions potencies might ap-
pear which now seem to be absent. The disappearance of a potency field
is evidently associated with the progressive differentiation of its different

290 PATTERNS AND PROBLEMS OF DEVELOPMENT

regions in different directions. In many hydroids and planarians there
is no progressive limitation of potency except in the hydranth or the
head.

At present it appears that developmental fields are not sharply and
definitely bounded in their earlier stages, and it has often been suggested
that different fields may overlap so that a certain region may be part of
two fields at the same time, but direct proof is lacking. If fields are pri-
marily gradient systems, clearly defined boundaries are not to be ex-
pected, and extent of a field may increase or decrease with the activity
in it. If such changes occur, certain regions may be parts of different
fields at different times. Within the field, as in development in gen-
eral, stable structure and differentiation appear to be secondary and to
result from a pattern primarily nonspecific and labile with a graded meta-
bolic differential. That a field in its most general form is anything more
than a quantitative gradient system of a particular kind remains to be
proved and in the light of experimental evidence appears improbable.

HARMONIOUS-EQUIPOTENTIAL SYSTEMS

Driesch introduced the concept of harmonious-equipotential systems
and in one of his papers characterized them as follows: "Each part can
give rise to any part ['jedes Element kann jedes'] and each effect [i.e.,
developmental result] occurs only once or a definite number of times and
in a fixed relation to all other effects" (Driesch, 1899a, pp. 73-74)- He
regarded the Tiihularia stem and cleavage stages of the sea urchin as such
systems and later discovered what he beheved to be other systems of the
same sort. The term has since been applied to various developmental
systems, among them the amphibian limb field. The existence of these
harmonious-equipotential systems was regarded by Driesch as evidence
for "autonomy of vital processes," that is, for so-called "vitaUsm." It is a
fact amply confirmed by many investigators that any level of the Tubu-
laria stem can develop into any body-level of Tuhularia. This is also
true for Corymorpha and many other hydroids and for the postcephalic
regions of various planarians, nemerteans, and annelids. However, the
various body-levels, even the levels of the Tuhularia and Corymorpha
stems, are equipotential only specifically, not quantitatively, at any given
moment, that is, any of them can give rise to the various parts of the
body but scale of organization and rate of development decrease from
distal to proximal or from anterior to posterior levels. That is, individuals
developing from different levels are primarily different. The amphibian

GRADIENTS, FIELDS, AND DETERMINATION 291

limb field is not equipotential at any given moment. Any other area in
it than the presumptive limb area must undergo change to become ca-
pable of giving rise to a limb, and hmbs from some parts of it may be
less complete than those from other parts. In short, for every particular
development the system, whether Tiibularia stem or limb field, must be-
come another system; since this is possible, it is equipotential in one sense,
and since it has a pattern, or a new one is produced in it, it is harmonious.
According to this view, the harmonious-equipotential system appears to
be a gradient system in which the cUfferent levels have not become so far
specifically different that they cannot react to altered conditions with an
altered development, that is, a system in which a similar pattern may be
differently localized under different conditions. In these terms it is far
from constituting proof of the autonomy of vital processes; but if organis-
mic pattern consists primarily in definitely locahzed specificities, as
Driesch assumed, it becomes difficult to see how a part can become a
whole without the aid of Driesch's entelechy or some other equally capable
metaphysical agent. As a matter of fact, various evidences that inequi-
potentiahties appear in systems assumed to be equipotential are found in
Driesch's data; but either they were ignored, or in certain cases it was
maintained that they represented results of mere physicochemical con-
ditions without the controlling action of entelechy.

DETERMINATION AND DIFFERENTIATION

In experimental analysis of development the terms "determine," "de-
termination," and "differentiation" are so generally used that some con-
sideration of their usage and the basis for it seems necessary. To deter-
mine experimentally a certain developmental result is to provide the con-
ditions necessary for it. We can determine certain differential modifica-
tions of pattern by exposure of the entire developing organism to toxic
agents (chaps, v-vii). New polarities and symmetries can be determined
in various ways (see. chap. xi). In recent years, however, we have come
to speak of determination of regions or parts in the course of development.
A part not yet visibly differentiated, that is, not yet morphologically
characterized, is regarded as determined when its development continues
unaltered — at least for some time — after change in, or isolation from,
organismic environment. Such a part is said to be capable of self-differ-
entiation. The change in environment may consist in transplantation to
another region of the same individual or to other individuals of the same
or other species; that is, it may be autoplastic, homoplastic, heteroplastic,

292 PATTERNS AND PROBLEMS OF DEVELOPMENT

or xenoplastic, or it may consist in isolation in water or some other
medium. Different degrees of fixity, stability, or irreversibility of deter-
mination are recognized according to the development of the part in dif-
ferent environments. If it shows capacity for self-differentiation in cer-
tain environments, not in others, determination is said to be "labile."
Often however, conclusions concerning fixity or irreversibility of deter-
mination are drawn from a single change of environment. Self-differentia-
tion in the altered environment shows, of course, that determination or
segregation (F. R. Lillie, 1927, 1929) has taken place, but that it is fixed
or irreversible in all environments does not necessarily follow. It has been
pointed out by Harrison (1933) and by Gilchrist (1933) that determina-
tion is or may be relative and may be evident in one environment and not
in another. Transplantation experiments with urodele fin ectoderm pro-
vide an excellent example of the relative character of determination
(Twitty, 1939). Many regions or parts found by experimental alteration
of environment to be more or less determined are not morphologically
distinguishable from other parts at the time of alteration, but their later
development shows that differences of some sort must have been present
at that time.

The development of the concept of "formative substances" has led
many investigators to believe that determination in development results
from the presence in the part concerned of a substance or substances dif-
ferent from those in other parts. In the earlier stages of determination
formative effects of such substances may not yet have become sufficient
to be directly distinguishable. Development of morphological form con-
sists in local differences in growth rate, in cell movement, pressure, ten-
sion, turgor, viscosity, etc., and in the metabolic reactions of the proto-
plasmic system and the character of the substrate ; it is the expression of
an exceedingly complex action system. That the systems concerned in
determination of a hydroid tentacle, a planarian head, or an amphibian
limb differ from other parts of the individual — at least after a certain
stage of development — ^is evident; but the tentacle, the head, and the
limb result from a definite spatial activity pattern, not merely from pres-
ence of a particular substance.

With the progress of experiment it becomes increasingly evident that
the term "formative substance" is a misnomer. We find formative pat-
terns with metabolism as an essential factor. The reaction patterns, rather
than substance or substances, are the formative agents. A particular sub-
stance may determine a certain kind of metabolism, but it is the spatial

GRADIENTS, FIELDS, AND DETERMINATION 293

pattern of molar magnitude in which metabohsm occurs that is determi-
native and formative. Properly speaking, formative substances do not
exist. Even assuming that localization of a particular substance in rela-
tion to the polar pattern of a hydroid determines the level or region where
a tentacle shall develop, tentacle development results from a particular
spatial activity pattern in that region; this is no less true for the am-
phibian limb field. A chemodifferentiation of a field without an activity
pattern cannot determine morphological form. There is, at present, no
actual evidence that the character of metabolism in the region where a
tentacle or a limb will develop is different in earlier stages from that in
other regions. Even in later stages the same structures — skin, muscle,
connective tissue, bone — develop in the limb, in the tail, and in various
other parts of the amphibian body. The difference between these parts
is in the spatial developmental pattern rather than in substance. As re-
gards anteroposterior and dorsi ventral axes, the amphibian hmb field
represents a certain relation to general body pattern, different from that
of other parts. In this pattern a new gradient system arises and becomes
the longitudinal axis of the limb. The limb as a pattern is specific and
undergoes a definite orderly series of changes, with the limb as the re-
sult. Different substances are formed in different regions of the pattern,
as in other gradient patterns; but again it is the pattern in which the sub-
stances are formed, not the substances, which constitutes the limb.

Determination is commonly supposed to become increasingly stable
in the course of development, and this is usually considered to indicate
increase in specificity or chemodifferentiation, but determination is some-
times apparently merely a matter of gradient-level. A piece of Corymorpha
stem from a high gradient-level transplanted to a low level may dominate
its environment and develop into an apical region, as it would have done
if isolated, and may even induce other parts in the host; a similar piece
from a low gradient-level, capable of the same development as the other,
when transplanted to a high level may be incorporated and develop
merely as part of the stem. The one might be regarded as stably deter-
mined, the other as labile. The relative character of lability and stability
of determination is shown in many lines of experiment. Various data from
isolation and transplantation of parts of the sea-urchin embryo indicate
stable determination of the apical region as ectoderm, but in the extreme
forms of exogastrulae (Fig. gi,H,I) and with transplantation of micro-
meres to the apical pole (pp. 443-44) it is found that the apical region
can develop as entoderm.

294 PATTERNS AND PROBLEMS OF DEVELOPMENT

Transplantation experiments with amphibian material have shown in
many cases more or less labiUty as regards region of the body formed
but complete or relatively high stability as regards species or group char-
acteristics of organs formed (pp. 457,499). On the other hand, certain
species and group characteristics of larval form in echinoids and asteroids
can be altered to an extreme degree and in different directions by differ-
ential inhibition, conditioning, and recovery, that is, by altering the gra-
dient pattern (chap. vi).

Doubtless many cases of determination do represent increase in specifi-
city over the undetermined condition and may be regarded as the begin-
ings of differentiation, an "invisible differentiation" (Gilchrist, 1937a, b).
The view that determination is not a gradual development of specificity
but a relatively abrupt restriction of potency of a part, and that it is al-
ways dichotomous, has been advanced by F. R. Lillie (1927, 1929). So far
as determination represents attainment of a certain degree of specificity in
a certain region of a gradient pattern, its gradual origin and development
seem equally possible, and Lillie 's critical period may represent the thresh-
old of attainment of a certain degree of specificity permitting self-differen-
tiation. Lillie regards determination as an independent variable in devel-
opment; but if it is a resultant of gradient pattern, it is by no means in-
dependent.

The concept of differentiation is perhaps the most indefinite and most
loosely applied of any concerned with development. In a morphological
sense a part of a cell, a cell, or a cell group is commonly regarded as dif-
ferentiated when it differs visibly in structure from an earlier "embry-
onic," supposedly undifferentiated condition; sometimes a change of shape
has been regarded as differentiation. On the other hand, certain cells,
although visibly different in appearance from embryonic cells, are often
assumed to be indifferent or undifferentiated because they give rise in
reconstitution to parts other than those which they formed in the original
individual. Moreover, as noted above, chemodifferentiation is often in-
ferred in parts capable of self-differentiation, though other evidence is
lacking. The more advanced stages of histological and organ differentia-
tion are directly and clearly distinguishable in the higher animals, and
the morphological differentiation is generally paralleled by specificity in
chemical constitution and in character and products of metaboHsm; but
exactly when or how a particular differentiation begins, we do not know.
Conclusions as to presence or absence of differentiation are often purely
matters of opinion determined by more general opinions concerning de-

GRADIENTS, FIELDS, AND DETERMINATION 295

velopment. For example, one who believes that, once differentiation is
initiated, regression is impossible is likely to ignore appearance, structure,
and function of certain cells and to regard them as undifferentiated be-
cause they differentiate into new organs in reconstitution. The so-called
''formative cells" in certain organisms are examples. Others, who believe
that dedifferentiation is possible, are likely to regard such cells as more
or less differentiated and as undergoing dedifferentiation in the activa-
tion and new development of reconstitution.

It is an interesting question whether differentiation is entirely cytoplas-
mic or may involve the nucleus. Certainly, the nucleus takes on many
different forms and appearances in different cells. It becomes polymor-
phic, sometimes highly branched, in other cells much condensed, and its
staining properties may differ greatly in the same cell at different times.
The nucleus of the ovarian oocyte, for example, usually differs greatly
in appearance and staining from that of maturation and later stages.
Nuclei of most spermatozoa differ greatly from other nuclei of the same
species. Certain blastomeres of Ascaris undergo diminution of chromatin
in early cleavage stages; and following this change, number of chromo-
somes is greater, size less, and appearance very different. Moreover, local-
ization of diminution in certain cells in relation to centrifuging and dis-
permy indicates that regional cytoplasmic differences determine which
nuclei undergo this change and which do not (Boveri, 1910&; Hogue, 1910).
Whether these and many other nuclear changes appearing in the course
of development of gametes and other cells are to be regarded as differenti-
ations is at present largely a matter of opinion. A cytoplasmic environ-
ment seems to be necessary for long-continued nuclear life, and the
nucleus is certainly not wholly insensitive to change in this environment ;
that nuclear differentiation may be induced by the cytoplasm in some
cells seems probable, and perhaps nuclear self -differentiation is possible.

The question of the basis of differentiation has interested biologists
since the study of development began. The Roux-Weismann theory of
qualitative nuclear division as the basis of differentiation is now discarded.
In the ordinary mitosis each daughter cell is supposed to possess the
same genie constitution as its parent. As Morgan once put it (1919,
p. 241), "each cell inherits the whole germ plasm." But cells and cell
groups become increasingly different from each other in the course of de-
velopment. How is this possible?

Undoubtedly, there is interaction between nucleus and cytoplasm, and
supposedly different genes become activated or in some way come to

296 PATTERNS AND PROBLEMS OF DEVELOPMENT

play a part in determining the differences in different cells; but unless
the genie system is similar to Dreisch's entelechy, differences in gene effect
must be based on conditions in the cytoplasm. We can scarcely conceive
that substances produced by the genes can arrange themselves in an
orderly definite pattern resulting in the regional differentiation characteris-
tic of the individual and species, and there is nothing in nuclear pattern
that suggests the spatial pattern of differentiation. An orienting and or-
dering factor of some sort giving rise to a cytoplasmic pattern independ-
ently of the nucleus appears necessary for development. This pattern
must be the basis of physiological polarity, symmetry, asymmetry, and
of morphological form and differentiation in general. In preceding chap-
ters the presence of physiological gradients involving differences in meta-
bohc rate has been shown to be a feature of developmental pattern and
to be definitely related to the course of differentiation, but whether such
gradients are the primary factors of pattern or results of a still more funda-
mental pattern may still be questioned.

An "intimate structure," an orientation of molecules or of colloidal par-
ticles crystalline in structure, or a space lattice have been assumed by
various authors to be the basis of developmental pattern. As regards such
hypothetical structures, it may be noted, first, that there is no evidence
of their existence as general characteristics of pattern. Orientation of mol-
ecules undoubtedly occurs in relation to interfaces, phase boundaries, etc.,
and many highly differentiated structures give evidence of orientation of
molecules or particles; but there is no evidence that the polarity originat-
ing in a hydroid or planarian piece undergoing reconstitution results from
orientation of molecules in all the cells involved or from a space lattice
extending through the whole. It is difficult to believe that such a struc-
ture could persist even in eggs and embryos through all the changes in
position and form of cortical as well as other parts of the cytoplasm.
Second, if such structure exists, its orientation presumably originated in
relation to something. The free pole of the ovarian oocyte becomes the
apical or animal pole of the egg and embryo in many forms. If this polar-
ity consists in a molecular structure, how was that structure oriented in
that particular direction? That a metabolic differential is present between
free and attached poles of the oocyte is sufficiently evident in many cases
and appears beyond question in others. If such a differential is present,
it is highly improbable that a polar molecular orientation could occur in-
dependently of it; but, if it is the orienting factor, it, rather than the mo-
lecular orientation, is the primary polarity and the postulated molecular

GRADIENTS, FIELDS, AND DETERMINATION 297

structure appears unnecessary and, if present, is a result, not a cause, of
physiological polarity. Third, the question how such a molecular structure
can bring about organization and morphological differentiation remains.
It is suggested by Harrison (1937) that the polar, symmetrical, or asym-
metrical structure of protein molecules brings about the localization of
different substances at opposite poles of the egg or in relation to sym-
metry or asymmetry. Developmental pattern, according to these views,
is primarily structural and static. Recent X-ray diffraction photographs
of various embryonic tissues of amphibians and the chick have failed to
show any evidence of molecular orientation that might constitute a basis
for spatial developmental pattern. The authors, however, point out that
these negative results do not prove the absence of such orientation (Har-
rison, Astbury, and Rudall, 1940, "An attempt at X-ray analysis of em-
bryonic processes," Jour. Exp. Zool., 85).

Turning to the gradient concept, the following points are of interest in
relation to the problem of differentiation. The physiological gradients are
real, not hypothetical; they are activity gradients involving protoplasmic
dynamics as well as substrate; when they are altered, the course of differ-
entiation is altered (chaps, ii, v-vii). When they are obhterated, axiate
development and differentiation do not occur, even after the inhibiting
factors have been removed. New patterns can be initiated by external
differentials which affect metabolic rate (see chap. xi). A relation between
the gradient pattern and differentiation is evident, but there is still the
question whether this pattern provides an adequate basis for differentia-
tion. The differences distinguishable at different gradient-levels in early
stages appear to be primarily quantitative ; but what is quantitative and
what qualitative in an activity gradient in a living protoplasm is perhaps
a somewhat academic question. Assuming, however, that a gradient may
be primarily quantitative, at least as regards its dynamic characteristics,
can specific or quahtative differences originate at different levels of it?
Uptake of oxygen, intake and transformation of nutritive material already
present, breakdown of certain molecules, synthesis of others, and dis-
charge of CO. and other metaboUtes are factors of the metabolism of living
protoplasms. In a region of high rate of metabolism, transformation of
nutrition may occur as rapidly as it becomes available, in part perhaps
by complete oxidation, in part by partial breakdown, recombination, and
synthesis of new molecules. In a region of lower rate concentrations of
nutritive material and of oxygen available in relation to rate of transfor-
mation are undoubtedly different, and metabolism there may result in

298 PATTERNS AND PROBLEMS OF DEVELOPMENT

formation of different products. Differences in concentration of electro-
lytes and t^heir ions must also occur at different gradient-levels. Electric-
potential differences are apparently characteristic of gradients. Different
rates of metabolism involve differences in enzyme activity. Present knowl-
edge of molecular constitution of proteins and the part which they play
in enzymes suggests that many of them may be extremely sensitive to
differences such as these and that positions and relations of groups in a
molecule may differ at different levels with resulting difference in reac-
tions. The chain of reactions concerned in oxidations is probably not
the same with different concentrations of reacting substances. Certain
substances tend to concentrate in regions of greater or less surface energy.
These and doubtless many other factors may be concerned in originating
differentiations at different gradient-levels. Relation and interaction be-
tween levels appears also to be an important factor in differentiation.
There is no theoretical difficulty as regards origin of specific or qualita-
tive differentiations at different levels of a primarily quantitative gra-
dient. In fact, it is difficult to believe that such a gradient can remain
without some differentiations for any considerable time. The progressive
increase in specificity of particular parts so generally characteristic of
development also suggests that the primary pattern may be without re-
gional specificity. If a gradient extends over more than a single cell, cells
along its course represent different levels, and these differences provide a
basis for difference in gene action, certainly an essential factor in differ-
entiation. The character of metabolism in a gradient within a single cell
is undoubtedly also determined by interaction between nucleus and cyto-
plasm. According to this conception, determination and differentiation
of parts are earlier and later stages of a continuous series of changes: the
primary pattern which initiates these changes and determines their or-
derly relation along a physiological axis is a gradient in which differences
in rate of metabolism constitute the effective factor.

The assumption that axiate developmental pattern consists primarily
of a static structure or of localized specific substances seems to involve
confusion of the dynamic and material aspects of living protoplasms and
of development. Developmental pattern appears primarily as an activity
pattern with localization of specific substances and morphological struc-
ture as a result. Investigation of organization and development has been
largely in the hands of those with morphological, I'ather than physiologi-
cal, training; and viewpoint and theories of development have usually
been based on embryonic development alone. This situation is perhaps

GRADIENTS, FIELDS, AND DETERMINATION 299

in part responsible for the character of much of developmental theory.
But the primary developmental pattern is no longer present in most ani-
mal eggs at the beginning of embryonic development; more or less struc-
tural organization is usually already present. Many buds and reconstitu-
tions of individuals from isolated pieces or cell aggregates involve origin
of a new developmental pattern, which is accidental as regards both time
and place, so far as the original individual is concerned. That these forms
of development are not merely the further expressions of organizations
already present in consequence of the particular past history of a single
cell, as embryonic development is, appears evident. How completely in-
dependent of pre-existing organization they may be will appear more
clearly in later chapters. It is to these forms of development, rather than
to the egg and embryo, that we must look for the beginnings of develop-
mental pattern and the physiological bases of determination and differ-
entiation of parts.

DEDIFEERENTIATION

Diflferences of opinion concerning the possibility of dedifferentiation
are, of course, associated with the question whether certain cells are or
are not differentiated. If cells which take part in the reconstitutional de-
velopment of other organs than those of which they were originally parts
are differentiated, they presumably undergo more or less dedifferentia-
tion. Those who regard them as differentiated usually conclude that they
dedifferentiate and redifferentiate in reaction to the altered environment.
Others apparently conclude that, because they reconstitute other organs,
they must be undifferentiated. The question of reversibihty or regressi-
bility of development in general is, of course, essentially the same ques-
tion and open to this difference of opinion.

Weismannian theory necessarily assumed that reversal or regression of
differentiation is impossible, that a cell which has once begun to differen-
tiate can never return. Only the germ plasm does not differentiate. As a
matter of fact, however, the egg and the spermatozoon of most animals
appear to be very highly differentiated cells, perhaps the most highly
differentiated cells of the individual. They appear to have approached
or attained the limit of possible differentiation in two different directions.
The yolk is both a chemical and a structural differentiation; the motor
apparatus of spermatozoa can scarcely be regarded as undifferentiated
cytoplasm; and the sperm nucleus is certainly in very different condition
from an embryonic nucleus. When fully differentiated as gametes, both

300 PATTERNS AND PROBLEMS OF DEVELOPMENT

egg and sperm are near death and, except for parthenogenetic eggs, in-
evitably die unless fertilization occurs. Naturally parthenogenetic eggs
show, in general, less extreme morphological differentiation than those
requiring fertilization (Child, 1915^, chap. xiii). The sperm cytoplasm
may not take part in the changes following fertilization; but when em-
bryonic development begins, the egg begins to lose its egg characteristics,
and sooner or later the resulting cells attain what is commonly called
"embryonic condition." The pattern of organization in the egg, however,
becomes the basis of embryonic pattern. The earher stages of embryonic
development appear to involve a considerable dedifferentiation from the
egg condition with progress of a new differentiation within the general egg
pattern. If these changes are not dedifferentiation, there is probably no
dedifferentiation in any cell.

That developmental determination can often be experimentally altered
has been abundantly demonstrated. If such labile determination repre-
sents an early stage of differentiation, its alteration must be a dediffer-
entiation. Under certain conditions synthesis and accumulation, and
under other conditions decomposition and loss of a substance or sub-
stances, take place in cells. These changes appear to be differentiation
and dedifferentiation.

The assumption that certain cells are undifferentiated because they
give rise to new organs in agamic and reconstitutional development is
open to criticism on two grounds. First, it ignores the changes in struc-
ture and behavior usually evident in these cells; second, it involves the
assumption that these cells have been insensitive to their physiological
environment during the preceding development but suddenly become sen-
sitive to conditions resulting in budding, fission, or reconstitution. The
epidermal cells of the begonia leaf are very different from embryonic cells
in structure and behavior. They have formed cellulose membranes and a
large vacuole and under the usual conditions would never divide again ;
but under the conditions initiating bud formation their cytoplasm changes
in structure and staining properties, they begin to divide and grow, cel-
lulose membranes disappear, and a gradient system with vegetative tip
in its high region develops and gives rise to all the structures of the plant
axis (pp. 17^19). Fission and reconstitution in ciliate protozoa apparently
involve extensive ectoplasmic dedifferentiation and redifferentiation. Old
cilia and cirri "melt down" into the ectoplasm, and new ones develop
from other regions." Either the ectoplasmic structures of these animals

" E.g., Dembowska, 1925, 1926; Lund, 191 7. Many other papers give similar data.

GRADIENTS, FIELDS, AND DETERMINATION 301

do not represent differentiation, or dedifferentiation and redifferentiation
occur. Huxley and De Beer (1923) have described what they regard as
dedifferentiation in the hydroid Obelia under inhibiting conditions. The
cells of stem pieces of Corymorpha seem to become less differentiated under
conditions that obliterate polarity and lead to development of new axes
from lateral regions. Cell changes in sea-urchin larvae, regarded as de-
differentiation but resulting in degeneration and death rather than re-
differentiation, have also been described by Huxley (1922). The paren-
chyma cells of planarians and nemerteans are very different in appearance
from embryonic cells of the species, but in reaction to section they undergo
change in structure and behavior and develop into various organs. These
cells, or some of them, are assumed by some to be undifferentiated "forma-
tive cells," apparently because of the part they play in regeneration, while
others maintain that dedifferentiation takes place.''' Regeneration of the
annelid central nervous system from ectoderm involves active prolifera-
tion and loss of epithelial character. In the intact animal these cells se-
crete cuticle and have a certain structure and form different from that of
nerve cells. According to Nuzum and Rand (1924), cells of the pharyngeal
epithelium can also give rise to nervous tissue. Pharyngeal and nerve
cells certainly appear to be differentiated in different directions. Cells usu-
ally regarded as of mesodermal origin, the neoblasts, play a considerable
part in regeneration of other annelid organs and are regarded by some
authors as embryonic or undifferentiated cells.'^ Faulkner (1932) main-
tains that these cells do not come from the coelomic wall, as described
by others, but from outside it, that is, from the blastocoel, and that they
give rise to germ cells as well as regeneration cells. Sayles (1927) and
Weitzmann (1927) regard them as mesodermal cells that have resumed
active proliferation. They certainly undergo change in appearance and
behavior after activation following section and develop into other organs
than the coelomic wall. R. G. Stone also regards them as mesodermal in
origin and finds that X-rays inhibit their activation but also inhibit acti-
vation of other cells in Tubifex and consequently inhibit regeneration;
but this is far from proving that neoblasts are specifically "formative

'3 Formative cells: Curtis and Schulze, 1924; Curtis and Hickman, 1926; Curtis, 1928;
Collings, 1932; Coe, 1934a, b. Dedifferentiation: Nusbaum, 1912; Kenk, 1922; Steinmann,
1926; Prielgauskiene, 1933. See also Goetsch, 1929, 1931; Bandier, 1936, and the general
papers; Schultz, 1908; Stolte, 1936. Also Curtis, 1940, "The histologic basis of regeneration
and reassociation in lower invertebrates," Atner. Nat., 74; Hyman, 1940, "Aspects of regenera-
tion in annelids," Amer. Nat., 74.

'•t See, e.g., Hammerling, 1924a, b; Probst, 1931, 1932.

302 PATTERNS AND PROBLEMS OF DEVELOPMENT

cells. '"5 According to Probst (1932), the annelid Owenia regenerates with-
out neoblasts. In short, neoblasts are apparently not necessary for an-
nelid regeneration, and most of the evidence indicates that when they
are present they are activated and dedifferentiated cells from the coelomic

wall.

A similar difference of opinion exists concerning reconstitutional de-
velopment of ascidians, either from winter buds or cell masses remaining
after degeneration of zooids or directly from pieces of zooids or stolons.
However, even though many cells degenerate and die when reduction of
zooids or isolated pieces occurs, the chief reason for regarding the cells
that remain and give rise to new individuals as undifferentiated seems to
be the fact that they do remain and develop. In these cells form, appear-
ance, activity, and relation to other cells undergo change in connection
with the new development, apparently a regression to a less differentiated
condition and a new progressive differentiation.'^

Butler and, later, Thornton have presented very definite evidence of
dedifferentiation in regeneration of the amphibian limb. Thornton main-
tains that the regenerating tissue, the blastema, originated by the dediffer-
entiation of cells of the limb stump.'' That dedifferentiation is involved
in lens regeneration from the dorsal margin of the iris in adult amphibia
appears beyond question. The iris cells lose their pigment, resume divi-
sion and growth, and differentiate into a lens (p. 396). Retinal cells may
also give rise to cells resembling lens cells. Changes suggesting dedifferen-
tiation have been observed in various explanted tissues by many investi-
gators, but some maintain that it never occurs. In cultivation of embry-
onic rat tissue on the chorioallantois of the chick there seems to be con-
siderable dedifferentiation (Nicholas and Rudnick, 1933). Many other
cases of apparent dedifferentiation might be cited.

General discussions of the subject show the same difference of opinion
as special papers. Schultz (1908) and Nusbaum (191 2) maintained that

'5 R. G. Stone, 1932, "The effects of X-rays on regeneration in Tiihifcx lubife.w" Jour.
Morphol., 53; 1933, "The effects of X-rays on anterior regeneration in Tubifex tubifex," ibid.,
54. See also L. H. Hyman, 1940, ".\spects of regeneration in annelids," Amer. .\at., 74.

'^Schultz, 1907, 1908; Huxley, 1921a, 1926; Spek, 1927.

•7 E. G. Butler, 1933, "The effects of X-radiation on the regeneration of the fore limb of
Amblystoma larvae," Jour. Exp. Zool., 65; 1935, "Studies on limb regeneration in X-rayed
Amblystoma larvae," Anat. Rec, 62. See also C. S. Thornton, 1938a, "The histogenesis of
muscle in the regenerating fore limb of larval Amblystoma punctatitm," Jour. Morphol., 62;
19386, "The histogenesis of the regenerating fore limb of larval Amblystoma after exarticula-
tion of the humerus," ibid.

GRADIENTS, FIELDS, AND DETERMINATION 303

dedifferentiation occurs very often and have given many examples.
Schaxel (191 5, 192 1) and Hammerling are certain that it never occurs.
Weiss (1939) makes a distinction between dedifferentiation, which he be-
Ueves does not take place, and "modulation," which is believed to occur.
However, it is not evident from his discussion that modulation is any-
thing but a lesser degree of dedifferentiation. According to Schotte, de-
differentiation occurs in many cases. '^

Doubtless; the differences of opinion will continue until we know more
about what constitutes differentiation in any particular case. At present,
however, there is a large body of evidence indicating that more or less
dedifferentiation does occur in many organisms and many tissues, even
to some extent among the higher vertebrates. Probably cells previously
subjected to a developmental environment producing less stable differen-
tiation are more capable of reaction to altered environment than others.
Dedifferentiation does not necessarily involve complete loss of determi-
nation of cells but is alteration of structure and behavior in the direction
toward the more general characteristics of cells of earher stages of devel-
opment. Such changes certainly do occur in many cells. The assumption
that certain cells in adult organisms are undifferentiated, irrespective of
their structure and function, because they are able to become, or give
rise to, cells with other structure and function requires not merely asser-
tion but rigid proof.

'8 O. Schotte, 1939, "The origin and morphogenetic potencies of regenerates. First sym-
posium on development and growth," Growth, Suppl. 1939.

CHAPTER IX

PHYSIOLOGICAL INTEGRATION: DOMINANCE AND
PHYSIOLOGICAL ISOLATION

THAT the organismic individual is a more or less closely inte-
grated whole is evident. Certain questions concerning origin and
nature of the integrating factors, their limitations in some organ-
isms, as indicated by physiological isolation, and their relation to develop-
mental patterns are considered in this chapter. The examples are taken
in large part from agamic and reconstitutional development because phys-
iological isolation is often conspicuous in relation to these forms of de-
velopment and more accessible to experiment than is usually the case in
embryonic development.

It was pointed out in chapter i that physiological dominance or control
may be effected either by transmission of energy changes or by mass
transport by diffusion or otherwise of chemical substance. In earlier stages
of reconstitutional and various other forms of development the dominant
region is generally the "high" region of a gradient, and by inhibiting its
activity the range and effectiveness of dominance is decreased and obliter-
ation of the gradient obliterates dominance. It is often possible to estab-
lish a dominant region and a gradient in relation to it by establishing
a localized region of increased activity. Moreover, in reconstitution in
the simpler animals and in many other cases the central nervous system
becomes an important factor in dominance. These and various other facts
brought to attention in the following pages suggest that in the primitive
form of developmental dominance the dominant region acts essentially
Hke a region of excitation and that dominance of such a region is effected
by transmission rather than transport. Such transmission is perhaps pri-
marily electrical and a result of potential difference between the dominant
region and other parts, or it may be transmission of a protoplasmic ex-
citation in which electrical factors are undoubtedly concerned. In the sim-
pler organisms this type of dominance is limited in range of effectiveness,
but the range varies with activity of the dominant region and can be
altered experimentally. The evidence indicates that there is a decrement
in effectiveness with increasing distance from the dominant region and

304

DOMINANCE AND PHYSIOLOGICAL ISOLATION 305

that at a certain variable distance dominance becomes ineffective as a
factor determining development or maintenance of parts already devel-
oped. In protoplasms lacking specialized conducting paths the effects of
local excitation are apparently transmitted with a decrement; and even
in cases in which some part of the central nervous system appears to
be the dominant region, evidences of limited range of dominance often
appear in the simpler animals; but whether this limitation results from an
actual decrement in nervous transmission in primitive nerve complexes
or from other factors — for example, incomplete differentiation of the nerv-
ous system in regions of rapid growth in length — is uncertain. Protoplas-
mic excitation and transmission attain their highest development in nerv-
ous tissue. The early development of the nervous system, the localization
of the chief masses of nervous tissue in, or in close association with, the
high regions of gradients, and the close parallelism between relations of
dominance and subordination within the nervous system and gradients
of earlier stages suggest that the nervous system is a relatively direct de-
velopmental expression of the primary factors in organismic integration.

A part of an individual which, for any reason, comes to lie beyond the
range of effective dominance is physiologically isolated, that is, it is no
longer subjected to the factors which were concerned in determining its
development as a part of the individual or its persistence as a particular
part. If the part is not so stably determined or differentiated that it
cannot react to this isolation, it tends to lose more or less completely its
characteristics as a definite part of the individual and may, under certain
conditions, reconstitute a new individual.

This limitation in the effective range of dominance is a factor in limiting
length of the individual or zooid in many organisms, but the limit of domi-
nance and of length varies with conditions. An intensely active dominant
region determines, in general, a greater length than one less active — for
example, in various flatworms. Also, a greater length of individual or
zooid is attained in planarians, Stenostomum, and various other forms with
slow than with rapid growth, probably because differentiation of the
longitudinal nerve cords in the regions of most rapid growth more nearly
keeps pace with a slow, than with a rapid, increase in body length. In
planarians and Stenostomum the length of a single zooid increases with
advance in development of the head. If growth in length of a Stenostomum
chain is rapid, a new fission zone and head region arise at a shorter dis-
tance from a zooid head in early developmental stage than from a fully
developed head.

3o6 PATTERNS AND PROBLEMS OF DEVELOPMENT

In the more highly differentiated animals range of effective dominance
is not limited, except perhaps in earlier developmental stages. With at-
tainment of all-or-none conduction of nervous impulses nervous domi-
nance is effective over an indefinite distance. The limit of individual size
in these forms is determined by factors limiting growth rather than by a
limited range of dominance. Dominance effected by mass transport of
substance is not necessarily limited in range, though it may be so limited
if the substance decreases in concentration by spreading over a greater
area or is progressively used up or altered in the course of transport.

In many of the simpler organisms physiological isolation with resulting
reorganization and reconstitution of the isolated part may result from
any one or more of four factors: (i) increase in length of the polar axis,
so that a part comes to lie beyond the range of effective dominance; (2)
decrease in effective range of dominance in consequence of decrease in
activity of a dominant region; (3) blocking by local inhibition between
dominant and subordinate region; (4) alteration of a subordinate part so
that it becomes more or less insensitive to dominance, by activation or
stimulation from some other source and probably in many cases by pro-
gressing determination and differentiation.

A suf!icient degree of physiological isolation of a part may result in
its development into a new individual, as in cases of budding in coelen-
terates and other forms and of fission in planarians and annelids. Re-
duplication of parts, segments, etc., from a growing region probably also
involves at least partial successive physiological isolations. In general,
agamic reproduction in axiate organisms appears to be a reconstitution re-
sulting from physiological isolation, as reconstitution of pieces results from
physical isolation.

Physiological isolation of a part by its determination or differentiation
leads to a different result, that is, independent or self-differentiation.
Moreover, attainment of a certain degree of specificity by a part may re-
sult in its activation in spite of dominance. The orderly appearance of
localized regions of increased developmental activity, increased rate of
dye reduction, and increased susceptibility in connection with develop-
ment of particular organs, as in the chick embryo (pp. 159-63), suggests
this sort of self-isolation. When the region from which the optic vesicle
develops attains a certain physiological stage, it undergoes an activation
irrespective of any general dominance. Similarly, the otic region, the
appendage region, etc., undergo activation at a certain stage. Such physi-
ological isolation may or may not be so complete that the part can con-

DOMINANCE AND PHYSIOLOGICAL ISOLATION 307

tinue its development unaltered when physically isolated, but it has cer-
tainly become so far isolated that it can develop in a way different from
preceding stages.

Dominance of the primitive type may be effective in establishing a
gradient and later in maintaining it, and so in determining the course of
development of its different levels. Such a region is an inductor in earlier
stages and later may determine persistence of the induced development.
Even a part capable of self-differentiation after a certain developmental
stage may be incapable of continued existence after differentiation with-
out influence of a dominant region; muscular tissue, even in higher verte-
brates, is an example.

In recent years much has been learned concerning the roles of specific
chemical substances as factors of physiological dominance in develop-
ment and of maintenance in adult life. As regards the great significance
and the exceedingly complex interrelations of chemical dominance, there
can be no question. However, most of the investigations in this field con-
cern advanced stages of development or functional relations in mature
individuals and, except for the recent work on plant hormones, are
largely concerned with the higher vertebrates. It has often been pointed
out that the earlier the stage of development the less evidence it affords
of specific interrelations of parts or of chemical dominance. If quantita-
tive gradients are the primary factors in axiate pattern, it follows that
chemical dominance by production, transport, and effect of specific sub-
stance is not the primary form of physiological dominance but is possible
only after different regions have become, to some degree, specifically dif-
ferent. The high region of a gradient may produce more of a certain sub-
stance than other levels, and its transport may alter the concentration
of this substance at other levels and so influence their condition; but this
is a nonspecific dominance directly related to gradient differences. A cell
membrane may, in a sense, dominate the cell interior by its specific perme-
ability to substances; but even in this case the membrane has become
different from the interior in consequence of exposure to an external
medium, and it is, in general, merely selective, not productive.

With origin of specific differences in development of the individual,
chemical dominance becomes possible and with progress of differentiation
evidently plays an increasingly important part in determining and influ-
encing the further course of development, attaining its highest develop-
ment in the hormone interrelations of the higher vertebrates. Even in
these organisms, however, nervous dominance is still the chief integrating

3o8 PATTERNS AND PROBLEMS OF DEVELOPMENT

factor and influences hormone production; but hormones, once produced,
also influence the nervous system. It is perhaps of some interest to note
that the hypophysis, apparently a highly important factor in the hormone
complex, is, like the chief aggregation of nervous tissue, a development
from the higher levels of the chief gradient.

In some cases transmissive and transportative factors may combine in
dominance. A gradient established in the earliest developmental stages
may influence direction of transport of chemical substance, as will appear
in the following section. A transmitted nervous impulse may set free at
the end of its path a particular substance, a "neurohumor," which deter-
mines the final effect. Physiological isolation from chemical dominance
involves essentially the same factors as isolation from the primitive type
of dominance: decrement in concentration or alteration with transport;
decrease in production by the dominant region; blocking of transport;
alteration of the subordinate part, making it insensitive.'

DOMINANCE IN PLANTS

Extended discussion of plant dominance is beyond the present purpose,
but attention is briefly called to a few points because of their interest in
relation to dominance in animals. Experiments on plant dominance be-
gan with the early grower of plants, who learned to prune and trim in
such manner that certain results were obtained; and the botanist, by ex-
tensive and varied experimentation, has thrown much further light on the
problems concerned. The most famihar example of dominance in plants
is that of the vegetative stem tip over lateral buds at stem-levels below it.
In plants which give rise to lateral bud primordia (potentially new axes) ,
the vegetative tip of the primary axis of some forms prevents the out-
growth and development of these buds, unless inhibited in activity or
removed, but its dominance may decrease in the course of the life-cycle;
in other forms it may retard their development and determine their
growth form as lateral branches. The bean seedling is an example of the
first type ; the second type appears in many conifers and numerous other
plants. In the bean seedling removal, inhibition of the tip, or blocking its
effect results in outgrowth of the previously inhibited buds, those of the
uppermost node reacting most rapidly and inhibiting more or less com-
pletely those of lower levels. In the conifer removal of the tip is followed
by the turning-upward of one or more of the uppermost lateral branches
and a change from the bilateral pattern of secondary branching character-

' See Child, ig2ia; 19246, chaps, x, xi; 19296 for earlier discussions of dominance.

DOMINANCE AND PHYSIOLOGICAL ISOLATION 309

istic of many conifers to the radial pattern of the primary axis. It is obvi-
ous that the growth forms of axiate plants must depend, to a great extent,
on the degree or persistence of dominance of the primary stem tip and of
other tips. If dominance is highly effective and persistent, the axis re-
mains unbranched or may bear short lateral branches, usually different
in pattern of secondary branching from the main axis. With less effective
or decreasing dominance the plant may become a highly branched spread-
ing form with several or many equivalent or nearly equivalent axes. Many
trees and other plants show the first form in earlier stages; later, some
degree of physiological isolation of branches evidently occurs, and some
or all of them become more or less equivalent. Growing leaves, as well
as stem tips, inhibit bud development. The root tip is apparently, also,
to some extent a dominant region inhibiting formation of a new root tip
within a certain distance or determining its development as a lateral root.

Various hypotheses concerning the mechanism of this dominance of
stem tips and other active regions have been advanced. It has been held
by some that there is in plants something analogous to nervous control;
others have maintained that dominance is a matter of nutrition, the domi-
nant region taking so large a part of the available supply that other parts
are unable to obtain enough for their development; another view is that
the dominant region produces substance inhibiting development of buds
in other regions. Recently, however, discovery of the substances now
known as "auxins" and the rapid development of analytic investigation
concerning their production, distribution, and functions have thrown
light on some aspects of the problem of dominance in plants.'

A biological method for comparative estimation of auxin amounts in
terms of their effects on cell elongation has been developed with the
coleoptile of the Avena (oat) seedhng, which also played an important
part in the discovery of auxin. The coleoptile is a sheath surrounding the
young shoot. After an early stage its growth is by cell elongation without
division, and the region of maximum elongation is some distance below
the tip. Auxin is produced by the tip, is transported basipetally, and is
concerned in the elongation. When coleoptile tips, stem tips, or other
parts are placed on agar blocks, auxin, if present, diffuses into the agar;
and if the block is placed on one side of a decapitated coleoptile, auxin is

2 This field of investigation is developing so rapidly that any general survey is practically
out of date by the time of publication. The book P/iyk>kormones by Went and Thimann (1937)
is the chief authority for the few points mentioned here. The book includes an extensive
bibliography. See also Boysen-Jensen, 1936; Thimann and Bonner, 1938.

3IO PATTERNS AND PROBLEMS OF DEVELOPMENT

transported basipetally on that side, and cell elongation, with curvature
of the coleoptile toward the opposite side, results. With standardization
of the procedures involved and establishment of a biological unit, this
method permits comparative assays of auxin activity with considerable
accuracy.

Two auxins, A (auxentriolic acid) and B (auxenolonic acid), are distin-
guished; and indole-3-acetic acid (heteroauxin) has been found to act in
the same manner as natural auxins. Many other organic substances,
among them a substance from human urine, have more or less auxin-hke
action, and certain relations between chemical structure and activity
seem to be established.

Auxins are apparently highly versatile substances. They, or inactive
precursors which undergo activation, are produced in active stem tips,
growing leaves, etc., and are transported, chiefly basipetally, in living
tissues of plant axes. By differential distribution or production in relation
to external factors and by inducing cell elongation they are concerned in
tropisms. They induce root formation; in extremely low concentrations
they accelerate growth of roots; but in higher concentrations they in-
hibit root growth. Apparently roots are either more susceptible to auxin
than stems or do not require it for growth.

Auxin is transported rapidly in the living plant by a mechanism funda-
mentally different from diffusion but otherwise still obscure. Direction of
transport in physiological concentrations is chiefly or wholly basipetal,
that is, in a definite relation to the polarity of the axis concerned. This
direction is maintained even against a concentration gradient of auxin.
With concentrations far above the physiological range some acropetal
transport may occur, probably in the transpiration current. With nar-
cotization of coleoptiles by ether vapor, transport becomes essentially
diffusion without polarity; with sufficiently low ether concentrations, this
obliteration of polarity is reversible. At 0° C. transport also approaches
diffusion but is still polar. It seems evident that the axiate pattern of
vital activity is in some way concerned in auxin transport. Streaming of
protoplasm, movement along interfaces, and electric potential gradients
have been suggested as possible factors in directed transport, but none
of these appears entirely adequate. According to Clark (1937), the elec-
tric-potential gradient can be obliterated experimentally without affecting
polar transport, and this can be abolished with persistence of electric
polarity. Association of auxin activity with oxidations is suggested by
Thimann and Bonner (1938). That production and transport are associ-

DOMINANCE AND PHYSIOLOGICAL ISOLATION 311

ated in some way with cell metabolism and so with physiological gradient
pattern seems to be indicated by the evidence.

As regards dominance of a stem tip, it has been found that in Viciafaba
seedlings the stem tip produces the most auxin, the leaves less, and dor-
mant axillary buds almost none ; but these buds become producers when
they develop. Auxin in agar applied to the cut end of the stem after re-
moval of the tip is as effective in inhibiting axillary buds as the tip.^
Further experiment has given similar results with many other plants
and with various auxins and substances showing auxin action. The mech-
anism of inhibition of axillary buds is still obscure. There is some evi-
dence that auxin effect differs according as it moves with, or against, the
polarity of the axis, being inhibitory when applied basally and moving
acropetally and having the opposite effect when applied apically. If it
passes from the stem into axillary bud axes, it moves acropetally in these
axes. Certain lines of experiment indicate that inhibiting action of auxin
may be indirect, through its effect on other factors.^

Physiological isolation of axillary buds can be brought about by inclos-
ing the stem tip in an atmosphere without oxygen or in plaster, the tip
remaining alive for some time under these conditions but auxin produc-
tion being presumably inhibited. A zone of low temperature about the
stem between tip and buds is effective in blocking dominance (Child and
Bellamy, 1919, 1920). On the basis of these experiments it was suggested
that a transmissive factor of some sort might be concerned. However, in
the light of the data on the role of auxin in dominance and those showing
that auxin transport is almost stopped by chilling a zone of stem below
5° C. (Cooper, 1936), blocking of dominance by low temperature appears
to be a blocking of auxin transport; but these experiments also suggest
that cell metabolism is in some way concerned in the transport.

This dominance of stem tips and growing leaves over buds which have
already attained a certain developmental stage is a secondary, not a pri-
mary, type of dominance involving relations in a multiaxiate pattern al-
ready present. In the higher plants new axes, with leaf primordia as their
first developmental expression, are localized in an orderly spatial pat-
tern within the embryonic tissue of the tip itself. Each new bud ap-
pears at a certain distance from the apex of the tip and in a definite
spatial relation to other buds already present. This spatial pattern shown

3 Thimann and Skoog, 1933, 1934; Skoog and Thimann, 1934.

■< See the discussion in Went and Thimann, 1937, chap, xii; also Went, 1936; Le Fanu, 1936;
Snow, 1937.

312 PATTERNS AND PROBLEMS OF DEVELOPMENT

by arrangement of leaf primordia (phyllotaxis) is species-characteristic
under natural conditions — alternate, opposite, whorled, etc. — but often
alterable experimentally.^ This pattern suggests, first, that a dominance
of the extreme apical region of the main stem tip is effective over a very
short distance in preventing new buds from originating; and that beyond
that distance, at most a few millimeters, some degree of physiological
isolation occurs, permitting initiation of lateral bud development and
often continued development of a leaf; and second, that each actively
developing bud or its leaf also dominates a certain area about itself and
so prevents other buds from arising within this area, that is, each center
of activity gives rise to a gradient system. That of the main tip has be-
come axiate; that of the lateral bud is at first more or less radial, or
symmetrical with respect to the axiate pattern of the stem tip, and be-
comes axiate by differential growth, like other buds. Nothing is known
concerning the factors effective in determining and maintaining these
orderly patterns within the stem tip. That they depend on auxin seems
at present rather improbable. Doubtless, electric-potential gradients are
associated with each growing primordium, but whether they are con-
cerned in determining the spatial and chronological order remains to be
determined.

The activity of the root system of higher plants inhibits more or less
completely development of roots elsewhere, although physiological isola-
tion may be brought about in many plants by subjecting other regions to
conditions favorable to root formation. Experiment indicates that this
general dominance of a root system depends largely on transport of water
and salts from the roots. Depletion of these in other parts of the plant
favors root formation, chiefly at the more basal levels of the stem, unless
experimentally or otherwise inhibited there, presumably because deple-
tion is most rapid there. This also is a secondary dominance depending
on a pattern already differentiated. There is, however, some evidence of a
local dominance with limited range at the more apical levels of a root axis.
In the presence of an actively growing root tip a new root primordium
appears only at a certain distance from the tip, and removal or inactivation
of the tip destroys the dominance. Here, as in the stem tip, there appears
to be some degree of physiological isolation with increasing distance along
the root axis from the tip.

Evidences of dominance and physiological isolation also appear in the

5 See, e.g., Mary and R. Snow, 1931, 1934; R- Snow, 1929; and earlier work on phyllo-
taxis cited by these authors.

DOMINANCE AND PHYSIOLOGICAL ISOLATION 313

lower plants. For example, the very definite order of appearance, the rate,
and the direction of growth of lateral branches of the thalli of many
multiaxiate algae, even when the whole thallus is a single cell, as in
Bryopsis, indicate presence of a definite spatial pattern involving domi-
nance of the apical region of the main axis and physiological isolation at
a certain distance from it. Development of new axes from cells that do
not normally give rise to buds has often been induced by removal of the
apical region in various algae and fungi. In the prothallia of liverworts
and ferns dominance of the apical region has also been experimentally
demonstrated. Many of these plants also afford evidence of decreasing
effectiveness of dominance with increase in distance from the dominant
region.

The so-called "gills" of the mushroom constitute a very definite spatial
pattern; but when they are removed, leaving a flat surface, outgrowths
may arise from any part of this surface, their localization being entirely
irregular and apparently determined by local chance differences in activ-
ity. Each outgrowth dominates a certain area about itself, and an out-
growth somewhat in advance retards or completely inhibits development
of other outgrowths within that area (Magnus, 1906) ; but beyond a short
distance it is ineffective. Similarly, when numbers of adventitious buds
develop near together, any one which develops more rapidly may retard
or inhibit further development of others within a certain distance from
itself. Whether dominance in these cases depends on auxin transport
within the gradient system of each outgrowth or bud or on some other
factor is not known.

DOMINANCE, BUDDING, AND AXIAL RELATIONS IN
CERTAIN COELENTERATES

The hydra bud appears at the most proximal levels of the body capable
of reacting rapidly to physiological or physical isolation. Removal of the
apical region proximal to the tentacles accelerates development of buds.
In animals in good condition buds appear only after the body has at-
tained a certain length, and in a stock in uniform environment this length
is fairly constant. Under depressing conditions, in "senescent" and in sex-
ual animals, buds often appear at more distal levels than the usual bud-
ding zone, and more or less persistent "colonies" may result from delayed
separation. After removal of the body distal to the budding zone a bud
may inhibit reconstitution of a distal region and, with gradual change in
position, replace the part removed. Apparently a developing bud inhibits

314

PATTERNS AND PROBLEMS OF DEVELOPMENT

development of a second bud within a certain distance transversely in
the parent body.^

The hydroid Tuhiilaria gives rise to buds and branches to a varying
degree in different species. The young animal, emerging from the gono-
phore, attaches to the substratum by its basal end; this spreads and be-
comes somewhat flattened; and, as the unbranched stem increases in

1

B

D

Fig. io8, A-D. — Tubularia. A, young individual with developing stolon; B, transformation
of stolon tip into hydranth after physiological isolation by increase in distance from dominant
hydranth; C, reconstitution of stem piece with hydranth distal and stolon proximal; D, bi-
polar reconstitution ("axial heteromorphosis").

length, one or more buds develop from the base and grow out as stolons.
The stolon represents a new gradient, with high end at the tip, but differs
from the stem in that it grows in contact with the substrate and remains
a growth gradient without evidence of differentiation as long as it remains
a stolon (Fig. io8, A).'' Both stem and stolon elongate; and when a cer-
tain length, varying with physiological and external conditions, is at-

^ The experimental literature on hydra is extensive. The following papers are in part di-
rectly concerned with the problem of dominance and budding: Hyman, 1928; Weimer, 1928,
1932, 1934; Rulon and Child, 1937a; they also give references to earher literature.

' Child, 1907(7, igigd, ig2id, 1923a; Child and Watanabe, 1935/^ Watanabe, 1935c.

DOMINANCE AND PHYSIOLOGICAL ISOLATION 315

tained, the stolon tip alters its reaction, turns away from the substratum,
and its tip becomes a new hydranth (Fig. 108, B). In relation to this
hydranth a new stem develops, and later one or more new stolons grow
out from its base, their tips becoming hydranths when suflliciently isolated
physiologically from the dominant hydranth. In some species successive
stems develop in this way at almost equal distances in series; in others
the sequence is less regular. With still further increase in length of the
original stem hydranth buds may arise in some species along the stolon
and later from the proximal region of the stem itself. Each of these de-
velops a stem and becomes a branch. Buds may also appear if the apical
hydranth is removed and kept from reconstitution ; but when it is present
and in good condition, new hydranth development begins only at a certain
distance from it. From the distal end of a stem piece several centimeters
long a hydranth develops rapidly, and usually a second smaller hydranth
develops more slowly from its proximal end (Fig. 108, D). Sometimes,
however, particularly in certain species, a stolon appears at the proximal
end (Fig. 108, C), and its tip transforms later into a hydranth. Stolons
develop more frequently from the more proximal levels of the stem (Child,
1907a) . In various hydroids apical regions can be transformed into stolons
by inhibiting conditions (pp. 172-75) : in Tubularia the stolon is evidently
an axis somewhat inhibited by the dominant hydranth or by other con-
ditions. It develops as a bud from the lower levels of the stem gradient
in consequence of partial physiological isolation, and with increased isola-
tion its tip becomes a hydranth. When the hydranth, the region of chief
dominance, is removed, the degree of isolation at the proximal end of a
piece may be sufhcient to permit development of a hydranth there at
once, or a short stolon may develop first (Fig. 108, C, D). So-called "axial
heteromorphosis" in Tubularia differs from the natural agamic reproduc-
tion only in that removal of the dominant hydranth permits a sufficient
degree of isolation for hydranth development at a shorter distance from
the distal end than in the intact animal. In pieces the rate of hydranth
development decreases from distal to proximal levels, except in very long
pieces, in which the proximal end is already more or less physiologically
isolated, and a rapidly developing oral hydranth retards still further the
development of the aboral hydranth. This is shown by the more rapid
development of the aboral hydranth when development of the oral hy-
dranth is retarded or inhibited and also by the more rapid development of
a hydranth at the distal end of a proximal piece than at the proximal end
of the piece immediately distal to it. If oral and aboral hydranth develop-

3i6 PATTERNS AND PROBLEMS OF DEVELOPMENT

ment begin about the same time, neither inhibits the other. The great
mass of experimental data on reconstitution in Tubularia offers no diffi-
culty to interpretation in terms of gradients and a dominance associated
with them, with physiological and physical isolation possible, both experi-
mentally and under natural conditions.

Corymorpha, also a tubularian hydroid, never gives rise to new hy-
dranths by budding. The stolons are holdfasts, threadhke outgrowths,
developing in large numbers from longitudinal series of buds in the basal
region. Those nearest the basal end develop first, are dominant, and in-
hibit others until they have become so long that their dominance is not
effective. On removal of the buds nearest the basal end by section, those
adjoining the section grow out very rapidly and inhibit others. Each sto-
lon is a gradient with high end at the tip (Child, 19286; Child and Wata-

nabe, 1935^) •

In the reconstitution of stem pieces Corymorpha resembles Tubularia,
except that the proximal fifth, more or less, of the stem, the only part
secreting perisarc in mature animals, reconstitutes a basal region from its
proximal end more frequently than Tubularia; this is according to expecta-
tion, since dominance in Corymorpha is evidently more effective than in
Tubularia, as the absence of buds and branches shows. Pieces from the
naked four-fifths of the stem reconstitute essentially like Tubularia pieces
with hydranths at both ends, the proximal developing more slowly than
the distal, except when pieces are so short that gradient difference is prac-
tically absent (Child, 19266).

Watanabe (1935c) has made an experimental analysis of the dominance
of the original hydranth and the development of dominance by a develop-
ing hydranth. With increasing delay in removal of the original hydranth
its effectiveness in inhibiting hydranth development at the proximal end
of the piece increases, as is shown by the increase in frequency of unipolar
forms with the original hydranth at the distal end and proximal hydranth
reconstitution completely inhibited. The experimental procedure and the
result are shown in Figure 109. Frequency of unipolar forms increases
from practically zero with section at both ends at the same time to 86
per cent with 72 hours delay of proximal section; that is, if the original
hydranth remains 72 hours after the proximal section, it so completely
inhibits hydranth development at the proximal end that, even after its
removal, only 14 per cent of the pieces show hydranth development there,
and most of the other pieces develop a basal end. If the original hydranth

DOMINANCE AND PHYSIOLOGICAL ISOLATION

317

does not remain in good condition, it is much less effective in inhibiting
proximal hydranth development.

With removal of the original hydranth and delay of proximal section,
unipolar frequency with hydranth at distal end increases from practically
zero to 92 per cent with 24 hours delay and to 100 per cent with 48 hours
delay of proximal section (Fig. no); that is, if proximal section is made
48 hours later than distal, hydranth development at the proximal end is
completely inhibited by the hydranth developing at the distal end. The

100

Fig. 109, A, B. — Dominance of original hydranth oi Corymorpha. A, experimental pro-
cedure; proximal section at o hr., distal section at X (o, 24, 48, 72) hr.; B, graph of results;
ordinates, percentages of unipolar frequency; abscissae, hours; fifty pieces in each lot (from
Watanabe, 1935c).

developing distal hydranth is apparently more effective as a dominant
region than the original, fully developed hydranth.

When distal and proximal section are made at the same time and the
hydranth primordium is removed from the distal end after various periods
of delay, unipolar frequency increases from zero with no delay to 70 per
cent with 48 hours delay (Fig. in). In this case, however, the hydranth
is at the proximal end of the piece, hydranth development at the distal
end is inhibited, and a base may develop there later; a complete reversal
of polarity has resulted from localization of the dominant region at the
proximal, instead of at the distal, end. In these pieces the dye-reduction
gradient is also reversed. In general the dye-reduction gradient or gra-
dients show close parallelism to the dominance, as indicated by hydranth

100-

FiG. no, A, B. — Dominance of developing hydranth ol Corymorpha at distal end of piece.
A, experimental procedure; distal section at o hr., proximal section at A" (o, 6, 12, 18, 24, 48)
hr.; B, graph of results; ordinates, percentages of unipolar frequency; abscissae, hours; fifty
pieces in each lot (from Watanabe, 1935c).

100-

A

Fig. Ill, A, B. — Development of dominance by hydranth of Corymorpha at proximal end
of piece. A, experimental procedure; distal and proximal section at o hr., hydranth pri-
mordium at distal end removed at X (o, 8, 16, 24, 32, 40, 48) hr.; B, graph of results; ordinates,
percentages of unipolar forms; abscissae, hours; fifty pieces in each lot (from Watanabe,
1935c).

DOMINANCE AND PHYSIOLOGICAL ISOLATION 319

development and inhibition. In the case of reversal of dominance and
polarity (Fig. iii) the reduction gradient was found to be completely re-
versed in 86 per cent of pieces from which the distal hydranth primordium
was removed 48 hours after isolation of the pieces, the dye-reduction test
being made several hours later. In general the experiments along this line
indicate that a little more time is required for attainment of complete
dominance and reversal of the gradient by a hydranth developing at the
proximal end than for attainment of complete dominance by a hydranth
at the distal end.

A very similar increase in unipolar frequency appears in pieces of the
stalk of the sessile scyphozoan Haliclystus with delay of proximal or
distal section for different periods. With simultaneous distal and proximal
section the stalk pieces show a high bipolar frequency (Fig. 113, J-M,
p. 334), but this can be reduced to zero with a certain period of delay of
proximal section and reversal of polarity in the whole piece can be brought
about with increasing frequency in the same way as in Corymorpha pieces
(Fig. III). Moreover, gradients in the indophenol blue reaction a few
hours after section show almost complete correspondence to the bipolarity
or unipolarity developing later (Watanabe, 1937).

The branching hydroids show spatial and chronological orders very
similar to those of multiaxiate plants. Many tubularian (gymnoblast)
hydroids are monopodial, that is, the primary axis persists and the first
hydranth retains more or less dominance, later buds giving rise to lateral
branches in radial or spiral order about the main axis, or opposite or
alternate in a single plane. At a sufficient distance from the dominant
region or in its absence lateral branches may transform into main axes.
In many species of campanularian (calyptoblast) hydroids the axial pat-
tern is sympodial, that is, each new bud on each axis becomes temporarily
dominant, is later subordinated to the next bud, and becomes a lateral
branch of similar sympodial character. In some other hydroid species the
axes are more or less equivalent, and the group resulting from budding
consists of similar zooids (e.g., Clava). The spatial relations between dom-
inant hydranths and new hydranth buds indicate a more or less definite
range of effectiveness of dominance but varying with physiological and
external conditions. In various hydromedusae other medusae develop
from buds on the manubrium. Dominance and physiological isolation are
evidently concerned in determining spatial and chronological order identi-
cal with that of phyllotaxis in certain plants. According to Wood- Jones
(19 1 2), the apical zooid is dominant and radial in form in certain branch-

320 PATTERNS AND PROBLEMS OF DEVELOPMENT

ing corals; lateral zooids bud from about its base and are dorsiventral in
the direction of the branch axis, or, perhaps more correctly, in the direc-
tion of the radius of the apical zooid in which the bud appears. At a
certain distance from the dominant apical zooid a lateral zooid may trans-
form into a radial apical zooid and become the apex of a new branch; or
if the apical zooid is removed, one or more of the uppermost lateral zooids
may become an apical zooid. The resemblance to the dominance of the
stem tip over lateral buds in plants is evident, but the mechanism of domi-
nance is undoubtedly different. In certain other coral species the zooids
are equivalent; and in some an apical dominant zooid may be present in
favorable, and all zooids equivalent in unfavorable, environment.

It is suggested that the nerve net is the chief factor in dominance in
hydroids and other coelenterates; but whether the limited range of domi-
nance in these forms depends on transmission with a decrement in the net
or on the relation between rate of increase in length of the body and of
differentiation of the net, the experimental data do not show. If the nerve
net shares in the gradient difference along the stem of such a form as
Tiibularia, for example, there may be a transmission decrement from
higher to lower levels. The time required for attainment of dominance
by a new hydranth in reconstitution suggests that reorganization of the
nerve net may be concerned.

A different conclusion has been reached by Barth (1938a). He finds
that an oil drop introduced into the coenosarcal cavity of the Tuhularia
stem blocks dominance and concludes that the factor determining domi-
nance is transported in the circulation in the gastrovascular cavity of the
stem. If this is the case, the factor must be inactivated or disappear in
the course of transport, for the range of dominance is limited, and under
given conditions the limit is at a rather definite distance from the domi-
nant hydranth. However, since Tuhularia is highly susceptible to decrease
in oxygen (Barth, 1938ft) and the perisarc is not highly permeable to
oxygen, it seems possible that the oil drop may decrease oxygen supply
sufficiently in the coenosarc in contact with it to block transmission
through the nerve net. Relations of dominance and physiological isola-
tion are essentially similar in Corymorpha and Tuhularia, but in Cory-
morpha a large number of small longitudinal canals lying just beneath
the ectoderm represents the gastrovascular cavity of the stem. Regular
or uniform circulation in these appears impossible because the naked stem
contracts, extends, and bends almost continuously, even in isolated
pieces; but localization, length, and time of appearance of hydranth pri-

DOMINANCE AND PHYSIOLOGICAL ISOLATION 321

mordia show a high degree of regularity. In the Tubularia stem and prob-
ably in Corymorpha circulation is in both directions. Why does not the
circulating factor inhibit the dominant hydranth? Also, how is it possible
for hydranths to develop simultaneously at both ends of a short piece
without greatly inhibiting each other if each produces an inhibiting sub-
stance?

FISSION AND BUDDING IN FLATWORMS AND ANNELIDS
PLANARIANS

In certain species of the triclad genera Planaria and Dugesia ( = Euplan-
aria) fission usually occurs in individuals above a certain length at a more
or less definite body -level posterior to the mouth. Morphological evidence
that a new individual is developing in the posterior region is usually lack-
ing in planarians, but repetition of the copulatory organs in the posterior
zooid region of D. dorotocephala has been reported (Kenk, 1935a). Ani-
mals showing this dupHcation were prevented from undergoing fission
by maintaining them at low temperature. According to Kennel (1883,
1888), cephahc ganglia, eyes, and a new pharynx develop before separa-
tion in the posterior zooid of a planarian from Trinidad. In D. paramensis,
which divides anterior to the pharynx, a new pharynx develops in the
anterior zooid preceding fission, suggesting some degree of physiological
isolation of the posterior zooid.

In D. dorotocephala, the material for most of the experimental work on
fission, difference in susceptibility, respiration, dye reduction, head fre-
quency, and motor reactions indicate that the region posterior to the
fission zone represents one or more zooids, partially isolated physiologi-
cally from the dominance of anterior regions.^ There is no evidence of a
posterior zooid in very small animals soon after hatching or in individuals
reconstituted from short pieces, but with increase in length the posterior
zooid region becomes distinguishable by various methods, though fission
does not usually occur until animals are 15 mm. or more in length. With
rapid growth in length the posterior zooid appears and fission takes place
in shorter animals than when growth is slow, and with increasing degrees
of inhibition of head development the length attained before fission de-
creases (Child, 191 It/). Evidently the range of effective dominance in-
creases with increase in length of the animal, but in general less rapidly
than the length, so that progressive physiological isolation of the posterior
region results.

^ See pp. 42, 43, 45, 109, 112, 181.

322

PATTERNS AND PROBLEMS OF DEVELOPMENT

Fig. 112. — Elonga-
tion and permanent
alteration of form in
Dugesia doroloceph-
ala resulting from re-
peated independent
reaction of posterior
zooid without fission.

The act of fission in large intact animals may begin
with independent motor reaction of the posterior zooid
region, consisting in attachment to the substratum, while
the anterior zooid attempts to advance. In the struggle
the body anterior to the fission zone often becomes
greatly elongated until rupture results. Following rup-
ture, both zooids undergo reconstitution. Fission some-
times results from sudden longitudinal contraction an-
terior and posterior to the fission zone. With slow growth
in length in the laboratory animals may attain more than
twice the usual length without fission; in these growth
beyond a certain length is almost entirely in the posterior
zooid region (see Fig. i6, p. 42). In many of them the
posterior region shows independent reaction; but fission
is not completed, either because the animals are not ac-
tive enough or because the tissues rupture less readily
than in smaller, younger animals. By frequent repetition
of the independent reactions the posterior part of the
anterior zooid becomes permanently much elongated and
very slender (Fig. 112), and there is complete or almost
complete absence of control of the posterior zooid region
by more anterior levels. It is merely dragged about as the
anterior zooid advances and often interferes with the ad-
vance by attaching itself, but it is usually detached sooner
or later by the struggles of the anterior zooid or may re-
lease itself after the anterior zooid has ceased to struggle.
In these animals physiological isolation of the posterior
zooid has become directly visible.

Fission is readily induced by removal of the head, even
in animals far below the length at which it usually occurs
(Child, 1910a). Low concentrations of anesthetics and
other inhibiting agents also induce fission, apparently
by decreasing dominance; and a second division of the
posterior fission piece often follows the first. Under these
conditions fissions are usually limited to the period when
the regenerating head has developed far enough to deter-
mine more or less motor activity but apparently has not
attained full dominance. If there is no fission during this
period, it usually occurs only after increase in length or
another removal of the head. When animals with differ-

DOMINANCE AND PHYSIOLOGICAL ISOLATION 323

entially inhibited heads are fed, the length attained before fission is
less than in animals with normal heads and decreases from teratoph-
thalmic, through teratomorphic, to anophthalmia head forms (p. 178).
Acephalic forms often divide near, or even anterior to, the middle when
stimulated to contraction, though only 3 or 4 mm. in length, dominance
being almost absent.

STENOSTOMUM

In the rhabdocoel families Stenostomidae and Microstomidae morpho-
logical differentiation of zooids precedes fission; and since increase in
length is more rapid than differentiation and separation of zooids in well-
fed, rapidly growing animals, new zooids develop before separation of
earlier generations; consequently, long chains of zooids in various stages
of development result (see Fig. 10, p. 27). Each zooid in the chain under-
goes division when it attains a certain length, and in Stenostommn the
length attained by any zooid before a new fission zone appears in its
posterior region increases with advance in development of its head, but
the length of a zooid when it first becomes distinguishable is approxi-
mately the same at different levels of the chain under uniform conditions.
Apparently the length of body that a particular head can dominate in-
creases as development of that head progresses. This relation is evident
in Figure 10.

With each new zooid new cephalic ganglia develop and reorganization
of the nerve cords posterior to each new head must result. It is probable
that the progressive extension of dominance posteriorly with advance in
head development is associated with progress posteriorly of this reorgani-
zation. In animals living at low temperatures with little food, conse-
quently with slow increase in length, single zooids may become as long as
well-fed chains at higher temperatures.

Stenostomum is a protandric hermaphrodite. When a chain attains the
male stage, the zooids already present in the chain develop and separate,
but there is no further development of new zooids. When the individuals
attain the later female stage, they cease to feed and the pharynx degen-
erates; but they continue to elongate until they are as long as many of
the agamic chains, but without development of new zooids. Apparently
increase in range of dominance keeps pace with the elongation; it is less
rapid than in well-fed chains. After section of the elongated females at
postcephalic levels the posterior piece remains acephalic' In this respect

9 These observations and experiments are chiefly with Stenostomum grande, but other species
show similar relations of zooids in the chains. Sexual individuals of other species have not
been obtained by the writer. See Child, 1902; Van Cleave, 1929.

324 PATTERNS AND PROBLEMS OF DEVELOPMENT

the female stage of Stenostomum resembles rhabdocoel species without
agamic fission ; heads do not regenerate from levels any considerable dis-
tance posterior to the cephaHc ganglia in those species.

ANNELIDS

Agamic reproduction by transverse fission and, in some species, by
budding appears rather commonly among annelids; zooid chains with
morphological development of zooids preceding separation are formed in
various polychete and oligochete species, and in the polychete syllids
several types of budding appear. In Aeolosoma, Nais, and some other
genera of microdrilous oligochetes, each zooid elongates, a new zooid
arises in its posterior region, and the length attained by each zooid pre-
ceding this division is apparently related, as in Stenostomum, to the stage
of development attained by the head region of the zooid. Development
of each new head must involve more or less reorganization of the central
nervous system posterior to it; consequently, the length of body domi-
nated by a developing head region is probably correlated with progress
of this reorganization, as suggested above for Stenostomum.

The position of the fission zone in the oligochete Pristina longiseta can
be altered experimentally by nutrition, temperature, and condition of
medium (Van Cleave, 1937). With conditions favoring rapid growth of
new segments — abundant food, temperature of 2o°-25° C, and fresh cul-
ture medium; also with good conditions at 10° C. — new fission zones ap-
pear at shorter distances from a head than with poor nutrition at 20°-
25° C. or with previously used culture medium and good nutrition at the
same temperature. In other words, under conditions permitting rapid
growth physiological isolation occurs at shorter distances; under condi-
tions unfavorable to growth, at greater distances from a dominant head
region. Position of fission zones in various other species is probably simi-
larly determined (see, e.g., Eckert, 1927). Among the polychetes, the
syllids show various types of budding as well as fissions. Physiological
isolation is apparently concerned in these, but experimental data are
lacking.

AGAMIC REPRODUCTIONS IN OTHER ANIMAL GROUPS

Division in many protozoa is apparently nothing more than cell divi-
sion, but in some of the axiate forms there are indications of dominance
and physiological isolation. Cytoplasmic reorganization of the body of
Paramecium into two zooids begins before nuclear changes are visible pre-

DOMINANCE AND PHYSIOLOGICAL ISOLATION 325

ceding division, and the location of the fission plane is apparently deter-
mined soon after the preceding fission and several hours before it be-
comes visible (Peebles, 1912). Under certain experimental conditions Par-
amecium forms chains of zooids (Hinrichs, 1927), and some protozoan
species form chains naturally. Some of the cihates that form multiaxiate
complexes give rather definite evidence of dominance and physiological
isolation in the localization and sequence of new zooids — for example,
Zoothamnium (Faure-Fremiet, 1930; Summers, 19380, h). Some of the
remarkable forms of agamic reproduction among the Suctoria present ex-
tremely interesting problems for the future; at present the determining
conditions are quite unknown (see pp. 609 14). In other suctorian species
forms of budding apparently essentially similar to those in other groups
appear.

Budding sequences in bryozoa afford beautiful examples of definite
spatial patterns, indicating dominance and physiological isolation, and
reconstitution also indicates dominance in axial relations.'" Development
of buds from stolons in sessile ascidians suggests that physiological isola-
tion is a factor in their origin. The remarkable types of budding in the
pelagic tunicates — for example, the migration of cell groups from the
ventral to the dorsal stolon and the difference between median and lateral
buds on the dorsal stolon of Doliolum, and the periodic arrangement of
buds in blocks or "wheels" in certain of the Salpidae — also suggest exist-
ence of very definite spatial patterns of ordering and control of these
phenomena; but here, too, experimental data are lacking." The relation
between agamic reproduction and regression of the original individual,
as described by Berrill (1935) for various ascidian species, indicates that
physiological isolation by decrease or elimination of dominance is con-
cerned.

AUTOTOMY AND FRAGMENTATION

The triclad Fonticola velata kept at or above 20° C. with abundant
food, on attaining a certain length, ceases to feed, the pharynx degen-
erates, and the body undergoes repeated fragmentation from the posterior
end anteriorly. Fragmentation at each successive level is preceded by in-
ternal degenerative changes. The isolated fragments, usually less than a
millimeter in diameter, encyst and, after extensive degeneration and re-

■° See, e.g., Brien, 1936, and his citations of earlier work; Brien et Huysmans, 1937, for pat-
terns of budding; Otto, 192 1, for axial relations in reconstitution.

" In this connection see M. E. Johnson, 1910; Ritter and Johnson, 191 1.

326 PATTERNS AND PROBLEMS OF DEVELOPMENT

constitution of organs, emerge as minute individuals. A considerable de-
gree of control of fragmentation is possible. Animals kept at low tem-
peratures increase in length slowly and do not fragment, even when they
have attained much greater length than the fragmentation length at
higher temperatures. Occurrence and course of degeneration and fragmen-
tation are also influenced by character of food supplied." It is not certain
in this case whether physiological isolation initiates the degenerative
changes at the posterior end or whether the degeneration results in physio-
logical isolation, but the latter appears more probable. It is evident to
observation that the posterior region about to separate as fragment has
ceased to behave like a normal posterior end, shows little or no reaction to
stimuli, and approaches spherical form.

Autotomy or fragmentation results from irritation in various nemer-
teans and some annelids (e.g., Lumhriculus) , apparently in consequence
of independent muscular contractions in different regions. Excitation of
postcephalic regions evidently isolates them physiologically for the mo-
ment, and independent contraction and physical isolation follow. A local
dominance of an excited region may perhaps be a factor in determining
length of pieces, that is, excitation at one level may prevent independent
reaction within a certain distance, which will probably differ with body-
level and other conditions.

REDUPLICATION OF PARTS

An actively growing part — for example, a tentacle — is apparently able
to inhibit development of another tentacle within a certain distance of
itself; but, if two or more such parts develop simultaneously, they appear
to have Httle or no inhibiting effect on each other. These relations appear
very clearly in the scyphozoan Haliclystus. After removal of the margin
of the umbrella by transverse section the eight tentacle groups and mar-
ginal organs appear simultaneously and develop rapidly, but after ob-
lique section reconstitution at the more distal levels retards or completely
inhibits that at more proximal levels (pp. 50-52 and Fig. 24). A similar
dominance of distal over more proximal levels appears in the scyphistoma
of Amelia, though somewhat less effective. Tentacle development is
greatly retarded at the more proximal levels of oblique section; but new
tentacles appear at these levels and develop slowly, the normal number,
sixteen, being gradually approached and perhaps attained finally (Child).

The activation at a certain stage of development of the segment-forming

" Child, 1913c, 1914^; W. A. Castle, 1928.

DOMINANCE AND PHYSIOLOGICAL ISOLATION 327

region in front of the anal segment in annelids suggests a partial physio-
logical isolation from anterior dominance at this stage, and the repetitive
formation of segments appears to be a reproduction resulting from some
degree of physiological isolation of each segment primordium from the
slight and short-range dominance of the growing region. The segment re-
mains a segment, instead of becoming a complete individual, either be-
cause it is subordinated as it develops to anterior dominance, probably
through the nervous system, or because capacity for head development
is absent. In some annelids subordination to anterior dominance is evi-
dently the factor maintaining segmental character, for, when more an-
terior parts are removed, a head regenerates from the segment or segments
adjoining the level of section. In other species the segment remains es-
sentially unaltered when more anterior parts are removed, but it is still
able to react to removal of posterior parts by regeneration of a posterior
end and a new segment-forming region. Successive physiological isola-
tions in the mesoderm may be involved in segment formation in other
segmented animals. Probably physiological isolation, perhaps local in ec-
toderm, mesoderm, or entoderm, plays a part in determining localization
and order of various other reduplications of organ systems and parts.
Such isolations may be temporary, occurring at certain stages of develop-
ment and being followed by reintegration.

DOMINANCE AND PHYSIOLOGICAL ISOLATION IN THE FUNCTIONAL
ACTIVITIES OF CERTAIN ORGANS

The ctenophore plate row affords an extremely interesting example of
functional dominance and physiological isolation. '•^ Each row is a gra-
dient with high end toward the apical (aboral) pole of the body, but each
plate in a row is capable of independent movement. Ordinarily, the whole
row is dominated by rhythmic impulses transmitted through the nervous
system from the central nervous tissue about the apical pole. When the
whole animal is subjected to anesthetics, cyanide, or various other depres-
sing agents in gradually toxic concentrations, the impulses from the apical
nervous system become less frequent and perhaps weaker, and in the
plate row susceptibility to the toxic agent decreases from the apical to
the oral end. After some time under the depressing conditions a new
rhythmic impulse, independent of, and more rapid than, the depressed
apical impulse, begins to appear in the less susceptible and less inhibited
oral two-thirds, half, or less of the row, according to experimental condi-

'•3 Parker, 1905; Child, 1917c, 1921a, pp. 212-20, 1933a, and literature cited in these papers.

328 PATTERNS AND PROBLEMS OF DEVELOPMENT

tions. This region is now physiologically isolated from the apical region,
and the region of origin of the new impulse dominates all levels of the row
oral to it. Later the continued toxic action of the agent may inhibit this
dominant region; and a second physiological isolation, with again a new
dominant region, may appear still farther orally in the row. In the long
plate rows of Mnemiopsis three successive physiological isolations, each
with its new dominant region, have been observed under inhibiting con-
ditions.

These isolations are similar, as regards relation to a gradient pattern
and a dominance, to establishment of posterior zooids in planarians and
other forms, except that in the ctenophore isolation results from decreased
dominance while in the planarian under natural conditions it usually re-
sults from increase in length but may be brought about experimentally
by decrease in dominance.

On recovery from the toxic effect a gradual functional reintegration of
the plate row takes place, the physiologically isolated regions being again
finally subordinated to apical dominance. In Pleurohrachia the extreme
oral end of the row may become temporarily dominant after certain de-
grees of toxic action and direction of transmission of impulse reverses
over the whole row. This is a reversal of polarity by differential inhibi-
tion, but with recovery the original dominance and direction of transmis-
sion gradually reappear, replacing the reversed transmission step by step
along the row.

Physiological isolation at any level of the row may also be brought
about by blocking the impulse from the apical organ by low temperature
or mechanical pressure or by direct stimulation which initiates impulses
independent of the apical impulse. By continued mechanical stimulation
at a particular level of the row a new dominant region can be established.
Impulses initiated in this region are transmitted orally, that is, down the
gradient, over the whole length of the row oral to the level of stimulation,
but acropetally only a short distance at first; with continued stimulation
acropetal transmission extends farther and may even reach the apical end
of the row, obliterating the apical impulse. After cessation of the stimula-
tion the dominant region established by it gradually disappears, and the
whole row is again integrated under apical dominance.

The vertebrate heart resembles the ctenophore plate row in many re-
spects as regards dominance and physiological isolation. In the fully de-
veloped functional heart the sinus region is the pacemaker, that is, the
dominant region, and it develops from the high end of the primary heart

DOMINANCE AND PHYSIOLOGICAL ISOLATION 329

gradient. In early development, however, contraction begins in the ven-
tricular region and appears progressively toward the sinus and finally at
the sinus. This probably does not represent a progressive shift of domi-
nance from the ventricular to the sinus region but results from the course
of growth and differentiation of the myocardium (see, e.g., Copenhaver,
1939). Dominance of the sinus does not become evident until contractility
has developed over the whole heart. If the sinus of the fully functional
heart is inhibited, physiological isolation results, and a beat is initiated
at a lower level. As in the ctenophore plate row, physiological isolation
of subordinate regions with resulting independent contraction can also
be brought about by blocking the impulse from the sinus by direct stimu-
lation of the part concerned. By transplantation of the dominant sino-
atrial region to the conus region in the embryonic chick heart in vitro,
reversal of direction of beat and determination of ventricular rhythm by
the transplanted sino-atrium have been brought about (Paff, 1936).

In animals with tubular heart and periodic, or occasional, reversal in
direction of passage of contraction along the heart — for example, ascidians
and various arthropods — the pacemaker or dominant region is not fixed
in position (Gerould, 1931, 1933; Wolf, 1932). The end which is domi-
nant at a certain time loses its dominance sooner or later, in consequence
either of fatigue or of some other change in condition; and the other end,
becoming physiologically isolated, becomes independent, initiates a
rhythm, and becomes dominant and the pacemaker before the region
originally dominant regains its original condition. Later the same decrease
in dominance occurs in the new pacemaker, and reversal in direction of
beat again results. As already noted, a similar reversal in direction of
transmission in the plate row of the ctenophore Pleiirobrachia can be in-
duced by differential inhibition and recovery or by direct stimulation of
subordinate regions. In differentially inhibited fish embryos the heart
may remain tubular with reversible beat (Gowanlock, 1923). In all these
cases with reversal of beat the heart gradient is apparently readily re-
versible, either periodically under natural conditions or experimentally
by physiological isolation of a previously subordinate part in consequence
of differential inhibition or of direct stimulation of the part.

At certain definite levels of the mammalian alimentary tract are regions
which, under ordinary conditions, dominate levels posterior to them for
a certain distance and may be more or less independent of each other or,
under other conditions, may be subordinated to the dominance of more an-
terior regions or to the most anterior dominant region. Under pathological

330 PATTERNS AND PROBLEMS OF DEVELOPMENT

or experimental conditions subordinate regions may become dominant,
in consequence either of decrease in the pre-existing dominance or of irri-
tation or direct stimulation of a subordinate region, and the direction of
dominance may be reversed locally or over a considerable length of the
tract. A more or less complete physiological closure of the morpho-
logically completely open intestine may result from local reversal in
direction of dominance and contraction produced by a local irritation
or lesion. The regions normally dominant represent the high ends of gra-
dients, as indicated most distinctly in the small intestine, by graded dif-
ferentials in rate of rhythmic contraction in isolated pieces, rate of res-
piration, length of latent period, irritabihty, tone, etc.'-* Here, again, the
similarity to the ctenophore plate row and to the heart is evident, as Alvarez
1928, chap, vii) has pointed out. A similar functional dominance is ap-
parently present in the ureter and will probably be found in various other
elongated organs in which functional activity progresses in a definite direc-
tion.

CONCLUSION

The evidence from agamic reproduction in both plants and animals and
from reconstitution of isolated pieces indicates that physiological domi-
nance tends to maintain the individual, zooid, axis, or part as an inte-
grated unit, while physiological isolation, like physical isolation, antag-
onizes this unity and paves the way for origin of new individuals, zooids,
axes, or parts. For the new development of the physiologically or physi-
cally isolated part, however, a new dominance and establishment of a
new gradient or gradients are necessary; and these changes are possible
only when the isolated part, or some of its cells, are capable of reacting
to the isolation by activation and attainment of a more generalized be-
havior and so of initiating the reorganization.

The importance of the nervous system in dominating and integrating
activities in later development and mature life is a familiar fact. Its early
differentiation and the localization of the chief aggregations of nervous
tissue in the high regions of the gradient pattern present in early stages
suggest that, as regards the general features of its pattern, the central
nervous system represents the morphological expression of the higher gra-
dient-levels and is physiologically a further development of the primitive
type of dominance. The influence of the nervous system in maintaining

'4 Most of our knowledge of these relations in the alimentary tract is due to Alvarez and
his co-workers. See Alvarez, 1928, and papers cited there.

DOMINANCE AND PHYSIOLOGICAL ISOLATION 331

structure and determining regeneration need not necessarily be specific
but may merely serve to maintain metabolism in the part concerned at a
level which maintains or makes possible the development of a certain
structure. For example, muscle may differentiate in the embryo, but later
it atrophies in the absence of innervation because in the earlier stages the
intrinsic metabolism of the cells is sufficiently intense to determine muscle
structure but later is not sufficient to maintain it in the absence of innerva-
tion. In the simpler animals dominance is apparently chiefly or wholly
nervous or neuroid, so far as axiate pattern is concerned. In the embry-
onic development of so-called "mosaic eggs" and of vertebrates the re-
gional cytoplasmic differentiation of the egg at the beginning of develop-
ment or chemical relations resulting from this differentiation may play a
part in bringing about the dominance of the nervous system.

CHAPTER X

PHYSIOLOGICAL DOMINANCE AND ORGANIZATION
IN RECONSTITUTION

PROGRESSIVE DETERMINATION OF POLARITY

DEVELOPMENT of bud and root axes in plants, of buds in ani-
mals, and the reconstitution of pieces shows that the new polar
pattern is not established all at once but begins with the domi-
nant region and extends progressively from it, as far as range of domi-
nance, size of isolated piece in relation to scale of organization, or pres-
ence of other dominant regions and polarities permit. The conclusion
seems to be justified that polarity itself develops progressively in these
cases. At an early stage of reconstitution, budding, or fission the apical
or anterior region may be the only part of the new axis present. What-
ever the level of section in the stem of Tuhularia or Corymorpha, the apical
region of the hydranth originates adjoining that level, except under in-
hibiting conditions which prevent attainment of a gradient-level high
enough for development of an apical region; then the level adjoining the
level of section sometimes gives rise to a more proximal region of the
hydranth, and the extreme apical region may be permanently absent. Al-
ways, however, the new axis develops from this level. Whatever the level
of section in the planarian body, if head regeneration is possible, it be-
gins at that level and determination of the new polar axis progresses
posteriorly from it. In embryonic development the so-called law of an-
teroposterior development indicates that, even though regional cytoplas-
mic differentiations are present in an egg, a physiological dominance and
gradient factors may determine the order of their developmental activity
though polarity is already present.

That a new polarity may originate gradually and progressively from a
dominant region in nonembryonic forms of development which involve
determination of new pattern seems evident from the available data.
Moreover, the evidence concerning origin and progressive extension of a
new gradient from a dominant region is in accord with this view. And,
finally, the evidence pointing to this conclusion has a very important

332

DOMINANCE IN RECONSTITUTION 333

bearing on the question of the nature of polarity and developmental pat-
tern in general. If polarity results from a molecular orientation or an
"intimate structure" of some sort, it is not easy to account for the gradual,
progressive origin of a polar axis and pattern, beginning at one end. If
we assume some factor which determines molecular orientation or inti-
mate structure progressively from one pole, that constitutes the primary
polarity; and the molecular or other structural orientation is a resultant,
an expression of polarity, like other features of polar pattern. But as re-
gards determination and progressive extension of a dynamic gradient from
an activated region, no assumptions are necessary, for these phenomena
are visible or can be made visible by various methods. If a structural
orientation or a localization of different substances is present along a
polar axis, it seems highly probable that it results from, rather than de-
termines, a physiological gradient involving dynamic factors. Without
operation of such factors it is not evident how the changes in structural
orientation or localization of substances required to account for changes
in polarity under experimental and other conditions are determined.

CAN A DOMINANT REGION ORIGINATE INDEPENDENTLY
OF OTHER PARTS?

If gradual progressive determination of a polarity, beginning at a domi-
nant region, does take place, the dominant region must originate inde-
pendently of other parts of the polar pattern. That it does originate inde-
pendently in various kinds of buds and in many reconstitutions was
briefly pointed out above. But the question whether it can develop inde-
pendently of other parts of the polar axis, whether it is a self-differentiat-
ing system, remains to be considered. Among animals the most striking
evidence of a high degree of independence is found in the partial uni-
polar, bipolar, and multipolar forms in hydroid, planarian, and annelid
reconstitution. When short pieces of Tiihularia or Corymorpha develop
into partial forms, these, except occasionally under inhibiting conditions,
represent the apical region and more or less of the normal polar pattern,
according to length of piece, scale of organization, and presence or ab-
sence of other partial polarities (Fig. 113, A-I)} The parts of the axis
formed in such cases appear normal in pattern, fully developed, even
though they consist of nothing but a hypostome or a hypostome with
distal tentacles. In partial forms developing from aggregates of dissoci-

' For other bipolar and multipolar forms see Fig. 15, B, p. 42; Fig. 116, p. 346; Fig. 120,
p. 3()o; and Fig. 124, p. 366.

Fig. 113, A-M. — Unipolar and bipolar coelenterate reconstitution. .4-7, apical partial
forms of Tuhularia or Corymorpha from pieces of different length (from Child, 1915c, and un-
published data); J-M, bipolar forms from stalk of the sessile scyphozoan Haliclyslus auricula
(from Watanabe, 1937).

DOMINANCE IN RECONSTITUTION 335

ated Corymorpha cells the apical region shows a similar independence of
other parts (Fig. 117, p. 347). It is possible, of course, that these partial
apical forms fall short of complete development in some, as yet unrecog-
nized, structural or functional characteristic, but no such lack is evident.
The hypostome reconstituted from a short piece is capable of reacting to,
and taking in, food, although the food immediately passes out at the
proximal end.

Figure 113, J-M, shows interesting bipolar forms reconstituting from
the stalk of the sessile scyphozoan Halidystus (see Fig. 23, p. 50). The
stalk differs rather widely in structure from the umbrella, but tentacle
groups and marginal organs develop directly from the healed surface
of section, and with simultaneous distal and proximal section frequency
of bipolar forms is high (Watanabe, 1937). Further changes involved in
reorganization of the stalk into an umbrella are much less rapid; and
sometimes dominance is too weak to bring them about, and the piece
remains essentially a stalk with tentacle groups at one or both ends
(Child).

It has already been pointed out that a hydranth or head, reconstituting
at the distal or anterior end of a piece from any level except that immedi-
ately adjoining the original hydranth or head, is "out of place," as truly
"heteromorphic" as a hydranth or head developing at the proximal or
posterior end of a piece. At either end they are hydranths or heads of new
individuals and determine reorganization of pattern over a greater or less
distance. When the hydranth or head develops at the distal or anterior
end of the piece, the new polarity is in the same direction as the old and
determines development of parts apical or anterior to the level of origin
of the piece; but when it develops from the proximal or posterior end, it
determines a polarity opposite in direction to that originally present. The
evidence supports the conclusion that hydranth or head is not determined
by other parts of the piece, for both can develop when other parts are not
present or in a relation to other parts quite different from the normal.
On a postoral planarian piece, for example, a head begins to develop be-
fore pharyngeal, oral, and prepharyngeal levels. If reconstitution were
determined anteriorly from more posterior levels, we should expect that
oral, pharyngeal, and prepharyngeal regions would develop successively,
and finally the head; but this is never the case. The same relations appear
in annelid reconstitution. Whatever the level of section, in species capa-
ble of head regeneration a head regenerates first, and in most species
only a certain number of segments characteristic for the species regener-

336 PATTERNS AND PROBLEMS OF DEVELOPMENT

ates posterior to the head, even though more were removed (e.g., Hy-
man, 1916a). In general it appears that a necessary condition for hy-
dranth or head formation on an isolated piece is an activation of the
cells concerned sufficiently intense to bring about a high degree of
physiological isolation from other parts. When they become independ-
ent of other parts, they reconstitute a new dominant region and so
begin development of a new individual. Any level of the Tuhularia or
Corymorpha stem, any postcephalic level of the planarian, and all, or the
more anterior, postcephalic levels of many nemerteans and annelids will
develop as hydranth or head if not dominated by some other part. In
other words, so far as other parts of the body are concerned, the hydranth,
or even its apical portion, and the head, or probably the nervous tissue
of the head, appear to be self-determining and self-differentiating sys-
tems. They may be regarded as expressions of the primary or funda-
mental action system of the species, at least so far as reconstitution is
concerned. In the reconstitution of various minor or subordinate and spe-
cialized axes, such as appendages of various sorts, relations as regards
the longitudinal or polar axis of the part are apparently essentially similar,
at least in earlier stages. The distal tip of the starfish arm regenerates
"out of place" as regards relations to other parts and basipetal deter-
mination of the axis follows.^ Regeneration of a whole distal part of an
amphibian limb may take place from a half-transverse section (p. 370),
and a distal part may even regenerate from a proximal surface of section
of a limb implanted with longitudinal axis reversed (p. 390).

Position of a hydranth or head at the distal or anterior end of a piece
may be determined by the original polarity of the piece, though it is not
necessarily so determined; but the polarity of the developing hydranth
or head is evidently independent of the original polarity, even when in
the same direction, for it does not "fit" into that polarity but makes it
over more or less completely, according to body-level of the piece. The
original symmetry or asymmetry of the piece may play a part in deter-
mining symmetry or asymmetry of the reconstituting dominant region
(pp. 387-89) but does not necessarily do so, for radial symmetry of hy-
dranth and dorsiventrality of head may develop in various positions with
respect to the original symmetry.

The actinian Harenadis affords interesting examples of development of
symmetry in reconstitution. Partial disks develop after partial transverse
section of body wall and esophagus (pp. 387-88) ; here radial pattern of

=! King, 1898, 1900; DawydofiF, 1901; Morgulis, 1912; Schapiro, 1914.

DOMINANCE IN RECONSTITUTION

337

H

H

H

the reconstituted partial disks and the complete disk at the distal end
after complete transverse section are evidently directly determined by
the original radial pattern. Under other conditions,
however, reconstitution of a radial or of a bilateral
pattern is possible in this animal quite independent-
ly of the original radial pattern (pp. 371-74). In
planarians the anterior tip of a head regenerating
from a complete transverse section is directed away
from the piece, and its dorsiventrality coincides with
that of the piece ; but a complete head may regener-
ate from a surface of section half, or less than half,
the width of the body or even from a longitudinal
cut surface (Beyer and Child, 1930). A complete
head may also regenerate from a partial transverse
section in various annelids (Morgan, 19026; von
Haffner, 193 1, etc.).

Von Haffner's experiments on Lumbriculus are of
special interest. Following removal of lateral pieces
from different body-levels with injury of axial or-
gans, heads developed from the more anterior levels,
both heads and posterior ends from middle levels,
the posterior ends anterior to the heads, and only
posterior ends from posterior levels (Fig. 114). In
the middle region, where both heads and posterior
ends develop, the heads decrease in size and devel-
opment and posterior ends increase in size from more
anterior levels of section posteriorly (Fig. 114, A',
F, Z). In this region posterior ends are anterior to
heads because heads develop from the cut surfaces
facing anteriorly, posterior ends from those facing
posteriorly. As von Haffner suggests, these relations
are doubtless due to the polarity of the original
animal, but it is perhaps possible to go a step farther
and suggest that differences in degree and proba-
bly in rate of activation and in dominance at differ-
ent body-levels are concerned. At the more anterior
levels the regenerating head dominates the new tissue completely, and
only an anterior end develops. At the middle levels dominance of the re-
generating head is less effective, and the dominance of the original ani-

FiG. 114. — Lumbricu-
lus variegalus, diagram-
matic, indicating char-
acter of regeneration in
relation to body-level
after removal of small
lateral pieces with in-
jury of axial organs; H,
head; P, posterior end
(after von Haffner, 1 93 1 ) .

338 PATTERNS AND PROBLEMS OF DEVELOPMENT

mal determines tail development; and, as activation and dominance of
the tissue regenerating head decrease from X to Z, tail development
increases because not antagonized so strongly by dominance of the de-
veloping head. At more posterior levels activation and dominance of
regenerating tissue are not sufficient to overcome the original domi-
nance, and only posterior ends develop. Both heads and posterior ends,
however, though developing from partial transverse sections, are dorsi-
ventrally and bilaterally complete.

In general, the polarity of a regenerated part, head, posterior end, or
appendage may coincide with, or be oblique or opposed to, that of the
original body; but the symmetry or asymmetry is usually in the same
direction as the original and may be complete, though regeneration is
from a partial section. If physiological axes are gradients, these features
of reconstitution are very simply accounted for. There is no polarity dif-
ferential on a transverse section; consequently, polarity in the reconstitut-
ing part is determined by the gradient arising in relation to activation
following section and outgrowth. On a section oblique to the original
polar axis there is a polar differential, and reconstitution in relation to
obhque section is often asymmetrical (pp. 50-53). Ventrodorsality and
dorsiventrahty do present a differential on a transverse plane of section,
and this must influence reconstitution from that plane; also, a partial
differential on the plane of section may be sufficient to initiate develop-
ment of the whole differential in the reconstituted part.

THE ORIGINAL NERVOUS SYSTEM IN RELATION
TO NEW DOMINANCE

Much experimental work has been done in the attempt to determine
whether or to what extent the nervous system of the original animal is a
determining or essential factor in reconstitution. The fundamental ques-
tion, as regards this problem, concerns the role of the nervous system in
determining a new dominant region, or, more specifically, the new nervous
system in that region. For example, do the parts of the nerve cords pres-
ent in a headless planarian piece play a part in determining the cephalic
ganglia in the regenerating head?

In hydroid reconstitution the hydranth is determined basipetally from
the apical region; and since reorganization in a short piece may involve
the whole piece in formation of the apical portion of a hydranth, irrespec-
tive of level of origin of the piece, it is highly improbable that the part
of the nerve net originally present in the piece plays any essential role in

DOMINANCE IN RECONSTITUTION 339

the reconstitution. Moreover, an aggregate of dissociated cells of Cory-
morpha and other hydroids can reconstitute a complete individual. Rate
of regeneration of the scyphomedusa Cassiopca is decreased by removal
of the marginal sense organs; but this is believed to indicate merely a
nervous influence on the general metabolic level, since respiration is also
decreased by removal of the sense organs (Gary, 19 16).

As regards head regeneration in planarians, Flexner (1898) and Keiller
(19 10) have maintained that the cephalic ganglia may develop independ-
ently of the nerve cords in the piece. In a land planarian the new ganglia
are said to develop in continuity with the anterior ends of the old nerve
cords (Bandier, 1936). However, the distance between the cut ends of
the nerve cords and the locus of the new ganglia is so small and nerve
fibers so difficult to distinguish in planarians that a definite conclusion
seems impossible. Even if the ganglia develop in continuity with the
nerve cords, it does not necessarily follow that their development is de-
termined by the cords; the ganglionic development apparently deter-
mines reorganization of the region of the nerve cords where it develops.
The fact that heads can develop from a partial transverse or even a
longitudinal cut surface lateral to the nerve cords leaves little doubt that
ganglia can develop independently of the old nerve cords, though lateral
branches of the cords, if present at the cut surface, may play a part in
localizing head development. The ganglia of the regenerating head of the
nemertean Lineus develop in complete independence of the old nerve
cords and later become connected with them by outgrowth of fibers from
the ganglia; in some other nemerteans the gangha develop at the anterior
end of the cords (Coe, 1934a, h).

It seems to be conclusively established that in the earthworm and some
other annelids head regeneration can occur in the complete absence of the
nerve cords at the cut surface.^ Nevertheless, the old nervous system may
play a part in localizing head regeneration. When the ventral cord is
present at the cut surface, it is evidently a factor in localizing head re-
generation in annelids.'' Regeneration of a posterior end does not take
place or is incomplete in absence of the nerve cord, and implanted pieces
of cord may localize posterior regeneration.^ In view of the data, it is

3 Goldfarb, 1909, 1914a; Siegmund, 1928; Bailey, 1930, 1939; Avel, 1932; Kropp, 1933;
Okada, 1934; Crowell, 1937; Painter, 1938.

■• Morgan, igo2h; von Haflner, 1931; Zhinkin, 1935; Sayles, 1937. See also Hyman, 1940,
"Aspects of regeneration in annelids," Amer. Nat., 74; Sayles, 1940, "Buds induced by im-
plants of posterior nerve cord, etc., "Biol. Bull., 78.

s G. E. Holmes, 1931; Zhinkin, 1936; Sayles, 1939.

340 PATTERNS AND PROBLEMS OF DEVELOPMENT

evident that the old nervous system is not an essential factor in establish-
ment of a new dominant region in many species of hydroids, planarians,
nemerteans, and annelids.

In certain triclads (e.g., Dendrocoelidae) and in various nemerteans
and annelids head regeneration occurs only anterior to a certain body-
level, characteristic for the species. So far as data are available, those
rhabdocoels that do not undergo fission, poly dads, and some nemerteans
and annehds do not reconstitute cephalic ganglia unless some portion of
the original gangHa remains; that is, determination of an entirely new
dominant region apparently does not occur.^ We do not know whether,
or to what extent, nonnervous cells are concerned in regeneration of
cephalic gangHa in these cases, but it is evident that only cells of the more
anterior body-levels, whether nerve cells or others, are able to give rise
to ganglia under known experimental conditions. This inability to re-
generate ganglia and head does not mean that no cells capable of regenera-
tion are present, for complete posterior regeneration occurs at levels
incapable of head formation. It may be suggested that rate or intensity
of activation following section at these levels is not sufficient to render the
cells independent of other parts, and so to make possible development of
new ganglia and head, but is sufficient to permit development of posterior
ends under the dominance of the old nervous system.

The hypothesis that absence of head regeneration posterior to certain
body-levels is due to decrease in number of undifferentiated "formative
cells" from anterior to posterior levels fails to account for the rapid and
extensive posterior regeneration at those same levels.^

Presence of the radial nerve at the cut surface, even if it is separated
by section farther proximally from the rest of the nervous system, is es-
sential for regeneration of the ophiurid arm, according to MorguHs (191 2).
Schapiro (19 14) finds that regeneration of the distal part of the starfish
arm is possible even when a rectangular piece, including both dorsal and
ventral body wall, is removed at a more proximal level of the arm, leaving
only lateral connection with other parts. In this experiment a piece of

'' For examples see the following: F. R. Lillic, 1901, a dendrocoelid triclad; Child, 1904ft,
19106; L. V. Morgan, 1905, the polyclad Leptoplana; Coe, 1932, a nemertean; T. H. Morgan,
1897, 19026; Hyman, 1916a; Sayles, 1936, annelids. Various other papers give similar data,
and unpublished data on a number of rhabdocoel species (Child) show absence of head re-
generation in absence of ganglia.

' For this hypothesis see Curtis and Schulze, 1924; Curtis and Hickman, 1926; Curtis, 1928.
Steinmann, 1926, describes extensive dediflerentiation in planarian reconstitution, and
Bandier, 1936, is unable to distinguish "formative cells" from other parenchyma cells in the
reconstitution of land planarians.

DOMINANCE IN RECONSTITUTION 341

the radial nerve remains at the distal cut surface but is isolated from
central parts, except for possible lateral connections. Whether the nerv-
ous system is essential to regeneration of a complete disk and arms from
a single arm in certain starfishes (e.g., Kellogg, 1904) is not known.

Reconstitution of the bryozoan from a statoblast and of certain ascid-
ians from winter buds seems to be essentially development of a new indi-
vidual from what is primarily little, if anything, more than a cell aggre-
gate; the new nervous system develops in complete isolation from the old.

The question of nervous influence on regeneration of amphibian ap-
pendages has been the object of extensive experiment, and the literature
shows marked difference on the conclusions reached, although most au-
thors find some nervous influence.^ Some find merely retardation of re-
generation and failure to attain normal size after ehmination of regions of
the spinal cord innervating the leg, and it has been repeatedly suggested
that this effect is due merely to lack of function in the regenerating leg.
However, the possibility of some degree of innervation by other than the
normal nerves has not always been excluded. Goldfarb denies nervous in-
fluence. Walter finds that absence of motor innervation is without eft'ect
but holds, as does Locatelli, that regeneration does not take place in ab-
sence of nervous connection and suggests that connection with spinal
ganglia is essential. According to Weiss, there is proliferation but no true
regeneration in absence of nervous connection. He regards the nervous
influence not as formative or determinative but as a tonic excitation from
the sympathetic system. Schotte's experiments lead him to a similar con-
clusion. Hamburger, working with anura, finds legs normal in form but
of small size, regenerated in apparent complete absence of innervation,
but admits that the possibihty of sympathetic innervation is not ab-
solutely excluded.

At present there is general agreement that absence of motor innerva-
tion does not prevent leg regeneration but may retard or prevent growth
to full size through absence of function. The experiments of Weiss and
Schotte indicate the necessity of sympathetic innervation, and these and
various other authors agree that the nervous influence is not formative.
Apparently it is a factor in determining and maintaining the physiologi-
cal state, the level of physiological activity, of the cefls which makes pos-
sible a sufficient activation following section to bring about the degree of
physiological isolation necessary for establishment of a new dominance

8 Goldstein, 1904; Wintrebert, 1904; Goldfarb, 1909; Wolff, 1910; Walter, 1911; Weiss,
1925a, 1930, p. 124; Schotte, 19266, c; Brunst, 1927; Hamburger, 1928; Locatelli, 1929, and
literature cited by these authors.

342 PATTERNS AND PROBLEMS OF DEVELOPMENT

and gradient in the outgrowing tissue. That the nervous influence is not
specifically formative is shown by experimental alterations of the course
of brachial or sciatic nerves. Nerves thus diverted may determine Hmb
development at other than the normal position within the limb field ; but
the sciatic nerve, led into the tail region, determines a tail.^ Transplanted
primordia of chick appendages develop irrespective of innervation (Ham-
burger, 1939), but this development does not involve origin of a new
dominance.

SERIAL HETEROMORPHOSIS OR HOMEOSIS

Among arthropods regeneration of an appendage characteristic of a
particular segment usually results in an organ more or less similar to that
removed, though perhaps not completely normal; but sometimes an ap-
pendage resembling more or less closely that of another segment, usually
the next posterior, develops." In the case of development of antenna-like
structures in place of eyes, as observed in various Crustacea, Herbst,
Janda, and Kfizenecky find that this type of regeneration occurs after
removal of the optic ganglia, even though only the distal part of the eye
stalk is removed. Removal of antennae near the base apparently favors
regeneration of leghke structures; but, according to Suster (1933), normal
regeneration of antennae is possible in the orthopteran S phodromantis
after removal of the ganglia. In all these cases the regenerated structure,
when not normal, resembles the appendage of a segment posterior to
that on which it regenerates. In the light of the data on reconstitution
in hydroids and planarians and on differential modification of develop-
ment by external agents, it may be suggested that here, also, degree or
intensity of activation of cells following section is an essential factor in
determining the segmental character of the regeneration. If the ganglion
most directly related to the part is present, it is probably a factor in
bringing about sufficient activation to determine normal segmental regen-
eration; in absence of the ganglion there is less activation and in many
cases no regeneration, or an appendage may regenerate, but its cells fail
to attain the physiological level necessary for development of an ap-
pendage normal to the segment concerned, and a structure resembling

'Locatelli, 1925; Guyenot et Schotte, 1926. Guyenot, 1928, reports similar results in a
reptile, Laceria.

"Antenna-like organs in place of eyes: Herbst, 1896, b, c, 1899, 19016; Janda, 1913;
Kfizenecky, 1913. Leglike organs in place of antennae: Friza and Przibram, 1910; Przibram,
1910, 1917, 1919a, b; Brecher, 1924; Borchardt, 1927; Suster, 1933. Crayfish legs, Zalpeter,
1927. Homeosis in general, Przibram, 1910, 1920.

DOMINANCE IN RECOXSTITUTION 343

the appendage of a segment representing a lower level of the polar gra-
dient of earlier stages results. Further evidence that degree of activation
following section is concerned is the increasing frequency of homeosis
with advancing age of the animal and its occurrence at low, but not at
high, temperatures in certain cases. MetaboUc rate and regenerative po-
tency decrease, in general, with advancing age in postembryonic or post-
larval stages and certainly decrease with decrease in temperature.

According to Zalpeter (1927), however, regenerating legs of the crayfish
may show characteristics of either more anterior or more posterior legs."
The similarity of the crayfish legs suggests that gradient differences in the
leg segments are not great and that slight variations in activation follow-
ing section may be sufficient to determine development of a more anterior
or more posterior leg on a certain segment. ^Moreover, a secondar\- gra-
dient with high end posterior is probably present at certain developmental
stages of the crayfish; if so, more anterior legs probably represent lower
levels of this gradient, and their development from more posterior seg-
ments may also result from unfavorable or inhibiting conditions.

After removal of uropods and caudal ganglion Herbst ''igiy) found no
regeneration of uropods in most cases; but in a few individuals uropods
were recognizable, even though the caudal ganglion was not reconstituted.
This is not a case of homeosis; but, like those cases, it suggests individual
differences in degree of activation at the cut surface. Occasionally it is
sufiicient to determine uropod development in absence of the gangUon.
Differences in physiological age or other differences in physiological con-
dition, variation in degree of injur>-, maimer of healing, etc., may account
for the var}'ing results. Supposedly, the cells concerned in homeotic re-
generation are determined as appendage cells, perhaps as appendage of
a more or less hmited number of segments; but they are evidently not
fixedly determined as cells of the appendage of a particular segment.

DOMINANCE AND SCALE OF ORGANIZATION

By "scale of organization" is meant the spatial order of magnitude of
the developmental pattern in relation to an axis. In reconstitution of hy-
droids, planarians. nemerteans, annelids, and various other forms scale
of organization may vary greatly wdth physiological condition of the origi-
nal individual, with body-level from which the part is isolated, and with

" Przibram mentions cases of replacement of posterior by an anterior wing in insects and
some other cases of appendages characteristic of more anterior segments, but conditions of
origin of these are unknown.

344 PATTERNS AND PROBLEMS OF DEVELOPMENT

natural and experimental environmental conditions. Proportional scale
of different pattern components may also show similar variation (sec
chaps, v-vii). Reconstitution of parts of eggs, of isolated blastomeres
or blastomere groups, and of parts of later embryonic stages shows in
some forms considerable variation in scale of organization with size of
isolated part; but experiments directed toward alteration and control of
scale are few. In many eggs scale is so stably determined at the beginning
of embryonic development that no great alteration appears under condi-
tions thus far employed. Although data of many authors show differences
in scale, their importance for problems of embryonic, as well as reconstitu-
tional, development seems not to have been fully recognized.

At present it appears difficult to account for many of the expressions
of scale of organization except in terms of dominance and gradient pat-
tern. According to these terms, a gradient established in an isolated part
in good condition will, in general, represent a more intense activity at its
high end, will extend over a greater distance, and will perhaps be less
steep than one established under unfavorable conditions; consequently,
the local fields developing along its course, within which the particular
parts develop, will also be longer and the scale of organization therefore
larger. Such an interpretation must, of course, remain general until we
know more about metabolism and its changes at different levels of a
physiological axis in reconstitution and embryonic development and about
the relation of dominance and determination to such factors. A localized
active region, however it originates, may determine a gradient or gradient
system and so become a dominant region, or a gradient may arise as a
direct reaction to an environmental gradient. In the latter case the en-
vironmental gradient is merely the initiating factor; protoplasmic consti-
tution and condition are the factors determining the final character of the
gradient; and the high end apparently is more or less dominant. Domi-
nance and the gradient establish a physiological basis for definite axiate
pattern. In short, the dominant region is or may be an inductor or "or-
ganizer," that is, it determines the organization or reorganization of other
parts. A few examples of alteration of scale of organization in reconstitu-
tion will serve to indicate some of the relations between scale and certain
factors.

SCALE OF ORGANIZATION IN HYDROID RECONSTITUTION

Many investigators have noted the great variation in length of the
hydranth primordium in reconstitution of stem pieces of Tubularia.'' The

" See, e.g., Bickford, 1894; Driesch, 1897, 1899; Morgan, igoib, 1902(7, 1903(1; Child, 1907a.

DOMINANCE IN RECONSTITUTION

345

formation of this primordium and its decrease in length from distal to
proximal stem-levels were discussed in chapter ii. Physiological condition
of the individual from which the piece was taken is also
a factor in determining primordium length. Pieces from
animals in good condition, as indicated by size and ac-
tivity of the original hydranth, generally develop long-
er primordia at given stem-levels than pieces from ani-
mals with small degenerating hydranths or none. Pri-
mordium length can also be altered by inhibiting and
accelerating conditions. It is increased by increase in
temperature, up to a certain limit, and by slightly
hypotonic sea water and is decreased by inhibiting
agents, anesthetics, KCN (Child, 1931), decrease in
temperature, etc. It is also decreased by the domi-
nance of another hydranth within a certain distance
of it. Figure 115 shows the difference in length under
different conditions in stems of the same diameter and,
so far as could be determined, in similar physiological
condition preceding experiment. Similar differences in
primordium length appear in the same stem with dif-
ference in length of piece, stem-level, or with develop-
ment at opposite ends of a long piece. Similar dif-
ferences in scale of hydranth primordium appear in
Corymorpha; but since perisarc is absent and the
pieces change their shape by contraction and ex-
tension, measurements of primordium length are less
exact.

In both these hydroids primordium determination
very often occurs from both cut ends of a piece, and
bipolar or multipolar forms result. In these the portion
of each axis developing is determined by the scale of
organization and the portion of the piece occupied by
the other axis or axes. Each axis is complete from its
apical end basipetally, as far as determination and
development extend. In short pieces which become
unipolar with hydranth at the distal end, the activation at the proximal end
following section may be sufficient to shorten the gradient and hydranth
primordium at the distal end, that is, to decrease scale of organization
there. Under conditions resulting in difference in scale, short pieces of

B

Fig. 115, A, B. —
Extreme differences
in length of hydranth
primordium in Tubu-
lar ia pieces. A, high
temperature or hy-
potonic sea water; B,
physiological or ex-
ternal inhibiting fac-
tors.

346

PATTERNS AND PROBLEMS OF DEVELOPMENT

the same length may give rise to forms ranging from a complete individual
on a small scale (Fig. ii6, A) through unipolar and bipolar partial forms
(Fig. 1 1 6, B F) to unipolar or bipolar hypostomes (Fig. ii6, G, H).^^
In short pieces bipolar frequency may be increased and scale of organiza-

H

Fig. ii6, A-H. — Corymorpha. Forms resulting from differences in scale of organization in
pieces of the same length.

tion decreased by exposure to inhibiting agents for a short time after
section and unipolar frequency, and scale of organization may be increased
by differential conditioning to low concentrations of inhibiting agents or
by inhibiting development at one end of the piece. This is accomplished
in Tiihularia by closing the end with parafhn or sticking it in sand, and

'•5 For earlier observations on unipolar and bipolar forms of Tubularia see authors cited in
footnote 6, p. 36. For partial forms of Corymorpha see Child, 19266, 1927a, b.

Fig. 117, A-D. — Corymorpha. Differences in scale of organization in aggregates of disso-
ciated cells remaining undisturbed in contact with glass bottom of container (from Child,
1928c).

348

PATTERNS AND PROBLEMS OF DEVELOPMENT

in Corymorpha by contact with the glass of the container or by low oxygen
provided in some other way.

Differences in scale in development of aggregates of dissociated cells
from Corymorpha stems are also extreme, not only with size of aggregate
but with external conditions (Child, 1928c). Aggregates remaining un-
disturbed in contact with a glass surface usually (79 per cent) develop
apical ends from the free surface and a base from the surface in contact
and become either completed individuals or "mosaic" forms, consisting

Fig. 118, A-D. — Corymorpha. Aggregates of dissociated cells moved about and turned over
from time to time (from Child, 1928c).

of apical and basal parts without intervening regions. The range of dif-
ference in scale in such forms with difference in size of aggregate is indi-
cated in Figure 117, A-D. Aggregates frequently moved about and turned
over develop predominantly unipolar or bipolar apical parts of hydranths
on a much larger scale (Fig. 118, A-D). There is no question as to the
capacity of these stem pieces and aggregates to give rise to complete in-
dividuals. What actually develops in any particular case depends on scale
of organization, as determined by physiological and external conditions.
It is evident that scale of organization determined in a piece may greatly
exceed length of piece, or in cell aggregates, size of aggregate, or in bipolar
and multipolar forms, length to which each axis can develop. Each gra-

DOMINANCE IN RECONSTITUTION 349

dient is then only a partial gradient, and partial axial development re-
sults. It is also evident that development along each axis is determined
progressively from its apical end, for development in the partial forms
includes as much of the hydranth from the apical end basipetally as scale
of organization and available length permit. In the naked Corymorpha
pieces and aggregates the gradients corresponding to the partial axes can
be rendered directly visible by differential dye reduction soon after section
and before there is any indication of hydranth morphogenesis (pp. 97-98).
Occasionally a new axis, more or less inhibited by the dominance of a
hydranth or by external conditions, determines development of only prox-
imal parts of a hydranth — for example, a circle of proximal tentacles
without manubrium, or a few proximal tentacles, or sometimes only a
single tentacle at the apex of the axis. Apparently, the high end of the
inhibited partial gradient in these forms is not high enough to determine
the more distal parts of the hydranth. In most cases, however, either the
apical part of the hydranth develops or hydranth development is com-
pletely inhibited. Similar changes in scale of organization appear in the
reconstitution of other hydroids, of the scyphozoan Haliclystus, of the
scyphistoma of Aurelia, and of actinians.

SCALE OF ORGANIZATION IN PLANARIAN RECONSTITUTION

A wide range in scale of organization appears in planarian reconstitu-
tion. It has long been known from work of many investigators that iso-
lated pieces of different length may reconstitute complete individuals of
different size. A long and a short piece may be of the same width, but a
smaller head regenerates on the short, than on the long, piece; and the
individual resulting gradually approaches normal proportions, though it
may not attain them unless it is fed after reconstitution. On the other
hand, pieces of the same length from different body-levels differ as re-
gards relative size or length of parts. Size of head and distance between
head and pharynx, that is, length of prepharyngeal region, decrease from
anterior to posterior levels of origin of piece, or in species with fission zone,
to this zone, and both increase again in the posterior zooid (pp. 44-46).

Histological studies of planarian reconstitution have led to different
conclusions concerning the method of formation of the new tissue regen-
erating at anterior and posterior ends of a piece. Some have maintained
that both the new tissue and the changes in shape and proportions of the
piece result largely or wholly from migration of cells with little or no
proliferation and growth, while others have observed mitoses following

35© PATTERNS AND PROBLEMS OF DEVELOPMENT

section, and still others have recorded occurrence of amitosis in the new
tissue/^ Those who hold that cell migration is the chief or only factor in
planarian reconstitution apparently believe that some factor brings about
migration not only of cells but of whole organs, such as the pharynx, until
normal form and proportions are approached or attained (morphallaxis,
Morgan) ; but the fact that these changes in form and proportion are much
less rapid and less complete when the animals are not fed than in those
well fed as soon as they are able to feed after reconstitution suggests that
proliferation and differential growth are concerned. Growth of cells certain-
ly takes place in the new tissue of Dugesia. In earlier stages of head and
tail regeneration nuclei are much closer together, and there is much less
cytoplasm than in fully developed parts. Whatever the role of cell migra-
tion in planarian reconstitution, other factors are certainly concerned in
determining the differences in size of head and scale of organization in
relation to body-level and length of piece. In pieces of equal length rate
of head regeneration and length of prepharyngeal region decrease pos-
teriorly to the fission zone (Watanabe, 1935a; Rulon, 1936a), or in species
without posterior zooid, to the posterior end or as far posteriorly as head
regeneration occurs. In pieces below a certain length, which differs with
body-level, head regeneration is partly or completely inhibited by a stimu-
lation from the posterior cut surface and the activation of cells following
section; delay of posterior section decreases or abolishes this inhibition
(pp. 181-83). Nutritive condition of the animal is also concerned; in
pieces below certain lengths head regeneration is more inhibited in starved
than in well-fed animals (Child, 1920a).

Form and proportions of "normal" well-fed animals apparently repre-
sent approach to, or attainment of, an equilibrium of supply and demand
of material at the different physiological levels of the pattern. Reconsti-
tuting pieces differ at first from this norm according to length of piece
and level of origin, and the relative requirements of parts also differ, so
that some increase in size more rapidly than others and more rapidly in

'■t As regards cell migration, see Stevens, 1907; Steinmann, 1926; Bandier, 1936, and cita-
tions by these authors. Bandier finds no evidence of cell division of any kind in reconstitution
of a land planarian. Flexner (1898) observed mitoses following section in a planarian. In
pieces of Dugesia dorotocephala numerous mitoses appear during the first day or two following
section, both adjoining level of section and in the region in which the new pharynx develops,
and scattered mitoses appear in other parts of the piece (Child, unpublished). Bardeen (1902)
and P. Lang (191 2, 1913a, b) believe that amitosis occurs, but Steinmann and Bandier find no
amitosis. Murray (1927) observed amitosis in living parenchyma cells of Dugesia in tissue
culture and was able to follow the stages of division from beginning to complete cytoplasmic
separation. The writer has also observed these divisions in Murray's cultures.

DOMINANCE IN RECONSTITUTION 351

fed than in starving animals, until the equilibrium is approached or at-
tained. Partial individuals approach an equilibrium different from entire
reconstituted individuals. In acephalic forms, for example, a posterior
zooid develops, even in very short pieces, because the dominant head re-
gion is absent, and this zooid increases in length at the expense of the
region anterior to it in starving acephalic forms, often becoming relatively
very long. In starving intact animals different parts do not undergo re-
duction at the same rate. Apparently the most active, or most contin-
uously active parts decrease least rapidly because they are able more
nearly to maintain themselves at the expense of other parts than less
active regions. The head decreases in size less rapidly than the body;
but most of the digestive tract, in which the level of functional metabolism
is undoubtedly very low in the absence of food, may disappear com-
pletely. With repeated reconstitution of starving animals scale of organi-
zation may become very small (S. J. Holmes, 191 1). In short, with the
progress of experimental analysis reconstitution of large individuals from
long, and smaller individuals from shorter, pieces and the changes in form
and proportion appear somewhat less mysterious than they have appeared
to certain authors in the past. They are evidently expressions of physio-
logical factors that can be altered and controlled experimentally. Scale
of organization and morphological type of head depend upon degree of
activation or inhibition of cells concerned in its formation, and this affects
its rate of development and its dominance ; differential growth in relation
to gradient-level and available nutritive supply appears to be the essen-
tial factor in the change in form and proportion.

The distance between regenerated planarian head and pharynx, that is,
scale of organization of the prepharyngeal region, can be altered experi-
mentally in pieces from the postpharyngeal region. In pieces representing
this region (Z of Fig. 119, ^) the position of the pharynx reconstituting
in normal environment is approximately that indicated in Figure 119, B,
that is, somewhat anterior to the middle. Under inhibiting conditions —
low concentrations of anesthetics and other toxic agents, or culture water
fouled by dead planarians — head regeneration is inhibited, the head is
small, and scale of organization of prepharyngeal and pharyngeal regions
decreases according to degree of inhibition of the head (Fig. 119, C, D).
In extreme cases no pharynx develops, and there is no reorganization of
a prepharyngeal region (Child, 191 if, 19296). Reconstitution at higher
temperatures (26°-28° C.) is more rapid; heads are larger; and scale of or-
ganization of prepharyngeal and pharyngeal regions is increased. In forms

352

PATTERNS AND PROBLEMS OF DEVELOPMENT

like C and D of Figure 119, most of the piece represents one or more
posterior zooids, and fission often occurs anterior to the middle; that is,
the dominance of the inhibited head does not extend over the whole
length of the piece, often not over the anterior half.

Short planarian pieces, one-tenth or less of the body length, from levels
near the head (A' and Y of Fig. 119, A) often develop as tailless forms or
forms with very small posterior outgrowths and no pharynx or mouth

zzx

Y

Fig. 119, A-G. — Scale of organization in reconstitution of Dugesia dorotocephala. A,
outline, indicating body-levels of pieces X, Y, Z; B, reconstitution of piece Z in normal en-
vironment; C, D, decrease in scale of organization of preoral region of Z-pieces under inhibiting
conditions. E-G, physiological inhibition of posterior development in pieces shorter than scale
of organization; E, head+.Y; F and G, X or I' {B-D, from Child, 19296).

(Fig. 119, F, G). Short pieces, including the original head, are also often
tailless (Fig. 119, E).

Table 8 gives percentages of forms, either tailless or with tail inhibited,
reconstituting from 1/16 pieces: A" and Y (Fig. 119, A), 1/8 pieces;
X + Y, without the original head; and A' and A' + F with the head.
According to the table, posterior development is absent or inhibited in
64 per cent of A'-pieces, in 30 per cent of F-pieces, and not at all in pieces
X + Y and head + A' + Y. Moreover, in A^-pieces with regenerating
head posterior development is much more inhibited (64 per cent) than in
pieces with posterior ends at the same level but with the original fully
developed head present (8 per cent). Apparently, in these short pieces the

DOMINANCE IN RECONSTITUTION

353

anterior region activated by section or beginning regeneration is more
effective in determining a scale of organization longer than the piece than
is the original head. Actually, however, the effective length of the A'-pieces
for reconstitution of more posterior levels is somewhat less than the length
of the piece, for the cells of the anterior part of the piece are directly con-
cerned in head regeneration, while in head+A'-pieces this is not the case.
From isolated heads posterior ends never develop, perhaps because of
lack of potency ; but the table shows that at the levels A' and F length of
piece, rather than presence or absence of potency, determines presence
or inhibition of posterior development.

TABLE 8

Alteration of Scale of Organization in Relation to
Length of Piece and Body-Level in Short Anterior
Pieces of Dugesia ( = Enplanaria) dorotocephala (loo PIECES
in Each Lot

(Data from Rulon and Child, 19376)

Tailless

Tail
Inhibited

Normal

Dead

X

41
22

23
8

32
60

100
92

100

4

F

10

X-\-Y

Head+A'

Head-i-A'+r

3

5

Posterior development is also inhibited by delaying posterior section
of A'-pieces for 24 hours after anterior section, that is, by activation and
initiation of head regeneration before posterior section. In all these cases
of posterior inhibition the scale of polar organization determined is ap-
parently longer than the piece. Dominance and the high gradient-levels
are so effective in maintaining or determining the piece as an anterior
region that not only is activation and outgrowth of cells at the posterior
end largely or wholly inhibited but the piece includes no level corre-
sponding to the pharyngeal region.

Exposure to KCN for a time after section decreases scale of organiza-
tion in these pieces, so that after return to water the frequency of pos-
terior development increases. In Table 9 posterior development is absent
or inhibited in 71 per cent of the A'-controls and in 41 per cent after KCN.
In the F-controls it is absent or inhibited in 31 per cent, and after KCN

354

PATTERNS AND PROBLEMS OF DEVELOPMENT

in 21 per cent. The few forms listed as apolar in Table 9 are without
either head or tail, and those Hsted as bipolar are bipolar heads. Percen-
tages of deaths in the different lots do not differ sufficiently to affect the
general result. With increase in frequency of posterior development after

TABLE 9

ALTERATION OF SCALE OF ORGANIZATION IN Dugesia PIECES BY EXPOSURE TO KCN

m/lOO,000 FOR 72 HOURS FOLLOWING SECTION (lOO PIECES IN EACH LOT)

(Data from Rulon and Child, 193 yi)

Tailless

Tail
Inhibited

Apolar

Bipolar

Tail and
Pharynx
Present

Dead

f^

Controls in water ■!

50

24
35
17

20
6

5
I

I

20
56
42
63

9

I

13

I ■*

{X

KCN m '100,000 <!

11'

I
3

17

2

14

TABLE 10

Differential Inhibition of Head Development in Those Pieces

OF Table 9 Which Develop Pharynx and Tail*

(Data from Rulon and Child, 19376)

Number

OF

Pieces

Head Forms

Normal

Teratoph-
thalmic

Terato-
morphic

Anoph-
thalmic

Acephalic

{X

Controls in water \

\y

f^

KCN m/ 100,000 •)

\y

20
56
42
63

20
48
30
26

5
3

I

I
4
4

I

3

33

*For the characteristics of the diEEerentially inhibited heads see pp. 177-80.

KCN there is also increase in frequency of differential inhibition of head
development, as shown in Table lo. That inhibition of head development
decreases scale of organization in long posterior pieces was shown by the
experiments described above. Table lo gives additional, though some-
what less direct, evidence of the same relation between head develop-
ment and scale of organization in these short anterior pieces.

DOMINANCE IN RECONSTITUTION 355

SCALE OF ORGANIZATION IN OTHER RECONSTITUTIONS AND
AGAMIC REPRODUCTIONS

Evidences of a very considerable range in scale of organization appear
in many organisms, but only a few further examples are noted here. Pieces
of Stentor about 1/27 of the volume of the original individual reconstitute
complete individuals (F. R. Lillie, 1896); but whether minimal size of
piece capable of such reconstitution varies with level of body from which
it is taken is not known. According to Morgan (1901c?), however, the
peristome is "too large" in individuals reconstituting from anterior pieces,
"too small" in those from posterior pieces. These differences are evi-
dently expressions of the longitudinal gradient shown by other methods
to be present. As in planarian pieces, these differences gradually decrease,
and with feeding the usual proportions may be attained.

Certain nemerteans show a very wide range of scale of organization in
reconstitution. Coe regards the regenerating nemertean head region as
dominant and inducing the further reorganization of the piece. '^ In an-
nelid reconstitution a wide range in scale of organization also appears.
In most annelids capable of anterior reconstitution this is only in part
regeneration, no more than a certain number of segments being regen-
erated, even though more were removed, the intervening parts being
formed by reorganization of old segments into segments characteristic of
more anterior levels. In some of the more transparent microdrilous oligo-
chetes this reorganization can be observed in the living animal, and in
some other cases its occurrence has been noted, but in many studies on an-
nelid reconstitution the question is not considered. The number of segments
reorganized may vary greatly in some species, but experimental analysis
of the extent of reorganization in relation to dominance of the regenerating
anterior regions is largely lacking. The case of Sahella is of special interest
because reorganization of "abdominal" into "thoracic" segments in re-
constitution is directly visible, since the two regions differ morphologi-
cally, and also because the reorganization may take place in relation to a
level of section before head regeneration; and in some cases reorganization
of a few segments occurs anteriorly from a posterior level of section. Evi-
dently the activation following anterior section, and sometimes that fol-
lowing posterior section, is sufficient to initiate the reorganization of other
segments, that from the posterior section being on a small scale, as might
be expected, since it opposes the polarity present. Evidently the reorgani-

'5 For data bearing on the question of scale in nemertean reconstitution see Nusbaum und
Oxner, 1910, iqii, 1912; Davydov, 1915; Coe, 1929, 1930, 1932, 1934a, b.

3S6 PATTERNS AND PROBLEMS OF DEVELOPMENT

zation induced by the anterior activation is not specifically different from
that induced anterior to a posterior section/'^ In a few annelid species a
large, and apparently indefinite, number of segments may regenerate an-
teriorly as well as posteriorly, as, for example, in Criodrilus (Janda, 1912a,
h; Tirala, 191 2). These differences in anterior reconstitution are similar
to those in different planarian species, that is, in some only a head region
regenerates, other parts developing by reorganization of levels posterior
to the level of section, while in others parts posterior to the head also
regenerate and there is little reorganization of old parts. These differences
doubtless depend in both groups on physiological conditions determining
rate of regeneration and rate or capacity for reorganization. In planarians
it has been possible, by means of inhibiting conditions, to bring about
head development almost wholly by reorganization with regeneration only
at the extreme tip (see Fig. 72, H, /, p. 191). As in planarians, headless
annelid pieces can determine development of posterior ends, but nothing
anterior to the level of origin of the piece develops except in relation to
a developing head region or a region activated by section and isolation
from more anterior levels. Some species show great variation in scale in
relation to length of piece, complete individuals developing from a single
segment in certain forms, in others only from longer pieces.'^ The varia-
tion in position of regenerated sex organs in reconstituted Criodrilus
(Janda, 1912a, h) suggests variarion in scale determined by differences in
dominance of the developing head region.

Short annehd pieces, including the head, or from levels near the head,
often fail to regenerate posterior ends, again like Euplanaria. This has
commonly been regarded as indicating lack of potency to develop a pos-
terior end in this anterior region, but the experiments on scale of organi-
zation in short anterior planarian pieces suggest that in the annelid, as in
the planarian, scale of organization may sometimes be "too long" for the

piece.

In buds, both plant and animal, the dominant apical region is the first
part of the new polar axis to appear. Adventitious plant buds may origi-
nate within a single cell or involve a group of cells (see pp. 17-19), but
the scale of the bud pattern increases with progress of development, and
successive axial levels are progressively determined. Buds of multicellular
animals apparently involve a considerable number of cells in their eariiest
stages, and considerable increase in size may occur before scale of organi-

■6 Berrill, 1931; Berrill and Mees, 1936; Gross and Huxley, 1935.
'7 Morgulis, 1907; Korschelt, 1919; Dehorne, 1932; Martin, 1933.

DOMINANCE IN RECONSTITUTION 357

zation is determined. That range of dominance and scale of organization
differ with conditions in buds, as elsewhere, is highly probable; but ex-
perimental analysis concerned with this problem has not been undertaken.
In fission the length attained by an individual before appearance of a
fission zone may vary with development of the head and with external
conditions, and length of a new zooid when it first becomes distinguish-
able may also vary. However, the posterior zooid of a pair may not rep-
resent the entire polar axis in early stages; the axis may be determined
progressively from a dominant region, as in buds (p. 332).

Questions of dominance and scale of organization in early embryonic
development will be taken up in later chapters, but attention is called to
a case of embryonic echinoderm reconstitution because it resembles so
closely the changes in scale in hydroid and planarian reconstitution. The
apical half of the sea-urchin embryo usually develops in water into an
entirely ectodermal blastula-like form; but when treated with hthium, it
may form entoderm and gastrulate.'^ The scale of organization deter-
mined in water is apparently too large for the apical half, but with decrease
of scale by hthium a complete individual develops. Short pieces of Tuhu-
laria and Corymorpha stem and short pieces of planarians show essentially
identical changes under inhibiting conditions, that is, scale of organiza-
tion decreases, so that a piece of a given length develops a larger part,
or the whole, of the polar pattern, instead of only the apical portion.
The basal half of the echinoderm embryo, however, can reconstitute an
apical region and become a complete larva. This is exactly what happens
in the hydroid stem pieces and in planarian and anneHd reconstitution.
Should we not attempt to account for all these essentially similar recon-
stitutions in essentially similar terms rather than by special hypotheses
for the sea-urchin embryo and others for hydroids, planarians, and an-
nehds? In many eggs and embryos scale of organization is more or less
fixed at the beginning of embryonic development, or certain cytoplasmic
regions must be present for complete development; these are evidently
beyond the stage of simple gradient pattern.

CONCLUSION

The evidence indicates that in both plants and animals origin of a new
dominant region is independent of other parts. Reconstitution of hy-
dranth or head is not determined by other parts of an isolated piece; on
the contrary, a certain degree of physiological or physical isolation from

'S See pp. 506-S; also von Ubisch, 1925ft, 1929; Horstadius, 1936a.

358 PATTERNS AND PROBLEMS OF DEVELOPMENT

those parts is necessary for such reconstitution. In other words, reconsti-
tution of a hydranth or head occurs in spite of the rest of the piece, not
because other parts of the piece determine the replacement of parts re-
moved. Only when activation of the cells concerned is suflEicient to bring
about a certain degree of physiological isolation from other parts does a
hydranth or head develop from a stem piece or a postcephaHc piece. The
only demonstrated action of other parts on determination of a new domi-
nant region is inhibitory. Determination of the hydranth or head repre-
sents the first steps in determination of a new polar axis. The hydranth,
or its apical part, and the head, or probably the nervous tissue of the
head, are primary developmental expressions of the reaction system of
the species. Every stem level of Tuhiilaria, Corymorpha, and other hy-
droids, every postcephalic level of Dugesia and many other forms, will
develop as hydranth, or apical part of a hydranth, or as a head if its de-
velopment is not otherwise controlled or inhibited by a dominant region
or by external conditions.

A region dominant in development is an inductor or "organizer." It
determines the spatial pattern and localization of parts along the axis
concerned on a larger or smaller scale, according to conditions. This de-
termination does not necessarily, probably not usually, take place all at
once, but progressively from the dominant region. Scale of organization
apparently depends primarily on range and effectiveness of dominance.
Differences in scale represent differences in spatial localization and deter-
mination of parts. It seems evident that they depend on a factor or fac-
tors operative and effective over a certain spatial range which varies with
the activity of the piece or system concerned. In view of the various other
hues of evidence, the simplest assumption is that of a dynamic gradient
resulting from the activation following section and from the physiological
isolation from more apical or anterior regions in isolated pieces, or deter-
mined by an external differential or in some cases by localization of an
activated region experimentally or otherwise without section. Such a gra-
dient is, of course, only the initiating factor in pattern. However it is
determined, its character, the changes it undergoes, and the kind of pat-
tern and organism that results depend on the specific constitution, nuclear
and cytoplasmic, of the living system in which it appears.

CHAPTER XI

RECONSTITUTIONAL PATTERNS IN RELATION TO
EXPERIMENTAL CONDITIONS

RECONSTITUTIONAL development provides a wealth of mate-
rial showing establishment of new developmental patterns in
- definite relations to experimental conditions. New patterns can
be determined experimentally not only in isolated pieces of mature indi-
viduals but in eggs and early embryonic stages of some forms. From some
of these experimentally determined patterns and the ways in which they
are determined we can learn something about the real beginnings of de-
velopment and how these beginnings can be initiated. The egg at the be-
ginning of embryonic development usually has already an established
pattern, often with considerable regional differentiation and relatively
stable. Except in those cases showing capacity for reconstitution of pat-
tern, most eggs and early embryonic stages tell us little concerning the
beginnings of developmental patterns.

NEW PATTERN IN RELATION TO SECTION

In the preceding chapter it was shown that in coelenterates, planarians,
nemerteans, and annelids new dominant regions may be determined by
the activation adjoining level of section, that they are self-differentiating,
and that they may act as inductors of a new polarity. Even when this
polarity is in the same direction as that of the original individual, it is a
new polarity, at least as far as alteration of the pre-existing gradient and
pattern extends from the dominant region. The unipolar reconstitutions
of Tuhularia and Corymorpha in Figure 113, A-D (p. 334), and Figure 116,
A, B, E, G (p. 346), are cases in point. The bipolar forms of Figures 113
and 116 are cases in which activation at both ends of the piece was suffi-
cient to determine a new polar pattern over a greater or less distance.
Corymorpha pieces often give rise to multipolar forms, particularly in
short pieces from levels near the basal end, from animals in poor physio-
logical condition, under experimental inhibiting conditions, and some-
times from differential exposure. At least some of the polarities in these
are localized in a definite relation to the region of section, but some may

359

36o PATTERNS AND PROBLEMS OF DEVELOPMENT

not be. The most frequent forms of multipolarity, appearing in absence
of purposely applied experimental conditions other than isolation of
pieces by section, are multiphcations of manubria (Fig. 120). Each manu-
brium is normal and complete in form from its apical region as far basip-
etally as it develops, and each can be shown by the various methods

H I

Fig. 120. — Multipolar forms of Corymorpha (from Child, igibb, 192 ;«)

available to constitute a gradient pattern with high end apical. Forms
with numerous manubria usually develop from proximal stem-levels where
the diameter is greater than distally. Section at these levels often involves
more crushing and laceration, and closure of the cut end is less rapid and
regular than distally. The multiple manubria are apparently determined
by local differences in activity of cells or cell groups on the end of the
piece. There is some evidence that the longitudinal entodermal canals

RECONSTITUTIONAL PATTERNS IN EXPERIMENT 361

just beneath the ectoderm may sometimes play a part in locahzing more
active regions. In pieces of smaller diameter closure of the cut end pre-
ceding development is more rapid and uniform, and a single gradient usu-
ally results.

Multiplication of basal gradients and structures is also frequent (Fig.
120, C, F, G, H, I), either at the end of the piece opposite that developing
apical parts or elsewhere, even between multiple manubria (Fig. 120,
H, I). As previously noted, basal axes are somewhat inhibited axes, and
their development is favored by slightly inhibiting or depressing condi-
tions. Many of the multipolar forms are mosaics of partial axes (Fig. 1 20,
A, B, C, I), that is, each axis, whether manubrium, hydranth, or basal,
develops from its apical region as far as its scale of organization, presence
of other developing axes, and size of piece permit. These cases give fur-
ther evidence that each axis develops basipetally.'

Physical isolation of the piece from other parts, of course, plays a part
in initiating reconstitution in pieces of hydroid bodies; but physiological
isolation is also a factor in determining bipolar and many multipolar
forms. In long pieces without the original hydranth, dominance of the
distal region is insufhcient to prevent development of a dominant region
at the proximal end; in short pieces there is little gradient, the two ends
are physiologically isolated from each other, begin development at about
the same time, and neither can inhibit the other. Appearance of new
dominant regions adjoining the level of section and induction of new axes
in relation to them, with development of unipolar, bipolar, or multipolar
forms, is very generally characteristic of coelenterates.''

In some hydroids outgrowth of tissue from the level of section precedes
hydranth development, but the outgrowth represents establishment of a
new gradient, as is readily demonstrated; and the high region at the tip
gradually attains hydranth development if not inhibited, or grows as a
stolon if it does not attain hydranth-level in consequence of dominance
of other regions or inhibition by external conditions. ■^

' For other multipolar forms of Coryniorpha see Child, 1926&; 1927a, b.

2 Weimer, 1928, 1932, and many earlier papers by various authors on Hydra; numerous
papers on many hydroid species; Child, 1933^, and Watanabe, 1937, the scyphozoan Hali-
clystus; Gilchrist, 1937c, the scyphistoma of Aurelia; Child, 1903a, b, 1904a, 1908, 1909a, b,
the actinians Cerianthus and Harenactis. See also Figs. 14, 15 {B}, 24, 25, in chap. ii.

3 The algae Acetabularia mediterranea and A . wettsteinii, forms with a polar axis possessing
a considerable degree of organization and with whorls of branches, are single uninucleate cells.
They are mentioned here because of their remarkable similarity to hydroids in reconstitution.
Shorter pieces may even develop as bipolar partial forms (Hammerling, 1934, 1936). Exist-

362 PATTERNS AND PROBLEMS OF DEVELOPMENT

As regards establishment of a dominant region and induction of other
parts by it, reconstitution in planarians, nemerteans, and anneUds does
not differ essentially from that in coelenterates, except that regeneration
of new tissue from the cut surface is a characteristic feature. In absence
of inhibiting conditions the activation following section and isolation re-
sults in outgrowth of tissue, apparently more or less embryonic in char-
acter and undergoing differentiation. Usually the new dominance develops
from this regenerating tissue at the higher end of that part of the original
polar gradient present in the piece, and the tissue becomes wholly or in
part a head. The length of outgrowth and the postcephalic parts develop-
ing from it differ in different forms and in planarian pieces from different
body-levels of the original individual (Figs. 17-22 [pp. 43~46]). In some
planarians the outgrowth may be greatly decreased by inhibiting condi-
tions, so that the head develops in large part by reorganization posterior
to the level of section (Fig. 72, H [p. 191]; Fig. 119, D [p. 352]). Under
these conditions planarian head reconstitution approaches the type of
reconstitution of hydranth in Tuhularia and Corymorpha. Anterior out-
growth of new tissue and head development are increasingly inhibited
with decrease in length of piece in several planarian species; and in pieces
below a certain length, which differs in a definite way with level of origin
of piece, size and physiological condition of animal, and external condi-
tions, there is complete inhibition of head development, but development
of a posterior end is possible at the same level under the same conditions
(pp. 180-90). In absence of the tissue outgrowth from which head de-
velops, httle or no reorganization of body-levels anterior to level of origin
of the piece takes place, but all parts posterior to that level can still de-
velop. For example, pieces from the prepharyngeal region, cut short

ence and movement in opposite directions of two formative substances are postulated by
Hammerling, but it seems evident that dominance and a gradient are present along the axis
and that determination of new dominances and gradients by activation following section are
concerned in the reconstitution of these plants. Moreover, the postulated movement of forma-
tive substances in opposite directions from the nucleus, which is situated near the basal end,
in unipolar reconstitution of pieces, seems to require the presence of an axial differential of
some sort. The fact that nonnucleated pieces possess considerable capacity for reconstitution
appears to support this view. Whatever the situation as regards movement of substances,
orderly movement in opposite directions seems to require presence of a pattern of some sort,
and the data of reconstitution of these algae suggest that that pattern is a gradient. Sub-
stances from the nucleus may, of course, play an important part in determining physiological
condition and the character of later differentiation of parts, but how they can determine the
general spatial pattern and axiate order and the localization of secondary axes as lateral
branches does not appear.

RECONSTITUTIONAL PATTERNS IN EXPERIMENT 363

//

enough to be acephalic (/ of Fig. 121, A), develop pharynx and mouth
and all parts posterior to the mouth (Fig. 121, 5); but acephalic pieces
from a level just posterior to the mouth (7/ of Fig. 121, .4) do not develop
mouth, pharynx, or prepharyngeal region (Fig. 121, C). However, if there
is even rudimentary head development
(anophthalmic [p. 179]) in pieces from
postoral levels, the anterior part of the
piece reorganizes into a prepharyn-
geal, pharyngeal, and oral region (Fig.
121, D). Moreover, such forms, even
though head development never goes
beyond this stage, are distinctly more
like normal animals in motor behavior
than are acephalic forms.

Many investigators have recorded de-
velopment of bipolar forms from short
planarian pieces.'' Such forms may con-
sist merely of heads alone or of heads
and more or less of the anterior post-
cephalic region, and sometimes a phar-
ynx develops in one or both axes. Oc-
casionally also one or two posterior ends
develop from lateral regions, usually
after more or less elongation in conse-
quence of the opposed locomotor activ-
ity of the two heads (Fig. 122). Both
heads may be normal; or one, usually
the posterior, may be more or less chf-
ferentially inhibited (Fig. 122, E). Oc-
casionally both head and posterior end
develop from a posterior cut surface,
particularly if it is somewhat obhque (Fig. 122, G; also Fig. 25, D [p. 52]).
Bipolar forms may apparently develop from any body-level in sufficiently
short pieces, but in the species with a definite postoral fission zone they
seem to appear most frequently from the region of the fission zone, with
pieces of given length. In this region there is a shght rise in gradient-
level from the posterior end of the anterior zooid to the anterior end of

4 See, e.g., Morgan, 1898, 1900a, 1904(7, h; Bardeen, 1902, 1903; Child, 1911&; P. Lang,
1913a; Rustia, 1925; Lus, 1926; and other authors.

Fig. 121, A-D. — Development of
acephalic planarian forms from different
body-levels. .-1 , outline indicating levels
of sections / and //; B, prepharyngeal
acephalic form from level /, pharynx and
mouth reconstituted; C, postpharyngeal
acephalic form from level //, does not
reconstitute pharyn.x or prepharyngeal
region; D, postpharyngeal piece with
rudimentary anophthalmic head, recon-
stitutes pharynx, mouth, and prepha-
ryngeal region {B-D, from Child, 1911c,
1929&).

364

PATTERNS AND PROBLEMS OF DEVELOPMENT

the posterior zooid; consequently, there is Httle or no gradient difference
at the two ends of a short piece from this region, and neither is dominant.
A planarian head sectioned transversely at the level of the cephaHc ganglia
sometimes becomes bipolar by development of a second head opposed in
orientation from the cut surface (Fig. 122, F).

Bipolarity in short planarian pieces from other levels than the fission
zone can be greatly increased by differential inhibition with various
agents (Rustia, 1925). According to the gradient concept, the two cut
ends of such pieces differ but little because only a small fraction of the

Fig. 122, A-T. — Bipolar and apolar forms from short planarian pieces. A-F, Dugesia;
G, Cercyra papulosa, a marine triclad (after Lus, 1926); H, I, apolar forms.

polar gradient is included between them. Differential inhibition decreases
this difference still further, so that after return to water the two ends are
so nearly alike that neither can dominate the other; consequently, the
activation at each end is independent of that at the other, and the new
tissue develops into a head and determines a new polarity as far as length
of piece, scale of organization, and presence of the other polarity permit.

Under the usual conditions and under most experimental conditions
thus far employed, bipolar heads do not develop on pieces 1/4 or more
of the body length of large, mature planarians; but in somewhat hyper-
tonic modified Ringer solutions bipolar heads frequently appear on such
pieces (Child, unpublished). Since loss of motor co-ordination and more
or less complete paralysis occur in these solutions but regeneration is not
much inhibited, the decrease in dominance results in physiological isola-
tion of the posterior cut end to a degree permitting the activated cells
there to develop a head instead of a posterior end.

Extremely short planarian pieces often appear to be completely apolar.

RECONSTITUTIONAL PATTERNS IN EXPERIMENT 365

By contraction of the cut ends they become circular in dorsal or ventral
aspect, new tissue merely fills in the contracted cut surfaces without out-
growth, and definitely directed locomotion is absent (Fig. 122, H). In
consequence of contraction of the cut surfaces the longitudinal muscles
function much like the circular muscles in a medusa, their contraction
decreasing the circular diameter of the piece. Since the ventral muscles
are stronger, the central regions are forced dorsally when they contract,
and in time a permanent change in form with more or less elongated dorsal
outgrowth results (Fig. 122, /). These forms finally appear completely
radial in form and motor activity. It is sug-
gested that outgrowth of tissue is absent in
these pieces because the whole piece is in-
volved equally or almost equally in the
changes following section; that is, the
pieces are so short that a differential be-
tween regions adjoining the cut surfaces
and other parts of the piece is practically
absent, the whole piece adjoins a cut sur-
face, and consequently there is no differen-
tial developmental behavior of any kind
except that finally resulting from the radial
muscular activity.

Bipolar posterior ends may develop from short posterior pieces of cer-
tain triclad species that do not undergo fission (Morgan, 1904a; Lus,
1926). In these cases activation is not sufficient at either level of section
to determine dominance and independence of other parts, and reconstitu-
tion does not involve independent origin of a new pattern but develop-
ment of a subordinate part determined by other parts of the piece. Un-
doubtedly differences in physiological condition at different body-levels
may be concerned in determining whether bipolar heads or bipolar tails
develop, but lack of potency for head or tail development is not necessarily
involved. Bipolar heads or tails may be determined by a difference in
relation between the activated cells and other parts: if they become inde-
pendent, heads develop; if they remain subordinate, tails develop.

Planarian pieces from the region lateral to the ventral nerve cords
show increase in frequency of head regeneration from the longitudinal cut
surface on the median side of the piece rather than from the anterior
transverse surface, with decreasing length of piece. The head may regen-
erate on the angle between the two cut surfaces (Fig. 123, A); or in very
short pieces, so strongly contracted that transverse and longitudinal cut

Fig. 1 2T,. — Reconstitution of short
lateral planarian pieces (from Beyer
and Child, 1930).

366

PATTERNS AND PROBLEMS OF DEVELOPMENT

surfaces are indistinguishable, one or two heads may develop on the
whole surface of section (Fig. 123, 5, C). Apparently, with decrease in the

B

Fig. 124, A-C. — Planarian forms with multiple heads after partial posterior longitudinal
splitting. A, two heads from anterior regions of sides of split; B, diiplicitas cruciata; C, split
twice, the second split dividing each of the parts separated by the first; cruciate after first
split, multiple heads from longitudinal surfaces of section of lateral posterior regions after
second split {B, C, from Silber and Hamburger, 1939).

anteroposterior differential, as length of piece decreases, the mediolateral
differential plays an increasing part in localization of the head.^

5 Child, 1915c, p. 164; Olmsted, 1918; Beyer and Child, 1930.

RECONSTITUTIONAL PATTERNS IN EXPERIMENT 367

Certain cases of bipolarity and multipolarity in planarians are of special
interest. One or two heads occasionally appear at the anterior end of a
longitudinal split extending over most of the body length from the pos-
terior end (Fig. 124, A, B).^ Probably a certain degree of physiological
isolation is concerned in the development of these heads; such isolation
results when there is little direct nervous connection with more anterior
regions. In these cases the longitudinal split is between the nerve cords;
consequently, only cut transverse commissures can be present at the sur-
faces of section. If they are present, they may play a part in localizing
head development. As regards the new tissue developing on the longi-
tudinal cut surfaces, the heads are anterior, perhaps because gradient-
level and reactivity are higher there, though head development may occur
at more posterior levels of a longitudinal split, as will appear below.

Cruciate forms with a single head at the anterior end of the longitudinal
split (Fig. 124, B) and a variety of multipolar forms resulting from the
same operation have been described for Dugesia tigrina by Silber and
Hamburger (1939). One of these, resulting from a second longitudinal
split of each part separated by the first split, is shown in Figure 124, C.
This form became cruciate after the first split; but after the second, heads
regenerate from the longitudinal cut surfaces of the two lateral posterior
ends. Three distinct heads and five other pairs of eyespots in early stages
are present on these surfaces. The ventrodorsality of these heads coincides
with that of the body, and their bilateral form is evidently due to the
fact that each one represents a localized region of outgrowth with a ven-
trodorsal pattern. The nervous system in these lateral posterior ends must
be undergoing or have undergone extensive reorganization; but whether
parts of it are concerned in localizing the outgrowths which become heads
does not appear. In some species of planarians more readily than in others
a head may regenerate from a lateral region partly separated from the
body by an oblique section extending posteriorly and with its anterior
end removed to provide an anterior cut surface; from this the head de-
velops. Provided the longitudinal separation persists, an indefinite num-
ber of heads along the sides of the original body, with more or less induced
reorganization posterior to each, may be produced. This is essentially a
variation of the experiment of Figure 135, ^ (p. 400).

As regards establishment of new dominant regions, determination of

* Van Duyne, 1896; Morgan, 1900a; Goetsch, 1921, 1922, 1928; Keil, 1924; Beissenhirtz,
1928; Silber and Hamburger, 1939. _^

368 PATTERNS AND PROBLEMS OF DEVELOPMENT

new gradients and induction of reorganization, annelid, and planarian
reconstitution do not differ greatly. It was noted earlier that in most
annelids only a certain number of segments, characteristic for the species,
regenerates anteriorly after removal of a larger number, the other remain-
ing segments removed being formed by reorganization of old segments
induced by the new dominant region. A few annelids regenerate a large
number of segments anteriorly.^ Attention is called again to the case of
Sahella in which "abdominal" and ''thoracic" segments differ morpho-
logically. Regeneration from anterior ends at abdominal levels induces
reorganization of some abdominal into thoracic segments. This reorgani-
zation may also occur after section and before head regeneration. Also,
in some cases reorganization of segments anterior to a posterior section in
the abdominal region takes place ; this represents an approach to bipolar-
ity. Apparently the activation following section and isolation is sufhcient
to bring about induction without actual head regeneration.^ According
to Berrill and Mees, visible light increases greatly the number of segments
reorganized; in some way it apparently makes the dominance established
following section effective over a greater distance. The authors suggest
that the effect of light involves a photochemical reaction and release of
electrical energy rather than diffusion of a chertiical substance.

Regeneration of posterior ends in planarians and anneUds involves es-
tablishment of a growth gracUent opposite in direction to the primary
polar gradient. In short planarian pieces and apparently also in Lumhricu-
lus the posterior activation may inhibit head regeneration (pp. 183, 406).
After completion of posterior regeneration in planarians this gradient per-
sists chiefly at the extreme posterior end and is associated with growth in
length; but in the species with one or more posterior zooids these grow
in length more rapidly than the anterior zooid. In the annelids the pos-
terior gradient becomes a gradient of segment development and growth,
with high end posterior, but anterior to the anal segment, and may per-
sist throughout life in some species or disappear later in others.

Bipolar partial forms with heads at both ends develop in pieces from
anterior regions, and bipolar posterior forms in pieces from posterior re-

' For numbers of segments regenerated anteriorly and reorganization of old segments see
Ivanov, 1908; Hyman, 19160, and citations; also Morgulis, 1907; Korschelt, 1919; Martin,
1933; also Hyman, 1940, "Aspects of regeneration in annelids," Amer Nat., 74.

8 Berrill, 1931; Berrill and Mees, 1936a, h; Gross and Huxley, 1935. In his earlier paper
Berrill regards the regenerated head as organizer, but it seems evident from the later papers
that the activation following section, rather than the head itself, is the inducing factor.

RECONSTITUTIONAL PATTERNS IN EXPERIMENT 369

gions, of various annelids.'^ As in planarians, level of origin of piece, gra-
dient differential in it, and intensity of activation at the cut ends may be
the factors determining at least some of these bipolar forms rather than
specilic differences in potency for head or posterior development at differ-
ent levels.

Pieces of the ascidian Clavellina from the "esophageal" region may re-
constitute complete individuals or become bipolar partial forms by de-
velopment of siphons, branchial chamber, and other distal parts at both
ends (Brien, 1934). Stolons attached to the body develop new individuals
at their distal ends; stolon pieces isolated by section reconstitute individ-
uals from their proximal ends, probably because activation of cells at this
end by section and isolation is sufficiently intense to establish dominance
and a gradient which inhibits the less rapid bud development at the distal
end. The bipolarity of Clavellina pieces is certainly not due to restriction of
potency, for experiment has shown that proximal, as well as distal parts,
can develop from the proximal ends of pieces like those that become
bipolar, and there is evidently no limitation of potency in the stolon.
Brien's interpretation of his experimental data is in terms of gradient
relations.

The relative roles of reorganization of old parts and regeneration of
new tissue vary in different species and groups among the lower inverte-
brates; in some planarians they differ in definite ways with body-level of
origin of piece and with experimental conditions. In the higher inverte-
brates and in vertebrates, reconstitution in mature individuals is chiefly
or wholly by regeneration and is, in general, limited to regeneration of
subordinate parts, such as appendages, or in higher vertebrates almost en-
tirely to regeneration of a particular tissue by cells of that tissue.

Regeneration of legs of arthropods and amphibians evidently involves
origin of a new longitudinal axiate pattern. The regenerate is at first bud-
like in appearance; but, as regards gradient pattern, little is known. In
the regenerating legs of certain insect nymphs (Agrionidae) the tarsal
claws usually become distinguishable as definite, localized outgrowths in
very early stages, long before muscle attachments and articulations ap-
pear; that is, the first visible evidence of structural differentiation is at
the tip of the regenerate. Development of muscles, tendons, and articula-
tions, however, generally progresses from the proximal region distally;
and when regeneration begins in later nymph stages muscles and articula-
tions are often absent from the distal region, or the whole leg may have

'Morgan, 1902; Korschelt, 1904, 1927-31; Gates, 1927; von Haffncr, 1931.

370 PATTERNS AND PROBLEMS OF DEVELOPMENT

small tarsal claws but no muscles or articulations. These data suggest
an ectodermal gradient with high end at the tip, at least in earlier stages,
and mesodermal development progressing from the proximal region dis-
tally (Child and Young, 1903).

In the regenerating amphibian leg skeleton and skin develop, even
though they have been more or less completely removed from the proxi-
mal stump." The longitudinal half of the leg of Triton does not regener-
ate the half removed; but the distal part, regenerated from a transverse
surface of section of such a half-leg, may be complete and normal as
regards polarity and asymmetry." The distal portion of a leg may re-
generate from the proximal cut surface of part of a leg implanted by its
distal cut end (p. 390). Apparently the primary polar pattern of the re-
generate is not determined by the part from which it develops but origi-
nates in the budlike outgrowth. However, the mitotic index in earlier
stages of regeneration is highest near the base and shifts distally as re-
generation progresses (Litwiller, 1939). Also, structural differentiation,
practically all mesodermal, progresses from the proximal region distally.
Except for these data, nothing is known concerning gradient pattern in
the regenerating or the embryonic amphibian leg. Undoubtedly there is
a radial gradient system, decreasing from a center in the early stages;
but whether the high central region becomes the distal tip or remains
proximal or the high region becomes distal in the ectoderm and proximal
in the mesoderm, or whether the gradient pattern changes in the course of
original development and regeneration, remains to be determined. Early
stages of appendages in the chick embryo show dye reduction decreasing
radially from a center, but information concerning dye reduction in earlier
stages of the amphibian leg is lacking, partly because of pigmentation
and the difhculty of staining with the oxidized dyes.

In all these cases of reconstitutional development of a new axis in rela-
tion to section it originates either as a direct alteration of a pre-existing
axis by a new dominance and gradient, as in Tubularia and Corymorpha,
or as an outgrowth of new tissue much like a bud in that developmental
activity at first decreases more or less radially from a center and the
radial decrement becomes longitudinal by differential growth. In hydroids
the radial symmetry of the new axis may originate anew in the bud or
may develop in relation to the radial pattern of a stem piece. For ex-

" Weiss, 1925, 1927a; Bischler, 1926; Liosner, Woronzowa, und Kusmina, 1936; and cita-
tions by these authors.

"Weiss, 1925c, 19266. See also Graper, 1926a, b.

RECONSTITUTIONAL PATTERNS IN EXPERIMENT 371

ample, in Corymorpha pieces or aggregates radial pattern may develop
quite independently of the original radial pattern, but in stem pieces
isolated by transverse section the longitudinal entodermal canals of the
stem which are parts of the original radial pattern are factors in localizing
tentacle development.

SOME SPECIALLY SIGNIFICANT CASES OF NEW PATTERNS
IN RELATION TO SECTION

Under natural conditions the dominance of the apical hydranth of
Corymorpha prevents development of lateral hydranth buds, and the hy-
dranth-stem axis remains permanently unbranched. A simple lateral
transverse section part way through the Corymorpha stem closes in a few
hours but may become the locus of development of a new hydranth and
stem, provided the original hydranth is removed. In experiments involv-
ing several hundred animals the frequency of hydranth development from
a simple incision, with removal of the apical hydranth at the same time,
was 56.7 per cent at a level near the distal end, 26.7 per cent at a middle
level, zero at a proximal level; with the apical hydranth present and in
good condition it was zero at all levels (Child, 1927^, 1929a, 1932a). How-
ever, even a small lateral incision, with edges lacerated by repeated small
cuts so that heahng does not occur so rapidly and smoothly, becomes a
new hydranth-stem axis after healing in almost 100 per cent of the cases
if the original hydranth is removed (Fig. 125, A-C). Later an outgrowth,
and in some cases a differentiated basal region, may develop from the
opposite side of the original stem (Fig. 125, D). In these forms, as in
other Corymorpha reconstitutions, hydranth development is not from a
cut surface but from a region activated by the lacerated incision ; develop-
ment begins only after the closure of the incision. The activated region
determines a new gradient, at first radial, as in buds generally, and be-
coming longitudinal by differential growth, and acts as an inductor, some-
times being effective across the original stem in inducing development of
a basal region where there is no incision or other injury. Under slightly
inhibiting conditions the lacerated incision may give rise to a basal region
with perisarc and holdfast buds instead of a hydranth (Fig. 125, E).

New axes have been localized in a somewhat similar manner in the
elongated actinian Harenactis attenuata (Child, 1909c, 1910c). After re-
moval of muscles and mesenteries from the inside of short pieces isolated
by transverse sections from levels proximal to the esophagus contraction
brings distal and proximal cut surfaces of the body wall of the piece to-

372

PATTERNS AND PROBLEMS OF DEVELOPMENT

gether about the whole circumference, and union results, forming a closed
ring like the inner tube of an automobile tire. Usually these rings undergo
a rather remarkable orientation which brings the line of union of distal
and proximal ends onto the upper surface as the ring lies. Along or ad-
joining the band of new tissue uniting distal and proximal ends tentacles
and tentacle groups, varying in number and arrangement of tentacles,
develop from either distal or proximal side of the union or from both, and
in some cases there is approach to development of an actinian axis

^1^
//[\\

A

B

Fig. 125, A-E. — Determination of new axis by a lacerated lateral incision in Corymorpha.
A-C, incision, later outgrowth and early hydranth; D, induction by the new dominance of a
basal region from the opposite side of the original stem; E, development of a basal end with
perisarc and stolon buds from a lacerated incision under slightly inhibiting conditions (from
Child, 1927c).

(Fig. 126, A-C). Some of the tentacle groups are bilateral, others radial,
and still others irregular. Examination shows that the tentacles of these
groups are localized between the regenerating mesenteries and that the
form of the group depends on the positions of the mesenteries on the two
sides of the line of union and the directions in which particular mesenteries
happen to grow toward the line of union. When there is definite axiate
outgrowth, as in Figure 126, C, some of the mesenteries reconstitute with
the outgrowth and determine the number of tentacles at its apical end.
The groups appear to be localized where reconstitutional activity is
greater.

After local laceration of distal and proximal cut ends of the body wall

RECONSTITUTIONAL PATTERNS IN EXPERIMENT

373

by a number of longitudinal incisions close together and so placed that,
when contraction and distal-proximal union occur, distal and proximal
lacerations will come together, much more new tissue is formed in the
course of union, and in some cases a complete new disk with normal
number of tentacles and with mouth and esophagus develops from the
lacerated region, and development of more or less of the body follows

B

C D

Fig. 126, A-D. — Development of tentacle groups and new axiate patterns in "rings" of
Harenactis attemiata; further explanation in text (from Child, 1909c, 1910c).

(Fig. 126, D). The activation following laceration and union has deter-
mined a bud which becomes an entirely new polar axis and gives rise to
the distal parts of a normal individual. Here the old mesenteries are not
concerned, but the normal number of new mesenteries of a young indi-
vidual develops. It seems beyond question that the larger amount of new
tissue developing and probably a more intense and extensive activation
following the laceration are factors in determining the entirely new axis
of completely normal character as far proximally as it develops. Evi-

374 PATTERNS AND PROBLEMS OF DEVELOPMENT

dently the pattern of the new disk of Figure 126, Z), is not determined by
the original pattern but has arisen, Hke other bud patterns, from a region
of local activation involving both distal and proximal levels of the piece.
The tentacle-bearing outgrowths of Figure 126, C, apparently represent
less intense activation and less complete dominance; in them mouth and
esophagus do not appear. The signiticance of these cases is twofold : they
provide further evidence to show how completely independent of the old
pattern the new pattern may be, and must be for development of a new
polarity; and they agree with other data in indicating that differences in
intensity and probably in extent of activation may give different morpho-
logical results. They do not differ essentially from cases of reconstitution
of new dominant regions but are merely rather striking examples of es-
tablishment of new dominance by localized activation, with more or less
induction resulting, in Corymorpha, even across a pre-existing pattern
(Fig. 125, D).

Bud formation in Hydra, resulting from local cautery (Tepliakova,
1937), and determination of head development in a planarian by local
cautery, laceration, and application of certain chemical substances, as
reported by Goldsmith (1932, 1933, 1934, 1937, 1940), are similar cases
in which the local injury brings about sufhcient activation to initiate de-
velopment of a new dominant region and polar pattern.

QUESTIONS OF PATTERN AND ITS ALTERATIONS IN EARLY EMBRYONIC
RECONSTITUTIONS OF ISOLATED PARTS

Data of early embryonic reconstitutions are discussed in the following
chapters. Here it is desired to call attention to certain questions concern-
ing effects of section and isolation on egg and embryonic pattern, sug-
gested by the reconstitutions of isolated pieces of adult animals. First,
it may be noted that in isolated parts of eggs and isolated blastomeres or
blastomere groups there is no evidence of regeneration in the strict sense,
no outgrowth in relation to a surface of section. Whatever reconstitution
occurs is reorganization within the part. What happens to the pattern
present in the part when it reconstitutes a whole? We know even less
about what happens in these cases than we do concerning reconstitution
in an isolated piece of an adult hydroid or planarian, but at present there
seems to be no good reason for regarding embryonic and adult reconsti-
tution as essentially different. Gradient patterns have been demonstrated
in many eggs and embryos. These are unquestionably altered in one way
or another in reconstitution, as they are in adult reconstitution. Direc-

RECONSTITUTIONAL PATTERNS IN EXPERIMENT 375

tion of polar pattern is usually not altered in the early embryonic recon-
stitutions, as far as known; but there are cases in which it cannot be de-
termined whether it is altered or not. There is often considerable change
in scale of organization when half an egg or an isolated blastomere becomes
a complete individual. What happens to the pattern in these cases? Vari-
ous interpretations have been offered, in terms of hypothetical formative
substances, of substance gradients and their changes, etc. But if sub-
stances do change their positions in definite and orderly manner in these
cases, these changes must be determined by a pattern of some sort. What
is the character of that pattern? Does the reconstitution of an apical re-
gion on the basal half of an eight- or sixteen-cell sea-urchin embryo differ
fundamentally in any way from reconstitution of a hydranth in a piece
of Tubiilaria or Corymorpha or a head on a planarian piece? Does not
failure of the apical halves of the same stages to develop mesenchyme and
entoderm present the same problems as failure of isolated hydranths or
heads to reconstitute more proximal or posterior parts? With a large scale
of organization a short hydroid or planarian piece may develop only the
apical or anterior part of an individual. With decrease in scale it may
become an entire individual. The same is true for the apical half of the
sea-urchin embryo. Under natural conditions it forms only the more api-
cal portions of an individual. With experimental decrease in scale it may
become a complete individual (p. 357).

Ventrodorsality and dorsiventrality may be reconstituted in definite
relation to the pattern already present, though in many cases of blasto-
mere reconstitution it is not known whether or not this is the case. Sup-
pose, for example, we isolate the blastomeres of the two-cell stage of a
form in which ventrodorsal pattern of some sort is present at this stage but
the first cleavage plane has no definite relation to it. What happens when
either or both blastomeres develop as entire individuals? Is the entire
ventrodorsal pattern present on one side and reconstituted on the other,
or is only a part of it present and the rest reconstituted, or is a new ventro-
dorsality established on a smaller scale? If ventrodorsality is primarily
a gradient pattern, any one of these effects is possible. If ventrodorsaHty
or dorsiventrality has developed beyond the simple gradient stage, recon-
stitution of a complete individual may be possible only with isolations at
or near the median plane.

Embryonic reconstitution is narrowly limited or practically absent in
some forms, apparently in consequence of highly stable regional differ-
ences in the cytoplasm. Similar limitations appear in the adults of various

376 PATTERNS AND PROBLEMS OF DEVELOPMENT

species. It has often been noted, sometimes as something remarkable and
difficult to account for, that embryonic stages of certain forms appear
almost completely incapable of reconstitution, while adults of the same
species have high reconstitutional capacity. This is the case in many an-
nelids. The only difficulty here is the failure to recognize that the egg
may have a higher or more stable regional differentiation than different
levels of the adult body and that much of the egg differentiation may be
lost during development. In the annelid the embryonic differentiations
give rise mostly to larval parts and to head regions; it has never been
shown that the cells from which postcephalic regions develop possess any
such differentiation. In fact, there is evidence from their reconstitution
that they do not.

With the progress of experimental investigation and analysis it appears
increasingly evident that essential features of developmental pattern in
many eggs and early embryonic stages are fundamentally similar to fea-
tures of pattern in adults of many of the simpler organisms, as far as
spatial order and relations are concerned. If this is true, reconstitutional
development in adult material, being usually more accessible to experi-
mental analysis than that in parts of eggs and embryos and bringing us
nearer the beginnings of developmental pattern than the beginning of
embryonic development in many forms, should be regarded as an invalu-
able aid in throwing light on problems of embryonic development and
reconstitution.

A few data concerning modifications of embryonic pattern in relation
to section and isolation and involving points of special interest are briefly
discussed. Alteration by section of the polar pattern in the egg of the
sea urchin Lytechinus variegatus has been reported. Using position of
polar bodies and micropyle for identification of the original polar axis, it
was found that "animal" and "vegetal" halves, meridional halves, and
halves resulting from section in random directions of unfertilized eggs
may, after fertilization, reconstitute into wholes. Pieces down to 1/20 of
the total egg volume, irrespective of the region concerned, may become
normal blastulae with primary mesenchyme ; micromere formation is not
necessary for formation of primary mesenchyme. Section of unfertilized
eggs of Strongylocentrotus piirpuratus, S. franciscanus, and of the starfish
Patina miniata is followed after fertilization by first and second cleavages
in a plane vertical to the plane of section ; and when plane of section can
be identified at the stage of gastrulation, invagination occurs on the sur-
face of section. Moreover, in Lytechinus micromeres form on the surface

RECONSTITUTIONAL PATTERNS IN EXPERIMENT 377

of section." These results give little evidence of any pattern of localized
substances or regional cytoplasmic differentiation. Even polarity is al-
tered by section, according to these authors, the surface of section becom-
ing the basal region. With isolated pieces of unfertilized and fertilized
eggs of Arhacia somewhat different results are obtained by Harnly (1926).
Localization of micromere material between nucleus and center of egg is
indicated, and micromeres do not form on the surface of section. Accord-
ing to Plough (1927, 1929), development of isolated 1/2 blastomeres of
Arhacia, Echinarachnins, Echinus, and Paracentrotus indicates that micro-
mere material is localized at the time of first cleavage; but neither this
localization nor that postulated by Harnly agrees with the results of
Taylor, Tennent, and Whitaker. According to results obtained with other
species, polarity is not altered by section.'-'

These various results cannot, at present, be brought into agreement,
and it is not likely that all of them represent species differences. Perhaps,
however, if the polar pattern of the sea-urchin egg is primarily a quantita-
tive dynamic gradient system rather than a pattern of localized specific
substances, the different results may prove less difficult to reconcile. The
high temperature of the water in which Lytechinus lives (Tortugas) may
conceivably account for the apparent susceptibility of the polar pattern
to section. As regards development of micromeres from the surface of
section in Lytechinus, it has been found that in Strongylocentrotus and
Dendraster the micromeres appear at the low end of the primary polar
dye-reduction and susceptibihty gradients. They, together with prospec-
tive entoderm, apparently undergo an activation preceding gastrulation
(pp. 134-40). In the light of these data it appears possible that micro-
meres might form on the surface of section in consequence of a depression
resulting from the injury to the cell.

Sectioning sea-urchin eggs or early embryos in or near the frontal plane
results in reversal of ventrodorsality in the dorsal but not in the ventral
portion. Ligaturing in a meridional plane may determine ventrodorsality
vertical to the plane of ligature. With ligature tightly drawn, two op-
posed ventrodorsalities result, dorsal sides of both being toward the liga-
ture.'^ If the ventral side is the high end of the ventrodorsal gradient,
as it appears to be, there is apparently depression or inhibition in the plane

" Taylor and Tennent, 1926; Taylor, Tennent, and Whitaker, 1926; Tennent, Taylor, and
Whitaker, 1929.

'3 H5rstadius, 1936(7, b, 1937; Horstadius und Wolsky, 1936.

'•< Horstadius, 1936a, h, 1937, 1938; Horstadius und Wolsky, 1936.

378 PATTERNS AND PROBLEMS OF DEVELOPMENT

of isolation with reversal of ventrodorsality in the dorsal part, perhaps
in consequence of physiological isolation (p. 306). But only by further
experiment can we hope to discover what really happens in these cases.
It seems evident, however, that ventrodorsality is not fixedly determined
in these early stages but can be altered, as was maintained by Boveri
(1902).

INDUCTION OF RECONSTITUTION BY IMPLANTS IN MATURE ANIMALS

It was shown by Browne (1909) and confirmed by later workers that
even small pieces of Hydra from the region about the base of the tentacles,
implanted laterally, can persist and determine new apical regions and
polarities, largely from host tissue ; that is, the implant is dominant and
acts as an inductor. Similar small pieces from other regions of the body
are incorporated in the body wall or resorbed; but large pieces may per-
sist, reconstitute an apical region, and induce. Pieces of stalk, especially
those bearing the foot, tend to persist and grow without modification.
It is perhaps of interest in this connection to recall that dye reduction in
low oxygen occurs early in the- foot of the fully developed animal and that,
if the stalk contracts actively, it reduces more rapidly than the body
(p. lOl).

Autoplastic or homoplastic implantation of very small fragments, even
1/4 sectors of very short transverse stem pieces of the hydroid Corymor-
pha in small lateral incisions in the stem, may determine new hydranth-
stem axes, the graft itself forming only more or less of the apical region of
the hydranth and inducing other parts from the host tissue (Fig. 127,
A-C). Development of the new axis sometimes occurs when the apical
hydranth is present; but its frequency ranges from 35 to 75 per cent
higher, according to level of origin and of implantation, when the apical
hydranth is removed. Grafts from a distal donor-level, that is, from a
higher gradient-level, are more effective in inducing a new axis than those
from proximal levels, whether level of implantation is distal, middle, or
proximal (Child, 1929a, 1932a). In series of comparable experiments with
hosts and grafts and donors of approximately the same length and those
of each lot collected at the same time, pieces from a distal donor-level
gave a total frequency of new axes at distal, middle, and proximal host-
levels of 16.6 per cent with the apical hydranth present and in good con-
dition; pieces from a proximal donor-level, only 3 per cent with apical
hydranth present. With removal of the apical hydranth at the time of
implantation percentages were, respectively, 73.3 and 47.1. In short, the

RECONSTITUTIONAL PATTERNS IN EXPERIMENT 379

dominance of the host stem and the gradient of the donor stem are both
factors in determining effectiveness of the implant as inductor. In these
experiments the implants were 1/4 sectors of transverse pieces as short as
could be cut and were implanted without any attempt at orientation.
When isolated, such pieces usually develop into partial apical forms but

A

Fig. 127, A-F. — Determination by induction of new axiate patterns by small grafts in
Corymorpha. A-C, developmental stages, part developing from graft shaded; D, new induced
axis consisting of proximal tentacles and stem; E, hydranth-stem axis and outgrowth on oppo-
site side of host stem induced by graft of distal half of fully developed manubrium, reconstitut-
ed distal host hydranth inhibited and host stem reduced distal to graft; F, determination of
bases at both ends of a piece of host stem by dominance of a new axis induced by a small
graft {A-C from Child, 1932c;; F, after Child, 1935).

sometimes into complete individuals of extremely small size. When
grafted, they usually form only more or less of the apical part of the
manubrium. Rarely, however, development of the hydranth is so far in-
hibited that only proximal tentacles without manubrium develop (Fig.
127, D); in these cases the graft takes part in the tentacle development.
With more extreme inhibition only a stemlike outgrowth without hy-

38o PATTERNS AND PROBLEMS OF DEVELOPMENT

dranth may develop; and at proximal host-levels 10-25 per cent of grafts
from proximal donor-levels induce development of basal ends but can
also induce hydranth-stem axes. Since the grafts are not oriented when
implanted, but the new axis develops approximately at right angles to the
host axis, it appears highly probable that the original polarity of the graft
plays no essential part in determination. Apparently the graft and per-
haps host tissue about it act primarily as a group of active cells and, as
such, determine a new dominance and gradient extending into the host
tissues. Whether a complete hydranth with stem, proximal tentacles, and
stem, stem only, or basal end develops apparently depends on degree of
activation of the cells of the graft and surrounding tissue, as determined
by level of origin of graft, level of implantation, and inhibiting effect of
host dominance. Occasional cases of duplication of hydranth occur in
which the graft forms the apical part of one hydranth, the other being
formed entirely from host tissue. One case among several hundred has
been observed in which the graft lay between two new hydranths and
took no part in the formation of either. These duphcations are almost
completely limited to distal stem-levels, where new axes are determined
by simple partial section without implant more frequently than elsewhere
and are apparently cases in which sufficient activation of host tissues oc-
curred to determine one, or even two, new axes. In the latter single
case, the graft, though not inducing, apparently served as an obstacle to
closure of the wound and so brought about more activation and growth
of host tissues.

Fully developed hydranths and apical parts of hydranths, when iso-
lated, do not reconstitute proximal parts; but even the distal half of a
fully developed manubrium, implanted laterally, can induce the proximal
parts of a hydranth and a stem from host tissues (Fig. 127, £). Implanted
transverse pieces from the middle of fully developed manubria develop
hypostome from the distal end, or sometimes from both ends; and both
unipolar and bipolar hypostomes may induce a single proximal hydranth
region and stem (see also Beadle and Booth, 1938).

In many animals with a new axis developing laterally from graft or in-
cision, the reconstituting host hydranth is more or less inhibited, and the
host stem distal to the new axis becomes slender, as in Figure 127, E.
After it has developed, the new axis can completely inhibit development
of a host hydranth at more distal levels within a certain distance and can
even induce development of a base from both distal and proximal ends
of a piece of host stem isolated with it (Fig. 127, /^) ; or, if it is less effec-

RECONSTITUTIONAL PATTERNS IN EXPERIMENT 381

tive, it inhibits hydranth development to a greater or less degree at the
distal end of the piece (Child, 1935). Development of a base at the distal
end of the piece involves complete reversal of polarity; in those forms the
dye-reduction gradient is also reversed in direction. Development of a
base at the proximal end of the piece also gives evidence of the dominance
of the graft hydranth, for it occurs in much higher frequency than in con-
trol pieces without graft hydranth. A similar dominance of the hydra
bud has been repeatedly observed. If the parent body is cut off a short
distance distal to a bud or if a bud forms soon after section, reconstitution
is often inhibited, and the bud may become the apical region of the indi-
vidual (Weimer, 1938; Rulon and Child, 1937).

Reversal of polarity in short pieces of hydroids, planarians, and an-
nelids grafted in inverse position on longer pieces^^ has usually been re-
garded as resulting from an action of the longer on the shorter component.
In view of the evidence for independence of the dominant region it ap-
pears improbable that any such action of the longer component is involved
in development of an apical region or head on the basal or posterior end
of the short piece grafted to the longer. The development probably oc-
curs at the free end simply because the grafted end is not dominant. The
case does not appear to be essentially different from development of hy-
dranth or head on the proximal or posterior end of a completely isolated
piece that becomes bipolar. On the other hand, development of a basal
or posterior part from the distal or anterior end of a short piece grafted
to a longer in inverse position is undoubtedly determined by the domi-
nance of the longer component.

The work of Moretti (1911), Gebhardt (1926), and Goetsch (1929) has
given evidence of determination of new axes in planarians by grafts, and
more recent experiment has demonstrated conclusively the dominance and
inductive capacity of both homoplastic and heteroplastic grafts from the
region of the cephalic ganglia and from certain other regions."" Here only
a few of the varied results can be presented. Postcephalic parts do not
develop from isolated heads, even when entire ; and small pieces from the
ganghonic region with cut surfaces on all sides, such as have been used

'5 E.g., Peebles, igoo; King, 1901; L. V. Morgan, 1906; Ruttloff, 1908; Leypoldt, 1910;
Korschelt, 1927-31, 1929.

'^Santos, 1929, 1931, homoplastic, Dugesia { = Enplariaria) dorotocephala and D. tigrina,
heteroplastic between these species; Okada and Sugino, 1934, 1937, and Sugino, 1938, homo-
plastic, Planaria gonocephala, a Japanese species known by that name; J. A. Miller, 1938,
heteroplastic, D. dorotocephala, D. tigrina, and D. tigrina novangliae.

382 PATTERNS AND PROBLEMS OF DEVELOPMENT

for grafts, die when isolated; but when they are implanted in a planarian
body, other parts of the head may regenerate and induction of post-
cephalic regions in the host body may follow. With complete union of
dorsal and ventral epithelia of graft and host on all sides the graft may
induce a cylindrical, tubular outgrowth with eyespots variously situated
about the closed tip and with either dorsal or ventral epithelium inside
or outside; if the outgrowth is dorsal, ventral epithelium is inside; if
ventral, it is outside. Under certain conditions, when part of the cut sur-
faces of graft and host fail to unite, a tubular or funnel-hke outgrowth,
open at both ends, develops, eyespots may appear in it, and it may induce
postcephalic parts in the host. Also, when union between graft and host
is incomplete, a normal head may develop from the graft and act as in-
ductor. This is a common result with implantation at postcephalic levels.
Heteroplastic grafts of ganglionic pieces into the ganglionic region give
certain results of interest. Complete union with host usually results
with ganglionic grafts of D. dorotocephala into the ganglionic region of
D. tigrina. With grafts in normal orientation there is no outgrowth, little
development of new tissue, no evidence of induction, and locomotion and
reaction to food are normal within 12 hours after operation, though most or
all of the host ganglionic region was removed. In controls, with a window
in the ganglionic region similar to that in which the graft is implanted, nor-
mal reactions appear only after a week or more of reconstitution. Appar-
ently, the implanted ganglionic region of the one species attains domi-
nance very rapidly in the other. With reversed orientation of the hetero-
plastic grafts complete union is less frequent, tubular outgrowths with
eyespots may develop, some with short postcephalic induction, and a
secondary head may develop from a free cut surface of the host. These
forms often show further modifications resulting from partial distintegra-
tion or resorption of host head or graft outgrowth and may finally ap-
proach more or less closely a normal animal with single head. Appar-
ently either the graft head or the host head may dominate the other
component and inhibit it. Even with dorsi ventral reversal of hetero-
plastic grafts in the ganglionic region complete union may result. Such an-
imals advance with head raised and bent posteriorly. With incomplete un-
ion a graft head or combinations of graft and host tissue may develop. Ex-
tra eyespots frequently appear with ganglionic grafts in the ganglionic
region, even though there is no outgrowth, and with other than normal
orientation of the graft its polarity may be altered or reversed. These
results with ganglionic grafts in the ganglionic region are from Miller.

RECONSTITUTIONAL PATTERNS IN EXPERIMENT

383

Homoplastic and heteroplastic grafts of ganglionic pieces into the an-
terior prepharyngeal region (Santos, Okada, and Sugino) may induce
short postcephalic regions, the longest about equal in length to the dis-
tance between level of implantation and host head (Fig. 128, A-C).
Grafts at the pharyngeal level usually induce a prepharyngeal outgrowth
and a pharynx in the host body posterior to the outgrowth (Fig. 128,

U

Fig. 128, A-E. — Ganglionic grafts of Dugesia ( = Euplanaria): regeneration of graft and
induced reorganization of host. A, homoplastic, D. dorotocephala, in anterior prepharyngeal
region, has induced cylindrical outgrowth with dorsal epithelium on outer surface; B, hetero-
plastic, D. tigrina novangUae to D. dorotocephala, host decapitated twice; C, heteroplastic,
D. tigrina novangliae to pharyngeal region of D. dorotocephala with removal of host pharynx
with old pharynx remaining, induces prepharyngeal region and pharynx; E, heteroplastic,
D. tigrina novangliae to pharyngeal region of D. dorotocephala with removal of host pharynx
and head, induces prepharyngeal region and pharynx posterior to graft, and regenerating host
head induces another pharynx anterior to the graft (from Santos, 1929, 193 1).

D, E). Other examples of induction by heteroplastic grafts in the pharyn-
geal region, showing the induced modifications of the alimentary tract,
appear in Figure 129 from Miller's experiments. Details are given in the
legends. In A, B, and C of Figure 129 a pharynx is induced posterior to
the graft, and in B and C regeneration of a host head after later decapita-
tion anterior to the graft-level is completely inhibited; the two branches
of the alimentary tract at the original anterior end after this decapitation
in B indicate its reorganization into a posterior end. In Figure 129, D,
the graft has induced a pharynx posteriorly, another anteriorly with re-

B

Fig. 129, A-J. — Heteroplastic grafts in pharyngeal region. A, E, F, Dugesia iigrina to
D. dorotocephala; B, G-J, D. Iigrina novangliae to D. dorotocephala; C, D, D. doroiocephala to
D. Iigrina novangliae. A, DDAA* union complete, 42 days; tubular outgrowth developed
ventrally, covered with ventral epithelium and with two supernumerary eyespots, one shown
at S. B, DDAP, union complete, 35 days; original eyespots disappeared and four pigmented
areas developed, tail-like outgrowth from right side of host anterior to graft; host, decapitated
at X, failed to regenerate. C, DDAP, union complete, 42 days, ventral view; tubular out-
growth with terminal eyespots from large graft; host, decapitated about 3 mm. from graft,
regenerated only small amount of new tissue at A. D, DDAP, large graft union, complete, 28
days; pharynx with reversed polarity induced anterior to graft; extent of graft dominance
anteriorly indicated by anterior reorganization of digestive tract; after section at A^Y host
failed to regenerate. E, DDAA, large graft, 31 days, union complete, but graft given a free
posterior surface by section 7 days after implantation ; with development of head from posterior
side of graft its polarity is completely reversed. F, the animal of E at 74 days; the graft has
retained specific characters and has reversed polarity in a part of the host, as indicated by
ciliary movement, orientation of pharynx, and regeneration of a tail after fission from a level
originally anterior to the graft. G, DDAP, 14 days; anterior surface of graft failed to unite
with host; head (0) has developed from host portion of tubular outgrowth and small pharynx
has been induced anterior to the graft. //, the posterior part of the animal shown in G after
section at XX, with posterior end regenerated from the originally anterior surface of section.
/, DDAA, 21 days; left side of graft failed to unite with host and host lateral to graft tore
apart; modification of graft polarity indicated by position of auricles in relation to the eye-
spots. /, following decapitation of / at XX, tail and second pharynx were induced in host
anterior to the graft (from J. A. Miller, 1938).

* In this figure and in Fig. 131 orientation of the graft in relation to the host is given in the notation adopted
for amphibian limb grafts by Harrison (p. 390); "DDAA" means that dorsal and anterior sides of graft are toward
dorsal and anterior of ho.st, that is, the graft is normally oriented; "DDAP," graft in normal dorsiventral, reversed
anteroposterior orientation.

RECONSTITUTIONAL PATTERNS IN EXPERIMENT

385

versal of host polarity and reorganization of digestive tract. After re-
moval of the anterior region of the host at A' A', following development of
the reversed pharynx, head regeneration was inhibited. In Figure 129,
E and F, is shown a case of complete re-
versal of host polarity and also reversal
in the graft, because it was given a pos-
terior free surface after union, by remov-
al of regions posterior to it. In G and H
a pharynx is induced both posterior and
anterior to the graft, and later removal
of the host anterior end is followed by
development of a posterior end instead
of a head. In / polarity of the graft is
altered go° because its free surface, from
which the head develops, was originally
lateral, a pharynx is induced posteriorly,
and later section of the host at A' A' is
followed by induction of another pharynx
at a host-level originally anterior to the
graft-level and development of a posterior
end from the originally anterior cut sur-
face (/).

Ganglionic grafts in the posterior zooid
region very generally induce a postce-
phalic outgrowth and reorganization of
the host body posterior and for a certain
distance anterior to the graft. Two ex-
amples are shown in Figure 130, and
others with heteroplastic grafts and show-
ing reorganization of the digestive tract
in Figure 131. A and B show induction
in both directions, though the graft head
is far from normal in both; C is the poste-
rior product of fission following induction
by the graft both posteriorly and anteriorly, with development of posterior
end instead of head from the originally anterior surface of fission {A). In
this individual gradual fusion of the two posterior ends resulted in the con-
dition of Figure 131, /), with two pharynges at the same body-level and
with a duplicated postpharyngeal digestive tract. Later fissions in cases

Fig. 130, A, B. — Induction by gan-
glionic grafts in posterior zooid region.
A, homoplastic, Diigesia tigrina novan-
gliae; induced postcephalic tubular out-
growth and reorganization of host body
with development of pharynx posterior
and anterior to the graft-level. B, het-
eroplastic, D. tigrina novangliae to D.
dorolocephala; induction by graft of
postcephalic outgrowth and of reor-
ganization of host with pharynx de-
velopment posterior and anterior to
graft-level (from Santos, 1929, 1931).

386

PATTERNS AND PROBLEMS OF DEVELOPMENT

of this kind may show dupUcation of eyespots and of prepharyngeal, as
well as of postpharyngeal, regions. One case observed showed duplication
through sixteen successive fissions (Miller).

A point of considerable interest is that the induced reversal of polarity
in the host is a gradual progressive process, extending over several weeks.

A

D

Fig. 131, A-D. — Heteroplastic planarian ganglionic grafts in posterior zooid region. A, B,
Dugesia tigrina novangliae to D. dorotocephala; C, D, D. tigrina to D. dorotocephala. A, DDAP,
union complete, 21 days; outgrowth, reorganization of digestive tract and development of
pharynx posterior and anterior to graft; after section at .YA' tail instead of head developed on
piece containing graft. B, DDAP, union complete, 14 days; induction of reorganization and
pharynx development posterior and anterior to graft. C, DDAA, 133 days; induced reorgani-
zation and pharynx development posterior and anterior to graft. After fission at A , originally
anterior to the graft, a tail instead of a head reconstituted, and this part became a second
posterior end, as in the figure, and fusion between the two posterior ends A and P gradually
took place. D, same animal as C at 216 days, after fusion of the two posterior ends; only main
branches of digestive tract are shown (from J. A. Miller, 1938).

Section anterior to the graft a few days after it is implanted is usually
followed by regeneration of a normal head. Section at a later period may
result in a differentially inhibited head or complete inhibition of regenera-
tion, and after still later section a posterior end instead of a head de-
velops. Evidently there is a gradual extension anteriorly of the graft
dominance until it is blocked by the original dominance of the host. In
the light of other data this gradual extension of dominance suggests a

RECONSTITUTIONAL PATTERNS IN EXPERIMENT 387

gradual reorganization of the nervous system, progressing from, and in-
duced by, the dominant region of the graft. Doubtless a similar, though
less extreme, reorganization is induced posterior to the graft; but since
it is in the direction of the original polarity, its progress is not so readily
shown. According to Miller, the epithelium and pigment of heteroplastic
grafts are, in most cases, gradually replaced by epithehum and pigment
characteristic of the host. To what extent other tissues are replaced is
not known.

The extensive experiments of Okada and Sugino with a Japanese pla-
narian species show essentially the same relations of dominance and in-
duction by the grafts, but they have also shown that the head region an-
terior to the eyespots and pieces from the lateral head region, including
an auricle, are capable of dominance and induction and that postcephalic
pieces from the prepharyngeal region can induce a pharynx even without
head regeneration. Both the anterior and lateral pieces of the head include
some of the central nervous system, and the inducing capacity of pre-
pharyngeal pieces is what might be expected from the fact that isolated
prepharyngeal pieces that remain completely acephalic reconstitute all
parts posterior to their level of origin in the parent body (p. 180) but
nothing anterior to that level.

Working with European planarians, Br^ndsted (1939) finds that en-
tire, fully developed heads grafted on anterior cut surfaces of planarian
pieces do not induce reorganization. He maintains that the new tissue
of regeneration, the blastema, is the inducing agent. With the American
species, however, very little new tissue is formed in many cases in which
the grafts induce, and in the experiments of Santos and Miller no evident
relation between amount of new tissue and induction has been found. It
is entirely possible, however, that fully developed heads of some planarian
species are not active enough to induce reorganization.

MIRROR-IMAGING IN VARIOUS RECONSTITUTIONS

The two opposed components of bipolar forms of Tubularia, Corymor-
pha, planarians, and annelids are, in general, mirror images of each other
because, although their polarities are opposed, the radial, ventrodorsal,
or bilateral patterns are essentially the same. Under certain conditions
reconstitution may result in triplication of axes. The actinian Harenadis
affords an excellent example. A partial transverse section involving both
body wall on one side and one side of the esophagus results in union of
the cut surfaces of body wall and esophagus both distal and proximal to

388

PATTERNS AND PROBLEMS OF DEVELOPMENT

the level of section, leaving a lateral opening into the esophagus. As Fig-
ures 1^2, A and B, show, there is a complete disk and tentacles at the
distal end of the body (D), either the original or a reconstituted oral end.
Distal to the lateral opening a partial disk and partial circle of tentacles
are reconstituted with a polarity opposed to the original {D') , and proxi-
mal to the opening another partial disk and tentacles with polarity in
the original direction but, of course, with an alteration of the original
polarity {D"). Distal to the opening, one side of the animal is bipolar,
and the partial oral end D' is a mirror image of D" and also of the cor-

FiG. 132, yl-C— Mirror-imaging in reconstitution, A, B, in Harenactis attenuata (from
Child, 19096); C, schematic outline for Britchdreijachhildtingen; explanation in text.

responding part of D (Child, 19096). Similar reconstitutions can, of
course, be produced in other forms.

A large number of triphcations with similar mirror-imaging, mostly
appendages, have been described by Bateson (1894, chap, xx) and Przi-
bram (192 1 and earlier papers). Some of these have been produced ex-
perimentally by lateral partial section; others have been found as anoma-
lies. These triplications have been called Bruchdreifachbildimgen by Przi-
bram. He regards them as resulting from a lateral injury with reconstitu-
tion of distal parts from both distal and proximal surfaces of the injured
region, essentially like the case of Harenactis. Figure 132, C, indicates
diagrammatically the character of these triplications, the curvatures of
the triplicate distal parts representing the asymmetries. The mirror-imag-
ing is the same in D, D', and D" in Figure 132, ^, 5, and C, the only
difference between the Harenactis triplication and that of the appendage

RECONSTITUTIONAL PATTERNS IN EXPERIMENT 389

being that in the former D' and D" are partial symmetries developing
in relation to the anatomy of the original individual, while in the latter
the asymmetries of the reconstituted parts are complete, even though
they develop from surfaces of partial section.

Przibram interprets these triplicate forms in terms of a hypothetical
space lattice {Raumgitter) in the protoplasm. In terms of the gradient con-
cept, activation of the two surfaces by section or injury is sufficient to
determine a new polar gradient on each. With outgrowth of new tissue
each becomes the distal part of an appendage, the parts regenerating
being those normally distal to the level of the appendage from which re-
generation takes place, though in some cases not all of these develop.
Evidently the polarity of D' originates independently of the polarity of
the region from which it develops; and if the preceding analysis of recon-
stitution is not wholly mistaken, the polarity of D" is also independent
of that of the region from which it develops. The asymmetries of the re-
generates, D' and D" , however, are the same as the asymmetry of the
region from which they develop, except that their asymmetry may be
complete and normal, although they develop from a surface of partial
asymmetry. Apparently the partial asymmetry differential at the surface
from which regeneration takes place is sufficient to determine a complete
asymmetry in the new tissue.

The regenerated axes, D' and D" , do not usually develop parts proxi-
mal to the level from which they arise. In the case of D" these parts
are already present; as regards D' , if the high end of the gradient is distal
and development progresses proximally, it is balanced and its further
progress prevented by the similar opposed axis D; if development pro-
gresses distally from the surface of section or injury, development of distal
parts is probably determined by the activation adjoining that level; if
the high ectodermal region is distal and that of the mesoderm proximal,
both these factors may be concerned in determining the result. The scale
of organization of D' is often smaller than that of D, probably in conse-
quence of more or less inhibition of the new gradient by D. A similar
difference in scale appears very often in bipolar partial forms of hydroids
and planarians, the proximal hydranth or partial hydranth and the pos-
terior head being smaller and less fully developed and sometimes repre-
senting a smaller part of the polar axis than the distal hydranth or the
anterior head.

39© PATTERNS AND PROBLEMS OF DEVELOPMENT

ASYMMETRY AND MIRROR-IMAGING IN RECONSTITUTION
OF AMPHIBIAN LIMBS

A new limb polarity, opposite in direction to the original, may originate
in regeneration of amphibian limbs. A piece of limb including the knee
and part of the upper and lower leg implanted in a host body by its distal
cut end may show regeneration from its free proximal end, but this re-
generate gives rise not to parts normally proximal to the level of the prox-
imal cut end on which it appears but to the distal portion of a limb,
which may include all levels normally distal to the proximal cut end from
which it arises. Its polarity is opposite in direction to that of the original
piece and evidently independent of it, but the asymmetry is the same as
that of the proximal stump. Evidently the asymmetry provides a pattern
at the cut end, and this affects or persists in the regenerating tissue ; but
there is no such pattern for the polarity of the regenerated part, and this
evidently represents a new pattern originating in the bud of new tissue.
The regenerated part is a mirror image of the reversed piece except that
it develops distal parts absent from the piece.'' Similarly, pieces of am-
phibian tails implanted in the dorsal region by their posterior cut ends
may become bipolar by regeneration of a new posterior end from the free
anterior cut end (Milojevic et Burian, 1926). Here, also, the symmetry
of the regenerated posterior end is the same as that of the implanted
piece, but the polarities are opposed; consequently, there is mirror-imag-
ing, so far as parts are present in the implanted piece.

Development of limbs following transplantation of limb primordia in
amphibia presents a number of interesting problems and an indefinite
range of experimental possibilities; a great variety of experiment and an
extensive literature have resulted.'^ According to the terminology adopted
by Harrison (1921a), transplantation is orthotopic to the normal location
of the limb, on either side of the body, or heterotopic to some other loca-
tion— for example, the flank, homopleural to the same side of the body
as origin, heteropleural to the opposite side. As regards orientation of the
transplanted primordium, dorsodorsal position is with dorsal border dor-
sal, dorsoventral, with dorsal border ventral with respect to the host.
The anteroposterior axis of the transplant coincides in direction with that

'7 Delia Valle, 1911, 1913; Kurz, 1922a; Graper, 1922; Milojevic et Grbic, 1925.

'* Harrison, 19210, 19250, and Przibram, 1924, 1927, give numerous references; the general
review by Mangold, 1929a, gives a very complete bibliography. See also Swett, 1923, 1926,
1927, 1928a, b, c, 1930, 1932, 1937a, b, 1938a, b, c, 1939; Oka, 1934, and references cited by
them. Experimental data concern chiefly the forelimb.

RECONSTITUTIONAL PATTERNS IN EXPERIMENT 391

of the host when transplantation is homopleural dorsodorsal or hetero-
pleural dorsoventral and is reversed in homopleural dorsoventral and
heteropleural dorsodorsal. These relations give four positions of trans-
plant with respect to host axes in orthotopic location, as indicated in Fig-
ure 133, and four similar positions in lateral heterotopic location.

Experimental results with Amhly stoma punctatum led Harrison to for-
mulate three rules, of which the essential points are as follows: (i) rever-
sal in relation to the host of the anteroposterior axis of the limb bud re-
sults in a limb with asymmetry of the side opposite that on which the
bud was implanted (disharmonic) ; (2) when the anteroposterior axis is
not reversed, the asymmetry of the resulting limb is that proper to the

HOM.
D-D

HOM.
D-V

HET.
D-D

HET.
D-V

Fig. 133. — Diagram of positions of amphibian limb-bud transplants in relation to host
axes. A, anterior; P, posterior; D, dorsal; V, ventral, of host outside, of limb bud inside circle;
R, right; L, left; Horn. D-D, homopleural dorsidorsal position, etc. (from Harrison, 19210).

side on which it is located (harmonic) ; (3) the original member (the first
to begin development) of limb duplications has an asymmetry according
to (i) or (2), and the secondary member is a mirror image. In later ex-
periment these rules have been found to hold very generally, few real or
apparent exceptions having been observed.

Duplication, with mirror-imaging of the two limbs, is frequent, and in
some cases there is triplication, in which case each limb is usually a mir-
ror image of the one next it; that is, there are two planes of mirror-imag-
ing, as indicated in Figure 134, but certain cases have been observed in
which this relation apparently does not hold (Swett, 1924; Oka, 1934).
In orthotopic location duplications are more frequent when axial relations
between bud and host are disharmonic; in heterotopic locations they are
more frequent when relations are harmonic; but little is known concerning
the physiological factors that bring about duplications and triplications.

As regards the asymmetry relations, Harrison holds that anteropos-

392

PATTERNS AND PROBLEMS OF DEVELOPMENT

teriority is definitively established in the primordium or bud at the time
of transplantation but that dorsiventrality is not yet established or is re-
versible until later stages, the dorsiventrality of the transplant being de-

np2(u)

MP,(R)

P.DU

Fig. 134, /I -C— Diagrams of limb reduplication. .4, PR, primary limb; P. DU, posterior
reduplicating member; A. DU, anterior reduplicating member; MPi{R), primary (radial)
mirror plane; MPz{U), secondary (ulnar) mirror plane; 1-4, digits; 5, level of section shown
in B and C. B, C, sectional diagrams of reduplication; in B mirror planes radial (MP.) and
ulnar (MP.); in C they are radiodorsal (MP.) and ulnopalmar (MP^). D, dorsal; PAL, palmar;
R, radial; U, ulnar (from Harrison, 1921a).

termined by the dorsiventrality of the host. When double or triple limbs
develop, the mirror-imaging results, according to Harrison, from develop-
ment of more than one growth center within the range of mutual influ-
ence. The center developing first, or most advantageously placed, re-
verses the asymmetry of the other or others. Localization of growth cen-

RECONSTITUTIONAL PATTERNS IN EXPERIMENT 393

ters beyond the range of mutual influence may result in duplication with-
out mirror-imaging.

A very different interpretation has been advanced by Przibram (1924,
1925, 1927). His view is that axiate pattern is definitively established in
the primordium at the time of transplantation, that apparent reversal of
asymmetry is not due to an influence of the host body or pattern but to
development from the proximal, instead of the distal, part of the trans-
plant. With persistence of the original asymmetry such a limb will be a
mirror image of a limb developing from the distal region of the limb bud,
just as a proximal hydranth or a posterior head or an isolated hydroid or
planarian piece is a mirror image of that at the distal or anterior end.
Such a limb would correspond to D' of Figure 132, C. Duplication would
result from development of a limb both distally and proximally from the
transplant, that is, from bipolar development with original asymmetry
in both polar axes. Triplication of hmbs Przibram interprets in terms of
the Bruchdreifachbildungen (p. 388), as limb development from distal and
proximal regions of the transplant and also from the limb bud region of
the host. It may be noted that Delia Valle (191 1, 1913) obtained the
axial relations of Figure 132, C, in regeneration from a partial transverse
section of the amphibian limb.

As in so many biological controversies concerning interpretations, it
seems entirely possible that both views may be correct, that is, that
dupHcations and triplications of hmbs and their asymmetries may occur
in both ways and perhaps also a new center of limb development, that
is, a region of activation may sometimes originate in a region of less com-
plete union of transplant and host or of more intense irritation by the
operation. There are plenty of cases in which such factors do initiate
axiate development. It is well known that parts of limb primordia sec-
tioned in various directions may give rise to whole limbs. '"^ Limb-bud
mesoderm isolated from ectoderm and implanted with reversal of the
mediolateral axis, the longitudinal axis of the limb, results in development
of whole limbs which accord with the rules (Harrison, 1925a). Redupli-
cated limbs developing in heteroplastic transplants between Amhly stoma
species always have donor characteristics, according to Swett (1932). On
the other hand, Oka (1934) maintains that in Hynohius some cases of
duplication represent distal and proximal regeneration from the trans-
plant and that some triplications develop entirely from the transplant.

'"Harrison, iQ2ia; Swett, 1924, 1Q26, 1928; Graper, 1926.

394 PATTERNS AND PROBLEMS OF DEVELOPMENT

Harrison regards the asymmetry of the Hmb as an expression of an
"intimate structure" and suggests an analogy to compounds composed of
asymmetric molecules. Przibram (192 1) postulates a space lattice which
is assumed to reverse its asymmetry as required by experimental data.
At present, however, there is no evidence of the existence of such an
oriented intimate structure or space lattice as the basis of axiate pattern;
and the mechanism of reversal of such a structure in a limb-bud trans-
plant by some effect of the host pattern, or in relation to another bud,
requires further hypothesis. Interpretation in dynamic terms seems to
offer less difficulty. There is an anteroposterior and a dorsiventral differ-
ential in susceptibility and in rate of dye reduction in the amphibian em-
bryo at stages corresponding to those used in limb transplantation. Pig-
mentation has made it impossible with the material at hand to demon-
strate a gradient in rate of dye reduction in the limb bud, but in the chick
limb bud such a gradient is clearly evident (Rulon, 1935). The limb pri-
mordium must share in the gradient pattern of the body, but the dorsi-
ventral differential in it, though apparently present in gastrula stages
(Swett, 1938c, 1939), can be altered by change in orientation with respect
to the host pattern up to much later stages. The dorsiventrality of a small
transplant is alterable up to a later stage than that of a large transplant
(Lovell, 1937). Larger limb-disk transplants of Triturus are more likely
to develop than smaller. With transplantation to the mid-ventral line
duplications are usually ulnar and in a plane transverse to, but on the
flank are usually radial and in the plane of, the longitudinal body axis.
Inclusion in the transplant of parts about the limb primordium, particu-
larly the region dorsal to it, results in better limb development (Takaya,
1938). Here, as in Hydra and other invertebrates, size of transplant ap-
pears to be a factor in its persistence, and the body pattern is apparently
a factor in determining not only plane of duplication but direction of
radio-ulnar asymmetry.

The frequent rotations of limb-bud transplants in other than normal
orientation noted by various authors (e.g., Nicholas, 1924) suggest pres-
ence of effective dynamic factors. If such factors can determine rotation
in many cases, it is not difficult to conceive that they may determine a
physiological axiate pattern in a transplant that does not rotate. More-
over, since a dynamic factor of some sort is apparently required to bring
about reversal of an intimate structure or space lattice, it is at least a
pertinent question whether it is necessary to postulate the structure. As
regards duplication and mirror-imaging, the effects of two dynamic fields

RECONSTITUTIONAL PATTERNS IN EXPERIMENT 395

on each other are perhaps of interest. These effects have been pointed
out by various authors in attempts at analysis of the mitotic figure. For
example, if two like magnetic poles are brought near together, the field
of each is altered so that the two become mirror images of each other. If,
then, a third like pole is brought near on one side, so that it acts equally
on both fields, both acquire, so to speak, a dorsiventrality but are still
mirror images of each other. Similarly, two equal fields of fluid, each
flowing radially from a center, when near together alter each other so that
each becomes a mirror image of the other. A current flowing at right
angles to the line joining the two centers will determine a new asymmetry
in both fields, but they wifl still be mirror images. In both the magnetic
and the fluid fields a morphological asymmetry and mirror-imaging is es-
tabhshed and persists as long as the dynamic factors act; or if the mag-
netic field is made visible by iron particles, the pattern may persist after
removal of the poles. These examples of field action and resulting pattern
are offered merely as suggestions, not as models.

Apparently reversal of asymmetry in one limb primordium by another
may depend on relative positions of the two. With transplantation in
certain orientations to a position anterior to the host limb bud, the asym-
metry of the transplant is reversed; with transplantation in normal orien-
tation posterior to the host limb bud, asymmetry of the host limb is re-
versed. In these combinations the posterior member always develops more
rapidly and the anterior member is reversed, apparently by the domi-
nance of the other (Takaya, 1936). These experiments suggest that posi-
tion in the anteroposterior body gradient may influence rate of develop-
ment and that developmental activity, rather than a structural pattern,
is the essential factor in influence of one hmb bud on another. In this
connection mirror-imaging of sea-urchin larvae developing from blasto-
meres of the two-cell stage in contact after separation (Driesch, 1906)
and the cases of situs inversus and mirror-imaging in twinning and double
monsters are of interest, as probably involving much the same factors as
amphibian limbs. In hydroid reconstitutions two hydranths developing
close together often show more or less inhibition on the facing sides and
so are mirror images.

REGENERATION OF THE LENS IN AMPHIBIA

The regeneration of the lens from the dorsal region of the iris in various
amphibia has presented a problem of considerable interest which has led

396 PATTERNS AND PROBLEMS OF DEVELOPMENT

to repeated experiment and various interpretations.^" This regeneration
occurs not only after the lens is removed from the eye but after it is
displaced so that direct continuity with the iris is lost and after lens de-
generation by X- radiation (Politzer). The fact that the origin of the re-
generated lens is different from its embryonic origin and the apparent
limitation of potency to the upper or dorsal part of the iris led Wolff to
a teleological interpretation which was opposed by Fischel, but many
investigators have concerned themselves chiefly or wholly with presenta-
tion of data. Sato (1930), transplanting pieces of the upper iris to another
eye, found a gradient of decreasing lens potency from the most dorsal iris
region. In further experiments he implanted optic primordia of different
embryonic stages with dorsiventral reversal, so that the originally dorsal
side of the optic cup was ventral; and at a later stage shortly before
metamorphosis, he removed the lens (Sato, 1933a, b). According to the
stage of dorsiventral reversal, the choroid fissure developed ventral in
position, dorsal in origin, or two fissures developed, or the original dorsi-
ventrality persisted. Apparently, dorsiventrality of the optic primordium
is not irreversibly fixed in earlier stages ; in later stages the original dorsi-
ventrality apparently determines one fissure, the imposed dorsiventrality
another; still later, the fissure develops in its normal position with respect
to the original dorsiventrality. The lens regenerates from the iris region
opposite the choroid fissure or, when two fissures are present, opposite
the more completely developed ; and a gradient of lens potency decreasing
from the iris region opposite the fissure is present. Sato concludes that the
choroid fissure exercises an inhibiting action on lens regeneration, so that
the lens arises from the iris region farthest from it.

Sato's data, as well as those of others, suggest a more general inter-
pretation, that is, that both position of choroid fissure and region of lens
regeneration are expressions of dorsiventral pattern, evidently a dorsiven-
tral gradient, either the original gradient or that imposed on the im-
planted, dorsiventrally reversed eye. The high region of the dorsiventral

" Colucci, 1891; Wolff, 1895, 1901, 1913; E. Miiller, 1896; Fischel, 1900a, 1902; Spemann,
1905; Wachs, 1914, 1920; Ogawa, 1921; Beckwith, 1927; Adelmann, 1928; Sato, 19330, b,
1935; Toro, 1932, Kesselyak, 1936; Politzer, 1936, Monroy, 1939. Lens regeneration from the
iris has been observed in adults of certain urodeles; but in Amblystoma punctatum and A. ti-
grinum and in the anurans Rana clamitans, R. pipiens, and R. sylvatica it does not occur even
in larval stages; in these forms, however, lens fragments implanted or remaining after lens
removal may regenerate lens (Stone and Sapir, 1940, "Experimental studies on the regenera-
tion of the lens," Jour. Exp. ZooL, 85).

RECONSTITUTIONAL PATTERNS IN EXPERIMENT 397

gradient in vertebrates is dorsal, and the developing eye may be expected
to possess a dorsiventral differential representing a part of this gradient.
According to this suggestion, lens regeneration occurs normally from the
dorsal iris margin, and lens potency decreases ventrally from this region,
because the dorsal region represents the highest gradient-level present in
the eye and reacts most rapidly to removal of the lens, and probably also
exercises some degree of dominance over lower levels. In earlier stages
the dorsiventral differential is reversible, but perhaps becomes specific
later, though it is possible that even in the adult the difference is one of
susceptibility or reactivity. Tests of sectors of the iris of reversed eyes
for lens potency show that the potency of the dorsal region (originally
ventral) is not very high and not widely different from that of other
parts. This suggests that the original dorsiventral gradient and that im-
posed secondarily by reversal partially neutralize or obliterate each other,
and Sato's experiments in general indicate that either may become the
effective one. That a dorsiventral differential is involved in lens regenera-
tion is further suggested by the relation of orientation of the lens fibers
to the dorsiventrality of the optic cup (Dragomirow, 1930).

Interpreted in these terms, lens regeneration does not appear to differ
fundamentally from the decreasing capacity for head regeneration from
higher to lower levels of the polar gradient in various planarians and an-
nelids and the disappearance of the capacity at certain levels in some
forms. In all these cases it still remains to be determined whether the
apparent decrease and disappearance of potency at lower gradient-levels
results from specific differentiation or from a difference in susceptibility
or intensity of reaction to the experimental conditions. The reversal of
dorsiventrality of the eye implanted in dorsiventrally reversed position
is paralleled by the reversal of dorsiventrality in similarly reversed, trans-
planted limb buds.

DOMINANCE AS AN INHIBITING FACTOR

In both plants and animals a dominant region tends to inhibit develop-
ment of another similar region within a certain distance of itself. Various
instances of this effect of dominance have already been discussed in other
connections. Here a few further cases are considered in which dominance
either prevents a certain kind of reconstitution, inhibits it to a greater
or less extent, or brings about regression or destruction of parts already
morphologically distinguishable.

398 PATTERNS AND PROBLEMS OF DEVELOPMENT

INDUCED REGRESSION IN CERTAIN HYDROIDS

Removal of the distal half of the hydranth primordium of Tubularia
(see Fig. 13 [p. 36]) in its earlier stages may lead to complete regression
of the proximal part and development of a new complete primordium in
the usual relation to the end of the piece. In this case the gradient is
altered by the activation following the second section, and a new hydranth
pattern is determined, partly in the cells forming the proximal part of
the first primordium, partly in cells farther proximal. The new pattern
obHterates that which had already begun to develop. Results of such op-
erations, however, differ with level of section and stage of development
of the first primordium. If section is near the distal end, only more or less
reorganization of the distal region of the primordium may occur; or if
the primordium is advanced in development, regeneration of the part re-
moved is the usual result. Regression of the primordium sometimes re-
sults in both Tubularia and Corymorpha from section immediately proxi-
mal to it in early stages or in its proximal region. Disappearance of the
original primordium is followed by development of a shorter one in the
distal region of the piece and, in some cases, of another at the proximal
end. In these cases activation following section at the proximal end of
the early hydranth tends to establish a new dominance and reverse the
gradient pattern over more or less of its length; and the reversal obliterates
the former pattern, even though it has become morphologically visible,
and determines a new pattern at one or both ends.^'

REGRESSION OF FISSION ZONES

Fission may be inhibited and the fission zone obliterated in planarians
by section and regeneration of a head a short distance anterior to the
zone. Here there is no directly visible regression, but the region which
represented the anterior part of a posterior zooid is reorganized into a
prepharyngeal region. If or when fission occurs in such an individual after
reconstitution, it is at a level much farther posterior. In some species of
Stenostomum section and reconstitution of a head a short distance anterior
to a fission zone already visible, but in early developmental stage, will
bring about regression of the zone and reorganization of the region con-
cerned into a part of the body of the reconstituted individual. Similar
regression and reorganization of early fission zones has been brought about
in the same way in several species of microdrilous oligochetes by various
experimenters. A section anterior to the zone is much more effective than
posterior section (E. H. Harper, 1904).

" Driesch, 1897, 19026; Peebles, 1900; Child, n)ogd.

RECONSTITUTIONAL PATTERNS IN EXPERIMENT 399

INHIBITION OF RECONSTITUTION AND DESTRUCTION OF ZOOIDS
AS A RESULT OF ALTERED DOMINANCE

According to Stevens (1902), pieces from the basal stem region of
Antennularia ramosa usually reconstitute hydranth-stem axes; those from
more distal levels, stolons. This difference appears to be due to different
relations of dominance at the different levels. At the proximal levels hy-
dranths are often absent from the lateral branches or in poor condition,
while farther distally they are present and in better condition, the more
distal branches being physiologically younger and their dominance more
effective in inhibiting development of hydranth-stem axes than that at
more proximal levels.

In planarian reconstitution a developing head region may, under cer-
tain conditions, inhibit development of another head. For example, in
the well-known experiment involving decapitation and partial anterior
longitudinal spHtting of the body, a head develops on the anterior end
of each separated part, even if the two anterior ends represent different
body-levels (Fig. it,5, A, B). However, if a half-transverse section is also
made at the posterior end of the longitudinal split, thus removing one
of the two separated parts, head development is inhibited on the half
anterior cut end resulting from this section (Fig. 135, C). Apparently, in
this case the more posterior level of section on the one side is dominated
and determined as lateral, instead of as head, by the dominant region
at a more anterior level on the other side. In the case of B this dominance
is not effective in inhibiting head development, probably because it cannot
overbalance the gradient already established and intensified by the activa-
tion following section in the shorter anterior end. This experiment has
been used repeatedly by the writer in laboratory class work (see also
Goldsmith, 1939). The conditions in forms like Figure 135, A and B,
are comparable to those in bipolar forms of hydroids and planarians;
neither dominant region affects the other because the two opposed gra-
dients are more or less nearly equal. In the case of Figure 135, C, one
of the opposing gradients has been removed, and head development is
inhibited. It has been shown earlier in this chapter that planarian head
reconstitution can also be inhibited by grafts.

A striking case of the effect of altered dominance appears in the inhibi-
tion, disintegration into cells, and complete destruction of zooids already
morphologically distinguishable in certain species of the rhabdocoel
Stenostomum (Child, 1903&; Van Cleave, 1929). In Figure 136, A, a chain
of S. grande sectioned at A' A' leaves the posterior part of zooid i.i. and

400 PATTERNS AND PROBLEMS OF DEVELOPMENT

the whole of zooid 1.2. with head in early developmental stage, anterior
to zooid 2.7, with more advanced head. Although the anterior end of the
posterior part of zooid i.i. is capable, under other conditions, of head
development, it is completely inhibited, and a gradual reduction and sepa-

A B C

Fig. 135, y4-C.— Planarian head regeneration at different levels, following decapitation
and anterior partial longitudinal splitting; explanation in text.

ration into cells, first of the part of zooid i.i., then of zooid 1.2., results
(Fig. 136, B-E). The cells from these zooids accumulate in the pseudocoel
and gut of the dominant zooid 2.1. and in zooid 2.2. posterior to it and
gradually disappear, apparently serving as nutritive material, for these
zooids grow rapidly. That the head region alone of an older zooid can
bring about this destruction of headless and younger complete zooids

//.

/2.

22.

2.1.

2.2.

2.2.

A D E

Fig. 136, A-E. — Stenostomiim grande. Reduction and destruction of zooids resulting from
alteration of dominance; zooids and fission zones numbered according to origin and order of
appearance. A, chain of four zooids, sectioned at XX; B-E, stages in reduction and destruc-
tion of headless part of zooid /./. and of zooid 1.2. (from Child, 1903&).

402 PATTERNS AND PROBLEMS OF DEVELOPMENT

anterior to it is evident from Figure 137. In Figure 138, a chain of
S. tenuicauda, there is even more extensive destruction. After section in
the anterior zooid (A'A' of Fig. 138, A) the posterior part of this zooid
(i.i.i.i.) and the next two zooids (1.1.1.2. and 1.1.2.) are progressively
reduced and destroyed by zooid 1.2.1. (Fig. 138, B-F), and this zooid
finally remains as a single individual, fission having occurred posterior
to it at fission zone I. (Fig. 138, A). This reduction and destruction of
zooids occur only when their heads are in early stages or have been re-
moved. Sometimes the headless part of an anterior zooid will reconsti-
tute a head rapidly enough to prevent its destruction, though it may
undergo some reduction before the head region attains a sufficient domi-
nance to prevent it. The farther anterior the level of section in the an-
terior zooid, the more frequently does head reconstitution result; that
is, the higher the gradient-level of section, the more rapidly does head
reconstitution take place and the less likely is obliteration of the zooid
gradient and destruction of the zooid to occur. On the other hand, an
already visible reconstituting head at a cut end may finally be inhibited
and destroyed like the developing head regions of younger zooids.

In the normal development of the Stenostomum chain each new zooid
originates from the posterior region of a zooid already present; conse-
quently, in all cases a more advanced head region is anterior to one less
advanced. The older anterior component of a pair is evidently able to
prevent extension of dominance anteriorly from the younger head pos-
terior to it. When a headless zooid is anterior to a head, or an earlier
zooid anterior to a more advanced head region, without an older head
region still farther anterior, the dominance of the posterior head evidently
extends anteriorly and obliterates or perhaps reverses the gradient, and
reduction and disintegration follow.

Whether a zooid is able to maintain itself or is destroyed in these ex-
periments probably depends chiefly on the developmental stage of its
cephalic ganglia and the extent to which they have attained connection
with and have brought about reorganization in, the nervous system pos-
terior to them. Until a certain degree of nervous dominance over pos-
terior regions is attained by a zooid, it may be destroyed by a more ad-
vanced head region posterior to it, provided there is not a more advanced
head region at a still more anterior level. If these suggestions are correct,
it follows that the dominance of an older head region must still be more
or less effective in a zooid developing from its posterior region, at least
until the head region or the ganglia of this zooid have attained a certain

///.

MJ.

112.

1.2.

I.

2.1

D

Fig. 137, yl-Z>.— Reduction and destruction of zooids by a more advanced head region in
Stenostomum grande. A, portion of chain between XX and X'X' used; B-D, stages of destruc-
tion; cellular material from destroyed parts in pseudocoel indicated by dots (from Child,
19036).

B

n.il-

121

A D E F

Fig. 138, A-F. — Destruction of zooids in Stenoslonnim lenuicauda. A , section at A'A', fission
later at zone /., still later at zone 111.2.; B-F, stages of destruction (from Child, 1903/)).

RECONSTITUTIONAL PATTERNS IN EXPERIMENT 405

stage of development. And if this is the case, the physiological isolation
making possible development of a new zooid is not complete isolation.
That it is not complete is further indicated by the fact that contraction re-
sulting from stimulation of an advanced head may extend posteriorly be-
yond an early fission zone but may be blocked at an advanced zone.

In 5. leucops head development from an anterior cut end is very fre-
quently completely inhibited, apparently by the rapid development of a
fission zone and head at a more posterior level. This development may
have begun before section but not have become visible, or may be initiated
by isolation of the piece and, being more rapid than reconstitution at the
anterior cut end, inhibits the latter (Van Cleave, 1929). No case of reduc-
tion and destruction of zooids or of head reconstitution from anterior cut
ends was observed in more than a hundred pieces of a Japanese species
resembling S. leucops, cut in various relations to visible fission zones; but
new fission zones developed very rapidly in the pieces, and their domi-
nance apparently inhibited reconstitution at the anterior cut end. These
species differences suggest that if head reconstitution at a cut end is slow,
as compared with head development at a fission zone, the latter may
inhibit the former.

Destruction of zooids and inhibition of head reconstitution at a cut
end in an animal in which head development occurs rapidly at a fission
zone are of considerable interest as effects of obliteration or reversal of a
gradient by change in position of the dominant region. In S. grande
this reversal determines not only inhibition of development but complete
loss of the epithelial character of the body layers and their separation
into isolated cells. In S. leucops and the Japanese species the reversal of
dominance has less extreme effects; it inhibits head reconstitution with-
out destruction of zooids. The question may be raised whether the pres-
ence of a gradient is necessary for maintenance of the epithelial character
of the body layers in S. grande. The isolated cells are not killed but have
evidently lost pre-existing relations to each other. In this connection cer-
tain other suggestive cases may be recalled. When the gradient of the
axis of the alga Griffithsia is obliterated by differential inhibition with
chemical agents, the cells separate but are not necessarily killed, and with
return to normal environment may reconstitute new axes (Child). With
decrease or obliteration of the gradient in the blastula of Phialidiiim loss
of epithelial character and immigration into the blastocoel of cells from
all parts of the blastula wall, instead of from the basal region only, occur
(p. 167). The entoderm of echinoderm gastrulae and exogastrulae shows

4o6 PATTERNS AND PROBLEMS OF DEVELOPMENT

a similar loss of epithelial character with obHteration of gradient by dif-
ferentially inhibiting conditions (chap. vii). However, complete reversal
of gradient and dominance is possible in Corymorpha and many other
coelenterates and in planarians without isolation of cells. Even if it should
be found that a gradient is necessary for the origin of epithelial order, its
persistence is evidently not always essential for maintenance of epithehal
character.

EFFECTS OF OTHER PARTS ON DOMINANT REGIONS

Although a new dominant region originates independently of other
parts in many plants and animals, its scale of organization and in some
forms its morphological pattern may be affected by other parts. In recon-
stitution this effect is usually inhibition.

In pieces of Tubularia or Corymorpha stem below a certain length, but
still two or three times the length of the hydranth primordium developing
on longer pieces with distal end at the same level, the primordium length
decreases with decrease in length of piece (Driesch, 1899; Child, i907e).
Driesch regarded this decrease as indicating a teleological relation between
length of piece and primordium length. Actually, however, the primor-
dium length decreases less rapidly than length of piece, so that, with suffi-
cient decrease in length of piece, only a hydranth develops; and in still
shorter pieces, apical partial forms (Fig. 113, A-I [p. 334])- Primordium
length does not adjust to length of piece, either with the aid of an entel-
echy, as Driesch maintained, or in any other way; but the activation fol-
lowing section at the proximal end of the piece, even though not sufficient
to determine a hydranth there, does determine, at least temporarily, a
gradient, as dye reduction shows; and in pieces below a certain length this
opposes the distal gradient and decreases scale of organization of the distal
primordium somewhat. That this is the case is readily shown by delaying
proximal section for different periods after distal section. Driesch's teleo-
logical assumptions are not only unnecessary but fail entirely to account
for the apical partial forms.

Certain planarian species show an inhibiting action on head develop-
ment originating from the posterior end in pieces below a certain length
and increasing in effectiveness with decrease in length of piece at a given
body-level. In the discussion of differential inhibition of the planarian
head (pp. 177-90) it was pointed out that this physiological inhibiting
factor originates at the posterior cut end; that it is effective only below a
certain length of piece, which differs in a definite graded manner with

RECONSTITUTIONAL PATTERNS IN EXPERIMENT 407

body-level and physiological condition of animals; and that section of the
nerve is almost as effective in inhibiting the head as section across the
whole body. Apparently the inhibiting factor is, at least in large part,
nervous in character, a stimulation resulting from section of the nerve
cords at the posterior end of the piece (Watanabe, 19356) and probably
from effects upon them of the activated tissue there. Since active tissue
and nervous system are apparently the chief factors in dominance in the
mature planarian and the inhibiting factor also appears to be nervous,
the conclusion seems justified that it acts by decreasing the independence
of the cells concerned in head formation.

At first glance it may seem improbable that a nervous stimulus from
more posterior levels can inhibit head regeneration, but a brief analysis
will show the grounds for this conclusion. It is evident that the cells at
any body-level are determined and continue to function as that particular
body-level only as long as they represent a certain relative gradient-level
and are in physiological relation with other parts, chiefly the dominant
region. This determination is reversible; and when they are isolated from
dominance, every body-level can become a dominant region and give rise
to a head and so determine reorganization of other parts. In other words,
every body-level will develop cephalic ganglia and a head when freed from
dominance and relations with other parts. Head regeneration is a self-
differentiation, and nervous tissue is apparently the primary differentia-
tion in head development. In fact, planarian parenchyma cells in tissue
culture give rise to long fibers and become indistinguishable from cultures
of ganglionic cells (Murray, 193 1). Nervous stimuli from other parts of
the body than the dominant head region probably play some part in
maintaining the characteristics of a particular body-level. When a given
level is isolated by section from more anterior levels, the cells adjoining
the level of section tend to lose their characteristics as cells of a particular
level, undergo a change in condition, and become a new dominant region.
But if another section is made posterior to the first, but within a certain
distance from it, the stimuli resulting from section of the nerve cords and
the activation of cells adjoining the posterior cut end tend to maintain
the characteristics of the body-level represented by the cells at the an-
terior end of the piece. They represent factors in the physiological rela-
tions of these cells with other parts and, while not sufficient to prevent
head formation in the absence of posterior section, are evidently intensi-
fied by such section and become effective in partly or completely inhibiting
head regeneration because their effect is to keep the cells functional parts

4o8 PATTERNS AND PROBLEMS OF DEVELOPMENT

of the body. These stimuli apparently undergo a decrement in effective-
ness with transmission anteriorly; consequently, posterior section beyond
a certain distance from the anterior section has no effect on head forma-
tion. However, the decrement may be only apparent and is perhaps ac-
tually a block rather than a decrement. In pieces from anterior body-
levels in which activation and development at the anterior end of a piece
is more rapid than farther posteriorly, posterior section must be at a very
short distance from the anterior end to be at all effective in inhibiting
head development. At more posterior levels of the anterior zooid it is
effective over a greater distance, perhaps because less effectively blocked
by the less intense anterior activation. This relation to body-level and
length of piece is shown in Figure 66 (p. 182) ; it is evidently an expression
of differences in condition at different gradient-levels.

Demonstrative evidence that the physiological factor inhibiting head
regeneration is associated with effects of posterior section is provided by
the results of different periods of delay of either anterior or posterior sec-
tion (Child and Watanabe, 1935a). Pieces of Dugesia dorotocephala, in-
cluding the region XP (Fig. 139, A), develop 100 per cent normal heads;
but pieces XY , with anterior ends at the same level but only 1/8 or less
of the body length of animals 16-20 mm. long, are almost or quite 100
per cent acephalic. If anterior section at X is made first, the posterior
section at Y later, head frequency increases with increasing delay of the
posterior section (Fig. 139, B). Even i or 2 hours' delay is usually suffi-
cient to increase head frequency. With 12 hours' delay of posterior section
head frequency increases from almost complete acephaly to almost 100
per cent normal, and with 24 hours delay it is 100 per cent normal.

For delay of anterior section pieces AY are cut, and later pieces XY
are taken at intervals from successive lots. Figure 140 shows the change
in head frequency with delay of anterior section up to 96 hours. There is
usually a slight decrease in frequency with 12-24 hours' delay. Appar-
ently not merely the posterior cut surface but the cell activity following
section plays a part in inhibiting head regeneration. Comparison of Fig-
ures 139 and 140 shows that with delay of anterior section the inhibiting
effect of posterior section persists over a much longer period than with
delay of posterior section.

Inhibition of head development in relation to length of piece and body-
level occurs in different degree in different planarian species." Even in

" Cf . Child, 1913c, 1914J; SiviCkis, 1923; Buchanan, 1933; Watanabe, 19356; Child and
Watanabe, 1935a; Abeloos, 1930.

RECONSTITUTIONAL PATTERNS IN EXPERIMENT

409

extremely short pieces of some species posterior section has httle or no
effect; in others there is inhibition, but less than in D. dorotocephala.
The species differences probably result chiefly from differences in rate or
intensity of anterior activation following section, in relation to the effec-
tiveness of the inhibiting factor.

Fig. 139, A, B. — Effect of delay of posterior section on planarian head frequency {Dugesia
dorotocephala). A, outline indicating levels of section; B, graph of increase in head frequency
with delay of posterior section; ordinates, head-frequency indices (see Appendi.x VII), abscissae,
hours of delay; curves ah and cd, twenty-five pieces for each period of delay, from data obtained
by Child with different stocks; ef, from combined data obtained by students in laboratory
class work, more than seven hundred pieces. Irregularities are chiefly due to differences in
length of pieces (from Child and Watanabe, 1935a).

Head regeneration in the Dendrocoelidae and some other forms does
not occur posterior to a certain body-level, irrespective of posterior sec-
tion and length of piece, probably because cells of more posterior gradient-
levels do not react sufficiently to isolation by section to develop a head.
They are capable, however, of giving rise to posterior ends.

4IO

PATTERNS AND PROBLEMS OF DEVELOPMENT

This inhibition of planarian head development by posterior section,
Hke inhibition of head development and destruction of zooids in Stenosto-
mum, represents an interference by a region posterior to the level of re-
constitution with a dominance already present or with the attainment of
the independence necessary for head formation. The decrease in length

100

Fig. 140. — Increase in head frequency with delay of anterior section {Dugesia doroto-
cephala). Ordinates, head-frequency indices (see Appendix VII); abscissae, hours; curve ab
from data obtained by Child with fifty pieces for each period of delay; curve cd from com-
bined data of students, including more than six hundred pieces (from Child and Watanabe,
1935a)-

of hydranth primordium in Tuhularia with decrease in length of piece
evidently represents a similar, but less conspicuous and effective, inter-
ference. Moreover, Hyman (1916a) has shown somewhat similar relations
in Lumhriculus.

In the reconstitution of a new dominant region from other body-levels
than the apical or anterior region there is, so to speak, a conflict between
the factors concerned in the reaction initiating reconstitution and those
tending toward maintenance of the original condition. The new domi-

RECONSTITUTIONAL PATTERNS IN EXPERIMENT 411

nant region may be said to develop in spite of the isolated piece. The
reconstitution of a subordinate part, such as a posterior end, apparently
involves no such conflict. Nerve stimuli not only do not inhibit it but
may be necessary for its occurrence, probably by their effect on metabolic
level of the cells concerned. Local and temporary dominance may exist
in the new posterior end, but it does not induce extensive reorganization
in more anterior regions. Regeneration of appendages apparently re-
sembles more closely posterior reconstitution than it does head reconsti-
tution. There may be local dominance in the regenerating tissue, either
distally or proximally, but it does not induce any considerable reorganiza-
tion in the parts from which regeneration takes place; for example, the
regenerating amphibian hmb has little effect on the limb stump, except
adjoining the level of section.

We have seen that new dominant regions can develop in hydroids,
planarians, and annelids in the absence of other parts; and there is at
present no evidence that other parts of the body, when present, contribute
to the completeness of their development; they tend, rather, to inhibit
it, if they have any effect. In the later course of development, more par-
ticularly in the higher animals, relations of parts may change, local pat-
terns of dominance and subordination may arise, and parts originally
developing independently of certain other parts may later be dependent
on, or affected by, those parts in one way or another. For example, it has
been shown by many investigators that in vertebrates absence or addition
by implantation of peripheral parts in embryonic or later stages may
bring about hypoplasia or hyperplasia in some part or parts of the central
nervous system.--' Such effects, of course, have nothing to do with pri-
mary developmental pattern; but they indicate that peripheral parts may
acquire, after they appear, some degree of dominance over the further
development of parts previously independent of them.

DOMINANCE IN COMPENSATORY REVERSALS OF ASYMMETRY

An experimentally reversible dominance of an appendage on one side
of the body over an appendage of the same segment on the other side
has been found in certain annelids and arthropods. The serpulid poly-
chete Hydroides dianthus possesses two opercula, one highly developed
and functional in closing the tube when the animal withdraws into it,
the other rudimentary. Removal of the functional operculum is followed

^•J Shorey, 1909; Detwiler, 1921, 1923, 1926, 1927, 1936, particularly chap, viii and citations
given there; Hamburger, 1934, 1939; Kappers, 1934; May, 1927, 1933.

412 PATTERNS AND PROBLEMS OF DEVELOPMENT

by development of the rudimentary into a functional operculum, and the
originally functional one regenerates as rudimentary. When only the rudi-
mentary operculum is removed, it regenerates as rudimentary. When
both are removed, both may develop the functional form or reversal, or
persistence of the original asymmetry may result. Also, when the anterior
part of the body is removed in the thoracic region, two opercula of func-
tional type develop in the anterior regeneration (Zeleny, 1902, 1905a,
1911).

Many decapod Crustacea show heterochely; that is, one chela is large,
the other small; in some species only in one sex; or the two chelae differ
in form and function, being distinguished as crushing and pinching or as
snapping and cutting. In different species the asymmetry ranges from a
high degree of uniformity as regards right-left position of the two types
to approximate equality of positions. In some forms asymmetry is not
reversed by removal of either claw. In others removal of the large claw
is followed by development of the small into a large and by regeneration
of a small in place of the original large one. In still others, like the lobster,
with two claws differing in type, reversal of asymmetry does not occur in
the adult but does in the young animal (Emmel, 1908); and when both
claws are removed, even in the adult, both may regenerate as crushers
(Emmel, 1906). Experimental reversal of asymmetry apparently has no
relation to the constancy of a particular asymmetry in the species. There
is no experimental reversal in some forms with approximate equality of
right-left positions, or it appears only in young animals.^'*

In the forms with experimentally reversible asymmetry the subordi-
nate member attains dominance when the other is removed; but if both
are removed at the same time, both may regenerate as dominant mem-
bers, or an asymmetry may result. The experimental results are much
like those with hydroid and planarian pieces. Unipolarity or bipolarity
results according to the difference or similarity in condition at the two
ends. So with the appendages asymmetry represents a difference in con-
dition on the two sides of the segment concerned, and experiment shows
that one side is dominant, as in unipolar forms; but if both sides begin
development at the same time, both may develop as dominant but with-
out inhibiting each other, like bipolar forms.

How the dominance of one appendage becomes effective on the other
is not certainly known. Some evidence was obtained by E. B. Wilson

2^ For further data sec Przibram, 1Q02, 1905, 1907, 1908, 1918; E. B. Wilson, i903(;;
Morgan, igo4C, 1924; Zeleny, 1905; Emmel, 1906, 1908.

RECONSTITUTIONAL PATTERNS IN EXPERIMENT 413

that section of one or both nerves might abolish the dominance in the
crustacean Alpheus; but because of rapid nerve regeneration, autotomy
of appendages following nerve section, and high death rate it was not
regarded as demonstrative. It seems not improbable that afferent im-
pulses from one appendage may inhibit efferent impulses to the other and
so prevent its complete development, but with an equal start in regenera-
tion neither may be able to inhibit the other. Perhaps the most interesting
question in connection with these asymmetries is that of their origin in
egg or embryo.

EXPERIMENTAL DETERMINATION OF PATTERN AND DOMINANCE
BY ENVIRONMENTAL FACTORS

The differential effects on pattern already present of general uniform
exposure to various chemical and physical environments were discussed
in chapters v-vii. The present section is chiefly concerned with experi-
mental determination and alteration of pattern by differential exposure
to environmental factors.

POLARITY IN RELATION TO DIFFERENTIALS IN OXYGEN TENSION

Many hydroids are extremely susceptible to low oxygen tension in sea
water. It has been shown that change from flowing to standing sea water
brings about degeneration and death of hydranths and reconstitution of
stolons instead of hydranths from apical as well as other regions in a num-
ber of hydroid species (pp. 172-75). Stolons can live and grow at much
lower oxygen tension than hydranths. With other species, less susceptible
to low oxygen in water, low concentrations of cyanide and various other
inhibiting agents determine reconstitution of stolons instead of hydranths.
Plumularia pieces in very low concentrations of cyanide and also in stand-
ing water reconstitute nothing but stolons from both distal and proximal
cut ends and from cut ends of lateral branches; but in flowing, well-aerated
water hydranth-stem axes develop from all cut ends. According to Barth
(19386), decrease in oxygen tension below 4.5 cc. per liter decreases mark-
edly the rate of hydranth reconstitution in Tubularia, and an increase
above that level increases rate and size of primordium (see also Torrey,
191 2). Local removal of perisarc from pieces of Tubularia stem may
bring about reconstitution of a single hydranth or of two with opposed
polarities from the exposed coenosarc (Zwilling, 1939). In long pieces of
Tuhularia stems with proximal ends exposed to high oxygen tension, distal
ends to boiled water, nitrogen, CO2 or a mixture of 90 per cent oxygen,

414 PATTERNS AND PROBLEMS OF DEVELOPMENT

or lo per cent CO2 all hydranths developed at the proximal ends. With
short pieces results are similar but "less striking." With the same oxygen
tension at both ends circulation of water at one end determined hydranth
development there (J. A. Miller, 1937, 1939). Corymorpha resembles
Tubularia as regards susceptibility to low oxygen. The hydranths soon
die in standing water. When individuals in good condition are stained
with methylene blue and placed in standing water, oxygen tension soon
becomes so low about the basal regions of the crowded tentacles and
about the manubrium that the dye is rapidly reduced there in well-
aerated sea water with large surface open to air. In rapidly flowing water
the dye is not reduced. Various attempts have been made to subject the
two ends of Corymorpha pieces to different oxygen tensions; but by means
of contraction, extension, and changes in diameter the naked stems usually
make their way out of the opening in the partition between the two sea
waters, or, if the opening fits tightly about the stem, separation usually
results. However, other experiments with this species make it highly
probable that an oxygen differential can determine which of two polarities
shall develop or can actually determine polarity.

Corymorpha pieces only a few millimeters long show relatively little
motility during the earlier stages of reconstitution. Such pieces lying un-
disturbed on the bottom of a container are freely exposed to water on
one end or side, while the other is more or less closely in contact with the
underlying surface and diffusion is more or less interfered with there.
That this is the case as regards oxygen is readily shown by the rapid dye
reduction on the surface in contact of pieces stained with methylene blue
or Janus green and a gradient of decreasing rate of reduction from the
surface in contact toward the free surface.

In 73 per cent of two hundred Corymorpha pieces 5-10 mm. long, sup-
ported on thin gauze or loose absorbent cotton near the surface of the
water and lying on their sides so that both ends were equally exposed,
hydranths developed on both ends. Of a like number of pieces 5 mm.
long with one end, either distal or proximal, in contact with the bottom
of the container, 30 per cent developed hydranths at both ends, and most
of these were pieces that fell over one side in consequence of contractions
and extensions. In another experiment hydranths developed at both ends
in 66 per cent of fifty similar pieces frequently moved about and turned
over by water currents and reversed individually. A similar lot undis-
turbed with one end in contact gave 22 per cent bipolar hydranths, most

RECONSTITUTIONAL PATTERNS IN EXPERIMENT 415

of these also in pieces that failed to stand continuously on end.^* In these
experiments many of the unipolar forms finally developed basal ends from
the ends determined as proximal by the differential exposure, whether
originally proximal or distal.

In Corymorpha pieces lying on the side, particularly in bipolar pieces,
a basal region with perisarc and stolon buds frequently develops quite
independently of section from the side in contact (Fig. 1^1, A, B); or two

Fig. 141, A-D. — Development of bases in bipolar forms of Corymorpha (from Child,
19266).

basal ends sometimes develop, producing forms of cruciate type (Fig. 141,
C). The fact that these basal ends developing from lateral regions in con-
tact are localized midway between the two hydranths of bipolar forms if
these are approximately equal suggests that the two hydranth-stem gra-
dients determine this region as the lowest gradient-level in the piece and
that the differential between side in contact and free side plays a part in
determining the basal region laterally instead of about the whole circum-

^5 The naked stems of large Corymorpha individuals are several millimeters in diameter
when contracted, and short pieces usually remain more or less contracted for some hours after
section. Reconstitution is so rapid that the hydranth primordium usually becomes directly
visible within 24-30 hours; consequently, this length of time is usually sufficient to determine
position of the hydranth or hydranths. Some of these experiments were reported in an earlier
paper (Child, 19 266). For further discussion of hydroid reconstitution in relation to o.xygen
see Barth, 1940, "The process of regeneration in hydroids," Biol. Rev., 15.

4i6 PATTERNS AND PROBLEMS OF DEVELOPMENT

ference. In short pieces lying undisturbed on one side with httle move-
ment, a hydranth frequently develops from the upper, and a base from
the lower, side of one or both ends (Fig. 141, D). Here the new gradient
and axiate pattern are determined across one or both ends of the piece
by the differential between upper and lower surfaces.

In all these experiments with contact-free-surface differential it is pos-
sible that accumulation of CO2 at the surface in contact, as well as low
oxygen, is concerned in determining the developmental pattern. Increase
in hydrogen-ion concentration in consequence of CO2 accumulation is
probably not sufficient to have any appreciable effect. In the light of the
other experiments it appears highly probable that the oxygen differential
is the chief, if not the only, determining factor.

Exposure to various inhibiting agents — ethyl ether, ethyl alcohol, ethyl
urethane, chloretone, and HCl-sea water (CO2?) — preceding or following
section or both, results, after return to water, in great increase in fre-
quency of determination of new axes by the differential between upper
and lower sides and in frequency of multiple polarities from a piece
(Child, 19270, h). Little or no movement occurs in such pieces after return
to water until development is more or less advanced, and with sufficient
exposure to the agent they often become more or less flattened on the
glass or lose their characteristic structure to some extent. In Figure 142,
A and B, traces of the longitudinal entodermal canals remain and make
it evident that the new polarities are at right angles to the old. In ^ a
hydranth develops from the upper, a base from the lower, side of each
end. In Figure 142 {B, early, and C, later, stage of the same individual)
the hydranth develops from the upper, the base from the lower, side,
and reconstitution from the ends does not take place. In D the piece lost
its structure and became an almost hemispherical mass, and the new
axes arose as buds from its upper free surface, the lower side becoming a
base and secreting perisarc. In E both hydranths and bases develop from
the upper surface, and a large basal area from the lower. The positions
of the two bases on the upper side suggest that they are determined by
the dominance of some or all of the hydranths. The larger base develops
between the larger hydranths, much as in many bipolar forms, and the
smaller appears below the small hydranth.

The new multiple hydranth-stem axes developing after decrease or
obliteration of the original gradient by differential inhibition are ad-
ventitious as regards position, that is, their localization has no definite
or constant relation to a pre-existing pattern. Their general localization

RECONSTITUTIONAL PATTERNS IN EXPERIMENT 417

on the upper surface is evidently determined by the contact-free-surface
differential. Localization of any one of the multiple axes appears to be
fortuitous but probably results from localization of regions of greater
cell activity after return to water. With disappearance of the entodermal

Fig. 142, A~I. — Determination of new pattern in Corymorpha pieces by the contact-free-
surface differential after decrease or obliteration of the original polarity by differential in-
hibition. A-C, ethyl alcohol 2 per cent for 48 hr. following section, development in water; D,
ethyl ether 0.25 per cent for 26 hr. preceding section, five new a.xes from upper side after return
to water; E, chloretone 0.02 per cent for 26 hr. following section; F-I, dorsiventrality deter-
mined by the contact-free-surface differential after slight inhibition (from Child, 1927(2).

canals their cells tend to accumulate irregularly in small groups, and
there is some evidence that these cell groups just beneath the ectoderm
may play a part in localizing the new axes.

Short pieces lying on one side with httle or no movement or change in
position in well-aerated standing water and without preceding inhibition

4i8 PATTERNS AND PROBLEMS OF DEVELOPMENT

usually develop hydranths at one or both ends; but after temporary ex-
posure to slightly inhibiting conditions that retard reconstitution but do
not obHterate the original polarity or the proximal end as factors in deter-
mination, the hydranths often show a dorsiventrality in relation to the
contact-free-surface differential. Tentacles may show a gradient in length
and rate of development decreasing toward the side in contact, and one
or more bases may develop from the side of the hydranth or hydranths
in contact (Fig. 142, F-I).

It is sufficiently evident that development of bases or stolons, either in
Corymorpha or in many other hydroids, is not a specific reaction to con-
tact, for these parts develop under the dominance of apical regions or
under somewhat inhibiting external conditions quite independently of con-
tact. In the experiments described above, the contact is apparently mere-
ly a factor in determining low-gradient level.

Aggregation of cells into masses, following dissociation by pressing
through bolting-cloth or otherwise, and development of individuals and
partial forms from the aggregates have been observed in numerous sponge
species and in several hydroids.'^ How developmental pattern originates
in these aggregates is of considerable interest. They result from chance
contact of cells and may differ greatly in size ; and the larger aggregates
may be cut into smaller ones, as desired. If an inherent persistent polarity
is present in the cells, orientation to each other or to some external factor
would have to be assumed to account for polarity of the whole. Actually,
however, the sponge cells chiefly concerned in the development of aggre-
gates do not appear to have a definite polarity, and the polarities of hy-
droid cells do not coincide with the polarity of the whole. If the cells
orient to form epitheha, their orientation must be according to their
polarities rather than according to a superpolarity of the whole; and if
they orient to some external factor, the same difficulty arises. If polarity
is primarily a gradient imposed on the aggregate, there are two possibili-
ties of origin: it may be determined by an external differential — for ex-
ample, an oxygen differential — or perhaps in some cases, at least in the
hydroids, by a chance group of cells with higher metabolism than others,
such as cells from the distal region of a Corymorpha stem. In sponge ag-
gregates canals develop in relation to a region of greater thickness between

^' For aggregation and development from dissociated cells of sponges see H. V. Wilson,
1907, 1911a, b; K. Miiller, 191 1; Huxley, 1921a, b; Galtsoff, 1925; De Laubenfels, 1932. De-
velopment of sponges from aggregates has also been observed by the writer in a consider-
able number of species (unpubhshed). For hydroids see H. V. Wilson, iqiib; C. W. Hargitt,
1915; Child, 1928c.

RECONSTITUTIONAL PATTERNS IN EXPERIMENT 419

surface in contact and free surface as a center, and the osculum forms in
this region on the free surface. Gravity is not concerned; probably the
contact-free-surface oxygen differential is the chief factor, but demonstra-
tion seems to be lacking. Swimming sponge larvae in certain developmen-
tal stages aggregate when brought into contact, and sponges may de-
velop from the aggregates with osculum on the free surface and a pattern
without any conceivable relation to the pattern of individual larvae
(H. V. Wilson, 1907).

Experiments with Corymorpha aggregates, consisting of cells from many
individuals, indicate that the contact-free-surface differential is the chief
factor in determining the polarity of the resulting individual or partial
form. Aggregates remaining undisturbed with one region continuously in
contact usually (79 per cent in characteristic lots) give rise to unipolar
complete individuals on a small scale of organization. Aggregates of ap-
proximately the same size, moved about and turned over repeatedly at
somewhat irregular intervals of one to several hours, usually (86 per
cent) develop only apical parts of the manubrium on a larger scale but
are mostly unipolar. Probably polarity is determined by the chfferential
exposure in intervals between change of position. Experimental attempts
to determine the time necessary for determination of polarity by the con-
tact-free-surface differential have not yet been performed, but it is cer-
tainly not long. In the aggregates moved about at intervals the changes
in position permit development on a much larger scale of organization;
consequently, apical partial forms are the characteristic result (Child,
1928c). Aggregates kept free from continued contact by water currents
or by continued change of position with respect to gravity by slow revolu-
tion on a vertical wheel in closed containers usually remain spherical, do
not develop at all, and finally, after several days, become completely in-
closed in perisarcal secretion and remain without further change, except
decrease in size of the living tissue inside the perisarc, until death. How-
ever, some aggregates develop partial or complete axes, even when kept
entirely free from continued contact on the revolving wheel, probably
because of the presence in some region or regions of the aggregate of a
higher level of cell activity, perhaps a few cells from distal stem-levels.
A few such cells near together by chance may initiate a new dominance
and gradient, and, if more than one such group is present, more than one
axis may develop. At present there is no evidence of determination of the
polarity of an aggregate by an inherent persistent polarity of its cells and
their orientation to each other.

420 PATTERNS AND PROBLEMS OF DEVELOPMENT

In most of the experiments described thus far in which the oxygen
differential is certainly or probably a factor in determining where a hy-
dranth or hydranths shall develop, it is primarily a regionally selective
factor. In the pieces with transverse cut ends it determines which of two
possible polarities shall develop; the activation following section and the
resulting dominance and gradient do not depend on the oxygen differen-
tial, but they are the factors directly concerned in determining the axiate
pattern of the hydranth that develops. The oxygen differential determines
that this dominance and gradient at one end shall be adequate for hy-
dranth development, and in determining the intensity of activation at the
end or surface of a piece it plays a part in determining the length of the
resulting gradient and the scale of organization of the hydranth, and to
that extent is concerned directly in determining the polar pattern. In the
aggregates of dissociated cells the contact-free-surface differential, prob-
ably chiefly or wholly an oxygen differential, may apparently be directly
concerned in determining the polar gradient. However, whether the dif-
ferential acts merely as a regional activator or directly as determiner, the
final result is determination of the polarity of the whole piece or aggregate.

These experiments with hydroids have been discussed at some length
because they appear to be particularly significant for the problems of
developmental pattern and polarity. If a differential in oxygen tension
can determine a physiological polarity that can be distinguished as a
gradient and becomes the basis of an axiate pattern with definite differ-
entiations along its course, the most logical and obvious conclusion is
that physiological polarity in its simplest terms originates as a gradient
involving differences in rate of the basal metabolism characteristic of the
protoplasm concerned.

DETERMINATION OF RECONSTITUTIONAL PATTERN BY OTHER
ENVIRONMENTAL FACTORS

Light has been shown to be an important factor in determining axiate
pattern in many plants, but influences chiefly the course of vegetative
development. In regeneration of the alga Bryopsis, development of a
new thallus axis instead of rhizoid axis can be induced by light at the
proximal cut end, although the plant has a well-defined polarity (Noll,
1900; Winkler, 1900a). This is one of the few plants in which reconstitution
occurs directly from the region of injury, as it does very generally in ani-
mals. According to Loeb (1895) and Goldfarb (1906, 1910), light is neces-
sary for reconstitution, or for continued reconstitution of hydranths in
certain hydroids; but this effect is apparently on axiate patterns already

RECONSTITUTIONAL PATTERNS IN EXPERIMENT 421

present rather than a determination of the patterns. Doubtless it does
alter gradients that may be present, and perhaps with proper procedure
it may determine axiate pattern.

Single cells of the red alga Griffithsia hornetiana, subjected to electric
current, develop rhizoids toward the anode, irrespective of the original
polarity; and chromatophores migrate toward the anode. This localiza-
tion of rhizoid formation is regarded as resulting from migration of charged
particles (Schechter, 1934).

Internodes, that is, pieces of stem between successive branches, of the
hydroid Ohelia in sea water usually give rise to hydranth-stem axes at
both ends, development distally being more rapid than proximally. In
this hydroid reconstitution involves outgrowth of tissue from the ends,
rather than reorganization without outgrowth, as in Tuhularia and Cory-
morpha. Electric current of proper density, flowing longitudinally, deter-
mines a high frequency of hydranth-stem development on the end toward
the anode and delays or completely inhibits similar development at the
end toward the cathode (Lund, 1921c). With certain current density, hy-
dranth development is inhibited at the cathodal ends of Tuhularia pieces,
as reported by Lund for Obelia; but with a sufficient increase in current
density the inhibition is reversed and becomes greatest at the anodal
end. In Eudendrium pieces inhibition is greatest at the cathode end,
but in Pennaria pieces it is greatest at the anodal end with all current
densities used (Barth, 1934a). According to Barth, effect of current in
these cases appears to be largely a difference in degree of inhibition at the
two ends, although in Tuhularia pieces hydranth frequency is higher at
the anodal end in current densities which determine cathodal inhibition
than in controls. This, however, may result from the cathodal inhibition,
for it has been repeatedly shown that hydranths develop more rapidly at
proximal ends of pieces when distal hydranth development is inhibited.

Strictly speaking, electric current, like the oxygen differential, deter-
mines in these cases which of two possible polarities shall express itself
in development. The activation following section determines the polarity
or polarities in the piece, and the current obviously affects the activation
at cathode and anode differentially or in different degree. The result is,
of course, determination of polarity of the piece, but perhaps the current
should be regarded as selecting the polarity rather than as determining it.
Whether electric current can determine a new polarity in hydroid tissue
independently of a cut end has not yet been discovered, but that it can
do so seems highly probable. Undoubtedly, a sufficient difference in many

422 PATTERNS AND PROBLEMS OF DEVELOPMENT

other physical or chemical factors at the two ends of Tubularia or Cory-
niorpha pieces would have effects very similar to those of electric current ;
but, except for temperature difference, reported to be effective in Cory-
morpha by Gilchrist and Schmidt (1932) and in Tubularia by J. A. Miller
(1939), and for difference in H-ion concentration, also stated to be effec-
tive,^'^ experimental evidence seems to be lacking.

In many plants gravity is a factor in localizing regions from which
rhizoids or roots develop and in determining what bud primordia shall
develop, but apparently it is not an essential factor in determining the
root or bud gradient. Its effect in these cases appears to be determination
in certain regions of conditions favorable for initiation of rhizoid or root
formation. Such determination may, of course, play a part in establishing
a general polarity of the whole plant or piece, but the polarity of each
rhizoid or root apparently results from the local activation and initiation
by it of a gradient system. In buds activated by gravity the gradient
pattern is already present, and gravity appears to be merely an activator.
Gravity is supposed to bring about differential distribution of substances
of different specific gravity and may determine gradients in concentration
of such substances, but these gradients are not essential to the axiate
patterns of particular rhizoids, roots, or buds.

Prolonged low-speed centrifuging of isolated cells of the alga Griffithsia
tends to localize development of new apical cells and resulting thallus
axes centrifugally, where heavier substances are concentrated; but the
concentration of substance is regarded as stimulating or activating rather
than as directly determining the new polarity (Schechter, 1935).

According to Loeb (1891), pieces of the hydroid A ntennularia antennina,
suspended in various positions, develop stolons from parts toward the
earth, hydranth-stem axes from parts extending in the opposite direction.
Stolons develop even from the apical ends of lateral branches or of the
main axis when these point downward. Further experiments by Morgan
(1901) and Stevens (1902, 1910) do not entirely confirm Loeb's results.
These authors find that stolons often develop from both ends of sus-
pended pieces and that pieces in various positions on a slowly revolving
vertical wheel in water usually develop hydranth-stem axes. As in the
case of Plumularia, this latter result is probably due to better oxygen
supply and removal of CO2 with the constant change of position on the

'^ According to Komori (1933), pieces of Tuhularia reconstitute stolons or nothing from the
distal ends at pH 6 and hydranths from proximal ends at pH 8.45. This difference in H-ion
concentration is not great and raises the question whether hydrogen ion or CO^ is the effective
factor in this case.

RECONSTITUTIONAL PATTERNS IN EXPERIMENT 423

revolving wheel. That Loeb's results may have been due to an oxygen
gradient decreasing downward is perhaps possible. In general, gravity
or, as will appear below, even high-speed centrifuging does not seem to be
a very important factor in determining or altering axiate pattern in ani-
mals, though there are some cases in which one or the other factor is
effective; and further experiment, particularly with the very high centri-
fuge speeds now possible, may bring new evidence on this point.

B

Fig. 143, A-H. — Early development of Fiicus. A, normal development with plane of cell
division at right angle to axis indicated by rhizoid outgrowth, separating a rhizoid cell and a
thallus cell; 5-Z), altered relations of rhizoid outgrowth and division plane resulting from
change in direction of illumination; E-H, bipolar forms resulting from periodic change of 180°
with respect to direction of light.

EXPERIMENTAL ALTERATIONS OF EMBRYONIC AND AGAMIC
PATTERN BY ENVIRONMENTAL FACTORS

So far as data are available, pattern in most eggs and early embryonic
stages appears to be relatively stable in relation to external differentials,
but in some forms new pattern can be determined experimentally by such
factors. Perhaps the most interesting case is the egg of the alga Fucus.
This egg apparently possesses a polarity when shed, since it can develop
the axiate rhizoid and thallus in complete darkness without any evident
relation to external factors. The first indication of development in the
originally spherical cell is a bulging or outgrowth from a part of its sur-
face, followed by a cell division in a plane transverse to the axis of the
outgrowth (Fig. 143, A). The outgrowth represents an early stage of the
first rhizoid and, with various methods, shows a gradient with high end

424 PATTERNS AND PROBLEMS OF DEVELOPMENT

at the tip. The other cell gives rise to the thallus; and, as this develops,
a gradient, with high end apical gradually appears in it. It has long been
known that polarity of this egg can be determined by light, that is, by
differential illumination, the rhizoid developing on the side away from
the light, with the first division plane transverse to the direction of illumi-
nation.^^ More recently a certain range of wave length toward the blue-
violet has been shown to be chiefly effective (Hurd, 1919, 1920). With
repeated (e.g., hourly) change of position with respect to direction of in-
cident sunlight, the axis indicated by outgrowth of rhizoid and the divi-
sion plane may form any angle with each other (Fig. 143, B-D). From
changes of 180° in position bipolar forms frequently result, either with
rhizoids at or near opposite poles (Fig. 143, E-G) or with equal division
and no rhizoids, like Figure 143, i^ (Child, unpublished). These varia-
tions suggest that rhizoid outgrowth and cell division require different
periods or intensities of illumination for determination of the usual rela-
tion, but further experiment on this point is desirable.

Polarity of this egg may also be determined by electric current, the
rhizoid developing on the side toward the anode (Lund, 19236). When
several eggs lie close together, there is mutual determination of polarity,
the rhizoids arising toward the center of the group, thus indicating a
chemical differential dependent on presence of the eggs. It has been fur-
ther shown that increase in H-ion concentration increases this group effect
and that a H-ion gradient induces rhizoid development on the more acid
side, up to a certain concentration, and above that on the less acid side.
However, even with removal of excess CO2 from acidified sea water the
possibility that it is concerned in this effect is not entirely excluded.^' In
another line of experiment it was found that the rhizoid tends to form on
that side of the egg to which a sufficient concentration of the potassium
salt of indole acetic acid (heteroauxin) is applied, and in the light of these
results it was suggested that certain agents which determine polarity in
this egg may do so by influencing in some way the activity of auxin in
the egg.''° Whitaker points out that this hypothesis will account for most
effects of increased H-ion concentration, since acidity increases auxin ac-
tivity. On this basis effects of high H-ion concentration result from in-
crease in auxin activity to the point at which it becomes inhibitory.

^* Farmer and Williams, 1898; Winkler, igoob; Kiister, 1906; Kniep, 1907. The first
division plane and the polarity of the Equiselum spore show a similar relation to direction of
incident light (Stahl, 1885).

^"Whitaker, 1935, 1937a, 1938a; Whitaker and Lowrance, 1937.

i" Du Buy and Olson, 1937; Olson and Du Buy, 1937.

RECONSTITUTIONAL PATTERNS IN EXPERIMENT 425

Polarity of Fucus eggs may also be determined by centrifugal force.
After a certain intensity and period of centrifuging eggs suspended in firm
agar-sea water to prevent possible orientation, 93-99 per cent of eggs with
persisting stratification develop rhizoids at or near the centrifugal pole,
but redistribution of egg substances abolishes this effect (Whitaker,
19376). According to Beams (1937), ultra-centrifuging does not affect
polarity of eggs of another species of Fucus. Whitaker (19386) finds, how-
ever, that at pH 7.9-8.1 the H-ion concentration of sea water, all ultra-
centrifuged eggs form rhizoids on the centrifugal halves, while at pH
5.8-6.0, well on the acid side, 90 per cent form rhizoids on centripetal
halves. He makes the further suggestion that auxin may be adsorbed on
heavy substances and attain inhibitory concentrations at the centrifugal
pole in acid sea water, or its transport may be affected by amphoteric
substances, with reversal when the isoelectric point is passed. Eggs ex-
posed to a slight temperature gradient develop rhizoids on the warmer
side (Lowrance, 1937a, h). When recently fertilized eggs are drawn into
a capillary pipette with lumen small enough to elongate the egg, the
pectin or cellulose wall "sets" and the egg retains ovoid form after extru-
sion: in darkness 96 per cent (of 114 eggs) form rhizoids at or near one
end of the longitudinal axis. At pH 6, however, rhizoids tend to form on
the surface in contact, where diffusion is less rapid, and this differential
supersedes the effect of shape (Whitaker, 1938c, 1940).^^ Whatever the
physiological factors concerned, it is evident that, even though a polarity
may be originally present in the egg, the polar pattern can be reconsti-
tuted by various external differentials, entirely without section or other
injury to the egg.

In many animal eggs cytoplasmic differentiation is apparently so far
advanced at the beginning of embryonic development that axiate pattern
appears highly stable, but alteration is still possible in some eggs. When
polarity is obliterated by differential inhibition in the blastula of the
hydromedusa Phialidium (pp. 168-69), the blastula becomes solid and
spherical, and after return to water rolls about on the bottom of the con-
tainer for perhaps several days, ciliary activity being no longer co-ordi-

3' For determination of dorsiventrality in an echinoderm egg in a similar way and the
interpretation suggested see p. 427. More recent papers on experimental determination of
polarity in Fucus and the related form, Pelvetia, are as follows: Whitaker, 1940, "The effects of
ultra-centrifuging and of pH on the development of Fucus eggs," Jour. Cell. Comp. Physiol.,
15; Whitaker and Lowrance, 1940, "The effect of alkalinity upon mutual influences deter-
mining the developmental axis in Fucus eggs," Biol. Bull., 78; Lowrance and Whitaker, 1940,
"Determination of polarity in Pelvetia eggs by centrifuging," Gro2vth, 4.

426

PATTERNS AND PROBLEMS OF DEVELOPMENT

nated. It gradually comes to rest, apparently in any position, there being
no evidence of orientation when it is rolled about passively. After coming
to rest, it flattens to a more or less hemispherical mass and secretes peri-
sarc about itself (Fig. 144, A). With further recovery a hydranth-stem
axis may develop from its upper surface (Fig. 144, B) ; or two, three, or
even four new axes, one a hydranth-stem axis, the others stolons (Fig. 144,
C), or all stolon axes (Fig. 144, D) may develop. These stolon axes often
give rise later to hydranth-stem axes from their upper surfaces (Child,
19256). The contact-free-surface differential is evidently a factor in de-
termining the new patterns: hydranth-stem axes always develop from

Fig. 144, A-D. — Development of new polarities in blastula of the hydromedusa PhiaJidium
after obliteration of polarity by differential inhibition. A, solid blastula flattened on the sub-
strate after cessation of ciliary activity; B, hydranth-stem axis developing from upper sur-
face; C, hydranth-stem axis from upper surface and three stolon axes below; D, three stolon
axes (from Child, 1925c).

the upper free surfaces of the masses, and stolon axes always in contact
with the substrate. Particular stolon axes are probably localized by slight
chance differences in condition in different regions, though perhaps some
unrecognized external differential may contribute to their localization.
In development from planulae under natural conditions stolon axes do
not appear until considerably later stages, after the hydranth-stem axes
has developed. Differentially inhibited planulae give very similar results,
though their polarities are not usually completely obliterated, and stolon
axes develop from one or both ends but may appear elsewhere.

Polarity of animal eggs is usually not altered by centrifuging, though
position of polar-body formation may be altered by displacement of ma-
turation spindle or of nucleus, but ventrodorsality apparently can be
altered by centrifugal force in some eggs. In eggs of certain sea urchins,

RECONSTITUTIONAL PATTERNS IN EXPERIMENT 427

centrifuged before or after fertilization, the centrifugal region tends to
become ventral. ■'" According to Lindahl, however, in centripetal frag-
ments of these eggs the ventral side is centripetal, and in another species
he finds the ventral side centripetal. Also, in eggs stretched by being
drawn into a capillary tube the end in advance becomes ventral, because
greater stretching of this end renders plasma colloids less stable. Deep
staining of this end with Nile blue sulphate makes it dorsal. Lindahl sug-
gests that the ventral region is an inductor and has a higher metaboHsm
than the dorsal region and that the region where protoplasmic inclusions
are aggregated becomes ventral because the inclusions determine higher
metabolism in it. Differential dye reduction and susceptibility agree in
indicating or suggesting higher metabolism in the ventral region in normal
sea-urchin embryos.^^ But why the ventral region should be centrifugal
in the whole egg and centripetal in a centripetal piece is not entirely clear.

Unfertihzed Dendraster eggs centrifuged 6-8 minutes at 45,000 times
gravity develop into plutei with an undivided centripetal lobe, situated
ventrally at the angle between ventral and anal surfaces in 93.6 per cent
of the larvae. Of these, 56.9 per cent have the lobe lateral on the ventral
surface, in 14.9 per cent it is nearly median, in 22 per cent intermediate
(Pease, 1939). Ventrodorsality is believed to be present in the egg and
to be shifted by centrifuging. A "ventral-determinant" gradient is postu-
lated with highest concentration or activity ventral, probably cortical,
perhaps an enzyme, requiring a substrate probably dififuse in the ento-
plasm but concentrated centripetally by centrifuging and so partially ro-
tating the ventrodorsal axis. In centrifuged eggs of the gephyrean Urechis
the centripetal region also tends to be ventral and without relation to
point of sperm entrance or first cleavage (Pease, 1938). The "determi-
nate" cleavage pattern of ultra-centrifuged eggs of the pelecypod Cumin-
gia and the polychete Chaetopterus is related to the stratification, though
with wide variation in Cumingia; and polarity and bilaterality are ap-
parently determined in relation to the cleavage pattern. ^^

Eggs of the sea urchin Arhacia suspended in sugar solution of proper
density can be separated by strong centrifuging into two parts, the cen-
tripetal part colorless, the centrifugal granular and pigmented; and each
of these parts can be again separated by further centrifuging. These egg

32 Runnstrom, 1925c, 1926a; Lindahl, 1936.
3-5 See pp. 134-38 and chap. vi.

i"^ Pease, 1940, "The influence of centrifugal force on the bilateral determination and the
polar axis of Cumingia and Chaetopterus eggs," Jour. Exp. Zool., 84.

428 PATTERNS AND PROBLEMS OF DEVELOPMENT

fragments can be fertilized and show more or less development, the color-
less halves becoming plutei, the pigmented halves containing only the
male nucleus, sometimes forming blastulae and occasionally plutei. Un-
fertilized centrifuged halves without nucleus, vv^hen artificially activated,
may cleave and form blastulae. Centripetal halves, artificially activated,
usually develop normally. When centrifuged after fertilization, centripetal
halves, containing both nuclei, develop, but plutei are not normal; cen-
trifugal halves do not develop. ^^ Evidently there is extensive reconstitu-
tion in those fragments which develop, but exactly how the pattern is
altered is not known. These, like many other centrifuge experiments, sug-
gest that the pattern in these eggs is chiefly or primarily cortical but
that protoplasmic content inside the cortex, as distinguished from granu-
lar inclusions, is perhaps of significance in maintaining sufficient cortical
activity for development.

In most animal eggs a considerable degree of stratification by centri-
fuging may occur without essential alteration of developmental pattern,
though, as noted above, it may alter position of polar-body formation by
displacing the nucleus or maturation spindle. However, eggs of certain
ascidians stratified by centrifuging show dislocation of pattern of tissues
and organs (Conklin, 1931).

Perhaps the most interesting, certainly the most extensively investi-
gated and discussed, case of alteration of animal developmental pattern
by gravity and centrifugal force is that of the amphibian egg and early
embryo. Only brief consideration of some of the more important points
is attempted here. In amphibian eggs and early cleavage stages, main-
tained in inverted or partly inverted positions with respect to gravity, or
centrifuged in these positions, more or less complete reversal in position
of heavier and lighter substances occurs. The heavier yolk accumulates
in the original apical region, displacing the lighter parts of the cytoplasm
to the upper, originally basal region, and certain alterations of pattern
result. ^^

Development of double monsters from the inverted two-cell stage, first
observed by Schultze, was usually interpreted as the result of decrease

35 Harvey, 1932, 1936, 1939.

36 Born, 1885; Schultze, 1894; Wetzel, 1895, 1896; Chiarugi, 1898; Tonkofi, 1900, 1904;
Bagini, 1923; Schleip und Penners, 1925, 1926; Wiegmann, 1926, 1927; Hammerling, 1927;
Penners und Schleip, 1928; Penners, 1929; Wittmann, 1929; Dalcq et Pasteels, 1938; Pasteels,
1938; also, omitted from bibliography, Pasteels, 1938, "Recherches sur les facteurs initiaux
de la morphogenese chez les Amphibiens Anoures, I," Arch. Biol, 49. For a general discus-
sion with bibliography see Schleip, 1929, pp. 584-90, 696-715.

RECONSTITUTIONAL PATTERNS IN EXPERIMENT 429

or elimination by the cytoplasmic movements of relations between the
two blastomeres, so that reconstitution resulted in a more or less com-
plete, instead of a half -embryo, from each. Later experiments showed that
this interpretation was not adequate, for it was found that double em-
bryos might develop from inverted undivided eggs, from four-cell, and
even from eight-cell stages, and occasionally triple monsters developed
from two-cell stages (Wetzel, 1896; Schleip und Penners, 1925). Most
authors are agreed that the region of the gray crescent, which normally
becomes the dorsal lip of the blastopore, the ''organizer" region, does not
change its position in inverted eggs; and, according to Penners and Schleip
(1928), sinking of the white yolk occurs in different ways in different
eggs, and portions of it may remain in the original basal region or at the
surface along the cleavage furrows present. These authors maintain that
gastrulation and organization are not necessarily dependent on the origi-
nal dorsal lip region but may result from localization of such regions where
a band or mass of white yolk is in contact with other cytoplasm. Double
embryos arise by localization of dorsal lip regions and invagination in op-
position directions on both sides of a band or streak of yolk, left behind
as the yolk sank. Gastrulation and organization may also be localized
in relation to other yolk masses that failed to sink or reached the surface
at the lower, originally apical, pole. Evidently the region of the gray
crescent is not definitively determined as the only dorsal lip region and
organizer up to four-cell and eight-cell stages. Other regions are capable
of becoming organizers and of determining gastrulation and position and
axial direction of the neural plate. Locahzation of gastrulation and neural
plate are related to gravity or centrifugal force only in so far as these
agents alter local relations of yolk masses and cytoplasm.

Further analysis by means of partial inversion confirms the work of
Penners and Schleip as regards the significance of contact of cytoplasm
and yolk masses for localization of invagination and of the neural in-
ductor or organizer and leads to the further conclusion that localization
of the blastopore or blastopores is as near the original prospective dorsal
lip as the relations of cytoplasm and yolk permit (Pasteels, 1938; Dalcq
et Pasteels, 1938). In frog eggs maintained in inverted position by com-
pression between slides a blastopore lip forms at the edge of peripheral
yolk but, because of the compression, does not invaginate, fades out in a
few hours, and another or others may be formed (Pasteels, 1939). Dalcq
and Pasteels conclude that the essential pattern of amphibian develop-
ment is a primary apicobasal cytoplasm-yolk gradient and a dorsal cor-

430 PATTERNS AND PROBLEMS OF DEVELOPMENT

tical area with gradient of potency for gastrulation and inductor develop-
ment decreasing from the mid-dorsal region of the gray crescent in the
frog egg, that is, from the region of primary invagination in normal de-
velopment. They regard the metabolisms resulting from, and determined
by, local concentrations of cytoplasm and yolk and from relative differ-
ences in amounts of each as essential factors in development. In the
single, double, or triple embryos and monsters developing from these in-
verted eggs polarity, symmetry — in fact, the whole axiate pattern — may
be the result of a reconstitution and entirely a new pattern ; but the locali-
zation of new dorsal lips, blastopores, and neural plates shows a relation
to the original dorsal region, suggesting a graded differential in that re-
gion. It is of some interest to note that results of experiments on differ-
ential susceptibihty,. differential dye reduction, and distribution of SH-
proteins and some of the data on oxygen consumption are in general
agreement with these conclusions (see. pp. 151-58).

As regards the vegetative reproductions of parts in plants, it was
pointed out above that external factors may determine the regions in
which certain parts develop, but not their axiate patterns. It has long
been known that pattern of branching in the thalli of certain algae can
be determined by direction of illumination in relation to the main axis
of the thallus. Branching is radially symmetrical when illumination is
equal on all sides, bilateral or dorsiventral when it is unequal (Berthold,
1882). Other external factors — for example, gravity — may affect pattern
of arrangement of axes in many plants though not determining the pat-
tern of the particular axiate parts. In various plants, however, external
factors may determine dorsiventrality directly. For example, it is well
known that light can determine dorsiventrality in certain algae, mosses
and liverworts, and in the prothallia of ferns. In some of these forms a
dorsiventrality once determined can be reversed in the later growth by
reversing direction of illumination; in others it is stable. Rhizoids and,
in fern prothallia, sex organs develop from the side regarded as ventral,
and in many liverworts the thallus itself becomes dorsiventrally differ-
entiated. It has been shown for many of these forms that the entire dorsi-
ventral pattern of further growth is experimentally reversible by reversal
of illumination. The gemmae of various liverworts possess bipolar pat-
tern with apical growing cell at each tip and each gives rise to two thalli
that grow in opposite directions and finally become separated. Extensive
experiments with gemmae of Marchantia and Lunularia have shown that
their dorsiventrality can be determined by light, by gravity, or by a

RECONSTITUTIONAL PATTERNS IN EXPERIMENT 431

nutritive agar, or even a purely mineral substrate. In general, the less
illuminated side or the side toward the earth becomes ventral, but the
side toward the substrate becomes dorsal. In Marchantia, however, the
effect of light depends on temperature. With germination at high tem-
perature the less illuminated side tends to become dorsal instead of ven-
tral ; also, with light and gravity acting in opposite directions, the effect
of gravity tends to overbalance that of light at high temperature. With
equal action of external factors in opposite directions on the two sides
both may differentiate in the same way (Fitting, 1935, 1937). In some
of the higher plants dorsiventrality may be determined by Hght, and
gravity may also be a factor. It is an interesting question how the region
of the circumference of the plant vegetative tip where the first lateral bud
primordium develops, or the first two, three, or more in forms with oppo-
site or whorled lateral buds, are determined. Perhaps a more complete
acquaintance with botanical literature would show that this question has
been answered, but essentially similar questions regarding animal develop-
ment await an answer.

Physiological, rather than external, factors are usually concerned in
localizing agamic reproductions of new individuals and development of
parts and organs in animals, but there are some cases in which external
factors may be concerned and regarding which the question raised above,
concerning lateral bud primordia in plants, must be asked. For example,
how is the region of the circumference determined where the first lateral
bud appears on a Hydra individual? The same question arises with re-
spect to the first medusa bud on the manubrium of Pennaria or other
hydroids. In development of the hydranth from the planula of Corymor-
pha a single tentacle often appears before others. How is its localization
on the circumference of the planula determined? This planula does not
swim but creeps on the substrate with one side in contact; the first ten-
tacle often develops on the upper side, but that it always does is not cer-
tain. In other individuals two tentacles on opposite sides, or three, equi-
distant from each other, develop apparently simultaneously. The apical
region of the planula often turns away from the substrate before tentacles
develop; but even if equal exposure determines simultaneous develop-
ment of two or three tentacles, the question of how they are localized on
the circumference remains. With localization of one Hydra bud, medusa
bud, or tentacle bud, its local dominance may play a part in determining
localization of another by determining that another cannot develop within
a certain distance from itself. A question of somewhat different character

432 PATTERNS AND PROBLEMS OF DEVELOPMENT

is that of the factor or factors determining development in one plane of
the sea fans (Gorgonacea). In fans developing on more or less vertical
rock faces with water movement chiefly vertical, the plane of develop-
ment is usually, if not always, vertical. These and many other cases of
spatial pattern and order of zooids or parts present problems of funda-
mental significance for our conception of development.

CONCLUSION

Some of the ways in which axiate patterns can be experimentally de-
termined have been considered in this chapter. So far as the examples
given have been analyzed, they appear to involve alteration, obliteration,
and determination of dominance and a gradient or gradients as the earliest
distinguishable feature of the change in pattern. In the light of the experi-
mental data it appears that a factor operative in the reorganization of
other parts in reconstitution is associated with the high end of a gradient
or gradient system. In isolated pieces the region or regions most intensely
activated following section or otherwise become more or less dominant
and alter a pre-existing gradient or determine a new one in a different
direction from that already present. A new gradient and dominance may
also be determined by environmental differentials or gradients of various
kinds. However a gradient is initiated, the specific constitution and phys-
iological condition of the protoplasm concerned are undoubtedly the chief
factors determining its physiological characteristics, its length, steepness,
and the effective range of dominance.

The determination of polarity in Corymorpha cell aggregates by the
contact-free-surface gradient, probably an oxygen gradient in the absence
of any cut surface, the effectiveness of ganglionic planarian grafts in induc-
tion of reorganization in the host body, the obHteration of fission zones
by reconstitution of heads a short distance anterior to them, the destruc-
tion of headless parts of zooids and whole zooids in early stages by a more
advanced head region posterior to them in Stenostomum pieces, and in
general the very evident relation between dominance and dynamic, rather
than structural, conditions— all support the view that dominance and the
gradient or gradient system associated with it are primarily dynamic in
character, not structural, except in so far as activity and structure of
some sort cannot be dissociated in living protoplasms. Vital activity, me-
tabolism, appears to be the primary factor determining pattern rather
than a pre-existing structure determining metabolism.

Reconstitution of a hydranth or head is not a replacement or restitu-

RECONSTITUTIONAL PATTERNS IN EXPERIMENT 433

tion of a part removed but the development of a new axiate pattern, be-
ginning with the apical region or head, which develops from the high
region of the gradient determined by section and isolation. Reconstitu-
tion of hydranth or head at the distal or anterior end of a piece from any
body-level except one immediately adjoining the original hydranth or
head is just as truly a heteromorphosis as a hydranth or head developing
at the proximal or posterior end. The one is "out of place" as much as
the other. The parts normally present between such a developing hy-
dranth or head and the level of section where it develops are absent until
later, when reorganization is induced by the new dominance and gradient.
At either end of the piece or elsewhere the reconstituting hydranth or head
represents beginnings of a new axiate pattern.

Development of hydranth or head on an isolated piece can occur only
with a certain degree of physiological isolation of the cells concerned from
other parts of the piece ; this isolation results from the activation of cells
following section and isolation; this, in turn, alters or obliterates the old
polar gradient pattern and determines a new one. It is no exaggeration
to say that development of a hydranth or head or, in general, determina-
tion of a new gradient in reconstitution occurs in spite of the rest of the
piece, that is, in spite of the pre-existing gradient and organization. The
old pattern is more or less completely made over. On the other hand,
development of a basal or posterior end is primarily development of a
more or less subordinate part, determined either by the dominance of
parts anterior or distal to it or by inhibiting external conditions. The
hydroid stolon is an axiate pattern with its own dominance and gradient,
but still in some degree under the dominance of the hydranth-stem axis
or directly determined an external inhibiting factor.

New pattern can be determined not only by the activation following
section and isolation but by implants and by external differentials —
light, gravity, centrifugal force, electric current, temperature, an oxygen
differential, H-ion concentration — in many organisms, both plant and
animal, by more than one of these factors. Even change of shape, prob-
ably involving local or regional stretching of the cortex and consequent
alteration of its physiological condition, is effective in some eggs. But
however the new pattern is initiated, it is, of course, the specific constitu-
tion of the protoplasm in which it appears that determines its char-
acteristics as developmental pattern. Induction by a dominant region of
reconstitution in other parts is very generally characteristic of reconstitu-
tional development, but a new gradient may be directly determined by

434 PATTERNS AND PROBLEMS OF DEVELOPMENT

an external gradient ; in those cases dominance probably develops second-
arily from the high region. It appears sufficiently evident that physiologi-
cal and external factors effective in determining new developmental pat-
tern are, in general, factors influencing the physiological activity, the
metabohsm, of Hving protoplasm. Usually they are activating factors,
though in some cases pattern of one kind can be transformed into another
by depressing or inhibiting factors — for example, the transformation of
hydranth-stem axes into stolon axes by low oxygen tension or by inhibit-
ing chemical agents. In view of all the evidence developmental pattern
appears to be primarily an expression of the dynamics of hving proto-
plasms.

CHAPTER XII

INDUCTORS AND SO-CALLED ''ORGANIZERS" IN
EMBRYONIC DEVELOPMENT

THE demonstration by Spemann and his co-workers that t'he re-
gion of the urodele amphibian embryo which becomes the dorsal
lip of the blastopore and on invagination forms chorda-mesoderm
is a dominant region and can induce or determine development of other
parts provided experimental evidence in support of the conclusion that a
relation of dominance and subordination may be of fundamental signifi-
cance in the development of the amphibian embryo. Such a relation had
already been shown to exist for certain organs of later developmental
stages — for example, in induction of a lens by an optic vesicle or cup.
Dominance of certain parts and physiological or physical isolation from
that dominance had also been shown to be fundamental factors in agamic
and reconstitutional development (see chaps, ix-xi). However, the dis-
covery of a regional dominance and of the presence of a so-called ''organ-
izer" or "organization center" in amphibian development has exercised
a sort of dominance over the field of experimental embryology and has
resulted, during the last fifteen years, in a tremendous amount of investi-
gation on various aspects of the problem and in the discovery or postula-
tion of many other "organizers," concerned with one feature or another of
development, not only in amphibia but in many other organisms. In
chapters ix-xi it was shown that dominant regions in many reconstitu-
tions and agamic reproductions are apparently primarily the high regions
of gradients and that their dominance results from their activity rather
than from specific differentiation. Dominant regions resulting from sec-
tion and isolation alter existing gradients or determine new ones. In view
of the evidence concerning the relation of gradients and dominance to
axiate pattern, it appears probable that the gradient determined by an
activated region is the real organizing factor. According to this concep-
tion, the region of primary activation is an organizer only indirectly, by
initiating and determining a gradient pattern; conditions at different
levels of this pattern determine the orderly localization of parts along an
axis. In short, these experimentally determined dominant regions in the

435

436 PATTERNS AND PROBLEMS OF DEVELOPMENT

simpler animals are inductors of gradients, and the gradients are the real
organizers. Since the concepts of inductors and organizers have developed
from experiment on embryonic stages, it is necessary to raise the question
whether, or to what extent, similar factors are concerned in the action of
embryonic inductors and the inductors concerned in reconstitutional de-
velopment of isolated parts of adult animals. Experimental data bearing
directly on this question from the embryonic side concern chiefly certain
echinoderms, fishes, amphibia, and birds; but embryonic development of
various other forms affords some data which are also suggestive or posi-
tive.

INDUCTION IN ECHINODERMS

Before turning to the experimental data attention is again called to a
few points in the earlier discussions of echinoderm developmental pattern
(chap. vii). The hypothesis of two opposed, overlapping gradients of
concentration of substances, an "animal" and a "vegetal" gradient, the
one decreasing basipetally, the other acropetally, advanced by Runn-
strom and accepted and developed by his co-workers, Horstadius and
Lindahl, has been discussed (pp. 241, 243). A third, ventrodorsal gradient
has also been postulated, and Horstadius has suggested two opposed gra-
dients in the right-left axis.' Since animal and vegetal gradients overlap
according to this hypothesis, they must be specifically different; Runn-
strdm (1933) and Lindahl (1936) have advanced further hypotheses con-
cerning the character of animal and vegetal metabolisms. Moreover, to
account for their experimental results these authors find it necessary to
assume that one of these gradients in the polar axis may "suppress" the
other more or less completely, that both may be altered in extent or other-
wise, and that certain external agents affect one or the other specifically.

Von Ubisch has arrived at somewhat different conclusions." He also
postulates existence of two substances, ectodermal and entodermal, de-

1 References p. 241, footnote 20. In discussion of the experiments of these authors it will
be convenient to follow their usage of the terms "animal" and "vegetative," except that
"vegetal" is used instead of "vegetative." However, these terms are antiquated and have
little or no meaning for echinoderm development, except as indicating position or direction in
relation to the polar axis. The terms "apical" and "basal," and "acropetal" and "basipetal,"
seem equally applicable and somewhat more convenient and are used except where the others
seem to be required. The terms "pole" and "antipole" have been used by some authors for the
two ends of the polar axis, but"polar" and "antipolar" are less satisfactory than "apical"
and "basal," and the use of "polar" to indicate one pole or the region adjoining that pole,
instead of the entire physiological axis, is likely to lead to confusion.

2 Von Ubisch, 1925a, b; 1929; 1931; 1932(1, b; 1933; 1934; 1936a, b; 1938a.

EMBRYONIC INDUCTORS AND ORGANIZERS 437

creasing in concentration respectively basipetally and acrope tally, but as-
sumes that after isolation of parts ectodermal substance accumulates
apically, entodermal substance basally, and that transplanted micro-
meres are not organizers, as Horstadius maintains, by production of "ento-
dermal" substance, but merely attract it.

These hypotheses are attempts to interpret sea-urchin development and
its experimental modifications in terms of formative substances. They
present interesting similarities to the attempts made by Morgan to ac-
count for reconstitution in Tubularia and other forms by assuming two
opposed material gradients and the changes in them required by the ex-
perimental data (Morgan, 1905, and various earher papers). It will be
recalled from earlier chapters that differential dye reduction in three sea-
urchin and one starfish species and differential susceptibility, as indicated
by differential death and differential modification of development, sug-
gest a somewhat different view. These data indicate presence in early
stages of only one gradient, corresponding in direction to the animal gra-
dient of Runnstrom; but preceding gastrulation a second gradient with
the same characteristics as the primary, as regards susceptibihty and dye
reduction, appears in the basal or vegetal region. The high end of this
gradient is distinctly higher than the high apical end of the primary gra-
dient ; and, as immigration of primary mesenchyme occurs, its cells become
the most rapidly reducing cells of the blastula. A ventrodorsal gradient
also becomes evident before gastrulation. The secondary gradient does
not overlap the primary but simply obliterates and reverses gradient di-
rection for a greater or less distance acropetally. Again it must be noted
that these susceptibility and reduction gradients do not constitute evi-
dence either for or against the overlapping substance gradients, though
they perhaps indicate that the difference of apical and basal metabolism
is not as great as Runnstrom and Lindahl believe ; they certainly indicate
that change in gradient pattern is a feature of normal development. That
different substances are present at different levels of the polar axis is indi-
cated by the pigment band of certain species (Boveri, 1901) and by dark-
field observations (Runnstrom, 1928a), but visible granules and sub-
stances appear, in general, to be primarily results of more fundamental
physiological differences along the axis. Unquestionably, a change in phys-
iological condition in the basal region takes place preceding gastrulation,
and immigration of mesenchyme and invagination of entoderm appear
to be associated with this change. Both susceptibility and dye reduction

438 PATTERNS AND PROBLEMS OF DEVELOPMENT

suggest rather intense activation in the basal region; a developmental
activation certainly occurs.

The first two divisions of the sea-urchin egg are meridional ; the third
is equatorial, apical and basal cells being approximately equal. In the
fourth cleavage the four apical cells divide meridionally, forming a ring
of eight cells (mesomeres), the four basal cells transversely and very un-
equally into the four micromeres at the basal pole and four large cells
(macromeres). Next the eight apical cells divide transversely and equally,
forming two rings of eight cells each, designated an^ and an2 by Hor-
stadius (Fig. 145, A). Somewhat later the four macromeres divide me-
ridionally, forming a ring of eight cells, then divide transversely, forming

arif
an 2

Fig. 145, .4, B. — Cleavage stages of sea urchin with Horstadius' designations of the four
rings of cells used in experiment; atii, auz, apical (animal) and subapical ring derived from the
four apical cells of eight-cell stage; vegi, vegz, rings derived from the four macromeres of sixteen-
cell stage which are basal (vegetal) except for the micromeres.

two rings of eight cells each, veg^ and veg, (Horstadius), as indicated in
Figure 145, B. Earlier workers disagreed as regards the portion invagi-
nated as entoderm, but by means of vital staining of particular rings of
blastomeres it seems to be demonstrated that an^, aHi, and vegi all nor-
mally form ectoderm and that only the cells of veg, invaginate (von Ubisch,
1933; Horstadius, 1935, 1936a).

With the aid of temporary exposure to calcium-free sea water Hor-
stadius has isolated blastomeres and groups of blastomeres as desired —
even single, two, three, or four micromeres — and, since the cells adhere
readily on contact, has been able to make various combinations of blasto-
meres or rings of blastomeres and to implant different numbers of micro-
meres in various relations to other cells. Staining of particular cells or
groups provides a means of identification, and local staining of isolated
parts serves to indicate axes, surfaces of separation, etc. In these experi-

EMBRYONIC INDUCTORS AND ORGANIZERS

439

ments the possibility exists that calcium-free sea water, isolation of micro-
meres or other blastomeres or groups, staining of particular cells or groups,
and the manipulation necessary for transplantation may alter physiologi-
cal condition of the cells concerned, either in the direction of activation
or of depression; but whether such change occurs is not known. That it

Fig. 146, .4 -i/.— Development of isolated apical halves of ParacentroUis. A, isolated apical
half; B, C, earlier and later stages of e.xtreme apical partial development; D, E, earlier and later
stages of less extreme apical development; F, G, H, essentially normal development follow-
ing implantation of four micromeres in basal region of apical half of early blastula (after
Horstadius, 1935).

may be a factor in the variation of results of a particular experiment
seems not impossible.

Apical halves, isolated in cleavage stages (Fig. 14^, A), usually develop
into thick-walled, blastula-Hke forms with extension of the apical tuft of
long stiff cilia over a large part of the surface (Fig. 146, B). These are
extreme apical partial forms, comparable to the apical partial forms from
short pieces of Tuhularia and Corymorpha (Fig. 113, A-I [p. 334])- Later

440 PATTERNS AND PROBLEMS OF DEVELOPMENT

the long cilia disappear, and they become uniformly ciliated spherical
forms (Fig. 146, C). Some apical halves, however, show less extreme api-
cal partial development (Fig. 146, D), and some may finally develop
ciliated band and stomodeum (Fig. 146, E). A.pical halves differentially
inhibited by lithium salts develop mesenchyme and entoderm (von
Ubisch, 19256, 1929) and occasionally do so in supposedly normal environ-
ment (von Ubisch, 1936a), though these cases may perhaps represent slight
differential inhibition by some uncontrolled factor. Evidently their fail-
ure to form mesenchyme under the usual conditions is due not to lack
of potency but to a scale of polar organization longer than the piece.
When the scale is decreased by lithium, basal parts develop. The case is
completely parallel to reconstitution of short Tuhularia and Corymorpha
pieces in relation to scale of organization (pp. 345-49). Animal halves
with four micromeres implanted basally develop into practically normal
plutei (Fig. 146, F-H); that is, the vegetal gradient is increased by the
micromeres (Horstadius), or the primary gradient is partly obliterated
and reversed by the activation of the micromeres.

Figure 147 indicates results of implanting one, two, or four micromeres
in the isolated blastomere rings. Afii requires four micromeres for normal
development; a«2 with one micromere approaches pluteus form, is normal
with two micromeres, and less completely developed apically with four.
Development of vegi is not complete in any case but becomes less com-
plete as more micromeres are implanted, and entoderm is relatively "too
large"; veg2, with one or more micromeres, develops exogastrulae with
large entoderm and small ectoderm,

Horstadius admits that this schematic representation is not entirely
adequate, and von Ubisch (1936a) has called attention to the selection of
particular experimental results on which it is based. Assuming, however,
that it does show, in a general way, effects of difference in apicobasal level
of blastomere rings and of number of micromeres present, it indicates, as
regards an^, awa, and vegi, that the effect of implanted micromeres in-
creases with the number implanted and with increasingly basal level of
the blastomere ring. If the micromeres undergo activation sooner or later,
as they apparently do preceding normal gastrulation, they probably oblit-
erate the primary gradient and reverse gradient direction over a greater
or less distance from the region of implantation; and this secondary gra-
dient extends farther and is more effective at lower than at higher levels
of the primary gradient, that is, in the more basal blastomere rings.
Activation of micromeres increases the scale of basal or vegetal organiza-

.5 T3

^ :0

^ o

£ *^

S .S

^ t-

442 PATTERNS AND PROBLEMS OF DEVELOPMENT

tion, and apical regions are correspondingly decreased. In the case of
o«i, normal development with four micromeres seems rather difficult to
account for on the basis of substance gradients. An^ represents the high-
est concentration of animal substance; and, since it is less than a fourth
of the whole polar pattern, there can be no great difference in concentra-
tion in it. The micromeres supposedly represent the highest concentra-
tion of vegetal substance. To produce an effective vegetal gradient this
substance must diffuse from the micromeres. To balance the animal gra-
dient it must approach the concentration of normal development; but in
that case we should expect no great difference in concentration of either
substance, since only the two ends of the polar axis are represented.
Evidently there must be mutual ''suppression" between the two gradients
as the vegetal substance diffuses from the micromeres, if anything like
normal development is to result, or else the accumulation at the two
poles of the two substances must be postulated; and these assumptions
raise further questions and difficulties. Similar difficulties arise with re-
gard to aw2 and vegi,; each represents only a fraction of the gradients, and
the micromeres another fraction. How do whole gradients arise from
these? The concept of dynamic gradients which can be altered in length,
height, or slope and obUterated or reversed by quantitative metabolic
differences seems to present fewer difficulties and is in Hne with observa-
tions on normal sea-urchin development and on reconstitution in many
animals. If a substance diffuses from the micromeres, may it not be simply
an activating substance rather than one producing some specific differ-
entiation?

Effects of the micromeres with veg^ appear anomalous, in terms of con-
centration gradients. The ring, veg2, is supposedly entirely entodermal,
but there is reconstitution of ectoderm from its more apical region and,
with micromeres present, exogastrulation. But, according to Horstadius'
experimental records seven of thirteen larvae from isolated vcg^ without
micromeres developed exogastrulae ; and of seventy-three with one or
more micromeres, thirty-seven showed different degrees of exogastrula-
tion, and thirty-six died early with little development. In normal develop-
ment veg^ does not form ectoderm at all; but when isolated, with or with-
out micromeres, ectoderm is reconstituted. According to the substance
hypothesis, it contains only a low concentration of animal substance but
a high concentration of vegetal substance. How can the animal substance
become effective in the presence of the high concentration of the other?
The reconstitution of ectoderm in veg2 is apparently quite similar to re-

EMBRYONIC INDUCTORS AND ORGANIZERS 443

constitution in hydroids and planarians. Isolation from higher levels of
the primary gradient results in activation of the more apical parts of
veg2 and raising of gradient-levels there, so that ectoderm, instead of
entoderm, develops; but the ectoderm does not attain full development,
probably because inhibited by the basal activation and gradient, much
as hydranth and planarian head are inhibited by an opposed activation
gradient.

It is a question of some interest whether exogastrulation in veg,, with-
out or with micromeres, results merely from the large proportional size
of entoderm or from some other condition. From Horstadius' data it does
not appear that the micromeres play any considerable part in determining
exogastrulation. In this respect Figure 147 appears somewhat misleading,
but it does indicate a greater degree of exogastrulation with larger ento-
derm and less ectoderm with increasing number of micromeres. Even if
the micromeres increase concentration of vegetal substance, it is not evi-
dent how such increase determines failure of entoderm to invaginate.
Since veg^, with or without micromeres, is retarded in development and
about half the number isolated die early, inhibition, rather than the micro-
meres, is probably the factor determining exogastrulation; but whether
early isolation, reconstitution of ectoderm, or some other factor is the
inhibitor remains uncertain. If these isolates are inhibited, exogastrula-
tion is probably determined in the same way as with chemical agents.

After implantation of four micromeres between an^ and an., of whole
embryos gastrulation and entoderm formation from presumptive ectoderm
occur at the region of implantation, as well as basally ; and plutei with two
guts and usually with extra skeletal elements result. Results are essen-
tially similar after implantation of four micromeres between an, and veg,,
except that the induced entodermal invagination is nearer the original
entodermal pole. Implantation of four micromeres in the apical pole of
whole embryos may induce a second small entodermal invagination in
that region, persisting in the apical region of the pluteus (Fig. 148, A-C),
and in some cases extra skeletal elements develop. Implantation of four
micromeres into the apical pole of apical half-blastulae results in a wide
range of forms. In some individuals development of the apical region is
more or less inhibited, but the original polarity persists; in others the
implanted micromeres may give rise to extra skeleton in the apical region ;
in still others a small apical invagination is induced, and another occurs
in the basal region of the apical half, that is, from presumptive ectoderm ;
and finally, in some there is a small apical invagination with complete

444

PATTERNS AND PROBLEMS OF DEVELOPMENT

reversal of polarity (Fig. 148, D-F). As far as the experimental evidence
goes, the invaginations induced by implanted micromeres may conceiv-
ably be due to an activation above the ectodermal level, such as appar-
ently occurs normally at the time of immigration of primary mesenchyme,
or to a specific effect. Incidentally, it may be noted that apical invagina-
tions, in appearance much like those figured by Horstadius, are present
not infrequently in secondary modifications of differentially inhibited
forms, even in exogastrulae. Although they appear at the extreme apical

Fig. 148, A-F. — Implantation of four micromeres into apical pole. A-C, implantation in
whole embryo with development of small entodermal invagination, persisting in apical region
of pluteus; D-F, implantation of micromeres into apical region of apical half-blastula with
complete reversal of polarity; stained micromeres and apical region of half indicated by shad-
ing in D (from Horstadius, 1935).

pole, they have been regarded as stomodeal invaginations. To account
for reconstitution of an apical region from the original basal region of
the apical half-blastula (Fig. 148, D-F), either in terms of activity gra-
dients or concentration gradients, requires special assumptions or hypoth-
eses. It may conceivably represent a physiological isolation, resulting
from alteration in the original apical region. When formed, the micro-
meres apparently represent the low end of the primary gradient; their
implantation apically may interfere considerably with the primary gra-
dient in early stages before their supposed activation and so may permit
physiological isolation of the basal region and reconstitution of an apical
region there. In terms of concentration gradients interpretation seems at

EMBRYONIC INDUCTORS AND ORGANIZERS 445

least equally hypothetical. Micromeres are supposed to increase the vege-
tal gradient. How do they determine an animal gradient opposite in di-
rection to the original in the basal region of the half-blastula? The gra-
dient changes required can, of course, be assumed to take place; but such
assumptions are certainly no less hypothetical than those concerning
physiological isolation, for physiological isolation is demonstrated in many
cases.

As already noted, development of mesenchyme and entoderm from the
basal regions of apical halves is brought about by certain exposures to
Kthium. The most basal regions of lithium-treated apical halves, marked
by local staining at the time of isolation, when implanted in the basal
regions of other apical halves, prevent the extreme apical partial develop-
ment characteristic of isolated apical halves and induce entoderm forma-
tion, and development of essentially normal plutei may result. These im-
planted cells are regarded as a secondary organizer by Horstadius (19366).
Cells from the apical regions of lithium-treated blastulae, similarly im-
planted in apical halves, perhaps have slight effect in inhibiting extension
of the apical tuft; but, since they do not induce entoderm, they afford
evidence that the effect of the "secondary organizer" is not due to pres-
ence of hthium in it. Horstadius regards the action of hthium as similar
to that of implanted micromeres, but the evidence from chapter vi indi-
cates that lithium is generally inhibitory ; and the production of entoderm
in the lithium-treated apical halves and in those with implanted basal
cells from hthium-treated apical halves is apparently quite similar to the
reconstitution of more basal parts in short pieces of Tuhulana and Cory-
morpha when scale of organization is decreased by inhibiting conditions
(pp. 344-48). Activation of mesenchyme and entoderm in the lithium-
treated half occurs in recovery after return to water, not as a direct effect
of lithium. Heteroplastic transplantation of micromeres has shown that
the inducing action and the reactions of the host ectoderm to the skeleton
are not species-specific and that the skeleton in the resulting chimeras
may possess characteristics of the donor species.'

In normal asteroid development there are no micromeres, nor is any
mesenchyme formed preceding gastrulation ; but invagination of ento-
derm occurs in essentially the same way as in the sea urchin, though the
most intense activation of the entoderm apparently takes place after in-
vagination (p. 137). Even in the sea urchin presence of micromeres or of
primary mesenchyme, resulting from reconstitution of other parts of the

3 Von Ubisch, 1931, 1932(7,6, 1934, igi^^h; Horstadius, iq^ed; F. J. Schmidt, 1936.

446 PATTERNS AND PROBLEMS OF DEVELOPMENT

embryo, is apparently not necessary for invagination. Stained cells from
vegi, implanted in the apical poles of whole embryos, may invaginate with-
out reconstituting primary mesenchyme; but a part of the primary mesen-
chyme of the host aggregates about the invaginated cells (Horstadius, 1935) .
In view of all the evidence, not only that from the transplantation ex-
periments but also that from differential dye reduction, differential sus-
ceptibility, and the differential modifications of development resulting
from it, it may still be questioned whether induction by implanted micro-
meres is primarily anything more than a nonspecific effect depending on
physiological condition rather than on specific differentiation; but even
if the micromeres produce a specific substance, its effect may be primarily
activation. In his interpretations of experimental results Horstadius uses
terms implying dynamic factors. He speaks of the "conflict" of the gra-
dients, the "strength" or "weakness" of one or the other, the "suppres-
sion" of one by the other, the "weakening" of one, permitting the other
to become stronger and gain the upper hand. In terms of primarily non-
specific gradients of rate or intensity sea-urchin development appears not
as a conflict of gradients but rather as an orderly and definitely deter-
mined sequence of gradient changes which can be altered in definite ways
by changes in relations of parts and in environment.

FORMAL PATTERNS OF VERTEBRATE DEVELOPMENT

Before discussing induction in vertebrate development some considera-
tion of the formal regional maps and of the cell movements or migrations
by which parts of the early embryo attain their final positions is neces-
sary. In the study of embryonic development it soon became evident,
for vertebrates as for invertebrates, that observation and description of
development, necessary though it was as a first fine of attack, could give
us httle more than a timetable of events as a basis for discussion and
speculation. The observation that localization of the embryo in a par-
ticular region involved considerable changes in position and apparent mi-
gration of cell materials raised the question of the fate in development,
the prospective significance, of different regions; and the theory that em-
bryo formation resulted in large part from concresence of two lateral
halves was advanced and debated pro and con. In the attempt to de-
termine more exactly the fates of different regions methods of local mark-
ing by puncture or other local mechanical injury, insertion of hairs, local
cautery, and, later, local electrolytic injury and radiation were employed.
To all these the objection has been raised that the injury or obstruction
resulting from it may interfere with the cell migrations. The method of

EMBRYONIC INDUCTORS AND ORGANIZERS

447

local vital staining by application to the surface of egg or embryo of small
pieces of agar impregnated with dye^ is open to the objection only that
the dye may be, to some extent, toxic; but results obtained thus far indi-
cate that, as the method is used, toxicity is not sufficient to constitute
a serious objection to it.

This method has made it possible to map more or less accurately pro-
spective or presumptive embryonic regions and to follow the cell migra-
tions that bring them to their definitive positions. Such maps have, in
general, little more than formal significance, since isolation and transplan-

FiG. 149, A , B. — Map of prospective or presumptive urodele embryonic regions at beginning
of gastrulation. A , lateral, B, dorsal view. Denser broken lines, neural plate; less dense broken
lines, general ectoderm; coarse stippling, notochord; fine stippling, mesoderm; /, beginning of
invagination; b, later blastopore; //, limit of invagination; p, basal pole; g, gills; /, lateral meso-
derm; pi, pronephros and mesoderm of forelimb; I, caudal region; i-io, somites (after Vogt,
1929).

tation experiments have made it clear that identification of a particular
embryonic area, as giving rise in normal development to a particular
organ or organ system, gives no information concerning its potentialities
or potencies. It shows only that, under the conditions which we call "nor-
mal," a certain region of the egg or early developmental stage becomes a
certain part.''

The accompanying amphibian, teleost, and chick maps indicate formal
regional pattern and regional migrations, as determined by local vital
staining. The amphibian maps, as given by Vogt (1929), are the most
complete. Figure 149 shows the urodele regional map at the beginning

■•Vogt, 1925; 1926a, b; 1929.

5 The following papers are concerned with this problem of formal pattern. Fishes: Kopsch,
1927; Oppenheimer, 1936c; Vandebroek, 1936a; Pasteels, 1936a. Amphibia: Vogt, 1925,

448

PATTERNS AND PROBLEMS OF DEVELOPMENT

of gastrulation, viewed laterally (A) and from the basal pole (B), Fig-
ure 150, the map of the early anuran blastula, viewed laterally (A) and
dorsally (B). The general similarity is evident, the chief difference be-
tween the two being that in the anuran the regions of the axial organs do
not extend so far toward the apical pole as in the urodele, the greatest
difference being in the neural plate. In Figure 151 directions of cell mi-
grations (A and B) are indicated as they take place in gastrulation. With

Fig. 150, A, B. — Regional map of early anuran blastula. A, lateral, B, dorsal, view;
denser broken lines, neural plate; less dense broken lines, general ectoderm; coarse stippling,
notochord; fine stippling, mesoderm; a, apical, v, basal, pole; /, beginning of invagination; b, later
blastopore; //, limit of invagination; g, gills; //, posterior limit of head ectoderm; mf, medullary
(neural) fold; e, field of eyes and chiasma; s, sucker; /, lens; an, auditory vesicle; pi, pronephros
and mesoderm of forelimb; somites indicated by lines in mesoderm (after Vogt, 1929).

approach of gastrulation there is an increase in area and a stretching,
particularly in the apicobasal direction, in cells of the apical hemisphere,
apparently greatest in the dorsal region; and gastrulation begins with
the "rolling in," the invagination of cells in the median dorsal region at
the boundary between the future chorda-mesoderm and yolk. This in-
vagination extends laterally and ventrally until the blastopore becomes

19260, b, 1929; Pasteels, 1932. Birds: Assheton, 1896; Peebles, 1898, 1904; Wetzel, 1925a, b,
1929, 1931, 1936; Kopsch, 1926a, b, 1934a, b; Graper, 1929; Pasteels, 19266, 19370; Jacobson,
1938 . For citation and discussion of the literature, including earlier papers concerning con-
crescence, gastrulation, and vertebrate developmental pattern in general, see the above
authors, particularly Vogt, 1929; Wetzel, 1931; Pasteels, 1937a; also Holmdahl, 1925, 1926,
1933, 1935; and textbooks of vertebrate embryology.

EMBRYONIC INDUCTORS AND ORGANIZERS

449

circular, gradually decreases in diameter, and overgrows the yolk. The
invaginated cells migrate anteriorly beneath the ectoderm and give rise
to notochord and mesoderm, and continued increase in area of the super-
ficial cells brings more material to the blastopore lip and so to invagination.
Isolation and transplantation experiments have brought to light various
facts of interest concerning these regional changes in position of cellular
material. Vogt (1923) maintains that they are not the summation of ac-
tivities of single cells but essentially regional amoeboid movements in

A B

Fig. 151, A, B.— Directions of cell migrations in amphibian gastrulation. A, lateral, B,
dorsal view; alternation of broken and unbroken lines merely for ease in following directions
(from Vogt, 1929).

which the individual cells are passive. Pieces of the future blastopore lip
transplanted to other regions undergo the same "stretching" and invagi-
nation as in normal environment, even when abnormally oriented with
respect to the region in which they are implanted, and different regions
of the blastopore lip stretch and invaginate independently. When invagi-
nation is prevented in transplanted pieces or united halves of embryos,
hornlike or irregular outgrowths arise. Also, stretching and invagination
may occur independently of each other. According to Holtfreter, how-
ever, the abihty to stretch is intrinsic in individual cells; isolated blastula
cells undergo the change of shape in the same way as larger cell groups.'^

Thus far no physiological interpretation of these changes in shape and
migrations of cells has been offered, but it is perhaps of interest to note

6 0. Mangold, 1920; Vogt, 1922&; Spemann und H. Mangold, 1924; Spemann, 1931, 1936,
pp. 63-71, 1938, pp. loi-io; Holtfreter, 1939'^-

4SO PATTERNS AND PROBLEMS OF DEVELOPMENT

that they apparently occur first and are greatest in those regions of the
embryo which show the highest differential susceptibility, the most rapid
dye reduction, and perhaps the highest respiration (pp. 151-58), and the
stretching is greatest in the longitudinal direction. These relations sug-
gest that they are related to the gradient pattern of the stages concerned.
It appears difficult to account for stretching in a particular direction unless
there is a differential of some kind in this direction, affecting individual
cells. The isolated and transplanted pieces, or even individual cells, rep-
resent a part of this differential and behave accordingly. As Vogt and
Spemann have pointed out, the movements constitute an orderly, defi-
nite series of events which determine a definite result. If the cells were all
alike and there were no regional directive factor, this would be impossible.
Such evidence as is available indicates that the cells initiating invagina-
tion in the intact embryo represent the highest levels of the gradient
system of the dorsal lip region. In the absence of this region in trans-
planted pieces the highest gradient-level present apparently initiates in-
vagination. Increase in surface area of cells occurs in early embryonic
development of many other forms and apparently begins in the high
regions of the gradient pattern present. For example, the ectoderm of
the echinoderm blastula and gastrula undergoes increase in surface area,
beginning apically and progressing basipetally, and the developmental
activation of the entoderm is followed by increase in its surface area. The
relations of these changes to gradient pattern are evident; and when the
gradient pattern is experimentally altered or obliterated, they are also
altered.

In the amphibian the regional differences in these cell activities appear
to be quantitative rather than limited to particular regions, but they
may be regarded as representing a certain stage of development from the
primary gradient pattern, this stage being attained first and in greatest
degree by cells of the higher gradient-levels. Whether, or to what ex-
tent, the cells concerned have become specifically different from others
which attain the condition later or not at all remains to be determined.
The autonomous invagination of pieces of the blastopore lip suggests a
"determination" of some sort. The factors determining invagination are
unknown, but, in general, gastrulation appears to represent a certain
stage in the progressive complication of the earlier gradient pattern.
This seems particularly clear in echinoderm development.

The regional map of the teleost blastoderm, as determined by Pasteels
(1936a), is shown in Figure 152, A; and the directions of cell migrations

EMBRYONIC INDUCTORS AND ORGANIZERS

451

in relation to gastrulation, in Figure 152, B. Maps by Oppenheimer
{igs6d) are very similar. Regional maps of the chick blastoderm by
Graper (1929), Wetzel (1929), Waddington (1932), Pasteels (19366), and
Jacobson (1938) are, in general, much ahke but differ in certain minor
points which cannot be considered here. Maps of four stages by Pasteels
are shown in Figure 153, A-D; and in Figure 154, A-D, directions of cell
migrations are indicated. Superficial migrations in formation of the primi-

A

Fig. 152, A, B.—A, map of prospective embryonic regions of teleost blastoderm just pre-
ceding gastrulation; B, directions of cell migrations in gastrulation. Vertical lines, prechordal
plate; lightly stippled, lateral and ventral mesoderm; heavily stippled, caudal unanalyzed
regions; b, brain; t, trunk ectoderm; c, notochord; ce, cephalic ectoderm; e, extraembryonic
ectoderm; n, postcephalic nervous system; s, somites (from Pasteels, 1936a).

tive streak, as determined by Wetzel, are shown in Figure 155, A-C.
Jacobson beheves that the longitudinal migration recorded by these au-
thors does not take place. In spite of the general similarity of results.
there has been disagreement as regards significance of the primitive streak.
The more generally accepted view is that the streak represents a part
of the gastrulation process in which mesoderm is invaginated, entoderm
being invaginated earlier from the posterior border of the blastoderm
(e.g., Patterson, 1909). Wetzel, however, maintains that the streak has
nothing to do with gastrulation.

It has become evident that the formal regional pattern and the migra-
tions concerned in embryo formation are, in general, similar in all verte-
brate groups. Even mammalian development, although it exhibits certain
features associated with intrauterine environment, evidently does not dif-

452

PATTERNS AND PROBLEMS OF DEVELOPMENT

fer fundamentally in general pattern of embryo formation from that of
other vertebrates. Moreover, pattern of embryo formation in Amphioxus
and ascidians shows certain general resemblances to that of vertebrates.

B

Fig. 153, A-D. — Regional map of chick embryo at four stages; dotted area, lateroventral
mesoblast; oblique or vertical lines, somite region; coarse stippling, cephalic mesoblast; hori-
zontal lines, notochord; neural region indicated by line marking its anterolateral boundary,
.1, early blastoderm; B, early primitive streak, parallel dotted lines indicating streak; C J).
later stages, showing superficial map on left side, map of invaginated parts on right (after
Pasteels, 19366).

As already pointed out, the regional patterns indicated by the maps
give no information concerning determination or differentiation of the
regions distinguished. At best they are maps of future embryonic parts.

EMBRYONIC INDUCTORS AND ORGANIZERS

453

The regions mapped are not, in general, coextensive with embryonic fields.
Usually they are more limited than the fields, that is, they show the
regions which will actually develop into certain organs, not the regions in
which development of those organs is possible. Physiologically, such a
map may represent nothing more than a quantitative gradient pattern,

Fig. 154, A-D. — Cell migrations in chick embryo at four stages, according to Pasteels;
arrows drawn in unbroken line indicate movements involving both cell layers; those in broken
line, movements of surface layer only (after Pasteels, 1936&).

the developmental fate of the regions being determined by their position
in the gradient system. In other words, the regional maps of early stages
do not represent actual patterns but are projections of the patterns of
later stages back onto a stage in which the actual pattern is simpler and
more general in character. It is not necessary to think of the various parts
of the embryo as spread out on the blastula or blastoderm, as the maps

454

PATTERNS AND PROBLEMS OF DEVELOPMENT

seem to indicate. The parts are not necessarily present in these stages,
but the future pattern is projected on a present pattern, from which it
originates. The cell migrations constitute the basis for projection. That
dynamic factors of some sort play a part in determining and directing
these movements seems probable, and the possibility that electric-poten-
tial gradients may be concerned in determining the differential in the
cells suggests itself. An anteroposterior potential gradient is reported in
Ambly stoma and chick embryos, with increase in potential difference with
progress of development and demonstrable without direct contact with
the embryo. Also, the dorsal lip region of Amhlystoma is electronegative

Fig. iss, .l-C— Cell migrations in chick embryo, according to Wetzel; change in position
of embryonic regions indicated by gradations of stippling. A, directions in which migrations
will occur in formation of primitive streak; B, C, later streak stages (from Wetzel, 1929).

to the apical pole. The developing embryo apparently gives rise to a
steady-state electrodynamic field with definite axiate and regional inten-
sities (Burr and Hovland, 1937^, h).

The recent studies of cell migrations by means of local staining have
finally shown that the much debated concrescence theory of vertebrate
development, that all except the most anterior parts are formed by union
of two originally separated lateral halves, is not entirely in accord with
the facts. The axial organs are primarily median, but there is concres-
cence or convergence of certain lateral areas toward the median plane.''

NEURAL INDUCTION IN AMPHIBIAN DEVELOPMENT

The work on induction and so-called "organizers" in amphibian de-
velopment has attracted much attention and has become widely known

7 Sec the discussion by Vogt, 1929, pp. 668-78.

EMBRYONIC INDUCTORS AND ORGANIZERS 455

through its consideration in handbooks, textbooks, general reviews, and
addresses.^

The induction which has aroused most interest and received most at-
tention, supposedly the earliest induction in amphibian development, is
the induction of the neural plate by the dorsal region which invaginates,
underhes the ectoderm which becomes neural plate, and itself becomes
chorda-mesoderm. Stages of invagination of chorda-mesoderm and ento-
derm and the neurula, with neural plate formed after invagination, are
shown in Figure 156. As is now well known from the work of many in-
vestigators, a piece of the presumptive chorda-mesoderm region (Figs. 149,
150), transplanted before its invagination to a region not normally in-
volved in neural-plate formation, or implanted in the blastocoel of another
embryo, can induce a new neural plate in ectoderm of the host which
would normally form only epidermis (Figs. 157, 158). The secondary em-
bryo thus induced was at first regarded as a product of the implant rather
than an induction (Spemann, 19 18), but heteroplastic transplantations
from the dorsal lip of the unpigmented Triton cristatus to the pigmented
T. taeniatus made it evident that the neural plate developing in relation
to the implant was wholly, or almost wholly, derived from host tissue,
while the underlying chorda-mesoderm was derived from the implant.^

Transplants of presumptive ectoderm had no such inducing effect but
were incorporated and developed according to their position in the host,
although preserving their pigmentary species-characteristics. The forms
resulting from these heteroplastic transplantations are chimeras, consist-
ing of tissue of two species. These experiments led Spemann to regard the
inducing region as an "organizer" or "organization center." Only some
of the more important points in the further investigation of various as-
pects of this induction by Spemann and many others are considered here.

* See the following books: Morgan, 1927, a brief account; fuller consideration with ex-
tensive bibliographies in Schleip, 1929; Huxley and De Beer, 1934; Dalcq, 1935, on chordate
egg organization in general, with bibliography; Spemann, 1936, 1938, lectures concerned with
the problem of induction in embryonic development, with bibliography; Weiss, 1939, a general
textbook of experimental embryology; also the most recent book on the subject, Waddington,
1940, Organizers and Genes; a consideration, from the viewpoint of the Cambridge group, of
evocators and organizers in vertebrate, and chiefly in amphibian development, with discussions
of competence, individuation, and organization in general, and an attempt to bring genes
into the picture, but without anything new in the way of synthesis or further light on the
problems concerned. Numerous reviews of the subject have also appeared in Ergehnisse,
Jahresberichte, and review journals, notably those of Mangold, 1928a, 1929a, 1931a, with
bibliographies. See also De Beer, 1927; Gilchrist, 1929c; Weiss, 1935.

'Spemann, 1921; Spemann und H. Mangold, 1924. See also Marx, 1925; Geinitz, 1925a.

Fig. 156, A-H. — Amphibian gastrulation and neural plate formation; diagrammatic;
chorda-mesoderm indicated by deeper, ectoderm by lighter line-shading, entoderm by cell
boundaries. A~E, median sections; F, same stage as E in transverse section; G, H, neurula in
dorsal view and in transverse section (from Spemann, 1936, Experimentelle Beitrdge zu eitur
Theorie der Entwicklung. Springer, after chart by V. Hamburger and B. Meyer; also Spemann,
1938)-

EMBRYONIC INDUCTORS AND ORGANIZERS

457

The earlier experiments were on urodeles, chiefly Triton; but work with
anura and heteroplastic and xenoplastic transplantations between dif-
ferent urodele species, between anuran species, and even between urodeles
and anura have shown beyond question that the presumptive chorda-
mesoderm acts as inductor in amphibians generally and that its action is
not species-, genus-, or even order-specific; but tissues and even embry-

FiG. 157, A-C. — Induction of a secondary neural plate and embryonic axis in ectoderm by
implantation of dorsal lip material. A, B, two views of Triton taeniatns neurula with secondary
neural plate induced by heteroplastic implant from dorsal lip of T. cristatns; C, later stage of
same embryo with secondary embryo on left side (from Spemann und H. Mangold, 1924).

onic extracts of some forms are toxic for some other species.'" Since ac-
tion of the living dorsal inductor, is, to such an extent, independent of
species, the question at once arises as to the nature of its action, whether
material or dynamic. This question has been raised and discussed by

" Geinitz, igasa; Bytinski-Salz, 1929a, b, c; H. Mangold, 19296; O. Mangold, 1929&;
Schotte, 1930; Raven, 1931, i933&,- G. A. Schmidt, 1933, 1936a, b; Holtfreter, 1935a, 6, 1936.
See also Twitty, 1937, "Experiments on the phenomenon of paralysis produced by a toxin
occurring in Tritiints embryos," Jour. Exp. ZooL, 76; Horsburgh, Tatum, and Hall, 1940,
"Chemical properties and physiological actions of Trikirtis embryonic toxin," Jour. Pharmacol.
E.xp. Therapeutics, 68.

458

PATTERNS AND PROBLEMS OF DEVELOPMENT

Vogt, Spemann, Mangold, and more recently by others. But quite apart
from this question, the lack of species-specificity suggests two possibilities
in this induction: production by different species and even by urodeles
and anura of the same, or closely similar, substances, with essentially the
same specilic effects; or an activating action dependent on the high physio-
logical level of the inductor and nonspecific. In other words, does the
inductor determine wholly or in part the specific character of the neural

Fig. 158. — Transverse section through middle region of embryo shown in Fig. 157, D.
Primary embryo at left, secondary induced embryo at right of figure: fn, primarj^ neural tube;
in, induced neural tube; sc, secondary notochord; 5.y, secondary somite; sp, secondary pro-
nephric duct; sg, secondary gut. The implant, shown in lighter shading, forms the secondary
notochord and part of the right somite (after Spemann und H. Mangold, 1924).

plate, or does it bring the ectoderm of a particular developmental stage to
a physiological level at which development of a neural plate takes place?
Consideration of further data must precede discussion of this question.
Both capacity of presumptive chorda-mesoderm to induce and of ecto-
derm to react change in the course of development. Exactly when induc-
tive capacity appears in the living embryonic inductor is diflicult to de-
termine because pieces from earlier pregastrula stages implanted in other
embr^^os continue to develop and may induce, and ectoderm is developing
at the same time. The reactivity of ectoderm gradually decreases in later
gastrula stages (Holtf refer, 1938a).

EMBRYONIC INDUCTORS AND ORGANIZERS 459

THE LATERAL DIFFERENTL\L IN INDUCING CAPACITY

Inducing capacity of the dorsal lip region before invagination decreases
laterally from the median region, and the same is true of the inductor
after invagination; the somite region is less effective than the chorda
primordium, and the lateral mesoderm still less effective." The lateral de-
crease in ability to induce neural plate is steeper in anura than in urodeles;
lateral and latero ventral parts induce tail, not neural plate (G. A.
Schmidt, 1936a, b; Schechtman, 1938). There is, in short, a gradient in
inducing power, decreasing from the median region laterally; susceptibil-
ity of the dorsal lip region shows an essentially similar gradient (p. 152).
In this connection an experiment of Waddington's (1936a) is interesting.
He finds that transplanted, newly formed lateral-plate mesoderm is able
to induce an extra neural plate from presumptive ectoderm. Median dor-
sal lip region, substituted for presumptive lateral plate, also induces an
extra neural plate; but lateral mesoderm does not induce a neural plate
in normal development. These results suggest a relation of dominance and
subordination between higher and lower levels of the mediolateral gra-
dient of the presumptive chorda-mesoderm, the high median level being
dominant. When isolated from this dominance, the lateral region is able
to induce, perhaps because of some degree of activation following isola-
tion, as in invertebrate reconstitution. Lateral symmetry or asymmetry
of the inductor does not determine symmetry or asymmetry of the in-
duced plate. A right or left half of the dorsal lip region induces a. whole
neural plate, not a right or left half; and median or lateral pieces less
than half also induce a whole plate, though not always a completely sym-
metrical one. The implanted lateral half of the inductor from an early
gastrula reconstitutes to a bilaterally symmetrical system with median
chorda. Halves from more advanced stages show less reconstitution, the
notochord developing more or less at one side, and the other side being
formed by more or less appropriation (induction) of host tissue. The later
the stage of the implant, the more asymmetrical is the induced neural
plate.'' These facts again suggest that the mediolateral differences are pri-
marily quantitative gradient differences but become increasingly specific
with progress of development.

" Bautzmann, 1926, 1928, 19296, 1933. See also Ruud, 1925; Mangold und Seidel, 1927;
and various other papers include evidence on this point.
'^ Spemann, 1918; Weber, 1928; B. Mayer, 1935.

46o PATTERNS AND PROBLEMS OF DEVELOPMENT

THE LONGITUDINAL DIFFERENTIAL IN INDUCTOR
AND IN REACTING ECTODERM

The region of the dorsal inductor lying nearest the level at which blasto-
pore formation begins, and invaginating first, migrates anteriorly after
invagination and finally comes to underlie the future head region, while
parts which invaginate later underlie successively more posterior regions.
In the attempt to determine whether a longitudinal differential or specific-
ity in inducing capacity exists, pieces of the dorsal lip at different stages
of gastrulation have been implanted. Pieces from the earliest stages of
gastrulation implanted at the level of the future head induce neural plates
with well-developed head region, but when implanted at trunk-levels also
induce well-developed heads. Pieces from the dorsal lip of later stages
implanted at head-levels may also induce neural plates with anterior
head region approaching normal; but when implanted at more posterior
levels, the anterior region of the induced neural plate is deficient or absent.
Pieces of the dorsal inductor invaginating last and lying farthest posteriorly
after invagination may induce tail.'-' These results have led to the desig-
nation of different levels of the inductor as "head-inductor," "trunk-
inductor," and "tail-inductor." Essentially similar results have been ob-
tained with implants of different regions of the chorda primordium alone
(Bautzmann, 1928, 1929&). However, as Spemann has pointed out, the
longitudinal differences in results depend not only on regional difference
in the inductor but also on a difference in the reacting ectoderm, for
"trunk-inductor" implanted at head-level induces head more or less com-
pletely and is completely trunk-inductor only at more posterior ectoder-
mal levels.

Before invagination the physiological gradient in the inductor decreases
anteriorly from the region invaginating first, and in the ectoderm of the
presumptive neural plate and in other ectoderm there is a gradient differ-
ential decreasing from the region about the apical pole (pp. 151-58).
After invagination the two gradients coincide in direction and more or
less closely in extent, in the region developing as neural plate, the one
underlying the other. When the high gradient-level of the inductor is
implanted at the high ectodermal level, a well-developed head results;
but the high inductor level is also able to induce more or less complete
head development at lower levels of the ectodermal gradient, and a lower
inductor level may induce head when implanted at the high ectodermal

'3 Spemann, 1931, 1936, pp. 167-74, 1938, chap, xiii; Holtfreter, 1933^, 1936; Bytinski-
Salz, 1931; O. Mangold, 1932^; Lehmann, 1932; Schechtman, 1938.

EMBRYONIC INDUCTORS AND ORGANIZERS 461

level. Are the longitudinal differences in induction and reactivity prima-
rily anything more than these gradient differences?

The question of longitudinal regional specificity of the inductor has
been much discussed. The chorda-mesoderm is regarded as a mosaic of
specific inductors by Holtfreter, and Lehmann believes he has obtained
evidence of regional specificity by the use of LiCl.'^ In fact, Lehmann
holds that action of chemical agents on embryonic developmental stages
is, in general, regionally specific. As regards his experiments, however,
it must be noted that the possibilities of differential recovery and of altera-
tion of nonspecific differential susceptibilities in the course of development
by alteration of activity are completely ignored.

This hypothesis of longitudinal regional specificity in the inductor pre-
sents difficulties. Neither inductor nor presumptive neural plate gives any
evidence of specific regional differences in earlier stages, though this does
not prove their absence. Moreover, different levels of the central nervous
system all differentiate into cells and fibers; the specificity of the different
levels appears to be in the relations of these rather than a chemical speci-
ficity, but how the inductor can determine these relations does not appear.
Head-inductor and trunk-inductor induce trunk or head according to level
of implantation, and head is not present or absent but shows various de-
grees of gradation between complete development and absence. This does
not suggest specificity of induction. Head and trunk, of course, become
widely different in later stages, and various other inductions occur in
their development; but that the neural inductor is anything more than a
factor in determining relative physiological levels in the reacting ectoderm
remains to be proved. The regional differences in the inductor are re-
garded as quantitative by Dalcq and Pasteels.'^

The inductor region may be specifically different from other regions
of the egg cytoplasm at the beginning of development; but, as will appear,
a great number of tissues, living and dead, tissue extracts, and chemical
substances are also neural inductors; consequently, any close relation be-
tween whatever specificity may be present in the natural inductor and
its inducing capacity does not appear probable. According to Holtfreter
(1939a), there is complete lack of specificity in aggregations of isolated
amphibian cleavage cells, but with isolation at later stages pure ectoderm
and pure entoderm may aggregate.

'■* Holtfreter, 1938, and various earlier papers; Lehmann, 1936a, b, 1937a, c, 19386, c.

'5 Dalcq, 1935, 1938a; Dalcq et Pasteels, 1938; Pasteels, 1938. See also Pasteels, 1938,
"Recherches sur les facteurs initiaux de la morphogenese chez les Amphibiens Anoures. I,"
Arch. Biol., 49.

462 PATTERNS AND PROBLEMS OF DEVELOPMENT

IS INDUCTION NECESSARY FOR NEURAL-PLATE FORMATION?

From his experiments Goerttler (1925, 1926, 1927) concluded that
orientation of an inductor implant may be favorable or unfavorable, ac-
cording to its relation to the regional cell migrations; consequently, he
regarded these migrations as factors in determination of the neural plate
and the presence of underlying chorda-mesoderm as not absolutely neces-
sary. Holtfreter (19330, b) found orientation without effect and con-
cluded that induction, not the migrations, is the essential factor in de-
velopment of the neural plate. However, Goerttler has described and fig-
ured cases in which more or less development of neural plate and neural
tube occurred when the whole dorsal lip region had been removed before
its invagination. According to Lehmann (1926, 1928, 1929), there is a
quantitative relation between degree of defect in chorda-mesoderm, re-
sulting from experimental removal of part of the dorsal lip region before
invagination, and degree of development of the overlying region of the
neural plate; but this relation differs at different levels. Anteriorly only
the most extreme defects in the inductor determine defects in neural
plate, but farther posterior the neural plate is less independent. Lehmann
concludes that there is labile determination of the neural plate at the
beginning of gastrulation, decreasing from the anterior region posteriorly,
and therefore independent of underlying chorda-mesoderm.

Isolation of parts of the embryo before, or at beginning of, gastrulation
give different results. Apical regions isolated in water never develop neu-
ral plate, nor do two such regions unite (Spemann, 1936, p. iii). Cul-
tured in balanced salt solutions, entoderm and mesoderm show a high
degree of independent differentiation (W. Erdmann, 1931; Holtfreter,
igsia, b), and Erdmann found that presumptive neural plate similarly
cultured develops into neural tissue, both in urodeles and anura; but
Holtfreter maintained that it always forms epidermis, with at best only
minute traces of neural tissue. However, he did obtain neural and other
tissues from a group of four cells from the lower part of the apical hemi-
sphere, isolated by destroying the other cells and developing in salt solu-
tion, surrounded by the debris of the killed cells. Erdmann and Holt-
freter agree that presumptive neural plate cultured in explant does not
give rise to chorda-mesoderm. Material of the dorsal lip shows a con-
siderable capacity for reconstitution; an explanted half may give rise to
a bilaterally symmetrical complex, consisting of a chordal strand with
muscle and, in the anterior region, neural tissue and epidermis, that is,
ectodermal derivatives from presumptive mesoderm (Holtfreter, 1933c).

EMBRYONIC INDUCTORS AND ORGANIZERS 463

Explanted presumptive entoderm develops as entoderm and may attain
a high degree of differentiation.

Various eariier attempts to culture isolated parts of amphibian blastu-
lae and gastrulae in organic media proved unsuccessful, but it was dis-
covered that pieces introduced into the coelom of an older larva would live
and develop. Development of these coelomic cultures differs in certain
respects from that in salt solutions, as regards presumptive epidermis and
neural plate (Holtfreter, 1929, 1931a). Presumptive neural plate may de-
velop either into pure nervous tissue or into epidermis with practically equal
frequency; and presumptive epidermis, into epidermis with differentiation
of gland, ciliated, and pigment cells or into nervous tissue. Implantation
in the eye cavity of older individuals after removal of the eye again gives
different results; chorda-mesoderm may develop from almost all regions
of the early gastrula of Triton (Kusche, 1929), even from the region of the
apical pole in Amhly stoma (Bautzmann, 1929a).

Developing eggs of the urodele Tritums, subjected to a lateral tempera-
ture gradient, may develop on the warm side thickenings resembling neu-
ral tissue in structure from regions of presumptive epidermis not under-
lain by chorda-mesoderm and typical neural folds from regions with un-
derlying mesoderm (Gilchrist, 1928, 1929a, 1933). Other experiments with
localized high temperatures have not produced clearly identifiable neural
tissue (Castelnuovo, 1932; Margen and Schechtman, 1938). Inhibition of
one-half of the embryo by low temperature or by lack of oxygen results in
asymmetry and alteration in developmental fate of various parts. It
shows, further, that different regions of the neural plate may develop in-
pendently of each other and that the anterior head region may develop
when gastrulation is almost completely suppressed (Vogt, 1927; 1928a, b).

Naked embryos (axolotl) developing from early blastula stage in a
modified Ringer solution undergo exogastrulation, the mesentoderm evag-
inating instead of invaginating and in the complete exogastrula never
underlying any part of the ectoderm (Holtfreter, 1933^, e). Directions of
cell migrations in exogastrulation are indicated in Figure 159, A, B; the
differentiation, in C. Ectoderm of these exogastrulae forms an irregular
cell mass, not a definite layer, and shows no traces of neural plate or
neural tissue and no evidence of any definite developmental pattern. It
may become ciliated, but the ciliary beat is not definite in direction, as in
the normal animal. The evaginated mesentoderm, on the other hand,
undergoes a high degree of differentiation into notochord, muscles, kidney
tubules, and gut and may approach the normal embryo in general form.

464

PATTERNS AND PROBLEMS OF DEVELOPMENT

but is, of course, inside out (Fig. 159, C; Fig. 160). All gradations between
complete exogastrulation and normal embryos occur. In partial exogas-
trulae evagination is followed by more or less invagination; and, if this
is sufficient to bring chorda-mesoderm beneath ectoderm, neural tissue
may develop, varying in form and size with form and size of invaginated
chorda-mesoderm and differentiating into the more posterior parts of the
neural tube, or only the caudal portion when only a small part, represent-
ing the posterior portion of the chorda-mesoderm, invaginates. Pieces of

Epldfixmb

Kopfmesoderm
Munddarm

Fig. 159, A-C. — Exogastrulation in axolotl. A, B, directions of regional migrations; C,
diagrammatic longitudinal section, indicating differentiation of mesentoderm and lack of
differentiation of ectoderm. Parts indicated: coelom, intestine {Diinndarm) , pronephric
tubule {Nierenkanal), trunk muscles (Rumpfmuskel), notochord {Chorda), pharynx (Kienien-
darm), head mesoderm {Kopfmesoderm), oral entoderm {Munddarm), rectum {Enddarm),
epidermis (from Holtfreter, i933e, Biol. Zbl., 53, H. 7, 8, Thieme).

ectoderm from early gastrulae, transplanted to the surface of the evagi-
nated chorda-mesoderm, differentiate according to the region of the in-
ductor on which they lie. On the head inductor they may develop into
brain, with eyes, olfactory pits, and ear vesicles; on the trunk inductor
they develop spinal cord; and at levels still farther posterior a tail may
develop, with growth of chorda-mesoderm into the ectoderm and induc-
tion of a neural tube. These exogastrulae are regarded by Holtfreter as
providing complete proof that there is no determination in the ectoderm
before invagination of the chorda-mesoderm and, consequently, that its
development is entirely the result of induction in the normal individual.
This conclusion ignores, or regards as erroneous, evidence from earlier

EMBRYONIC INDUCTORS AND ORGANIZERS 465

work which is not in accord with it, as Huxley and De Beer (1934, P- 493)
have pointed out; and there are other points to be considered. The fact
that the presumptive neural-plate region of the exogastrulae undergoes
more stretching than presumptive epidermis indicates a difference of some
sort, though it may be a nonspecific gradient difference. Moreover, the
Ringer solutions that bring about exogastrulation may inhibit develop-
ment to some extent; various Ringer modifications inhibit planarian re-
constitution. If this is the case, the exogastrulae do not provide the final
proof of complete dependence of neural development on induction.

The results with presumptive epidermis and neural plate cultured in
salt solution or in vivo in the coelom or eye cavity show that a piece of

Fig. 160. — An exogastrula with mesentoderm (left) somewhat similar in form to an embryo
and connected with the irregular ectodermal mass (right) only by a slender strand (from
Holtfreter, 19336, Biol. Zbl., 53, H. 7, 8, Thieme).

either may develop as epidermis, neural tissue, or chorda-mesoderm.
Holtfreter regards the eye cavity as having complex inductive capacities,
but the possibility is not yet excluded that the primary differences in
these tissues are nonspecific. The doubtful results of the temperature ex-
periments are not conclusive evidence against this view. In developing as
neural tissue ectoderm apparently undergoes a rise in physiological level,
relative to the other ectoderm, and apparently a further rise is necessary
for development as chorda-mesoderm. The isolated pieces, implanted in
coelom or optic cavity, are isolated from the gradient pattern of the em-
bryo and apparently under favorable conditions. Development of chorda-
mesoderm from almost all regions of the early gastrula, implanted in the
optic cavity, does not seem to be very different in principle from develop-
ment of a hydranth or head from any level of the hydroid or planarian

466 PATTERNS AND PROBLEMS OF DEVELOPMENT

body when it is isolated from higher levels of the gradient pattern and
sufficiently activated. In the temperature experiments the whole gra-
dient pattern of the parts subjected to the higher temperature is elevated
to a higher level, but presumably there is little or no change in the rela-
tions of parts and consequently little or no alteration in the course of
development.

The conclusion of Goerttler and Lehmann that the anterior region of
the presumptive neural plate is more capable of independent differentia-
tion than more posterior levels is in accord with the evidence concerning
gradient pattern. This region is the region about the apical pole, the high
end of the primary gradient. Under certain conditions its gradient-level
is apparently high enough, relative to other parts, to permit independent
development as neural plate or neural tissue, while at more posterior
levels there must be activation, relative to surrounding ectoderm, either
by the inductor or, in isolated pieces, by environmental conditions.

Vogt and Spemann agree in admitting a Bahnung toward determina-
tion in the presumptive neural plate, independent of the invaginated in-
ductor. But that the determination of presumptive epidermis or neural
plate has not become highly specific and fixed, even in the gastrula, is
indicated by other results of transplantation. Presumptive ectoderm at
the beginning of gastrulation, implanted in presumptive mesodermal or
entodermal regions, becomes mesoderm or entoderm (O. Mangold, 1923).
Transplanted to dorsilateral regions of older embryos, both presumptive
epidermis and neural plate may develop, the neural part ranging from
brain with nose and eye to posterior parts of the spinal cord, according to
level of implantation. It may also take part in balancer formation or in de-
velopment of gill or leg, and its deeper layers may become mesoderm. In
the pronephric region it may give rise to pronephric tubules; in contact
with muscle it may form muscle, this most frequently at more posterior
levels, or notochord, or even gut wall (Holtfreter, 19336).

The relative significance of gradient-levels as such and of specific re-
gions or fields. in bringing about these inductions is not known. The ques-
tion whether tissue of presumptive neural plate can become neural tissue
independently of the inductor and of other parts of the embryo appears
to have been finally settled by the experiments of Barth (1939c). Explants
of presumptive neural plate with presumptive epidermis from stages of
Ambly stoma punctatum in which these tissues still form the roof of the
blastocoel and before the inductor underlies the presumptive neural re-
gion, when fused by their anterior ends, form neural tubes in the region

EMBRYONIC INDUCTORS AND ORGANIZERS 467

of fusion. Fusion of the anterior half of such an explant with the anterior
end of a whole explant gives similar results. Lateral fusion with antero-
posterior axes in opposite directions does not give neural tubes. More-
over, a single explant can develop neural tube if lateral edges are brought
together to form a tube having an anteroposterior axis. If anterior and
posterior margins are united, a neural tube rarely develops. Barth holds
that neural-tube development in these explants results from maintenance
of the gradient pattern by proper fusion and healing. If the gradient is
obHterated or weakened by the fusions, epidermis, instead of neural tube,
results. In view of all the evidence it appears probable that any determi-
nation or Bahmmg toward determination in the presumptive neural plate
independently of the inductor is predominantly quantitative and highly
susceptible to change in environment.

Presumptive chorda-mesoderm is apparently somewhat more stably
determined. Pieces transplanted to other regions of embryos of the same
stage as the donor usually invaginate, form chorda-mesoderm, and induce;
but under certain conditions this region may take part in formation of
ectoderm.'''' The apparently more advanced determination of this tissue is
similar to the relatively high stability of high gradient-levels of hydroids,
planarians, etc., when transplanted to other regions. They are more able
to persist and induce than material from other gradient-levels (chap. xi).
The presumptive inductor region behaves in transplants exactly as would
be expected if it represents a sufficiently high gradient-level to render it
more or less independent of other parts. The question of its metabolism
has already been discussed (pp. 153-58).

ORIENTATION OF THE IITOUCED EMBRYO IN RELATION TO THE HOST

The longitudinal axis of the neural plate induced from presumptive
epidermis by implanted inductor tissue coincides with that of the implant,
and this, in turn, is definitely related to the direction of invagination and
migration of the chorda-mesoderm material. When invagination in the
normal direction in the intact embryo is prevented by injection of gelatin
into the blastocoel, it may occur in the opposite direction across the basal
region, and the neural plate develops there. Under these conditions part
of the inductor region may fail to invaginate, and neural plate may be in-
duced in it; also, part of the presumptive neural plate may become epi-
dermis (Eakin, 1933, 1939a).

The secondary embryonic axis induced by inductor implants is very
commonly parallel to the host axis, even when the inductor tissue is im-

'<> Vogt, i922(z; Mangold, 1923; Mangold und Seidel, 1927; Bruns, 1931; Lopaschov, 1935.

468 PATTERNS AND PROBLEMS OF DEVELOPMENT

planted in different orientation in relation to the host axis.'^ With vitally
stained inductor implants, oriented transverse to the host axis, Lehmann
observed that invagination began in the direction of the implant axis but
changed its direction toward the anterior end of the host. But when the
implant axis was directly opposed to the host axis, direction of invagina-
tion was highly variable, and the axis of the secondarily induced embryo
might be more or less opposed to that of the host embryo. In such cases
the secondary neural plate was broad and often did not close normally,
and the death rate was high. Apparently there is in the host a directive
action of some sort, influencing the direction of invagination and migra-
tion of the implanted inductor; and when orientation of implant and in-
duced axis are opposed to the host axis, form and development of neural
plate suggests, as Spemann has noted (1936, p. 97), that it is opposed by
some factor in the host. In these cases two opposed polarities are involved,
one established, the other introduced by experiment. Since the neural
plate in normal animals shows a general anteroposterior gradient in early
stages, it is probably that the inductor implanted with axis opposed to the
host axis induces in some cases a new gradient opposed to that of the host,
but this does not attain full development. These results are apparently
very much like the relation of the original gradient to the new gradient re-
sulting from section at the proximal ends of stem pieces of Tuhularia and
Corymorpha; the proximal hydranth gradient is shorter than the distal,
the scale of organization of the proximal hydranth is smaller than that of
the distal, and its development is slower. Ganglionic planarian grafts in
posterior regions often show similar differences in induced reorganization
posteriorly and anteriorly. The alterations in direction of invagination of
the amphibian inductor implant by the host are also closely paralleled in
ganglionic planarian grafts ; the polarity of the graft may be altered to con-
form with that of the host, particularly with implantation in the head re-
gion, but at other body-levels the polarity of the implant usually persists
(p. 382). Frequency of induction by Corymorpha implants depends on
level of origin of implant, level of implantation, and dominance of the host
hydranth (p. 378). At present interpretation in terms of gradients in-
volving dynamic factors serves as well for these relations in amphibia as in
hydroids and planarians.

HOMEOGENETIC INDUCTION

Induction of a neural plate by chorda-mesoderm is heterogenetic in
character, that is, something different from the inductor results from the

'' Geinitz, 1925c; Spemann, 1931; Lehmann, 1932.

EMBRYONIC INDUCTORS AND ORGANIZERS 469

induction. But induction of parts like the inductor — homeogenetic induc-
tion— also occurs. Implanted pieces of the dorsal inductor may dominate
adjoining host tissue, even of a different species, and induce mesoderm
formation from it, with resulting development of a complete chorda-meso-
derm; in heteroplastic transplants this consists in part of one species-
tissue, in part of another.'^ As noted above, presumptive ectoderm,
whether neural plate or epidermis, transplanted to the presumptive meso-
dermal region, invaginates with it and becomes mesoderm. Here, again,
the similarity to dominance in invertebrate reconstitution appears. A
piece of inductor dominates adjoining regions of the host, and they become
parts of the system concerned. These cases have been regarded as homeo-
genetic induction, but it may be questioned whether they are really ho-
meogenetic except in the general sense that mesoderm induces mesoderm.
The regions induced become parts of the chorda-mesoderm system; but
they supplement, rather than duplicate, the parts already present, so that
a harmonious whole results; consequently, they must become different in
some way, either in gradient-level or specific constitution, from the parts
of the system present in the implant. Mesoderm induced by mesoderm
from presumptive ectoderm is also capable of inducing neural plate and is
not species-specific in action.'*^ Presumptive neural plate or epidermis,
transplanted to entodermal regions, becomes entoderm. In some of these
cases it is perhaps not entirely certain whether or to what extent induction
by the region receiving the implant or the general physiological environ-
ment is concerned in determining the result.

A case of homeogenetic induction concerning which there is no doubt is
the induction of neural plate by neural plate after it has itself been induced
by chorda-mesoderm. Not only neural-plate naterial of the normal em-
bryo but that of a supernumerary neural plate induced by implanted in-
ductor can induce neural plate when implanted in the blastocoel, and its
action is not species-specific."' This result, although unexpected by the
investigators (see Spemann, 1936, p. 137), and justifiably so if the inductor
is primarily specific in its action, is exactly what might be expected if the
inductor is primarily an activator, whatever the manner in which activa-
tion is brought about. Neural-plate material has supposedly become dif-
ferent in some way from chorda-mesoderm ; but if induction is by means of
a specific substance, both must possess or produce it, and neural-plate ma-

'8 Spemann und H. Mangold, 1Q24; O. Mangold und Seidel, 1927; Lehmann, 1932; B.
Mayer, 1935.

''O. Mangold, 1923; Spemann und Geinitz, 1927; Holtfreter, 19336; Raven, 19356.
2" O. Mangold und Spemann, 1927; O. Mangold, 1929c.

470 PATTERNS AND PROBLEMS OF DEVELOPMENT

terial of one species must provide the substance necessaty to induce neural
plate in another. But, if neural-plate induction is primarily activation of
ectoderm, plate material, having itself been activated, should be capable
of inducing neural plate within a certain range of its developmental stages.
Actually the capacity is present not only in the open neural plate but per-
sists after closure of the folds and probably up to hatching. There is ap-
parently no regional correspondence transversely between inducing parts
of the plate and the induced plate. The right half of the anterior third of
the brain region from an axolotl neurula with closed neural folds, implant-
ed in the blastocoel of a Triton gastrula, develops, according to its origin,
as a half-brain region but induces a bilaterally symmetrical, very complete
head (O. Mangold, 1932a). There is, however, a longitudinal differential
in inductions by neural plate. Brain and head structures are induced most
frequently by anterior parts of the plate, and, the more posterior the part
of the plate used as inductor, the more posterior is the character of the in-
duced structure (O. Mangold, 1933). Different levels of the plate, like
different levels of chorda-mesoderm, are apparently ''head-inductors" and
"trunk-inductors." Also, implants of caudal portions of the plate may
bring about development of a supernumerary tail with neural tube and
mesodermal somites, induced in host tissue (Bytinski-Salz, 1931; O. Man-
gold, 1932a). It appears, however, that the extreme posterior part of the
plate normally gives rise in part to mesoderm of the tail (Bijtel und Woer-
deman, 1928; Bijtel, 1931); consequently, its inducing capacity should be
similar to that of caudal mesoderm.

THE PROBLEM OF THE NATURE OF THE INDUCING FACTOR

Early in the course of investigation of amphibian induction by chorda-
mesoderm material the same question arose that had arisen earlier with
respect to other cases of dominance and induction in development, that is,
the question of the manner in which induction is brought about. One hne
of experiment concerned with this question was the attempt to determine
whether other tissues, not only of amphibians but of other animals, would
induce. Apparently first was the discovery that the developing amphibian
limb bud had some inducing power (O. Mangold, 1928); the regenerating
tissue of the urodele leg and tail gave clear-cut induction (Umanski, 1932,
1933). These cases suggest a relation between induction and tissue activ-
ity. On the other hand, narcotized dorsal-lip tissue was found to be effec-
tive (Marx, 1930). However, the whole problem of induction began to ap-
pear in a new light when it was discovered that the urodele dorsal lip and

EMBRYONIC INDUCTORS AND ORGANIZERS 471

the neural plate remain capable of induction after killing by heat, freez-
ing, drying, after several hours in 20 per cent HCl, several days in ether,
several hours' extraction by ethyl or petrol ether, 6 months in alcohol, fol-
lowed by impregnation with xylol and paraffin, etc."' This line of experi-
ment was most extensively developed by Holtfreter, using the method of
implantation in the blastocoel of a host and that of culture in a modified
Ringer solution of pieces of presumptive epidermis placed on, or inclosing,
the treated tissue. Equally interesting was the discovery by Bautzmann,
and its confirmation and extension by Holtfreter, that regions of the em-
bryo which do not induce when alive can become inductors when killed in
various ways. Even fragments of boiled undivided eggs and centrifugates
of ovarian eggs are able to induce. It was also found that various adult
tissues from many animals and even some plant tissues have more or less
inducing power on urodele embryos, some both hving and dead, others
only when dead.'" Apparently, however, hquid tissue extracts, to be effec-
tive, must be coagulated or more or less solidified by addition of sub-
stances proved not to be inductors, but solid bodies are not necessarily
inductors. The heterogeneous character of these foreign inductors is evi-
dent from the following incomplete list, taken in large part from Weiss

(1935)-

Plants: Cambium of birch, induced small neural plate; growing tip of potato tuber,
induction neural of tissue but no distinct plate

Coelentera: Hydra tissue, boiled

Annelids {Enchytmeus) : Body fragments

Mollusca {PlanorUs, Limnaea) : Muscles of foot, hepatopancreas

Crustacea {Daphnia) : Coagulated body extract.

Lepidoptera {Deilephila) : Haemolymph and ganglia of pupa

Odonata {Libellula larva) : Fat body, ganglia

Fishes: Gasterosteus: heart, liver, ovarian eggs, muscle, spleen; Danio: presump-
tive chorda-mesoderm

Amphibia {Triton, Salamandra, Rana): Liver, heart, ovarian eggs, nuclei and
cytoplasm of unfertilized eggs, nuclei more effective, muscle, cartilage, brain, retina,
regenerating tissue of tail

Reptiles (Lacerta) : Liver, kidney, testis

Birds: Liver, kidney, testis, thyroid, fat body, brain, retina, coagulated chick
embryo extract, fragments of primitive streak

Mammals (mouse) : Heart, liver, kidney, adrenals, brain lens; also calf liver

Man: Liver, brain, kidney, thyroid, tongue, sarcoma, carcinoma

2' Bautzmann, Holtfreter, Spemann, and Mangold, 1932; Mangold, 1932b; Holtfreter,
1933c, 1934a; Spemann, Fischer, and Wehmeier, 1933; Wehmeier, 1934.

" Fischer und Wehmeier, 1933a, b; Woerdeman, 1933c; Holtfreter, 1934^/ Wehmeier, 1934;
Hatt, 1934; Waddington and Wolsky, 1936; Ragosina, 1936, 1937; and other papers.

472 PATTERNS AND PROBLEMS OF DEVELOPMENT

It is sufi&ciently evident from this list that abiUty to bring about the
change in urodele tissues that constitutes induction is not hmited to any
particular kind of tissue from any single group of organisms but is a very
general property of tissues, either alive or dead or both. Holtfreter con-
cluded from his experiments that induction is effected by a chemical sub-
stance or substances. If a single specific substance is the inducing factor
in all cases, it must be a substance present in tissues of many kinds from
many groups, but the experiments give no clue as to what substance or
substances may be concerned. They do not show whether living and dead
inductors act in the same way or whether any of the foreign inductors in-
duce in the same way as the natural inductor.

Quite aside from the demonstration of the widespread occurrence of in-
ducing power, these experiments, particularly Holtfreter's extensive data,
are of much interest in relation to certain other aspects of the problem of
induction. Since certain of the inductors bring about extensive inductions,
while others under similar conditions have only slight inducing action, it
appears probable that size and completeness of development of the in-
duced structure may indicate, in some degree, the intensity or effectiveness
of the inducing action. Frequency of induction also varies with different
inductors and may likewise indicate difference in effectiveness. In general,
inductors which become such only after killing are apparently weaker in
action than the natural inductors. Heating to ioo° C. usually weakens
inducing action of tissues that induce at lower temperatures, and heating
above 120° abolishes it in various cases. Invertebrate tissues are usually
less effective than those of vertebrates. Certain tissues — for example,
those of the internal glands, such as the liver — are more effective than
various others. Almost all the tissues used induce neural tissue, but only
certain of them induce mesoderm. The inductions differ greatly in char-
acter, ranging from mere epidermal thickenings or sometimes a single lens
or balancer to large masses, consisting of various supernumerary organs,
several neural tubes being formed in some cases. The neural induction
may resemble a part of the spinal cord or may become brainlike ; and gang-
lia, nose, eyes, otic vesicles, and balancers may develop in relation to it.
Supernumerary legs may also develop, and induction of muscle and noto-
chord from ectoderm or from mesoderm sometimes occurs.

The apparent absence of any definite relation between character and
origin of inductor and character of parts induced suggests that the foreign-
tissue inductors are primarily activators rather than definitely specific for
particular organs or complexes and that other factors — physiological con-

EMBRYONIC INDUCTORS AND ORGANIZERS 473

dition, regions of origin, and developmental stage of ectoderm or other
tissue in which induction occurs — play an essential part in determining the
result, particularly when pieces of tissue are explanted with the inducing
tissue. The effects of foreign inductors are found to be quantitatively
greater at anterior than at posterior levels of the hosts. Organs of the
head region, brain, sense organs, and balancers appear chiefly in tissue
from more anterior levels, while posteriorly atypical neural structures and
tails are more frequent; induced kidneys and legs appear more frequently
at or near the levels where they normally develop. Also, axial orientation
of induced parts, if they are axiate, commonly coincides in direction with
longitudinal host axis (Holtfreter, 1934/^). The anteroposterior differences
and the orientation point to the longitudinal or polar gradient of the host
as the factor concerned, and the regional differences in character of organs
induced may also be determined by this gradient or by the regional or or-
gan fields resulting from it. If this is the case, an adequate activation of a
region of the host body within a particular field may be expected to result
in development of the organ or organ system characteristic of that field.
However, correspondence between expectation and actual character of
induced parts is by no means complete. Apparently the fields, if present,
are not stably determined, and activation may have different results in
the same region. Brain and head structures are said to appear more fre-
quently with strong than with weak inductors. If the inductor is primarily
an activator, this is to be expected. A weak inductor may activate only
slightly and so induce only some slight differentiation of neural tissue or
only a thickening that cannot be certainly identified as neural, even in
anterior host regions. Conversely, a strong inductor may activate more
posterior levels to such a degree that they develop as brain. Still more in-
tense activation may perhaps induce formation of notochord and meso-
derm. It is difiicult to account for the experimental results except in some
such terms as these, unless we assume specifically different inductor sub-
stances for each of the different inductions.

A recent example of such assumptions appears in the work of Chuang
(1938, 1940; also further data in another paper, 1940, Arch. Entw'mech.,
140). He finds that mouse kidney induces in isolated gastrula ecto-
derm exclusively brain parts, sense organs, and other head structures
and that Triton liver induces trunk regions, notochord, muscles, and tail.
These differences are regarded as specific effects of the two foreign in-
ductors. With implantation of the kidney or liver in the ventral side of a
gastrula these differences in character of inductions are less evident; both

474 PATTERNS AND PROBLEMS OF DEVELOPMENT

tissues may induce mesoderm and secondary appendages, but in general,
mouse kidney induces trunk parts less frequently than Triton liver. Ac-
cording to the author, a "host effect" masks more or less completely the
specific differences in inducing action of the two tissues. This host effect
consists in decrease in frequency of head parts from anterior levels poste-
riorly and in trunk parts in the opposite direction. Moreover, with mouse
kidney the character of induction differs after different periods of boiling:
with boiling for a few seconds, induction, especially of mesodermal parts,
is increased; after 15 minutes of boiling no mesodermal parts are induced,
and frequency of brain inductions is decreased; after i hour of boiling
there is further decrease in brain inductions, but sense organs and bal-
ancers are still induced. These results are regarded as due to different sub-
stances or as specific effects of different periods of boiling on some sub-
stance or complex. However, the evidence for specific action does not ap-
pear any more conclusive in these than in other experiments. The differ-
ences may be due to differences in intensity or rate of inducing action and
stage of development of host. The "host effect" is obviously an expression
of the longitudinal gradient. If the inducing action of kidney and liver
were actually specific, how could the host effect alter it to a mere differ-
ence in frequency of particular parts? The results after different periods of
boiling may also indicate merely different intensities or rates of inducing
action, effective at different stages of host development with different re-
active capacities of host tissue to the same inducing action.

The conclusion that organ fields of the host determine what parts re-
sult from action of foreign inductors (Weiss, 1935) seems not entirely in
accord with the data. Very probably the fields are concerned in certain
cases — for example, when a lens or a leg is induced — but whether or to
what extent the various fields are present and sufficiently developed to de-
termine the result of induction is uncertain. While certain parts, such as
brain, appear more frequently in certain regions of the host, they may ap-
pear elsewhere. In many cases the region activated by the inductor ap-
parently develops organ fields of its own, as might be expected if fields re-
sult from a gradient pattern.

In the light of the evidence considered thus far the question is pertinent
whether or to what extent these inductors, natural or foreign, living or
dead tissues, are actually organizers? It was pointed out above that the
longitudinal gradients of ectoderm and inductor coincide in direction after
invagination and more or less closely in extent. Does the chorda-meso-
derm do anything more in normal development than "reinforce" and per-

EMBRYONIC INDUCTORS AND ORGANIZERS 475

haps steepen the primary ectodermal gradient? Assuming that the in-
ductor does act in this way, it is only indirectly concerned in organization
in that it alters the gradient pattern in the ectoderm, that pattern being
the real organizer. The chorda-mesoderm itself develops longitudinal and
transverse organization, but in its earher stages there is no definite evi-
dence of anything more than a gradient pattern in it, and with differen-
tial inhibition the higher levels of this pattern, that is, anterior and median
regions, are most inhibited.^^ Implanted into other than the normal posi-
tion, it may sometimes induce a new gradient, as in the case in which the
induced embryo is more or less opposed in orientation to the host. Here
the inductor gradient determines the gradient pattern of the induced em-
bryo, and so its organization, and comes nearer being a true organizer than
in normal development; but even here, according to this view, the gradi-
ent pattern induced in the ectoderm is the real organizing factor.

The foreign inductors, if living, may or may not possess gradient differ-
ences; whether living or dead, their action is usually more or less local
and does not induce an orderly whole but merely certain tissues or organs
in more or less unordered complex. If these inductors are primarily activa-
tors, the character of the induced part depends chiefly on two factors-
level of host gradient in which it arises and degree or intensity of activa-
tion. Either of these may be predominant in determining the result. These
inductors, however, give us no information as to the manner in which or-
ganization occurs. Weiss (1935) admits this as regards foreign inductors
but regards the natural inductor not only as activator but as organizer.
Actually, however, there is often much organization in the complexes in-
duced by foreign inductors, but it is not normal. If these inductors are not
organizers, the organization must result from pattern of some kind already
present in the part subjected to inductive action, and the only sort of pat-
tern that seems adequate to account for the varied results is a gradient
pattern, wholly or predominantly nonspecific regionally. But, as already
noted, foreign inductors can probably determine new gradient patterns
in the host tissue; if so, they are indirectly organizers.

Needham, Waddington, and Needham (1934) hold that two factors,
"evocation" and "individuation," are concerned in induction. Evocation
consists in bringing about development of an embryonic axis in the ecto-
derm; individuation, in determining the regional character of that axis.
The foreign inductors— living, dead, or extracts— are, in general, only
evocators. This conception of induction is open to certain criticisms.

^3 See pp. 257-65; also Bellamy, 1919; Lehmann, 19370; Cohen, i(

476 PATTERNS AND PROBLEMS OF DEVELOPMENT

First, there is the interesting question: What is the nature of an embry-
onic axis without regional character? Second, the foreign inductor does
not always evoke an embryonic axis; it may induce merely a mass of tis-
sue or an unordered complex; and when an axis does appear, it usually
coincides with the host axis. Another interesting question is: How can a
piece of some tissue — for example, liver — induce an embryonic axis?
Third, according to these authors, an axis without individuation may
range from a neural tube with notochord and somites in proper position,
but one end not distinguishable from the other, to histologically differen-
tiated neural, chordal, and mesodermal cells without definite order. But
development of neural tube with somites and notochord in normal rela-
tions represents a very definite axiate pattern. Even if the two ends are
not distinguishable as different regions, it probably did not arise all at
once; and if it did not, a pattern with regional character of some sort must
have been present. The concept of an embryonic axis without regional
character seems somewhat metaphysical. Does not an axiate pattern, par-
tial or complete, constitute individuation?

The concept of an individuation field originating from the inductor in
normal development (Waddington and Schmidt, 1933) also presents difii-
culties. The inductor does not originate embryonic pattern; it is a part
of that pattern. In normal development it does not determine the polar
axis of the embryo but merely plays a part in determining certain develop-
mental events in relation to that axis. Does not the pattern of the whole
amphibian egg constitute the individuation field, within which orderly
changes precede and bring about invagination and induction?

CHEMICAL ASPECTS OF THE INDUCTOR PROBLEM

That induction may be due to a chemical substance was suggested by
Spemann and Mangold (1924). The later discoveries that dead tissues and
tissue extracts, when associated with a more or less solid carrier, can in-
duce, and that boiling, treatment with alcohol, ethyl ether, acetone, ace-
tone followed by several days in water, or glacial acetic acid, does not de-
stroy inducing power and in some cases increases it, seemed to indicate
that any inducing substance present in the material is not soluble with
these treatments, though it must presumably reach the overlying ecto-
derm in some way. Further experiment showed, however, at least for
acetone, that both the extracted residue and the acetone extract could
induce.^''

^■'Spemann, Fischer, und Wehmcier, 1933; Fischer und Wehmeier, 1933(7, b; Wehmeier,
1934; Holtfreter, 1934a, b.

EMBRYONIC INDUCTORS AND ORGANIZERS 477

According to Woerdeman (1933^^, b, d), the glycogen content in the
apical hemisphere of the amphibian blastula is high and more or less uni-
form but decreases in the dorsal lip region as it invaginates. Similar ob-
servations were made by Tanaka (1934)- In transplanted pieces of the
dorsal lip the same decrease occurs on invagination (Raven, 1933a, 1935a).
On the basis of his observations Woerdeman advanced the view that de-
crease in glycogen in the inductor region at the time of gastrulation may
be associated with induction, presumably through glycolysis. He found
also that implants of malignant tumors known to have high glycolytic
activity induce (Woerdeman, 1933c). Fischer and Wehmeier (1933a;
Wehmeier, 1934) also suggested, independently of Woerdeman and on
other grounds, that glycogen might be an inductor substance and obtained
neural-plate inductions with glycogen from liver, but in later experiment
they found that inducing power decreased to zero as purification of the
glycogen preparations became more complete (Fischer und Wehmeier,
1933&). Induction was also obtained with the unsaponifiable material
from ethereal extracts of crude glycogen.'^ Both Woerdeman (1933c) and
Holtfreter (19346) found glycogen inactive. More recently the conclusion
has been drawn from experiments with glycogen from rabbit liver that an
agent associated with glycogen is a highly effective inductor (Heatley,
Waddington, and Needham, 1937). Pasteels (1936c), using a different
method from that of earlier workers for demonstrating intracellular gly-
cogen, maintained that the supposed decrease in the region of the dorsal
lip does not occur and that, consequently, increased glycolysis cannot be
responsible for the inducing power of the chorda-mesoderm. However,
with still another method Heatley and Lindahl (1937) found that during
gastrulation glycogen decreases in all parts of the embryo but that de-
crease is greatest in the invaginating material. The high anaerobic gly-
colysis of the inductor region (three times that of other parts) and its high
respiratory quotient (about unity), indicating carbohydrate metabolism,
have already been noted (pp. 154-55). However, according to Brachet
(1939), carbohydrate metaboUsm is not necessary for induction ; induction
may take place in the presence of agents inhibiting glycolysis. He sug-
gests that proteins may be important factors in induction. Evidently
still further investigation along these lines is necessary. Whatever the
final conclusion, the glycogen hypothesis is interesting, as pointed out by
Weiss (1935, p. 664), in that it suggests that induction is mediated by

^5 Waddington, Nowinski, Needham, and Needham, 1934; Waddington, Needham, Nowin-
ski, Lemberg, and Cohen, 1936.

478 PATTERNS AND PROBLEMS OF DEVELOPMENT

metabolic activity with glycogen as source; that is, the activity, rather
than the substance, is the real inductor.

Other attempts were made with other methods to isolate a particular
inductor substance. The material of neurulae, crushed in a small volume of
water and centrifuged, separated into three layers — the cellular debris, an
aqueous layer containing protein but not entirely fat-free, and a fatty
layer. Induction of neural tubes was obtained by implantation of coagu-
lated parts of the aqueous, protein-containing layer. Also, neurulae,
ground with anhydrous sodium sulphate, the mass extracted with ethyl
ether or petroleum ether and the extract mixed with solid carriers shown
by experiment not to be inductors, gave, when implanted, neural tubes or,
more commonly, solid rods and other masses, "probably neural in char-
acter." Petroleum ether extracts of adult amphibian viscera were also in-
ductive. From these experiments the conclusion was drawn that the in-
ductor is a definite chemical substance, soluble in ether and probably
lipoidal.-^

Continuing work with ethereal extracts, Waddington, Needham, and
co-workers obtained induction in Triton and axolotl with the unsaponiti-
able fractions of ether extracts of whole newt bodies and mammalian liver
and with the parts of these fractions precipitated by digitonin, implanted
after emulsifying in egg albumin and coagulating by heat. Minute doses
of certain synthetic polycyclic hydrocarbons, massive doses of certain
acids, and certain other substances may also induce. Recently methylene
blue has also been found to induce, and it has been suggested that it acts
by setting free an inductor substance. Nuclei of unfertilized amphibian
eggs are more potent inductors than the cytoplasm. The earlier sugges-
tion of these workers was that induction was due to a sterol-Uke substance ;
but further experiment seems to indicate that, even if such substances are
inductors, other substances also induce.^^ The possibility that methylene
blue, a respiratory catalyst in certain concentrations, induces directly by
activation may be noted. Methylene blue in low concentrations increases
oxygen uptake 45 per cent in amphibian embryos, in high concentrations
is inhibitory ; and Janus green and neutral red in low concentrations also

"^ Waddington, Needham, and Needham, 19330, b; Needham, Waddington, and Needham,
1934-

^'Waddington, Needham, ei al., 1934, 1935; Waddington and D. M. Needham, 1935;
Waddington, Needham, et al., 1936; Waddington, Needham, and J. Brachet, 1936; Wadding-
ton, 1938a, b. For a quantitative study of induction by a water-soluble hydrocarbon see
Shen, 1939. See also Needham, 1939, "Biochemical aspects of organizer phenomena," Groictli,
Suppl.

EMBRYONIC INDUCTORS AND ORGANIZERS 479

induce neural development in ventral gastrula ectoderm isolated into the
solution.-^ '^

It was found by Fischer, Wehmeier, and associates that pure fatty
acids of animal and plant origin, also synthetic acids, emulsified in agar
and implanted, induce and that exhaustive extraction of amphibian em-
bryos or tissues of other animals with aqueous or organic solvents does not
abolish inductive power. Moreover, the nucleoprotein fractions derived
from these tissues are the active agents.^''

Induction, owing either to cephalin or some impurity, was obtained
with cephahn fractions of mammalian brain (Barth, 1934c). However,
acetone extracts of brain, containing sterols but no cephalin, and the un-
saponifiable fraction of cephalin also induce. Cephalin preparations that
induce also produce more or less cytolysis; and other cytolytic agents — •
for example, digitonin — with acid or alkahne buffers, also induce. The
protein residue of calf brain is more potent than the lipoid extract. ■'*'

Since various tissues induce after killing but not while living, and cer-
tain foreign inductors — for example, cephalin and digitonin — produce
more or less cytolysis in the host, it has been suggested by Barth and
others that this cytolysis of host cells may set free inducing substance.
Neural induction by fuller's earth, silica, and CaC03 has been reported by
Okada (1938). These substances do not give off chemical inductors, but
they do produce injury and cytolysis of tissues. Okada regards substance
or substances set free by the cytolyzed tissue as the inductor in this case.
Induction by microcautery in the blastocoel is reported by Cohen (1938).

Evidently the chemical problem of induction is by no means solved. If
induction or "evocation" is primarily activation, it seems not at all im-
probable that many chemical substances may be inductors and that irrita-
tion may also induce. It is also by no means certain that any of the for-
eign inducing factors is identical with the natural inductor. It has not
even been demonstrated that the natural inductor is a substance. In any
case, it is not evident how a particular chemical substance can organize
a neural axis unless its concentration or the amount set free is graded in
the inductor, but such a gradation provides an axial differential independ-
ent of the chemical constitution of an inducing substance. In any case,
induction by a particular chemical substance, whether produced by the
natural inductor or of other origin, throws no light on the problem of or-

2* Beatty, De Jong, and Zielinski, 1939.

^' Fischer, Wehmeier, Lehmann, Jiihling, und Hultzsch, 1935.

i° Barth, igs4d, 1937, 1939a; Barth and Graff, 1938.

48o PATTERNS AND PROBLEMS OF DEVELOPMENT

ganization. For organization a spatial pattern of some sort is essential,
and it is not at present evident how a sterol or other particular substance
can, of itself, originate an orderly and definite spatial pattern. The prob-
lem of organization is still the problem of the spatial pattern in the em-
bryo and any other developmental sytsem. In the amphibian the natural
inductor does not originate this pattern but is a part of it and plays a role
in modifying it. The foreign inductors merely modify it by local activation
or otherwise. The spatial pattern, whether we call it a gradient, a gradient
system, or something else, is the real organizer. In this connection Wad-
dington's recent discussion of organization (see p. 455, footnote 8) is of
particular interest because of his inability to throw any real light on
the problem.

INDUCTION IN EARLY DEVELOPMENT OF OTHER CHORDATES

ASCIDIANS

The ascidian egg was earlier regarded as an extreme "mosaic," but re-
cent experiment has shown that a considerable capacity for reconstitution
is present (p. 577). Moreover, according to the most recent work with
isolated blastomeres, an inductor is concerned in early development. The
posterior blastomeres of the four-cell stage and the apical blastomeres of
the eight-cell stage are found to be incapable of independent differentia-
tion when isolated, but the anterior basal blastomeres of the eight-cell
stage can differentiate independently and are necessary for the differen-
tiation of other parts. These blastomeres represent presumptive noto-
chord, nerve cord, entoderm, and some mesoderm, but ascidian anterior
half-embryos differ from amphibian dorsal half-embryos in that they do
not reconstitute (Rose, 1939).-''

According to other experiments, removal of various parts of the ascidian
embryo some time before gastrulation gives no certain evidence of induc-
tion (von Ubisch, 1940, and another paper, 1940, Arch. Entw'mech., 140).
After removal of prospective notochord material the nervous system and
other organs are normal. After removal of all prospective muscle and
most of mesenchyme other organs develop without evidence of induction
by mesoderm. When prospective neural material is removed, no nervous
system develops; evidently there is no neural induction in other ectoderm

3' Comparison with amphibian development raises the question whether the two sides of
the egg and embryo designated by ConkHn (1905a) and others as anterior and posterior are
not more nearly dorsal and ventral; neural plate and notochord develop from the "anterior"
basal cells of the eight-cell stage.

EMBRYONIC INDUCTORS AND ORGANIZERS 481

by chorda-mesoderm or any other part. After removal of entoderm there
is apparently some reconstitution of entoderm, probably from ectoderm;
and entoderm may perhaps have some inductive action on the papillae of
attachment. Rose's data suggest induction at a very early stage, but how
far it is similar to amphibian induction is not yet evident; conceivably,
it may have taken place before the stage of von Ubisch's experiments.

FISHES

The posterior part of the teleost embryonic shield, representing the
dorsal lip of the blastopore, possesses inducing capacity, as in amphibians.
When pieces of it are transplanted to other regions, even the extraembry-

A B

Fig. 161, A, B. — Induction in the teleost, Fundulus. A, induced embryo (/) in primary
embryonic shield; B, induced embryo (/) outside primary shield (from Oppenheimer, 1936c).

onic regions of the blastoderm, they may induce secondary embryos. ^^
When transplantation is within the embryonic shield, the induced and the
primary embryo are closely similar, and corresponding parts are at the
same levels; but embryos induced by transplants to extraembryonic re-
gions show no definite relations to the primary embryo, and corresponding
parts may develop at different levels (Fig. 161). The polarity of the em-
bryo induced in the extraembryonic region may even be opposed to the
host polarity. Evidently, within the embryonic shield the longitudinal
pattern of the host determines direction of the induced secondary em-
bryonic axis and the levels at which its organs develop. In extraembryonic
regions gradient pattern is slight or absent; consequently, the inductor
gradient may become the chief or only factor in determining axial direc-
tion. However, blastomeres explanted to salt solution may undergo more

■i- Oppenheimer, 1934, 1936a, b, c; W. Luther, 1935, 1936a, h.

482 PATTERNS AND PROBLEMS OF DEVELOPMENT

or less differentiation without gastrulation (Oppenheimer, 19366). More-
over, up to the end of the blastula stage a piece from any region of the
blastoderm, including part of the margin, may, when isolated in the yolk
sac of an older individual, give rise to a more or less complete embryonic
primordium. Neural tissue develops in most cases, notochord almost as
frequently, even from extraembryonic regions containing none of the pre-
sumptive chorda-mesoderm; and very frequently gut and kidney also de-
velop. This capacity of parts of the blastoderm for independent reconsti-
tution decreases rapidly in extraembryonic regions with gastrulation, and
inductor action becomes necessary for embryo formation (Luther, 1936a).
In its high capacity for reconstitution in earher stages the teleost blasto-
derm differs from the amphibian blastula, although under certain condi-
tions neural, or even mesodermal, tissue may develop from presumptive
amphibian epidermis or neural plate, as noted above. Luther regards the
inductor as an activator, differing quantitatively, not qualitatively, from
other parts, and holds that the capacity of isolated pieces of the blastula
blastoderm to reconstitute embryonic primordia results from levels of
activity high enough in all regions at this stage to determine embryo for-
mation without aid of an inductor. As gastrulation begins, this activity
decreases in extraembryonic regions, and embryo formation occurs there
only by induction. This view agrees closely with that suggested above for
the amphibia, and the data on differential susceptibihty in teleosts also
show a high susceptibility in the inductor region as gastrulation ap-
proaches (p. 150). Perhaps the point of greatest interest in these experi-
ments on teleosts, as compared with those on amphibians, is the recon-
stitution of new inductor regions in isolated pieces, even from extraem-
bryonic regions without any of the original inductor. This does not sug-
gest any considerable specificity in the inductor. The fact that the extra-
embryonic regions do not develop inductors and embryos in normal de-
velopment suggests that the dominance of some other region, presumably
the more active embryonic region, or later the inductor region itself, pre-
vents such development, and that an activation resulting from isolation is
sufficient to bring about development of neural tissue or even of inductor
region. In short, the data give further support to the conception of neural
induction as primarily an activation and appear difficult to interpret in
other terms.

Following division of the archenteric roof of the trout embryo into
anterior, middle, and posterior pieces, inclosure of each in indifferent ecto-
derm from the extraembryonic region, and implantation in the yolk sac,

EMBRYONIC INDUCTORS AND ORGANIZERS 483

the anterior piece induces no neural tissue in the ectoderm, the middle
piece induces brainlike parts, the posterior piece chiefly spinal neural
structure and occasionally otic vesicle. The middle piece is regarded as
''head organizer," the posterior piece as "trunk organizer" (Eakin,
1939^) ; but the experiments do not show that there is anything more than
a quantitative difference between them.

BIRDS

Following chscovery of induction by the region of the dorsal blastopore
lip in amphibians, the question whether a comparable induction occurred
in the chick naturally arose. It was suggested that the region of the node
at the anterior end of the primitive streak is an "organizer," or at least
essential to formation of the embryo. ■'•' However, it was found that por-
tions of the streak not including the node can continue development when
isolated in vitro, and implantation experiments showed that not only the
node region of the streak but more posterior regions without any part of
the node can induce neural tissue or plate or a more or less complete em-
bryonic axis. Implants from more anterior levels of the streak differenti-
ate into notochord, somites, and neural tissue; those from levels farther
posterior form only mesodermal structures; but both induce. Both ante-
rior and posterior derivatives of the streak, the head process, and the sinus
rhomboidalis, and also the neural plate after its formation, can induce
neural plate. ^"^ The action is not species-specific; heteroplastic implants
between duck and chick are effective. Pieces of primitive streak killed and
coagulated by boiling water also induce. The host "individuation field"
(Waddington) influences axial orientation of induced parts, as in am-
phibians; but sometimes the induced axis is opposed to that of the host
and is then supposedly determined by the inductor. Apparently there may
be considerable reconstitution of the graft into more anterior parts or a
larger part of the longitudinal axis than would have been formed in normal
development. Reaction of the host to implanted inductor decreases as de-
velopment progresses. Neural plate is induced up to a later head-fold
stage, but at this stage only by implantation in anterior regions; at mid-
trunk levels only epidermal thickening results. In neural-fold stages only
thickening occurs at all levels, and at early somite stages the epidermis
shows no reaction (Woodside, 1937).

33 Wetzel, 1925a, b, 1929; T. E. Hunt, 1929, 1931, 1932; Willier and Rawles, 1931.
^-t Waddington, 1930, 1932, 1933^, 1934, 1935, 1937; Waddington and Schmidt, 1933;
Waddington and Taylor, 1937.

484 PATTERNS AND PROBLEMS OF DEVELOPMENT

Pieces of extraembryonic ectoderm, presumptive epidermis, and pre-
sumptive axial or lateral mesoderm of primitive-streak stage implanted
under the primitive streak of the same stage are all capable of developing
into neural tissue. Their anteroposterior axes may be altered or reversed
by the host, but dorsiventrality persists (Abercrombie, 1937). The neural
development of these implants is considered to be due to induction by the
host streak. The polarities of pieces of primitive-streak pieces implanted
beneath primitive streak may also be reversed by the host and their re-
gional character altered to conform to host-level. In some cases the host
axis shifts to one side of the implant, and the latter induces either extra
neural tissue or a new axis, or the implant may be incorporated in the host
body (Abercrombie and Waddington, 1937). By implantation into the
primitive streak presumptive ectoderm may be converted into mesoderm
(Waddington and Taylor, 1937). In these cases the effect of the host on
the implant is evident in varying degree. The implant may retain its in-
dividuality and induce, or be altered as to polarity and body-level, or be-
come a part of the host body.

An interesting apparent induction by the entodermal axis has been
obtained by separating epiblast and entoderm in early primitive-streak
stages of duck and chick embryos and replacing them with longitudinal
axes opposed (Waddington, 1933a). Effects on development of this pro-
cedure vary. In cases of axial induction two embryos may develop with
heads together and longitudinal axes opposed, one representing the epi-
blast axis, the other apparently induced by the entodermal axis; or the
epiblast axis determines an embryo, the entodermal axis a transitory prim-
itive streak. In still other cases the original primitive streak disappears,
and the embryo determined by the entodermal axis persists. Even when
a new embryonic axis is not induced, the original axis may be inclined to-
ward the direction of the entodermal axis or be semicircular, or the em-
bryo may be very short. Waddington holds that the entoderm merely in-
duces the cell migrations that result in development of the primitive
streak but admits that a gradient system or other axiate pattern is neces-
sary to account for the orderly character and definite directions of the mi-
grations. However, it does not appear from the data that the entodermal
axis differs essentially in its inducing action from implanted parts of the
primitive streak. It is apparently a rather effective inductor, for it is able
to determine a new polarity opposed to the original. That this polarity
results from the imposition on the epiblast of a new gradient pattern cor-
responding to the entodermal pattern seems probable. The persistence of

EMBRYONIC INDUCTORS AND ORGANIZERS 485

both or either one of the two axes doubtless depends on incidental factors,
levels of activity of epiblast and entoderm, and perhaps their alterations
by operation ; it suggests varying relations of dominance of one or the other

axis.

Neural-plate induction in the chick has also been obtained by implanta-
tion of the anterior part of a two-somite rabbit embryo (Waddington,
1934), affording still another example of the nonspecific character of these
inductions.

However, it appears from many experiments that neural tissue can de-
velop independently of an inductor in the chick. In various isolation ex-
periments forebrain has been found to develop independently of noto-
chord, and grafts of notochord have proved ineffective as inductors.^s
Moreover, explants of purely ectodermal pieces from anterior regions of
very early primitive-streak stages may give rise to neural tubes (Rudnick,
19386).

MAMMALS

The few data available show that neural induction can occur in a mam-
malian embryo but give no information as to the part it plays in normal
development. Chick primitive streak implanted in the rabbit embryo cul-
tivated in vitro may induce neural-plate formation, but thus far implanta-
tion of rabbit primitive streak in rabbit embryos has given only uncertain
results as regards induction, although the implants may differentiate into
neural tissue and notochord (Waddington, 1934, 19366, 1937).

NEURAL induction: CONCLUSIONS AND QUESTIONS

From the data it appears that many different substances can induce,
directly or indirectly, development of neural tissue in amphibians and that
under certain conditions neural tissue may develop without an underlying
inductor. It may perhaps be questioned whether there is a specific chem-
istry of induction in the sense that a particular substance or group of sub-
stances is the inducing factor. It also appears that the natural inductor
does not originate axiate embryonic pattern but is a part of it, and that the
pattern, either of the inductor or of the reacting ectoderm or both, rather
than the induction itself, is the real organizer. Induction in vertebrates
throws no light on the essential problem of organization, the problem of
the origin and nature of the pattern within which the inductor acts.

The question whether neural induction is primarily an activation or a
specific action, or in normal development an orderly axiate series of spe-

35 T. E. Hunt, 1931; Waddington, 1932, 1933; Wetzel, 1936.

486 PATTERNS AND PROBLEMS OF DEVELOPMENT

cific actions, has been raised and discussed above. With the progress of
experiment evidence for specific action of the inductor seems progressively
less conclusive, even though specific substances may be concerned. The
data on neural induction do not conflict with the evidences of gradient pat-
tern given by other hnes of experiment. On the other hand, they may be
regarded as throwing some light on the manner in which development
proceeds from gradient patterns of certain sorts.

An interesting question in relation to the problem of organization is
that of the origin of the inductor region. In amphibian embryos under
experimental conditions a region of invagination may appear in other than
the normal position, but evidently in relation to a pattern (pp. 259, 429).
These cases suggest that in normal development the dorsal inductor may
originate within a more general pattern, perhaps primarily as a local ac-
tivation, perhaps as the earliest determination or differentiation; but how
its localization on one side of the egg is determined is not known (p. 686).
That the inductor region is not a primary feature of pattern is also indi-
cated by the appearance of new inductors in reconstitution of isolated
extraembryonic parts of the fish blastoderm. The problem of organization
in vertebrates involves not merely the results of inductor action but the
origin and pattern of the inductor and the origin and nature of the pattern
within which it appears and acts.

The question of the role of induction in normal development is not fully
answered as yet. In the teleosts isolation of a part of the blastoderm is suf-
ficient to bring about development of new embryonic axes in the entire
absence of the inductor. This being the case, does the inductor play any
essential part in normal neural development? Under certain conditions
explants of presumptive amphibian epidermis can develop into neural tis-
sue or even into mesoderm in absence of the inductor, though such devel-
opment probably results in some cases from activating conditions or sub-
stances which perhaps should be regarded as inductors. As regards birds
and mammals, we know at present little more than that new embryonic
axes may result from implantation of certain parts.

If induction is essential to the development of a morphological axis in
the ectoderm and if it is primarily an activation, it is an activation in a
definite graded axiate pattern, and this is the first step toward axiate de-
termination and differentiation. Induction by the chorda-mesoderm does
not represent the origin of axiate pattern but is merely an expression of
pattern already present. That cells or regions at different relative levels
of physiological activity behave difTerently in development appears cer-

EMBRYONIC INDUCTORS AND ORGANIZERS 487

tain. In various hydroids hydranths degenerate and stolons develop, even
from apical, as well as from basal, ends, in standing water, in KCN, and
with other inhibiting agents, while in flowing, well-aerated water hy-
dranths develop (pp. 172-75). Is this fundamentally different from devel-
opment of presumptive amphibian epidermis as epidermis under certain
conditions, as neural tissue, or as mesoderm under others?

LENS INDUCTION

Among the organ inductions of later embryonic stages that of the lens
by the optic cup in amphibians was known before induction of neural
plate by chorda-mesoderm and has been the subject of much experiment
and discussion. ■5''

DEVELOPMENT OF THE AMPHIBIAN LENS

In the anuran map (Fig. 150) the presumptive lens region is at some
distance from the region of the optic primordium, but in the course of
development the optic cup comes to lie beneath the presumptive lens
ectoderm. Under natural conditions the lens develops from the epidermis
as the optic cup comes into contact with it. The amphibian optic vesicle
arises as a hollow lateral outgrowth from the developing brain and re-
mains attached to it by a stalk. Its wall, at first with little regional differ-
ence in thickness, becomes, in part, the thick retinal layer, in part the
much thinner tapetum ; the retinal layer becomes concave lateroventrally,
forming the optic cup, and lens formation occurs by a proliferation and
thickening of the inner layer of the epidermis overlying the concavity of
the optic cup. As the concavity becomes deeper, the developing lens ex-
tends farther into it and separates from the epidermis. Earher stages of
lens formation are indicated in Figure 162. This course of development
very naturally suggests that the optic cup may have something to do
with lens development. This is also indicated by the fact that, in cases
of approximation of eyes and cyclopia, lenses develop in normal position
with respect to the optic cup, but presumably from other epidermis than
the presumptive lens-forming region. Decisive evidence concerning the
role of the optic cup in lens formation has been sought by various lines
of experiment: replacement of presumptive lens epidermis by other epi-
dermis, removal of optic vesicle preceding lens formation, implantation
of optic primordium in other regions, etc.

•5* For fuller discussion of the data than is possible here and for the literature see O. Man-
gold, 193 i(x; Spemann, 1936, 1938.

488

PATTERNS AND PROBLEMS OF DEVELOPMENT

EXPERIMENTAL LENS INDUCTION IN DIFFERENT SPECIES

In the earliest experiments with Rana fiisca as material neither lens
development nor the loss of epidermal pigment characteristic of corneal
development occurred when the presumptive optic region was removed at
the stage of early neural plate, and a similar dependence of lens formation

Fig. 162, A-C. — Stages in development of amphibian lens. A, early optic cup; r, retinal;
t, tapetal layer; and /, epidermal thickening of early lens development; B, C, later stages, in C
formation of fibers beginning (schematized after Rabl, 1898).

on presence of optic cup in other vertebrates seemed probable (Spemann,
1901a). But lack of uniformity soon became evident: lenslike structures
were found to develop in R. palustris after killing the optic primordium
(King, 1905); further experiment showed that a well-formed lens with
fibers may develop in R. esculenta after removal of the optic region from
the neural plate; also, in Bombinator, R.fusca, and R. cateshiana apparent
beginnings of lens development (lentoids) sometimes appear in similar

EMBRYONIC INDUCTORS AND ORGANIZERS 489

experiments." No indications of lens development in absence of the optic
cup have been observed thus far in the urodeles, Pleurodeles (Pasquini,
1927), Triton taeniatus, and T. alpestris (O. Mangold, 1931a); and in
Amhly stoma results of earlier experiments were negative or only slightly
positive (Le Cron, 1907). Another form of experiment hkewise indicates
dependence of lens formation on presence of optic cup in Triton. Removal
of the head mesoderm in the early neurula usually results in approximation
of the eyes to the median plane or in complete cyclopia, and lenses de-
velop in normal relations to the optic cups. This experiment is regarded
as doubly significant: first, because the condition determining cyclopia
occurs at a relatively late stage of development; second, because the
cyclopia results from a local mesodermal defect, while in cyclopia pro-
duced by inhibiting chemical agents, or occurring in the reconstitutional
duplications resulting from ligature in early stages, the whole embryo is
affected, and the locus of the presumptive lens region, as well as position
of the optic cup, may conceivably be altered. When the optic primordium
is reduced in size, either by reconstitution from a part or otherwise, the
lens of R. esculenta is not correspondingly reduced (Spemann, 19126), but
in other forms for which data are available there is more or less close cor-
respondence in size of optic cup and lens.-^^

Transplantation experiments bearing on lens formation may consist in
(a) transplantation of presumptive lens epidermis to other regions of the
body; (b) transplantation of optic vesicle or cup to other regions or ex-
plantation with epidermis from other regions; (c) substitution of other
epidermis for presumptive lens-forming epidermis. All these experiments
have been performed on at least some amphibian species, some of them
on several species.

Transplantation of presumptive lens epidermis of Amblystoma puncta-
tum to other head regions in early and late neurula results in lens forma-
tion in the new location (Harrison, 1920); but other species, including
R. esculenta, have given negative results (Spemann, 1912a). Transplanta-
tions of the optic region of the neural plate (optic plate) or the optic
vesicle of a later stage to other body regions have been made in various
species with induction of lens. The transplanted optic primordium de-
velops and differentiates to an advanced stage, and transplanted pieces
may reconstitute to optic cups of small size. Substitution of other epi-

3' Spemann, 1907, igiia, b; von Ubisch, 1923, 1924, 1925c, 1927; Pasquini, 1931.
3^ W. H. Lewis, igoja, b; Spemann, 1912(7; Wachs, 1919; von Ubisch, 1924; Filatow, 1925;
O. Mangold, 1931a.

490 PATTERNS AND PROBLEMS OF DEVELOPMENT

dermis for that normally overlying the eye by reversing the anteropos-
terior axis of head epidermis, by transplanting epidermis from other re-
gions of head or trunk, or by regeneration of epidermis from adjoining
regions after removal of the presumptive lens epithelium has also given
positive results in many species. ^^ Regenerating tissue of urodele leg and
tail and of anuran tadpole tail transplanted to the eye may form a lens
after removal of the original lens (Schotte and Hummel, 1939). Even
optic vesicles explanted with skin pieces may induce a lens in other than
the presumptive lens epidermis (Perri, 1933).

As the data stand at present, lens can be induced in head epidermis
not immediately adjoining the eye region and in trunk epidermis by the
optic cup in R. fiisca, R. sylvestris, R. temporaria, R. catesbiana, R. ridi-
bunda, R. pipiens, Hyla arborea, in trunk epidermis, probably also in
head epidermis but data are lacking, and in T. taeniatus and Pleurodeles
waltlii. According to earlier experiment (Spemann, 1912a), lens induction
occurred only in head epidermis in Bombinator pachypus and R. esculenta,
but more recently it has been obtained from both head and trunk epi-
dermis in both species. Earlier experiment on A. pundatum gave posi-
tive results only with epidermis adjoining the presumptive lens epidermis,
but Stone and Dinnean (see footnote 39 below) report lens induction in
ventral ectoderm.

In general, reactivity appears to be less in trunk than in head epider-
mis, particularly in regions near the eye. Lenses induced in trunk epi-
dermis are usually smaller and less developed, and apparently less fre-
quent, than in head epidermis, suggesting a relation between the effect
of induction and position of the epidermis in the anteroposterior gradient.
Although nasal and otic primordia of Hynobius are capable of independ-
ent differentiation, the optic cup can induce lens in them by overcoming
whatever determination exists (Ikeda, 1937).

Mangold and Spemann regard the data as indicating that in some
species — for example, R. esculenta and A. pundatum — the presumptive
lens region is determined as lens before the optic cup comes into contact
with the epidermis, while in others it is not so determined or is less deter-
mined, and that, in general, capacity of other epidermis to react to the

39 W. H. Lewis, 1904, 1907a, b; Spemann, 1905, igiia; Ekman, 1914; Harrison, 1920;
Filatow, 1925a, 1934; Beckwith, 1927; Pasquini, 1927, 1932; von Ubisch, 1927; Gostejewa,
1935; etc. Mangold, 1931a, p. 263, gives in tabular form results obtained with different species
on lens formation by other than the presumptive lens epidermis. See also Stone and Dinnean,
1940, "Origin of the lens by induction in the salamander, Amblystoma piinctaUim,'" Proc. Soc.
Exp. Biol. Med., 40.

EMBRYONIC INDUCTORS AND ORGANIZERS 491

optic cup by lens formation is less limited in those species in which the
presumptive lens region is not already determined, and vice versa. Von
Ubisch, on the other hand, maintains that all amphibian species are es-
sentially alike at corresponding developmental stages, as regards lens de-
termination/" It seems that final conclusions concerning this point must
await further investigation. Inhibiting effects of the operative procedures
on transplanted epidermis or optic primordium or on the host tissues,
rate of healing, and various other factors may differ with different ma-
terial and in different experiments. Since reactivity of the epidermis to
the inductor decreases and becomes regionally more limited as develop-
ment progresses, slight difference in stages used by different experimenters
may be responsible for some of the differences in results. Inductive in-
tensity of optic cup may differ in different species; in some it may be
sufficient to induce lens formation only in more anterior levels, in others,
from any level. Epidermal reactivity also appears to differ regionally, and
the anteroposterior differential may be greater in some species than in
others. It is generally agreed that the optic cup influences the lens after
the original inducing action. The lens primordium, or the lens after fiber
development has begun, may develop somewhat further after removal of
the optic cup, when placed in the blastocoel of a gastrula or transplanted
to other regions; but sooner or later its structural pattern is lost and it
degenerates.^'

Recent experiments have placed the whole question of lens potency
and its restriction in a new light. Optic primordia implanted in regenerat-
ing tissue of tadpole or adult urodele tails can induce lens from the re-
generating cells. Regenerating leg tissue of urodele or regenerating tail
tissue of urodele or anuran tadpole transplanted to the lensless eye of an
adult urodele or a large anuran tadpole can give rise to a lens."*^

In other experiments a Nile blue stained ectoderm from the side of a
larva is transplanted to the eye of a larva from which all corneal and
orbital epidermis has been removed. Later implanted epidermis is removed
from over the eye, permitting regeneration from the Nile blue stained
edges of the implant, and lenses or lentoids may develop from the regen-

"t" For a discussion of this point see O. Mangold, 1931a, pp. 288-89. Spemann, 1936, pp.
24-57, and 1938, chap, iii, discusses lens induction at length with numerous figures and refer-
ences.

•" Le Cron, 1907, Amhlystoma; Fischel, 1917, Salamandra; Filatow, 1925c, Triton; Kriiger,
1930, Triton.

-•^ Schotte, 1937, 1938; Schotte and Hummel, 1939.

492 PATTERNS AND PROBLEMS OF DEVELOPMENT

erated tissue (Schotte, 1940). Unless Nile blue has spread from cells of
the implant to those of the host and host cells are involved in the regenera-
tion, it appears that the regenerating flank ectoderm from a hatched larva
is capable of developing lens under inductive action of the eye. Potency
for lens formation, apparently lost in the course of development, may re-
appear in cells undergoing reconstitution. This seems to be a case of de-
differentiation and indicates clearly the relative character of so-called
"determination."

THE QUESTION OF SPECIFICITY IN AMPfflBIAN LENS INDUCTION

In amphibian lens induction, as in neural induction, there is no tissue-,
organ-, or species-specificity. Nonlentogenous epidermis of Bujo vulgaris
transplanted over the optic cup of R. esculenta develops lens. Hetero-
plastic transplantations between Triton species and between Triton and
axolotl also give positive results (Mangold, 1929c; 1931a, p. 277). In
heteroplastic transplantations between Triton species with lenses of dif-
ferent size the induced heteroplastic lens maintains, in general, donor
size, though it may sometimes be smaller in consequence of inhibiting con-
ditions. The optic cup undergoes more or less adjustment to lens size,
apparently through influence of the lens on its growth (Rotmann, 1939).

Various tissues, living or dead, can induce lens — for example, boiled
posterior neural plate, entodermal cells killed by alcohol, fresh liver, and
boiled heart of salamander (Holtfreter, 1934a, h). In Triturns pyrrhogastcr
nasal primordia, otic vesicle, brain, heart, liver, neural plate, dorsal ar-
chenteric wall, ectoderm, mesoderm, and entoderm of head region im-
planted in place of optic primordium induce lens in the presumptive lens
epithelium (Okada and Mikami, 1937). The optic cup can induce other
organs than lens. Implanted below the skin of regenerating tails of large
frog tadpoles, it may induce lens, also olfactory parts, otic vesicle, and
mouth cavity (Schotte, 1937). In this case the optic cup apparently in-
duces something resembling the head region in the regenerating tissue;
doubtless the high level of activity and the slight degree of differentiation
in the regenerating tissue are concerned in the result.

A chemical substance produced by the optic cup or other tissue has
commonly been supposed to be the lens-inductor, but the positive results
with transplantations between species and genera and lens induction by
various tissues, living and dead, raise the same questions as regards speci-
ficity of the inducing factor that arose concerning neural induction. To
what extent activity-level of epidermis or attainment of a certain degree

EMBRYONIC INDUCTORS AND ORGANIZERS 493

of specificity dependent on developmental stage and differing in different
species and at different body-levels may be concerned in reaction of epi-
dermis by lens formation to an optic cup or other living or dead inductor
is not known, and practically nothing is known concerning intensity or
degree of inducing power of optic primordia or of the effects of operative
procedures, transplantation, explantation, etc., upon it.

ORIGIN OF PATTERN IN THE LENS

The lens possesses a polarity normally coincident in direction with the
optic axis. Mangold (1931a, pp. 272-74) regards the optic cup as deter-
mining this polarity but holds that in lens developing independently of
the optic cup the polarity is already determined in the lens primordium.
Spemann (1936, p. 56) believes that the polarity is determined by the
inducing action from one side. Since the lens in early stages is essentially
similar to a bud, its polarity may result from this form of development
rather than from the induction itself. The central region of the bud, pre-
sumably the most active region in early stages, becomes the lens proper,
the peripheral regions forming the capsule. Instead of becoming an elon-
gated axis like most buds, direction of growth of the originally central
region appears to undergo reversal in direction and to be directed toward
the cavity of the lens primordium and the epidermis, whether because of
pressure on the retina or some other factor (Fig. 162). In any case it is
evident that a polarity may originate in relation to the differential activity
represented in the differential growth of the lens primordium and quite
independently of an inductor. The lens primordium buds from the epi-
dermis, and the lens proper buds into the interior of the lens vesicle formed
by separation of the bud from the epidermis, but the axis of the lens is
in the same direction as that of the original bud, though direction of
growth is reversed. The only relation of an inductor to the lens polarity
may be that of initiating a region of activity grading off from a center.
Alterations of lens polarity under experimental conditions offer no diffi-
culties to this conception. Experiments by Dragomirow (1929) show al-
teration in relation to other developing parts than an optic cup. Early
optic cups with presumptive lens epidermis were so transplanted that the
epidermis was between the developing otic primordium and the optic
cup. Lenses developed with two axes at differing angles to each other,
sometimes opposed, one in the usual relation to the optic cup, the other
vertical to the surface of the otic vesicle. Here the effect of the otic ves-

494

PATTERNS AND PROBLEMS OF DEVELOPMENT

Fig. 163. — Diagrammatic, indicating
course of lens fibers and in heavier lines
the two sutures in which the ends of the
fibers meet (after Rabl, i

icle may be merely a differential activation, decreasing radially from the
region most affected.

The lens fibers extend between the two polar regions, forming concen-
tric layers; at each pole they come together along a line, forming a short

suture (Fig. 163). In the amphibian
eye the inner suture, toward the
retina, is anteroposterior, the eye be-
ing regarded as lateral; the outer,
dorsiventral in direction. If presump-
tive lens epidermis, optic plate, or
early optic vesicle is turned 90°, it
is possible, up to a certain stage, to
obtain lenses normal in relation of
sutures to the optic cup; but during
closure of the neural folds, that is,
before the lens develops at all, the
suture pattern is fixed and not al-
tered by turning (Woerdeman, 1934).
Apparently, then, although it may
be altered by an inductor up to a
certain stage, it is determined in relation to the general pattern quite in-
dependently of the inductor, perhaps in relation to dorsiventral or antero-
posterior gradient pattern or both.

LENS DEVELOPMENT IN OTHER VERTEBRATES

Lenses appear frequently, either in the normal location or in other
head regions without relation to eyes or contact with nervous tissue, in
teratological embryos of teleost fishes, both those occurring occasionally
in nature and those experimentally produced. They may even develop
in anophthalmic forms, and several supernumerary lenses may be present
in an individual ; they vary in size and degree of development but may
be large and advanced in differentiation.-^' It seems beyond question that
some of these lenses develop from other than presumptive lens epidermis.
The factors concerned in their origin are not known, but their develop-
ment suggests a condition in the epidermis associated with stage of de-
velopment and region of body rather than with action of a specific induc-
tor. According to Werber, however, blastolysis, that is, a dissociation of
parts of organ primordia, occurred in consequence of exposure to inhibit-
« Mencl, 1903, 1908; Gemmill, 1906a, b; Stockard, 1909, 19106, c; Werber, 1916a, b, 1918.

EMBRYONIC INDUCTORS AND ORGANIZERS 495

ing chemical agents, in his experiments chiefly to butyric acid and acetone.
The free lenses were assumed to be induced by fragments of the optic
primordia scattered through the head region in consequence of blastolysis;
and the hypothesis that the vertebrate lens is induced in all cases was
advanced, but conclusive evidence supporting it is lacking. Lenses also
appear in normal relation to the optic cups of teratological forms when
these reach the epidermis, but experiments demonstrating induction seem
to be lacking. That the normally lentogenous teleost epidermis can de-
velop lens in absence of optic cup is evident from some of these teratologi-
cal forms, and there is little doubt that the supernumerary lenses develop
from other epidermis. The teratological forms concerned are differential
modifications of development. It may be suggested that in those experi-
mentally produced the differentially inhibiting action of the agent de-
creases or perhaps abolishes the differential in the head region. With the
activation associated with recovery following return to water, the inte-
grating and ordering factors in the head region being less effective than
normally and different regions being more alike, local activations in the
epidermis at certain stages, either induced by underlying parts or per-
haps originating in the epidermis, may bring about formation of several
lenses independent of eyes. The appearance of supernumerary adventi-
tious polarities in Corymorpha with essentially similar treatment may be
recalled (pp. 416-17).

Data bearing on lens induction in the chick are few, but it has been
shown that in stages from primitive streak to four somites the optic cup
can induce lens in ectoderm of head, neck, or trunk, in later stages only
in head and neck. Up to five lenses or lentoids may develop in relation to
one transplant (Alexander, 1937). Earlier experiments also show or indi-
cate induction in the normally lentogenous or adjoining epidermis.-"^

OTHER INDUCTIONS IN VERTEBRATE DEVELOPMENT

As development progresses, other inductions of organ systems and
organs, most extensively studied in amphibians, take place in particular
parts of the body or under experimental conditions at certain develop-
mental stages. Some of these have been mentioned in connection with the
field concept (chap, viii), and certain others are briefly considered here.

RETINA AND CONJUNCTIVA

Although retinal and tapetal parts of the optic primordium become
visibly different in early stages (Fig. 162, A), the tapetum may still give
■tt DanchakofE, 1924, 1926; Reverberi, 1929; Willier and Rawles, 1931; Rawles, 1936.

496 PATTERNS AND PROBLEMS OF DEVELOPMENT

rise to retina when brought into contact with epithehum of the developing
otic vesicle (Dragomirov, 1937; Ikeda, 1937). The thinning and loss of
pigment of the epidermis in development of the conjunctiva is induced
by an optic vesicle or cup, by a fragment of the retina, or by a lens. De-
velopment of conjunctiva from epidermis of other regions can be induced,
either by transplantation of optic vesicles or by grafting other epidermis
over the optic primordium, provided there is contact between the two.^s
Moreover, the reaction is not species-specific, at least not for various uro-
dele species (Mangold, 1929c; Harrison, 1929a), and a hmb bud implanted
in place of the optic cup at certain stages apparently induces conjunctival
development (Durken, 19 16), though this has been questioned (H. Peter-
sen, 1924).

THE EAR

Pieces of hindbrain or medulla implanted near the ear region or even
in the flank induce development of otic structures; this induction is effec-
tive between R. pipiens and yl. punctatum in either direction.-"^ Presump-
tive medulla, but not forebrain or spinal cord, implanted in regenerating
tissue of the tadpole tail, induces vesicles resembhng otic vesicles (H. S.
Emerson, 1939). The ability to induce ear is itself induced in presumptive
hindbrain by the underlying mesoderm, and this mesoderm can induce
ear development directly in foreign ectoderm in absence of hindbrain
(Harrison, 1935). The optic primordium induces otic vesicles among other
organs in regenerating tadpole tail tissue, but perhaps only indirectly
(Schotte, 1937). After removal of the presumptive otic region the ear
may develop from regenerated ectoderm or from ectoderm transplanted
from other body regions and even from anurans to Triton^'' On the other
hand, the earlier stages of ear development from presumptive otic ecto-
derm of neurulae are apparently independent of other organs, but induc-
tion is necessary for full differentiation (Kaan, 1926). If this is true, the
ear is primarily determined as a locus in the general pattern of the ecto-
derm without induction; but induction can determine ear development in
ectoderm of other regions, even in different species, genera, and orders of
amphibia, raising once more the question of specificity of the inducing
action. If otic induction is primarily an activation, it is activation at a
certain period of development when the reacting tissue has doubtless
attained a condition dift'erent from that of early stages. The same induc-

« Fischel, 1900Z), 1917, 1919; Spemann, 1901a; W. H. Lewis, 1905; Groll, 1924.
■i^L. S. Stone, 1931; Albaum and Nestler, 1937; Ponomarewa, 1936.
^T Amblystoma: Tokura, 1925; Kaan, 1926; Yntema, 1933; Harrison, 1935, 1936, anura to
Triton: G. A. Schmidt, 1936c.

EMBRYONIC INDUCTORS AND ORGANIZERS 497

ing factor may apparently have different effects on the same ectodermal
region at different developmental stages.

The general gradient pattern of the body is apparently a factor in the
results of heterotopic transplantations. The farther ventral or posterior
to the otic region the place of origin of ectoderm transplanted to the
otic region, the less otic development takes place. Transplants of pre-
sumptive otic ectoderm develop more completely at a level between ear
and eye than at levels some distance posterior and ventral to the otic
region, and the most complete development at the anterior level occurs
at an earlier host stage than at the more posterior and ventral levels
(Yntema, 1933). The anterior level apparently attains a condition favor-
able to, or less inhibitory to, ear development earlier than the posterior
and ventral levels. These data certainly suggest definite relation to the
anteroposterior and dorsiventral gradient pattern. The differences in otic
development in ectoderm from other than the otic region have commonly
been regarded as indicating differences in potency, but they may be
merely differences in susceptibihty or reactivity to the inducing factor.
Otic development in regenerating tail tissue, following implantation of
optic primordium or presumptive medulla, suggests, first, that the re-
generating tissue has undergone a considerable dedifferentiation and pos-
sesses, or has regained, potencies supposedly lost, and, second, that even
here different embryonic regions may serve as inductors.

It is generally agreed that the otic vesicle induces aggregation and
flattening of mesenchyme cells about itself and that these give rise to the
cartilaginous capsule. Thai this mesenchymal reaction is not primarily
a specific determination and may be altered by change in conditions is
emphasized by Filatow (1927). He regards it as not widely different
from the inflammatory reaction following implantation of a foreign body
or, in later stages, even of an otic vesicle. Under influence of an implanted
otic vesicle primordia of other organs — a vertebra (Yntema, 1933), and
in the fish, Acipenser, shoulder girdle (Filatow, 1930) — show more or
less capsular reaction. However, it appears that the capsular reaction
may be absent when otic primordia are transplanted to other than the
normal otic region.''^ This induction may be chemical in character, as far
as accumulation of mesenchyme is concerned; but, in view of the wide
occurrence of cartilage development, it can scarcely be regarded as specific
for the otic vesicle, except perhaps mechanically in determining form of
the capsule.

Development of the amphibian tympanic membrane is normally de-

48 L. S. Stone, 1926; Luther, 1927; Balinsky, 1929; Kaan, 1930; Yntema, 1933.

498 PATTERNS AND PROBLEMS OF DEVELOPMENT

pendent on the annular tympanic cartilage. When this is removed, the
membrane does not develop; when it is transplanted to other regions,
membranes form over it. This induction occurs in any region of the skin
before and after metamorphosis. The transplanted quadrate will induce
tympanic membrane, and the suprascapula shaped into annular form is
probably slightly inductive (Helff, 1938; 1934^^, b).

Skin of fully developed dermal plicae with mucous and other glands
from postmetamorphic R. palustris transplanted to the backs of larvae
of R. cateshiana undergoes complete regression of the plicate structure to
generaHzed skin. Retransplanted after this regression to the otic region
of metamorphosing R. clamitans or R. cateshiana, it forms fully differen-
tiated tympanic membrane (Helff, 1934c). Whether, or to what extent,
mechanical or chemical factors, activating or specific, are concerned here
is not yet evident.

According to the data, three successive inductions are involved in de-
velopment of the amphibian ear: the primary neural induction, induction
of the otic vesicle by hindbrain regions, and induction of the capsule by
the vesicle; probably other inductions are involved in later development
of its parts. But, as in case of the lens of at least some amphibians and
apparently of the neural plate, some degree of definition or determination
as the locus of particular developmental events occurs in the presumptive
otic region in relation to the general pattern of earlier stages before these
inductions take place. If that general pattern is a gradient pattern, the
otic region is primarily defined by its gradient co-ordinates, and induction
by hindbrain or other inductor may be merely an acceleration or intensi-
fication of changes already initiated by position in the gradient pattern.
The inductor, according to this view, is primarily an assisting factor in
bringing the reacting region to a physiological level at which a certain
kind of differentiation, already initiated, can continue to a certain result.
As inductor and reacting region differentiate, there may be definite spe-
cific relations between them.

BALANCER, ORAL REGION, GILLS

Early larval stages of most urodeles possess so-called ''balancers," slen-
der, elongated, tentacle-like organs, slightly club shaped at the tips, con-
sisting of epidermis and mesodermal core derived from the neural crest,
and situated ventral and somewhat posterior to the optic region. From
early neurula stages of Amhly stoma on, presumptive balancer ectoderm
can develop a balancer when transplanted alone to other head regions.

EMBRYONIC INDUCTORS AND ORGANIZERS 499

but not until later stages when transplanted to the trunk. This difference
is attributed to the mesoderm of the head region (Harrison, 1925^). The
anterior part of the archenteric roof placed in the blastocoel of the early
Triton gastrula comes to lie ventrally and induces balancer there. Al-
though the axolotl possesses no balancer or only traces, inductor trans-
plants to Triton may induce balancer development (Mangold, 193 1&).
Heteroplastic transplants of presumptive balancer ectoderm from
A . punctatum to the corresponding region of A . tigrimim, which has no bal-
ancer, may develop balancer; but ectoderm of A. tigrinum., transplanted
to the balancer region of A. punctatum, does not give rise to balancer
(Harrison, 1925&). Heteroplastic transplants of presumptive balancer ec-
toderm and of ventral ectoderm to the balancer region between Triton
species give balancers characteristic of the donor species. The reaction,
but not the inducing factor, is species-specific (Rotmann, 1934, 19356).
Neural crest material of R. fusca in the blastocoel of Triton and axolotl
may induce neural plate and various modifications in the host, among
them often supernumerary balancers, but always in the balancer region
(Raven, 1931, 1933&).

The urodele mouth has small teeth, the anuran mouth horny jaws;
in anuran larvae two suckers are present ventral to the mouth, but no
balancers. The entoderm underlying the presumptive oral ectoderm is
necessary for formation of the oral region in Triton and axolotl. Presump-
tive oral ectoderm of neurulae, transplanted or explanted alone, does not
form a mouth (Stroer, 1933). Presumptive abdominal ectoderm of the
anuran gastrula, transplanted to the urodele mouth region, gives rise to
an anuran mouth with horny jaws and suckers; abdominal urodele ecto-
derm in the anuran mouth region develops a urodele mouth with teeth,
and, if the transplant includes the balancer region, balancers, sometimes
supernumerary, develop. Both balancer and sucker may develop in some
cases, one from the host, the other from the donor ectoderm; and, if the
boundary between urodele and anuran ectoderm is in the region of the
anuran sucker, part of a sucker develops on the anuran side, none on the
urodele side, the development ceasing sharply at the junction.''^

Presumptive anuran gill ectoderm, transplanted alone to other regions
in neurula or tail-bud stages, develops gill stumps, and the gill pattern
in transplants turned 180° is similarly turned. Ectoderm adjoining the
gill region is also capable of gill development (Ekman, 1913a, b; 1922).

■"Spemann und Schotte, 1932; Schotte, 1932, 1933; Rotmann, 19350.; Holtfreter, 1936;
G. A. Schmidt, 1937a.

500 PATTERNS AND PROBLEMS OF DEVELOPMENT

Transplanted gill ectoderm of Amhlystoma shows less evidence of gill de-
velopment (Harrison, 192 1&), but after removal of gill ectoderm and meso-
derm and implantation of a limb bud in the gill region some gill develop-
ment may occur in ectoderm adjoining and usually ventral to the gill
region (Detwiler, 1922). After removal of presumptive branchial ento-
derm in tail-bud stages the gill does not develop, and this entoderm,
transplanted beneath ectoderm and mesoderm slightly posterior and ven-
tral to the gill region, can determine gill development (Severinghaus,
1930; Ichikawa, 1934, 1938). Induction of gill development may occur
in presumptive neural plate or in ventral ectoderm transplanted hetero-
plastically to the gill region between Triton species (Spemann und Rot-
mann, 1931), or from axolotl to Triton (Rotmann, 1935c), or even xeno-
plastically, from an anuran, Bombinator igneus, to Triton (G. A. Schmidt,
1937&). The induced gills show donor characteristics more or less clearly
in time of appearance, size, time of degeneration, and, when differences
between donor and host are considerable, by lack of correspondence be-
tween gills and underlying mesentoderm. However, some modification by
the host of size and perhaps of period of persistence apparently takes
place. In general, the experimental results indicate that presumptive
branchial ectoderm of the neurula and tail-bud stages possesses some ca-
pacity for independent gill development, apparently less in urodele than
in anuran species; but for complete gill development the underlying ento-
derm or mesentoderm seems to be necessary.

"double assurance" in amphibian development
Implanted pieces of the dorsal lip can induce development of a neural
plate from ectoderm entirely outside the presumptive neural region, and
induction by the underlying chorda-mesoderm is supposedly a factor in
neural-plate formation in normal development. However, as we have seen,
there is considerable evidence that the presumptive neural plate, or at
least its anterior part, is in some degree "determined," predisposed to the
course of development resulting in neural plate, and under certain condi-
tions can undergo more or less development along this line independently
of an underlying inductor. Similarly, lens development can be induced in
ectoderm of various regions by an optic cup; but in some amphibians a
lens can develop in the presumptive lens region quite independently of
an optic cup, and in others some degree of determination or predisposition
toward lens formation is apparently present in the presumptive lens epi-
dermis. These and other cases of apparently twofold determination Spe-
mann regards as examples of the principle of double assurance or the

EMBRYONIC INDUCTORS AND ORGANIZERS 501

synergetic principle in development. s" Development of the part is doubly
assured, on the one hand, by its own determination and, on the other, by
the inductor. It must be pointed out, however, that there is no double
assurance unless the part concerned can actually develop independently
of, as well as in consequence of, the action of the inductor. As far as we
know, a complete, normal neural plate does not develop independently of
an inductor. The neural tissue or neural tube developing independently
of an inductor could not serve an individual as a central nervous system.
Its independent development is without value to the organism. The lens
of some amphibian species apparently constitutes a better example of
double assurance, but in other species double assurance is apparently
lacking. In the teleost embryo under inhibiting conditions several lenses
may develop independently of an eye, seemingly a multiple assurance
but without significance for the normal organism.

But even if it be granted that both independent differentiation and de-
velopment by induction are concerned in certain parts, the principle of
double assurance throws no light on the problem of the organization which
initiates independent development or of the factors concerned in induc-
tion. It is a teleological, not a physiological, principle. We must, then,
ask the question: What does double assurance mean physiologically?
Does it mean anything more than that the pattern within which a greater
or less degree of independent development or determination takes place
and the inductor act on the part concerned in very much the same man-
ner, though perhaps in different degree or intensity? Assuming, for the
moment, that the primary pattern of organization is a gradient pattern,
the presumptive neural plate represents a certain region, a certain range
of levels, in that pattern; and whatever independent development is pos-
sible results from its position in, and relation to, that pattern. The inductor
apparently represents a higher range of gradient-levels and, by raising
the level of presumptive neural plate, makes possible further development.
But apparently almost any stimulating or irritating factor acts in some
degree like the natural inductor, though those factors are usually without
axiate pattern, and any axiate pattern developing must be of ectodermal
origin. These suggestions do not necessarily involve the assumption that
the invaginated chorda-mesoderm differs only quantitatively from the
presumptive neural plate. It may already be specifically different when

so Spemann, 1931, pp. 508-10; 1936, p. 59; 1938, p. 92. For earlier biological applications
of this principle see Rhumbler, 1897; Braus, 1914. For the principle of kombinative Einheits-
leistuHg, a somewhat broader concept than double assurance, see Lehmann, 1928, and later
papers cited in Bibliography.

502 PATTERNS AND PROBLEMS OF DEVELOPMENT

it invaginates, but the data of neural induction do not support the view
that specific characteristics of the chorda-mesoderm are directly concerned
in determining that overlying ectoderm shall become neural tissue. They
may play a part in determining features of the pattern of its further de-
velopment and differentiation; but other tissues, living or dead, from
many organisms, tissue extracts, and residues and various synthetic sub-
stances can induce development of neural tissue from ectoderm. Again
the question arises: Is this induction anything more primarily than acti-
vation? If this induction is not specific, the chorda-mesoderm is only in-
directly neural inductor; it activates ectoderm, and the activated ecto-
derm becomes neural tissue. Later in development, when ectoderm has
attained a somewhat different condition, an activation may result in de-
velopment of a lens, at least in head ectoderm. The epidermis has very
probably undergone some degree of differentiation which makes lens de-
velopment possible with sufficient activation by an inductor or otherwise.

PRESENT STATUS OF THE INDUCTOR PROBLEM

By way of conclusion to this chapter the attempt is made to indicate
briefly the conception of induction to which the evidence, at present
available, points. It is impossible to draw a hard and fast line between
what is quantitative and what is specific in living protoplasms; but, ac-
cording to the evidence considered in preceding chapters, axiate organis-
mic pattern in its early stages appears to be predominantly or wholly a
quantitatively graded pattern, always, of course, in a protoplasm of spe-
cific constitution. Nonspecific alterations of that pattern alter the course
of development. Even a difference in oxygen tension may determine the
physiological and morphological differences between a stolon and a hy-
dranth in certain hydroids (pp. 172-75).

In general, the natural inductors of early stages are, or at certain de-
velopmental stages become, high gradient-levels, and their inducing action
apparently consists in alteration of condition in regions representing
lower levels. This alteration evidently involves activation; and, with the
progress of experimental analysis, it appears increasingly probable that
from the hydroids to the vertebrates the primary factor in the inductions
of early development, whether reconstitutional or embryonic, is activa-
tion. Amphibian head-, trunk-, and tail-inductors, for example, represent
different gradient-levels of the inductor region ; and there is no conclusive
evidence that at the time of gastrulation they differ in other ways, though
they may already be different from the ectoderm. Moreover, their indue-

EMBRYONIC INDUCTORS AND ORGANIZERS 503

tive effects are not strictly regionally specific but show gradations depend-
ing on gradient-levels in the reacting ectoderm as well as in the inductor
region. The head-inductor may induce trunk; the trunk-inductor, head;
and between a normal head and a trunk without head there is a graded
series of defective heads, as in the planarian, though not of the same
types. Moreover, the great number and variety of inductors of neural
tissue in amphibians does not support the view that the inductor is a
specific chemical substance with a specific effect; in fact, this possibihty
seems to be excluded. If neural induction is primarily a localized, non-
specific activation of ectoderm, effectiveness of many different inductors
is to be expected and seems to have been demonstrated.

In so far as an inductor is a nonspecific activator, it is not directly an
organizer or organization center. The natural inductor is itself a product
of pattern and organization, and as inductor it alters pattern relations in
other parts; but the pattern thus altered is the real organizer or reorgan-
izer. The inductor comes nearer to being a real organizer when it deter-
mines an entirely new pattern, as in hydroids and planarians, where the
region activated by section at the proximal end of an isolated piece or a
small graft determines a new polarity at right angles to, or opposed to,
the original polarity. In these cases the inductor apparently acts solely
as activated region without pattern and determines a gradient pattern on
a smaller or larger scale. Even here, however, the activated region or
graft is only indirectly organizer through the gradient pattern. When an
implanted piece of the dorsal lip induces in a vertebrate an axis in a dif-
ferent direction from, perhaps opposed to, the host axis, this is not en-
tirely new but is apparently an imposition on the ectoderm of the axiate
pattern of the inductor, and this pattern is the organizer.

Even if the earliest inductions in development are nonspecific activa-
tions or begin as such, it does not follow that all inductions have this char-
acter. With progress of differentiation induction may become more spe-
cific, both as regards inducing factor and reacting system. Certain hor-
mones are evidently inducing factors. They are specific substances pro-
duced by specifically differentiated organs and producing certain effects
on other specifically differentiated organs, but even in some of these cases
the question whether the effect is primarily an activation of the reacting
organ or direct determination of specific differentiation is open. In any
case, these specific inductors of later stages are scarcely to be regarded
as organizers. Their effects result from presence of a specific organization,
not from its absence.

CHAPTER XIII

CERTAIN EMBRYONIC RECONSTITUTIONS IN RELATION
TO PRE-EXISTING PATTERN

MANY of the regional differences in reconstitution of isolated
embryonic parts are evidently expressions of the axiate pat-
tern already present and therefore of interest in relation to
the problem of its nature. This chapter is concerned with some of these,
particularly in forms with relatively high capacity for reconstitution in
earlier stages.

COELENTERATES

Although embryonic stages and planulae of this group should be in-
teresting in this connection, experimental data are few. According to Zoja
(1895), hydroids may develop from 1/2, 1/4, 1/8, and even from 1/16
blastomeres of certain forms, but nothing is known concerning differences
in capacity in relation to the original pattern. However, if 1/8 and 1/16
blastomeres can develop into planulae that give rise to hydroids, there
is probably little or no regional differentiation in the pattern. How a 1/16
blastomere attains the pattern of the whole is an interesting question.
In view of the instability of polarity in the embryo and planula of Phia-
lidium (pp. 167, 425), it seems possible that in some of these isolated blas-
tomeres a new polar pattern may be determined by the differential be-
tween free surface and surface in contact.

In some other forms pieces of undivided eggs above a certain size and
1/2 and 1/4 blastomeres can become wholes, but 1/8 blastomeres result-
ing from equatorial division become abnormal larvae or show little de-
velopment.' Pieces of some hydrozoan planulae give rise to hydroids
(Maas, Child), but pieces of planulae of the anthozoan Renilla do not
develop beyond the swimming stage. These fragmentary data throw little
light on the question of the original pattern in these forms. Limitation
or absence of reconstitution in 1/8 blastomeres or in planula pieces sug-
gests some degree of polar differentiation but may perhaps be due to scale
of polar organization too large for the blastomere or piece rather than to

I Maas, 1901, 1905, 1908; E. B. Wilson, 19036; Conklin, 1908.

504

EMBRYONIC RECONSTITUTIONS 505

restricted potency. Further experiment with developmental stages of this
group is necessary for definite conclusions.

ECHINODERMS

Some of the earliest experiments on embryonic reconstitution were per-
formed with sea-urchin eggs and cleavage stages, and their high reconsti-
tutional capacities aroused great interest and led to extensive further ex-
perimentation. These early experiments appeared particularly significant
because they came at a time when the predeterministic Weismannian
theory of development was current and when studies of cell lineage in an-
nelids and mollusks indicated a high degree of predetermination in those
forms. ^

Early stages of sea-urchin development were regarded as constituting
a harmonious-equipotential system by Driesch (1899, 1901), but with the
progress of experiment some data seemed to point to a different conclu-
sion. In early experiments parts of sea-urchin eggs and blastomeres were
isolated by section or by shaking, and it was shown by various experi-
menters (Boveri, Driesch, Loeb, Morgan, Delage, and others) that nu-
cleated and even nonnucleated fragments of unfertilized eggs above a cer-
tain size could be fertilized with sperm of the same or of another species
and might develop and that whole eggs and nucleated fragments could
be activated to parthenogenetic development by various agents. These
experiments, however, gave little or no information concerning develop-
mental pattern because the portion of the egg represented by a particular
fragment was usually not known. They did indicate, however, that con-
siderable reconstitution was possible. In recent years development of
techniques of isolation of blastomeres and of local vital staining has made
it possible to obtain much more exact information concerning regional
differences in reconstitution in these eggs and embryos than earlier meth-
ods permitted. Single blastomeres and groups can be isolated as desired,
locally stained for orientation or identification, and transplanted in de-
sired locations and combinations.-'

^ In view of the more definite results obtained in recent years with improved techniques,
discussion of the earher work appears unnecessary. Among the numerous workers who con-
tributed to this field, the following are mentioned with dates indicating approximately the
periods covered by their papers, but the papers are given in the Bibliography only when par-
ticularly referred to in the text. Driesch, 1891-1923; Boveri, 1889-1914; Morgan, 1894-1908;

Ziegler, 1896-1925; Garbowski, 1904-10; Jenkinson, 19116; von Ubisch, 1913 ; Runnstrom,

1914 . For bibliography to 1928 see Schleip, 1929.

i See, e.g., von Ubisch, 1925a, b, 1929, 1932a, b, 1936; Horstadius, 1928a, b, 1935, 1936a, b,
1937a; Horstadius und Wolsky, 1936; also pp. 436-46.

5o6 PATTERNS AND PROBLEMS OF DEVELOPMENT

Experimentation along these lines has given results of much interest
in relation to the problem of pattern, though it has not yet provided a
basis for complete agreement as to the character of that pattern. Discus-
sion of the regional differences in reconstitution in relation to the axiate
pattern of the sea-urchin egg is based largely on the extensive experiments
of Horstadius from 1928 on. Here again, use of the terms "animal" and
"vegetal" will be convenient in certain connections. Also, it will be re-
called that, according to the data on differential susceptibility and differ-
ential dye reduction, the apical or "animal" pole is the high end, the basal
or "vegetal" pole the low end, of the primary polar gradient. According
to the Runnstrom hypothesis, animal and vegetal poles are, respectively,
the regions of highest concentrations of animal and vegetal substances.''

Animal halves, the eight mesomeres of the sixteen-cell stage, or
on, + an, of a later stage (Fig. 145 [p. 438]) represent about two-thirds
of the presumptive ectodermal region. They usually give rise, when iso-
lated from basal halves, to apical partial forms without entoderm or mesen-
chyme but with variation in scale of organization from the extreme apical
tjqDe, with the long cilia normally characteristic of the apical region ex-
tending over most of the surface, to less extreme forms (Fig. 146, A-E
[p. 439]). Isolated awi rings, that is, approximately the most apical fourth,
become extreme apical partial forms, usually completely covered in earher
stages with the long ciha and later uniformly ciliated (Fig. 147, an,
[p. 441]). Reconstitution of isolated an, rings is less extremely apical and
very similar to that of the apical half (Fig. 147, an, [p. 441])- These iso-
lated apical halves and an, and an, are more apical in reconstitutional de-
velopment than in normal development. They may develop as extreme
apical regions, and some of them are apparently almost or entirely apolar
or anaxiate. According to Horstadius (1935), this is because the animal
gradient gains the upper hand over the vegetal; but how this takes place
is not made clear. Presumably concentrations of hypothetical animal and
vegetal substances are not immediately altered by isolation. What deter-
mines the assumed alteration? The results suggest that some effect of
regions farther basal is present in normal development of the whole em-
bryo but absent in the isolated apical parts. The observed primary polar
gradient, decreasing basipetally, and the activation in the basal region
preceding gastrulation, giving rise to a secondary gradient opposite in
direction to the primary and partly obliterating it (pp. 134-40) provide,
a somewhat different basis for interpretation of the experimental results.

■» See pp. 241, 243, and chap, vi in general.

EMBRYONIC RECONSTITUTIONS 507

In the isolated apical half, or an, or an^, the primary gradient is not partly
obliterated or reversed by the secondary, resulting from basal activation ;
consequently, the scale of polar organization becomes larger than in nor-
mal development and larger than the apical half or the fourths, and they
develop as more or less extreme apical partial forms. These apical partial
forms are essentially similar in origin to the apical partial forms in the
reconstitution of Tuhularia and Corymorpha pieces. Short pieces of these
forms with a single gradient reconstitute unipolar apical partial forms
which usually have a much larger scale of organization than bipolar partial
forms developing from pieces of the same length with two opposed gra-
dients (Fig. 116 [p. 346]). The later the developmental stage at which
isolation of apical halves or fourths occurs, the less is the change in scale
of organization from normal development (Horstadius, 1936a). This, of
course, is to be expected if there is a secondary basal activation and altera-
tion of the primary gradient.

That the apical overdevelopment of apical halves and fourths is really
due to the large scale of organization determined in the isolates, rather
than to restriction of potency for entoderm and mesenchyme, is shown by
the fact that after temporary exposure to differentially inhibiting condi-
tions (lithium salts) apical halves may develop primary mesenchyme and
entoderm, gastrulate, develop skeleton, and become plutei; or in some
cases they may even become exogastrulae (von Ubisch, 1925a, b, 1929;
Horstadius, 1936a). Runnstrom and Horstadius regard lithium as exert-
ing a specific effect in increasing the vegetal gradient or substance, but it
has been shown that after the secondary basal activation occurs lithium
produces modifications in the opposite direction and that other agents be-
sides lithium produce entodermization (pp. 228-33). Exposure to lithium
salts involves two factors — the direct, differentially inhibiting action, and
the recovery after return to water, which may also be differential. In
early stages hthium inhibits the higher apical levels of apical halves more
than the basal levels; consequently, scale of organization of these regions
is decreased and extent of extreme apical development is therefore less
than without the inhibiting action. Basal regions of the apical halves
are also inhibited to a lesser degree, but sufficiently to bring them down
to the gradient-levels of presumptive entoderm and mesenchyme in early
stages, that is, entodermization of presumptive ectoderm results. With re-
covery after return to water apical dominance is inadequate to prevent
the secondary basal activation, and mesenchyme formation and gastrula-
tion take place much as in the normal individual. With sufficient inhibi-

5o8 PATTERNS AND PROBLEMS OF DEVELOPMENT

tion exogastrulation may result. According to this suggestion, the effect
of Hthium in these cases is merely lowering of the levels of the primary
gradient and decreasing dominance of the apical region, so that the basal
activation is possible within the smaller scale of organization. The de-
crease in scale of these apical halves is quite similar to the decrease in
scale by inhibiting conditions in the reconstitution of Tuhularia and
Corymorpha pieces. Pieces giving rise to apical partial forms in well-
aerated water may develop as complete individuals after certain degrees
of inhibition. At present the evidence for a specific effect of lithium on a
vegetal substance or gradient does not appear convincing. Apical halves
of sea-urchin embryos occasionally reconstitute entoderm and mesen-
chyme and gastrulate in normal environment, according to von Ubisch
(1936a, h). He suggests that this takes place only under the most favor-
able conditions, but in the hght of other experimental evidence it appears
probable that such development results from slightly inhibiting conditions
producing lower levels of metabolism and consequently smaller scale of
organization.

Reconstitution of isolated basal halves is essentially similar to reconsti-
tution of a piece of hydroid stem or planarian body after removal of apical
or anterior regions, but Runnstrom and Horstadius have interpreted it in
terms of the two opposed concentration gradients.^ Basal halves include
the most basal part of the presumptive ectoderm {veg^, Fig. 145), the
whole presumptive entoderm {veg^, Fig. 145), and the primary mesen-
chyme. According to Horstadius, they give rise to a variety of forms
ranging from normal plutei, through various modifications of pluteus form
and more or less ovoid larvae with gut proportionally too large and with
abnormal, often excessive skeleton, to exogastrulae. The teratological
forms resemble forms resulting from differential inhibition by various
agents. In the forms which approach or attain normal proportions there
is evidently considerable reconstitution of ectoderm. In apical halves the
animal gradient is supposed to gain the upper hand, according to Hor-
stadius; one might expect, then, that in the basal halves the vegetal gra-
dient would gain the upper hand, but it certainly does not in the forms
that reconstitute ectoderm. These cases also seem to be similar to post-
embryonic reconstitutions in the lower invertebrates. After isolation from
the higher gradient-levels of the apical half more or less activation of the
more apical levels of the basal half apparently increases scale of the ecto-
derm and raises gradient-levels so that apical regions can develop. This

s Runnstrom, 1928a; Horstadius, 1928a, h, 1935.

EMBRYONIC RECONSTITUTIONS 509

alteration appears essentially similar to the activation of levels of the hy-
droid stem or the planarian or annelid body after isolation to a condition
making possible hydranth or head development.

On the other hand, removal of the higher levels of the primary gradient
may result in earlier development and greater extent of the secondary
gradient. The forms with too large gut and excess skeleton and the
exogastrulae probably represent cases in which the secondary gradient
has "gained the upper hand" to greater or less extent and partly inhibited
or obliterated the primary gradient. This series of forms is interesting in
comparison with the series of inhibited head forms in planarian reconsti-
tution. There the inhibition of head development results from the activa-
tion following section at the posterior end of the piece. In the basal sea-
urchin half it apparently results from activation and development of a
secondary gradient in the basal region, but the effects on the primary
gradient show interesting similarities in the two cases.

Reconstitution of isolated vegi and veg2 (see Figs. 145, 147) has also been
followed by Horstadius (1935). The veg^ ring represents the most basal
third of presumptive ectoderm and may develop as an entirely ectodermal
blastula but less extremely apical than the apical half; or it may reconsti-
tute primary mesenchyme and entoderm, gastrulate, and sometimes ap-
proach normal pluteus form, but development is slow. Horstadius ac-
counts for the differences in development by the assumption of variation
of the third cleavage plane from equatorial to subequatorial, so that
vegi is more basal in some individuals than in others. There may be such
variation, but the assumption is necessary only for the hypothesis of
overlapping substance gradients. In terms of primary and secondary met-
abolic gradients the range of variation in form is accounted for by quan-
titative differences in metabolism and in degree of dominance and of basal
activation, either characteristic of different eggs or resulting from differ-
ences in experimental procedure and environmental conditions.

Isolated veg2 rings, though entirely entodermal in normal development,
reconstitute ectoderm but do not develop apical tuft or stomodeum which
are characteristic of the more apical ectodermal levels. Apparently the
activation in the primary gradient following isolation is less than in veg^,
again a parallel to hydroid and planarian reconstitution. Here, again, the
vegetal gradient does not "gain the upper hand," though it might be ex-
pected to do so, according to Horstadius' interpretation of various other
cases. However, these veg2 rings develop relatively large entoderm and
some become exogastrulae; but skeletal development is slight, probably

5IO

PATTERNS AND PROBLEMS OF DEVELOPMENT

because ectodermal development does not proceed far. Reconstitution of
ectoderm is difficult to account for in terms of the Runnstrom hypothesis,
but in terms of a quantitative primary gradient it is again essentially
similar to what happens in isolated pieces from the lower levels of hydroid
stems and planarian bodies.

Isolation of the whole presumptive ectoderm {an^, an^, veg„ Fig. 145)
results in forms either wholly ectodermal, with ciliated band surrounding
an oral region, a stomodeum, but no oral lobe or arms and no skeleton
(Fig. 164, A, B), or with small entoderm and rudimentary skeleton (Fig.
164, C, D). These isolated parts represent approximately three-fourths

Fig. 164, A-D. — Range of forms resulting from reconstitution of whole presumptive ecto-
derm (ani, anz, vegi) of Paracentrotus. A , B, with large scale of organization along primary
gradient; C, D, with smaller scale of organization (after Horstadius, 1935).

of the polar axis; consequently, concentration of vegetal substance should
be considerable in their basal regions, but many of them develop no trace
of entoderm. Moreover, veg-, alone develops entoderm more frequently
and may approach the normal pluteus form, as shown above. How is the
very different development of the same axial level in the two cases to be
accounted for? According to Horstadius, there is lack of balance between
the two gradients in the isolated whole ectoderm, and the animal gradient
may suppress the vegetal, while in isolated vegi the two are more nearly
balanced. In terms of metabolic gradients the ectodermal gradient and
dominance may persist in the apical three-fourths because with good
physiological and external conditions there is nothing to determine ento-
dermal levels. With less vigorous animals or with slightly inhibiting con-

EMBRYONIC RECONSTITUTIONS 511

ditions — perhaps contact of the basal region with bottom of the container,
resulting in a decrease in the oxygen supply — a small entodermal region
may be determined. Primary and secondary gradients are more nearly
balanced in isolated vegi than in isolated whole ectoderm; but in terms
of primary and secondary activity gradients the balance results from the
changes following isolation rather than from pre-existing concentration
gradients. It was noted earlier that Horstadius' interpretations often seem
to be in terms of metabolism, rather than of concentrations, and to imply
something very similar to dominance.

Apical halves with the ring veg2 added gastrulate before forming mesen-
chyme but develop into plutei with relatively small entoderm, although
only two-thirds of the presumptive ectoderm and the whole presumptive
entoderm are present. Horstadius accounts for the small size of entoderm
by the fact that the presumptive entoderm also forms mesenchyme. Mes-
enchyme formation, however, occurs late, "at the end of invagination,"
when the size of the entoderm must be more or less determined; and it
seems improbable that the small number of cells forming mesenchyme
would decrease entodermal size very greatly. It is suggested as another
interpretation that in this case vego, being brought into direct relation with
more apical levels of the primary gradient, is partly ectodermized.

Whole ectoderms (anj, a«2, vcgi) with four micromeres added basally
differ markedly in development from whole ectoderms alone. They give
rise to plutei, some scarcely distinguishable from controls, but most with
entoderm somewhat too small. In this combination presumptive ento-
derm is absent, but entoderm of normal or almost normal proportions is
reconstituted. Apparently the implanted micromeres bring about this en-
todermization of presumptive ectoderm, as they do when implanted else-
where in ectodermal regions. Horstadius holds that they increase the veg-
etal gradient, presumably by producing vegetal substance; von Ubisch,
that they attract vegetal substance. There is also the possibility that,
whether they are or are not specifically different from other cells in early
stages, they initiate basal activation and the secondary gradient, which
is apparently a factor in gastrulation and differentiation of entoderm.
They may even be more effective in this way in consequence of isolation
from their original relations to other cells.

Apical halves with vcg, and micromeres added include all the presump-
tive entoderm but only two-thirds of the presumptive ectoderm. Thev
give rise to plutei with somewhat too large entoderms and thus differ
from apical halves plus veg2, in which the entoderm is small. Here, again,

512 PATTERNS AND PROBLEMS OF DEVELOPMENT

the micromeres apparently bring about increase in size of entoderm,
whether by increased activation of the basal region or otherwise.

Other experiments show entodermization of presumptive ectoderm in
somewhat different ways. A small group of cells derived from arh, stained
for identification and implanted in the basal region, may become part
of the entoderm. Combination of the eight apical cells of the sixteen-cell
stage (stained) by their basal surfaces with the median surface of a merid-
ional half results in larvae in which stained cells of presumptive ectoderm
form almost half of the entoderm (Horstadius, 1928a, b; 1935). Here
either vegetal substance presumably diffuses from the basal region of the
meridional half into presumptive ectoderm of the apical half and ento-
dermizes it or the activation in the basal region of the one induces the
secondary gradient in presumptive ectoderm of the other, either dynami-
cally or by diffusion.

The preceding experiments have been concerned with apicobasal pat-
tern only. Certain others, bearing on the problem of ventrodorsality, are
also of interest. According to earlier authors, 1/2 and 1/4 blastomeres
reconstitute whole larvae; but certain defects, particularly in skeletons
of larvae from isolated 1/2 blastomeres, were described by Plough (1927,
1929). Halves and quarters of the earlier cleavage stages of Paracentrotus,
isolated in planes of first and second cleavages, may give normal or prac-
tically normal plutei — in fact, four almost completely normal plutei may
develop from a single egg (Horstadius und Wolsky, 1936). Orientation by
staining of the surfaces of separation of these halves and quarters leads
the authors to conclude that first and second cleavages may have any
relation to the median plane, that ventral halves and quarters develop
more rapidly than dorsal, and that the original ventrodorsality persists
in ventral but is inverse in dorsal halves and quarters, the original dorsal
side becoming ventral, the original ventral, dorsal. The more rapid de-
velopment of ventral isolates is to be expected since the high region of the
ventrodorsal gradient is ventral (pp. 134-40). Lindahl (1932) and Hor-
stadius have also suggested a higher metabolism ventrally than dorsally.
Inversion of ventrodorsality in the dorsal half or quarter apparently rep-
resents activation of the low region of the ventrodorsal gradient in conse-
quence of isolation; possibly inhibition by staining of the surface of sepa-
ration may also be concerned in bringing about the inversion. Right and
left halves, stained on the surfaces of separation, are usually defective on
the stained side. Isolated meridional eighths of thirty-two-cell stages
without micromeres reconstitute larvae approaching pluteus form more

EMBRYONIC RECONSTITUTIONS 513

or less closely, but usually with defective skeleton and incomplete develop-
ment or absence of oral lobe and anal arms; they may attain practically
complete bilateral symmetry of form.

GENERAL CONSIDERATIONS CONCERNING THE
SEA-URCHIN EXPERIMENTS

Reconstitution of a whole of small size from an isolated part involves
decrease in scale of organization, that is, a decrease in scale of the gradient
system. Development is usually slower in these reconstitutions than in
controls, and, according to Tyler (1933), more energy is required to attain a
particular developmental stage than in the normal whole. That the slower
development and greater energy requirement of the isolated parts is not
a matter of the time required for "regulation" is indicated by the fact
that giants resulting from fusion of two fertilized eggs develop more rap-
idly than embryos from single eggs. The range of decrease in scale of
organization is rather narrowly limited in the sea-urchin embryo; conse-
quently, isolated parts frequently develop into forms differing in pro-
portions of differentiations from higher and lower gradient-levels or par-
tial forms in which scale of organization is larger than the isolated part.

The paralleHsm between disproportionate and apical partial forms in
sea-urchin reconstitution and in postembryonic reconstitutions of hy-
droids and planarians and many annehds is striking, and decrease in scale
of organization by inhibiting conditions has essentially similar effects in
both. In fact, the isolations and transplantations of early sea-urchin de-
velopment seem to show that the primary apicobasal patterns concerned
in them and in the postembryonic reconstitutions of lower invertebrates
are not fundamentally very different. The embryonic reconstitutions of
the sea urchin bring to Hght little that dift'ers essentially from the data
of reconstitution of mature individuals.

It has already been pointed out that two opposed substance gradients
may be associated with a single activity gradient and that the latter may
be the effective factor in development (p. 241). The assumption of two
opposed specifically different metabolisms coexisting in the same region,
each tending to "suppress" the other but each supposed to persist unless
completely suppressed, and appearing as required to account for experi-
mental results, involves difficulties. Even if two opposed concentration
gradients of substances are present, a single metabolic gradient may be
expected as a resultant. Moreover, the evidence from other reconstitu-
tions indicates that physiological activity, rate or intensity of metabolism,

514 PATTERNS AND PROBLEMS OF DEVELOPMENT

is more nearly a primary factor in developmental pattern than specific
substance differences.

Between the data of embryonic sea-urchin reconstitution and those in-
dicating existence of a primary, and later of a secondary, activity gra-
dient there is no conflict. Different substances may be present in apical
and basal regions or be formed as development progresses; and as far as
they are present, they are doubtless factors in differentiation of ectoderm
and entoderm, but entodermization and ectodermization in consequence
of isolation and combination of parts and of differential susceptibility to
chemical and physical agents appear to be more directly related, at least
as regards their initiation, to nonspecific differences in metabolic activity
than to specific substances.

In the echinoids for which data are available the secondary basal re-
gion of activation becomes more effective as a dominant region and appar-
ently as an inductor than the apical region. This accords with the fact
that dye reduction and susceptibihty suggest that it represents a more in-
tense activation than any other part of the embryo at the stages con-
cerned. Also, entodermization of presumptive ectoderm seems to occur
more frequently in Horstadius' experiments than ectodermization of pre-
sumptive entoderm. Probably future experiment will show differences in
different echinoid species in activation of the basal region. No data of
differential dye reduction have been obtained for Arbacia punctulata,
and the capacities of the micromeres as inductors are not known, but the
characteristics of gastrulation suggest that basal activation may be less
intense in that species than in some others. It seems to be less in the
asteroid Patiria, preceding invagination, than in echinoids (p. 134).

RECONSTITUTION IN INSECT DEVELOPMENT

Investigation of reconstitutional potencies in eggs and embryos of most
arthropods is, to some extent, limited by the character of the chorion and
the impossibility of removing the earlier stages from the egg envelope.
However, by means of Hgature at different levels, localized cautery, and
ultra-violet radiation it has been possible to obtain some information
concerning potencies of early developmental stages and the progress of
physiological differentiation. Earlier experiments on chrysomehd beetles
indicated absence of reconstitution at the developmental stages concerned
(Hegner, 1908, 191 1). Results of ligaturing and microcautery of early
stages of fly development also indicated that at the earliest stages avail-
able after eggs were laid reconstitution did not take place (Reith, 1925;

EMBRYONIC RECONSTITUTIONS 515

Pauli, 1927). It appears from recent work that loss of a certain amount
of ocplasm in consequence of puncture very soon after egg-laying does
not prevent normal development in Drosophila melanogaster, but with
similar loss at somewhat later stages abnormal imagines may result.*^

According to the conclusions reached by Seidel from ligaturing, local
cautery, and ultra-violet radiation of developmental stages of the libel-
lulid Platycnemis pennipes, the developmental pattern of this species dif-
fers from any pattern known in other animal groups. ^ The embryonic
primordium of this species becomes visible first as two lateral areas in
which nuclei are more numerous than elsewhere (Fig. 165, .4). These two
areas come together on the ventral surface; the cephalic lobes appear
anteriorly as still more densely nucleated areas and at the posterior end of
the primordium the level at which inversion of the embryo into the yolk
begins is also indicated by denser nucleation (Fig. 165, B). Seidel main-
tains that the region between the level at which the posterior end of the
embryonic primordium normally forms and the posterior end of the egg
is an initiative or formative center {Bildungszenirmn) and that a substance
essential to embryo formation passes anteriorly from it. Complete separa-
tion by ligature or kilhng of this region, during stages after cleavage
nuclei have reached the egg surface and before blastoderm formation,
prevents embryo formation but does not prevent further cell prohfera-
tion and formation of a cellular layer without definite pattern over the
whole surface anterior to the level of injury, a nonembryonic blastoderm
(Fig. 165, C). With complete separation by ligature at more anterior
levels, embryo formation occurs only posterior to the ligature and a uni-
form anaxiate cell layer anteriorly (Fig. 165, D, £); but with incomplete
hgature that does not destroy protoplasmic continuity, more or less de-
velopment may occur both anterior and posterior to the ligature, as in
Figure 165, F, in which a large head region develops anterior to, and an
elongated embryo posterior to, the hgature. As development progresses,
separation of the posterior region has progressively less effect on develop-
ment anterior to it; in late blastoderm stages separation or killing of
posterior parts has no inhibiting effect anteriorly. Incomplete ligature
which does not destroy protoplasmic continuity does not prevent or in-
hibit development anteriorly even in early stages, because, as Seidel main-
tains, the substance from the Bildungszentrum can still pass.

As far as embryonic development is concerned, the primary effect of the

^Howland and G. P. Child, 1935; Rowland and Sonnenblick, 1936.
7 Seidel, 1926, 1928, 1929a, b, 1931, 1934, i935-

Fig. 165, A-I. — Normal embryonic primordium of the libellulid insect Platycnemis pennipes
and results of ligaturing in early stages. A, two lateral areas more densely nucleated than
other areas in early stage of normal development; B, later stage with cephalic lobes and region
of beginning inversion of posterior end of primordium indicated by denser nucleation; C, re-
sult of complete separation of posterior region by ligature in early stages, inhibition of embryo
formation; D, E, complete ligature at more anterior levels, embryo formation only posterior
to ligature; F, incomplete ligature, head formation anterior, and embryo posterior to ligature;
G, incomplete ligature in blastoderm stage anterior to differentiation center, embryo only
posterior to ligature; H, incomplete ligature at blastoderm stage in differentiation center,
embryos both anterior and posterior to ligature; /, ligature posterior to differentiation center,
embryo only anterior to ligature (after Seidel, 19296, 1934).

EMBRYONIC RECONSTITUTIONS 517

Bildungszentrum at the posterior end of the egg is to determine a "differ-
entiation center," corresponding in position to the pro thoracic segment.
The aggregation and more rapid division of nuclei begins there, and de-
velopment proceeds both anteriorly and posteriorly from this region.
Seidel (1934) regards the differentiation center as in some way an organ-
izer, but not by means of a substance, since incomplete ligatures block its
action. With incomplete hgature anterior to this center in early blasto-
derm stages no embryo formation occurs anterior to the ligature (Fig. 165,
G) ; but with ligature in the differentiation center embryo formation may
occur both anterior and posterior to it (Fig. 165, H), and with hgature
posterior to the center an embryo forms only anterior to the ligature
(Fig. 165, /).

With other insects as material other investigators have found that pos-
terior ligature or cautery in very early stages prevents embryo formation
but at later stages may be followed by normal or partial development,
according to species.** As regards the differentiation center, however, there
is not complete agreement. According to Keith, the head region is the
differentiation center in Camponotus, but the evidence concerning the
beetle Sitona is inconclusive. He suggests that the differentiation center
is the region of "greatest physiological activity." The embryo of the honey-
bee, which occupies the whole length of the egg, shows very considerable
reconstitutional potency, even 24 hours after laying. Dwarf complete em-
bryos may develop posterior to a ligature separating the anterior fifth of
the egg, but there is usually no development anterior to the hgature.
With ligatures further posterior there may be partial development both
anterior and posterior. Seidel's hypothesis of a prothoracic differentiation
center is regarded as supported by these experiments (Schnetter, 1934a, h;

1936).

Extensive duplications of any or all parts have been produced in an
orthopteran, Tachycines asynamoriis, by local injuries and by separation,
by glass needles, of parts of the small embryonic primordium situated at
the extreme posterior end of the egg; and presence of a thoracic differentia-
tion center is inferred, chiefly, it appears, from a single case of duplication
(Krause, 1934). In a later paper Krause (1938) finds the high point in
cleavage and sequence of differentiation at the level of the cephalic lobes;
evidence for a differentiation center in this orthopteran appears only in
the space-time order of differentiation posterior to the head.

Early developmental stages of Drosophila subjected to certain ranges
8 Reith, 1931, an ant, Camponotus ligniperda; 1935, a beetle, Sitona lineata.

5i8 PATTERNS AND PROBLEMS OF DEVELOPMENT

of intensity and duration of ultra-violet radiation show no defects in em-
bryonic or larval stages, but teratological modifications and defects ap-
pear in the imago (Geigy, 1931). Radiation of different early stages shows,
according to this author, a sensitization progressing ''in the craniocaudal
or, better, thoraco-abdominal" direction, followed by a densensitization
progressing in the same direction. Modifications of the imaginal head
were not obtained in any case. These experimental results lead the author
to accept the hypothesis of a prothoracic differentiation center.

In the lepidopteran Ephestia posterior cautery does not prevent axiate
development, though it may produce defects of posterior regions. Cell
formation, as nuclei reach the egg surface, progresses anteriorly and pos-
teriorly from the differentiation center (Maschlanka, 1938).

Experiments with a beetle, Bruchus quadrimaculatus , show prevention
of axiate development by posterior constriction or cautery in early stages,
with decreasing effect as development progresses. The differentiation cen-
ter is maxillar-prothoracic (Brauer and Taylor, 1936). In further experi-
ments with the same species on developmental modification in relation to
susceptibihty to KCN, differential inhibitions of development have been
obtained. With certain periods of exposure to cyanide in early stages, the
median ventral maxillar-prothoracic region shows the greatest injury with
decrease anteriorly, posteriorly, and laterally from the anteroventral posi-
tion of this region (Fig. 166, A-C). With inhibition of development, vari-
ous degrees of embryonic dupHcation may result, beginning at the level
of the differentiation center and extending anteriorly and posteriorly from
it, according to degree and extent of the differential inhibition, occa-
sionally resulting in complete duplication (Fig. 166, D, E). In these cases
the less susceptible lateral regions are physiologically isolated by inhibi-
tion of the median region and with recovery develop independently. At
stages when the embryonic plate is elongating, inhibition of the elongation
is the most conspicuous feature of modification by cyanide (Brauer, 1938).

In embryos of the beetle Tenebrio molitor after posterior cautery in
early stages, heads, or even anterior parts of heads, may develop in com-
plete absence of all other parts (Ewert, 1937); the thoracic differentiation
center is not evident.

All workers in this field agree, however, in finding that major reconsti-
tutions occur only in early developmental stages, if at all ; but, according to
Seidel's account, the pattern of early insect development is unique. The
hypothesis of a differentiation center activated or determined by another
center in the posterior region of the egg raises certain questions. First, as

EMBRYONIC RECONSTITUTIONS

519

regards the Bildungscentnim, how does a substance diffusing from this
center activate or determine an axiate pattern with region of greatest de-
velopmental activity near the anterior end of the future embryo, that is,
farthest away from the source of the substance? A complete Hgature in

^ Porsal Surface.

•ApleriOrSujAlCe \ ,— Cauda/ P.'ale
^ L ^

Meoa/

Max. ilfae
ProfJ^orax

?fetal};oratc
Tfesofjporax

^-Veplral Sur/ctce

Fig. 166, ^-£.— Differential modification of development by KCN in the coleopteran
Bruchus quadrimaculatus. A, diagram of a median longitudinal section showing surfaces and
presumptive embryonic regions; B, ventral view of early stage, indicating by density of shad-
ing the differential susceptibility to KCN, decreasing from the median maxillar-prothoracic
region posteriorly and laterally; C, embryo with maxillar-prothoracic defect resulting from
KCN treatment in early stage (unbroken lines), superimposed on early embryo fixed 2 hours
after KCN treatment (dotted lines) with susceptibility, as indicated by injury, decreasing
from inner to outer elliptical lines; D, later embryo, showing partial duplication by KCN
treatment, complete in thoracic region and extending anteriorly into the head and posteriorly
through the thorax; E, complete duplication by cyanide treatment (from Brauer, 1938).

the posterior region is supposed to block the movement of this substance;
but it also breaks the egg cortex, and a wound stimulus must result and
extend anteriorly from what is essentially a level of section. This may ob-
hterate more or less completely the anteroposterior gradient in the em-
bryonic region. With incomplete ligature continuity of the cortex is not

520 PATTERNS AND PROBLEMS OF DEVELOPMENT

interrupted, though in the Hgature experiments in relation to the differ-
entiation center at somewhat later stages the susceptibility of the cortex
may have increased to such an extent that incomplete ligature is sufScient
to produce a stimulus decreasing or obliterating the gradient of more an-
terior regions. The effect of cautery should be essentially similar to that of
section of the cortex by ligature. The insect egg is apparently highly sensi-
tive. Seidel observed contraction of the yolk about cauterized regions, and
Reith (1925) saw a wavelike contraction {Zucken) pass over the egg of
Musca when it was cauterized locally. The assumption that a part of the
insect egg can be pinched off by ligature or killed by burning without ac-
tivation or excitation extending to other parts seems, at present, unwar-
ranted, in view of the data on postembryonic reconstitutions and some
other embryonic reconstitutions. With complete obliteration of the an-
teroposterior gradient by an opposed gradient the embryonic region must
become apolar, as it actually does after complete ligature and cautery in
early stages. It gives rise to a nonembryonic blastoderm without differen-
tiation. In early stages the activation or excitation resulting from poste-
rior hgature or cautery may be sufhcient to obliterate the whole embryonic
gradient, or its effect may be partial and involve only abdominal, or ab-
dominal and thoracic, regions (Brauer and Taylor, 1936). In Tenebrio a
head, or only the anterior part of a head, may develop after posterior
cautery (Ewert, 1937). Later, after the anteroposterior gradient and de-
termination along its course have progressed farther and become more
stable, posterior ligature or cautery or ligature has little or no effect. It
may also be questioned whether inhibition of development anterior to a
ligature just anterior to the differentiation center is due to isolation from
the center or to obhteration of the gradient in the anterior region by the
opposed gradient resulting even from incomplete ligature in the more
susceptible anterior region. In short, the experimental data suggest that
inhibition of development anterior to a region of ligature or cautery may
be due, not to isolation from a center, but to obliteration of an antero-
posterior gradient by an opposed activation or stimulus resulting from the
injury. Absence of development posterior to a ligature in early stages is
doubtless due to isolation from a dominant region. These suggestions are
in line with results of postembryonic reconstitutions in lower inverte-
brates. Section a short distance proximal to a hydranth primordium in
early stages may inhibit it completely, and posterior section within a cer-
tain distance of the anterior end of a piece inhibits planarian head develop-
ment. No Bildungszentrum is involved in these cases. Apparently there is

EMBRYONIC RECONSTITUTIONS 521

not complete agreement as regards presence of a Bildungszentrum in all
insects investigated. According to the gradient concept, it is to be expect-
ed that ligature and cautery will be more effective in some species than in
others in obhterating the polar gradient at more anterior levels. This is
the case with posterior section in different planarian species.

As regards the prothoracic or maxillar-prothoracic differentiation cen-
ter and its significance, or even its presence in certain forms, there seems
also to be difference of opinion. If heads or anterior head regions can de-
velop in its absence, it does not appear to be essential unless it is present in
the cortical cytoplasm before nuclei reach the cortex, as some believe it to
be. The question whether it possesses inductive capacity or is merely a
region in which cell formation is more rapid and susceptibility greater than
elsewhere seems still to be open. Certainly prothoracic regions can be re-
constituted in other regions of the egg than the original presumptive pro-
thoracic region. If there is a prothoracic region of primary dominance, the
head is apparently secondary in origin; that is a unique polar pattern.
However, axiate pattern may be present in the cortex before nuclei reach
it; the head region may be a primary feature of that pattern; and the pro-
thoracic differentiation center a secondary development associated with
nucleation of the cortex. Early insect development remains an extremely
interesting problem.

RECONSTITUTION IN EARLY FISH DEVELOPMENT

Development of isolated blastomeres of the holoblastic lamprey egg in-
dicates presence of a dorsal inductor region more or less similar to that of
amphibians.'' Early embryonic stages of the meroblastic fish eggs show
high capacity for reconstitution. The cleaving blastoderm of Raja, sepa-
rated by section into fourths, may give rise to several embryonic primor-
dia, even though the separated parts gradually unite again to a single disk
(Eismond, 1910). However, a differential appears to be present in the
blastoderm, for the primordia develop from median and lateral regions of
one half, none developing from the other half ; and retarded development
and early degeneration of the lateral primordia after reunion of parts sug-
gests that their further development is inhibited by dominance of more
nearly median primordia, which probably represent higher gradient levels
of the blastoderm.

In teleosts normal embryos may develop from parts of the germinal
disk, from 1/2 blastomeres, from two 1/4 blastomeres, from six 1/8, and

^Bataillon, 1900a, b; Montalenti e Maccagno, 1935.

52 2 PATTERNS AND PROBLEMS OF DEVELOPMENT

from twelve 1/16 blastomeres in various combinations, and even after re-
moval of the whole embryonic area up to a certain stage preceding gastru-
lation.'" Removal by localized puncture or cautery of parts of the embry-
onic shield after its formation indicates, in general, progressive stabiliza-
tion of course of development; but apparently normal, or only quantita-
tively defective, development is possible after removal of small amounts,
even of axial material. Injury of the posterior end of the embryonic region
in earlier stages may completely inhibit embryo formation, perhaps by
obliteration of the polar gradient. In later stages it results in absence of
posterior parts still to be formed at the stage of injury. Recent experi-
ments, using the method of implantation of isolated parts of the blasto-
derm into the yolk sacs of other embryos, show similar development of all
sectors of the blastoderm in pregastrula stages (Mangold, 193 it; W. Luth-
er, 1936a, ^). Sectors including the marginal region, when thus implanted,
are capable of developing neural tube, notochord, muscle segments, in-
testine, auditory vesicles, eye lenses, blood vessels, blood corpuscles, kid-
ney canals, and liver tissue. Reconstitution of the extraembryonic sectors
apparently represents an approach to development of a new embryonic
primordium. Further experiments on removals and transplantations of
parts of the blastoderm of the trout also show high capacity for reconstitu-
tion in earlier stages (W. Luther, 1937a, b). Combination of two extra-
embryonic halves in the blastula stage by replacing the embryonic half by
an extraembryonic half results in reconstitution of complete, or almost
complete, embryos in 90 per cent of the cases. In early gastrula stages an
embryo develops after removal of a lateral half of the embryonic sector,
either by movement into its place of material still farther lateral or when
it is replaced by a piece from the extraembryonic half. Reconstitution fol-
lows removal of either lateral or median part of the embryonic region. In
the early gastrula the region capable of development extends about 45°
right and left of the longitudinal embryonic axis, but reconstitution some-
times occurs after replacement of somewhat more than 90° of the em-
bryonic sector by an extraembryonic piece. With progress of gastrulation
the reconstitutional capacity of extraembryonic regions of the blastoderm
gradually disappears, first in the median region opposite the embryonic
sector and from this region progressively on both sides. In the later gas-
trula capacity for development corresponds approximately to the pros-
pective organ regions.

Luther's interpretation of these data is that capacity for development

"Morgan, 1895a; Kopsch, 1896, 1904; Sumner, 1903; W. H. Lewis, 1912a, b; Hoadley,
1928.

EMBRYONIC RECONSTITUTIONS 523

depends on level of metabolic activity, that up to the gastrula stage all
parts of the blastoderm have a sufficiently high level to reconstitute axial
embryonic organs, but this gradually decreases from the beginning of
gastrulation. There is, however, a gradient of decreasing activity laterally
and around the blastoderm from the region of the longitudinal axis. In
earher stages lateral extraembryonic parts are active enough to replace
median embryonic parts, but this capacity is gradually limited to the em-
bryonic quadrant and finally to the organ regions. In connection with this
interpretation the possibihty may also be noted that extraembryonic re-
gions may undergo activation when isolated, or by induction when im-
planted in contact with higher gradient-levels. The data on differential
susceptibility are in line with Luther's interpretation in that they show
that the embryonic quadrant attains a higher susceptibility than other
regions in early stages.

Certain experimental data suggest presence of a differential and some
degree of dominance, even before formation of the embryonic shield. Un-
der inhibiting conditions more than one embryo may develop from the
teleost blastoderm." Evidently reconstitution of an embryonic region,
inductor, and axis quite independently of the original is possible in tele-
osts, and the fact that normally only one embryo forms suggests that the
presumptive embryonic region or sector attains some degree of dominance
over other parts. However, as far as developmental capacities are con-
cerned, the regional differences in early stages are apparently nonspecific
quantitative differences in activity in a gradient pattern.

Pieces of Fundulus germ ring 90° and 180° from the dorsal lip of gas-
trulae at any stage, when implanted in extraembryonic regions or in the
pericardium, form only epidermis, blood, and chromatophores but, when
implanted in the embryonic shield, may develop almost any embryonic
part, according to position, or sometimes parts not in accord either with
position or with presumptive fate. Some influence, a dominance of some
sort by the embryonic shield, is necessary for differentiation of these pieces
of germ ring into definite embryonic organs. No evidence of gradation of
developmental potencies around the germ ring has been found in this ma-
terial (Oppenheimer, 1938).

CERTAIN RECONSTITUTIONS OF EARLY AMPHIBIAN DEVELOPMENT

When one blastomere of the two-cell stage is killed or injured so that it
cannot develop, but is left in position, the other may develop as a half-
embryo (Roux, 1888; A. Brachet, 1927); but, according to Roux, the half-

" Stockard, 1921; Hinrichs, 1925; see also F. Schmitt, 1901, 1902.

524 PATTERNS AND PROBLEMS OF DEVELOPMENT

embryo may later become more or less a whole, either by reconstitution or,
in case the other blastomere was not completely killed, by spread of nuclei
into it and by cell formation from its cytoplasm (postgeneration). This
apparent reconstitution of a half to a whole was the subject of much con-
troversy. Combining killing of one blastomere with maintenance in in-
verted position, it was found that the living blastomere might develop as a
whole, supposedly in consequence of a more complete isolation of the liv-
ing from the killed blastomere through the rearrangement of materials by
gravity (Morgan, 1895^). The blastomeres are too fragile to permit direct
separation, but they may be separated by gradual ligature, or one may be
removed by suction. With these procedures the 1/2 blastomere may de-
velop as a whole, or with incomplete separation by ligature various de-
grees of twinning may result.'^

Partial separation of undivided eggs by ligature shortly after fertiliza-
tion leaves one of the two parts without a nucleus. The nucleated portion
develops, the nonnucleated does not; but if the ligature does not complete-
ly separate the two parts, a cleavage nucleus may sooner or later pass
through the constriction into the nonnucleated part, and this may then
begin to develop. Such cases may result in partial or complete twinning,
the originally nucleated portion being in advance of the other. Moreover,
it was found that complete separation of the first two blastomeres by liga-
ture or otherwise gave different results according to relation of plane of
first cleavage to the median plane. When the first cleavage plane coincides
with the median plane, both right and left half are able to develop as
wholes; but when dorsal and ventral halves are separated, the dorsal de-
velops as a whole, but the ventral forms only a rounded mass, showing
some development of germ layers but without axial organs. The dorsal in-
ductor region is responsible for this difTerence in development. With me-
dian first cleavage both blastomeres contain half of this region; and, since
the isolated half is able to induce whole development, each may develop
as whole. When the first cleavage is frontal, the inductor is wholly in the
dorsal blastomere. Apparently the same differences result from ligature
of the undivided egg in different planes.'-^ These differences in develop-
ment are regarded as indicating early locahzation of the dorsal inductor

"Endres, 1895; Herlitzka, 1896; Spemann, igoib, 1902, 19036; McClendon, 1910; Rand,
1925; G. A. Schmidt, 1930, 1933.

13 Spemann, igoib, 19036, 1914, 1928, 1936, pp. 16-18; Fankhauser, 1930a, b; also Streett,
1940, "Experiments on the organization of the unsegmented egg of Tritiiriis Pylirrogaster,"
Jour. Exp. Zool., 85.

EMBRYONIC RECONSTITUTIONS 525

region. Separation of halves by ligature in later cleavage, blastula, or even
early gastrula stages shows much the same differences, according to plane
of separation. After median separation the inner sides of the two indi-
viduals are, in general, increasingly less developed than the outer sides, the
later the stage of separation. The left member of the pair has normal situs
viscerum, but in the right member situs inversus is not rare.'^ Contrary
to the earlier conclusions of A. Brachet from experiments with local
cautery, ligature and removal of parts showed that early anuran embryos
have essentially the same capacity for reconstitution as the urodele (Vogt
und Bruns, 1930; G. A. Schmidt, 1933).

These reconstitutions in the early stages of amphibian development in-
dicate that the dorsal inductor region is in some way localized but that,
within it, relations of dominance and subordination and capacity for re-
constitution are present, and the symmetry pattern, if definitely present in
these stages, may undergo extensive reconstitution. However, the fact that
new dorsal inductors may originate in any part in inverted and partly in-
verted eggs (pp. 428-30) indicates that determination of the inductor re-
gion in early stages has probably not proceeded very far in specific differ-
entiation.

RECONSTITUTIONS IN THE AMPHIBIAN GASTRULA AND
LATER STAGES

The discussion of amphibian inductors in the preceding chapter was
largely concerned with reconstitution in consequence of altered relations
to an inductor. It was shown there that an inductor may determine the
course of development of other parts but that its own development may
be largely or wholly independent of them and may depend on relations of
dominance and subordination within the inductor region itself. When a
region, field, or organ system has attained a certain stage of determina-
tion, chemodifferentiation, or physiological stability, reconstitution of a
different region, field, or organ system from it does not occur under the
usual experimental conditions, though we do not know that it might not
occur under other conditions. In amphibian development progressive limi-
tation or restriction of developmental potencies under known experimen-
tal conditions becomes evident in gastrula and later stages, but extensive
reconstitution is still possible within various organ primordia and fields
when parts are removed. Some organ systems, like legs and tails of uro-
deles, are capable of regeneration even in the adult. Reconstitution in an

'•» Spemann und Falkenberg, 1919; Ruud und Spemann, 1922.

526 PATTERNS AND PROBLEMS OF DEVELOPMENT

organ field is not necessarily limited to the region in which the organ ac-
tually develops but may take place in other parts of the field. Examples
of this in the limb field, the optic field, and some others and their apparent
relations to gradient patterns were considered in chapter xi.

An extensive investigation of the capacity for development of isolated
parts of urodele and anuran gastrulae explanted in modified Ringer solu-
tion has shown that entoderm and parts of the mesoderm are capable of
advanced differentiation, even in small explants, but largely according to
prospective significance; that is, their development may involve more or
less reconstitution but is not greatly altered by isolation and explantation.
Parts of the dorsal inductor region similarly explanted show more exten-
sive reconstitution of parts beyond their prospective significance. Lateral
parts become bilateral; neural tubes and even epidermis may develop from
these pieces. Explanted ectodermal pieces show no capacity for any defi-
nite pattern of development. They form neural tissue only in relation to
an inductor. Also, gastrula ectoderm cultured for a time as explant and
then implanted in relation to an inductor shows decreasing reactivity, the
more advanced the stage from which it originates.'^ As regards the general
pattern of amphibian development, these experiments do not essentially
alter earlier conclusions. They confirm Holtfreter in his belief in a mosaic
of inductors in the dorsal inductor region, but this seems still to be a mat-
ter of opinion rather than a necessary conclusion from these or other exper-
iments. The data show great differences in capacity for independent de-
velopment ; but the question how far these differences depend on degree of
determination of parts, as Holtfreter believes, rather than on susceptibil-
ity to isolation and explantation is perhaps still open. It seems possible
that entodermal regions, for example, may be no more determined than
other parts in the sense of having attained a higher degree of specificity
before isolation but that, because of their low metabolism at the time of
isolation, they are not very susceptible to the altered conditions and con-
tinue development more or less according to the pattern of which they
were originally a part. Parts of the dorsal inductor region are much more
susceptible to isolation and show extensive reconstitution, even giving
rise to ectodermal parts. Explanted ectodermal pieces apparently never
attain a sufficiently high metabolic level to permit neural development.

A very considerable capacity for reconstitution appears in various or-
gan primordia of later embryonic stages; for example, a complete limb
may develop from a part of a limb bud or prospective limb region. In this

'5 Holtfreter, 1938^, b, c, and literature cited in these papers.

EMBRYONIC RECONSTITUTIONS

527

connection the eye is of particular interest. In the early gastrula of Triton
the whole presumptive epidermis can give rise to eyes, under influence of
the dorsal inductor; but eye potency decreases from apical to basal regions,
that is, down the primary gradient.'^' After the neurula stage only the
optic primordium gives rise to eye, but the capacity of the optic vesicle
and optic cup for reconstitution of a whole eye from parts homoplastically
transplanted to other embryonic regions supposedly without eye-deter-
mining factors — for example, the ear region — is great. Any part can recon-
stitute any part: even small pieces from any part of the tapetum can be-
come whole eyes. In general, capacity for organization increases with size
of the transplanted piece. The larger the piece, the more evident is the
heteropolar pattern. Extremely small fragments usually give rise only to
the pigment layer, but in larger pieces the retina may be massive.'^ The
orderly character of development in the transplanted pieces and the rela-
tion between organization and size of piece have led Dragomirow to re-
gard the optic vesicle as a gradient system in which the retinal region is
dominant. The pigment layer represents development from lower gradi-
ent-levels. He finds these embryonic optic reconstitutions closely analo-
gous to reconstitutions in hydroids and planarians. Pieces of iris of urodele
larvae of various ages implanted in the orbit after removal of eye or in the
abdominal cavity may reconstitute retina; and in the orbit, nerve libers
(Monroy, 1939).

In marked contrast to the high reconstitutional capacity of the optic
vesicle and various other parts in later embryonic stages is the failure of
the early tail bud to reconstitute parts removed,'^ although a high capac-
ity for regeneration is characteristic of later stages of tail development.
Two possible factors of this failure may be pointed out. For reconstitu-
tion presence of spatial pattern and reaction to removal of a part are neces-
sary : a mass of cells all alike cannot reconstitute anything if part of the
mass is removed; also, if level of activity is so high that removal of a part
does not bring about activation, there is no reconstitution, and the re-
maining part continues as before. The early tail bud is evidently a mass of
actively growing cells, perhaps with a slight radial gradient, like other
buds, but probably with so little differential and dominance and so high a
level of activity that removal of a part leaves the remaining part practi-

"■ See Mangold, 1931a, for literature.

'7 Dragomirow, 1932, 1933, 1935, and earlier literature cited in these papers. His experi-
ments include three species of Triton, Amblystoma mexkaniim, Pelobates fnsciis, and Bom-
binator igneus.

'* Schaxel, 1922; V'ogt, 1931; Svetlov, 1934.

528 PATTERNS AND PROBLEMS OF DEVELOPMENT

cally unaltered. In later stages, when gradient pattern has developed, fur-
ther removal of a part is followed by activation and increased growth
near the cut surface, and regeneration results.

Some parts of the amphibian and other embryos, isolated after a cer-
tain stage under proper conditions, are capable of independent or self-
differentiation, that is, they are able to continue for a time their original
course of development. It is commonly assumed that such parts also dif-
ferentiate independently when parts of the intact organism and that, in so
far as parts attain this condition, the organism becomes a ''mosaic" of in-
dependent parts. This conclusion does not seem entirely justified. Even
though differentiation is independent after isolation, it may not have been
before. In other words, self-differentiation of an isolated part may repre-
sent more than it accomplishes independently without isolation. The fact
that it can differentiate independently is not proof that it does so in the in-
tact organism.

RECONSTITUTION AND PATTERN IN THE AVIAN EMBRYO

Morphological and material aspects of developmental pattern in the
chick have been subjects of investigation by various methods and of con-
siderable controversy for many years. Local injury — mechanical, elec-
trolytic, or by radiation — has been extensively used in attempts to throw
light on the problem, and modifications resulting from subjection of em-
bryonic stages to various chemical and physical agents have been de-
scribed and analyzed. In recent years data on differential susceptibiHty
and differential dye reduction, transplantations of parts to other embryon-
ic regions or to the chorio-allantois, as a supposedly neutral site outside the
embryo proper, and explantations to plasma clots with embryonic extract
have given information concerning some of the physiological features of
pattern and concerning regional and chronological differences in develop-
mental capacity or potency under certain experimental conditions and
their relations to the general pattern. Some of the evidence from these ex-
periments is briefly reviewed.

Cell migrations are very slight during the first lo hours of incubation,
and extensive reconstitution is possible at these stages. Lesions produced
electrolytically or by ultra-violet radiation in the median posterior region
of the blastoderm result in various anomalies, among which are more or
less complete embryonic duplications and supplementary heads (Twiessel-
mann, 1938). The duplications are regarded as determined by the effect
of experimental procedure on the inductor — inhibition or killing of the

EMBRYONIC RECONSTITUTIONS 529

median parts — with reconstitution of more lateral regions. This author
also holds that his data support Dalcq's view that the inductor differ-
ences along the polar axis are quantitative, not quahtative.

Some of the earlier experiments on chorio-allantoic grafting were more
directly concerned with questions of the degree of organization or deter-
mination attained at particular stages than with spatial pattern and led
to the conclusion that progressive organization is arrested in a part by
isolation; consequently, the differentiation of the isolated part is a measure
of the degree of organization present at the time of isolation (Hoadley,
1926(3, 6; 1927). If this is true, it follows that early isolation of parts
should give less differentiation than isolation at later stages and that re-
constitution should not take place, but other experiments show extensive
reconstitution.

Transplantations of parts of the unincubated blastoderm show exten-
sive differentiation only when they include the intact posterior median
quadrant, the region from which the primitive streak later develops. This
quadrant alone, although much smaller than the anterior three-fourths,
gives rise to almost all the tissues of the anterior embryonic region and
shows as extensive differentiation as posterior halves or whole blastoderms
or the node-level of whole blastoderms of definitive primitive-streak stage
(Butler, 1935). This posterior quadrant is the most susceptible region of
the early blastoderm (p. 162). According to Butler, pieces lacking this
posterior region do not develop central nervous tissue and usually give
rise only to gut, smooth muscle, heart, liver, and skin tissues. Pieces in-
cluding part of this posterior region show higher frequency of graft devel-
opment than anterior pieces, and transverse fourths from different levels
show a posteroanterior gradient of decreasing frequency of development.
Fourths from the posterior half show much less development than the
whole half or the median quadrant; and longitudinal fourths show little
differentiation, giving only gut, smooth muscle, and heart, probably, as
Butler suggests, because of interference with the cell movements con-
cerned in development of the primitive streak.

In very early streak stages anterior, purely ectodermal regions, ex-
planted to plasma clots, may form neural tubes; pieces anterolateral to the
streak form chiefly heart in stages before the groove appears, after that,
heart and neural plate ; and the region of the streak gives rise in pregroove
stages only to erythroblasts, in groove stages to heart and erythroblasts
(Rudnick, 1938&). These results differ widely from those obtained by But-
ler with chorio-allantoic grafting of pieces from unincubated blastoderms.

530 PATTERNS AND PROBLEMS OF DEVELOPMENT

The earlier results of chorio-allantoic grafting with definitive primitive-
streak, head-process, and somite stages seemed to indicate that only pieces
from the node region were capable of extensive development and that this
region could give rise to all parts posterior to its level as far as, and includ-
ing, mesonephros; but, as the physiological condition indicated by the
node moved posteriorly, its development was progressively restricted to
more posterior parts; in general, it did not give rise to parts characteristic
of levels anterior to its position (T. E. Hunt, 193 1, 1932). Further experi-
ments, however, modified this view. Posterior parts of advanced primi-
tive-streak stages, containing no more than the posterior two-thirds of the
streak, cultivated in vitro, are able to continue differentiation (Wadding-
ton, 1935a). Dalton (1935) also found presence of the node region unessen-
tial for development of axial structures and regarded the technique of
grafting as an important factor in results obtained. Forebrain develops in
chorio-allantoic grafts from the head-process stage, even from pieces en-
tirely anterior to the head-process and without notochord (Stein, 1933).
Heteroplastic grafts between duck and chick of parts of the primitive
streak without the node may show more or less reconstitution and may
also act as inductors (Waddington and Schmidt, 1933).

Reconstitution of any part and even of the whole primitive streak after
removal occurs in blastoderms explanted to plasma-chick embryo extract
after some 20 hours of incubation with development of normal, or almost
normal, embryos (Waddington, 1932). The cell movements are regarded
by Waddington as the chief factors in this reconstitution, but both the cell
movements and the development of normal embryos in these experiments
indicate presence and effectiveness of a physiological pattern in relation to
which fates of parts concerned in reconstitution are altered.

Development of ectoderm, mesoderm, and entoderm in small pieces
from the different regions of the head-process blastoderm grafted on the
chorio-allantois is shown in the maps of Figures 167, 168, and 169 (Rawles,
1936). Evidently considerable reconstitution takes place in many of the
pieces, for various organ tissues develop in parts which, so far as known,
do not normally give rise to them. Many of the potency fields are more
extensive than the regions of actual differentiation in normal develop-
ment, and Rawles finds that in each field developmental potencies for the
part concerned decrease peripherally from a center. On the other hand,
each such field is a more or less restricted region or level of the blastoderm.
Some organs (heart, liver) develop only in lateral areas; others (notochord,
suprarenal, spleen) only in median areas; and still others in both median

EMBRYONIC RECONSTITUTIONS

531

and lateral areas but in higher percentage in median (brain parts, eye,
ear, mesonephros) . The highest frequency of graft development is in
median pieces including head process and node, but pieces from the left
side show a higher frequency than those from the right. Development of

P'iG. 167. — Map showing regional distribution of development of ectodermal structures in
chorio-allantoic grafts of the head-process blastoderm of chick; distances from the primitive
pit indicated in millimeters are average measurements (from Rawles, 1936; prepared by
Willier and Rawles).

the pieces indicates an anteroposterior and mediolateral physiological pat-
tern of some sort, in relation to which different developmental potencies
are realized.

A further study of eye development in chorio-allantoic grafts of late
primitive-streak and head-process stages has shown that the potentiaHty
for eye development is realized in the grafts from an area of field about the

532

PATTERNS AND PROBLEMS OF DEVELOPMENT

primitive pit in late primitive-streak stages and later at the anterior end
of the head process. Frequency of eye development is higher in grafts from
the median than in those from the lateral regions of this field, and in those
from the left side than in those from the right, as in Rawles's experiments.

Fig. i68. — Regional distribution of development of mesodermal structures in chorio-allan-
toic grafts of head-process blastoderm of chick; measurements as in Fig. 167; s, striated muscle;
H, nonstriated muscle (from Rawles, 1936; prepared by M'illier and Rawles).

In later stages capacity for eye development becomes greater laterally
than in the median region. The higher frequency in median than in lateral
grafts of organs normally lateral, such as eye and ear, is like eye frequency
in amphibian development at certain stages and, like that, suggests prece-
dence of the median region in attainment of a certain physiological condi-

EMBRYONIC RECONSTITUTIONS

533

tion (pp. 282-85). The map of Figure 170 shows in a different way regions

from which tissues of certain organs differentiate in chorio-allantoic grafts.

The results of grafting and explantation show a very considerable

capacity for reconstitution in many of the isolated parts and a general but

thyroid
liver

respiratory tract

aaterior gut

5tomacH

epltf^ellal tubules

liver

n;5pinitory tract

anterior gut

stomach

epittielial tutniles

thyroid

resp tract

anterior gut

stcmacK

epith tubules

epithelial tubules

thyroid

Uver

respiratory tract

aaterior gut

stomach

epithelial tubules

Uver

respiratory tract

anterior gut

stomach

epithelial tubules

Fig. 169.— Regional distribution of development of entodermal structures in chorio-al-
lantoic grafts of head-process blastoderm of chick; circles at margin indicate pieces from which
"primordial germ cells" were obtained; measurements as in Fig. 167 (from Rawles, 1936, pre-
pared by Willier and Rawles).

definite relation of character of reconstitution to a spatial pattern of some
sort. It seems evident that neural tissue can develop without induction
and that in earher stages the primitive streak may reconstitute from re-
gions lateral to those in which it normally originates; this is perhaps large-
ly a matter of continuation of the cell movements concerned in formation

534

PATTERNS AND PROBLEMS OF DEVELOPMENT

of the streak. It appears probable, however, that m many of the experi-
ments full realization of developmental capacities may be limited or in-
hibited by the experimental conditions. The chorio-allantoic grafts of the
smaller blastoderm pieces do not give rise to orderly definite embryos of

Fig. 170. — Map showing by shading certain organ areas of chick at head-process stage,
as defined by chorio-allantoic grafts. In the median region at the anterior end of the head proc-
ess is the eye area; heart develops from the large lateral areas; between them and median is
the mesonephric area and central within it a cross-lined area representing the adrenal-gonad
area; depth of shading in the various areas represents roughly intensity or degree of develop-
mental potency for the organ concerned under the conditions of experiment (prepared by Wil-
lier and Rawles, kindness of Dr. Willier).

small size or even to axiate partial forms, but to tissue complexes; and
many of them fail to develop at all, though they presumably possess the
same potencies as others that do develop. Such failures may, of course, be
due to incidental conditions, injury in isolation and grafting, inadequate
attachment, inadequate blood supply, etc. But other factors are appar-
ently concerned. Even when whole blastoderms of primitive-streak, head-

EMBRYONIC RECONSTITUTIONS 535

process, and early somite stages are grafted, parts posterior to the mes-
onephros do not develop in any case (Willier and Rawles, 193 1). Also,
grafted pieces from the more posterior levels of advanced primitive-streak
stages show, at best, only slight differentiation. The absence of posterior
parts in grafted whole blastoderms suggests that development is perhaps
unable to proceed beyond a certain stage on the chorio-allantois. Size of
piece may also play a part in determining occurrence and character of de-
velopment of grafted pieces, as is maintained by Murray and Selby (1930).
In postembryonic grafting in the lower invertebrates a larger piece very
generally persists and develops more frequently than a smaller. The small-
er fraction of gradient pattern in the smaller piece is less effective than a
larger fraction; but even small pieces from a high gradient-level, grafted
into a low level, may persist and develop. Similar questions arise with re-
gard to explantation. The explanted whole blastoderm may develop quite
normally up to a certain stage, at which development stops and death fol-
lows (WadcHngton, 1932). Interference with the cell migrations may also
be an important factor in hmiting realization of developmental capacities.
Explanted transverse strips of definitive primitive streak and head-process
blastoderms develop axial parts from the node-level and anterior regions,
but development of levels of the streak is interfered with by the trans-
verse sectioning (Rudnick, 1938a). The marked retardation of develop-
ment in explants of pieces of early blastoderms and the disappearance of
the streak structure from streak stages (Rudnick, 19386) suggests that de-
velopmental potencies of these pieces are far from realized in these experi-
ments. This may be due not to lack of organization but to the high sus-
ceptibihty of these stages to the conditions of explantation. According to
Butler (1935), the posterior quadrant of the unincubated blastoderm,
from which the primitive streak later develops, gives rise to axial organs
in chorio-allantoic graft; but in somewhat later stages the region of the
early streak in explants develops only erythroblasts (Rudnick, 19386). In
spite of the many positive results of experiment and the advance in knowl-
edge of avian development due to them, it still seems possible that devel-
opmental capacities of at least some parts of the blastoderm may be great-
er than experiment has shown. Some of the experimental data suggest
that the gradient pattern of later streak and more advanced stages de-
velops gradually in the blastoderm and that the apparent lack of organiza-
tion in early stages may be lack or inadequacy of this pattern, but they
throw no light on the problem of its origin.

536 PATTERNS AND PROBLEMS OF DEVELOPMENT

EMBRYONIC DUPLICATIONS AND POLYEMBRYONY

Differential inhibition of embryonic stages not infrequently results in
duplication or multiplication of parts or axes and sometimes in develop-
ment of more than one individual from a single egg or early embryo. Ex-
amples are the duplications in insect development resulting from differen-
tial inhibition by cyanide (p. 518), polyembryony in fishes after inhibition
by low temperature (Stockard, 192 1) and by exposure to ultra-violet light
(Hinrichs and Genther, 193 1), and dupHcations in chick embryos resulting
from various inhibiting conditions. Partial duplications in annelids and
other forms may also result from inhibiting conditions. In these cases
there is apparently a decrease in dominance with physiological isolation of
regions normally subordinate. Temporary inhibition sufficient to prevent
rapid recovery of the original dominance followed by return to natural
conditions, permitting parts originally subordinate to become dominant, is
apparently most effective in producing these duplications.

Duplications and multipHcations, ranging from bifurcations, through
all degrees of teratological duplications and multiplications and complete
twins from single eggs, to development of many, even hundreds, perhaps
thousands of embryos from a single egg, occur without experimental inter-
ference. Except in a few forms, complete and partial twinning appear only
occasionally, as do the teratological forms, and many of them probably re-
sult from inhibiting conditions. The extreme types of polyembryony are
usually normal characteristics of the species concerned. The natural
polyembryonies raise interesting questions concerning developmental pat-
tern, but at present it is possible in most cases only to call attention to
some of them and to suggest possibihties.

Polyembryony has been observed in various coelenterates as a conse-
quence of separation of blastomeres and blastomere groups, a "blastomere
anarchy," as Metschnikoff describes it.'^ These cases may be results of
slightly inhibiting laboratory conditions or other unfavorable conditions.
Low oxygen in standing water may be sufficiently inhibitory to obliterate
any pattern originally present in early embryonic stages, with resulting
isolation of cells or cell groups, and the differential between free surface
and surface in contact may determine new polarities in the isolates, or they
may reconstitute from the part of the original pattern persisting in them.

In several genera of bryozoa the blastomeres of earlier stages are ap-
parently completely separated, and follicle cells may lie between them.

"See, e.g., Busch, 1851; Haeckel, 1881; Metschnikoff, 1886.

EMBRYONIC RECONSTITUTIONS 537

Later they form a compact rounded mass of cells with more or less definite
outer layer but without any indication of definite embryonic structure or
polarity. Still later, outgrowths from this primary embryo develop, some-
times becoming elongated, finger-like extensions; from the tips of these
outgrowths cell masses, varying somewhat in size, separate as secondary
embryos. These may develop or in certain species give rise to further
outgrowths from which tertiary embryos develop.^" There is no indication
of polarity in the early stages of these masses. According to Harmer, how-
ever, the outgrowths develop only toward the distal end of the ovicell in
certain species of Crista. This perhaps indicates presence of a polar pat-
tern determined by some differential in the ovicell. In other species of
Crista Robertson finds outgrowths developing in any direction. Each out-
growth probably represents a localized region of increased cellular activ-
ity like a bud, the locahzing factor being apparently more or less fortui-
tous, and each mass separating from the outgrowth may develop polar pat-
tern from the part of the longitudinal pattern of the outgrowth persisting
in it; but, even if this is the case, the origin of ventrodorsality remains ob-
scure. That axiate physiological pattern is acquired at some time in the
course of this polyembryonic development and in relation to some differ-
ential or differentials in conditions and that a definite egg pattern is not
necessary for development of an axiate individual in these forms is highly
probable.

Development of two whole embryos from a single egg and various de-
grees of embryonic twinning, involving anterior or posterior duplications
or both, are not infrequent in ohgochete annelids.^' These duplications,
like some experimental duplications discussed in the following chapter
(pp. 556-61), are of special interest as indicating that annelid development
is not as completely a ''mosaic" of self-differentiating parts, as often
assumed.

In certain parasitic hymenoptera there are extreme degrees of polyem-
bryony, although maturation and early cleavage suggest presence of axiate
pattern of some sort and degree. If such pattern is present, there is no evi-
dence that it persists or is effective in later stages. By continued cleavage
the primary embryo becomes a solid mass of cells, apparently without
definite arrangement or visible pattern. Constrictions divide this embryo
into a number of cell masses, constituting the polygerm; and these pri-
mary masses divide to form secondary masses; these, in turn, to form

^"Harmer, 1893, 1896, 1898; A. Robertson, 1903.

" Kleinenberg, 1879; Vejdovsky, 1883-92; Weber, 1917; Welch, 1921; Penners, 1924a.

538 PATTERNS AND PROBLEMS OF DEVELOPMENT

tertiary; etc/' The cell masses may differ considerably in size, and, accord-
ing to Patterson (192 1), an embryo may develop from a single cell. On
the other hand, many masses fail to develop in some species, and some give
rise to asexual, nonviable individuals. There seems to be no common ori-
entation of the axes of embryos in the polygerm, nor do all embryos de-
velop at the same time or from cell masses of the same generation. There
is no evidence of persistence of axiate pattern through all the divisions of
the polygerm, and the apparently fortuitous character of the divisions sug-
gests that, if pattern was originally present, it has been obliterated. Can
axiate pattern originate autonomously in the final generations of cell
masses? If a gradient pattern, a molecular pattern, or a spatial pattern of
any kind determining axiate organization of the individual insect does
arise de novo, it is difficult to conceive how this is possible, genetically or
otherwise, except in relation to some initiating factor external to the mass
concerned. Differentials in oxygen supply or in CO2 accumulation and
perhaps potential differences in the parent body and between the masses
may be factors in initiating pattern. Some of the figures given in papers
cited above suggest that the final polarities of embryos may be determined
by a differential in their local environment — for example, between the
wall and the interior of the polygerm — but the question is not discussed by
the authors. Failure of some masses to develop may be due to absence or
inadequacy of axiate pattern rather than to nutritive conditions, as sug-
gested. Development of masses of any generation and failure of many to
develop suggests that growth and division of the masses continues until a
mass acquires a pattern adequate for development. In vitro cultivation of
the polygerms or cell masses, if found to be possible, may give some infor-
mation concerning their developmental physiology.

In most vertebrates polyembryony is only occasional under natural
conditions and is limited to various degrees of twinning, ranging from all
degrees of teratological duplication to conjoined twins, equal or unequal
in development, and completely separate duplicate or identical twins.
The descriptive Hterature of vertebrate teratology is voluminous, and the
question whether duplications and multiple forms result from fusion of
originally separate embryonic primordia or from division of a single pri-
mordium has been discussed for many years. However, continued study
of the teratological forms and of complete twins and the experimental
embryonic reconstitutions have estabhshed beyond question the origin

"Marchal, 1904; Silvestri, 1905, 1906, 1908; Patterson, 1915, 1921; Gatenby, 1918;
Leiby, 1922; and literature cited by these authors.

EMBRYONIC RECONSTITUTIONS 539

of many of these forms from a single egg or an originally single embryo.^''
Evidently such duplications are results of an agamic reproduction — a bud-
ding or fission with reconstitution of pattern — occurring in early embry-
onic stages; but conditions concerned in their origins can usually only be
inferred or guessed at.

Experiments, both on embryonic and postembryonic stages of plants
and animals, show that decrease of dominance by exposure to inhibiting
conditions may result in a greater or less degree of physiological isolation
of previously subordinate parts, in establishment of new dominant re-
gions, and, in some forms, with proper experimental procedure, in com-
plete obliteration of the original polarity and dominance. That somewhat
similar factors are concerned in many cases of nonexperimental embryonic
duplication, even in higher vertebrates, seems probable, as suggested by
Newman (1917^^, 1923). Toxic or other conditions inhibiting developmen-
tal activity may weaken dominance and polarity to such an extent that
new polarities and dominant regions may arise in less susceptible regions
or after the inhibiting conditions cease to act, either in relation to what
remains of the original pattern or in reaction to local differentials. The
high frequency of duplications with definite relation to the original sym-
metry pattern suggests that this often plays a part in determining locali-
zation of the new dominant regions. Partial axial duplication may, of
course, also result from mechanical division of a dominant region — for
example, split tails, limb buds, etc. Accidental or pathological conditions
may determine some duplications in this way.

Even among the mammals, however, polyembryony is not Hmited to
teratological forms and occasional identical twins. In two species of arma-
dillo development is normally poly embryonic. With few exceptions four
embryos develop from a single egg in the nine-banded armadillo, Dasypus
{=Tatusia) novemcinctus,^^ and from six to twelve, usually eight or nine,
in D. hyhridus (Fernandez, 1909). The quadruplets of D. novemcinctus
arise by two successive agamic reconstitutions of two embryonic primordia
from a single one. According to Patterson, these are primary and second-
ary buddings, but Newman regards them as fissions. Since they involve
the origin of new axiate patterns, they seem to resemble buds more closely
than fissions. The new patterns apparently originate in definite relation

'3 See, e.g., Klaussner, 1890; Dareste, 1891; Bateson, 1894; H. H. Wilder, 1904, 1908;
Schwalbe, 1907; Gemmill, 1912; Newman, 1917(7, 1923; and citations by these authors.

^■^ Newman and Patterson, 1909, 1910, 1911; Patterson, 1912, 1913; Newman, 1917a,
1923.

540 PATTERNS AND PROBLEMS OF DEVELOPMENT

to environmental factors, for the first duplication is right and left with
respect to the parent, on the two sides of the primary embryonic vesicle
toward the openings of the Fallopian tubes, and the second duplication is
in definite relation to the first.

Newman (1923) holds that this polyembryony is associated with occur-
rence of a quiescent period in development, during which the original
developmental pattern is, to a large extent, obliterated; and with renewal
of developmental activity the reaction to, and the determination of, a
new pattern by environmental differentials within the uterus results.
Even after this, however, the integrating factor is not adequate to pre-
vent a second physiological isolation of parts ; consequently, four embryos
are formed. The entire sequence of reconstitutions in D. hyhridus has not
been followed, but the irregular arrangement and variable number of em-
bryos in that species indicate that some of the embryonic primordia bud
or divide more than others. There seems to be no doubt that this type of
polyembryony involves repeated obliteration and determination of de-
velopmental patterns, and its occurrence in the mammals suggests that
under proper conditions pattern in various other eggs may not be as stably
determined as commonly assumed.

FUSIONS OF EMBRYONIC AND LARVAL INDIVIDUALS

Larvae of the sponge Lissodendoryx, merely brought into contact at
certain stages, fuse readily into masses consisting of indefinite numbers of
larvae in which all evidences of individual form and polarity disappear.
These masses, in contact with a solid substratum, may metamorphose
into perfect sponges (H. V. Wilson, 1907). Evidently there is complete
obliteration of the original larval patterns in these masses, and the polarity
of the resulting sponge is determined anew by an environmental differen-
tial, probably the differential arising between free surface and surface in
contact with the substrate, since development of the osculum or oscula
is on the free surface, as in aggregates of dissociated sponge cells (p. 418).
Like the cell aggregates, these fusion masses would be interesting material
for other experiments on determination of pattern.

Normal planulae from fused blastulae of the medusa Milrocoma have
been described, but nothing is known concerning possible changes in pat-
tern. Fused cell masses resulting from cleavage in calcium-free sea water
of eggs of the nemertean Cerehratidus may, after return to normal sea
water, develop into giant pilidium larvae with multiple organs (Yatsu,
1910c). Apparently in these the original pattern is not completely oblit-

EMBRYONIC RECONSTITUTIONS 541

erated; but whether, or to what extent, it may be altered, is not known.
Fusion among the several eggs in a capsule, either before or during cleav-
age, has been observed in another nemertean. Linens ruber; and fusion of
two eggs may result in normal gastrulae with cells approximately double
size (Nusbaum und Ochsner, 1913).

The giant eggs of Ascaris megalocephala undoubtedly result from fusion
of two or more eggs. Completeness of fusion varies widely ; the giant eggs
may be doubly fertilized, form two sets of polar bodies at different points,
and show more or less double cleavage; or in some cases they may de-
velop as a single individual. '^ Apparently the variations depend on the
relations of the apicobasal axes, on the degree of physical union, and per-
haps on other factors. Supposedly, development as a single individual
occurs only when polarities of fusing eggs are parallel and identical in
direction.

Fused unfertilized and fertilized sea-urchin eggs and early developmen-
tal stages give forms ranging through various degrees of twinning to single
individuals. Skeletal duplications or excessive skeletal development and
duplications of the archenteron are frequent.^'' Fusions of more than two
individuals are usually highly abnormal and die early. Single giant larvae
may develop from fusions with axes known not to be parallel, and twin-
ning may occur when axes are parallel. Extensive shiftings and rotations
of the components, either toward or away from paralleHsm, may take
place. The larger component may dominate the smaller to such a degree
that suppression of its development, reduction in size with translocation
of cells, and even complete incorporation into the body of the dominant
member may result. Translocation of mesenchyme cells to the dominant
member indicates that the regions of its ectoderm which determine locah-
zation of mesenchyme are more effective than those of the subordinate
member. At present it appears difficult to interpret the observed results
except in terms of dynamic factors. Investigation by means of differential
dye reduction of the gradient relations and the changes which they under-
go in these fusions, particularly in cases of suppression and absorption of
one member, would undoubtedly be of interest.

First cleavage stages of Ascidiella aspersa united in various orientations
develop as more or less completely double forms, often with organ anoma-
hes and dislocations, or as forms apparently single externally but more

^s Sala, 1895; Zur Strassen, 1898, 1906; Kautsch, 1913.

*' Driesch, 1893, 1900, 1903, 1910; Morgan, 1895c; Garbowski, 1904; Bierens de Haan,
1913d, b; Goldfarb, 19146, 1915, 1917; von Ubisch, 1925(2; Schleip, 1929, pp. 484-95.

542 PATTERNS AND PROBLEMS OF DEVELOPMENT

or less completely duplicated internally ; or in a single case among more
than a hundred developing, a single individual resulted from union with
apical pole of one component in contact with basal pole of the other (von
Ubisch, 19386). The author holds that the general organ-forming regions
develop according to their prospective significance but that within them
there may be extensive reconstitution. Two such regions may give rise to
a single organ system, or one may form two systems.

Homoplastic and heteroplastic fusions of two two-cell stages have been
accomplished with interesting results in species of the urodele Triton.
After removal of membranes the blastomeres of the two-cell stage become
almost spherical and are connected only by small areas. At this stage one
pair of naked cells is laid crosswise on the other, and the four cells come
gradually to lie in a plane, the blastomeres of one component alternating
with those of the other. From these fusions one to four axial systems may
develop. ^^ A single embryo resulting from a heteroplastic fusion is a chi-
mera, that is, composed of cells of two species. The results of these ex-
periments are interpreted in terms of the relation of the first cleavage
plane to the median plane and the consequent positions in the fused pair
of the dorsal inductor tissue, assuming that this inductor region is already
more or less definitely locahzed and determined at the two-cell stage.
Since the first cleavage is known to make any angle with the median
plane, the dorsal inductor region of each two-cell stage may be divided by it
into equal or unequal parts or be entirely in one cell, and the relation of
cleavage plane and median plane may be different in the two fused com-
ponents. Consequently, neural induction and embryonic axes will appear
in the fused pair in various positions and combinations of parts. The re-
sults agree with expectation and are in accord with other experiments
(chap, xii) in indicating that a part of the inductor may reconstitute to a
whole, that induction is not species-specific, and that such symmetry
pattern as may be present at the two-cell stage may undergo extensive
alteration in the reconstitutions resulting from the fusions.

CONCLUSION

It appears that embryonic reconstitutions in many animals do not
differ very greatly from reconstitutions in adults of the lower inverte-
brates, except that they are usually more narrowly Hmited. They com-
monly show definite relations to the original pattern, but sometimes that
is completely obliterated and new pattern determined. The embryonic

*7 Mangold, 1920; Mangold und Seidel, 1927; also Spemann, 1938, pp. 271-77.

EMBRYONIC RECONSTITUTIONS 543

reconstitutions provide further evidence in support of the view that the
effective factors in the developmental patterns of early stages must be
sought in the dynamics of living protoplasms rather than in concentra-
tions of hypothetical formative substances. As we pass from later to ear-
lier stages of embryonic development, there is progressively less evidence
of regional specificity, and in some eggs at the beginning of embryonic
development there is apparently little or nothing more than quantitative
gradient pattern. In others regional specificities are already present at
the time of the first cleavage, though evidences of gradient pattern may
still be present. Some of these, in which the original pattern tends to
persist with little or no change in isolated parts and development appears
to be more or less completely a mosaic of independent, self-differentiat-
ing systems, are discussed in the following chapter.

CHAPTER XIV
CLEAVAGE AND DEVELOPMENTAL PATTERN

IF NUCLEI are primarily alike as regards hereditary potentialities,
the physiological basis of embryonic developmental patterns must
be sought in the egg cytoplasm; and the question whether, or to what
extent, the widely different cleavage patterns of different animals are re-
lated to developmental pattern arises. Some cleavage patterns — for ex-
ample, those of annelids, mollusks (except cephalopods), ascidians, and
some other forms — appear to be intimately related to developmental pat-
tern; they are highly determinate, that is, certain cells give rise to certain
organs or parts by a definite cell lineage, and at least some of the cells,
when isolated, are capable of continuing development for a time with
little or no change. For this reason earlier stages of these forms have
often been regarded as "mosaics" of independently developing parts. At
the other extreme are completely indeterminate cleavage patterns with-
out any definite relation to developmental pattern, as in polyembryonic
bryozoa, in insects and probably most other arthropods, and in mero-
blastic vertebrate eggs. Between these extremes are various degrees of
determinate character. In other, less speciahzed forms of development,
budding, fission, and reconstitution of multicellular forms the single cell
has no definite relation to developmental pattern. This suggests that the
more highly determinate types of cleavage may be expressions of a certain
degree of determination or differentiation of regions of the undivided
egg. This chapter is largely concerned with the more highly determinate
cleavage patterns and with questions regarding their mosaic character
and relation to pattern of the organism.

SPIRAL CLEAVAGE PATTERN

As far as mitotic spindles and cleavage planes are concerned, so-called
''spiral cleavage" is actually oblique. To an observer in the egg axis with
head toward the apical pole the spindles are inclined with upper poles
to right (dexiotropic) or to left (leiotropic), and the upper cell is obliquely
dextral or sinistral to the lower. This type of cleavage has been called

544

CLEAVAGE AND DEVELOPMENTAL PATTERN 545

"spiral" because it may be conceived as resulting from a spiral twisting of
planes of radial or bilateral symmetry into surfaces forming a spiral about
the polar egg axis. There is, however, no evidence of such twisting, but
the term "spiral cleavage" has been so generally employed that it is used
here. In general, this pattern of cleavage gives way sooner or later to
more or less bilateral patterns; but as long as it persists, each spindle is
approximately perpendicular to that of the preceding division of the cell
concerned, and consequently successive cleavages are alternately dexio-
tropic and leiotropic. Because this form of cleavage occurs in the earlier
stages of polyclad turbellaria, annelids, gephyreans, and mollusks, except
cephalopods, and because it is sufhciently determinate in character to
permit the following of cell Hneage from first cleavage to certain regions
or organs of the larva or later embryos, it has received much attention and
has been largely responsible for the concept of ontogeny as a mosaic and
for theories of the phylogenetic significance of cleavage and of cell homolo-
gies.^

Various forms and stages of spiral cleavage are shown in Figures 171-76.
Different terminologies for designating the blastomeres have been used,
but the following has become more or less standard. The first cleavage is
meridional and may be equal (Fig. 171, A) or unequal (Figs. 172, A;
175,5). When it is unequal, the smaller cell is yl 5, the larger, CZ). Second
cleavage is also meridional and divides AB equally into A and B; and in
those forms in which the first cleavage is distinctly unequal, the second
divides CD into a smaller cell, C, and a larger, D. In polyclads and nemer-
teans the four quadrants are all alike and indistinguishable (Fig. 171, B).
In the third cleavage the first quartet of "micromeres" (la-id) separates
from four basal "macromeres" (lA-iD). The so-called "micromeres" are
usually smaller than the basal macromeres (Figs. 172, C; 173, A; 174,
A, C) but in some forms are equal in size to, or larger than, the latter
(Fig. 171, C); and definite size differences characteristic for the species
are often present, id and often ic being larger than la and ib (Figs. 172,

' For more or less extensive descriptive studies of spiral cleavage see the following: Poly-
clads: Hallez, 1879; A. Lang, 1884; E. B. Wilson, 1898; Surface, 1907. Nemerleans: Zeleny,
1904. Annelids: Whitman, 1878; Salensky, 1882-83, 1885; Vejdovsky, 1883, 1892; Hatschek,
1886; E. B. Wilson, 1892; von Wistinghausen, 1893; Wheeler, 1897; Mead, 1897; Eisig, 1898;
Child, 1900; Treadwell, 1901; Schleip, 1914a; Penners, 1922, 1923, 1924a, b, 1925. Gephyreans:
Griffin, 1899; J. C. Torrey, 1903; Gerould, 1906. Mollusks: Rabl, 1879; Hatschek, 1881;
Blochmann, 1882, 1883; Heymons, 1893; Kofoid, 1895; F. R. Lillie, 1895; Meisenheimer,
1896, 1901; Conklin, 1897, 1907; Drew, 1899; Heath, 1899; S. J. Holmes, 1900; Robert, 1903;
E. B. Wilson, 1904; Wiersejski, 1906.

546

PATTERNS AND PROBLEMS OF DEVELOPMENT

C; 173, A; 174, A). Division of both apical and basal quartets follows,
the apical usually preceding and giving rise to ia\ ia^-id\ and id\ the
basal to 2a-2d and 2A-2D. In most forms four or live quartets of micro-
meres are formed, each farther from the apical pole than the preceding,
before the spiral pattern is altered or obliterated in the cells concerned.
Since the spindles are alternately dexiotropic and leiotropic, the blasto-
meres interlock; and when the cleavages are equal or nearly equal, the
planes of contact are essentially similar to those resulting from surface
tension, as in a mass of soap bubbles (Robert, 1903). As far as known,

Fig. 171, A-C. — Early cleavages of the nemertean Cerebratuhis (after Zeleny, 1904)

certain descendants of the first quartet form the whole or part of the
prototroch; but numbers and origins of cells in prototrochs of different
species differ widely, and the series of divisions leading to its formation
also differ, or, when similar, are similar because they are spiral cleavages.
And in forms without prototroch, cells equivalent in origin to the trocho-
blasts form other parts of the ectoderm. Later divisions of the pretrochal
cells about the apical pole also differ in different species. The second
quartet of micromeres was regarded by Lang (1884) as entirely meso-
dermal in polyclads; but, according to E. B. Wilson (1898) and Surface
(1907), these cells in all four quadrants give rise to both ectoderm and
mesoderm. In the gasteropod Crepidula three of these cells {2a, 2b, 2c)

A

B

C

D

Fig. 172, yl-F.— Cleavage stages of Arenicola. A, two-cell stage; B, second cleavage; C,
eight-cell stage; D, second division of first quartet of micromeres and of the cell 2d (first somato-
blast) ; E, primary trochoblasts, shaded, and two-cell stage of 2d, surrounded by heavy line,
also the cell M {4d), the second somatoblast or mesoblast; F, later stage, showing development
of somatic plate, surrounded by heavy line, the two mesoblasts, outlined in broken line, al-
ready in blastocoel, and entoderm cells indicated by parallel broken lines (after Child, 1900).

B

Fig. 173, A, B. — Cleavage stages of Tubifex. A, eight-cell stage. B, later stage, showing
2da.nd4d{M) (after Penners, 1922). C-£, cleavage stages of leech, Clepsine { = Glossiphonia);
C, early stage, showing 2d (first somatoblast) and jD (second somatoblast or mesoblast) ; D
later stage, showing E, E, E, E, products of 2d, and M, M, the mesoblasts, products of 3D;
E, ectodermal teloblasts and germ bands from 2d, surrounded by heavy line, and mesoblasts,
M, M (after Schleip, 1914(7).

CLEAVAGE AND DEVELOPMENTAL PATTERN

549

give rise to ectoderm and mesoderm, and the fourth izd) is entirely ecto-
dermal (Conklin, 1897). In Unio three are wholly ectodermal, only one

Fig. 174, A-D. — Cleavage stages of mollusks. A, Unio, first and second quartets, 2d
largest (after F. R. Lillie, 1895); B, Crepidula, 2d no larger than other second quartet cells
(after Conklin, 1897); C, D, Fiilgur, first and second quartets, including 2d, all relatively very
small (after Conklin, 1907).

{2h) ectomesodermal (F. R. Lillie, 1895); while in polychetes all are ap-
parently wholly ectodermal, except for the doubtful origin of mesoderm
from cells of the second or third quartet of Aricia (Wilson, 1898). If the

550 PATTERNS AND PROBLEMS OF DEVELOPMENT

origin of mesoderm from cells of the second or of the third quartet is an
ancestral reminiscence, as has been suggested, it might be expected to
appear in greater degree in annelids than in mollusks, but apparently it
does not.

The cell 2d is similar in size to other members of the second quartet,
and its descendants apparently give rise to no more ectoderm than the
other cells of the quartet in polyclads and nemerteans; but in annelids
it is usually larger than other members of the quartet, is called "first
somatoblast," and its descendants constitute most of the trunk ectoderm,
giving rise in oligochetes and leeches to the ectodermal teloblasts from
which the ectodermal germ bands develop (Figs. 172, Z), £, /^; 173, B-E),
but the cleavage patterns by which these results are accomplished differ
in almost every species studied. Among the mollusks the cell 2d may be
much larger than other cells of the second quartet, as in Unio (Fig. 174,
A), 01 the same size as others (Fig. 174, B, D). In Unio shell gland and
pedal ectoderm are derived from it (Lillie, 1895); and that this is also
true for Crepidula is held by Conklin (1897, iQo?)) i^"^ spite of the great
difference in size of the cell and time of appearance of shell gland as
compared with Unio. The third quartet (ja-jd) is regarded as ectodermal,
except that jb is said to give rise to entoderm in the oligochete Tubifex
(Penners, 1922).

The cells 4a, 4b, and 4c of the fourth quartet become entoderm in cases
followed to this stage. In annelids and mollusks 4d, the second somato-
blast, is much larger than other cells of the quartet and becomes meso-
derm (M of Fig. 172, £, /^, and of Fig. ij^, B, D, E). In the leech. Clepsine,
according to Schleip (1914), 4D also becomes mesoderm (Fig. 173, D, E);
but in other forms it gives rise to entoderm, and what remains of the
"macromeres" after formation of the fourth quartet also becomes ento-
derm.

In many annelid and mollusk eggs certain cytoplasmic regions are vis-
ibly distinguishable, either by absence of yolk or by presence of certain
granules. These regions may apparently undergo definite changes in posi-
tion in connection with maturation and fertihzation and be distributed
to particular cells during early cleavages. Very generally the amount of
yolk and other granular inclusions increases basipetally from the apical
polar region; but in many eggs a well-defined zone, aggregation, or ring
of cytoplasm containing little or no yolk is present about the apical pole,
either before maturation or appearing later; a second aggregation often
appears about the basal pole. These polar plasms are, or become, very
definitely localized in eggs of Tubifex (Penners, 1922) and Clepsine (Whit-

CLEAVAGE AND DEVELOPMENTAL PATTERN 551

man, 1878; Schleip, 1914), which contain much yolk; in Dentalium eggs
they appear less definitely bounded and are separated superficially by a
pigmented, yolk-containing zone (E. B. Wilson, 1904). Some other eggs
show one or both of them in differing degrees. In consequence of the
definite character of cleavage these polar plasms are included in certain
cells; in Clepsine both are entirely or almost entirely included in the cell
CD by the first cleavage and in D by the second. Before the third cleavage
the basal polar plasm migrates apically and unites with the apical plasm,
and the entire mass remains in iD at the third cleavage and at the next
division is divided between 2d and 2D. The ectodermal germ bands de-
velop from descendants of 2d, and the mesoderm bands from 4d and 4D.
These derivatives of D constitute almost the whole mass of the cell, 3d
being small. Very similar behavior of the polar plasms has been described
for oligochetes, and eggs in which polar plasms do not appear fail to
develop or die in early stages and fail to develop somatoblasts (Vejdovsky,
1888; Penners, 19246). Penners apparently believes that failure to de-
velop and death result from absence of the polar plasms, but it seems
possible that absence of polar plasms may result from physiological or
pathological conditions and is only incidentally associated with develop-
mental failure and death.

A temporary cytoplasmic yolk lobe or polar lobe appears basally in
many mollusks and annelids at the first cleavage and in some species
again at second and third cleavages. The three successive lobes of the
first three cleavages of Dentalium are shown in Figure 175, A-D. Nor-
mally the lobe may become almost completely separated from the rest of
the egg by constriction ; but it is actually a part of one blastomere — CD
in the first cleavage, D in the second, and iD in the third. In Dentalium
it consists largely of the unpigmented basal plasm, but in some other forms
it contains mostly yolk. However, in lobe-forming centrifuged eggs the
lobe appears quite independently of the cytoplasmic stratification and
may consequently differ in content in different individuals, according to
the direction of stratification (p. 585). Its formation in connection with
cell division and the flow of cytoplasmic substance into it as it enlarges
led Boveri (1910a) to suggest that it represents apart of the cell not in-
cluded in the sphere of influence of the adjoining aster; but the periodic
form changes in isolated lobes of Ilyanassa, corresponding more or less
closely to periods of the division cycle, occur in cytoplasm entirely isolated
from the dividing cell (Morgan, 1933, 1935, 1936). Hydrostatic pressure
of 220 atmospheres, applied in early stages of lobe formation, brings about
withdrawal of the lobe and inhibition of cytoplasmic division in Chae-

552

PATTERNS AND PROBLEMS OF DEVELOPMENT

toptcrus. This result, like Morgan's centrifuge experiments, indicates that
the lobe is a cortical effect, associated with cytoplasmic division but
largely independent of the mitotic apparatus (Pease, 1940). Experi-
mentally the first lobe can be made to fuse with either of the cells of the
two-cell stage in the mollusks Dcntalium (Schleip, 1939, p- 208) and
Ilyanassa (Morgan, 1936) and in the annelid Sahellaria (Novikoff, 1940,
see footnote 8, p. 559), and the cell receiving it develops like a CD cell.

B

C D

Fig. 17s, A-D. — Early cleavages of Denlalium, to show the three successive polar lobes,
Li-Lj, and approximate boundaries of pigmented zone, indicated by dotted lines (after E. B.
Wilson, 1904).

The earlier, purely descriptive studies of cell lineage in forms with spiral
cleavage led to a further development of the hypothesis of "organ-forming
germ regions" advanced by His (1874). The definite cytoplasmic localiza-
tion, "precocious segregation," of different organ-forming substances and
their accurate distribution by cleavages was postulated, and the concept
of homology was applied to individual cells in the cleavage pattern, cer-
tain small cells in some forms being regarded as vestigial or rudimentary.
According to these views, development in forms with spiral cleavage is a

CLEAVAGE AND DEVELOPMENTAL PATTERN 553

mosaic, that is, it proceeds autonomously in the varous self -differentiating
parts.^

However, as data on different species accumulated, it became increas-
ingly evident that supposed cell homologies were in many cases less exact
than had been supposed. The first cleavage is said to be almost trans-
verse to the median plane in some annehds {Nereis, Chaetopterus) and in
some mollusks (Crepidula, Umbrella), while in certain polyclads (Thy-
sanozooyi, Lepto plana), other annelids {Clepsine, Amphitrite, Arenicola),
and mollusks {Planorbis, Unio, Trochus) it is approximately 45° from the
median plane. These differences necessitate the assumption of at least
some differences in cellular localizations of different formative substances.
The first cleavage may be equal or unequal, apparently without definite
relation to the fates of the cells. There is apparently a relation between
spiral cleavage and the later asymmetry in gasteropods: spirals of cor-
responding cleavages are reversed in direction in certain sinsistral species. ^
Conkhn (1903a, b) suggested that this might result from reversal of polar-
ity in eggs of these forms, but no evidence of reversal has been found.
Moreover, spiral cleavage is not accompanied by any general asymmetry
in polyclads, annelids, and pelecypods.

The later cleavages of the first quartet and the fates of some of the
cells differ in different species. Certain descendants of the first quartet
become cells of the prototroch in some species; while in others corre-
sponding cells form, but there is no prototroch. The first somatoblast,
2d, may be of the same size as other cells of the second quartet or much
larger, and this difference in size is apparently without definite relation
to the final relative size of the parts developing from it or to the stage
at which they differentiate. Also, the later divisions of 2d, as far as they
have been followed, differ as regards directions and sizes of products in
almost every species studied. Such differences are well shown in earlier
stages of the somatic plate in Arenicola, the part surrounded by heavy
line in Figure 172, £ and F, and the ectodermal teloblasts and early germ
bands of another annelid, Clepsine, similarly indicated in Figure 173,
D and E. Some cells of the second quartet are said to give rise to both
ectoderm and mesoderm in certain species; but in others, apparently only
to ectoderm.

Perhaps the cell 4d, the second somatoblast, which becomes mesoderm

' See, e.g., E. B. Wilson, 1892, 1894, 1896, 1898; F. R. Lillie, 1895; Mead, 1897; Conklin,
1897, 1898, 1907; Heath, 1899.

3 Crampton, 1894; S. J. Holmes, 1899, 1900; Wiersejski, 1906.

554 PATTERNS AND PROBLEMS OF DEVELOPMENT

in annelids and mollusks, presents the best case for cell homology in these
groups; but even here there are difficulties. According to Schleip, it gives
rise to only one of the mesodermal bands in Clepsine, the other being
formed from 4D, which is entodermal in other forms. Moreover, 4d is
apparently entodermal in polyclads. Actually there are very few, if any,
exact cell homologies in the groups with spiral cleavage. In order to apply
the concept of homology to the cells, we must, in any case, regard certain
cleavages as accomplishing segregation and others as nonsegregating. Vis-
ibly different regions are often more or less exactly segregated in different
cells, but the centrifuge experiments show that this segregation has little
or no significance for development in most cases. If the principle of homol-
ogy is applicable to egg cytoplasm, the homologies seem to be, at best,
regional; and their more or less exact coincidence with cell boundaries
appears to be an incidental or secondary result of regional cytoplasmic
differences of some sort rather than fundamentally significant for develop-
ment.

A particularly interesting feature of development with spiral cleavage
is the gradual appearance of bilateral pattern in the cleavage pattern and
its inexact character in early stages of most forms. Different species and
different regions of the embryos provide various examples. The annelid
first and second somatoblasts, 2d and 4d, and in Clepsine also 4D, are
cases in point. The cell plate (somatic plate) resulting from divisions of
2d in polychetes extends laterally on each side from the dorsal region and
gradually acquires a roughly bilateral form (Fig. 172, F), but up to an
advanced stage most of the cleavages in it have not been bilateral. More-
over, the median plane of the two mesoblasts does not coincide with that
of the ectoderm of the somatic plate (Fig. 172, F), but in later stages
coincidence is gradually attained. In Clepsine bilateral pattern appears
earlier in descendants of 2d (Fig. 173, D); but even after the germ bands
begin to form, the arrangement of the teloblasts is not necessarily com-
pletely bilateral, and the median planes indicated by the ectodermal telo-
blasts and by the two mesoblasts are farther from coincidence than in
polychetes (Fig. 173, E).

The pattern resulting from spiral cleavage is essentially quadriradial
in polyclads, in nemerteans, as far as known, and in some annelids and
mollusks. In this pattern there gradually appears a dorsiventral pattern
with bilateral symmetry or, in the gasteropods, asymmetry. These facts,
quite apart from experimental data, raise the question whether this form
of development is as completely a mosaic as has been believed. Is it pos-

CLEAVAGE AND DEVELOPMENTAL PATTERN 555

sible to account for the gradual replacement of the spiral and radial pat-
tern by a bilateral or asymmetric pattern except in terms of an ordering
and integrating factor, a wholeness of some sort which determines the
gradual appearance of the new pattern? It seems beyond the bounds of
probability that a pre-established harmony in a mosaic of independent
cells or cell groups resulting from spiral cleavage could be so complete as
to accomplish this result. Whatever the conditions determining spiral
cleavage, they are apparently different from those determining ventro-
dorsal or bilateral pattern. Even in the gasteropods, in which reversal of
cleavage is associated with reversal of asymmetry, the cleavage pattern
and the pattern of asymmetry are very different. Neither cell homologies,
nor regional homologies, nor precocious segregation throw any light on the
physiological factors concerned in spiral cleavage.

A physiological gradient pattern, at least in the polar axis, is indicated
in some forms by differential susceptibility, differential dye reduction, and
differences in hydrogen-ion concentration, and the differential in rate of
cleavage also suggests a polar gradient (pp. 119, 545) ; but in many forms
apicobasal pattern is apparently not merely a quantitative gradient but
a pattern of more or less definitely localized material differences, some-
times directly visible in the living egg. This pattern, however, is appar-
ently in large part or wholly an incidental and secondary result of the
essential pattern and can be altered to an extreme degree by centrifuging
without affecting development in most forms with spiral cleavage.

EMBRYONIC RECONSTITUTIONS IN FORMS WITH SPIRAL CLEAVAGE

NEMERTEANS

The nemerteans Cerehratulus lacteus and C. marginatus have a typical
spiral cleavage ; but development of isolated parts of eggs before and after
fertilization shows, first, that they are not primarily mosaics and, second,
that there is a progressive regional determination or stabilization with
early cleavages.'* Nucleated or nonnucleated pieces of unfertilized eggs
sectioned in various planes, can be fertilized, cleave like whole eggs and de-
velop into normal dwarf pilidium larvae. According to Wilson, however,
pieces less than about one-fourth of total egg volume do not give rise to
complete larvae. Here, as in other embryonic and postembryonic recon-
stitutions, scale of organization is decreased in the isolated pieces; and
in pieces below a certain size the scale may be larger than the piece, and
partial forms may result.

■f E. B. Wilson, 1903c; Zeleny, 1904; Yatsu, 1910^, c.

5S6 PATTERNS AND PROBLEMS OF DEVELOPMENT

Removal of parts of the egg between fertilization and first cleavage
shows increasing disturbance of cleavage pattern the later the removal,
but essentially normal larvae may develop. In isolated 1/2 and 1/4 blas-
tomeres the cleavage pattern is a half or fourth of the whole, except as
shifts in cell position occur; but normal larvae may still result (Wilson).
Developmental effects of removal of parts of one or both blastomeres dur-
ing or after first cleavage are slight (Yatsu). The apical quartet of the
eight-cell stage gives rise to larvae with apical flagellum but without
enteron; the basal quartet, to larvae with enteron but no apical organ
(Zeleny) . Apical and basal parts of blastulae isolated by section show much
the same differences (Wilson). The three experimenters conclude that
there is progressive apicobasal localization of formative materials during
early cleavages. More recent experiments agree in general with these re-
sults. Isolated blastomeres of the two-cell stage may form complete, sym-
metrical larvae, an extensive reconstitution. The four apical or the four
basal cells of the eight-cell stage, when isolated, give rise to apical and
basal partial forms; and isolated rings of cells an^, an^, veg„ and veg.,
of the sixteen-cell stage, ^ and various combinations of these rings, devel-
op as if the other blastomeres were present (Horstadius, 19376). Nothing
is known at present concerning gradients in the nemertean egg; but the
unfertiUzed egg evidently possess a polarity, and its high reconstitutional
capacity suggests that this polarity is a predominantly quantitative dif-
ferential at that stage but undergoes progressive regional differentiation in
early development. Certainly neither the egg nor early cleavage stages
are mosaics.

DUPLICATIONS AND OTHER RECONSTITUTIONS IN Tubifcx EMBRYOS

It was noted earher (p. 537) that development of complete or partial
twins from single eggs is not infrequent in certain oligochetes under sup-
posedly natural conditions. Partial duplications may involve either the
anterior or the posterior end or both, but in Tubifex these duplications orig-
inate in different ways (Penners, 1924a, b). Simple duplications of the
anterior end result from failure of the germ bands of the two sides to
come together anteriorly; the bands of each side give rise to an anterior
end, symmetrical as regards body wall, nervous system, and entoderm
but with only one dorsal and one ventral series of setae and one nephrid-
ium in each segment, seta-sacs and nephridia of the apposed sides not
developing. Penners suggests that these forms result from action of some

s The designations aiir veg2 are the same as used for the sea-urchin embryo (p. 438).

CLEAVAGE AND DEVELOPMENTAL PATTERN

557

unknown inhibiting factor which prevents normal union of germ bands
anteriorly. It may be suggested further that absence of setae and ne-
phridia on the apposed sides may result not from lack of capacity to re-
constitute these organs but from mutual inhibition of the two parts, as in
many cases of postembryonic rcconstitution. The other form of partial
twinning is a cruciate duplication, resulting from duphcation of the germ
bands and union of one band of each pair to form the two anterior ends
(Fig. 176). This type of duplication has also been produced experimen-
tally by subjecting eggs of Tubifcx to high
temperature together with low oxygen
content of water (Penners, 1924/'). Under
these conditions the first cleavage is equal
instead of unequal, as normally; and in
later cleavages two 2d- and two 4d-celh,
instead of one, appear. Equal division of
the cell CD at the second cleavage gives
the same result. Each of the 2d-ce\h
gives rise to a pair of ectodermal germ
bands, and each of the 4d-ce\h to a pair
of mesodermal bands. According to Pen-
ners, duplication of the somatoblasts
results from equal division of the polar
plasms (p. 550) by the equal first cleav-
age.

Development of certain blastomeres or blastomere groups of Tubifex,
after killing others by localized ultra-violet radiation, high temperature,
or strong shaking, has also been followed (Penners, 1926). The dead cells
soon separate from the living. Further cleavage patterns of isolated CD-
or Z)-cells containing the polar plasms is not altered; they give rise to
the two somatoblasts, these to teloblasts and germ bands, and normal
embryos result. The germ band of one side can develop and differentiate
when that of the other side is killed. When the polar plasms are killed,
germ bands do not develop, and the embryo consists of a compact mass
of entoderm with an ectodermal cap of variable size. Ectodermal germ
bands do not develop when the first somatoblast is removed, but meso-
dermal bands develop and cUfferentiate. Killing of the cells 2D or jZ),
which are in the mesoderm line, or of the mesoblast 4d or its two daughter
cells, results in absence of mesoderm. On the other hand, normally pro-
portioned embryos develop when only part of the entoderm cells are pres-

FiG. 176. — Posterior parts of ecto-
dermal germ bands in cruciate duplica-
tion of Tubifex (after Penners, 1924a).

5S8 PATTERNS AND PROBLEMS OF DEVELOPMENT

ent. In these cases the germ bands undergo less growth, and the embryo
is smaller than normal.

According to these data, the Z)-quadrant, containing the polar plasms,
is potentially a whole Tubifex embryo, or even two embryos; but the
A-, B-, and C-quadrants normally play a part in embryo formation. In
their absence the parts normally formed by them are reconstituted by
cells originating from the Z)-quadrant. Evidently, as regards this quad-
rant, development is not a mosaic. Penners regards the polar plasms,
which are contained in the Z)-quadrant, as "organ-forming substances"
and as determining the potentialities of the two somatoblasts.^ It is, how-
ever, an interesting question whether this conclusion, if correct, means
anything more than that the polar plasms represent all the potentialities
of the Tubifex egg cytoplasm, except perhaps those of the entoderm.
What sort of formative substance can it be which gives rise to the entire
ectoderm and mesoderm of an embryo or of two embryos? Are the polar
plasms anything more than active, yolk-free parts of the egg cytoplasm?
Do they differ from the entodermal cytoplasm except in being yolk-free?
As yolk-free cytoplasmic aggregations, they may be primarily regions of
more intense metabolism than other parts and consequently more or less
dominant.

Many other annelids and mollusks show more or less definite aggrega-
tions of cytoplasm apically or basally or both, but their fates in cleavage
differ. In some cases they are localized in the micromeres of the first
quartet, and new yolk-free aggregations appear apically in the macro-
meres preceding formation of each quartet of micromeres. In some forms
they disappear completely in early development. The case for specific
significance of the polar plasms does not appear entirely convincing. Un-
doubtedly the cells CD and D and the somatoblasts 2d and 4d differ in
some way from the other cells; but, since they are able to reconstitute a
whole embryo if a part of the entoderm is present, the difference seems to
be in the direction of less, rather than more, narrowly limited specificity.

DUPLICATIONS IN OTHER ANNELIDS AND MOLLUSKS

Duplications have been produced experimentally in the polychetes
Chaetopterus, Nereis, and SaheUaria and in the pelecypod Cumingia by
subjection of eggs to pressure, low temperature, high temperature, cen-
trifuging, anaerobiosis, and KCN during a definite period preceding cleav-

' See also Schleip, 1914a, h, 1929; von Parseval, 1922.

CLEAVAGE AND DEVELOPMENTAL PATTERN 559

age.'' According to Tyler, the usual condition resulting in duplication is
equal first cleavage with spindle at right angles to its normal position
but in, or parallel to, the equatorial plane of the egg. In centrifuged
Chaetopterus eggs, however, the polar lobe may be attached to the other-
wise smaller cell of an unequal first cleavage, and with incorporation of
the lobe into the cell the two blastomeres become equal. The two blasto-
meres resulting from equal cleavage have the potentialities of the CD-
blastomere of normal cleavage, and duplications result, as in Tubifex.
These vary in degree, some being clearly cruciate forms, the duplication
involving pretrochal, as well as posttrochal, regions; others, with only
posttrochal duplication. The apical organ is usually not duplicated, but
eyespots, mouth, and other organs may be. In general the completeness
of duplication is greater in the posttrochal, than in the pretrochal, region.
Both 1/2 blastomeres isolated following equal cleavage can develop ap-
parently normal trochophores. The pretrochal duplications indicate that
even in this region there may be some alteration in fates of some of the
cells in consequence of altered relations to other cells. The wide range in
degree of completeness of posttrochal dupHcation and the frequent de-
velopment of complete and separated posttrochal regions indicate that
the somatoblasts are totipotent for ectoderm and mesoderm of this region
and capable of reconstitution.

ISOLATIONS AND TRANSPLANTATIONS IN ANNELID AND MOLLUSK EMBRYOS

Isolation experiments with cleavage stages of the polychete Sahellaria
show that the formation of the apical tuft of cilia depends on the presence
of the first polar lobe and the C-cell; development of the posttrochal re-
gion, on the presence of D. Transplantations and unions of blastomeres
and of whole eggs show complete self-differentiation of all blastomeres.
Exogastrulae produced by treatment with alkahne isotonic NaCl show
completely independent development of ectoderm and entoderm with no
evidence of induction. However, duphcations result from KCN treat-
ment; evidently reconstitution of pattern is possible in the region from
which the somatoblasts form.^

After removal of the polar lobe of the two-cell stage of Ilyanassa the
second cleavage is equal; and the cell 4d, the mesoblast in normal develop-
ment, is the same size as other cells of the fourth quartet, instead of much

7 Titlebaum, 1928; Tyler, 1930; Novikoff, 1939.

*Novikoff, 1938,1939, and 1940, "Morphogenetic substances and organizers in annelid
development," Jour. Exp. Zool., 85. See also Hatt, 1931, 1932.

S6o PATTERNS AND PROBLEMS OF DEVELOPMENT

larger, as normally, and the resulting larvae develop a prototroch but die
before attaining the vcliger stage. In general, isolated blastomeres cleave
as if other blastomeres were present, and may develop into defective
forms with or without cilia but with entoderm overgrown by ectoderm
(Crampton, 1896).

The studies by E. B. Wilson (1904) give additional information on de-
velopment of isolated parts of eggs and early stages of the mollusks
Dentaliiim and Patella. In the living, unfertilized Dentalium egg two white
polar areas are visible, the intervening region being superficially pigment-
ed: the basal white area enters the polar lobe when that forms. Apical
pieces of the unfertilized egg, when fertilized, do not form a polar lobe;
the first two cleavages are equal; and cihated swimming larvae may de-
velop but almost always lack the apical tuft of long cilia. The basal piece,
when fertilized, forms a polar lobe, cleaves like the entire egg, and may de-
velop into a normal trochophore. Pieces isolated by apicobasal section
through the basal white area may cleave like whole eggs and give rise to
"nearly normal" trochophores. The apical piece of the fertilized egg de-
velops like that of the unfertilized; but the basal, nonnucleated piece forms
the three polar lobes successively without dividing. After removal of the
polar lobe in the two-cell stage, the following cleavages are essentially
equal, and 2d and 4d are no larger than other cells; and in the resulting
larvae the posttrochal region is absent or small and rounded and never
develops further, and the apical tuft is also absent. If the first polar lobe
is allowed to complete its normal cycle with return to the cell body and the
second lobe is removed, the resulting larvae also lack the posttrochal re-
gion, but the apical tuft is present. Larvae from the yl5-blastomere of the
two-cell stage or from ^ , 5, or C of the four-cell stage are similar, except in
size, to those from which the first lobe has been removed; but larvae from
the CD- or the D-blastomere, of which the polar lobe is a part, possess
both posttrochal region, usually "too large," and apical tuft. Wilson con-
cludes that the material of the lobe must be specific and the determining
cause of development of apical organ and posttrochal region. Centrifuge
experiments indicate, however, that the polar lobe of another mollusk,
Ilyanassa, represents activity of the cortex which is not displaced by cen-
trifuging and that altered distribution in the lobe of other cytoplasmic con-
stituents has no effect on its formation or on further development. Effects
of hydrostatic pressure in inhibiting lobe formation and cleavage point to
the same conclusion.'' If the lobe is cortical, the relation between the first

' Morgan, 1935, 1936; Pease, 1940. See also F. R. Lillie, 1906.

CLEAVAGE AND DEVELOPMENTAL PATTERN 561

lobe and the apical tuft probably results from an inhibition of development
following removal of the lobe rather than from presence in the lobe of a
specific tuft-forming substance. In any case, effects of removal of lobe in-
dicate that development in the forms concerned is not a mosaic. The rela-
tion of lobe or somatoblast and apical tuft suggests the possibility that the
first somatoblast may have some inducing action. Novikoff (1940, see
footnote 8, p. 559) maintains that a cytoplasmic substance present in
the polar lobe of Sabellaria directs development of the cell or cells into
which it enters and is to be regarded as both a morphogenetic substance
and an organizer.

The first two cleavages of Patella are almost or quite equal; consequent-
ly, the Z)-quadrant cannot be distinguished in early stages, and 2d, sup-
posedly the first somatoblast, is no larger than other cells of the second
quartet. Isolated blastomeres cleave as if the other cells were present; but
partial forms, even those from 1/8 and 1/16 blastomeres, usually form closed
embryos and may gastrulate if entoblast cells are present. Isolated 1/4
blastomeres and isolated micromeres of the first quartet from all four
quadrants give rise to larvae with apical tuft, differing in this respect from
Dentalium, in which only the Z)-quadrant develops the tuft and only if the
first polar lobe is not removed. Isolated cells of the first quartet develop
according to their lineage into cells of the apical organ, cihated prototroch
cells, and other ectoderm cells. Apparently dorsiventral pattern is further
advanced in development at beginning of cleavage in Dentalium than in
Patella.

GENERAL SUGGESTIONS CONCERNING SPIR.A.L CLEAVAGE

The very general decrease basipetally in rate of cleavage in early stages
suggests an apicobasal gradient of some sort, and other evidences of such
a gradient have been found (pp. 119, 545). The fixity or stabihty of pat-
tern apparently undergoes some decrease basipetally. The regions of most
stable determination or specific constitution, as indicated by the isolation
experiments, represent apical or anterior regions of the larva and are essen-
tially head regions. In this respect the forms with spiral cleavage resemble
many other forms. Isolated apical regions of adult hydroids and heads of
planarians and annelids do not reconstitute other parts, but apical regions
or heads can reconstitute from lower levels. Essentially similar conditions
appear in sea-urchin embryos.

At some time in annelid development a ventrodorsal gradient must ap-
pear, at least in the posttrochal region, for such a gradient is characteris-

S62 PATTERNS AND PROBLEMS OF DEVELOPMENT

tic of the adult polychete and oligochete. In the polychete this gradient
may be initiated by the coming-together ventrally of the active borders of
the somatic plate and in the oligochetes by the apposition ventrally of the
germ bands.

The highly stable condition of the more apical regions in early stages
and the high capacity for reconstitution of the adults of many annelids
has sometimes been regarded as presenting a puzzling problem. However,
the most stably determined cells of the apical region, such as the trocho-
blasts and apical tuft cells, apparently take no part in postlarval develop-
ment, and even in the adult the head region is apparently fixedly deter-
mined. But in the embryo the Z)-quadrant and the somatoblasts derived
from it are totipotent for ectoderm and mesoderm and are capable of re-
constitution. They are no more determined or differentiated than post-
cephalic regions in the adult, perhaps less so.

The most striking features of spiral cleavage are the definiteness of
cleavage pattern in a given species and the differences of pattern in rather
closely related species. However, cleavage pattern can be altered in vari-
ous ways, as will appear in following sections. The cleavage pattern is evi-
dently an expression of a cortical or general cytoplasmic pattern of some
sort in the egg, but the oblique form of cleavage apparently masks to some
extent the organismic pattern in early stages, so that it appears only grad-
ually and does not at first correspond in different cell groups. Apparently
the factors that determine cleavage planes oblique to the polar axis are not
very closely associated with organismic developmental pattern, but differ-
ences in cell size are probably related in some way to this pattern. The
gradual appearance of bilateral pattern, the lack of correspondence in the
bilateralities of different regions in earlier stages — for example, in the two
somatoblasts and their derivatives — and the gradual "adjustment" in the
course of development are difficult to account for in terms of a develop-
mental mosaic. The fact that isolated cells or cell groups are capable of
more or less differentiation in the same way as if other parts were present
does not prove that they are independent of other parts in the intact or-
ganism.

Although spiral cleavage pattern is relatively stable in those groups in
which it appears, the occurrence of widely different cleavage patterns in
related groups seems not without significance, as indicating that cleavage
pattern and organismic pattern are not very closely related. Among the
turbellaria spiral cleavage appears only in polyclads. Cleavage of the acoe-
lous Polychoerus differs from the spiral type, according to Gardiner, and

CLEAVAGE AND DEVELOPMENTAL PATTERN 563

cleavage of rhabdocoels and triclads is apparently not spiral; also, trem-
atodes and cestodes seem not to have spiral cleavage. Rotifer cleavage
pattern is apparently spiral in early stages but later departs from spiral
pattern (Zelinka, 1891; Jennings, 1896). If the rotifers are related to an-
cestral forms of the "Trochelminthes" and if the spiral cleavage pattern
has ancestral significance, we might expect to find it in well-developed
form in the rotifers. Cleavage in cephalopods is meroblastic and widely
different from the spiral type (Vialleton, 1888; Watase, 1891). Spiral
cleavage may appear to be closely associated with early differentiation of
free-swimming larvae of trochophore type, but in oligochetes and leeches
and in fresh-water pelecypods there are no such larvae, but cleavage is
spiral. Also, the piHdium larva of nemerteans is very different from the
trochophore type, and some nemerteans develop without pihdium larvae;
but spiral cleavage is present in both.'"

Whatever conclusions we may draw concerning the phylogeny of groups
with spiral cleavage and their relatives without, it seems probable that we
must distinguish between the obhque, or so-called "spiral," type of cleav-
age, as determined by physiological factors not closely associated with or-
ganismic developmental pattern, and the relative sizes of cells and their
differences in different groups, as expressions of that pattern. The obhque
form of cleavage corresponds closely to a surface-tension pattern, but rela-
tive cell sizes suggest regional differences in cortical activity or differences
in the deeper cytoplasm.

CLEAVAGE AND RECONSTITUTION IN CTENOPHORE DEVELOPMENT

Pattern of cleavage and early development of ctenophores differs from
those of other groups, and present knowledge serves chiefly to make evi-
dent the lack of any real insight into its physiology." In the undivided
egg of Plenrohrachia a thick ectoplasm or cortical layer, uniformly dis-
tributed about the egg, stains yellow with neutral red and rose with Nile
blue sulphate, as if alkaline in reaction, in sharp distinction from the in-
terior, which stains red or blue. The egg of Beroe shows no such differen-
tial staining, but with dark-field illumination a clearly defined green ecto-
plasm or cortex is visible (Spek, 1926). Polar bodies form at what is con-
sidered to be the basal or vegetal pole, that is, the later oral pole; and the

" Spiral cleavage has been followed through the earlier stages in a fresh-water nemertean
Stichostemma, with direct development (Child, unpublished).

" Chun, 1880, 1892; Driesch und Morgan, 1895; Fischel, 1897, 1898, 1903; Ziegler, 1898,
1903; Rhumbler, 1899; Yatsu, 1911, 1912a, 6; Spek, 1926.

S64

PATTERNS AND PROBLEMS OF DEVELOPMENT

nucleus remains near this pole. First and second cleavages are meridional
and equal, the cleavage furrows gradually progressing from the basal pole
(Fig. 177, A). The two cleavage planes correspond, respectively, to the
stomodeal or esophageal and the tentacular plane of the adult. The third
cleavages are somewhat oblique in opposite directions on each side of the
esophageal plane, so that four slightly smaller macromeres lie somewhat
apical to four larger ones (Fig. 177, B). Each of the eight cells then gives

B

D

Fig. 177, A-D. — Early cleavages of ctenophore egg. A, the first cleavage furrow passing
through egg, nuclei near basal pole; B, eight-cell stage; C, sixteen-cell stage; D, later stage on
larger scale (after Ziegler, 1898).

rise to a micromere from its apical region (Fig. 177, C); these micromeres
divide, and other generations of micromeres form (Fig. 177, D), resulting
in four more or less separated groups of micromeres, corresponding to the
four quadrants. According to Spek's observations with dark-field illumi-
nation, the green ectoplasm of the Beroe egg aggregates about the basal
region before each of the first three cleavages, and the cleavage furrow be-
gins there. The green ectoplasm forms the head of the cleavage furrow and
a layer on each side as cleavage progresses apically and, following first and
second cleavages, is again uniformly distributed over the surfaces of the
cells. Following the third cleavage, however, the green ectoplasm is local-

CLEAVAGE AND DEVELOPMENTAL PATTERN 565

ized in the apical regions of the macromeres, and the micromeres are
formed from it. Spek holds that this fixed localization results from a
marked increase in cytoplasmic viscosity following the third cleavage. His
observations lead him to the conclusion that the egg is not a mosaic,
though different substances are undoubtedly present in it. It seems evi-
dent that this orderly series of events in early cleavages does not indicate
independence of parts. The mosaic condition is brought about, according
to Spek, dynamically during cleavage through the activities associated
with mitosis and the changes in viscosity. To what extent the ectoplasmic
changes result from actual migration of the ectoplasm, as distinguished
from regional separation of cytoplasm proper from yolk or other inclu-
sions, in consequence of altered physical, associated with metabohc, con-
ditions, is perhaps still a question. Is the ectoplasm associated with the
cleavage furrow, or is that finally localized in the apical regions of the mac-
romeres, necessarily the same ectoplasm that originally extended over the
egg surface?

The four groups of apical micromeres spread over the macromeres and
give rise to the apical nervous system and sense organ, the swimming-
plate rows, two rows from the cells of each group, general ectoderm, and
stomodeum. The eight macromeres divide into sixteen entoderm cells, and
each gives off a micromere basally which contains the last traces of the
green ectoplasm (Spek) but becomes entoderm. Later an entodermal
pouch develops in each quadrant and divides into two canals, each leading
to a plate row.

There is no directly visible evidence of polarity in the cytoplasm of the
undivided egg, but the relation of the growing oocyte to the parent body
(Chun, 1880) suggests the possible determination of polarity by the dif-
ferential resulting from this relation. On the other hand, it is possible that
position of nucleus near the egg surface is not definitely predetermined
but is a matter of chance, and that polar-body formation there constitutes
a step in development of polarity, the orderly behavior of ectoplasm in re-
lation to this region another step, and its apical localization the final stage
in establishment of polar pattern. That the bisymmetrical pattern is also
determined dynamically in relation to the ectoplasmic activities associ-
ated with first and second cleavages seems not improbable.

Cleavage and development of isolated parts of the ctenophore egg,
blastomeres, and blastomere groups have been studied by a number of in-
vestigators.'' In isolated parts of undivided eggs relation of plane of sec-

•- References given in footnote 11, p. 563.

566 PATTERNS AND PROBLEMS OF DEVELOPMENT

tion to polar axis is not certainly known or can be inferred only from de-
velopmental results, but in eggs beginning first cleavage and later stages
definite orientation of plane of section is possible. After removal of an
apical portion by section more or less transverse to the apicobasal axis, de-
velopment differs according to amount and region removed. Removal of
parts by oblique section basal to the equator usually results in larvae with
one or more plate rows defective or lacking; but after removal of parts by
section apical to the equator, normal larvae may develop. After lateral re-
moval by section parallel to the apicobasal axis one or more plate rows
are usually defective or absent. Defects are greater after formation of po-
lar bodies than before (Yatsu). From the experimental data Fischel con-
cluded that plate-row material is localized in a ring near the basal pole, but
Spek regards regional differentiation of the ectoplasm before or at begin-
ning of first cleavage as improbable and maintains that defective develop-
ment or absence of one or more plate rows results from removal of a cer-
tain amount, rather than a particular kind, of ectoplasm. Since the ecto-
plasm aggregates basally before the first cleavage, removal by section of
cytoplasm about this region will remove a relatively large amount of ecto-
plasm and so interfere later with development of some of the micromeres.
The defective development of one or more plate rows following removal of
peripheral cytoplasm basal to the equator seems to indicate that the ecto-
plasm does actually migrate apically, as Spek asserts, for the plate rows
develop about the apical region.

Cleavages of isolated blastomeres, as far as followed, occur always as if
the other cells were present. Each 1/2 blastomere gives rise to four plate
rows, each 1/4 blastomere to two, each 1/8 to one. According to Yatsu,
the end cells and the middle cells of the eight-cell state (Fig. 177, 5) ap-
parently differ in some way, for in groups of middle blastomeres the num-
ber of plate rows developing is often less than the number of middle cells
present ; but isolated single or paired end cells give rise, respectively, to
one or two rows. In groups including both middle and end cells the num-
ber of plate rows developing is sometimes greater than the number of 1/8
blastomeres present, suggesting some reconstitution. Groups of 1/16
blastomeres, including more of the micromeres than of the macromeres,
usually develop plate rows equal in number to the micromeres, but some-
times fewer (Fischel, Yatsu).

The ectoderm of these partial forms incloses the entoderm completely ;
but the side of the larva representing the original surface of contact with
other cells remains flattened, and the stomodeum invaginates from the

CLEAVAGE AND DEVELOPMENTAL PATTERN 567

basal region of this side. The number of entodermal pouches is very com-
monly greater than the number of macromeres present, that is, a 1/2
form has three, a 1/4 form two, the extra pouch being smaller than the
others. Driesch and Morgan regarded this as a reconstitution, but Fischel
holds that the extra pouch results from mechanical division of entoderm
by the invaginating stomodeum.

Displacement of micromeres into two separate groups results in larvae
with two apical organs and four plate rows about each; but, although such
displacement almost always involves more or less displacement of macro-
meres, the larvae are single as regards entoderm and stomodeum. Isola-
tion experiments agree in general in showing that plate-row-forming cells
or cell groups are so far determined at an early stage that isolation does
not usually alter their course of development. The most interesting point
is the usual absence of reconstitution of plate rows in the partial forms.
The adult ctenophore shows high capacity for reconstitution of plate rows
from general ectoderm, as well as of other parts;'-' consequently, the gen-
eral ectoderm must be potentially capable of giving rise to plate rows.
Perhaps developing plate rows exercise some degree of dominance over the
general ectoderm, and this may be adequate in the small larval forms to
inhibit reconstitution of other rows. Since the plate rows are undoubtedly
regions of more intense physiological activity than the general ectoderm,
they probably influence it in some way. With increase in size of the indi-
vidual this dominance may not extend over the whole ectoderm between
rows, and reconstitution of other rows becomes possible. The fact that
partial forms sometimes develop more, sometimes fewer, plate rows than
expected from the number of micromere groups present is of interest in this
connection, as suggesting such relations. Also of interest are the observa-
tions of Chun (1892, 1895) on 1/2 larvae of Bolina found after a storm
which had presumably isolated blastomeres. These forms had developed
to larvae as 1/2 forms, but on metamorphosis they reconstituted to
wholes.

Except as regards the plate rows, the forms resulting from isolation and
displacement of blastomeres show considerable reconstitution. Duplica-
tion of apical organs and arrangement of plate rows about the duplicated
organs involves change in position and direction of some of the plate rows,
if not actual reconstitution. In 1/2, 1/4, and 1/8 forms ectoderm over-
grows the entoderm completely ; that this is entirely a mechanical matter
seems doubtful. Also, the stomodeum in such forms is a whole, not a par-

'^ Mortensen, 1913; Coonfield, 1936a, h, 1937a, b.

S68 PATTERNS AND PROBLEMS OF DEVELOPMENT

tial, stomodeum. That the extra entodermal pouch is always the result of
mechanical division of the entoderm by the stomodeum, rather than a true
reconstitution, may be questioned. And, finally, reconstitution in early
stages may be Hmited by viscosity of the cytoplasm. The fact that forms
developing from isolated blastomeres remain flattened on the side former-
ly in contact with other cells indicates a considerable physical rigidity.
Spek finds a marked increase in viscosity at the third cleavage. Physical
rigidity, as well as regional specificity, may limit reconstitution.

Evidently the undivided ctenophore egg is not a mosaic but a dynamic
system in which definite orderly and related changes take place. From the
data at hand it appears possible that before polar-body formation the
cytoplasm has no definite organization except a surface-interior difference
and that axiate pattern develops gradually. The ectoplasmic activities of
first and second cleavage may be factors in determining the biradial pat-
tern. In short, it seems possible that the polar-biradial pattern of the cten-
ophore can originate without any further "organization" than the surface-
interior, ecto-entoplasmic pattern, the eccentric position of the nucleus,
and the ectoplasmic activities during the first and second cleavages. Since
the plate row of the adult and the general ectoderm show a very definite
apicobasal gradient (p. io6) and the apical region is dominant in recon-
stitution (Coonfield, 1936a), gradient pattern must develop in the de-
scendants of the micromeres at some stage.

CLEAVAGE PATTERN AND DEVELOPMENT OF AscaHs

Cleavage pattern, cell hneage, and the "germ path" of Ascaris megalo-
cephala {^equorum) have been repeatedly and extensively studied.'^ The
germ path, that is, the persistence of a small number of large chromosomes
in the blastomere fine from which the germ cells develop, and the diminu-
tion of chromatin by exclusion from the nucleus of the terminal portions
of the large chromosomes and appearance of a larger number of small chro-
mosomes in other cells, has been regarded as a feature of particular inter-
est (Fig. 178). Essentially similar cleavage patterns have been observed
in various other nematodes, but chromatin diminution apparently does
not occur in all.

The egg and embryonic stages of Ascaris, inclosed, as they are, in a
hard thick shell, are less directly accessible to some experimental pro-
cedures than those of many other forms; but blastomeres have been killed

'^ Boveri, 1887, 1890, 1892, 1899, 19106; C. C. Schneider, 1891; Herla, 1893; Zur Strassen,
1896; Zoja, 1896; H. Miiller, 1903; Bonfig, 1925; et al.

S70

PATTERNS AND PROBLEMS OF DEVELOPMENT

by localized radiation, and pressure and centrifuging have been employed
in experimental analysis. The ovarian oocyte is more or less pear-shaped
and attached to the rhachis of the gonad by its slender end. Supposedly

D E

Fig. 179, A-E. — Early cleavages of Ascaris megalocephala. A, two-cell stage, spindles ot
second cleavages indicated; B, four-cell stage; C, the change in cell position to form the
"rhombus"; D, four-cell stage after change in position, the rhombus; E, seven cells, showing
asymmetry in position of the four dorsal cells. In the figures dorsal is above; anterior, right
(after Boveri, igioh).

the blunt free end becomes the apical pole. Normally the egg becomes
spherical before polar-body formation; but in eggs which retained the
elongated form until fertilization and shell formation, polar bodies were

CLEAVAGE AND DEVELOPMENTAL PATTERN 571

observed to form at the blunt pole in about three-fourths, elsewhere in
others. Most of these eggs developed abnormally, but in 91 per cent of the
normal cleavage stages polar bodies were at the blunt pole (Schleip, 1924).
After fertilization and maturation the yolk becomes more densely aggre-
gated basally, so that an apicobasal gradient is visible in the cytoplasm.

The cleavage pattern of Ascaris and of various other nematodes studied
differs from other known patterns. The first cleavage is transverse to the
apicobasal axis and is practically equatorial (Fig. 179, A), the apical cell,
AB, being regarded as dorsal, the basal cell, Pi, as ventral ; the two cells are
different, Pi containing more yolk. The second cleavage is meridional in
relation to the egg axis and is accompanied by diminution of chromatin in
the dorsal AB, transverse and without diminution of chromatin in the
ventral Pi (Fig. 179, B). A change in position of blastomeres follows this
cleavage (Fig. 179, C), resulting in a "rhombus" form (Fig. 179, D). Up
to this stage it has not been possible to distinguish future anterior and pos-
terior ends; but in the rhombus P. and the cell with which it comes into
contact {B) are posterior, the other two cells anterior, and the nuclei of
all four cells are in the median plane. According to Boveri, A and B give
rise to general ectoderm, largely of the anterior and dorsal regions; EMS
forms entoderm, mesoderm, and stomodeum; and P^, ectoderm, meso-
derm, and germ cells. Evidently there is no segregation of embryonic
layers except entoderm in particular cells. At the next cleavage A and B
divide equally in the median plane; but the four resulting cells shift to
somewhat oblique positions, so that in lateral view the right anterior cell
of the four is higher than the left anterior, producing a slight asymmetry
(Fig. 179, £). The cell EMS divides into posterior (entoderm) and ante-
rior (mesoderm and stomodeum) cells, both of which undergo chromatin
diminution. The spindle of this division is indicated in Figure 179, E.
P2 also divides into C and P3 without diminution, but at the next division
C undergoes diminution, giving rise later to ectoderm.

With change in position of cells of the four-cell stage in a direction at
right angle to the usual direction, a tetrahedral form (Fig. 180, ^), instead
of a rhombus, results. Intermediate directions of the shift occur, but
finally result in either rhombus or tetrahedral form. The tetrahedral form
supposedly results from injurious or inhibiting conditions; high tempera-
ture increases its frequency (Bonfig, 1925). In this form the two dorsal
cells divide transverse to the axis of the two ventral cells, that is, at right
angles to their divisions in the rhombus form ; and the cell on the left side
and anterior with respect to the ventral cells is higher in lateral view (Fig.

572 PATTERNS AND PROBLEMS OF DEVELOPMENT

1 80, B), instead of the right anterior cell, as in rhombus development (Fig.
179, E), a reversal of the asymmetry of the dorsal cell group. The adult
Ascaris is asymmetrical in that the nucleus of the unicellular excretory or-
gan is usually on the left side, but in one individual to about thirty or forty
it is on the right, and the proportion of reversed asymmetry in cleavage is
about the same; consequently, it is concluded that the asymmetry of
cleavage determines that of the adult (Zur Strassen, 1896).

In spite of the numerous studies on Ascaris development, the relation
of egg and cleavage pattern to axiate organismic pattern is still obscure.
Apparently egg polarity does not coincide with the longitudinal axis of the
animal. Granting this, how is the anteroposterior axis determined? The
two dorsal cells, A and B, of the four-cell stage are not distinguishable as

A B

Fig. 180. — A, tetrahedral four-cell stage; B, reversed asjonmetry of dorsal cells resulting
from tetrahedral form (after Bonfig, 1925).

anterior and posterior until the change in position results in the rhombus
form. Then the cell in contact with P^ is distinguishable as posterior (B of
Fig. 179, D) from A, the anterior cell. But the relative change in position
of dorsal and ventral cells is not always in the same direction ; it may be at
right angles to the usual direction, producing the tetrahedral form, and
the two dorsal cells are right and left, instead of anterior and posterior, in
relation to the ventral cells and apparently also in relation to the antero-
posterior axis of the animal. Is there normally a physiological difference
between the two dorsal cells which determines the direction of shift in
position in the four-cell stage but which is more or less completely obliter-
ated in shghtly inhibiting conditions? If this is the case, how are anterior
and posterior end determined in the tetrahedral forms? Or are the two
dorsal cells alike, and does the shift occur indifferently in either direction
to form the rhombus and at right angles to these directions to form the
tetrahedron? If this is the case, the ventral cells apparently determine the
anteroposterior axis.

CLEAVAGE AND DEVELOPMENTAL PATTERN 573

Although definite answers to these questions seem to be lacking, certain
features of Ascaris development are suggestive. There is an apicobasal
gradient, indicated by differential aggregation of yolk in the undivided
egg. From his study of dispermic eggs Boveri (igiob) concluded that the
polarity of the egg is represented by a cytoplasmic gradient and that nu-
clei coming to He in the apical levels of this gradient undergo diminution,
while those in the basal region do not. The behavior of dorsal and ventral
cells in normal cleavage suggests that the gradient differential is greater
in the ventral (basal) than in the dorsal (apical) half of the egg. The dor-
sal cell, AB, of the two-cell stage undergoes diminution as it divides into
A and B; the ventral cell does not. However, at the next division the cell,
E, evidently the more apical part of the basal half, undergoes diminution ;
but the cell, P2 representing the extreme basal region, does not. Evidently
there is sufficient gradient in the ventral (basal) half of the egg to deter-
mine this difference in nuclear behavior. Moreover, the dorsal cells, A
and B, form only part of the general ectoderm, while most of the organs
develop from the two ventral cells. The apical half of the egg probably
represents the higher levels of the apicobasal gradient, but with little axial
differential, most of the differential being in the basal half. The apico-
basal axis of the sea-urchin egg is apparently somewhat similar. If these
suggestions are correct, the ventral cells of the four-cell stage may deter-
mine the longitudinal axis.

In formation of the rhombus some unknown factor determines the
change in position of the cells, at least as regards plane of the shift. In
tetrahedral forms this factor has apparently been somewhat less effective.
The factors determining the asymmetry in position of the four dorsal cells
from the six-cell stage (Fig. 179, E), its reversal in tetrahedral forms (Fig.
180, B), and the postulated relation between this asymmetry and that of
the adult, which concerns a single cell, are also completely unknown.

Cleavage and development of blastomeres of two-cell and four-cell
stages, after killing or preventing development of others by localized radi-
ation, are stated always to be essentially as if the other cell or cells were
also developing (Stevens, 1909). The presence of the dead or inhibited
cells apparently does not determine the partial character of development,
since contact between them and the hving cells is often shght. There is no
conclusive evidence of reconstitution.

Eggs kept for several months without exposure to air show various evi-
dences of injury in their development on exposure to air. In some all
chromosomes pass into one cell at the first cleavage, diminution may take
place at the next division, and a blastula-hke cell mass is formed, as with

574 PATTERNS AND PROBLEMS OF DEVELOPMENT

an isolated AB-ccW, or the cell containing all chromosomes may develop
like a whole egg (Faure-Fremiet, 1913). It seems probable that these dif-
ferences in development result from alterations in the apicobasal gradient
pattern by the inhibiting conditions and the following recovery. Eggs
allowed to develop by exposure to air after some 3 months in CO2 give
forms ranging from normal to unorganized cell masses, but a third of them
show various degrees of what appears to be a differential inhibition, more
or less of the anterior region being present merely as an unorganized cell
mass and posterior regions normally developed (Painter, 19 15).

Sometimes a considerable part of the cytoplasm, usually containing the
nucleus of the second polar body, constricts off from the apical region but
does not divide. The remaining part of the egg divides into a dorsal and
a ventral cell, and these change their position with respect to this Neben-
zelle in the same way as the two ventral cells of the normal four-cell stage.
The following cleavages may be normal or more or less altered by presence
of the Nehenzelle; but, with a single exception, normal animals have not
been seen to develop from these eggs (Kautsch, 191 2).

Although an apicobasal gradient pattern is indicated in the fertilized
Ascaris egg, it does not necessarily follow that regional cytoplasmic dif-
ferences are merely quantitative, but conclusive evidence for regional
specificity at the beginning of development is lacking. The undivided egg
appears not to be a mosaic, and the change in position of cells at the four-
cell stage suggests physiological relations of some sort between the blasto-
meres. As regards cleavage pattern, regional characteristics are apparent-
ly determined early, but independent differentiation of isolated blasto-
meres does not proceed very far. If the ventral cells of the four-cell stage
do determine the anteroposterior axis, they perhaps induce some degree
of anteroposterior differential in the dorsal cells. It is evident, however,
that the cleavage pattern of Ascaris gives us little definite information con-
cerning developmental pattern.

ENTOMOSTRACAN CLEAVAGE PATTERN

The normal cleavage pattern of Cyclops is determinate, at least as re-
gards mesoderm, entoderm, and primitive germ cells. The germ path is
indicated by certain cytoplasmic granules with a definite cell lineage from
the first cleavage to the primitive germ cells. The plane of first cleavage is
said to be in the shortest diameter of eggs deformed by pressure in the egg
sac; and, according to Fuchs (19 14) and M. Jacobs (1925), the cleavage
axis has a definite and constant relation to developmental pattern and be-

CLEAVAGE AND DEVELOPMENTAL PATTERN 575

comes the dorsiventral axis of the embryo but is independent of the axis
indicated by polar-body position. If these observations are correct, or-
ganismic pattern appears to be determined by cleavage pattern or by the
same factor that determines cleavage pattern ; and if a polarity is deter-
mined in the ovary or by position of polar body-formation, it may be ob-
literated by the polarity of the cleavage pattern, which is, or may be, de-
termined by pressure. Development of centrifuged eggs seems to support
this conclusion (p. 586). There is, however, no evidence of a polar differen-
tial in earlier cleavage. The first two cleavages divide the egg into four al-
most equal quadrants, with polar cross-furrows, as in spiral cleavages, de-
termining a cleavage axis (Fig. 181, A), which supposedly becomes the

ABC

Fig. 181, yl-C— Cleavage stages of Cyc/oi'5. A, four-cell stage; B, sixteen-cell stage; C, later
stage, showing entoderm cell, E, and primitive germ cell, G, which are formed by division of
Z)'"^ of sixteen-cell stage shown in B; heavy line in C indicates boundaries of quadrants of
four-cell stage (after Fuchs, 19 14).

dorsiventral axis of the animal, but dorsal and ventral are distinguishable
only in later stages. The first cleavage plane is said to coincide approxi-
mately with the median plane of the embryo. The third cleavage is equa-
torial with respect to the cleavage axis determined by the first two cleav-
ages and is approximately equal; the fourth is meridional or apparently
sHghtly oblique (Fig. 181, 5). Entoderm develops from a single cell of the
fifth cleavage generation, and its sister cell is the primitive germ cell; both
lie at the ventral pole and approximately in the future median plane {E,
G, of Fig. 181, C); and cells immediately surrounding them give rise to
mesoderm, which is invaginated with entoderm and germ cell. This cell
lineage gives no information as to origin or character of organismic de-
velopmental pattern. If it is true that cleavage pattern is determined by
shape of the egg and that the cleavage axis becomes the dorsiventral axis,
it apparently follows that axiate organismic pattern is independent of any
pre-existing pattern or egg organization and originates from purely for-

576 PATTERNS AND PROBLEMS OF DEVELOPMENT

tuitous external conditions. But how cleavage pattern is determined in a
spherical undeformed egg does not appear. If the cleavage axis is the dor-
siventral axis of the animal, the origin of the anteroposterior axis remains
obscure. The only evidence of a differential in the accounts of cleavage is
the somewhat less rapid division of the quadrant from which entoderm and
primitive germ cell originate. When blastomeres are killed or inhibited by
localized ultra-violet radiation but not removed, development of remain-
ing blastomeres is not altered, except for more or less overgrowth of the
undivided cell or cells by ectoderm (M. Jacobs, 1925). These experiments
give further evidence concerning relation of embryonic axes to cleavage
pattern, confirming and extending that from normal development. In
Cyclops and also in Polyphemus the germ path may be either in a right or in
a left quadrant, that is, asymmetry of cleavage pattern in this respect may
be either dextral or sinistral.'^

In eggs of various other entomostracans studied a cytoplasmic polarity
is distinguishable, either before or after maturation, and becomes the axis
of differentiation ; or this axis coincides with the axis indicated by polar-
body position."* In Polyphemus, however, first and second cleavages are
obHque to the egg axis, but this axis becomes the axis of differentiation
(Kiihn, 1912); but in Lepas, according to Bigelow (1902), there is a shift
of the egg axis in relation to the oblique first cleavage. According to these
data, there are considerable differences in cleavage pattern and in its re-
lations to egg pattern and to pattern of differentiation among the ento-
mostraca; some of the conclusions suggest that further investigation is
highly desirable.

ASCIDIAN CLEAVAGE AND DEVELOPMENTAL PATTERN

Cleavage pattern and developmental pattern of ascidians are apparent-
ly closely related. In the freshly extruded egg of Styela {^Cynthia) po-
larity is indicated only by position of the nucleus close to the egg surface,
the region in which it lies becoming the apical pole. Following extrusion
the nucleus breaks down, an apical area of clear cytoplasm appears, and
the first polar spindle forms in it. The egg also shows a surface-interior
pattern, an ectoplasm containing yellow pigment, and an entoplasm with
yolk (Conklin, 1905a). Whether the nuclear position near the surface is
determined by a pre-existing cytoplasmic polarity or is itself the determin-
es Kuhn, 1912; Fuchs, 1914; M. Jacobs, 1925.

"■Grobben, 1879, 1881; Samassa, 1893; McClendon, 1906, 1907; Muller-Cale, 1913;
Kruger, 1922.

CLEAVAGE AND DEVELOPMENTAL PATTERN 577

ing factor of polarity is apparently not known. Sperm entrance in a region
about the basal pole suggests presence of a cytoplasmic pattern (cf . Dalcq,
1932c, 1935). Evidence of cytoplasmic dorsiventrality or bilaterality in
the unfertilized ascidian egg has been presented by van Beneden and Julin
(1884) and by Dalcq and Vandebroek (1937).

Following sperm entrance, an extensive and remarkable series of cyto-
plasmic movements takes place in Styela (Conklin, 1905a). The yellow
ectoplasm first streams basipetally and aggregates in the region about the
sperm nucleus. Dalcq and Vandebroek (1937) find that local vital staining
indicates movement away from the apical pole of the egg cortex about
this pole in Ascidiella (see also Vandebroek, igs6b). The sperm nucleus
migrates to a position near the egg equator, but on one side of the egg
axis; and the yellow cytoplasm goes with it and finally forms a yellow
crescent transverse to the polar axis and symmetrical to the position of
the sperm nucleus (Fig. 182, A).''' This crescent, designated as "meso-
plasm" or "myoplasm," gives rise to muscle tissue. The indophenol
blue, benzidin, and leucomethylene blue tests for oxidases and perocidases
show no localization in the unfertilized egg, but, after fertilization, locah-
zation in the myoplasm (Ries, 1939; Reverberi e Pitotti, 1939). The path
of the sperm nucleus and formation of the crescent suggest a dorsiventral
pattern as a determining factor. Six organ-forming regions, indicated
by differences in the cytoplasm, are distinguished by Conklin.'^

Cleavage is, to a high degree, determinate. The plane of the equal first
cleavage coincides with the median plane and divides the yellow crescent
equally ; the second cleavage is vertical to it and also meridional ; the third
separates smaller apical from larger basal cells and is slightly oblique
bilaterally (Fig. 182, B). Later cleavages, so far as followed, are strictly
bilateral, and the cell lineage of organs and tissues is definite (Fig. 182,
C, D, E)."^

The study by Chabry of development of isolated blastomeres and blas-
tomere groups showed that cleavage and development were in all cases
partial and gave little or no evidence of reconstitution, except that ecto-
derm formed a complete surface layer. Driesch (1895) maintained, how-

'7 The side of the egg on which the crescent Hes is regarded by Conkhn as posterior, but
comparison with amphibian development suggests that it is perhaps more nearly ventral or
posteroventral, since neural plate and notochord develop on the opposite side.

•* Conklin, 1905(7, 191 1. See also Duesberg, 1913.

'9 For earlier studies of ascidian cleavage see van Beneden et Julin, 1884; Chabry, 1887;
W. E. Castle, 1892.

578 PATTERNS AND PROBLEMS OF DEVELOPMENT

ever, that isolated 1/2 blastomeres gave rise to complete or only slightly
defective larvae. Cleavage of isolated blastomeres as halves was recog-

D E

Fig. 182, ^-£.— Cleavage stages of Siyela partita. A , fertilized, undivided egg, lateral view;
yellow crescent indicated by broken line; clear protoplasm at its borders, by dotted line. B,
eight-cell stage in lateral view; yellow crescent indicated. C, apical, D, basal, view of sLxteen-
cell stage; boundary of yellow crescent indicated. E, early gastrula; neural plate cells indi-
cated by dotting; notochord, by vertical lines; mesenchyme, by oblique lines; muscles, by hori-
zontal lines; partially invaginated entoderm and general ectoderm, white (after Conklin, 19050).

nized by Crampton (1897), but he also beheved that whole larvae resulted.
Chabry was later confirmed by Conklin,"" who followed the cell lineage of

" Conklin, 1905c, 1906, 191 1 ; experiments on Styela partita, Ciona iniestinalis, Molgiila man-
hattensis, and Phallusia mammillaia.

CLEAVAGE AND DEVELOPMENTAL PATTERN 579

partial forms and concluded that an isolated blastomere gives rise only
to those parts which develop from it in intact embryos, that the cyto-
plasmic regions visibly distinguishable represent organ-forming sub-
stances, and that ascidian development is a mosaic. However, both Conk-
lin and Chabry found that the sensory primordium might appear in right
or left partial forms, and Conklin regarded this as a regulation. Develop-
ment of one to three papillae in half-larvae, three being the number in
wholes, was noted by G. A. Schmidt (193 1) and regarded as indicating
regulation. Development of isolated blastomeres was also studied by Ber-
rill (1932), who agrees with Conklin; but later it was found that in half-
larvae of Ascidiella papillae vary in number from none to three (Cohen
and Berrill, 1936). These authors admit the possibility of papillar recon-
stitution but regard it as more probable that a papillar primordium is
variously divided or brought entirely into the right or left blastomere by
slight variations in plane of first cleavage. They also consider it probable
that presumptive neural tissue may become general ectoderm and take
part in the covering of the half -larvae; and they find, as did Chabry and
Conklin, that a sensory primordium may appear in right or left half-
larvae, but offer various suggestions to account for this in terms of mosaic
development rather than reconstitution.

Pieces of the unfertilized egg of Ascidiella can be fertilized and develop
(Dalcq, 1932a, b, c). Pieces resulting from meridional section may de-
velop half-embryos, complete symmetrical forms, or forms with more or
less complementary defects in the two halves. Dalcq, however, does not
regard even development of a whole larva from a half-egg as "true regula-
tion" but suggests that locahzation of cytoplasmic substances in the un-
fertilized egg gives it a bilaterally symmetrical structure and that merid-
ional section in the median plane divides these substances symmetrically,
so that all are present in each half and a whole larva develops from each.
However, what would normally give rise to half of the symmetrical pat-
tern gives rise in the half-eggs divided in the median plane to the whole
symmetry pattern. To maintain that reconstitution does not take place
in such cases seems scarcely in accord with the facts. Dalcq also finds
that small portions may be removed from apical or basal polar regions
without preventing normal development and that gastrulation occurs in
pieces down to one-fifteenth the egg volume; but in apical pieces there is
excess of ectoderm, in basal pieces ectodermal deficiency.

These results indicate a considerable capacity for reconstitution in the
unfertiHzed ascidian egg. More recently Dalcq (1935, 1938^) has con-
cluded that it is not possible to establish for ascidians a topographic locah-

58o PATTERNS AND PROBLEMS OF DEVELOPMENT

zation of hypothetical cytoplasmic substances agreeing with the data of
observation, that the occurrence of regulation in the ascidian egg must be
admitted, and that ascidian development, Hke that of other forms, must
be interpreted in dynamic terms. He regards the organization of the ascid-
ian egg as consisting of a polar gradient, the mesoplasm (Conklin's yellow
crescent), and a cortical, dorsiventral field with greatest concentration in
the region of the equator. These are regarded as dynamic factors and
morphogenesis is conceived as resulting from their interaction.

Eggs from which extrusion of cytoplasm has been brought about by
puncture of the chorion and pressure may give rise to more or less normal
larvae (Reverberi, 1931), but what region of the cytoplasm is removed
in these experiments is not known. Isolation of the four apical and the
four basal blastomeres of the eight-cell stage results in alteration of later
cleavage in the former but not in the latter, and the conclusion is drawn
that the factors controlhng bilaterality are locahzed in the basal region
(Reverberi, 1933). Pieces of the unfertilized egg of Ciona, when fertihzed,
cleave like the whole egg, irrespective of the region which they represent
(Reverberi, 1936). Pieces of fertilized eggs with known orientation of
plane of section give chfferent results. Pieces above a certain size, obtained
by section parallel to the polar axis, cleave like whole eggs. Pieces ob-
tained by section vertical to the polar axis cleave like wholes if they con-
tain a zone localized in three-fourths of the basal hemisphere ; but pieces
of the apical hemisphere, or of this plus the adjoining basal fourth, cleave
radially. These experiments also indicate that factors determining bilat-
erality are localized in the basal region (Reverberi, 1937).

These isolation experiments show that developmental potencies of at
least some regions of ascidian eggs are less strictly limited than most ear-
lier investigators believed. Experiments on translocation and combination
of blastomeres of Ascidiella aspersa provide further evidence that this is
the case (Tung, 1934). These experiments include rotations of the four
apical blastomeres (micromeres) of the eight-cell stage 90°, 160°, and 180°
on the four basal blastomeres, superposition of two groups of micromeres
with polar axes unchanged but with various degrees of rotation about this
axis, superposition of two two-cell stages with median planes at right
angles, and various isolation experiments. The only parts which appear
truly stable in Tung's experiments are the primordia of notochord, myo-
blasts, and probably of mesenchyme. Presumptive ectoderm and ento-
derm are "relatively equipotential." The four apical micromeres, nor-
mally ectodermal, can develop as entoderm; and the basal macromeres,

CLEAVAGE AND DEVELOPMENTAL PATTERN 581

which normally produce little ectoderm, give rise to forms completely
covered with ectoderm. Level of gastrulation is not predetermined; and
Tung suggests a double polar gradient, animal and vegetal, such as
Runnstrom has postulated for the sea urchin, and a quantitative deter-
mination of ectoderm and entoderm in relation to gradient-level. Here,
as in the sea urchin, the vegetal gradient may be secondary, at least as
far as dynamic factors are concerned.

According to Tung, the micromeres give rise to cerebral vesicle, and
macromeres to neural cord in many cases; and presence of notochord is
not necessary for formation of a cerebral vesicle, but the possibility of
induction by entoblast is not excluded. Presumptive neural tissue may
develop as general ectoderm, and supernumerary cerebral vesicles may
develop. Presence, absence, and variable numbers of pigmented cells in-
dicate that they, too, are not fixedly predetermined. The adhesive papil-
lae also are apparently determined in relation to other parts.

Results of fusions, killing, and isolation of blastomeres and blastomere
groups have led von Ubisch to conclude that the general organ-forming
regions of the embryo, except ectodermal, are determined at the sixteen-
cell stage. Presumptive neural ectoderm may become epithelium, and
presumptive epithelium may become entoderm. However, the embryonic
regions are not sharply separated from each other but are gradients which
merge into each other {ineinander iihergehen) at their boundaries; these
boundary regions may develop into organs of either of the gradient systems
concerned. Within the general organ-forming regions determination is
not fixed at the sixteen-cell stage, but extensive reconstitutions {Regiila-
tionen) are possible.^'

In short, it seems to be evident that ascidian development, often re-
garded as the most extreme case of mosaic development, is actually far
from a mosaic in early stages. As in various other eggs and embryos, cer-
tain parts are more stable than others and appear thus far to be definitely
restricted in potency; but the possibiUty remains that with other experi-
mental alterations of environmental relations of these parts other poten-
cies may be brought to light. Moreover, it appears that neither deter-
minate cleavage nor partial development of isolated parts provides an
adequate basis for the conclusion that ascidian development is, from the
beginning, a mosaic of independent parts. As development progresses,

2' Von Ubisch, 19386, 1940, also 1939, "tjber die Entwicklung Ascidienlarven nach friih-
zeitiger Entfernung der einzelnen organbildenden Keimbezirke," Arch. Entiv'mcch., 139; and
alaterpaper, 1940, "Regulation und Determination im Ascidienkeim," ibid., 140. See also the
evidence for presence of an inductor region in ascidians on p. 480.

582 PATTERNS AND PROBLEMS OF DEVELOPMENT

the mosaic condition is more nearly approached. Removal of various parts
of the embryo preceding gastrulation affords no certain evidence of re-
constitution, except possibly of entoderm, or of induction (von Ubisch,
1940).

CLEAVAGE AND PATTERN IN AmpMoXUS

Development of Amphioxus shows similarities in general regional pat-
tern to ascidian and amphibian development. The apical hemisphere is
chiefly ectodermal, the basal region entodermal; a mesodermal crescent
lies posteroventrally, and opposite this is the area from which notochord
and neural plate develop." In the separation of mesodermal and chordal
areas this pattern resembles the ascidian and differs from the amphibian
pattern. Evidence of dorsiventrality appears in the undivided egg, and
the first cleavage plane usually coincides with the median plane; but
cleavage is, in general, less distinctly determinate than in ascicUans and,
except for differences resulting from the smaller amount of yolk, gastrula-
tion resembles amphibian gastrulation.

Development of isolated blastomeres led E. B. Wilson (1893) to the
conclusion that Amphioxus development is not mosaic. Conkhn (1933)
agrees with Wilson that whole normal larvae may develop from 1/2
blastomeres but maintains that this is possible only if the blastomeres are
right and left halves. Partly separated blastomeres of the two-cell stage
show various degrees of twinning, the two partially united forms appar-
ently developing independently of each other, often with polar axes in
different directions in consequence of blastomere rotation. Partial separa-
tions at the four-cell stage develop into three or four blastulae or gastrulae,
but further development is rare and abnormal. Conkhn regards Amphi-
oxus development as essentially mosaic, except for reconstitution of 1/2
blastomeres into whole individuals. However, in the hght of the ascidian
experiments, it seems possible that further experiment may lead to some-
what different conclusions: an inductor region may be concerned in Am-
phioxus development, and transplantations and combinations of blasto-
meres may bring to light relations of parts not shown by isolations. »

CLEAVAGE AND DEVELOPMENTAL PATTERN IN OTHER FORMS

In many animals cleavage pattern is less definitely determinate than
the patterns discussed or apparently entirely or almost entirely indeter-
minate. In forms with total cleavage (holoblastic) the planes of early divi-

" Conklin, 1932. References to earlier literature given there.

CLEAVAGE AND DEVELOPMENTAL PATTERN 583

sions are usually definitely related to the polar axis, those of the first two
cleavages usually passing through this axis but often without definite re-
lation to a median plane. In most animals with meroblastic cleavage
little or no constant relation between cleavage pattern and pattern of
development has been discovered. Evidently developmental pattern may
be entirely independent of particular cells. Even in sea-urchin develop-
ment, although ectoderm, entoderm, and mesenchyme develop normally
from particular cells or regions, any one of the three parts can develop as
either of the others in the earlier stages. Normally the cleavage appears
to be, at least to a considerable degree, determinate; but nothing is defin-
itively determined in early stages. Ectodermization of prospective ento-
derm, entodermization of prospective ectoderm, mesenchyme formation
from prospective ectoderm or entoderm, and development of prospective
mesenchyme as ectoderm or entoderm, all occur under experimental con-
ditions. Dorsiventrality and polarity may be altered and even obliterated,
although the cleavage pattern of earher stages may remain entirely un-
changed. Evidently there is no necessary relation between cleavage pat-
tern and developmental pattern here. In reconstitutions in adult multi-
cellular forms the cell is evidently not an essential factor in developmental
pattern ; the pattern is supercellular. It is becoming increasingly evident
that, even in the so-called "mosaic forms," the relation of cleavage pat-
tern and the cell to developmental pattern is often much less definite than
has been assumed. In fact, even these forms are evidence of "the in-
adequacy of the cell theory of development" (Whitman, Woods Hole
Biol. Led., 1893). Most botanists are agreed that developmental pat-
tern in plants is not primarily a mosaic of cellular units of organization
and function. Years ago De Bary said: "Die Pflanze bildet Zellen, nicht
die Zelle bildet Pflanzen." The progress of botanical research has led
many others to essentially similar conclusions. A recent expression
of the view that plant organization is determined by the organ or organ-
ism, not by the cell, is the summary of studies in size and form of ovaries
and fruits of Cucurbitaceae in the symposium paper by E. W. Sinnott
(1939, "The cell-organ relationship in plant organization," Growth,
Suppl.).

EFFECTS OF CENTRIFUGAL FORCE ON CLEAVAGE AND
DEVELOPMENTAL PATTERN

Visible granules regionally localized in eggs and regional differences in
appearance of cytoplasm have often been regarded as representing forma-
tive substances, but centrifuge experiments have shown that in most

S84 PATTERNS AND PROBLEMS OF DEVELOPMENT

cases developmental pattern is independent of the stratified distribution
of these substances by centrifugal force. Stratification into three, four, or
five more or less distinct layers or zones of different specific gravity and
often of very different appearance results from centrifuging, according to
the material and intensity of force. On cessation of centrifuging a gradual
reconstitution of normal cytoplasmic structure usually tends to occur,
except after very high intensities. In consequence of displacement of nu-
cleus or spindle, one or both polar bodies may form elsewhere than nor-
mally, and cleavage pattern may be altered.

In the Crepidula egg, centrifuged at first polar spindle stages, the spindle
may be displaced as a whole or greatly elongated ; and when division oc-
curs, the cell which would normally be the polar body may be as large as,
or even larger than, the other cell. One or both polar bodies may be en-
larged in this way, but even with abnormal size and position of polar
bodies cleavage gives evidence of persistence of pattern, presumably the
original pattern (Conklin, 1917).

In some eggs containing much yolk, oriented with polar axis in the
direction of the force, apicobasal differences in rate of cleavage may be
increased or reversed with aggregation of yolk at one pole or the other.
Amphibian eggs centrifuged with heavier vegetal pole centrifugal may
approach or attain meroblastic cleavage because the dense aggregation
of yolk prevents cleavage in the centrifugal region. When centrifuged in
reverse orientation, cleavage may become more rapid in the basal than
in the apical region.^-' Such effects are primarily mechanical, and the ag-
gregation of yolk may make normal development mechanically impossible ;
but this gives no information concerning physiological pattern. However,
even with marked stratification in the amphibian egg, the blastopore is
normally localized, and development may be essentially normal (Morgan,
1906c). Apparently the region of the dorsal lip has not been displaced.

Many eggs do not become axially oriented to centrifugal force, or orien-
tation can be prevented; consequently, stratification may be at any angle
to the axiate pattern, but a normal developmental pattern may appear
even with a distribution of cytoplasmic components entirely different
from the normal and different in different individuals.

Developmental pattern in the few coelenterate eggs centrifuged is ap-
parently quite independent of degrees of stratification or alterations of
cleavage resulting (Conklin, 1908; Beckwith, 1914). Complete or almost
complete independence of developmental pattern and stratification is par-

2^0. Hertwig, 18986, 1904; Wetzel, 1904.

CLEAVAGE AND DEVELOPMENTAL PATTERN 585

ticularly clear in eggs of certain annelids and mollusks, in which both a
normal cleavage pattern and normal larval differentiation may be quite
independent of the axis of stratification.'" Eggs of the gasteropod Ilyanas-
sa, centrifuged in reverse orientation, form the temporary polar lobe, al-
though, except presumably for the cortex, the cytoplasmic contents of the
lobe are quite different from its normal contents. In eggs separated into
two parts by centrifuging, the lobe forms in the basal part, and rhythmic
changes occur, even though no nucleus is present. Lobe formation and
activity are apparently associated with conditions in the cortex, which is
not displaced in centrifuging (Morgan, 1935, 1936). With sufficient centri-
fuging, eggs of the gasteropod Physa become greatly elongated and sepa-
rate into pieces corresponding more or less closely to the stratification.
Nucleated pieces, consisting only of clear protoplasm, may develop to
young normal snails and hatch (Clement, 1938).

Effects of centrifuging on developmental pattern are apparently greater
in eggs of the oligochete Tubifex and the leech Clepsine. The polar plasms
(p. 550) become partly or wholly mixed with other cytoplasmic com-
ponents, and development is not normal. '^ These eggs orient in the centri-
fuge with apical pole centrifugal; but, according to Schleip, the ectoplasm
or cortical layer with the denser parts of the polar plasms attached to it
becomes secondarily oriented independently of the entoplasmic stratifi-
cation, so that a line joining the two polar plasms forms a right angle with
the axis of stratification.

High-speed centrifuging induces or influences ventrodorsality in the
egg of the gephyrean Urechis, a form with spiral cleavage, the centrip-
etal region tending to be ventral (Pease, 1938). In ultracentrifuged
eggs of Cumingia and Chaetopterus cleavage pattern is related to strati-
fication, though with wide variation in Cumingia, and polarity and
ventrodorsality are apparently determined in relation to cleavage pat-
tern (Pease, 1940, "The influence of centrifugal force, etc.," Jour. Exp.
Zool., 84).

With moderate centrifuging preceding cleavage, the Ascaris egg does
not orient, the cleavage pattern is entirely independent of the stratifica-^
tion, and the amount and kind of the stratified substances in particular
cells differ in different individuals and are without effect on development.
Even after separation of "yolk balls" of considerable size, cleavage and

^■' See, e.g., F. R. Lillie, 1906, 1909; Conklin, 1910; Morgan, 19106.

^s Tubifex: Penners, 1922, 19246, 1925; von Parseval, 1922. Clepsine: Schleip, 1914a, b,
1929, pp. 128-31.

S86

PATTERNS AND PROBLEMS OF DEVELOPMENT

developmental pattern in the remainder of the egg may show no funda-
mental alteration (Hogue, 1910). With stronger centrifuging continued
from the stage of separate pronuclei through the first cleavage, the eggs
are flattened by the force, the first cleavage spindle lies at right angles
to direction of the force, and the first cleavage plane coincides with direc-
tion of force. In many of these eggs a cytoplasmic ball containing the
heavier granules separates at the centrifugal pole during the first cleavage
(Fig. 183) because the cleavage furrow does not penetrate this dense region
(Boveri, 1910a, b; Hogue, 1910). Both cells resulting from first cleavage
of these eggs behave like the ventral cell, P^ (Fig. 179), as regards absence
of chromatin diminution and later cleavage pattern. According to Boveri

(1910&), occurrence or nonoccurrence of
chromatin diminution depends on level of
the cytoplasmic gradient at which nuclei
lie. In the "ball eggs" both nuclei result-
ing from first cleavage are at the same
level; consequently, they behave in the
same way, and the level is evidently suffi-
ciently basal so that diminution does not
take place. Since the two blastomeres of
these eggs include the whole, or almost
the whole, polar gradient, their cytoplas-
mically determined nuclear condition may
play a part in determining their behavior as ventral cells. Ultracentrifug-
ing Ascaris eggs usually suppresses the first cleavage, but nuclear division
continues, and all nuclei usually undergo diminution (King and Beams,
1938). To account for these results a specific chemical diminisher sub-
stance is postulated, formed from cytoplasmic substance, more concen-
trated apically, and produced slowly, so that normally the first cleavage
occurs before it diffuses throughout the egg. With suppression of first
cleavage it diffuses more or less throughout; consequently, with few ex-
ceptions, all nuclei undergo diminution. At present the hypothesis that
diminution may be a reaction to nonspecific differences in metabolism
seems equally plausible.

Eggs of Cyclops apparently do not orient in the centrifuge but become
stratified and elongate in direction of the force. The first cleavage spindle
is parallel to the stratification; consequently, the first cleavage is in the
greatest diameter of the egg rather than in the smallest, as normally, but

Fig. 183. — Separation of a cyto-
plasmic ball at first cleavage of strong-
ly centrifuged Ascaris eggs (after Bo-
veri, 1910a).

CLEAVAGE AND DEVELOPMENTAL PATTERN 587

development is normal (Spooner, 191 1). Does cleavage pattern determine
developmental pattern in these cases, as it seems to normally?

Ascidian developmental pattern may be much altered by centrifuging,
either by displacement of cytoplasmic components and resulting altered
localization of particular differentiations or by abnormal cleavage pattern
or by both. Neural tissue, sensory pigment cells, notochord cells, muscle
cells, and cells resembling entoderm may be found, after centrifuging, in
various regions quite different from the normal (Duesberg, 1926; Conklin,

1931)-

Regional differences in specific gravity in the egg of the sea urchin
Arbacia are not sufficient to determine orientation in the centrifuge. When
the egg is centrifuged after fertilization, the first cleavage plane is verti-
cal, the second parallel, the third again vertical to the plane of stratifica-
tion; and the micromeres, normally at the basal pole, form about the
intersection of two of the three cleavage planes, but their position may be
centrifugal or centripetal or lateral in relation to the stratification, and the
stratified constituents are differently distributed in different larvae, de-
velopment being normal. Polarity of egg and embryo is evidently inde-
pendent of the stratification.^^

With strong centrifugal force and a medium of proper density (sugar
solution), unfertilized eggs of Arbacia undergo stratification, elongate in
direction of the force, become dumbbell shape, and then separate into a
colorless and a pigmented portion. In spherical stratified eggs, the first
cleavage plane is usually perpendicular, in elongated eggs, usually parallel
to the stratification. The dumbbell-shaped eggs, when fertilized, give nor-
mal plutei. Both colorless and pigmented portions can be fertilized and
cleave and may develop to plutei. Both portions can be again separated
by further centrifuging into two parts; these cleave on fertilization, but
development does not proceed far. Activated anucleate halves or quarters
can cleave and form blastulae (E. B. Harvey, 1932, 1936, 1939).

That the primary developmental pattern of the insect egg is localized
in the superficial cytoplasm, or in a certain region of it, seems evident
from the superficial character of the blastoderm. Stratification of sub-
stances by centrifuging in eggs of certain Coleoptera (Hegner, 1909) and
muscid Diptera (Pauli, 1927) before blastoderm formation may be fol-

^*Lyon, 1906, 1907; Morgan and Lyon, 1907; Morgan and Spooner, 1909. Morgan and
Spooner believe that the original polarity may be somewhat altered by the effect of stratifica-
tion on cleavage but that position of micromeres and larval polarity is as nearly coincident
with the original axis as altered cleavage permits.

588 PATTERNS AND PROBLEMS OF DEVELOPMENT

lowed by normal or almost normal development. With extreme degrees
of stratification in the longitudinal axis of the beetle eggs, a dwarf larva
develops from the plasmatic part, the centrifugal yolk mass remaining
undeveloped or including some embryonic tissue without definite pattern.
At the blastoderm stage and later, centrifuging has little effect in altering
distribution of substances. Normal, or almost normal, embryos may also
develop from centrifuged muscid eggs, but with sufficient centrifuging the
embryonic zone is intermediate between a still lighter centripetal and a
heavier centrifugal zone; these may remain without development or be
taken into the embryo.

The cortical cytoplasm in which the blastoderm forms after nuclei mi-
grate into it from the interior of the egg may be displaced toward one or
the other end of the egg, but the fact that the embryos developing from
centrifuged eggs show normal axial orientation indicates that the original
pattern persists. Moreover, with centrifuging beginning in stages when
few nuclei are present, these may be displaced centripetally so that only
the centripetal part of the plasmatic zone becomes nucleated, and blasto-
derm and embryo are limited to this region. This is apparently develop-
ment of a whole pattern from a part of the original pattern, that is, a
reconstitution, and suggests, as do the other reconstitution experiments
with insect eggs, that pattern at this stage may be little or nothing more
than a quantitative differential.

It is sufficiently evident from the data mentioned that developmental
pattern in many eggs is, to a high degree, independent of stratification of
visible cytoplasmic constituents or inclosures. Alterations of cleavage by
centrifugal force may be due to displacements of nucleus of spindle or to
mechanical obstacles to cleavage resulting from aggregation of yolk. Ob-
viously, the stratified substances are, in most cases, not "formative sub-
stances," and formative pattern is independent of their distribution. This
pattern, then, must be a property of the "ground substance" (F. R. Lillie,
1909), supposedly not displaced by centrifuging; or it must be localized
in the cortex or superficial cytoplasm of the egg, in which displacement or
stratification of substances apparently does not result from centrifuging.
Differential susceptibility and differential dye reduction seem to indicate
that, in at least some eggs, a polar gradient pattern is present throughout
the egg before the cytoplasmic movements associated with maturation
and fertilization begin. If this is the case, this pattern perhaps persists
only in the superficial cytoplasm, which is less, or not at all, involved in
the streaming of deeper layers or in the stratification by centrifugal force.

CLEAVAGE AND DEVELOPMENTAL PATTERN 589

Gradient pattern and morphological pattern in ciliate protozoa are lim-
ited to the ectoplasm, and results of centrifuging eggs indicate that de-
velopmental pattern may be continuously present only in the cortex,
though perhaps extending to deeper cytoplasm in the oocyte. If centri-
fuging involves the cortex or cytoplasmic areas closely associated with it,
such as the polar plasms of Tubifex and Clepsine, developmental pattern
may be altered. It appears, however, that ultracentrifuging may alter
ventrodorsality and even polarity in some eggs with spiral cleavage, but
whether the alteration results from effect on the cortex or on the deeper
cytoplasm remains to be determined.

If a physiological gradient pattern involving metabolism is continu-
ously present in the cortical cytoplasm, the hypothesis of an elastic or
contractile network, which is merely stretched or distorted by displace-
ment of yolk and other substances in centrifuging and brings the particles
back into place after the force ceases to act, seems scarcely necessary.
Assuming presence of a gradient pattern, the original distribution of cyto-
plasmic granules or inclusions and its changes in maturation, fertilization,
and cleavage are undoubtedly related in some way to this pattern. It
appears probable that differences in condition at different gradient-levels,
perhaps extending to the deeper cytoplasm, may bring about the return to
normal structure and distribution. Electric-potential differences and dif-
ferences in surface tension and in colloidal dispersion are possible factors.

ALTERATION OF CLEAVAGE BY OTHER FACTORS IN RELATION
TO DEVELOPMENT

It is a well-known fact that mitotic spindles tend to orient vertical to
the direction of mechanical pressure and in the greatest cell diameter,
though other factors may prevent such orientation in many cases. The
stratification resulting from centrifuging is apparently very similar to
mechanical pressure in its effect on orientation of spindles in many cells.
Alterations of cleavage pattern and dislocation of blastomeres have been
produced in various eggs by pressure and in some by other means, such as
violent pipetting, shaking, and exposure to calcium-free sea water.

Cells of the early blastula of a medusa, Aegineta, after extreme dislo-
cation by pipetting, tend to reconstitute a spherical blastula (Fig. 184)
and to give rise to normal individuals (Maas, 1901). The apparent polar
difference in size of blastomeres in Figure 184, B and C, suggests that a
more or less normal arrangement may be regained. If a gradient pattern
is present in the blastula, as in other coelenterate blastulae (pp. 96, 167),

S90

PATTERNS AND PROBLEMS OF DEVELOPMENT

potential differences or differences in surface tension in cells of different
gradient-levels may bring the cells back to something like the original
order. Moreover, polarity in embryonic development of various coelen-
terates is highly labile, and it is possible in this case that cells of higher
gradient-levels may be sufficiently dominant to redetermine gradient-
level in other cells "out of placei." In certain other medusae {Geryonia,
Liriopc) calcium-free sea water and violent pipetting are necessary to

B

Fig. 184, A-C. — Dislocation of blastomeres of medusa, Aegineta, and reconstitution of
spherical blastula (after Maas, 1901).

bring about dislocation of blastomeres, and development is more or less
abnormal. The difficulty of dislocation is apparently due to physical prop-
erties of the blastomere surfaces, and these may also prevent rearrange-
ment, or blastomeres may be injured by the experimental procedure.

When the sea-urchin egg is subjected to a certain degree of pressure,
cleavage planes are in the direction of pressure, and eight- and sixtcen-cell
stages are flat plates of blastomeres only one cell thick ; micromercs may
not appear at the sixteen-cell stage. If the membrane is present, these
plates become more or less rounded after pressure is removed, apparently

CLEAVAGE AND DEVELOPMENTAL PATTERN 591

because the membrane tends to return to its original spherical form. In
absence of the membrane, removal of pressure results in divisions vertical
to the previous direction, so that the plate becomes two cells thick. Eggs
with cleavage pattern thus altered may develop into normal plutei.^^
From these modifications of cleavage pattern Driesch concluded that
blastomeres of the sea-urchin embryo are equivalent and can be mixed in
any order without preventing normal development. A critique by Braem
(1893) showed this conclusion to be unnecessary, and these experiments
are now generally regarded merely as evidence for independence of early
cleavage pattern and developmental pattern. If developmental pattern
is a general or a cortical gradient pattern, these changes in cleavage pat-
tern may not affect it at all, and even a regionally specific pattern might
persist essentially unaltered by the changes in cleavage.

Even in nemerteans, annelids, and mollusks — forms with spiral cleav-
age— ^pattern of early cleavage can, to some extent, be dissociated from
developmental pattern. In eggs of species of all three phyla under pres-
sure, cleavage spindles tend to lie in the greatest diarheters of the eggs, that
is, vertical to direction of pressure, and plates of eight or sixteen cells may
result, but the egg polarity is also a factor in determining cleavage pattern
in these experiments. For example, in eggs of the nemertean Cerebratulus,
with pressure at right angles to the polar axis, the first cleavage plane
passes through polar and pressure axis; the second, at right angles to the
first, also passes through the pressure axis, that is, it is equatorial with
respect to the polar axis, instead of meridional like the second normal
cleavage. If pressure is removed at this stage, the third cleavage planes
are meridional, like the second of normal development, and further de-
velopment is normal. Plates of eight or sixteen cells may also result from
pressure, but pilidia developing from these are more or less defective or
abnormal, sometimes partial duplications involving the apical organ
(Yatsu, 1910a, b). According to Dederer (19 10), polarity of the Cerebratu-
lus egg can be altered by pressure; but since no other experimenter has
observed such alteration in forms with spiral cleavage, this may be an
error.

Polar-body position is not altered in the egg of the mollusk Cumingia
by pressure in any direction, and the first cleavage plane usually passes
through the polar axis, though with pressure at right angles to this axis, it
is sometimes obhque or equatorial (Browne, 1910). With sufficient pres-
sure in the direction of the egg axis of some other forms with spiral cleav-

=' Driesch, 1892; Morgan, 1894; Ziegler, 1894.

592 PATTERNS AND PROBLEMS OF DEVELOPMENT

age, the third cleavage is meridional, instead of equatorial, and a plate
of eight cells is formed. On removal of pressure at this stage, the eight
cells, or most of them, particularly those including the region about the
apical pole, give rise to micromeres by divisions vertical to those that
formed the plate.^^ Normal trochophores may develop from these forms
(Wilson); but, according to Morgan, all are at least slightly abnormal.
Alteration of cleavage pattern by pressure in the ascidian Ciona prac-
tically always results in abnormal development (Morgan, 1910a). Here
cleavage pattern is apparently more closely associated with developmen-
tal pattern. Only the mutual relations of the first three cleavage planes
of the ctenophore egg have thus far been altered by pressure. Unequal
divisions occur under pressure, as normally, but later development has
not been followed (Ziegler, 1898). In the egg of Ascaris little change in
cleavage pattern has resulted from pressure (Girgolaff, 191 1; Bonfig,

1925)-

Since it was discovered by Herbst (1900) that in calcium-free sea water
blastomeres of sea-urchin eggs do not flatten against each other but re-
main spherical and may become completely separated, this procedure has
been widely used as a means of obtaining isolated blastomeres or blasto-
mere groups and also in some cases for dislocation of blastomeres. Cal-
cium-free sea water alters condition of the ectoplasmic layer which holds
the blastomeres together. On return to normal sea water the change is
more or less completely reversed, and blastomeres again adhere normally.
Herbst observed that cell division continued in the calcium-free water and
that even isolated cells might become ciliated. Somewhat later Driesch
(1902a) found that dislocation of blastomeres, varying from one individual
to another, could be obtained by exposure to calcium-free water and that,
on return to normal, sea water development proceeded. When not dis-
turbed, sea-urchin eggs in calcium-free water tend to form a cell plate;
and apical or basal cells, normally in contact, may be widely separated.
Dislocation of the four basal cells of the eight-cell stage into two sepa-
rated pairs results in formation of two groups of micromeres; and, if these
remain separated after return to normal sea water, two invaginations and
two archentera may develop. These results are in line with those of Hor-
stadius on transplantation of micromeres (pp. 440-45). In spite of ex-
tensive dislocations of blastomeres, however, many of the more or less
platelike blastomere groups develop into normal plutei. There is consid-
erable change in cell positions as the mass becomes rounded on return to

'^Nereis: E. B. Wilson, 1896; Morgan, 1910a. Crepiditla: Conklin, 1912.

CLEAVAGE AND DEVELOPMENTAL PATTERN 593

normal sea water, and how far return of cells to normal relations takes
place is uncertain. Driesch maintains, however, that normal development
may result, although blastomeres, except the micromeres, remain vari-
ously dislocated, and agrees essentially with Boveri (1901a, b) that the
vegetal region determines other parts. Since entodermization of prospec-
tive ectoderm, ectodermization of prospective entoderm and micromeres,
and induction of invagination from prospective ectoderm by implanted
micromeres have been shown to take place, normal development should
be possible with considerable dislocation of blastomeres.

Cerehratulus eggs from which membranes have been removed by shak-
ing give rise in calcium-free sea water to a ring of eight cells, or a plate
of two rows of four each, or in some cases to eight cells in a single series
(Yatsu, 1910^). In all these the third cleavage, like first and second, is
vertical to the underlying surface. On return to normal sea water at this
stage, all eight cells may give rise to micromeres and rings; plates and
single rows of cells may develop into pilidia, some of which appear nor-
mal, while in others the apical organ is absent or doubled. Either the dis-
located blastomeres regain a more or less normal arrangement, as Yatsu
suggests, or there is partial redetermination of pattern by one or more
dominant regions. Cerehratulus embryos, kept in calcium-free water for
longer periods, form solid irregular cell masses with little visible differen-
tiation, except an apical flagellum and ciliated cells in some. Returned to
normal sea water, these masses differentiate ectoderm and gut but do
not attain full pilidium development. Some of Yatsu's figures suggest
differential inhibition of development, greater in ectoderm than in ento-
derm.

ALTERATIONS OF CLEAVAGE BY DISPERMY AND POLYSPERMY

Effects of dispermy and polyspermy have been most extensively stud-
ied in sea-urchin eggs. The first cleavage of a dispermic egg may be with
two separate spindles into two cells, or with monasters or various forms
of multipolar spindles into three or four cells simultaneously, and the
patterns of following cleavages differ from normal. Normal plutei develop
from only a small proportion of these eggs, and often development does
not continue beyond blastula or gastrula stages. Many of the forms re-
semble those resulting from differential inhibition by external agents.
The blastocoel is often more or less filled with free cells, as in differentially
inhibited forms; and cells may also be given off externally, also as in dif-
ferential inhibition (chap. vi). Cytological studies by Boveri showed that

594 PATTERNS AND PROBLEMS OF DEVELOPMENT

the chromosomes are usually abnormally and variously distributed in the
first cleavage of dispermic eggs, so that some cells obtain less, some more,
than the normal number. ^^ Boveri pointed out that the abnormal develop-
ment could not be due to cytoplasmic factors because 1/4 blastomeres of
normal eggs can give rise to normal larvae, and concluded that it con-
stituted proof of qualitative difference of different chromosomes. How-
ever, according to Schleip (1929, p. 473), frequency of normal develop-
ment of isolated primary three or four blastomeres of dispermic eggs is far
below expectation on the basis of chromosome distribution. Boveri ex-
plained this low frequency as due to injury in separation of blastomeres;
but Schleip suggests that, in addition to abnormal chromosome distribu-
tion and incidental injury, poor condition of eggs favors dispermy and
may be in part responsible for abnormal development. But whatever the
factors determining abnormal or differentially inhibited development in
dispermic eggs, the occurrence of normal development in some whole
eggs and isolated primary blastomeres shows that cytoplasmic pattern is
not necessarily altered.

The modifications of cleavage pattern in dispermic eggs of the mollusk
Dentalium have been described by Schleip (1925). With simultaneous
cleavage into three or four cells, all may be in one plane or, in case of
four cells, may form a tetrahedron with three apical and one basal, or
one apical and three basal. Tetrahedral forms without definite relation to
the polar axis were not observed. The swimming forms which develop
from dispermic eggs of isolated blastomeres of these eggs are, according
to Schleip, too abnormal to permit exact analysis.

Dispermy has been observed occasionally in Ascaris eggs and usually
results in simultaneous division into four cells. Fusion of two or even three
of these cells may occur, supposedly because chromosomes are absent be-
tween certain poles of the tetrapolar spindle. When the four cells persist,
three types of second cleavage are distinguishable, as follows: three cells
undergo diminution of chromatin, one does not; two cells undergo diminu-
tion, two do not; one undergoes diminution, three do not. The three pat-
terns are indicated in Figure 185, A-C, the cells which undergo diminu-
tion being uppermost. In these patterns three, two, or one of the four cells
correspond to AB, and one, two, or three, to P^ of normal cleavage, not

^9 Boveri, 1902, 1904, 1905, 1907. Morgan had suggested earlier that the high frequency of
abnormal development in eggs dividing simultaneously into three or four cells, whether in con-
sequence of dispermy or other conditions, resulted from difference in number of chromosomes
in cells.

CLEAVAGE AND DEVELOPMENTAL PATTERN

595

only as regards diminution but also as regards products of later cleavages
(Boveri, 1910a, b). Boveri's conclusion that polarity in the Ascaris egg
is a gradient was drawn from these cleavage patterns of dispermic eggs.
Although the earlier normal cleavages of Ascaris are highly determinate,
it appears from the dispermic eggs that a physiological cytoplasmic pat-
tern persists independently of the alteration of cleavage pattern; but
whether it is quantitative or regionally specific remains uncertain. Since

Fig. 185, A-C. — Diagrammatic outlines of cleavage of dispermic Ascaris eggs. A, three
AB cells, one Pi cell; B, two AB, two P^; C, one AB, three Pi (modified from Boveri, 19106).

dispermic eggs give rise, not to normal individuals, but only to cell masses,
it is evident that something is wrong with developmental pattern. Ac-
cording to Boveri, eggs with one ventral cell (Pi) develop more nearly
normally than the other types.

Entrance of more than one spermatozoon occurs normally in eggs of
urodele amphibians, but the supernumerary spermatozoa and their cen-
ters take no part in development and disappear. Experimental poly-
spermy in anuran eggs, resulting from high sperm concentration, alters
cleavage because the number of centers equals the number of sperma-
tozoa, but only one sperm nucleus unites with the egg nucleus. Di- and

596 PATTERNS AND PROBLEMS OF DEVELOPMENT

trispermic eggs develop more nearly normally than others and may give
rise to complete embryos and larvae, but there is often regional cessation
of development and death of cells at some stage, and partial forms with-
out reconstitution result. With extreme polyspermy, most of the embryo
may die, or the egg may not cleave.^" Death of cells, of regions, or of the
whole embryo evidently results not from alteration of cleavage pattern
but from the anomalous relations of nuclei and cytoplasm, cells being
often multinucleate, and from abnormal and irregular distribution of
chromosomes by multipolar spindles.

From the facts at hand it appears that dispermy, like centrifuging and
pressure, may alter cleavage pattern without essentially altering cyto-
plasmic developmental pattern but that normal development rarely re-
sults, either because of pathological condition of eggs permitting dispermy
or because of anomalies in chromosome distribution. Chromosome
anomalies may alter or even obliterate developmental patterns by altering
quantitatively or qualitatively the metabolism of cells or cell groups.

DIFFERENTIATION WITHOUT CLEAVAGE IN ANNELID EGGS

Unfertilized or fertihzed eggs of the annelid Chaetopterus subjected for
an hour to certain concentrations of KCl added to sea water may undergo
some differentiation without cleavage or nuclear division (F. R. Lillie,
1902). The ectoplasm becomes vacuolated, as does the ectoderm of the
trochophore, and in some cases a ring of large vacuoles girdling the egg
resembles the ring of large vacuoles in the normal prototroch. Motile
cilia develop on the whole or a part of the surface; and slow or active, but
apparently undirected, swimming may occur. The yolk becomes aggre-
gated into a dense mass in the interior, sometimes separated from the
outer layer by a space. Cytoplasmic divisions may take place, but the
nonnucleated masses apparently fuse again with the nucleated portion.
Cilia develop somewhat later in these forms than normally.

Ciliation without cleavage and in one case a form with a girdle of cilia
resembling the prototroch have also been obtained by KCl treatment of
unfertilized eggs of the polychete Podarke (Treadwell, 1902). Unfertilized
eggs of another polychete, Amphitrite, treated with Ca(N03)2, KCl, or
CaClz or subjected to strong mechanical agitation, may also undergo dif-
ferentiation without cleavage and without, or with, nuclear divison, or
with few irregular or partial cleavages and few nuclei (J. W. Scott, 1906).

In these three forms normal cleavage is typically spiral, but some dif-

3° A. Brachet, 1910a, b, 1912; Herlant, 1911.

CLEAVACxE AND DEVELOPMENTAL PATTERN 597

ferentiation is evidently possible, not only in absence of the normal cleav-
age pattern but in absence of any cleavage or nuclear division. Again it
appears that, even in forms with spiral cleavage, cleavage pattern is not
a fundamental factor in differentiation.

CONCLUSION

It is evident that early cleavage pattern and organismic developmental
pattern may be completely independent or more or less closely associated.
In some cases cleavage may perhaps be a factor in determining develop-
mental pattern, but more commonly developmental pattern plays some
part in determining cleavage pattern when the two are related. Even
in some forms with determinate cleavage more or less alteration of cleav-
age does not necessarily alter developmental pattern. At present it ap-
pears highly improbable that any organism is, in any strict sense, a mosaic
of self-differentiating cells or cell grou-s. At certain stages some parts may
be more independent than others, but capacity for more or less independ-
ent differentiation when they are isolated is not proof that they are in-
dependent in the intact organism. Moreover, the postulated "organ-form-
ing substances" do not exist. An organ is the product of an action system
or various systems in relation to others. A multicellular organ involves
orderly relations of cells and a definite spatial pattern. A particular sub-
stance may bring about differentiation of a cell or tissue; but how can it
form an organ? Trochoblasts of the annehd embryo or the presumptive
notochord of the ascidian may differ specifically from other parts, and
their specific constitutions may determine their differentiation; but a
prototroch or a notochord can develop only in relation to other parts.
The formative factors in development are interrelated action systems.

Even in the most extreme cases early development approaches or attains
mosaic character only as regards certain parts, largely those giving rise
to temporary larval organs or apical or cephalic regions. Eggs of different
species differ widely as regards degree or stability of regional determina-
tion at the beginning of embryonic development; but in no case known
are all parts fixed and stable in their determination or differentiation,
even in relation to isolation alone.

Once more it may be recalled that in many forms with relatively mosaic
embryonic development this is only one of various ways in which indi-
viduals of the species may originate. For example, spiral determinate
cleavage is characteristic of embryonic annehd development; but in many
annehd species individuals can develop from buds, from fissions, or from

598 PATTERNS AND PROBLEMS OF DEVELOPMENT

experimentally isolated pieces of the adult body, in some species even
from a single segment. An ascidian can develop not only from an egg with
highly determinate cleavage and apparent regional cytoplasmic differ-
entiation but from buds of various origin, from isolated pieces of the
adult body, from small pieces of stolons, and in many species from aggre-
gations of cells, forming as other parts of the adult body undergo involu-
tion or degeneration. These forms of development may involve origin
of new polarities, symmetries, or asymmetries, and localizations of pri-
mordia in definite and orderly relations to the new axiate pattern; but an
organization and cleavage like that of the egg and early embryo are evi-
dently not essential in these forms of development. Development from
the egg is apparently the most highly specialized form of annelid and
ascidian development. Egg organization at the beginning of embryonic
development in such forms does not represent the real beginnings of indi-
vidual development. Only by comparative investigation and analysis of
the different forms of development can we hope to distinguish the funda-
mental factors in origin and development of individual pattern from those
incidental to a particular form of development.

If annelid or ascidian embryonic development were a rigidly determined
mosaic, budding, fission, and reconstitution of pieces would be impossible
in later life. If dediiTerentiation of cells is involved in these forms of de-
velopment, the cells which dedifferentiate were not rigidly determined.
If we assume that certain cells remain undifferentiated and are activated
in some way in budding, fission, and reconstitution of pieces, those cells
were certainly not rigidly determined.

In many organisms, even higher vertebrates, incapable, so far as we
know, of giving rise in adult life to new individuals by budding, fission,
or reconstitution of isolated parts, these forms of development are possible
in early embryonic stages and have proved valuable aids in physiological
analysis. It is, of course, far from true that every individual organism
originates from an egg, and some of the other ways in which individuals
originate may be more important than the egg and its development for
attainment of an adequate concept of origin and nature of developmental
pattern.

CHAPTER XV

QUESTIONS OF ORIGIN OF CERTAIN AGAMIC

PATTERNS UNDER NATURAL

CONDITIONS

CONCERNING various developmental patterns we know little or
nothing beyond the fact of their appearance under natural con-
ditions. They have been described, but experimental analysis is
lacking. Some of them are practically inaccessible to present experimen-
tal methods; and only suggestions, hypotheses, or guesses as to their
origin and nature are possible. These may serve, however, to make evi-
dent some of the problems they present and to indicate possibilities. Per-
haps it is as important to call attention in this way to some of the things
we do not know about development as to present established facts. More-
over, information concerning origin and nature of these patterns is no
less, perhaps in some cases even more, essential to an adequate theory of
development than information concerning embryonic development alone.

APPEARANCE OF AXIATE PATTERN FROM PLANT SPORES

Plant spores are usually unicellular but may be multinucleate without
cellular division of the cytoplasm. Many kinds of spores appear, with
different relations to the life-cycle: some apparently have nothing but
surface-interior pattern, others show definite axiate patterns, and in some
algae there seems to be little difference between spires and gametes.

The nonmotile spores of many fungi are spherical or ovoid bodies,
often with no indications of axiate pattern. Germination usually consists
in formation of a bud from one or more regions of the spore; this elon-
gates, forming a mycelial filament, and gives rise to further new axes by
budding, often without formation of separate cells. Germination appar-
ently results from a local activation, as in other buds — perhaps in relation
to some external differential or differentials. Presence of a gradient has
been shown, at least in the apical regions of mycelial filaments of various
fungi. In certain rusts (e.g., Puccinia) germination takes place through
pores in the spore coat on opposite sides, so that axes arise in two opposite

599

6oo PATTERNS AND PROBLEMS OF DEVELOPMENT

directions. Here the pores through which the spore protoplasm pushes
outward may serve to locahze the outgrowth and so the axiate pattern.

Spores of bryophytes and pteridophytes differ in form, those of some
species being spherical, others tetrahedral because of relation to sister
cells, still others bilateral, and in some the spore coat (exospore) may
rupture in a definite manner on germination; but whether a definite axiate
pattern is determined in the spore protoplasm seems not to be known.
Germination is apparently a budding, an activation, and an outgrowth of
a region of the spore protoplasm, determined either by the region of
rupture or by some local differential; or, if the spore possesses an axiate
pattern, that presumably determines the region of initiation of outgrowth.
The primary outgrowth may be a rhizoid, development of other axes oc-
curring later. In the true mosses the moss plant originates as a bud from
the filamentous protonema which results from spore germination. This
type of development resembles the development of the hydranth and stem
in many hydroids from stolonic outgrowths. Evidently the various bud-
dings involved in these forms of development represent new polarities
originating in local activations. Like buds of higher plants and of animals,
they are probably primarily radial gradient systems and become axiate
by differential growth of central and peripheral regions.

Spores of mosses and ferns develop from certain cells of sporangia,
which are usually axiate. Whether spores ever possess a polarity derived
from that of the sporangium seems not to be known; but if they do, it is
probably readily obliterated and a new polarity determined by local ex-
ternal conditions. Certain of the pteridophytes are heterosporous, and
the male gametophytes developing from microspores are reduced to a
few cells inclosed by the spore wall but show a definite orderly pattern,
suggesting a polarity. If a polarity is present, it is probably determined
by the relation to each other and to free surfaces of the four spores of a
tetrad. The female gametophyte of the heterosporous forms, Selaginella
and Isoetes, on the other hand, shows a definite polar pattern in general
coincident in direction with the polarity of the sporocarp axis on which it
originates (e.g., F. M. Lyon, igoi).

In certain water ferns (Hydropterineae) an axiate sporocarp, arising
essentially as a bud, consists of an outer covering (indusium) inclosing a
branching axis with megasporangium apical and one-celled microsporangia
at the tips of lateral branches. Development of any of the thirty-two cells
in the terminal megasporangium inhibits development of the microspo-
rangia; but if all cells of the megasporangium degenerate, the microspo-

ORIGINS OF AGAMIC PATTERNS 6oi

rangia develop (Pfeiffer, 1907). This is essentially a multiaxiate sporan-
gial pattern with apical region apparently dominant, as in multiaxiate
vegetative stages of many plants.

All seed plants are heterosporous with reduced gametophyte genera-
tion, the pollen grain representing the male, and the cells of the embryo
sac in the ovule, or, according to some botanists, the endosperm, repre-
senting the female gametophyte. Among the gymnosperms pollen grains
may consist of several cells in a definite axiate pattern, and the question
arises whether this is determined by the relation of the tetraspores to
each other or by some other factor. Pollens of certain conifers possess
two wings, developing on the outer surface of each tetraspore and sym-
metrical to the axis of the grain indicated by the cells composing it ; that
their localization and the axiate pattern of the grain are determined by
the relation of each spore to the others in the tetrad appears probable.

In most angiosperms the two divisions of the spore mother cell form a
linear series of four cells, coinciding in direction with the ovule axis, the
innermost of these giving rise to the embryo sac. Polarity of the ovule
apparently determines, in some way, the direction of the mitoses. The
long axis of the embryo sac and, in the gymnosperms, of the archegonia
developing in it also coincides with the ovule axis and apparently de-
velops as part of it. The ovule itself is, in certain respects, similar to a
bud in origin but has a symmetry pattern definitely related to the pattern
of the ovary.

Motile swarm spores (zoospores) with flagella are characteristic of vari-
ous algae and some fungi. The macrozoospore of Ulothrix, for example,
is formed by division within a cell of the axiate filamentous thallus, which
consists of a single cell series. One end of the zoospore is somewhat pomted
and bears four flagella; an axiate differentiation is also present, consisting
of colorless cytoplasm containing a contractile vacuole at the flagellate
pole and chlorophyll-bearing cytoplasm elsewhere; on one side of the
body is a red pigment spot, the so-called "eyespot," or stigma (Fig. 186,
A). In short, as often noted, these zoospores resemble flagellate proto-
zoa. The microzoospores (gametes) are similar in pattern, but they are
smaller and have two flagella.

The same questions as to origin of this axiate, asymmetric pattern
arise with reference to these zoospores as for the similar protozoan pat-
terns. Certain early figures (e.g., Dodel, 1876) show the polarities of two
macrospores formed within a thallus cell as parallel, oriented in the
same direction at right angles to the thallus axis, and with pigment spots

602

PATTERNS AND PROBLEMS OF DEVELOPMENT

adjoining, respectively, distal and proximal walls of the thallus cell (Fig.
iS6, B); but the spores in different mother cells of the same filament may
differ in axial orientation. Single thallus cells may give rise to more than
two spores, even to thirty-two or more microspores. In these cases the
spores apparently form at the surface of the cytoplasm adjoining the cell
wall, and their axiate patterns seem to be related to the differential from
surface to interior of the protoplasm. Very similar spores with two flagella
at one pole of an axiate pattern are formed in other green algae, and both
or one of the gametes of some algae have similar axiate pattern. They

> <

)—<

B

Fig. i86, y4-£.— Zoospores of algae. A, zoospore of Ulolhrix; cytoplasmic regions and pig-
mented "stigma" indicated; B, two zoospores developing in a thallus cell with axes transverse
to thallus axis, parallel and similarly directed (.4 and B, diagrammatic, after Dodel, 1876);
C, type of zoospore characteristic of brown algae; D, zoospore of Oedogonium; E, its develop-
ment with axis transverse to thallus axis {D and E after Pringsheim, 1858).

usually develop by repeated divisions of thallus cells, often with forma-
tion of a very large number from a single cell. How the axiate pattern
originates is an interesting question. In many forms the axes seem to be
in all possible directions; certain figures, however, show the flagellate
poles of spores lying superficially directed toward the surface of the
mother cell,' but in most figures no such relation is evident. An axiate
pattern related to, and determined by, directions of mitoses may perhaps
become the polar axis of the spore, the polar region of the spindle be-
coming apical.

It is stated that zoospores of this type attach by the colorless flagellate

' See, e.g., G. M. Smith, 1938, Fig. 59, Codiutn.

ORIGINS OF AGAMIC PATTERNS 603

pole, a rhizoid develops from it, and the chlorophyll-bearing portion be-
comes the thallus. Since the colorless flagellate pole is apical or anterior,
as far as locomotion is concerned, this course of development suggests that
the thallus filament represents a new polarity arising from the primarily
basal pole, as the hydranth-stem axis of calyptoblast hydroids develops
from the originally basal end of the planula (p. 96). These zoospores
and gametes are usually asymmetrical as regards position of the stigma
and apparently, at least in some forms, as regards nuclear position.
Among the brown algae zoospores and gametes occur with two flagella
arising from one side — the longer extending anteriorly, the shorter pos-
teriorly, in relation to direction of locomotion— and stigma closely asso-
ciated with point of flagellar origin (Fig. 186, C). Whether these asym-
metries show any definite relation to particular factors in the environ-
ment of the developing spores is apparently not known.

In certain other algae — for example, Oedogonium and Vaucheria — a
zoospore develops without division from an elongated cell of the thallus.
The zoospore of Oedogonium possesses a circle of cilia near one end and a
polar cytoplasmic differentiation into colorless and chlorophyll-bearing
zone at the level of the cilia (Fig. 186, D). It appears, however, that this
pattern develops at right angles to the thallus axis (Fig. 186, E), but the
factor determining it is not evident. This spore also attaches by its color-
less "apical" pole, from which rhizoids develop; and the thallus forms from
the chlorophyll-bearing part.

The large multinucleate zoospore of Vaucheria, developing at the tip
of a branch, becomes separated by a cell wall from the rest of the thallus,
in which cell walls are absent, and becomes cihated over the entire sur-
face, two cilia developing above each of the numerous superficial nuclei.
In germination, local budlike outgrowths, becoming new filaments, de-
velop from one or more regions. Apparently at least some of these out-
growths represent new axiate patterns, resulting from local activations,
determined by some internal or external differential.

In various algae the zygotes themselves or products of their division
become oospores, quiescent stages showing no indication of polarity and
giving rise on germination to a bud axis in the same way as the agamic
spores. Here, also, new axiate pattern apparently originates from a local
activation. The fertilized egg of Chara becomes an oospore; germination
consists in the origin of a bud from some point of its surface; the bud be-
comes a filament; and the axis of the plant develops as a branch of this
filament, that is, from another bud. It is sufficiently evident from these

6o4

PATTERNS AND PROBLEMS OF DEVELOPMENT

few examples that patterns of various plant spores present interesting
problems and also that development from spores often involves formation
of new axiate patterns originating as buds. These, like other buds, are
apparently primarily radial gradient systems resulting from local activa-
tions and become axiate by differential growth. Doubtless different gra-
dient systems of this sort, even in the same plant, may differ chemically
and physically and, consequently, may give rise to axiate patterns of
different kinds. The case of the Fucus egg (pp. 423-25) suggests that
many of these forms of reproduction and development still afford interest-
ing fields for experiment.^

<sz

A B

Fig. 187, A, B. — Schizogony of a schizogregarine, Schizocystis gregarinoides. A, multi-
nucleate stage; B, formation of merozoites by cell-budding and separation about each nucleus,
progressing from the posterior end anteriorly (after Leger, 1909).

PATTERN IN RELATION TO CELL FORMATION IN SPOROZOA

In many Sporozoa the formation of merozoites and sporozoites and,
in some forms, of gametes shows an interesting relation between pattern
and cell formation. The elongated uninucleate sporozoite of Schizocystis
gregarinoides, a schizogregarine, develops into a wormlike form (schizont) ;
this becomes multinucleate with nuclei coming to lie near the body sur-
face (Fig. 187, A); ^, cell body buds from the parent body about each
nucleus; separation into uninucleate cells results; and these elongate to
form merozoites (Leger, 1909). The separation into cells occurs progres-
sively from the posterior end anteriorly (Fig. 187, B), suggesting a slight

^ References have been omitted from this section almost entirely except in connection with
figures reproduced from papers, because the morphological data appear in botanical textbooks
and because the purpose here is merely to call attention to a few examples of these patterns and
the questions arising from them.

ORIGINS OF AGAMIC PATTERNS

605

and decreasing dominance with physiological isolation progressing from
the posterior end anteriorly. The merozoites may develop either into new
schizonts or into gamonts which pair, encyst, and give rise to male and
female gametes by nuclear multiplication and the budding-off of cells
from a central mass. Two gametes unite, and nuclear multiplication and
cell formation about each nucleus give rise to sporozoites which become
schizonts (Fig. 187, A).

Another schizogregarine, Caulleryella pipientis, has the form of Figure
188, A, in the nonreproductive stage but becomes spherical before repro-

FiG. 188, A-C. — Schizogony of a schizogregarine, Caulleryella pipietttis. A, uninucleate
parasite; B, cell-budding about each nucleus; C, merozoites and cytoplasmic remainder (after
Buschkiel, 1922).

ducing. Reproduction begins with nuclear multiplication; the nuclei ap-
proach the surface, mostly on one side of the spherical mass; and elon-
gated cell bodies bud out from the region about each nucleus, becoming
the merozoites and leaving behind a cytoplasmic remainder (Fig. 188,
A, B, C). Gametes also develop from the gamonts as cell buds about
nuclei, and sporozoite formation from zygotes is essentially similar to
merozoite formation (Buschkiel, 1922). If there is any appreciable domi-
nance in the uninucleate parasite, it is apparently lost before reproduction
begins; but position of nuclei at the stage of cell-budding (Fig. 188, B)
suggests a differential of some sort in the mass, perhaps persistence of
traces of the axiate pattern of the uninucleate form (Fig. 188, A). De-
velopment of merozoites (Fig. 188, B, C) also suggests relation to a polar
differential in the schizont from which they arise.

6o6 PATTERNS AND PROBLEMS OF DEVELOPMENT

Gamete formation in the gregarines commonly takes place in a cyst
containing two individuals. In each of these nuclear multiplication, some-
times to a great number, occurs, each individual in such cases becoming
an irregular mass with extensive surface. The nuclei become superficial,
and a gamete forms about each nucleus, one of the original individuals
giving rise to microgametes, the other to macrogametes. Figure 189 shows
a stage in formation of macrogametes (A) and microgametes (B) in a
eugregarine. Superficial nuclear position apparently results from reaction
to a surface-interior differential in the mass, and the outgrowth in rela-
tion to each nucleus is again a budding of cells from the common mass,
apparently initiated either by the nucleus or by cytoplasmic factors closely
associated with it, such as may be indicated morphologically by a centro-
some or blepharoplast.

A B

Fig. 189, A, B. — Stage in gamete development of a gregarine, Urospora lagidis. A, mega-
gametes; B, microgametes (after Brasil, 1905).

Cell formation by budding from a multinucleate mass in relation to a
nucleus is very generally characteristic not only of gregarines but also of
Coccidia and Haemosporidia, and not only in development of merozoites
and sporozoites but in that of gametes. A stage in development of micro-
gametes of a coccidian is outlined in Figure 190, A. The fully developed
microgamete is a greatly elongated, filamentous, biflagellate cell with
undulating membrane and nucleus also elongated to filamentous form.
The polarity arising in the bud which gives rise to this filamentous cell
apparently determines its axiate pattern. Developmental stages of the
microgamete of the tertian-fever parasite, a haemosporidian, indicated
in Figure 1 90, 5, show a similar cell-budding in relation to nuclear position ;
microgametes of other Haemosporidia develop similarly. In general, re-
action to a surface-interior differential and a local budding of cells from
a multinucleate mass are apparently concerned in origin of such axiate
pattern as these forms possess.

ORIGINS OF AGAMIC PATTERNS 607

The peculiar multinucleate spores of the subclass Cnidosporidia of the
Sporozoa exhibit great variety of pattern in different species (Dofiein-
Reichenow, 1929). They possess a distinct polar organization, some are
flattened in a plane of the polar axis, some are elongated at right angles
to the polar axis, others have long, tail-like extensions of the envelope,
and still others (order Actinomyxidia) develop a triradiate pattern about
the polar axis, the envelope in some species giving rise to long slender
processes and the whole spore resembling a three-armed or, in another
species, a six-armed grappling hook. These spores possess one, two, three,
or four polar capsules, each containing a spirally coiled thread which is
extruded under certain conditions — for example, in some forms on ex-
posure to intestinal fluids, or experimentally to various agents — and is
supposed to aid in anchoring the spore. These capsules resemble nema-

FiG. 190, A, B. — Development of microgametes of other Sporozoa. A, developmental
stage of a coccidian microgamete, Aggregata spinosa (after Moroff, 1908); B, developmental
stages of a haemosporidian microgamete, tertian fever parasite (after Schaudinn, 1902).

tocysts of coelente rates. They are situated at, or symmetrically about,
one pole, or in some species one at each pole. The spore contains one or
more cells from which the new generation develops.

Spores of these types develop in most species from cells descended by
division from a "pansporoblast," two, four, eight, or many spores de-
veloping from one pansporoblast, according to species. A definite number
of cells, different in different forms, attains a polar arrangement; certain
of them give rise to the polar capsules; others to the envelope, which
develops great variety of form, and to one or more "germs"; or a multi-
nucleate mass gives rise to the new individual. There is no visible indica-
tion of axiate pattern in the so-called "germ." In some forms the nuclei
of cells giving rise to different parts of the spore become visibly distin-
guishable early in spore development. In the order Myxosporidia the
cells from which spores develop arise inside an amoeboid body by delimi-

6o8

PATTERNS AND PROBLEMS OF DEVELOPMENT

tation of a part of the cytoplasm about certain "generative" nuclei, which
are thus distinguishable from other "somatic" nuclei. An example of myx-
osporidian spore development is given in Figure 191, with A showing the
two spores of a pansporoblast in a developmental stage, B a single spore of
somewhat later stage, and C the elongated form of the mature spore. This
spore is essentially a multicellular axiate organism, and its development
presents various problems. What determines the difference between gen-
erative and somatic nuclei in the multinucleate amoeboid mass, resulting

B C

Fig. 191, A, B. — Stages of spore development of a myxosporidian, Ceratomyxa drepaiwp-
settae. A , the two spores in amoeboid body, partially developed envelope, polar capsule cells
and polar capsules outlined; spiral thread present at this stage, but not shown in figure; bi-
nucleate germ stippled. B, later stage of a spore, spiral threads of polar capsules indicated in
section. C, form of mature spore (after Awerinzew, 1909).

in formation of delimited cell bodies about the former and not about the
latter? What determines the axiate pattern and the differences in be-
havior of cells, resulting in development of polar capsules, envelope, and
binucleate germ in a definite spatial pattern? These differences in fate
apparently involve differentiation of capsule cells and envelope cells along
quite different lines. The orientation of the two spores in Figure 191, A,
suggests axial determination in reaction to some factor external to the
pansporoblast, but various figures of this and other species show no indi-
cation of a general factor that might determine such pattern. As regards
certain features of cnidosporidian development, there is difference of
opinion among investigators; but there is no question concerning develop-

ORIGINS OF AGAMIC PATTERNS 609

ment of a multicellular axiate pattern in both myxosporidian and actino-
sporidian spores.

The extremely minute spores of the order Microsporidia possess a single
polar capsule or vacuole with long thread. Certain data suggest that in
some species the axiate spore pattern may be determined by factors in
the host cells, perhaps the physiological polarity of these cells, within
which the spores are formed. In the genus Ciirleya, however, the panspo-
roblast, containing four nuclei, gives rise to four spores by elongation in
four directions, the like poles of each of the four spores being, respectively,
peripheral and central as regards the whole group, as if determined by the
differential between surface and interior of the group.

As regards many species, there is no evidence that the "germ" of the
spore possesses or retains any trace of axiate pattern in the later amoeboid
stage. Some of the Myxosporidia are apparently axiate in the amoeboid
stage, but whether this pattern is determined in the spore is not known.
The small size and parasitic habit of these forms present difficulties to
experimental control, but the many problems concerning developmental
patterns in the group are no less important than those of embryonic de-
velopment.^

PATTERNS OF "eMBRYONIC" OR "lARVAl" STAGES OF SUCTORIA

The Suctoria show various types of budding, and the life-cycle involves
metamorphosis of a ciliated, free-swimming stage into a suctorial form,
usually sessile. Some particularly interesting and puzzling questions con-
cerning pattern arise in connection with origin and development of the
buds.-* The ciliated stages, sometimes called "embryos" — perhaps prefer-
ably "larvae" — show in different species considerable differences in axiate
and ciliary patterns and in manner of origin from the parent body. As
will appear, designation of the axes is, to some extent, a matter of view-
point. The disk of attachment of the later suctorial stage is usually visible
in the free-swimming larval forms; and since this constitutes one end of
the chief — often the only — axis of the suctorial form, it seems most con-
venient, though not necessarily otherwise significant, to regard it as rep-
resenting one end of the polar or chief axis in the ciliated stage. Collin,
however, regards this axis as dorsiventral.

3 It has been possible to mention only a few of the interesting features of this material.
Further data appear in Doflein-Reichenow, 1929, and in the literature cited there.

* The types of budding and forms of the ciliated stages are described at length in the mono-
graph by Collin (191 1, 1912), which has an extensive bibliography.

6io

PATTERNS AND PROBLEMS OF DEVELOPMENT

Larvae of Podophrya fixa are more or less ovoid in form and in trans-
verse section, and cilia are present over the whole surface. Before attach-
ment they become spherical and the cilia undergo regression, an equa-
torial band persisting longest. Another larval type is more or less dis-
coid, circular in outline, with disk of attachment in the center of one
flattened face and a marginal band of cilia (Fig. 192, A). This form may
be regarded as possessing a polar pattern vertical to the plane of flatten-
ing and as completely radial. A modification of this type appears in larvae

B

Fig. 192, A-D. — Larval forms of Suctoria. A, Discophrya cybislri, optical section, showing
granules about disk of attachment, meganucleus, two vacuoles, and ciliary band; B, Toko-
pkrya cyclopum, mature attached stage; C, "larval" stage of B, showing disk of attachment
with granules, cilia, meganucleus, and one vacuole; D, Acineta tuberosa, with chief a.xis
obliquely anteroposterior (after Collin, 191 2).

elongated vertically to this polar axis and therefore possessing a physio-
logical anteroposterior pattern and a bilaterality, the original polar axis
being physiologically dorsiventral with the disk of attachment ventral.
According to Collin, the physiologically posterior end of the larva is the
end of the bud last detached from the parent. These larvae usually move
with the "ventral" surface in contact with a substrate. In larvae of certain
other species the disk of attachment is at one pole of a longitudinal axis,
and this pole is anterior in locomotion. These forms possess an equatorial
ciliary band consisting of rows of cilia, different in number in different
species, and a group of cilia at one side of the pole opposite the disk of

ORIGINS OF AGAMIC PATTERNS

6ii

attachment. Except for this group of cilia the larvae appear to be com-
pletely radial (Fig. 192, B, C). Larvae of this type are free-swimming. In
still other larval forms the original polar axis, indicated by the disk of
attachment, is oblique to a dorsiventral pattern; consequently, bilateral-
ity is present, but a different bilateraUty from that mentioned above
(Fig. 192, D), in that the polar axis tends to become physiologically an-
teroposterior while in the other type it is physiologically dorsiventral and

B

D

Fig. 193, J-E.—Larval development from internal buds in Suctoria. A, B, two stages of
bud development in Tokophrya cyclopum, dividing meganucleus in B.; C, Choanophrya inftindib-
ulifera with one larva free in cavity, a second developing, and perhaps a third in still earlier
stage; D, early stage of Pseiidogemma fraiponti; E, Acineta tuberosa (after Collin, 1912).

at right angles to the anteroposterior axis. As regards the variety of axial
relations, these larval forms resemble somewhat the types of axiate pat-
tern in echinoderms.

Origin and development of the larval forms from the suctorial form
differ in different species and in some seem to differ from anything known
in other groups. Larvae of many species develop from "internal buds" in
a cavity of the parent body (Fig. 193). According to various authors,
their development in certain species shows two features of particular in-

6i2 PATTERNS AND PROBLEMS OF DEVELOPMENT

terest. First, the polar (longitudinal) axis of the larva has been repeatedly
described and figured as transverse to the chief axis of the parent suctorial
body; second, separation and differentiation of the larval from the parent
body progresses transversely to the larval longitudinal axis. Two stages
of this type of larval development are shown in Figure igs, A and B. In
A the larval axis is almost parallel to the parental axis, and in B it is not
transverse. According to Collin, this position is merely an anomaly, but
it suggests the possibility that the larval axis may be primarily parallel
to, and determined by, the parental axis and that the transverse position,
even though it may be assumed or approached very early in development,
is actually secondary. If this is true, the larval pole bearing the disk of
attachment develops toward the distal pole of the parent; and if a gra-
dient is present in the parent body, the distal pole probably represents
its high end. As already noted, the larval pole of attachment is in advance
in locomotion, that is, it is physiologically apical. In Figure 193, C, of
another genus, the position of the fully developed larva with the pole of
attachment more or less proximally directed with respect to the parent
has doubtless resulted from locomotion after separation. The axis of the
second developing larva is oblique, not transverse to the parental axis;
and the figure suggests that a third larva may be developing below and to
the right of the second, with its axis still more nearly in line with the
parental axis. The progress of development transversely in the larva is in-
dicated in Figure ig^,A,B, and C; but in A and B and in the second larva
of C the physiologically apical pole of attachment appears to be slightly
in advance of the opposite pole as regards separation from the parent
body, that is, there is some evidence of an apicobasal differential. Collin
gives figures of other species showing the larval axis parallel or almost
parallel to the parental axis (Fig. 193, D, E).^ However, even if the larval
polar axis is primarily coincident in direction with, and a part of, the
parental axis, the question of the factors determining the change to trans-
verse position in certain species and the plane in which it occurs remains.
If it is primarily transverse to the polar axis, as it appears to be from many
descriptions and figures of certain species, the manner of its establish-
ment and the progress of larval development from one side of the body
to the other present puzzling problems.

Larvae of some Suctoria develop from external buds. The larva of

5 See also Collin, Figs. Ill/, LXXXIII, LXXXIV, XCI a, XCV e. Still other figures sug-
gest that the polar axis of the larval type of Fig. 192, D, may also be primarily parallel to the
parental a.\is.

ORIGINS OF AGAMIC PATTERNS

613

P. fixa, for example, develops from the apical region of the parent, the
cilia covering its whole body replacing the suctorial tentacles progres-
sively from the free end proximally (Fig. 194, A). Larval development
from an external bud in a manner like the development from internal
buds shown in Figure 193, A-D, has also been observed in some species
(Fig. 194, B) and again raises the question of relation of bud axis and
parental axis. According to Collin, the bud axis in these forms is usually
transverse to the parental axis but may be oblique, as in Figure 194, B.
This figure suggests, however, that the bud axis may have been primarily
parallel to the parental axis.

Fig. 194, A-C. — Larval development from external buds in Suctoria. A, Podophrya; B,
Paracineta patula; C, Ephestia gemmipara in sagittal section, showing invaginated ciliary band
near free end, "cytostome" near attached end, and pole of attachment indicated by secretory
granules (after Collin, 19 12).

Larvae of some other species develop from external buds arising in a
circle about the periphery of the more or less flattened distal or apical
surface of the suctorial form. In a typical case the surface of the bud
toward the center of the parental disk becomes flattened ; the disk of at-
tachment and rows of cilia develop on it, the ciliated region becoming
invaginated ; and an infolding, regarded by Collin as a rudimentary cyto-
stome, also appears near the attached end of the larva (Fig. 194, C). In
such larvae the polar axis, indicated by the disk of attachment, either has
become physiologically dorsiventral, and is transverse or oblique to the
suctorial parental axis, or is bent in a right angle. The larva shows an
anteroposterior pattern parallel to the parental axis. It is not evident
from the descriptions and figures how this pattern originates, but the
following suggestion may possibly have some value. Assuming that the

6 14 PATTERNS AND PROBLEMS OF DEVELOPMENT

pole of attachment of the larva, indicated by stippling in Figure 194, C,
was originally determined in the apical region of the early bud, a bending
or folding-over of the growing bud toward the center of the parental disk
would result in the position and form of Figure 194, C. Larval form and
invagination of the ciHated band suggest that such a bending or folding
by differential growth of the two sides may have taken place. The "cyto-
stome," which apparently has no function and disappears later, may also
result from a bending or folding of the larval body in early stages. Accord-
ing to this suggestion, the original polar axis of the larva is primarily

A B

Fig. 195, A, B. — Ophryodendron reversum; early stages of vermiform individuals developing
from buds; in B, dividing meganucleus (after Collin, 191 2).

parallel to, and presumably determined by, either the parent axis or a
local bud activation and becomes bent so that the disk of attachment is
physiologically ventral, perhaps in consequence of a central-peripheral
radial differential in the parental distal region. The peripheral position
of the buds suggests that such a differential is present. In short, even the
larval pattern in these species may perhaps be determined by the relation
of the bud to the parent body.

Certain suctorial species give rise, by external budding, to extremely
elongated vermiform individuals (Fig. 195) with longitudinal axes parallel
to the parental axis and apparently determined by it or resulting from a
local activation. Species of the genus Hypocoma, ectoparasitic on other
protozoa, and some other forms develop only a single suctorial tentacle

ORIGINS OF AGAMIC PATTERNS 615

and apparently have no attached stage. They show a physiological and
morphological dorsiventrality with ventral ciliary bands in elliptical pat-
tern, except for slight lateral asymmetry. These forms undergo trans-
verse fission like most ciliates.

At present we have only the data of observation concerning these pat-
terns of suctorial development. The peculiar development of axiate larval
patterns from internal buds, with progress of differentiation from one side
to the other, differs from anything known in other groups. The pole of
attachment is physiologically ventral in some forms (Fig. ig2,A), physio-
logically apical in others (Fig. 192, C), and physiologically almost anterior
in still others (Fig. 192, D).

CERTAIN FEATURES OF PATTERN IN OTHER AXIATE PROTOZOA

Axiate patterns of individual cells composing the spherical colonial
flagellates, such as Pandorina, Volvox, and various others, appear to be
directly related to a surface-interior difference; but they may represent
polarities persisting through the successive divisions from the primary
cell of the colony, the divisions being parallel to the polar axes of the cells.
A relation between cell formation and pattern, apparently similar to that
so general among Sporozoa, appears in certain reproductions of Masti-
gophora. Swarm-spore development in Nodiluca is preceded by repeated
nuclear division near the surface of a part of the parent body, the apical
region. Following this period, budding of a cell body in relation to each
nucleus results in development of a swarm spore or perhaps a gamete with
flagella originating from the apex of the bud (Pratje, 192 1). The surface-
interior differential of the mother cell evidently plays a part in deter-
mining the position of the nuclei and may be concerned, together with
the budding, in determining the axiate pattern of the spore. The longi-
tudinal fission so generally characteristic of Mastigophora may be re-
garded as a sort of budding of the two cells from each other, with polarity
persisting but with reconstitution of symmetry or asymmetry in each
part in definite relation to the polar pattern as separation proceeds. The
transverse fissions characteristic of most ciliates involve, of course, polar
reconstitution. In many ciliates, however, extensive dedifferentiation of
the pre-existing pattern is associated with fission, so that what finally
results is actually a new pattern, but one that develops in direct relation
to the original pattern and presumably under its influence. For exar^ple,
the two new cirrus fields appearing in fissions, as well as in reconstitutions,

6i6 PATTERNS AND PROBLEMS OF DEVELOPMENT

of Hypotricha represent a new pattern appearing in a certain definite re-
lation to the parent pattern.^

In Vorticella the two products of the longitudinal fission differ. One
remains attached to the stalk and develops peristome and gullet. In the
other a ciliary ring develops about the originally basal pole, the individual
becomes free, swims with the originally basal pole in advance, and later
attaches by this pole; following attachment the ciliary ring disappears,
and the peristomial region develops. The question whether this develop-
mental history involves two reversals of polarity and of dominance arises
from the apparently opposed character of the two patterns. If one product
of fission in Vorticella is less directly connected with the stalk than the
other and if reduction of the peristome permits development of the basal
ciliary ring and attainment of dominance by the basal region in this prod-
uct, its separation would be expected. As in the planulae of calyptoblast
hydroids (p. 96) and some larvae of Suctoria (p. 612), the pole function-
ally apical in the free-swimming stage becomes basal on attachment.
Both products of fission usually remain attached in the colonial vorticel-
lids, but certain zooids may become free-swimming "ciliospores" with
change in pattern and behavior very similar to that in the free form of
Vorticella. The free-swimming microgametes, resulting from successive
fissions without intervening peristome development, also show similar
change in pattern and behavior. Evidently this "reversal" represents a
characteristic form of development for this genus and, to some extent,
for other members of the group.

SYMMETRIES AND ASYMMETRIES OF PROTOZOA

A high degree of morphological differentiation and an almost endless
variety of symmetries and asymmetries, highly species-specific in char-
acter, appear in the ectoplasm of flagellate and ciliate protozoa.^ Some
of the varieties of larval suctorial pattern have already been described.
In some forms the pattern is apparently strictly radial — for example,
Prorodon — but the body revolves about the polar axis in locomotion like
free-swimming developmental stages of many invertebrates, indicating

^ Wallengren, 1901; D. B. Young, 1922; Dembowska, 1925, 1926; Taylor, 1928.

" Figures are not given: first, because only a very large number would give any adequate
survey of the variety of these morphological patterns; second, because every student
of zoology is familiar with some of them, and some are figured and described in every textbook
of zoology and many others in the works on protozoa — for example, Kent, 1880-82; Biitschli,
1883-89; Doflein-Reichenow, 1929; Calkins, 1933, etc.; also a voluminous literature on par-
ticular species of groups.

ORIGINS OF AGAMIC PATTERNS 617

a spiral pattern of ciliary beat. Very commonly, however, the morphologi-
cal pattern shows a spiral asymmetry, even in forms such as Paramecium,
in which there is comparatively little regional morphological differentia-
tion; and spiral pattern is often much more conspicuous in the more
highly differentiated forms. The pattern of the peristomial cilia and mem-
branelles of ciliates is usually spiral, and in many species the whole ecto-
plasm apparently possesses spiral pattern. Among the parasitic flagellates
remarkable developments of certain types of spiral pattern appear. In
the genus Spirotrichonympha, for example, spiral bands at or near the
inner surface of the ectoplasm make a series of definite, regularly spaced
turns around the body, and the numerous flagella connected with them
form corresponding spirals. Of two species, S. polygyra and S. bispira,
recently described by Cleveland (1938) with spiral bands extending over
some three-fourths of the body length, the former shows forty-five, the
latter thirty-four, turns of the bands; and, according to Cleveland's fig-
ures, there is little deviation from geometrical regularity in the spacing
of successive turns and slope of spiral. In division of S. polygyra the
bands unwind, beginning at the anterior end, two going to each daughter
individual and two developing anew. In S. bispira a new band without
spiral arrangement develops from the point of origin of the parent bands
at the anterior end and migrates to the posterior end of the cell, which
becomes the anterior end of one of the daughter individuals. There the
band gradually becomes spiral, at first not completely regular, beginning
at the end originally anterior in the parent cell and now at the new an-
terior end of the daughter individual. The second flagellar band develops
from the anterior end of the first, extends posteriorly in the new individual,
and forms a spiral between the turns of the first band. Various other
features of pattern in these species — the axostyle and the elongated cen-
trioles — are of interest as indicating a high degree of differentiation, but
the description of these remarkable forms and their divisions serves to
emphasize the fact that we have not the remotest conception of the physi-
ological factors concerned in the origin and development of such pattern,
but that it originates and is localized in relation to regional differences in
physiological activity with a definite pattern seems evident.

A dorsiventral, as well as anteroposterior, pattern appears in the Hypo-
tricha, together with a great variety of morphological asymmetries in ar-
rangement of cirri and membranelles, most of which are genus- and
species-specific; but in cell division the asymmetries of daughter individ-

6i8 PATTERNS AND PROBLEMS OF DEVELOPMENT

uals originate as ectoplasmic fields in definite relation to the parental pat-
tern, and their orientation is presumably determined by it.

In many of these protozoan forms more or less morphological dediffer-
entiation occurs in connection with division, and the characteristic asym-
metries redevelop in the new cell generation ; but the essential physiologi-
cal factors of pattern evidently persist, or originate within the parental
pattern in an orientation with respect to it which is definite and char-
acteristic for the species. That the various asymmetries are in some way
associated with the specifically different physicochemical constitutions
of the species-protoplasms seems beyond question. Various suggestions
and possibilities will be considered in a later section after some other asym-
metries of unicellular individuals have been noted.

PATTERNS OF PLANT AND ANIMAL SPERMS

Antherozooids, and spermatozoids of plants and spermatozoa of ani-
mals show great variety of morphological pattern. A heteropolar longi-
tudinal pattern is almost always present; in some forms there is complete,
or almost complete, radial symmetry about the axis of this pattern; in
others a spiral pattern develops, either in the whole individual or in some
part of it; and in still others other asymmetries characteristic for the
species appear.

ANTHEROZOOIDS AND SPERMATOZOIDS OF PLANTS

The sperms of many algae are biflagellate forms similar to the zoospores
in general pattern (see Fig. i86. A, C) but often smaller than zoospores
of the same species. In some forms there appear to be gradations in size
and behavior as zoospores or gametes. The pigmented spot or stigma is
asymmetrical in position in forms with terminal flagella, as in similar
zoospores (Fig. i86. A); and in forms with lateral flagella both flagella
and stigma are asymmetrical in position (Fig. i86, C). As in the case of
zoospores, the question how these patterns originate or whether they show
relation to any particular factors is unanswered (see pp. 601-3).

The antherozooid or spermatozoid of Chara is an elongated, spirally
coiled biflagellate cell (Fig. ig6, A). The antheridial filament consists of
a single series of flattened cells, each of which metamorphoses into a single
antherozooid. After the final division the nucleus moves toward the side
wall; the blepharoplast appears adjoining the cell surface toward the base
of the filament, elongates, and becomes spirally coiled; and elongation and
spiral coiling of the nucleus follow, the axis about which the spiral develops

ORIGINS OF AGAMIC PATTERNS

619

coinciding with the axis of the antheridial filament.^ In this development
several points of interest as regards pattern appear. As regards direction

B

Fig. 196, A-E. — Plant spermatozoids. A, Chara (after Belajeff, 1894); B, C, D, stages in
development of Eqtiisetum spermatozoid (after Sharp, 191 2); E, spermatozoid of fern, Ne-
phr odium (from Yamanouchi, 1908).

of nuclear movement toward one side of the cell, both Belajeff and Mottier
give figures in which nuclei of successive cells of a filament have evidently
moved in the same direction, as if reacting to some factor external to the

* Belajeff, 1894; Mottier, 1904.

620

PATTERNS AND PROBLEMS OF DEVELOPMENT

filament, and also figures in which the two plane walls of successive fila-
ment cells diverge toward the same side and the nuclei are on the side
where the walls are farthest apart. The blepharoplast makes its appear-
ance adjoining the transverse cell wall toward the base of the filament, and
spermatozoids develop with the flagellate end also toward the base, sug-
gesting orientation of the pattern in relation to some factor in the fila-
ment as a whole, perhaps a gradient of some sort.

Spiral spermatozoid pattern is characteristic of bryophytes and pterido-
phytes and develops as elongation and spiral coiling, first of the blepharo-
plast, then of the nucleus. In certain of these forms what may be called

Fig. 197, A-E. — Male gametes of plants. A, pollen grain; B, section of developing sperma-
tozoid of Cycas revoluta (after Ikeno, 1898) ; C, spermatozoid of the cycad Zamia (from Webber,
1901); D, male nucleus of the sunflower Helianthiis (from Nawaschin, 1900); E, male nucleus of
Silphium (from IMerrell, 1900).

the "polar axis," that is, the axis about which the spiral coiling occurs,
coincides in direction with the axis of the final mitotic spindle, the
blepharoplast before elongation being at the pole of the spindle. For ex-
ample, in Equisetum (Sharp, 191 2) the two sister spermatids of Figure 196,
B, evidently possess an axiate pattern in relation to each other, and the
later spiral coiling of blepharoplast and nucleus is about this axis (Fig.
196, C), resulting in the spermatozoid (Fig. 196, D). Spermatozoid de-
velopment in the fern Nephr odium (Fig. 196, £) is apparently very similar
(Yamanouchi, 1908). In the cycads — a group related, on the one hand,
to pteridophytes and, on the other, to spermatophytes — spirally ciliated
spermatozoids also develop. The cycad pollen grain (Fig. 197, A) shows
a definite axiate pattern, consisting of three cells, two representing the

Fig. 198, A-F. — Animal spermatozoa. A, Plagiostomum, a turbellarian (from Bohmig,
1891); B, Castroda and C, Mesostomum, both turbellaria (from Luther, 1904); D, Ascaris
megalocephala, a nematode (from Scheben, 1905); E, Nereis, a polychete annelid (from F. R.
Lillie, 1912); F, Paliidina, a gasteropod (from Meves, 1903).

622 PATTERNS AND PROBLEMS OF DEVELOPMENT

male gamctophyte, the third giving rise to spermatozoids. This pattern
is presumably determined in the microsporangium. The polarity of the
grain is also evident in its germination and formation of the pollen tube
at the small-celled pole. In the tube the generative nucleus divides with
spindle transverse to the long axis of the tube, forming two hemispherical
cells. In each of these the blepharoplast, described as derived from the
centrosome, elongates, becomes spirally coiled anticlockwise, and gives
rise to a band of cilia (Fig. 197, B, C). The axis of coiling coincides with
the spindle axis of the preceding division ; the ciliated part of the sperma-
tozoid develops from the region about the pole of the last spindle; and
the rounded, nonciliated part from the flattened cell surface in contact
with the sister cell.^ The spermatozoid axis is at right angles to the axis
of the pollen grain and pollen tube and is apparently determined by some
factor correlated with division of the spermatogenous cell, since it coin-
cides with the axis of the spindle and is indicated by the hemispherical
form of the spermatid and by position of nucleus and centrosome; and
this division is definitely related to the axis of the pollen grain and tube.
In short, each step in development of the polar pattern appears in a defi-
nite relation to a pre-existing pattern. The spiral of the blepharoplast
appears only secondarily and in a definite relation to the polar axis.
During early development of the spiral band a beaklike extension of the
nucleus becomes closely associated with the blepharoplast, as if it were
about to elongate and become spiral, as in the spermatozoids described
above; but this association is only temporary, and the nucleus does not
become spiral. The male nuclei of various seed plants, however, do de-
velop a more or less pronounced spiral form preceding fertilization (Fig.
igj, D,E).

ANIMAL SPERMATOZOA

The great variety of form, the spirals, and specific asymmetries of ani-
mal spermatozoa have long been familiar to students of spermatogenesis.
The spermatozoa of Figures 198-200 show fully developed spermatozoa
of turbellaria (Fig. 198, A-C), a nematode (Fig. 198, D), a gasteropod
(Fig. 198, E), an annelid (Fig. 198, F), the more or less radial spermatozoa
occurring among Crustacea (Fig. 199, A-C), a spider (Fig. 199, D), a
beetle (Fig. 199,. £), a fish (Fig. 200, A),2l salamander (Fig. 200, B), and
a toad, with undulating membrane, resembling somewhat that of a try-
panosome (Fig. 200, C). In certain entomostraca the spermatozoa are
spherical cells; those of some others are amoeboid, according to Weismann

» Ikcno, 1898; Webber, 1901.

ORIGINS OF AGAMIC PATTERNS

623

(1880) and Zacharias (1885); and those of the cirripeds are described as
fibrillar without differentiation of head or tail.'"

D

Fig. 199, A-E. — Arthropod spermatozoa. A-C, Pinnotheres, Maja, Munidia, all Crustacea
(from Koltzoff, 1906); D, Agalena, an arachnid (from Bosenberg, 1905); E, Copris, a beetle
(from Ballowitz, 1890&).

Most animal spermatozoa are distinctly axiate, and spiral patterns and
specific asymmetries are related in a definite way to the axiate pattern.
Moreover, in many forms the polar pattern appears to be definitely re-

"* For further descriptions and figures of spermatozoa see Retzius, 1881; many papers by
Ballowitz, 1886-1908; Koltzoff, 1906, 1909; Korschelt und Heider, 1902, and literature cited;
E. B. Wilson, 1925, and literature cited.

624

PATTERNS AND PROBLEMS OF DEVELOPMENT

lated in the testis to a blastophore or cytophore or a particular kind of
accessory cell. The cytophore, characteristic of many invertebrates, has
somewhat different origin in different species. Successive divisions of

Fig. 200, /I -C— Fish and amphibian spermatozoa. .4, Raja; B, Triton {A and B from
Ballowitz, iSgoc); C, Bombinator (from Broman, 1900).

spermatogonia and spermatocytes are cytoplasmically incomplete in some
forms; and more or less spherical groups of cells, separate peripherally,
continuous centrally, result, the central cytoplasm becoming the cyto-
phore (Fig. 201, ^). In other species the cytophore develops from degen-
erated cells in the center of a mass (Fig. 201, B). The spermatozoa de-

ORIGINS OF AGAMIC PATTERNS

625

velop, either irregularly scattered over the surface of the cytophore (Fig.
201, C) or covering the whole surface like an epithelium (Fig. 201, D).
The point of particular interest is that the polar pattern of the sperma-
tozoon develops radially to the cytophore, the motile tail, when present,
being peripheral (Fig. 201, C, D). The tubular nematode testis contains
a rhachis to which the cells remain attached up to a late stage of sperm
development, and the sperm axis is apparently the axis between free and
attached pole of the spermatocyte.

Spermatozoa of many other forms, both invertebrate and vertebrate,
develop in bundles with axes parallel and heads imbedded in, or in contact
with, certain accessory cells. The spermatocysts, common among insects

A B C

Fig. 201, ^~Z>.— Sperm development in relation to a cytophore. A, spermatocytes and
part of cytoplasmic cytophore of earthworm (after Calkins, 1895); B, origin of cytophore of
the opisthobranch, Triopa clavigera, from degenerating cells (after O. S. Jensen, 1883); C, part
of cytophore and developing spermatozoa of the turbellarian, Plagiostomiim sulplmreum (after
Bohmig, 1891);/), cytophore of the annelid, Clitellio arenarius, containing degenerating nuclei
and showing a few of the spermatozoa which cover the whole surface (after O. S. Jensen, 1883).

and the lower vertebrates, consist of spermatogenous cells inclosed in an
epithelial cyst; and spermatozoa develop parallel with heads toward, or
imbedded in, one of the cells of the cyst wall, usually of considerable
size, and believed to show evidence of secretory activity. How this par-
ticular cell is determined does not appear. The relation of spermatid pat-
tern to the basal or Sertoli cells appears very clearly in the mammalian
testis (Fig. 202, A, B).

Cytophores and accessory cells to which the sperm axes are definitely
related have commonly been regarded as nutritive in function; and cases
like Figure 202, A and J5, have suggested the possibility of a tactic or
tropistic reaction of the spermatid to the Sertoli cell. But the question
arises whether the observed relations represent orientation of an axiate
pattern already determined, that is, a tropism in the strict sense, or de-

626

PATTERNS AND PROBLEMS OF DEVELOPMENT

termination of axiate pattern by the relation to the cytophore or acces-
sory cell (see Broman, 1902). It seems probable that in at least some forms
with cytophore the axiate pattern of the spermatozoon is determined by
the difference in conditions between the pole toward the cytophore and

Fig. 202, A, B. — Relation of mammalian spermatids to Sertoli cells. A, rat (after von
Lenhossek, 1898); B, man (after Broman, 1902).

the peripheral pole. In many cases there is cytoplasmic continuity with
the cytophore until development of pattern is far advanced. A tropistic
reaction is, of course, possible in the cases of secondary association with
an accessory cell, as in Figure 202, A and B, but does not appear very
probable. The polar position of the nuclei suggests a reaction to a factor

ORIGINS OF AGAMIC PATTERNS 627

external to the spermatid and certainly shows a definite relation to the
accessory cell, as does the axial fiber.

Studies of spermatogenesis have usually been concerned with the sperm
cell alone and have given little attention to the question of possible de-
termination of its polar pattern by factors external to it. However, in
many accounts of spermatogenesis extensive changes in position of cen-
triole, acroblast, etc., are described." Axiate pattern of the spermatozoon
in many forms develops after the final division by changes in relative
position (sometimes of almost 180°) of distinguishable parts of the cell.
If the pattern is inherent in the cell, these shifts in position seem difficult
to account for, and they have usually been merely described without any
attempt at interpretation. But if they represent reaction of different cell
constituents to a differential or other factor external to the cell, the diffi-
culty of interpretation is certainly lessened.

Cytophores or other accessory cells are entirely absent from spermato-
genesis of some forms; nevertheless, a spermatozoan pattern develops.
Cytological studies suggest that pattern may be determined by the final
cell division, as it apparently is in certain plant spermatozoids, the centro-
some-nucleus axis determined by direction of division becoming the sper-
matozoan axis, with the tail, when present, developing from the centro-
somal pole. This axis probably becomes the sperm axis unless an axis in
different direction is determined by reaction to some external factor, such
as cytophore or accessory cell.

UNICELLULAR ASYMMETRIES : QUESTIONS AND SUGGESTIONS

Development of a polar axiate pattern in animal spermatozoa appar-
ently precedes appearance of spirals and specific asymmetries. This
seems also to be true for plant spermatozoids with spiral patterns, but
in those the primary pattern is usually merely that resulting from the
final cell division. In general, the spiral patterns and specific asymmetries
seem to develop secondarily in a definite orientation with respect to the
primary pattern. Moreover, they appear to be associated with a high de-
gree of morphological structurization and differentiation rather than to
represent a general property of protoplasms. As regards the protozoan
asymmetries, it may appear at first glance that there is no such distinc-
tion between a primary axiate pattern and secondary asymmetries. In
protozoan fission the patterns of the daughter individuals are not entirely

" For figures and description of cases see E. B. Wilson, 1925, pp. 356-85. Many others are
given in the literature of spermatogenesis.

628 PATTERNS AND PROBLEMS OF DEVELOPMENT

new patterns but are reconstitutions from portions of the parent pattern
or in certain orientation with respect to it. In the longitudinal fissions of
flagellates the spirals and asymmetries are apparently remade in relation
to a polar pattern. Divisions of the two species of Spirotrichonympha, as
described by Cleveland, provide interesting evidence on this point (see
p. 617). In S. polygyra the four spiral bands unwind, two of them form
a new spiral about each new polar axis, and two new bands develop about
each axis. In S. hispira a new band begins to develop in the parental an-
terior region, migrates posteriorly, and becomes spiral only after it reaches
the original posterior end, which becomes a new anterior end. Moreover,
when the spirals appear, they are at first irregular and only gradually
attain the regularity characteristic of the fully developed animal. Cleve-
land's description suggests that the new spiral patterns are determined
by a more general longitudinal pattern already present in the new an-
terior regions. When Stcntor divides, the new peristomial band is not at
first spiral and only gradually attains its definitive form, apparently in
relation to an axiate pattern already present. According to a personal
communication from Dr. C. V. Taylor, the ciliate Colpoda duodenaria
at the beginning of excystment is almost completely radial, as far as
visible pattern is concerned. The ends of the ciliary meridians center
about the two poles with only slight indication of the spiral asymmetry
characteristic of the fully differentiated individual. The oral region be-
gins development as a shallow invagination near, and at one side of, the
anterior pole and progresses posteriorly in a slight spiral. In association
with this spiral invagination of the mouth the anterior portions of the
ciliary meridians undergo spiral twisting As the oral region migrates pos-
teriorly, the anterior polar region changes its position correspondingly,
with flexure of the polar axis, so that the original anterior pole comes
eventually to lie somewhat anterior to the mouth, but distinctly lateral
as regards the whole body. At the posterior end, however, the radial
arrangement of the ciliary meridians persists. Spiral pattern in this species
is apparently almost completely obliterated in the encysted stage and
reconstitutes in relation to a polar pattern on excystment.

Division in many forms, both flagellates and ciliates, involves what
seems to be more or less dedifferentiation of parental organs and develop-
ment of new organs, either in normal relation to each other or, in case of
the cirri of hypotrichs, in localized areas with a following change in posi-
tion, apparently in relation to a pattern.

In short, various lines of evidence indicate that in the protozoa, as

ORIGINS OF AGAMIC PATTERNS 629

well as in spermatozoa, the spiral features of pattern and the other spe-
cific asymmetries develop secondarily in a definite relation to a primary
pattern. In this connection it may be recalled that presence of a longi-
tudinal physiological gradient has been demonstrated in the ectoplasm
of a considerable number of ciliates by various methods, and at present
there is no evidence to indicate that the differentials in the protoplasmic
substrate and its activities which constitute these gradients are not the
primary axiate pattern. If they are primary, the spirals and asymmetries
represent secondary structural expressions and differentiations of the pri-
mary pattern in the various species-protoplasms. Whether axiate pattern
of the spermatid is based on a gradient pattern is not known, but the
apparent determination of the spermatid polar axis by an external differ-
ential in many forms and the changes in position of parts of the cell sug-
gest that a gradient pattern may be present.

Whether or not the primary pattern in these unicellular individuals is
a gradient, the problem of the nature of the spirals and other asymmetries
and their relation to it remains. The possibility suggests itself that many
of these features of unicellular pattern may result from, or be expressions
of, molecular or micellar structure or aggregation in definite orientation.
Researches of recent years with polarized light and X-rays have given
evidence of orientation of molecules or particles in cellulose membranes
and fibers, various animal fibrillar structures, connective tissue, muscle,
nerve, keratin of hairs and feathers, chitin, etc. Thus far this evidence
of molecular pattern in morphological structure concerns chiefly highly
differentiated and structurized protoplasms, proteins, keratin, cellulose,
and other nonliving products of protoplasmic activity rather than proto-
plasms in general. Various fibrillar structures appear in protoplasms un-
der various conditions, but many of these are temporary and disappear
completely when conditions change. Granting that proteins, cellulose,
and other organic substances are, or may become, indefinitely long chains
of chemical groups with definite polarity and symmetry or asymmetry,
that these chains may aggregate with definite orientation into larger units
(crystallites) which give evidence of crystalline structure, and that these
crystallites may also undergo further aggregation with definite orienta-
tion, there is, at present, no evidence that a definite and persistent struc-
ture of this kind is a primary and fundamental property of protoplasms.
It appears rather to be a feature of differentiation and structurization.
If this is the case, the definite orientation of molecules or chains is prob-
ably not autonomous but a reaction to something; and since these orienta-

630 PATTERNS AND PROBLEMS OF DEVELOPMENT

tions apparently originate in the course of development and differentia-
tion, they are presumably reactions to some factor of developmental
pattern. Such a factor may be local and related to cell surfaces, as the
cellulose pattern evidently is in various cells, or to local tensions or pres-
sures in a tissue, as in connective tissue, or within the cell, as perhaps in
development of the mitotic figure ; or it may be associated with organismic
pattern, particularly if that pattern is unicellular. In a multicellular pat-
tern local conditions may be more effective in determining molecular or
micellar pattern than the general pattern of the whole, though the local
factors are doubtless derivatives of the general pattern.

Most spermatozoa are certainly highly differentiated and structurized
unicellular organisms and contain little or nothing resembling even the
protoplasms of ordinary tissue cells, and still less those of embryonic cells.
If molecular pattern and orientation can become evident in morphological
pattern in any cells, it seems likely to be in spermatozoa. They are minute
and undergo perhaps a more extreme structural differentiation than any
other living cells. Many sperm heads and tails of some sperms are more
or less birefringent, but spherical sperm heads are apparently not bire-
fringent.'^ According to recent X-ray work, some sperm heads show char-
acteristics of fluid crystalline state. Separation of spermatozoan tails into
parallel fibrillae by treatment with various agents is also of interest here,
since it suggests a parallel orientation of elongated units (Ballowitz,
1890a, b; 1895). As already pointed out, however, earlier stages of sper-
matozoan development show only a general axiate pattern, presumably
involving the cell metabolism and perhaps primarily nothing more than
a gradient.

Although the ectoplasm of flagellates and ciliates appears to be less
extremely differentiated than many spermatozoa, it, or certain parts of
it, certainly undergo high degrees of structurization; and this is appar-
ently most extreme in forms with the most extreme specific spirals or
other asymmetries. The frequency of spiral patterns in spermatozoa and
axiate Protozoa and various other organisms is of particular interest.
They probably represent the reaction of many protoplasms to presence
of an axiate pattern, but we still lack definite information as to their
origin and nature. Except that they are spiral, they differ widely and
specifically in different species. The differences do not seem to have any
fundamental or necessary relation to the life of the individual. Corkscrew
heads or spirally twisted tails may be present or absent in spermatozoa,

"W. J. Schmidt, 1928, and his citations.

ORIGINS OF AGAMIC PATTERNS 631

even those of rather closely related forms; and radially symmetrical cili-
ates apparently succeed in life as well as the extreme asymmetrical forms.
It often appears as if the individual makes the best of the pattern it
possesses.

If the morphological asymmetries of these forms are expressions or
results of molecular orientations and aggregations, we still know nothing
about the relation between a particular morphological asymmetry and
a particular pattern of molecular character. It seems highly improbable,
however, that molecular or supermolecular patterns develop independ-
ently of the activities and conditions in the metabolizing protoplasm in
which they originate. In so far as they may be features of the morpho-
logical patterns of spermatozoa, protozoa, or other forms, they are evi-
dently localized and integrated in an orderly manner in relation to some
more general pattern, so that an individual with definite, species-specific
asymmetries results. That the ordering and integrating factors must in-
volve a spatial metabolic pattern seems beyond question, and the most
general evidences of such pattern are the gradients. Moreover, experi-
ment has shown so generally that these patterns do not arise autono-
mously in the cell or cell mass concerned but are induced by, or related
in a definite manner to, parental pattern or to some factor in intra- or
extraorganismic environment that it may at least be questioned whether
organismic patterns ever do originate autonomously.

PATTERNS OF CERTAIN CELL WALLS AND SURFACES

Work with the ultramicroscope and the polarizing microscrope and
X-ray analysis have led to the conclusion that cellulose consists of molecu-
lar chains of variable length. A parallel orientation of these to form a
definite pattern has been found in the cellulose layers of many plant cell
walls and in plant fibers. Microscopically visible striae in cellulose layers
of certain cell walls correspond closely with the pattern indicated by
X-ray analysis and are regarded as resulting from aggregation of the paral-
lel chains.

The cell wall of the alga Valonia is an interesting example. The Valonia
thallus is primarily a single multinucleate, more or less spheroidal, or
somewhat elongated cell, often attaining a length of several centimeters.
A layer of protoplasm adjoins the wall, and inside this is the large vacuole.
The cell wall consists of many cellulose layers, each of which shows
microscopic striations in a certain direction. Polarization and X-ray stud-
ies of this wall indicate that each layer consists of cellulose chains in

632 PATTERNS AND PROBLEMS OF DEVELOPMENT

parallel orientation, those of any one layer being inclined in direction to
those of the layer adjoining on each side at an angle averaging somewhat
less than a right angle. The chains of one set of layers constitute merid-
ional systems with reference to the two poles of the cell; those of alter-
nate sets, spiral systems centering at the same two poles. In short, the
X-ray cellulose pattern shows a definite relation to a cell polarity of some
sort.'-' The Fa/o;wa cell exhibits physiological polarity : small uninucleate
cells develop from the multinucleate protoplast in one polar region and
form holdfasts or rhizoids, and cells giving rise to thallus buds form from
the opposite polar region. This polar difference in behavior is similar to
that in many plants and animals in which a polar gradient has been
demonstrated; consequently, it is probable that a longitudinal gradient
system is present in Valonia. At any rate, the cellulose pattern
of the cell wall appears to be definitely related to, and probably deter-
mined by, a general protoplasmic pattern. Granting that such a relation
exists, it does not account for the periodicity represented by successive
layers or for the alternation of meridional and spiral orientations of cellu-
lose chains constituting the layers.

In cell walls of multicellular plants, as far as examined, the cellulose
micells or crystallites are, in general, parallel to the surface of the wall
but may be parallel or transverse to the long axis of the plant or un-
oriented. They are predominantly transverse in parenchyma cells of the
Avena coleoptile, according to the optical evidence; but a change toward
longitudinal orientation occurs when the cell wall is stretched longitudi-
nally, though not when growth elongation by action of auxin takes place.
In cell walls of the coleoptile epidermis the orientation is predominantly
longitudinal and on longitudinal stretching becomes more completely so.''*
Apparently cellulose orientations in these cells are determined by local
intracellular conditions and relations to adjoining cells. Experimental
work on effects of stretching, swelling, etc., on cellophane and plant fibers,
as well as on cell walls, indicates that mechanical factors play an im-
portant part in orientation.

Evidence of orientation of particles is also found in chitinous cuticles
and exoskeletons of various animals. An orientation of heteropolar mole-
cules vertical to surfaces or phase boundaries is another effect of local
factors in protoplasms. It does not appear, however, that the microscopic
or ultramicroscopic patterns of cell walls or protoplasmic surfaces consti-

■^ Preston and Astbury, 1937, and their citations of earlier literature.

'•• See Bonner, 1935, and his citations; also, for general discussion, Seifriz, 1936, chap. xv.

ORIGINS OF AGAMIC PATTERNS 633

tute the basis of spatial pattern of organisms. They are either patterns
of dead products of metabolism or reactions of protoplasmic molecules
to limiting surfaces, mechanical factors, and probably other factors. They
may be definitely oriented with respect to an organismic pattern already
present or a reaction to purely local conditions.

SYMMETRY IN CERTAIN MULTICELLULAR BUD PATTERNS

The terms "bud" and "budding" have often been used loosely by biolo-
gists to include forms of development representing fission rather than
budding. As regards axiate organisms, budding may perhaps be defined
and distinguished from fission as a localized activation in an organism,
resulting in development of a new polar pattern of individual, organ sys-
tem, or organ. In fissions of axiate forms the polar pattern is not an en-
tirely new pattern but a reconstitution in relation to a part of the parental
pattern. Buds apparently represent the beginnings of polar pattern. In
other types of axiate agamic development and in embryonic development
pattern is already present when development begins. In certain types of
agamic development, usually called "budding" — for example, intercalary
development of zooids in many annelids and the so-called "stolons" in cer-
tain Syllidae — the "bud" continues posteriorly the axiate pattern of the
parent. In origin it is essentially similar to a reconstitution, but it is
more or less physiologically isolated and may become a sexual individual,
differing in form from the asexual parent. Whether we regard these zooids
as products of budding or fission is perhaps a matter of opinion, for in
origin they show resemblances to both.

The only pattern at present distinguishable in early stages of the sim-
plest bud forms is a radial gradient pattern, representing the radial de-
crease of activation from a central region of primary or most intense ac-
tivation. In consequence of differential growth, or probably in some
cases differential cell migration, the radial gradient system becomes a
longitudinal axial system (pp. 16-21). There is no evidence that any
other pattern than this gradient system and its changes is necessary for
axiate development and differentiation from a bud. However, even
though the bud represents a new axiate pattern, certain features of its
pattern may be related to, and determined by, the pattern of the parent
individual. Lateral buds of many multiaxiate plants — for example, many
conifers — give rise to branches dorsiventral and bilateral in pattern in
consequence of their relation to the main axis. If the growing tips distal
to such a branch are removed, it becomes erect and radial in its further

634

PATTERNS AND PROBLEMS OF DEVELOPMENT

development. The dorsiventrality and symmetry or asymmetry of the
leaf bud and of parts of flowers develop in a particular relation to the
axiate pattern of the part on which they arise. The whole pattern of
the multiaxiate plant represents a series of such relations.

In certain coelenterates somewhat similar relations appear. Various
workers with hydra have observed that the pattern of tentacle appear-
ance on the bud of Pelmatohydra oligactis is not primarily radial but is re-
lated to the longitudinal axis of the parent hydra in such a way that

Fig. 203. — Order of tentacle appearance in bud of Pelmatohydra oligactis. Upper row,
side views of stages up to four tentacles; second row, apical views of stages from two to four
tentacles; lower row, lateral and apical views of later stages (from Rulon and Child, 1937a).

the bud is temporarily "dorsiventral."^^ The first two tentacles appear
simultaneously on opposite sides of the bud in a plane at right angles to
the polar axis of the parent; the third tentacle is midway between the
first two on the side toward the apical end of the parent; the fourth op-
posite the third; and the fifth and sixth between the first two and the
third, either simultaneously or one earlier than the other. Stages of this
development are shown in Figure 203. Gradually the tentacles become
equal in length and equidistant on the hypostome, and the dorsiventral
pattern becomes radial. Although the bud represents a certain degree of
physiological isolation and a new longitudinal axis at right angles to the
old, the parental pattern still plays a part in determining the order of

'5 Rulon and Child, 1937, and their citations.

ORIGINS OF AGAMIC PATTERNS 635

tentacle development. An interpretation suggesting that parental domi-
nance and gradient difference on the two sides of the bud are not entirely
obliterated in early bud stages has been offered (Rulon and Child, 1937,
p. 10). In some of the corals the apical zooid of the colonial axis is radial;
the lateral zooids, originating as buds from its basal region, are dorsiventral
and bilateral, with the median plane passing through the long axis of the
colony or branch. Each of these lateral zooids, like the lateral branch of a
multiaxiate plant, is apparently capable of becoming a radial apical zooid
when sufficiently isolated physiologically or physically (Wood- Jones,
1912).

Bryozoan buds represent new longitudinal axes, but their dorsiventral-
ity is determined in definite relation to pattern already present in the
region of the parent from which they arise. Statoblasts of phylactolaema-
tous bryozoa, often regarded as internal buds, develop from the funiculus,
a strand extending from the end of the aboral prolongation of the stom-
ach, the caecum, to the aboral wall of the zooecium. The funiculus is de-
scribed as consisting externally of mesoderm and internally of ectoderm,
which enters it from the wall of the zooecium in early stages. The stato-
blasts, more or less lentoid in form, develop successively from near the
caecal end of the funiculus. An axiate pattern coincident with the short
diameter and evidently determined at the time of their formation is pres-
ent, and pattern of later development has a definite orientation with re-
spect to it. The dorsiventrality of the polypid and localization of the
primary budding region are apparently also predetermined in the stato-
blast but are not evident until its development. The relation of stato-
blast pattern to the pattern of the parent zooid is not entirely clear. In
some forms — for example, Cristatella — the short diameter of the stato-
blast is said to be coincident with the long axis of the funiculus, and stato-
blast formation is a sort of strobilation of the funiculus. In other forms
(Plumatella) the short diameter is apparently at right angles to the funicu-
lar axis, and the statoblast is perhaps to be regarded as a lateral bud. In
either case it seems probable that the statoblast is primarily a bud with
a new polar axis, like other buds, but that certain other features of its
pattern are determined in relation to parental pattern.

Various types of budding appear in ascidians: some of the types of
development commonly called budding resemble fission rather than bud-
ding, and certain others approach development from cell aggregates. The
association of certain types of budding in many forms with a depression
or degeneration of a parent individual suggests physiological or physical

636 PATTERNS AND PROBLEMS OF DEVELOPMENT

isolation as a factor in the origin of buds. In the repetitive budding of
pelagic tunicates the orderly periodicity is difficult to account for, except
in terms of dominance and physiological isolations.

Apparently the longitudinal axis of the tunicate bud which originates
as a local activation represents transformation of a more or less radial
pattern into a longitudinal pattern by differential growth, as in other
buds; but dorsiventrality and asymmetry in these buds are definitely
oriented in relation to the parent pattern or to a local factor in the stolon
or other budding region. The winter buds of some forms are more like
fissions than buds."* They, or some of them, apparently reconstitute from
the part of the parent pattern persisting in them, but the reconstitution
of experimentally isolated pieces of ascidian bodies and stolons indicates
that new polarities may arise in such pieces in relation to section and
perhaps to other factors (p. 369). But whether the specific asymmetry
pattern of the ascidian can originate in relation to a new polarity and
independently of parental pattern seems not to be known.

The sponge gemmule is sometimes regarded as a sort of internal bud
but probably resembles more closely an aggregation of dissociated cells
than a bud, except that it does not contain all the differentiated tissue cells
that may be present in an aggregate. Like the aggregate, the cell mass
of the gemmule apparently does not possess a definite organismic pattern
until subjected to an environmental differential, usually that between
free surface and surface in contact, presumably an oxygen differential.
The canals center, and oscula develop on the free surface.

MULTICELLULAR AXIATE PLANT PATTERN FROM CELL AGGREGATES

In certain plants multicellular axiate patterns, often multiaxiate pat-
terns, develop from aggregates of myxamoebae, of diatoms, or of hyphae
of fungi. These patterns are orderly and definite in character, and some
of them attain considerable morphological differentiation. The features
and problems they present are, or should be, of much interest to the de-
velopmental physiologist.

DEVELOPMENT OF PATTERN IN ACRASIEAE

In this small group of plants, often regarded as more or less closely re-
lated to the myxomycetes, axiate pattern originates in definitely directed
motor reactions of separate amoeboid cells. Two forms, Polyspondylium

'^See, e.g., Berrill, 1935.

ORIGINS OF AGAMIC PATTERNS 637

and Dictyostelium, have received considerable attention /^ The following
account is based on the recent papers of Harper and Arndt. Cultured in
dung decoctions with agar or agar-bouillon and Bacillus coli, the spores
give rise to small amoeboid cells (myxamoebae) ; these move about in-
definitely, feed and divide for a time, but sooner or later begin to move
in more or less definite streams, at first in all possible directions, constitut-
ing a network of moving amoebae. According to Arndt, appearance of
directed streaming of the amoebae is associated with lack of nutrition in
the culture. As the streaming continues, centers of aggregation appear,
apparently entirely by chance, except that they are more likely to form
in dryer parts of the culture — for example, at the margins (Harper).
Harper and Arndt agree, however, that the amoebae react primarily to
a condition produced by other amoebae rather than to local differences
in substrate otherwise produced. Streaming of amoebae in the vicinity
of an aggregation becomes directed toward it; and as the aggregation in-
creases in size, the streaming becomes increasingly definite and the area
of centered streaming greater. Evidently the differential to which the
amoebae react represents a radial gradient system of variable scale, pre-
sumably due to diffusion of a substance produced by the amoebae.
Early-stage aggregations may be only temporary and may disappear com-
pletely by alteration in direction of reaction of the amoebae, or an ag-
gregation may change its position (Arndt). But once an aggregation is
well established, the amoebae continue to stream toward it and build it
up above the surface of the nutritive substrate, forming the beginning of
an axis. The further development of Polys pondylium, as described by
Harper, is of great interest. Figure 204, ^, is an approximate representa-
tion of the chief paths of streaming of amoebae toward a center, as shown
in a photomicrograph by Harper (1929). The developing axis, rising above
the substrate, is indicated at a in the figure. The amoebae continue to
arrive at the center of aggregation, move up the axis in or on the slime
secreted by those already there, and come to rest at the top, forming
an elongated cylindrical structure, the sorogen. From the basal part of
the aggregation a slender stipe forms by transformation (differentiation?)
of amoebae into cells resembling plant parenchyma with complete cessa-
tion of amoeboid activity. The stipe elongates by continued transforma-
tion of amoebae into stipe cells at its distal end. The resulting form is an

"Brefeld, 1869, 1884; van Tieghem, 1880; Olive, 1902; Potts, 1902; von Schuckmann,
1925; R. A. Harper, 1926, 1929, 1932; Arndt, 1937.

•■•*•»■ ■ f

-'Ji^ ..-."■'' ^" ?V'.-- "v?.*. "''i.. ''-V-*'-.-.....,

•^ oi -T'-Vv,. -.'{^s-. 'V.,-.;.
-■<" .^ i'^ ->' ;? * ^- T ^'V ^ • ■'.

^^

B

Fig. 204, A-E. — Development of axiate pattern in Polys pondyliiim. A, directed stream-
ing and aggregation of myxamoebae and beginning of axiate development at a; B, origin of
branches from masses of amoebae successively constricted from basal region of sorogen and
"left behind" as elongation continues; C, fully developed, unbranched individual; D, E,
branched individuals (after R. A. Harper, 1929).

ORIGINS OF AGAMIC PATTERNS 639

elongated cylindrical mass of amoebae borne on a slender stipe. This dif-
ferentiation into stipe and sorogen perhaps results from arrival of amoebae
at the summit of the aggregate more rapidly than they are transformed
into stipe cells, so that conditions in the mass at the summit become dif-
ferent from those in the developing stipe and perhaps retard the transfor-
mation.

Branching in Polys pondylium occurs as follows: At a certain stage a
constriction appears near the base of the cylindrical sorogen, and a mass
of amoebae is "left behind" on the stipe as it elongates further. In Fig-
ure 204, 5, three such masses have been isolated, and a fourth is constrict-
ing off at the base of the sorogen. Each of these masses on the stipe gives
rise to branches, usually a whorl. The branches develop in much the same
way as the main axis but at an angle to it; that is, the amoebae now
react plagiotropically. In Figure 204, C, D, and E show fully developed
individuals — unbranched (C), with two whorls (D), and with seven
whorls (£). Development of Didyostelium is, in general, similar except
that it usually does not branch. In these forms cell division is apparently
limited to the earlier stage of feeding and undirected movement of sepa-
rate amoeboid cells and has not been observed during morphogenesis.

At no stage in this development do the amoeboid cells form a Plas-
modium, as in myxomycetes. They stream and aggregate as separate
cells, and even the cylindrical sorogen consists of cells capable of resuming
amoeboid movement as separate cells, if again brought into contact with
the culture substrate, and of giving rise to a new sorogen.

As regards interpretation of the origin of this axiate pattern, Harper
holds that there is no reason "for assuming any superchemical or physical
stimuli or the action of organizing or regulating principles associated
with the multicellular organismal condition resulting from the aggregation
of the myxamoebae. The behavior of the pseudoplasmodium as a whole
is the sum of the behaviors and reactions of the individual myxamoebae"
(Harper, 1929, p. 237). Arndt does not go beyond the assertion that the
culture of myxamoebae is a harmonious-equipotential system.

While no data are at hand concerning the presence or absence of a
gradient in the developing axis, there is every reason to believe that a
gradient is present. Amoebae are being continuously added at the free
end; and when they arrive there, they are obviously in a different condi-
tion from earlier arrivals which have lost motility and have come to re-
semble ordinary plant cells. The stipe develops progressively as the amoe-

640 PATTERNS AND PROBLEMS OF DEVELOPMENT

bae continue to arrive and undergo change in condition. Presumably oxy-
gen tension is considerably lower in the interior or more basal parts of
the aggregate than at the free end. That this axis in its earlier stages is
a simple gradient in physiological condition of the constituent amoebae
and the cells developing from them appears highly probable. The change
from aqueous to aerial environment, as the aggregation rises above the
substrate, may also be a factor in bringing about the change in condition
in the cells; and its effect is probably, within limits, a function of time of
exposure. If this is the case, a longitudinal gradient in physiological con-
dition must result with the high end apical, where the active amoebae
are arriving. Moreover, the aggregation itself may determine an environ-
ment which alters the physiological condition of the amoebae; and any
such change probably progresses with time, so that older parts of the
aggregate have undergone more change than younger, and a gradient
results. In fact, it seems improbable that an aggregation such as occurs
in these forms is possible without development of a gradient pattern, and
the morphogenesis suggests such a pattern.

The successive separations by constriction of masses of amoebae which
give rise to branches suggests a dominance with limited range at the
stages concerned and successive physiological isolations, as continued ar-
rivals of amoebae in the apical region elongate the sorogen. The physio-
logically isolated mass apparently undergoes some activation, as in other
organisms, forms local aggregations, and these develop as branches. Har-
per's photomicrographs of the branch-forming masses suggest a slight
local dominance at the nodes; that is, the forms of the masses (Fig. 204,
B) and in many cases differences in size of sori and lengths of stipes in a
single whorl (Fig. 204, D, E) suggest that the branches of a whorl do not
develop simultaneously but in sequence.

This development differs in certain respects, though apparently not
fundamentally, from development of axiate pattern in cell aggregates of
sponges and hydroids (pp. 348, 418). Here the aggregation and the axiate
pattern result from a directed motor reaction and probably from differ-
ential exposure to certain environmental factors. In sponge and hydroid
aggregates aggregation supposedly results from chance contacts of cells,
and axiation is determined by a spatial environmental differential. That
either the myxamoebae or the dissociated sponge or hydroid cells possess
any inherent characteristics that enable them to originate an orderly and
definite multicellular axiate pattern independently of a chronological or
spatial differential in environmental factors is neither probable nor indi-

ORIGINS OF AGAMIC PATTERNS

641

cated by the data at hand. The early aggregate, such as in Figure 204, A,
is apparently not essentially different in pattern from an adventitious
plant bud, but its further development depends on addition rather than
on growth and division of cells.

THE PSEUDOTHALLI OF OTHER FORMS

Certain species of diatoms give rise to axiate systems resembling the
thalli of algae but consisting of diatoms, apparently not even in contact
but imbedded in a firm jelly-like substrate secreted by them. The growth
form of one of these diatomaceous pseudo thalli is shown in Figure 205.
The branches are flat and thin in the plane of the paper. Different
branches of a single pseudo thallus are approximately equal in width; and,

Fig. 205. — Part of a diatomaceous pseudothallus

as it attains a certain length, each branch undergoes dichotomous divi-
sion. This pseudothallus consists of diatoms, more or less regularly ar-
ranged in longitudinal rows in the jelly-like secretion, their long axes coin-
ciding in direction with the rows and the axiate pattern. This definite,
orderly axiate pattern must result from definite, orderly relations between
the diatoms giving rise to it. If all grew and divided at the same rate,
such a form could not arise. Evidently the chief, or only, region of growth
and division is apical; but the individual diatoms of the apical region
must be integrated in some way into a system resembling, in certain re-
spects, the growing tips of various algae, in which very similar thallus
patterns appear. Differential susceptibility and differential reduction in-
dicate a gradient in these pseudothalli with high end apical, and extending
for at least a short distance from the apical ends of the branches (Child,
19 1 9/). Some other diatoms give rise to branching systems of more or

642 PATTERNS AND PROBLEMS OF DEVELOPMENT

less orderly character in which the main axis and branches consist of the
secretion and the diatoms are in groups at the tips. Here, also, some degree
of integration and apparently some degree of dominance are involved as
regards growth and division of the different groups ; but these forms are
less regular in growth than the pseudothalli, and nothing is known con-
cerning possible physiological differences in the diatom groups of different
branches. Obviously, the patterns of these supercellular integrations can-
not be inherent in the individual diatoms but must originate in reactions
to physiological conditions determined by cells as environment of other
cells and also to the external environment.

DEVELOPMENT OF MULTICELLULAR AXIATE PATTERN IN CERTAIN FUNGI

In the mushrooms and toadstools among the basidiomycetes and in
some forms among the ascomycetes — the morel and related species — the
vegetative stage consists of indefinitely branching hyphae, each consisting
of a single series of cells; these hyphae usually form a subterranean felt-
work, the mycelium. Each hypha and branch represents an axis with
growing tip, and a gradient has been demonstrated in young, growing
hyphae of certain fungi; but hyphae of basidiomycetes have not been ex-
amined. Development of the fruiting body or sporophore, the mushroom,
toadstool, or morel, begins as an aggregation of hyphae at some region of
the mycelium. By more or less parallel upward growth of hyphae the
aggregation rises above ground as a small rounded structure. This elon-
gates by further, more or less definitely directed growth; and the form of
stipe and pileus or cap, inclosed in a loose feltwork of hyphae, the volva,
becomes distinguishable in it. In forming the pileus the hyphae change
direction of growth; and in development of "gills" or other parts from
which the basidia (modified ends of hyphae) arise and form spores, further
changes in direction of growth of hyphae in orderly and definite relation
to pattern of the whole are involved.'^ In some of the basidiomycetes
there is considerable further regional differentiation, and pigmentation
of the upper surface of the pileus is characteristic of some forms. Even
after development of pileus and stipe, every cell of Coprinus is capable,
on return to nutritive medium, of forming vegetative mycelium, and from
this, new sporophores. In certain forms evidence of dominance and physi-
ological isolation appears in the dependence of branching on inhibition of
the apical region (Brefeld).

The origin of axiate pattern in the mushroom or toadstool differs from

'* See, e.g., Brefeld, 1876; Magnus, 1906; P. Kohler, 1907.

ORIGINS OF AGAMIC PATTERNS 643

that in Polyspondylium in that pattern results from directed growth of
individual axes instead of directed movement and aggregation of individ-
ual cells, but in both cases pattern originates in the common reaction of
individual cells to some factor or factors in their environment.

CONCLUSION

It should be evident from this chapter that the problems of pattern
in development are not limited either to embryonic development or to
fissions, buds, and reconstitutions. While there are various types of de-
velopment about which we know little or nothing, the evidence from ob-
servation points, in general, in the same direction as the experimental
evidence in indicating that in some, if not in all of these, developmental
pattern is a reaction of a specific protoplasm to factors in its environment
within a parent organism or external to it and that pattern is primarily
dynamic in character rather than structural. Moreover, development of
a definite and orderly axiate pattern from directed motor reactions and
aggregation of primarily separate, amoeboid cells, from a multitude
of diatoms in a jelly-like secretion, and by directed growth reactions of
fungus hyphae suggests, like the experimental evidence from cell aggre-
gates, buds, and reconstitutions, that axiate organismic pattern is, or
may be, something different from the primary pattern of the cell or of a
protoplasm, often on a larger scale, and superimposed from without. But
whatever the initiating factor or factors, the character of development
within the pattern depends on the specific constitution of the protoplasm
concerned. Axiate pattern in the aggregates of Polyspondylium myxamoe-
bae develops as Polyspondylium; in an aggregate of Corymorpha cells, as
Corymorpha; and in a spermatid of a particular species, as a spermatozoon
of that species.

CHAPTER XVI

ORIGIN AND NATURE OF EMBRYONIC PATTERNS
THE PROBLEMS AND THE EVIDENCE

OUESTIONS of the origin and nature of pattern or "organization"
of the animal egg have been the subject of discussion and specu-
lation since the beginnings of embryological investigation. This
has been due in part to the fact that the ovarian oocyte is inaccessible to
most present experimental procedure but perhaps also in some measure
to the very general belief that embryonic development is of primary
significance and that all other types of development have only a secondary
interest. If we attempt to interpret embryonic pattern and development
in terms of what experiment has shown concerning the simpler forms of
development, we may perhaps at least come nearer agreement concerning
certain questions of embryonic pattern.

Preceding chapters have dealt chiefly with experimental evidence con-
cerning the characteristics of the spatial or regional pattern of individual
development, and the question of the origin of pattern under natural
conditions has received little attention. The earliest distinguishable pat-
tern appears to consist of regional differences of some sort on a molar or
morphological, rather than a molecular or micellar, scale. Protoplasms
are complex colloidal and crystalloidal physicochemical systems, and or-
ganisms appear to be systems of quantitatively or qualitatively different
protoplasms. The series of changes constituting what we call "living"
take place in many kinds of protoplasms, but an organism is an integra-
tion in an orderly and definite pattern of different rates or kinds of living.
The question of the origin and nature of this pattern is of fundamental
importance for our conception of the organism and of development.

If developmental pattern is primarily independent of environment,
either it must be continuously present as a spatial pattern through all
the cell divisions from one generation to another, or it must originate
autonomously in the cell or cell mass concerned. The assumption of an
"intimate structure" of some sort, adequate, ex hypothesi, as a basis for
pattern, is only a statement of the problem in terms of structure. The
Roux-Weismann theory of qualitative nuclear division states the prob-

644

ORIGINS OF EMBRYONIC PATTERNS 645

lem in terms of distribution of hypothetical hereditary units but tells us
nothing about the ordering and controlling principle determining the dis-
tribution so that an orderly organism of definite character results. It
would seem that this principle must be endowed with superhuman intelli-
gence and with knowledge of the end to be attained in order to get each
determinant into its proper cell and to supply the cell with accessory de-
terminants to provide for possible regeneration. The theory is implicitly
teleological.

Driesch, assuming that developmental pattern is independent of en-
vironment in origin, drew the logical conclusion that a metaphysical order-
ing and controlling principle, the entelechy, is necessary. Various other
authors have assumed an inherent, form-determining principle, teleologi-
cal in character. As a matter of fact, the conception of developmental
pattern as independent of environment seems to demand a teleological
principle. If we reject the hypothesis of a metaphysical origin of organis-
mic order and pattern, the question whether autonomy of pattern is pos-
sible on any other basis arises. Can the gene system alone originate a
spatial pattern in a cell or a cell mass? Or is the intimate structure of
molecular or other character, so often postulated as the basis of pattern,
an inherent property of protoplasms? If not, can it arise spontaneously,
that is, without relation to any external factor?

It was held by Rabl (1885) and others that the nucleus possesses polar-
ity coincident with the axis of the mitotic spindle, since the recurved
middle portions of the chromosomes lie toward the spindle pole. How-
ever, even if cell polarity is of this sort, it does not account for polarity
of a multicellular organism, for polarities of the two cells resulting from
division are in opposite directions. A nuclear polarity indicated by ag-
gregation of chromatin toward a certain region of the nuclear periphery
appears frequently under certain conditions or at certain stages but is ob-
viously a reaction to something outside the nucleus rather than a primary
factor of pattern. The general nuclear pattern and the chromosome pat-
terns are apparently entirely different from the pattern of the organism.
How a gene, a group of genes, or the whole gene system can originate
autonomously in the cytoplasm, a spatial pattern involving polarity and
symmetry or asymmetry, is not evident. Each nucleus supposedly con-
tains the whole gene system. "Each cell inherits the whole germ plasm"
(Morgan, 1919, p. 241). If different gene effects occur in different cells
or in different localized regions of the same cell, it seems that they must
be localized in relation to something, a differential or pattern of some sort

646 PATTERNS AND PROBLEMS OF DEVELOPMENT

independent of particular genes and of the nuclear or chromosomal pat-
tern. A discussion by Morgan (1919, pp. 241-46) under the heading, "The
Organism as a Whole or the Collective Action of the Genes," seems to
imply that organismic pattern is entirely a matter of gene action but does
not show how this is possible. In an earlier discussion a very different
view was advanced, that differentiation depends on reaction between
hereditary factors and regional differences in the egg and embryo which
are independent of them.'

With progress of experimental analysis it has become increasingly difh-
cult to separate organism and environment. Not only functional patterns
of mature individuals but developmental patterns are highly susceptible
to environmental conditions, and it is impossible to conceive an organism
except in relation to environment. If life is "the continuous adjustment
of internal relations to external relations," as Spencer put it, or a con-
tinuous equilibration in reaction to environment, then life is the behavior
of protoplasmic systems, and developmental pattern represents certain as-
pects of that behavior (Child, 1924&). Surface-interior pattern is so ob-
viously a reaction to environmental factors that it need not concern us
further here. In addition to surface-interior pattern, all but the simplest
organisms possess axiate pattern, and evolution is very largely a matter
of modifications of this pattern and of the reactions within it.

The view commonly held at present is that, whatever the nature of
embryonic pattern, it does not originate autonomously in the egg but
originates in relation to its intraorganismic environment, that is, to factors
in the relations of the oocyte to the parent organism.^ According to Gold-
schmidt, however, these factors are controlled by the gene system of the
parent. He regards the cytoplasmic pattern as a stratification, the differ-
ent levels provicUng substrata for activation of different genes. However,
cytoplasmic stratification involves quantitative or qualitative regional
metabolic differences; and these, rather than the stratification itself, are
the effective factors in development. On this basis heredity alone does
not provide a pattern for development but represents the potentialities
of the species- or individual-protoplasm ; for the realization of any of these
potentialities in development certain environmental conditions are neces-
sary. The course of embryonic development is altered when essential fac-
tors of the environment are altered; the potentialities realized under ex-

■ This view appears unaltered in a revised edition of the book by Morgan, Sturtevant, et al.,
1923 (first published in 1915).

' See, e.g., E. B. Wilson, 1925, pp. 1021-25; Goldschmidt, 1938, pp. 200-262.

ORIGINS OF EMBRYONIC PATTERNS

647

perimental conditions may give a pattern very different from that which
we call "normal."

SOME EMBRYONIC PLANT PATTERNS

Botanists have apparently been less generally concerned than zoolo-
gists with questions of the origin and nature of embryonic pattern, but
botanical literature contains many interesting and suggestive data and
raises various questions concerning pattern. A few of these are briefly
noted.

B

C

Fig. 206, A-C. — Embryonic pattern in moss and fern. A, embryo of a moss with part of
archegonium in longitudinal section (from Child, 1915c, after a preparation loaned by W. J. G.
Land); B, C, diagrammatic, indicating relations of embryonic regions in ferns to archegonium
and gametophyte, anterior end of prothallium to the right (after Coulter, Barnes, and Cowles
Textbook of Botany, 1910).

CRYPTOGAMS

The case of the alga Fuciis and related forms has been discussed in
another connection (p. 423). Among the bryophytes the polarity of the
embryo is generally coincident with the longitudinal archegonial axis, the
apical end being directed toward the neck of the archegonium. Figure 206,
A , the embryo of a moss in longitudinal section, shows this axial relation
and indicates the presence of an apicobasal gradient in physiological con-
dition by the basipetal decrease in depth of staining and increase in size
and vacuolation of cells.

Axial relations of the embryo to the archegonium and to the gameto-
phyte differ in different groups of the pteridophytes, suggesting that in
some the archegonium may be the factor determining the primary pattern,

648 PATTERNS AND PROBLEMS OF DEVELOPMENT

in others the axiate pattern, of the gametophyte. The gametophytes of
most ferns are flattened, axiate prothalha with an apical growing cell or
cell group, with dorsiventral differentiation, and with archegonia develop-
ing on the ventral side.^ The first division of the zygote is in a plane trans-
verse to the axis of the archegonium in most forms, but in Polypodiaceae
it is in a plane passing through the axis of the archegonium and transverse
to the long axis of the pro thallium, that is, in definite relation to its en-
vironment. When it is transverse to the archegonial axis, the second divi-
sion plane is at right angles to it; when it passes through the archegonial
axis, the second plane also passes through that axis at right angles to
the first, and the third is transverse. The four quadrants or four pairs of
octants resulting from these divisions are said to represent more or less
exactly four regions of the developing plant — cotyledon, stem, root, and
foot, the latter a temporary nutritive organ connecting the embryo with
the gametophyte. The fern embryo is then temporarily a bilateral form,
but the relation of the four regions to archegonial and gametophyte pat-
tern differs in different groups. For example, the diagrammatic Figure
206, B, indicates the relation believed to exist in certain forms (order
Marattiales), the foot developing from the two unshaded quadrants
next to the neck of the archegonium, stem, cotyledon, and root, from the
two shaded quadrants. The relations in the Polypodiaceae are indicated
in Figure 206, C, a lateral view of the eight-cell stage. The two dorsal
anterior cells (upper shaded quadrant of the figure) give rise to stem;
the ventral anterior pair, to cotyledon; the dorsal posterior pair, to
foot; the ventral posterior pair, to root. In Figure 206, B, the embryonic
pattern is apparently related to the archegonial axis, perhaps secondarily
to the longitudinal axis of the gametophyte; in Figure 206, C, the re-
lation is somewhat different, but since it is apparently constant it is
significant. In Equisetum the first division is transverse to the arche-
gonial axis, the cell next to the base forming the foot or the foot and root,
the other cell, stem and cotyledon.

In Lycopodium and Selaginella the first division is transverse to the
archegonial axis; and the cell next to the neck, or a descendant of it, be-
comes the suspensor, an embryonic organ not present in other pterido-
phyte groups; and cotyledon and stem develop from cells next to the
base of the archegonium; but in later development the embryonic axes
undergo change in direction, apparently in relation to the gametophyte

■' Archegonia are on the dorsal side of the prothallia in Ophioglossales, and in Marsilea
the single archegonium is apical on the apparently radial gametophyte.

ORIGINS OF EMBRYONIC PATTERNS 649

pattern. The development of Selaginella affords an interesting example
of these changes. The megaspore has an apicobasal axis; and development
of the apparently radially symmetrical megagametophyte is at first lim-
ited to the apical region, the basal region which contains much starch
becoming cellular only later. The apical region of the gametophyte in
which archegonia develop is exposed by rupture of the spore wall over it,
but most of the gametophyte remains within the spore wall (Fig. 207, A).
Archegonial and gametophyte axes are in the same direction. The first
division of the zygote in S. martensii is transverse to these axes, as indi-
cated in the two-cell stage in Figure 207, A (Bruchmann, 1909). From
the cell next to the neck of the archegonium the suspensor develops and
elongates, pushing the embryo deep into the gametophyte tissue; the
other cell gives rise to the embryo. This embryonic cell divides vertically,
and each resulting cell again vertically. Two of the embryonic cells are
shown in Figure 207, B. From one of the four cells resulting from the
first two divisions the apical cell of the stem is separated by a diagonal
wall (a of Fig. 207, C). As development progresses, the embryonic axis
gradually bends to one side (Fig. 207, D) and finally more or less toward
the apical region of the gametophyte, the foot develops toward the base
of the gametophyte, and the rhizophore opposite the stem (Fig. 207, E).
At this stage the embryo with suspensor and foot is a bilaterally sym-
metrical individual, but how change in direction of the embryonic axis
which results in definition of a median plane is determined seems not to
be known. The gametophyte is apparently radial; and, according to
Bruchmann, gravity is not concerned. Conceivably the change in form
from that of Figure 207, C, to that of Figure 207, E, may result from a
reaction of the stem axis to an apicobasal differential in the gametophyte,
or it may be associated with development of the large foot. But whatever
the factors concerned in these changes in form and direction of growth,
it seems evident that the pattern of early development shows a definite
and constant relation to archegonial or gametophyte pattern or to both.
The gametophyte of Isoetes is somewhat similar to that of Selaginella,
but relations of embryonic regions are apparently different. The first di-
vision of the zygote is transverse or inclined to the archegonial axis,
the cell toward the base becoming foot, the other embryo, suspensor
being absent. According to this account, the embryonic axis is oppo-
site in direction in relation to its gametophytic environment to that
of Selaginella. Perhaps early determination of the stem axis in Selaginella
and apparently much later determination in Isoetes are concerned in this

Fig. 207, A-E. — Stages in embryonic development of Selaginella niartensii; all figures
similarly oriented with respect to archegonium and gametophyte. A , outline of gametophyte
with advanced embryo and two-cell stage and a group of rhizoids at upper left, spore coat
still surrounding most of gametophyte; B, early embryo, suspensor and cells of archegonial
neck; C, later embryo, showing apical cell, a, of stem; D, still later stage in longitudinal sec-
tion; E, outline of advanced stage in median section; s, suspensor; r, rhizophore; /, foot; a,
apical cell of stem (after Bruchmann, 1909).

ORIGINS OF EMBRYONIC PATTERNS 651

difference. The parts of the Isoeies embryo are apparently determined in
relations not greatly different from those attained in Selaginella at the
stage of Figure 207, E.''

SPERMATOPHYTES

The gametophyte of the seed plants does not become an independent
individual but is represented by cells of the embryo sac developing from
the megaspore within the ovule. Its axiate pattern coincides in direction
with that of the ovule; and the ovule originates much like a burl but be-
comes surrounded by one or more integumentary layers with an opening,
the micropyle, at the free end. Earlier stages of gametophyte develop-
ment usually consist of nuclear division without formation of cell walls
in the cytoplasm of the megaspore within the developing ovule, cell
walls, in so far as they occur, appearing later.

Gametophytes of gymnosperms fcycads, conifers, etc.) consist finally
of many cells, and in most members of the group the eggs form within
archegonia; but there appears to be a progressively earlier individuation
of the egg in the evolution of the group, and in the most highly developed
forms archegonia do not appear. When archegonia are present, they are
usually at the micropylar end of the gametophyte with axes parallel to
the gametophyte axis, but in certain gymnosperms they appear elsewhere.
It has been suggested that their position is determined in reaction to the
position of the entering pollen tube.'^ In embryonic development the ear-
lier divisions of the zygote give rise to a proembryo. The "basal" region
of the proembr>^o, that is, the part toward the micropylar end of the
ovule or the neck of the archegonium, becomes the suspensor; and in some
forms other nonembr^'onic cells develop from it. The embryo develops
from the "apical" region of the proembr\'0 and stem tip, and cotyledons
from the apical region of the embryo. The suspensor usually elongates,
in some forms enormously (Fig. 208, A), and carries the embryo into the
nutritive endosperm of the gametophyte. But in Gingko it is massive and
many-celled; only the small-celled polar region is concerned in embr}^o
formation; and stem tip and cotyledons develop later from its apex (Fig.
208, B). In the embryo of Pinus (Fig. 208, A) only the small cells at the
tip of the suspensor form the embr\'0, and stem tip and cotyledons de-
velop from its apical region.'' An apicobasal gradient, indicated by dif-

t For more complete data on embryonic development of bryophytes and pteridophytes see
G. M. Smith, 1938; Goebel, 1930; Campbell, 1918; Coulter, Barnes, and Cowles, 1910.

5 Coulter and Chamberlain, 1910, pp. 420-21.

' In Fig. 208 and the following figures relating to spermatophyte development orientation
is uniformly with embryonic pole upward.

652 PATTERNS AND PROBLEMS OF DEVELOPMENT

ference in cell size, rate of division, vacuolation, staining properties (Fig.
208, B), or some of these differentials, very commonly appears in the
proembryo or later, often developing gradually. Although the physiologi-
cal factors concerned in determination of these axiate patterns in gymno-
sperms are not known, the patterns constitute a definite, orderly series,
each new pattern originating in definite relation to a pre-existing pattern.
That this relation indicates a physiological relation can scarcely be
doubted.

Fig. 208, A, 5.— Embryonic stages of gymnosperms. A, greatly elongated suspensor cells
of Pimis with embryo at tip (after Coulter and Chamberlain, Morphology of Gymnosperms
[1910]); B, proembryo of Ghigko (from Child, 1915c, reproduced from H. L. Lyon, 1904).

Development of the typical megagametophyte of angiosperms is dia-
grammatically outlined in Figure 209. Following the first nuclear division
of the megaspore, the two nuclei come to lie near the two ends of the
young embryo sac (Fig. 2og,A); each of these divides again (B), and each
of the four once more, giving four nuclei at each end of the sac (C). One
nucleus from each group migrates toward the middle of the sac; the two
usually fuse sooner or later, forming the endosperm nucleus. More or less
definite cell boundaries develop about the three nuclei at the micropylar

ORIGINS OF EMBRYONIC PATTERNS

653

pole, the three resulting cells being the egg and a synergid on each side of
it (Fig. 209, D). The fate of the nuclei at the opposite antipodal pole
varies in different families and species. They, or even the primary antip-
odal nucleus (Fig. 209, A), may degenerate; or more or less definite cell
walls may develop about them, and these cells may become very large
and undergo differentiation as nutritive organs or increase in number.
Axiate pattern of the embryo sac is primarily coincident in direction with
that of the ovule.

Fig. 209, A-D. — Diagrammatic outline of angiosperm gametophyte development. A-C,
nuclei of first, second, and third divisions; D, egg between two synergids at lower pole in figure,
antipodal nuclei at upper pole and members of primary endosperm nucleus in middle.

In development of the zygote the embryo is formed from the region of
the egg which protrudes into the embryo sac, the ''free" pole (upward in
Fig. 209, D), the suspensor from the region toward the wall of the sac.
Cotyledons and stem tip of dicotyledonous angiosperms develop from the
"free" pole of the embryo. Figure 210 shows the axiate pattern of a
dicotyledonous embryo and suspensor. The eight-cell stage of the embryo,
of which four cells are shown in Figure 210, B, consists of four terminal,
or apical, and four basal cells. From the four apical cells cotyledons and
stem tip develop. The plane in which the two cotyledons develop is ap-
parently also determined by factors external to the embryo.

The primary axis of the monocotyledonous embryos is oriented like

6S4 PATTERNS AND PROBLEMS OF DEVELOPMENT

that of dicotyledonous forms (Fig. 211, A-C); but in most members of
this group the single cotyledon develops from the terminal region, and
the stem tip from cells lying laterally (s of Fig. 211, C). Presumably the
side of the embryo on which the stem tip develops is determined by some
factor in the embryonic environment, but thus far no consideration of this
point has been found in the botanical literature consulted.

A B C

Fig. 210, A-C. — Three stages in early development of embryo and suspensor of a dicotyle-
donous angiosperm, Capsella biirsa-pastoris. A, two-cell stage of embryo; B, one side of eight-
cell stage; C, later stage (after Coulter and Chamberlain, Morphology of Angiosperms, 1903).

Polyembryony occurs in many spermatophytes with development of
embryos, not only from single early cells of the primary embryo but from
other cells of the gametophyte and even from cells of the nucellus, which
is sporophyte tissue surrounding the gametophyte. In various gymno-
sperms polyembryony is normal or usual. In the pine, for example, each
of the first four embryonic cells, products of two vertical divisions, has
a suspensor cell, and each may form an embryo; or a single embryo may
develop from all four. In other conifers also, single cells of early embry-

ORIGINS OF EMBRYONIC PATTERNS

655

onic stages may give rise to whole embryos, some degree of dissociation
being apparently the initiating factor. The proembryo of Ephedra forms
by nuclear division without cell formation up to eight nuclei, after which
cell formation occurs, and each of the eight cells gives rise to an embryo
and suspensor.

Among angiosperms numerous cases of polyembryony have been re-
corded. In some there is multiphcation of embryos by budding from the

Fig. 211, .4 -C— Three stages of embryo and suspensor of a monocotyledonous angiosperm,
Sagittaria variabilis. A, B, early stages with primary axis oriented as in dicotyledonous forms,
one synergid below embryo at left; C, later stage, the single cotyledon from the terminal re-
gion, the stem tip from the lateral cells, 5 (after Schaffner, 1897).

suspensor, but usually only one persists. In these cases all embryos de-
velop from cells of the zygote. In others a synergid may be fertilized and
give rise to an embryo or may develop without fertilization. Embryos
may develop from antipodal cells and apparently also from an endosperm
cell. They may also originate as buds from nucellar tissue outside the
embryo sac, more commonly from the region about the micropyle than
elsewhere. These cases of polyembryony present several points of inter-
est. First, they show that in at least some species the cells resulting from
division of the egg are not determined, or not irreversibly determined,

6s6 PATTERNS AND PROBLEMS OF DEVELOPMENT

in early stages as particular embryonic regions or parts. Second, the fact
that embryos may develop not only from the egg but from other cells of
the embryo sac shows that the distinctive characteristics of the egg are
not essential to embryonic development and also suggests more or less
dedifferentiation in embryo formation from these other cells. Third, it is
evident from embryo formation by diploid cells of the nucellus, cells of
the gametophyte being haploid and the fertilized egg diploid, that embryo
formation is independent of these chromosomal differences. Fourth, the
data of polyembryony suggest that conditions in, or associated with, the
embryo sac are factors in determining the embryonic type of develop-
ment, whether from the egg or other cells of the gametophyte or sporo-
phyte, and that there is usually a differential in these conditions such that
the egg forms at the micropylar end of the sac; the more frequent develop-
ment of embryos from nucellar tissue about the micropylar region is also
suggestive in this connection. '^

There are also indications that the micropylar region of the gameto-
phyte is or becomes physiologically dominant, the degree or effectiveness
of this dominance differing in different species and probably under dif-
ferent conditions. In the light of the data on polyembryony, development
of a single embryo from the egg alone suggests a dominance effective in
inhibiting embryonic development of other cells. The fact that in many
cases of polyembryony the egg is not fertilized or, if fertilized, fails to
develop or ceases development at an early stage suggests that in these
cases dominance is largely lacking and that physiological isolation may
be a factor in embryonic development of other gametophyte cells than
the egg. If dominance of the gametophyte is slight or lacking, cells of the
nucellus, usually from the micropylar region, may attain dominance and
give rise to embryos. It is perhaps of some interest to note that the
spermatophyte embryo resembles a bud in certain respects. It develops
not from the whole egg but from one polar region of it. Is there any es-
sential difference between bud formation and development of the embryo
from the small active cells about one pole of the Gingko proembryo (Fig.
208, B)? According to this suggestion, the proembryo appears more nearly
comparable to the animal embryo, and the plant embryo represents the
first bud of the sporophyte, originating from what was primarily the high
end of the gradient system of the proembryo.

And finally, in development of the gametophyte, position and polarity

7 For more complete data on spermatophytes see Coulter and Chamberlain, 1903, 1910;
Coulter, Barnes, and Cowles, 1910; Goebel, 1922.

ORIGINS OF EMBRYONIC PATTERNS 657

of the egg, and polarity of the embryo there is definite and characteristic
relation of each step in the development of pattern to the pre-existing
pattern. The data suggest that polembryony is associated with alteration
of the physiological relations that determine the usual sequence of events.

POLARITY OF ANIMAL EGGS IN RELATION TO THEIR ENVIRONMENT

The chief axiate pattern of animal eggs, so-called "polarity," may be
morphologically evident in a graded distribution, or a more or less sharply
defined separation of cytoplasm and yolk or other substance, in position
of nucleus or region of polar body formation, or it may be independent of
all these. In some eggs it is apparently present or, under experimental
conditions, persists only in the superficial cytoplasm or cortex; in others
it may appear as a graded distribution of materials throughout the egg;
in some eggs it is perhaps not estabHshed until the polar bodies form.

Ventrodorsality or dorsiventrality may also be visible in some eggs, at
least after fertilization— for example, in the mesoplasm of the Styela egg
(p. 577) and the gray crescent of the frog egg — and often, when it is not
visible, experiment indicates its presence even before fertihzation. In
some forms, however, ventrodorsality becomes evident only in the course
of development; and experiment does not give conclusive evidence of its
presence in the unfertilized egg, though some physiological basis for it,
presumably a slight regional difference of some sort, is probably present.
This seems to be the condition in the sea-urchin egg. In ascidian eggs
without visible dorsiventrality development of isolated egg pieces gives
evidence of a dorsiventral pattern. The specific asymmetries of later
stages of various animals are not usually evident in undivided eggs with
methods at present available, though difference in developmental be-
havior may indicate that difference of some sort is present.

All lines of evidence indicate that patterns of animal eggs are relatively
simple as compared with later stages, that is, they represent the most gen-
eral features of developmental pattern. This general pattern apparently
constitutes a sort of physiological co-ordinate system in relation to which
further development of pattern takes place.

Relations of the developing oocyte to its intraorganismic environment
differ widely in different animals, and the question of conditions in the
oocyte environment as possible factors in determination of its pattern is
best considered according to the character of this relation.*

« In an extended description of oocyte development in invertebrates Korschelt and Heider
(1902) distinguish diffuse and localized egg formation and in both of these types, solitary

658 PATTERNS AND PROBLEMS OF DEVELOPMENT

ISOLATED DEVELOPMENT

Oocytes of certain annelids separate from the ovary at an early stage
and pass their whole growth period as free cells in the coelomic fluid. The
oocyte of Arenicola crislata, an example of this type, is, when full grown,
flattened to a shape approaching a biconvex lens, the axis of flattening
apparently becoming the polar axis. The egg also differs somewhat in
diameter in two directions at right angles to each other and in a plane
vertical to the short axis, and these two diameters are probably parallel
to the first two cleavage planes. There is no visible regional differentia-
tion, not even localization of yolk.^ After separation from the ovaries
the environment of these oocytes is presumably not differential in any
definite or persistent way, since the waves of peristaltic contraction of the
body wall keep them almost continuously in motion. Nothing is known
of origin or nature of the primary developmental pattern of this egg.
For those who beheve that spatial organismic pattern is an inherent prop-
erty of a species-protoplasm or that it can originate in a single isolated
cell without any relation to external factors, the Arenicola egg provides
an apparent example of such origin. However, the possibility remains
that the basis of pattern may be established before isolation from the
ovary and that it persists because the uniform environment of the free
cell provides no differential to alter it.

IMBEDDED DEVELOPMENT

Coelenterates and sponges present examples of this sort of relation
between oocyte and body, the oocyte developing as a naked cell in the
tissues of the parent body and evidently obtaining nutrition from sur-
rounding cells or from intercellular fluids. These oocytes are often amoe-
boid in earlier stages and may migrate. In the hydrozoa, oocyte develop-
ment takes place in the ectoderm; in scyphozoa and anthozoa, in the
entoderm; and in sponges, in the mesogloea. Usually it is more or less
locahzed in certain regions of the body — for example, on the radial canals
or the manubrium in hydromedusae, on the mesenterial borders in an-
thozoa— but an ovary, as a special, definitely localized and bounded or-

development, that is, without special nutritive or other accessory cells, and alimentary develop-
ment, with relation to nutritive cells; when nutritive cells surround the oocyte development is
follicular, when they are single or several in a group it is ntitrimentary. The present considera-
tion, being primarily concerned with the question of the possible role of conditions of oocyte
development in determining pattern, employs a somewhat different classification of develop-
mental types.

' Child, 1900; A. marina, Child, 1898.

ORIGINS OF EMBRYONIC PATTERNS

659

gan, is not present. In many, if not in all, of these forms the oocyte is
subjected sooner or later in its development to an environmental differen-
tial. One pole of the ectodermal hydrozoan oocyte comes to be separated
from the water only by a very thin membrane and often protrudes from
the body surface as growth progresses (e.g., Hydra), while the other pole
is deeply imbedded and close to the entoderm, presumably the chief
source of nutrition. Usually the nucleus comes to lie near the outer pole
(Fig. 212, A, B). In the medusa Phialidium, in which the growing oocytes
form a columnar epithelium along the radial canals (Fig. 212, 5) and are

A C D

Fig. 212, A-D. — Hydrozoan oocytes. A, oocyte of Corymorpha in manubrium of medusa
bud with nucleus close to outer pole; 5, Phialidium oocytes, forming columnar epithelium; C,
D, gradient of Phialidium oocyte (from Child, 1925(7).

attached by their inner poles, it has been shown that a gradient, indicated
by susceptibility and by reduction, is present in the advanced oocyte
(Fig. 212, C, D) and that the outer pole becomes the apical pole of the
embryo (Child, 1925a). The data suggest a gradient determined by greater
respiratory exchange at the outer, and uptake of nutrition at the inner,
pole. There may be two opposed overlapping gradients in concentration
of different substances, but the data available indicate one activity gra-
dient decreasing basipetally. Entodermal oocytes of some scyphozoa and
anthozoa attain somewhat similar relations to the coelenteric cavity. In
certain scyphozoa the entodermal epithelium in contact with a region of
the growing oocyte becomes thickened, forming a "cell crown" regarded
as nutritive in function, and the oocyte nucleus is at this pole (0. und R.

66o PATTERNS AND PROBLEMS OF DEVELOPMENT

Hertwig, 1880). The Phialidiutn oocytes are almost pedunculate (Fig.
212, C, D), and the oocytes of some anthozoa do become pedunculate
at relatively early stages.

PEDUNCULATE DEVELOPMENT

Oocytes of many invertebrates become pedunculate sooner or later in
the growth period. The peduncle may develop from a regional attach-
ment of the oocyte itself or from follicular or other cells. The oocytes

A B

Fig. 213, A, 5.— Early and full-grown oocyte of Sternaspis sciUaia, the early stage showing
the blood vessel at the attached pole (from Child, 19156).

are primarily cells of an epithelium ; development of a peduncle is associ-
ated with their growth and protrusion into the ovarian cavity. An ex-
treme case of pedunculate development appears in the gephyrean Sternas-
pis. The peduncle attains considerable length, and a blood vessel extends
throughout its length, forming a loop in the cytoplasm of the attached
pole of the oocyte (Fig. 213, ^). From an early stage on, the nucleus lies
at the free pole (Fig. 213, 5), the polar bodies form there, and it becomes
the apical pole of an embryo with spiral cleavage.

Echinoderm oocytes show various degrees of peduncle development,
ranging from regional attachment with little elongation to peduncles of
considerable length. In the case of the sea urchin the attached pole be-

ORIGINS OF EMBRYONIC PATTERNS 66i

comes the apical pole of the embryo, according to Boveri (1901a, b);
the free pole is apical, according to Jenkinson (191 16). Recently the find-
ing of maturation spindles at the free poles in two species seems to have
settled this question in favor of the free pole as apical (Tennent, 1931;
Lindahl, 19326). There was a similar difference of opinion concerning
holothurians, but here also it has been found that the free pole becomes

apical/"

The nucleus of the asteroid oocyte usually comes to lie close to the
surface of the cell at some region nearer the free than the attached pole,
but without other definite relation to the free pole." Polar bodies form
in Asterias in the region where the nucleus is nearest the cell surface

Fig. 214.— Positions of polar body in attached eggs of the starfish Paliria miniata (from
Child, 1936a).

(Wilson and Mathews, 1895). According to Yatsu (1910a), eggs in which
free and attached poles can be distinguished by shape of the egg and
presence of follicular membrane invariably form polar bodies halfway be-
tween the equator and the rounded free pole. Polar-body formation in
Patina eggs still attached shows no constant relation to attached or free
pole but depends entirely on position of the nucleus, and this appears to
be determined by chance conditions, except that it is commonly nearer
the free than the attached pole and may sometimes be at the free pole
(Fig. 214). The diiJerential between attached and free pole may be less
in the starfish than in the sea urchin, or the nucleus may be less sensitive
to it, or it may be that the nucleus comes to lie near the surface where the
oocytes are less closely packed together and there is more circulation of
fluid. On the other hand, it is possible that the nucleus lies where it

" Gerould, 1896; Theel, 1901; Oshima, 1921; Inaba, 1930.
^' Asterias forbesii, Paliria miniata, author's observations.

662 PATTERNS AND PROBLEMS OF DEVELOPMENT

happens to be pushed by the accumulation of substance during growth
and that, if a polarity is determined by the differential between free and
attached pole, it is more or less completely obliterated by the activation
associated with polar-body formation. At the time of maturation a dis-
tinct dye-reduction gradient, decreasing from the region of polar-body
formation, appears, primarily in the egg cortex (Child, 1936a).

Regional ovarian attachment of the oocyte, often with development of
a peduncle, is characteristic of many other animals, perhaps most com-
monly among mollusks and annelids; and, as far as the relation has been
determined, the free pole becomes apical in most of these."

Oocytes of A scar is megalocephala are regionally attached to a rhachis
extending through the tubular ovary, the attached pole being more or
less pointed, the free pole rounded; but they usually become spherical
before maturation, and attached and free poles are not distinguishable.
Moreover, position of the second polar body, the only one adhering to the
egg surface, varies in relation to plane of cleavage, whether because posi-
tion of formation is variable or because its position changes after forma-
tion. Also, giant eggs formed by fusion of two may develop normally.
Zur Strassen (1906) and Boveri (1910a, b) believed that polarity in these
eggs is finally established late, after maturation or shortly before first
cleavage. Recently, however, it has been found that in certain eggs with
pear-shaped shells, presumably retaining the form of the ovarian oocyte,
polar-body formation usually occurs at the blunt pole, the free pole of the
oocyte, but may occur elsewhere. Moreover, the region of polar-body for-
mation, even when not at the free pole of the oocyte, may apparently
sometimes become the animal pole. These observations suggest that the
free pole normally becomes the animal pole and determines polar-body
formation there but that, when presumably abnormal conditions deter-
mine polar-body formation elsewhere, the ovarian polarity may be ob-
literated and a new polarity determined in relation to maturation.'^

A multicellular peduncle develops from epithelial cells of the ovary in
Limulus (Munson, 1898) and in arachnids (Balbiani, 1873). According
to Conklin (1932), the oocyte of Amphioxus and probably the ascidian
oocyte are attached by the pole which becomes apical or animal.

'2 E.g., the free pole is apical in the nemertean Cerehratulus (C. B. Wilson, 1899; E. B.
Wilson, 1903); in the mollusks Cyclas (Stauffacher, 1893), Unio (F. R. Lillie, 1895),
Musculium (Okada, 1935), and Dentalimn (E. B. Wilson, 1904); in the annelid Chaetopterus
(F. R. Lillie, 1906), and in Sagitta (Stevens, 1904).

'3 Schleip, 1924; 1929, pp. 230-33. See also Boveri, 19106.

ORIGINS OF EMBRYONIC PATTERNS

663

In general, the peduncle probably plays a part in the transport of
nutritive material to the cell. The oocyte often shows a special structure
radiating from it (Fig. 213, A), but among invertebrates the nucleus often
lies nearer the free pole. Respiratory exchange and elimination of prod-
ucts may be more rapid there. In forms with a well-developed circulatory
system and with oxygen supplied chiefly by the blood respiratory ex-
change may perhaps be greater at the attached pole. At any rate, condi-
tions are doubtless more or less different at free and attached poles; that
these differences may deter-
mine egg polarity in many cases
is indicated by the evidence.

The oocyte of the frog is sus-
pended from the ovarian wall in
an epithelial sac, the theca; be-
neath that is a thin follicular
membrane. The theca forms a
short peduncle, an artery and
vein develop in it, and a much
branched capillary circulation
surrounds completely the de-
veloping egg cell. In 75-80 per
cent of full-grown oocytes taken
at random from different ovari-
an regions the boundary be-
tween pigmented and unpig-
mented zones is within 20° of the peduncle. Injection of the vessels with
colored masses and direct observation of movement of corpuscles shows
that usually the greater part of the arterial circulation is over the pigmented
hemisphere (Fig. 215). There is, however, considerable variation, and
cases have been found in which the arterial circulation is largely over the
unpigmented hemisphere (Bellamy, 1919, 192 1). Nevertheless, the high
frequency of close correspondence between arterial circulation and the
region of the oocyte in which pigment develops, the apical or animal
hemisphere, is far above probability. Moreover, it is probable that full-
grown oocytes may change position in the theca in consequence of me-
chanical conditions resulting from the ovarian contractions or other
movements of the animal, for they are not in any way attached to it.
Bellamy has also suggested the possibility that the circulation may change

Fig. 215. — Full-grown ovarian oocyte of frog
{Ranapipiens), showing the usual relation between
arterial and venous circulation in the theca and
pigmented and unpigmented regions (from Bellamy,
1919).

664 PATTERNS AND PROBLEMS OF DEVELOPMENT

during the development.'^ The possibility that the "most arterial" region
where the artery enters the follicle and begins to branch becomes dorsal
may be noted. A study of the development of circulation in relation to
pigment development in earlier stages is needed to clear up the question.
Meanwhile it appears probable that polarity of the frog egg is determined
by the circulation differential.

FOLLICULAR DEVELOPMENT

Both imbedded and pedunculate types of oocytes in many animals
are surrounded by a follicle, at least in early stages of growth. Cells ap-
parently nutritive in function aggregate about the oocytes of certain
sponges; epithelial cells may be stretched into a thin follicular membrane
which ruptures and disappears sooner or later as the oocyte grows (Fig.
213, A; also various echinoderms), or may persist until it is full grown
(holothurians). The temporary follicular membranes probably have little
or no special function in some forms but are merely cells so situated that
they are stretched into a membrane by growth of the oocyte. In some
groups, however — for example, cephalopods, some insects, and chordates
— the follicle becomes an epithelium of considerable thickness, consisting
of a single cell layer or, as in most chordates, several cells thick. Since it
surrounds the oocyte, this must obtain nutritive material from or through
it. The follicular epithelium of cephalopods develops numerous folds
which grow into the cytoplasm of the oocyte. From the ascidian follicle
the "test cells" pass into the oocyte. The follicle may also secrete a chori-
on. Since the more highly developed follicles usually surround the oocyte
uniformly, they apparently do not provide an environmental difierential.
The growing oocytes of chordates within their follicles usually approach
and protrude from the surface of the ovary, and in many forms the follicle
becomes more or less pedunculate. This course of development may sub-
ject the oocyte to a differential between ovary and body cavity. The
possibility that the follicular circulation may provide a differential de-
terminating polarity in the case of the frog egg was noted above. The

'•< Bellamy has sometimes been cited as having abandoned, in the second paper (1921), the
view advanced earlier that polarity might be determined in relation to the circulation. He
did retract his earlier statement that "in every case observed the greater part of the arterial
blood supply was restricted to the pigmented hemisphere," because he found some cases in
which this was not true; but he emphasized the point that these concern only full-grown or
nearly full-grown oocytes and that they do not affect the earlier statement that the peduncle
is within 20° of the equator in 75-80 per cent of full-grown eggs with arterial circulation
largely over the pigmented hemisphere.

ORIGINS OF EMBRYONIC PATTERNS 665

follicular epithelium surrounding each oocyte in insect ovarian tubules
is probably a part of a longitudinal gradient in the whole tubule.

DEVELOPMENT WITH ACCESSORY OR NUTRITIVE CELLS

Oocytes of many animals grow at the expense of other cells, the so-
called "accessory" or "nutritive" cells. These may be more or less com-
pletely resorbed during growth of the oocyte. Very commonly the ac-
cessory cells are from the same region of the ovary as the oocyte and
are usually believed to possess the same potentialities as the cell that
becomes an egg. If this is the case, some factor in the environment or
relations to each other of the cells must determine their respective fates
as oocytes or accessory cells. Nutritive cells vary from one to a consider-
able number for each oocyte, but number and position in relation to the
oocyte are usually more or less definite for the species.

Association of oocytes and accessory cells is frequent among polychete
annelids. The oocyte of Ophryotrocha is accompanied by a single cell,
larger than the oocyte in early stages (Fig. 216, A, B)y= The paired cells
separate from the ovary and become free in the body cavity at an early
stage. Braem concludes that cells exposed to body fluid at the surface of
the ovary become accessory cells, each serving as a source of nutrition to
the cell immediately beneath it; the two cells become attached to each
other, the one originally below the ovarian surface becoming oocyte. In
the single observed case of polar-body formation with attached nutritive
cell, this cell was at the basal pole (Braem, 1894).

The primitive germ cells of the polychete Diopatra migrate from the
primary into a secondary ovary; there they divide to form chains, con-
sisting of a variable number of cells. In general, cell size decreases from
the middle toward both ends of the chain; one cell of the middle region
becomes larger than the others, protrudes from one side of the chain, and
develops as oocyte, all other cells of the chain being nutritive (Fig. 216,
C, D). As the oocyte grows with progressive ingestion of nutritive cells,
it protrudes increasingly from one side of the chain, so that finally remain-
ing nutritive cells of both terminal portions of the chain come to lie to-
gether at one pole (Fig. 216, E)\ this becomes the basal pole (Lieber,
193 1). Apparently there is some degree of physiological integration in
the chain; the gradation in cell size from the middle suggests a differential
of some sort in both directions as a possible factor determining which cell

'5 Braem, 1894; Korschelt, 1894.

666

PATTERNS AND PROBLEMS OF DEVELOPMENT

becomes oocyte. The data give no evidence for or against predetermina-
tion of the region of the oocyte which protrudes from the chain. Appar-

FlG. 2i6, A-F. — Association of oocytes and nurse cells in certain annelids. A, B, early and
late stage of oocyte development in Ophryoirocha puerilis (after Braem, 1894); C, D, E, early,
more advanced, and late stages of oocyte and accessory cells of Diopatra (after Liever, 193 1);
F, eight-cell group (one oocyte, seven accessory cells) of To7ttopteris (after Chun, 1888).

ently it may be on the convex side of a sUght curvature or be determined
by mechanical relations between the cells. If this is the case, direction
of the polar axis in the oocyte is a matter of chance. Two oocytes occa-
sionally develop in a chain but are separated by a number of nutritive

ORIGINS OF EMBRYONIC PATTERNS 667

cells, as if each, or the region of the chain in which it formed, were to some
degree dominant over a certain length of chain. Somewhat similar cell
chains develop in ovaries of certain parasitic copepods."^ In some of these
the oocyte develops at one end of the chain; and when it separates, the
next cell becomes an oocyte, etc. In others the oocyte develops elsewhere
in the chain, but whether polarity is determined in relation to the acces-
sory cells seems not to be known.

Two opposite regions of the Myzostoma oocyte, each associated with a
single nurse cell which is completely incorporated at an early stage, are
believed to become respectively apical and basal poles (Wheeler, 1896,
1897). Whether position of the accessory cells results from a pre-existing
polarity of the oocyte or oogonium or determines polarity is not clear.
If the nurse cells are alike, as they appear to be, their association with the
two poles of a predetermined heteropolar pattern is puzzling; on the other
hand, how they can determine a heteropolar axis if they are alike is not
evident. If they differ from each other, the question of the origin of these
differences arises. Moreover, they are sometimes together, and occasion-
ally there is only one.

In another annelid, Tomopteris elegans, groups of eight cells each, at
first all of the same size and similar in appearance, occur in the ovary
and later become free in the body cavity. One cell increases in size and
becomes the oocyte; seven remain accessory cells (Fig. 216, F) but are
apparently not incorporated into the oocyte. Presumably each group de-
velops from a single ovarian cell; but, in any case, development of one
cell as the oocyte must be determined by a pattern of some sort in the
group with dominance of this cell. The origin and nature of this pattern
and the relation of egg polarity to it remain to be determined.

In the oogenesis of the gephyrean Bonellia thin follicles are formed
about groups of cells at the surface of the ovary, become pedunculate, the
inclosed cells increase to a considerable number, a single cell at the proxi-
mal end of the follicle nearest the peduncle becomes oocyte, and all others
form a group at its distal pole. Since all cells of a follicle are apparently
alike in early stages, a differential must be present in the follicle to de-
termine the proximal cell as oocyte. With the follicular peduncle at one
pole and the group of accessory cells at the other the oocyte is probably
exposed to a differential adequate to determine its polarity.

Association of oocytes with nutritive cells is very common among ar-
thropods. In the genesis of the parthenogenetic egg of the daphnid Sida

^^ See, e.g., van Beneden, 1870; Kerschncr, 1879; Giesbrecht, 1882.

668

PATTERNS AND PROBLEMS OF DEVELOPMENT

I ' AJiOjS'o* 1 O

Fig. 217. — Ovarian tubule of
daphnid, Sida crystallina, show-
ing stage in genesis of parthe-
nogenetic egg. In each group of
four cells, g, the third from the
apical end of the ovary becomes
an oocyte, 0; the others are nu-
tritive cells (from Weismann,
i877,reproduced in Child, 19156)

crystallina, as described by Weismann (1877),
the primitive germ cells of the apical ovarian
region become successive series of four cells
each, the third cell from the apical end in each
becoming oocyte, the others nutritive cells
(Fig. 217). How differentiation of oocyte and
nutritive cells is determined does not appear.
The larger winter egg, containing more yolk
when full grown, uses several series of cells in
its growth, sometimes as many as twelve. In
Polyphemus the relation is different. All nu-
tritive cells associated with a single oocyte are
at one pole of the latter, but this may be either
pole as regards relation to the ovarian tubule.
Polar-body formation at the pole without nu-
tritive cells indicates that the other pole be-
comes basal whatever its position with respect
to the ovary (Kiihn, 1912).

In many insects oocytes and nutritive cells
develop from the primitive ovarian cells, all
apparently alike in early stages, with great
variation in number of nutritive cells associ-
ated with each oocyte. In other species an
epithelial follicle is in direct contact with the
entire surface of the oocyte, other nutritive
cells being absent. For convenience in descrip-
tion the end of the insect ovarian tubule con-
taining the primitive germ cells is regarded
as apical, the other end as basal. From early
stages cells in the ovarian tubule of the earwig
Forficula are associated in pairs (Korschelt,
1 891). The basal cell of each pair becomes
oocyte; the other, nutritive cell, develops a
large branched nucleus (Fig. 218, .4). The re-
lation between oocyte and nutritive cells in an
aphid, Melanoxanthum, is shown for the par-
thenogenetic egg in Figure 218, B, and for the
winter egg, requiring fertilization, in Figure
218, C. A protoplasmic strand connects oocyte

Fig. 2i8, A-F. — Oocytes and nutritive cells in insects. A, the earwig Forficula; one oocyte
and one nutritive cell, the latter apical, in each group; nucleus of nutritive cell becomes highly
branched (after Korschelt, 189 1). B, parthenogenetic, and C, winter, oocyte of plant louse,
Melanoxanthum, connected with nutritive cells by protoplasmic strand (from Tannreuther,
1907, reproduced in Child, 19156). D, Dytiscus; oocyte with intranuclear ring and seven of
the fifteen nutritive cells all cytoplasmically continuous (after Nusbaum-Hilarowicz, 19 18). E,
early, and F, later, stage of oocyte development in queen bee (schematized after Paulcke, 1901).

670 PATTERNS AND PROBLEMS OF DEVELOPMENT

and nutritive cells. Other Hemiptera show somewhat similar relations: in
some, each of several oocytes is connected with a different nutritive cham-
ber apical to it, the strands from the more distant, older oocytes passing to
one side of those farther apical and younger.

In the beetle Dytiscus four successive divisions of an ovarian cell pro-
duce sixteen cells, the basal, larger cell becoming the oocyte, the others
the nutritive cells. The nuclear content of the primary oogonium sepa-
rates into two visibly different portions preceding the first of the four divi-
sions. From one of these chromosomes form; the other passes entire into
the larger cell of the first division, and in each successive division it passes
into one cell, finally forming a large ring in the nucleus of the oocyte. This
has been regarded as a kind of diminution of chromatin in the nutritive
cells, analogous to that in Ascaris. A similar "diminution" has been de-
scribed for certain other insects.''

In Figure 218, Z), an oocyte of Dytiscus is shown with the intranuclear
ring and seven of the fifteen nutritive cells, all cytoplasmically continu-
ous. What Nusbaum-Hilarowicz (19 18) regards as streams of mitochon-
dria passing from the nutritive cells to the oocyte are indicated in the
figure.

The apical region of the ovarian tubule of the queen bee contains small
cells, all apparently alike; lower in the tubule the oocytes, apparently ir-
regularly distributed, become distinguishable by their larger size (Fig.
2 1 8, £) . With progress down the tubule they gradually become more regu-
larly arranged along the tubule axis and spaced at more regular intervals
with the small cells between them. At a still later stage each oocyte is
associated with, and extends into, a group of nutritive cells apical to it
(Fig. 218, F). A follicle develops about this complex, forming a thick
columnar epithelium about the oocyte, thinner about the nutritive cells
(Paulcke, 1901).

The polarity of the insect egg coincides in direction with the axis of
the ovarian tubule, the apical pole being directed toward the apical end
of the tubule. When nutritive cells are present, the oocyte is usually as-
sociated with a cell or group apical to it, that is, nutritive material from
these cells enters through that part which becomes the apical pole. In the
Crustacea, however, cells basal, as well as apical, to an oocyte function as
its nutritive cells. If the cells are all alike primarily — all potentially

^''Dytiscus: Giardina, 1901; Debaisieux, 1909; Nusbaum-Hilarowicz, 1918. Gryllus:
Buchner, 1909. Colymbetes (Coleoptera) : Gunthert, 1910. See also E. B. Wilson, 1925,
p. 326.

ORIGINS OF EMBRYONIC PATTERNS 671

oocytes, as is generally agreed — it may be a matter of chance in some
species, perhaps in the bee (Fig. 218, E), which become oocytes, which
nutritive cells. However, in forms such as Forficula (Fig. 218, A), in
which the basal cell of a pair becomes the oocyte, it is evidently not chance
but a definite spatial relation of some sort that determines the difference.
Also, the definitely directed association of oocyte with a cell or a group
of cells apical to it is not a matter of chance.

Recent observations on differential reduction in ovaries of Drosophila
hydei show a basipetal gradient of decreasing rate of reduction in the
ovaries as a whole and in each follicle string. The basal cell, that is,
the cell at the low end of this gradient in each follicle, becomes egg,
the others nutritive cells. The polar gradient of the egg is a part of the
ovarian and follicular gradient. A gradient from the ovarian surface in-
ward corresponds to the ventrodorsality of the egg, the ventral side
being at or toward the outer ovarian surface (see also p. 144; Child,
unpublished).

EGG POLARITY : GENERAL

Eggs of different species undergo their pre-embryonic development
under widely different conditions. The polar patterns of many are obvi-
ously related to differential conditions in their environment, but our
knowledge of environmental conditions affecting the developing egg cell
in the parent organism is a matter of inference from observation rather
than of experimental demonstration. In some forms there is apparently
a differential between free and attached pole, such that the free pole be-
comes apical, perhaps because respiratory exchange is more rapid there,
while nutritive material enters at the attached pole. In other forms the
pole through which nutritive material enters becomes apical, perhaps be-
cause of a gradient in the ovarian tubule of which the oocyte is a part,
as in insects, perhaps because a circulatory system is the chief mediator
of respiratory exchange as well as source of nutritive material. In still
other forms there is evidently relation to a factor external to the egg cell,
archegonium, gametophyte, and embryo sac in plants, nutritive cells,
follicles, etc., in animals; but we have even less basis for a guess as to
the nature of the effective factor. And finally, in some forms the oocyte
is isolated in a fluid medium during its development from a primitive
germ cell. Developmental stages of most egg cells are so inaccessible
to present experimental methods, and interest has been so largely cen-
tered in the egg cell itself, that physiological investigation of the origins
of egg polarity is almost completely lacking.

672 PATTERNS AND PROBLEMS OF DEVELOPMENT

In other forms of development polarity is determined by various ex-
ternal differentials, and polarity determined by one differential may be
obliterated and a new polarity determined by another in many animals.
In some cases of unicellular development polar pattern is evidently de-
termined by direction of the last preceding mitosis. It is possible that
every mitosis may determine some degree of cell polarity, perhaps evanes-
cent, or obliterated by the next mitosis in a different direction or by a
differential of one kind or another — for example, the general physiological
gradient and dominance in a multicellular organism, a differential in
oxygen supply, electric potential, light, etc. In the absence of an effective
differential external to the cell the polarity determined by the preceding
mitosis may perhaps persist and become a definitive axiate pattern. Per-
haps the polarities of some eggs are initiated at the final oogonial division.
In other eggs a polarity so determined may be obliterated by factors in
the intraorganismic environment. In the egg of Fucus the original polar-
ity, however determined, is evidently readily obliterated by various ex-
ternal factors (pp. 423-25). At present there seems to be no adequate
ground for concluding that polarity of the egg is different in origin and
nature from other physiological polarities of cells and multicellular sys-
tems.

SYMMETRIES AND ASYMMETRIES OF EMBRYONIC DEVELOPMENT

The features of developmental pattern commonly distinguished as sym-
metries and asymmetries from polarity have often been regarded as polari-
ties in other directions than the chief or primary polarity. They appear
in definite relations, characteristic for the species, to the polar, major,
or primary axiate pattern. Organismic pattern may be radial about the
polar axis, ventrodorsal or dorsiventral with a resulting bilateral pat-
tern at right angles to the polarity; or asymmetries — lateral, spiral, or
irregular and specific — may be present. In many animals a polarity is evi-
dent at or before the beginning of embryonic development; but ventro-
dorsality, dorsiventrality, or asymmetry becomes evident only at some
later developmental stage and perhaps gradually. The question how and
when the symmetries or asymmetries originate and what their physio-
logical basis is, has been and still is the subject of much discussion and
speculation. Experimental evidence indicates that the basis of symmetries
and asymmetries of the earlier developmental stages of various animals
is present in the\egg, apparently even before fertilization, though the
patterns mayjbe experimentally alterable. The point of entrance of the

ORIGINS OF EMBRYONIC PATTERNS 673

spermatozoon or its path in the egg has been regarded by some as deter-
mining the median plane in certain forms. The first cleavage plane is
definitely related to the median plane and so to the symmetry pattern
in some animals. It has been shown in earlier chapters that, at least in
certain organisms, symmetries and asymmetries show nonspecific differ-
ential susceptibility to external agents and can be altered, obliterated,
and determined in the same ways as polarity. These experimental results
suggest that polarity and symmetry patterns are not fundamentally dif-
ferent in nature. Moreover, the fact that obliteration or reversal of vari-
ous asymmetries by differential inhibition and recovery is not a specific
effect of a particular agent suggests that the differences of right and left
sides in the forms concerned results from a primarily quantitative physio-
logical difference. The physiological bases of axiate patterns appearing
at different developmental stages doubtless differ chemically, but as spa-
tial patterns they may be similar. For example, the polarity of a hydroid
tentacle doubtless differs chemically from the polarity of a hydroid plan-
ula, but they are both characterized by gradients which, as gradients, are
closely similar. In echinoderm development the pattern of differentiation
in relation to the polar axis differs from that in relation to ventrodorsality,
but at certain early developmental stages an apicoventral-basidorsal dif-
ferential or gradient is actually present (p. 134).

Some lateral asymmetries appear indifferently on the right or left. We
regard the laterality of these as determined by incidental or chance fac-
tors, but there is still the question as to the origin and nature of the
asymmetry. The asymmetry of the opercula of serpulid annelids and of
chelae of various decapods is reversible experimentally by removal of the
dominant member of the pair (pp. 411-13). Some others are more or
less constant in their laterality for the species but may be reversed experi-
mentally or occasionally in nature — for example, the vertebrate visceral
asymmetries and various other experimental and natural cases of situs
inversus. The larval coelomic asymmetry of echinoderms may be either
completely obliterated or reversed experimentally. Certain spiral asym-
metries, such as the coiling of gasteropods, result from differential growth
on the two sides of the body, but some species usually or always coil in
one direction. The question how these asymmetries originate and how
their laterality is determined is still unanswered. To say that they are
inherited, that genetic factors are concerned in their laterality, though,
of course, true, throws no light on the physiology of their origin in de-
velopment.

674 PATTERNS AND PROBLEMS OF DEVELOPMENT

The recent extended consideration of the problems of symmetry and
asymmetry by W. Ludwig (1932) and the data presented by Schleip (1929)
show the various types of these patterns and the manner in which they
become evident in development, but the attempts of these and other
authors to throw light on questions of origin and nature of these features
of developmental pattern, while suggestive as far as they go, are chiefly
significant in indicating our lack of definite information. If polar patterns
are determined by factors in ovarian environment, the question at once
arises whether there are factors in the ovarian environment of the oocyte
that might determine symmetry and asymmetry.

RADIAL SYMMETRIES

Radial symmetry may be primarily nothing more than absence of any
pattern except surface-interior about the polar axis. It is apparently no
more than this in many blastulae, sponge larvae, hydroid planulae, post-
tentaculate regions of many hydroids, and many cylindrical organs.
Often, however, differences appear in radii of the same body-level as de-
velopment progresses. In coelenterates, for example, certain radii may
become tentacle radii; in starfishes, arm axes. Some radial patterns de-
velop from patterns not primarily radial, and these from still earlier pat-
terns, apparently completely radial. The radii about the polar axis of the
Corymorpha planula are apparently alike ; but the succession of tentacles
differs in different individuals, as H. B. Torrey (1907) has pointed out and
the writer has also observed. Tentacle development may begin with ap-
pearance of a single tentacle at some point of the periphery of the planula,
a second developing later on the opposite side, still later two others, op-
posite and midway between the first and second. In other individuals two
opposite tentacles or three equidistant or occasionally four equidistant
develop simultaneously, and in still others the first two tentacles may be
not opposite but both in the same half of the periphery. Moreover, the
order of appearance of distal and proximal tentacles may differ in a single
individual. Evidently the final radial tentacle pattern results from de-
velopment of tentacles in the spaces between other tentacles at a certain
body-level. The individual differences in tentacle order perhaps result
from the fact that the Corymorpha planula does not swim but in earlier
stages lies with one side in contact with the substrate and only gradually
erects its apical end about the time that tentacles develop. Tentacle de-
velopment may be slightly retarded on the side on contact, as it is in re-
constitution of pieces; and different orders may result according as erec-

ORIGINS OF EMBRYONIC PATTERNS 675

tion of the hydranth region occurs earlier or later in relation to tentacle
development. Quite apart from this possible differential, there is evi-
dence of a spatial relation in the tentacle order. If one tentacle precedes
others, the second tends to develop as far as possible from it. If two ap-
pear simultaneously, they are almost always opposite, or almost opposite;
and if three or four appear simultaneously, they are approximately equi-
distant. The observations suggest that a developing tentacle is dominant
over a certain distance and that, about the periphery of the planula, not
more than three or occasionally four tentacles can develop. Only as this
periphery increases in size do more tentacles appear in the spaces be-
tween those already present. Evidently in this and various other hydroids
the radial symmetry of the mature hydranth is not predetermined but de-
velops gradually, often through an asymmetrical or a "dorsiventral"
stage, probably determined by external conditions, from a pattern in
which all radii at a given body-level are ahke. In general the number of
tentacles which can develop simultaneously about the periphery of a
coelenterate larva probably depends on the range of tentacle dominance
in relation to size of the periphery. Two different quadriradial patterns,
the perradial and adradial, appear in scyphomedusae ; and a third, multi-
radial tentacle pattern, is present in some. In certain actinians two oppo-
site radii become difTerent from others in mesenterial and esophageal pat-
tern, but the tentacle pattern remains multiradial. The apparently multi-
radial symmetry of tentacles and mesenteries in certain others — for ex-
ample, Cerianthus — results from a sort of ventrodorsality ; that is, ten-
tacles and mesenteries develop successively from one region of the pe-
riphery. In short, radially symmetrical patterns may result from radially
asymmetrical or may become asymmetrical; or certain regions may be
radial, others asymmetrical, ventrodorsal, or bilateral. How patterns like
the four radii of the scyphozoa or the biradial and ventrodorsal mesen-
terial and esophageal patterns of anthozoa are localized about the polar
axis we do not know at present.

VENTRODORSALITY AND DORSIVENTRALITY

In many animals ventrodorsal or dorsiventral pattern appears first dur-
ing the course of development as a graded differential in condition vertical
to the polar gradient. Bilaterality is, at least in most forms, merely an
expression of this apparently secondary pattern. Bilaterality, as well as
polarity, is determined in the ovary, according to E. B. Wilson (1925,
pp. 1021-25), but he does not suggest how it originates. If it is so deter-

676 PATTERNS AND PROBLEMS OF DEVELOPMENT

mined, there must be factors in ovarian environment that can determine
a second physiological axis at right angles to the polar axis. Is there evi-
dence or probability of such factors?

Oocytes of many animals are primarily cells of an epithelium. If this
epithelium is a part of the polar body gradient, as it is in various coelen-
terates, or if it develops progressively, as do the lobules of a branching
ovary or the ovarian rhachis and its epithelium in Ascaris, there is in it a
graded dififerential, at least during its development. This is parallel to
its surface, and the small fraction of this gradient in the oocyte is vertical
to a polar gradient between the free and the attached pole. Such a gra-
dient derived from the germinal epithelium may perhaps constitute the
basis of ventrodorsal pattern in some forms. Since the differential in the
single oocyte is very slight, the secondary axiate pattern may not be evi-
dent at the beginning of development and may appear gradually, as it
does in many animals. Even the tridimensional pattern of the Arenicola
egg (p. 658) may conceivably be determined in this way. The ovaries de-
velop in lateral regions of the body, and their cells share in anteroposterior
and ventrodorsal gradients of the body wall. If slight differentials in the
directions of these gradients are established in the oocyte at different
stages and a third arises between free and attached poles before the young
oocyte is isolated in the body cavity, a basis for the triaxiate pattern is
present.

The ovarian tubules of insects and some other arthropods presumably
share, during their development, in the ventrodorsal gradient of the body.
If this is the case, the ventrodorsahty of the egg may perhaps represent
the persistence and further development of the fraction of this gradient
in the cell. Lateral asymmetries of some forms may originate in similar
manner from a differential in the parent body, involving the oocyte.
Certain dorsiventraUties and asymmetries of later development seem to
originate in this way. The temporary dorsiventrality of the Pdmatohydra
bud appears to be an expression of the longitudinal gradient of the parent
body (p. 634). The anteroposterior and dorsiventral patterns of the am-
phibian limb are apparently expressions in it of anteroposterior and dorsi-
ventral body gradients, and up to a certain stage its dorsiventrality can
be altered by altering orientation of the hmb bud to the body gradients.
Probably patterns of various other appendages of other animals are simi-
larly related to general body gradients. Dorsiventrality in embryos of
fishes, reptiles, and birds coincides in direction with the primary polarity
of the egg but may perhaps originate in the differential between surface

ORIGINS OF EMBRYONIC PATTERNS 677

and interior of the blastoderm; this is perhaps also true for the mammals.
Differential dye reduction in Drosophila ovaries suggests that in this
form ventrodorsahty of the egg may be determined by a differential
between surface and interior of the ovary itself (p. 144).

SYMMETRIES AND ASYMMETRIES IN ECmNODERMS

Echinoderm development presents perhaps the most remarkable se-
quence of symmetries and asymmetries of any animal group. The early
embryonic stages of most echinoderms appear to be completely radial,
though ventrodorsahty is indicated in some species in early cleavage or
even before cleavage.^* Various lines of experiment also give evidence of
a ventrodorsal gradient in early stages and of a right-left gradient in
the apical archenteron.'^ Development of defective skeletons in plutei
from isolated 1/2 blastomeres led Plough (1927, 1929) to postulate locali-
zation at or before first cleavage, of skeleton-forming material in the basal
part of the egg, excentric in the ventrodorsal axis and bilateral. However,
according to W. Marx (193 1), skeletal defects of larvae from early blasto-
meres result from ectodermal conditions. An ectodermal bilaterality, pres-
ent at or before the two-cell stage, consists of two regions conceived es-
sentially as gradient systems, decreasing from centers of highest potency.
These, rather than localization of skeleton-forming material, determine
bilateral aggregation of mesenchyme and so the skeletal bilaterality; in
partial forms from isolated blastomeres this bilaterality is defective. This
conclusion is in accord with earlier work on the relation of skeletal locali-
zation to ectoderm. Skeleton-forming material is localized in the most
basal region of the unfertilized Arbacia egg, according to Horstadius
(1937a), not between the nucleus and center of the egg (Harnly, 1926).

The question of the relation of first and second cleavage planes to the
median plane of the larva has been investigated in a number of echino-
derms by intravital staining of one cell of the two-cell stage with Nile blue
sulphate.^" This procedure shows, in Echinus, Par echinus, and Paracen-
trotus, 50-59 per cent coincidence of first-cleavage plane and median plane;
17-42 per cent with first cleavage plane frontal, that is, at right angles to
the median plane; and the remainder mostly oblique, with small percen-
tages uncertain. In the clypeastroid Echinocyamus the first plane is usu-
ally frontal. In the asteroid Astropecten it is usually median (41. i per

'* J. Runnstrom, 1920; J. und S. Runnstrom, 1920.

■' See pp. 134-36, 219-21 .

" Von Ubisch, 1925a; Runnstrom, 19266; Horstadius, 192S6.

678 PATTERNS AND PROBLEMS OF DEVELOPMENT

cent) or frontal (44.6 per cent), rarely oblique (3.6 per cent), and some-
times equatorial (10.7 per cent). In Asterias it is frontal. The eggs of the
holothurians Cucumaria and Psolus possess a distinguishable ventrodorsal
pattern and are ventrodorsally elongated; the lirst cleavage plane is
frontal and in later cleavages a ventrodorsal difference in cell size appears
(J. und S. Runnstrom, 1920). These data suggest a difference in degree of
development of ventrodorsality at the time of early cleavage in different
echinoderms and perhaps also in different individuals of the same species.
In the regular sea urchins investigated ventrodorsal pattern apparently
usually prevents obliquity of the first two planes, and the first plane is
more often median than frontal; but in the holothurians the ventrodorsal
pattern, or perhaps merely the ventrodorsal elongation of the egg, deter-
mines orientation of the first cleavage spindle in the long axis, and a
frontal cleavage plane results. In general, both observation and experi-
ment seem to indicate that ventrodorsality develops gradually, becoming
an effective factor earlier in some species than in others. That ventro-
dorsality of the sea-urchin egg might be determined at fertilization, the
meridian of sperm entrance becoming median ventral, was suggested by
Jenkinson (191 1&). However, if sperm enters through the micropyle at
the apical pole, there is no such meridian. The physiological basis of ven-
trodorsal pattern is apparently present in the unfertilized eggs of some,
perhaps of all, echinoderms; but conclusive evidence as to manner and
time of its origin is lacking. From the blastula or early gastrula stage
the larvae of sea urchins and starfishes are bilateral in appearance and
show a distinct physiological ventrodorsality (pp. 134-39).

The lateral asymmetry of the hydrocoel becomes evident at different
developmental stages of the coelom. In some species earlier stages of the
coelom sacs appear alike on the two sides, even to formation of two
hydrocoel primordia, but one degenerates later. In others asymmetry be-
comes evident early in coelom development. The usual position of the hy-
drocoel on the left side, though readily altered experimentally, indicates
that genetic factors are in some way involved in the laterality; but how
the asymmetry originates and how the definite spatial relation to ventro-
dorsality and to polarity are determined, remain to be discovered. The
radial features of the adult echinoderm develop from the lateral asym-
metry. The left coelom, a part of which usually becomes the hydrocoel,
shows no evidence in its earlier stages of the later radial pattern. De-
velopment of the ambulacral canals from the hydrocoel probably involves
spatial relations of dominance, like tentacle development; but the pat-

ORIGINS OF EMBRYONIC PATTERNS 679

tern of these is primarily asymmetrical and becomes radial only seconda-
rily.

The oral-aboral polarity of the adult echinoid, asteroid, and ophiurid
is a new polarity, not coincident with that of the egg and embryo. In
holothurians the original polarity persists as the longitudinal axis of the
adult, and in the crinoids there is apparently a complete reversal of polar-
ity in metamorphosis. In addition to these changes in pattern, the radial
pattern of the adult becomes an anteroposterior polar pattern in clypeas-
troids and spatangoids; an anteroposterior motor pattern is present in
different degree in some other forms; and in certain holothurians the
radial pattern becomes ventrodorsal, that is, certain radii become differ-
ent from others and function as ventral side. If the various echinoderm
patterns originate as gradient patterns in relation either to the ovarian
environment of the oocyte or to gradient factors already present, it is
evident that evolution of spatial pattern in the group has involved changes
in the gradients. In this connection the approach of asteroid to echinoid
larval pattern with experimental differential inhibition and of echinoid
to asteroid pattern with secondary modifications of differential recovery
may be recalled (pp. 208, 217).

SYMMETRIES, ASYMMETRIES, AND DETERMINATE CLEAVAGE

It was pointed out in chapter xiv that in forms with spiral cleavage the
relation between the first two cleavages and the median plane is appar-
ently constant for the species but that different authors do not agree as
regards the relation in different forms. However, the species differences
in size of somatoblasts, lack of coincidence of the median planes in differ-
ent cells, and shifts in position of cells make it uncertain whether the
apparent differences are really significant. Dorsiventrality, as indicated
by difference in cell size, may be present or absent in early cleavage pat-
tern quite independently of symmetry or asymmetry of development. It
is not even certain whether a pattern of dorsiventrality or asymmetry
is always present before fertihzation or before cleavage. That the full-
grown oocytes of these forms possess a spatial pattern of some sort, at
least a polar pattern, appears beyond question; but this pattern may be
distinguishable only in somewhat eccentric position of the nucleus toward
the apical pole. Beginning with the breakdown of the oocyte nucleus, or
later with polar-body formation, or still later on with fertihzation, changes
in visible pattern very commonly occur, consisting in localization, usually
in definite relation to the polar axis, of visibly distinguishable cytoplasmic

68o PATTERNS AND PROBLEMS OF DEVELOPMENT

areas, or in change in localization or distribution of areas previously vis-
ible." The form of the unfertilized egg in some species — for example,
Arenicola (p. 658) — indicates that something more than polar pattern
is present. According to Just (191 2), the meridional first cleavage plane
in Nereis usually passes through the entrance point of the sperm, and this
may be any point of the egg surface. If this is true, a dorsiventral pattern
determining direction of cleavage is not present in this egg before fer-
tilization. Whether a dorsiventral pattern is present in the egg of Chaelop-
terus and some other forms before fertilization seems to be uncertain. ""

Although there is no general relation between spiral cleavage and
lateral asymmetry, since most of the groups with this type of cleavage
are not laterally asymmetrical, a relation apparently does exist in the
laterally asymmetrical gasteropods. In most gasteropod species the coil-
ing is dextral, with occasional sinistrally coiled individuals in some. Cer-
tain species, however, are completely sinistral with "situs inversus." In
these the cleavage is reversed, that is, cleavages dexiotropic in other
forms are here leiotropic, and vice versa.^'' Observation of spiral rays in
one aster of the first polar spindle in certain gasteropods led to unsuccess-
ful attempts to correlate direction of this spiral with cleavage and direc-
tion of coiling.'^ Moreover, a spiral aster appears at the inner pole of the
first polar spindle of the annelid Arenicola marina because the spindle
forms equatorially and turns through an angle of approximately 90° about
this pole (Child, 1898), hut Arenicola is not laterally asymmetrical. Spiral
asters are doubtless associated with protoplasmic movement, and it is,
of course, possible that direction of movement is correlated with an egg
pattern; but evidence of their association with gasteropod asymmetry is
lacking. The suggestion by Conklin (1903a, b) that reversal of cleavage
and of asymmetry in sinistral gasteropods results from reversal of polarity
in eggs of these species after they are freed from ovarian attachment is
not supported by evidence, even when position of polar-body formation
is altered by centrifuging (Conklin, 191 7).

The question of the manner of inheritance of sinistrality in gasteropods
has received considerable attention but is not yet fully answered. The
viviparous genus Partula produces either dextral or sinistral young ir-

" See, e.g., Conklin, 1902; E. B. Wilson, 1904; F. R. Lillie, 1906; Schleip, 1914; Penners,
1922.

" F. R. Lillie, 1906; Morgan, 1938; Morgan and Tyler, 1938.

23 Crampton, 1894, Physa; Holmes, 1899, Aitcylus, and 1900, Planorbis.

'^ Rabl, 1900; see also Schleip, 1929, pp. 111-12.

ORIGINS OF EMBRYONIC PATTERNS 68i

respective of its own asymmetry or may produce both dextral and sinistral
individuals (Crampton, 1916, 1924). Self-fertilized individuals of Ziwwaea
peregra or dextral or sinistral pairs give essentially similar results, but
some difference of opinion exists as to interpretation in Mendelian terms/s
Sturtevant suggested that the data indicate a simple Mendelian case with
dextrality dominant but "with the nature of a given individual deter-
mined, not by its own constitution, but by that of the unreduced egg
from which it arose." Later data, however, seem to require modification
of this hypothesis.

The coincidence of cleavage pattern with the biradial symmetry of the
ctenophore has been noted (p. 564); that a biradial pattern is present in
the unfertilized egg is maintained by Fischel (1903) and some other au-
thors, but it seems possible that the biradial pattern may develop gradu-
ally in relation to cleavage and the pre-existing polarity. According to
the generally accepted view, the apicobasal axis of the Ascaris egg be-
comes the dorsiventral axis of the animal, and the anteroposterior axis
becomes evident only after change in position of cells following the second
cleavage ; asymmetry in position of certain cells appears later (pp. 570-72) .
Except for the morphological polar differential usually visible in the egg
and the relation of chromatin diminution to it, nothing is known concern-
ing the pattern underlying these developmental events. The data at hand
concerning entomostracan Crustacea are particularly confusing. In cer-
tain of the forms studied a polar pattern is visible in the egg, but the axis
of the first two cleavages is oblique to the axis of this pattern but defi-
nitely related to the median plane of the animal. In some species egg
polarity and anteroposterior axis coincide; in others — for example, Cy-
clops— egg polarity is said to become dorsiventrality of later stages, with
basal pole ventral.^'' That the basal pole should be ventral seems rather
remarkable in view of the fact that in arthropods generally ventral pre-
cedes dorsal in development. Moreover, here, as in Ascaris, the antero-
posterior polarity of the animal apparently becomes evident only sec-
ondarily, and how it is determined does not appear.

The full-grown ascidian oocyte gives evidence of polarity in the ec-
centric position of the nucleus but no visible indication of cytoplasmic
localization, except a surface-interior difference. As described in chapter
xiv, a visible dorsiventrality results in the egg of Styela from cytoplasmic

^5 Boycott and Diver, 1923; Diver, 1925; Boycott, Diver, Hardy, and Turner, 1929; Boy-
cott, Diver, Garstang, and Turner, 1930; Sturtevant, 1923; Crabb, 1927.

^^ See pp. 574-76; also Schleip, 1929, pp. 295-321 and literature cited there.

682 PATTERNS AND PROBLEMS OF DEVELOPMENT

streaming, initiated by fertilization, and can be followed through the cell
lineage to particular organ systems. In eggs of some other ascidians cyto-
plasmic localizations are not visible, but the course of development sug-
gests that pattern is essentially similar to that of Styela (Conklin, 1905a,
h, c). The cytoplasmic localizations appear to be associated with fertili-
zation; but there is still the question whether fertihzation is actually the
determining, or merely an activating, factor, with localization resulting
from a dorsiventral pattern already present. According to Conklin, the
"mesoplasm," which gives rise to mesoderm, is a peripheral layer in the
oocyte; "ectoplasm," most of which becomes ectoderm, is in the oocyte
nucleus; and "entoplasm," later largely entodermal, is almost central.
The spermatozoon enters near the basal pole, apparently not at a definite
point; but its path in the egg leads Conklin to conclude that a symmetry
pattern is present before fertihzation.

On the basis of the character of development after fertilization of
pieces of unfertilized ascidian eggs a symmetry pattern was postulated
in the unfertilized egg by Dalcq (1935, chap. iii). This was regarded as
consisting of two bilaterally symmetrical, crescentic areas in the periph-
eral cytoplasm of the equatorial region on opposite sides of the oocyte,
each extending around four-fifths or more of the circumference and over-
lapping each other laterally. One, the "neurochordoplasm," lies slightly
deeper than the other, the "mesochymoplasm" (Fig. 2ig,A,B). A medio-
lateral gradient pattern is also postulated in each of these crescents. Al-
though, according to a personal communication. Professor Dalcq does
not now regard these figures as representative of the true organization
of the ascidian egg, they are given here with his permission because they
provide, in a measure, a scheme of pattern intermediate between that
postulated by Conklin and what Dalcq regards as a more adequate
scheme resulting from more extended investigation (Dalcq, 19385).
Through the personal kindness of Professor Dalcq, which is here grate-
fully acknowledged, a drawing provided by him showing this later scheme
is reproduced as Figure 219, C, and the explanation beneath the figure
is given in his words. Dalcq regards the cytoplasmic movements associ-
ated with fertilization as dependent on this pre-existing pattern, but the
origin of whatever pattern is present remains completely obscure. What-
ever its origin and nature, the symmetry pattern seems to be a highly
effective factor in development. The first cleavage plane is median, and
cleavage is bilateral, as far as followed in detail, but earlier cleavage
planes do not coincide with the boundaries of the cytoplasmic areas.

B

A

n

Fig. 219, .4 -C— Cytoplasmic symmetry pattern of the ascidian egg, as postulated by
Dalcq. A, B, scheme of pattern derived from earlier investigations: "neurochordoplasm"
unshaded; "myochymoplasm" black; A, polar view with median plane and plane of first
cleavage indicated by broken line; B, lateral view with polar axis and approximately the plane
of second cleavage indicated by broken line (after Dalcq, U organisation de Vauf chez les
Chordes, 1935). C, organization of the unfertilized egg, according to Dalcq's experiments on
double merogony. An and Vg, animal and vegetal poles; H, side of the future head; T, side of
the future tail; the plain semicircular lines represent the postulated cortical field with its decre-
ment from //; the dotted spots represent the mantle of yellow pigment lying just under the
cortex; the horizontal, interrupted lines represent the internal (yolk?) gradient with its
decrement from Vg (from a drawing provided by Professor Dalcq).

684 PATTERNS AND PROBLEMS OF DEVELOPMENT

SYMMETRIES AND ASYMMETRIES IN OTHER INVERTEBRATES

Concerning origins or relations to any particular factors of develop-
mental pattern in triclads, rhabdocoels, trematodes, cestodes, and bryozoa
practically nothing is known; in the polyembryonic bryozoa the prob-
lem appears in particularly interesting form (pp. 536-37). As regards most
other invertebrates, we know little more than that patterns of symmetry
or asymmetry appear. In monembryonic insects polar pattern is prob-
ably derived from the ovarian tubule, but whether a ventrodorsal pat-
tern is present in the tubule is not known. How patterns of the definitive
embryos of polyembryonic insects originate is a problem for the future.
Concerning possible relations of symmetry patterns of other arthropods
to any particular factors, there seems to be no definite information.

AMPHIBIAN DORSIVENTRALITY

Among the eggs of vertebrates those of amphibians have been most
studied with reference to the problem of origin of dorsiventrality. Eggs

of most amphibians show no definite evi-
dence of dorsiventrality preceding fertili-
zation. In various species a part of the pig-
mented surface adjoining the unpigmented
basal region becomes less deeply pig-
mented after fertilization, forming the so-
called "gray crescent" (Fig. 220). The
breadth of the crescent decreases bilateral-
ly from its broadest median region; this
region becomes the median dorsal region
Fig. 22o.~Diagrammatic outline of the embryo, the median plane passing
of frog egg viewed laterally with ap- through it and the apicobasal axis. The
proximate extent of one arm of gray crescent varies in extent and distinct-

crescent mdicateu by broken Imes. o ^ • i i

ness in different species, bemg most clearly
defined in certain anura, indistinguishable in some forms. Its median re-
gion is usually more distinct than lateral parts, and it is not sharply
bounded but shades off into the more deeply pigmented and the unpig-
mented regions; its visible characteristics suggest a gradient system.^^

Before the gray crescent was recognized as a visible feature of the
dorsal region, attempts had been made to determine whether a relation
between point of entrance of sperm or its path in the egg and the median

'7 For accounts of the formation and characteristics of the crescent see Vogt, 19266, 19286;
Banki, 1927; Weigmann, 1927; also the general account by Schleip, 1929, pp. 55'^-^-''-

ORIGINS OF EMBRYONIC PATTERNS 685

plane of the embryo existed. When the gray crescent was discovered,
such a relation seemed all the more probable, since the dorsiventrality
indicated by the crescent was believed to become visible only after fer-
tilization. Experimentally localized application of sperm to the egg and
histological studies of fertilization showed a high frequency of coincidence
or near coincidence of median embryonic plane and plane determined by
point of entrance of sperm or its path toward the female pronucleus
after its entrance. Jenkinson believed, however, that other factors,
gravity or light, might play a part.'^ The view that the spermatozoon
determines or is an important factor in localizing the median plane found
wide acceptance until the question was again opened by further investi-
gation. Polyspermic eggs of Rana fusca were found to develop a normal
gray crescent (A. Brachet, 1910a, b), and in parthenogenetic development
induced by puncture a normal gray crescent and dorsiventrality develop
without relation to the meridian of puncture (A. Brachet, 1911a, &; 1927).
Moreover, evidence that the crescent appears in some amphibian eggs
before fertilization was presented by Vogt (19266, 1928^)). Recent attempts
to throw further light on the question indicate that the median plane
may be determined independently of the meridian of sperm entrance.'^
According to Tung, entrance point of sperm is ventral, with rare ex-
ceptions, in normal development; but in dispermic eggs a definite rela-
tion between dorsiventrality and sperm entrance points does not appear.
Since gray crescent and dorsiventrality appear in polyspermic and in
parthenogenetic eggs, Brachet concluded that a labile dorsiventrality is
present in the unfertihzed egg but may be altered by the spermatozoon.
Another possibiHty, suggested by Spemann and Falkenberg (1919) and
regarded favorably by Dalcq (1935), is that the ventral region of a pre-
determined dorsiventrality is a preferential region of sperm entrance.
However, the earlier view that dorsiventrahty is epigenetically deter-
mined by the sperm still finds support (Wintrebert, 1933a, h; 1934). At
present it appears beyond question that a normal dorsiventrality and
crescent may appear in parthenogenetic and polyspermic eggs of various
amphibians. Moreover, the fact that, even when the fertihzation meridian
is in the ventral region, it is by no means always in the median plane and
may be at some distance from it indicates that other factors than the
sperm are concerned in determining dorsiventrality; but it does not ex-

=« Newport, Ellis, and Forbes, 1854; Roux, 1887; A. Brachet, 1903, 1904; Jenkinson, 1906(2,
1909.

^9 Banki, 1927; Wcigmann, 1927; Gilchrist, 1932; Tung, 1933; Pastcels, ig^h.

686 PATTERNS AND PROBLEMS OF DEVELOPMENT

elude the possibility, suggested by Brachet, that a predetermined dorsi-
ventrality may be altered by the sperm. It is not necessary to assume
that such alteration must always result in coincidence of median plane
and plane of fertilization meridian. It may only bring about more or less
approximation; or, if the angle between the two is large, the sperm may
have little or no efTect. The possibility that a dorsiventral differential
may be initiated in the oocyte by the pattern of follicular circulation has
already been noted (p. 664); this possibility is a tempting one, but it re-
mains for the future to discover whether it is a reaUty.

Some recent experiments on the eggs of R.fiisca are of particular inter-
est here. It is well known that after removal from the uterus unfertilized
amphibian eggs orient to gravity, with heavier basal pole down, as soon
as the swelling of the jelly and the appearance of the perivitelline fluid
permit rotation within the envelope. Some time after insemination an-
other partial rotation begins, amounting to some 15° in certain urodeles
and to 30° in R. fusca and some other anurans. This rotation displaces
the apical pole ventrally, the basal pole dorsally, in the future median
plane, and in R. fusca coincides in time with formation of the gray cres-
cent.

According to studies by Ancel and Vintemberger (1933, i935). the use
of local marks shows that this ''rotation of fertilization" is not rotation
of the whole mass of the egg but a movement of the "pellicle," that is,
of the superficial cortex over the deeper cytoplasm. On the dorsal side,
movement toward the apical pole is most extensive in the future median
plane and carries with it part of the underlying pigment, the region at
the lower border of the pigmented hemisphere thus partially deprived of
pigment becoming the gray crescent. In its median region, this is approxi-
mately 30° wide, and its width decreases laterally as the movement be-
comes less extensive. These authors find, however, that cortical move-
ment is not only toward the apical pole but from lateral regions toward
the median plane on the dorsal side, so that the cortex or "pellicle" be-
comes thicker dorsally, thinner laterally. Evidently the future dorsal re-
gion differs in some way at this stage from lateral and ventral regions,
and there is a graded difference between its median and lateral regions.

In further experiment dorsiventrality and median plane have been de-
termined experimentally in eggs of R. fusca by Ancel and Vintemberger
(1938). Eggs removed from the uterus to a slide without water, with
polar axis inclined 45° from the vertical and with basal pole upward, ad-
here to the slide by the jelly and are inseminated by application of sperm

ORIGINS OF EMBRYONIC PATTERNS 687

to the region in contact or elsewhere as desired and then brought into
water. The eggs remain in the obHquely inverted position for some 15
minutes until appearance of perivitelline fluid permits orientation to grav-
ity by partial rotation, the basal pole rotating downward, the apical pole
upward, in a vertical plane passing through the polar axis ("rotation of
orientation") ; that is, the heavier basal pole passes downward in the direc-
tion of its inclination from the vertical. In this orientation there is no
rotation about the polar axis. The later 30° rotation of fertilization is in
the opposite direction. In 87 of 100 of these eggs the median plane is
within 45° of the plane of rotation of orientation, in 9 eggs it is between
45° and 90°. The gray crescent forms on the side of the egg above the
basal pole as the rotation of orientation begins, and this side becomes
dorsal. The same results are obtained with electrically activated, as
with fertilized, eggs and with activation before or after orientation. In
eggs inseminated locally at the equator with basal pole directly upward
the angle between median plane and entrance point of sperm is less than
45° in 147 of 150 eggs and less than 15° in 78 of these. In these eggs the
sperm is a more or less effective factor in determining dorsiventraUty.
When sperm and rotation of orientation act in different directions, the
latter overcomes the former. The localization of dorsal regions and median
planes by change in position of cytoplasm and yolk by reaction to gravity
in eggs maintained in inverted and partly inverted position has already
been discussed (pp. 428-30). In general, the data indicate a physiological
basis for dorsiventrality in the unfertilized egg, but this is evidently al-
terable experimentally before the superficial movements associated with
formation of the gray crescent take place.

DORSIVENTRALITY IN OTHER VERTEBRATES

In the meroblastic eggs of fishes, reptiles, and birds the longitudinal
or polar axis of the embryo may probably be regarded as coinciding with
an egg meridian, though it is practically at right angles to the egg axis.
The dorsal side of the embryo is the side adjoining the free surface of the
blastoderm; consequently, the question of symmetry becomes the ques-
tion of how embryonic polarity is determined. A possible indication of
axiate pattern and symmetry, consisting in a regional difference in the
marginal periblast, was described by Ruckert (1892). One side of the
teleost blastodisc, supposedly the posterior side, is thicker than the other
before cleavage.^" Also, one side becomes more susceptible in early stages

3° Oellacher, 1872; Agassiz and Whitman, 1885; Kowalewsky, 1886.

688 PATTERNS AND PROBLEMS OF DEVELOPMENT

of some forms (p. 149), but how these differences originate is not known.
If they are expressions of axiate pattern, they indicate the radius or region
of initiation of axiate development, and bilaterality is apparently a sec-
ondary result of a longitudinal and dorsiventral differential in the develop-
ing region. This region is not the only one capable of forming an embryo,
as experiment has shown (pp. 521-22), but is probably living a little more
rapidly than other parts, so that certain events occur in it earlier than
elsewhere, and its dominance determines the course of development. How
it is localized is still to be discovered.

The same question, how the longitudinal embryonic axis and the median
plane are determined, arises with regard to reptiles and birds. Data on
reptile eggs are few: the embryonic axis of the gecko is said to be ap-
proximately at right angles to the long axis of the egg (Will, 1893). In
the chick the embryonic axis is more or less nearly at right angles to the
long axis of the egg shell; and, with the pointed end to the right, the
anterior end of the embryo is usually directed away from the observer.-''
This orientation, however, is by no means constant; deviation in either
direction is frequent but usually not great, though inverted orientation
sometimes occurs. ^^ The pigeon's egg shows a somewhat similar relation
between long axis of the embryo and of the whole egg, but the modal
angle between the two is about 70°, with variation from 8° to 135°, and
inversion is rare.^^

The early oocyte of the pigeon, according to Bartelmez, exhibits bi-
laterality. A polar axis is indicated by the eccentric position of the nu-
cleus nearer one pole, the future apical pole, and by positions of cyto-
plasmic granules and yolk nucleus. This polar axis is "not infrequently"
vertical to the surface of the ovary, with the apical pole attached. A
second axis, the long axis of the oocyte, is also indicated, and bilaterahty
results from position of the nucleus nearer one pole of the long axis, that
is, only one plane divides the oocyte into symmetrical halves. Whether

31 Von Baer's rule; von Baer, 1828.

s^Duval, 1884, found the two axes vertical to each other in 75 per cent; Kopsch, 1927, in
:iS per cent; and other figures range between these — e.g., Butler, 1935, 50 per cent. According
to Butler, deviation up to 45° to the right is 33.1 per cent; to the left, 10.6 per cent; more
than 45° to the right and left, 2.5 per cent each. See also Fere, 1900; Rabaud, 1908. Percent-
ages of inversions given by different authors vary widely: J. C. Dalton, 1881, 12 per cent;
Duval, 1884, 0.6 per cent; Bartelmez, 191 8, ^;i per cent.

" Bartelmez, 1912, 1918. See also Blount, 1909; Patterson, 1909. Bartelmez reports 0.67
per cent of inversion in one lot of 600 eggs and 2 per cent, all in eggs of two of some 90 birds
with a total of 400 eggs.

ORIGINS OF EMBRYONIC PATTERNS 689

this pattern is predetermined in the oocyte independently of ovarian en-
vironment, as Bartelmez seems to believe, or is related to some environ-
mental factor, as seems possible, it is not evident that it has any neces-
sary relation to bilaterality of the embryo, for the median plane of the
embryo forms an angle varying from 8° to 135° from the median plane of
the oocyte, with a mode at 70°. Bartelmez also finds that in some oocytes
the shorter diameter of the nucleus is inclined to the long axis of the
oocyte by about the same angle as the embryonic axis to the long axis
of the egg, and he suggests that the embryonic axis may be determined by
the nucleus. He regards the polar axis as predetermined in the oocyte,
and polarity and symmetry as inherent properties of protoplasm, per-
sisting from generation to generation. If this were the case, however, it
would seem that much greater constancy of angle between embryonic
axis and long axis of the egg might be expected, but the relation between
the two axes gives a probability curve with less variation in eggs from one
bird than in those from more than one. In short, the data concerning
chick, pigeon, and other birds do not exclude the possibility that embry-
onic axiate pattern is not predetermined but originates in reaction to
some more or less variable factor in the organismic environment of the
oocyte or egg. The possibility that its determination may be correlated
in some way with position of the ovary on the left side of the body seems
not to have been considered.

In mammals follicle development perhaps subjects the growing oocyte
to a definitely directed environmental differential, and a polarity may
result; but that polarity is not the polarity of the embryo. Oocyte polarity
may determine polarity of the blastocyst, that is, position of the embry-
onic area, which represents essentially a blastoderm; but whether it plays
any further role in determining developmental pattern is not certain.
Since embryonic dorsiventrality develops in a definite relation to the sur-
face-interior pattern of the embryonic area, the dorsal side being toward
the outer surface, it may represent either persistence of oocyte polarity
in the embryonic area or a direct reaction to surface-interior differences
or both. Since the apparently radially symmetrical blastocyst gives no
evidence of factors determining direction of the longitudinal embryonic
axis, the possibility suggests itself that this axis may be determined by
factors in the uterine environment, or, more specifically, in the region of
implantation. If a physiological differential is present in the uterine wall,
subjection of the embryonic area to it on contact with, or implantation
in, the uterine mucosa may perhaps determine that axis. Moreover, if

690

PATTERNS AND PROBLEMS OF DEVELOPMENT

determination docs occur in this way, direction of the axis in relation to
the uterus as a whole may differ, more or less, according to the region
with which the blastocyst becomes most closely associated. If there is a
longitudinal or a dorsiventral gradient system in the uterine wall, or a
gradient system in relation to each oviduct, the blastocyst adhering to,
or implanted in, the mucosa is subjected to the differential in the region
concerned. Even in the monotremes axial determination during the in-
trauterine period and in relation to intrauterine environment does not
appear impossible. ■''*

Fig. 221, A, B. — Polyembryony in the mnt-handtd. ^xmdiAiWo, Dasy pus novemcinctus. A,
outline reconstruction of blastocyst and vesicle with two primary buds, r and / and probable
early stage of a secondary bud at s. B, outline reconstruction of later blastocyst and vesicle
with four embryonic primordia. // and / supposedly from one primary bud; IV and III
from the other (after Patterson, 1913).

In this connection the case of the nine-banded armadillo {Dasypus
novemcindus = Tatusia novemcincta) , in which four embryos develop from
a single egg, requires special consideration. Here the ectodermal vesicle
formed by invagination of the ectoderm of the embryonic area into the
interior of the blastocyst gives rise to two "primary buds," each of which
in turn gives rise to two "secondary buds"; and each of these, together
with a part of the mesoderm and entoderm, becomes an embryonic pri-

34 Although no complete search of the literature has been undertaken in attempting to
discover evidence for or against the possibility of intrauterine determination of axiate pattern
in mammals, it appears that many accounts of early mammalian development do not con-
sider the question.

ORIGINS OF EMBRYONIC PATTERNS 691

mordium. The two primary buds arise on right and left sides of the
vesicle (Fig. 221, ^), facing the openings of the Fallopian tubes (Patterson,
1912, 1913). That factors in the uterine environment are concerned in
localizing the two primary buds and in orienting their opposed polarities
appears more probable than that they are determined independently of
environment by fissions of the blastocyst (Newman, 191 7, pp. 47-49)-
According to Patterson, two thickenings appear on each primary bud,
one at its tip, the other on its left side {s of Fig. 221, yl), as viewed from
the center of the free surface of the blastocyst. Later the four embryonic
primordia appear more or less equidistant from each other (Fig. 221, B),
two of them, // and IV, retaining the positions of the primary buds, the
others, / and ///, being respectively dorsal and ventral. How formation
of a secondary bud on the left side of each primary bud is determined re-
mains a problem; if it indicates an asymmetry in the primary bud, the
origin of that is entirely obscure.

In another species of armadillo^-^ tertiary buds arise in variable num-
ber, six to twelve embryos being formed (Fernandez, 1909). It has been
suggested that a number of radially arranged apical points are determined
in the ectodermal vesicle before visible budding and that those which
happen to lie toward right and left sides of the uterus have more room for
growth (Newman, 1917, pp. 50-51). This hypothesis is without support
of evidence and does not seem to be required by the data of observation.
Differences in rate and somewhat less uniformity in physiological domi-
nance and isolation would permit repetition of budding more frequently
in some of the primary or secondary buds than in others. Moreover, it is
difficult to conceive how such a radial system of presumptive embryonic
primordia could arise all at once from what Newman regards as origi-
nally a single embryo. The budding of the ectodermal vesicle results in
the determination of a new longitudinal axis for each embryo and, conse-
quently, a new bilaterahty (Fig. 221, 5); but dorsiventrality remains the
same in all.

VERTEBRATE ASYMMETRIES

All vertebrates are normally asymmetrical as regards position of vari-
ous organs — heart, stomach, liver, coiling of intestine, etc. — and exhibit
various functional asymmetries. In general, these asymmetries are highly
constant as regards laterality, but situs inversus occurs occasionally.
That genetic factors are concerned in some way is evident, but how the
constancy of laterality is determined is no better understood here than

i'^ Das y pus hyhridiis=Taliisia hybrida = Mulila hybrida.

692 PATTERNS AND PROBLEMS OF DEVELOPMENT

for invertebrate asymmetries. A few points that may have some bearing
on the problem are briefly noted.

In both urodele and frog developmental stages with open neural folds
removal and reimplantation in the same place, with anteroposterior re-
versal, of a dorsal rectangular piece of sufficient size from the middorsal
region, including presumptive neural tissue and underlying inductor ma-
terial, result in reversal of asymmetry; if the piece is merely removed,
asymmetry is not reversed. ^^ Other experiments show that altered orien-
tation of prospective neural tissue does not influence asymmetry; conse-
quently, it seems evident that the archenteric roof is a predominant factor
at the stage concerned in determining the laterality of asymmetry, but it
is not the only factor. These experiments throw no light on the origin of
asymmetry.

Reversal of polarity in the undivided frog egg by reversal of position
in relation to gravity does not reverse asymmetry (Hammerling, 1927),
nor does separation of right and left 1/2 blastomeres in a urodele by
gradual constriction with a hair ligature (Mangold, 1 92 1) . Gradual partial
constriction by ligature of the urodele blastula in the median plane gives
a somewhat different result. Monsters with more or less anterior duplica-
tion develop, the left member always with normal asymmetry, the right
usually with situs inversus. ^^ If constriction is continued to complete
separation of right and left halves, the left half develops with normal
asymmetry, the right about equally with normal and reversed asym-
metry (Ruud und Spemann, 1922). Division of the heart primordium into
right and left halves at the tail-bud stage results in development of two
hearts, the left with normal, the right with reversed, asymmetry (Ekman,
1925). These data bring out two important points: first, asymmetry ap-
parently develops gradually, for reversal of polarity in the undivided egg
and separation of right and left halves at the two-cell stage do not alter
asymmetry, but with separation at the blastula stage the right half is
often reversed; second, the laterality of asymmetry is apparently more
stable in the left than in the right half, as if there were on the left side a
region dominant as regards asymmetry. A gradually developing activity
gradient with high end on the left side is postulated by Huxley and
De Beer (1934, pp. 73-82).

3^ Spemann, 1906, 1918; Pressler, 191 1; R. Meyer, 1913.

37 Spemann und Falkenberg, 1919. Anterior duplications in fishes, if they do not extend
posteriorly beyond a certain level, often show situs inversus in the right member, as noted by
Stockard, 192 1, and by other authors.

ORIGINS OF EMBRYONIC PATTERNS 693

In various experiments two individuals or organ systems developing
side by side with cellular continuity between them very generally show
opposed symmetry or asymmetry patterns. Even two hydranths of Cory-
morpha developing close together become "dorsiventral" or bilateral in
relation to each other, and the mirror-imaging of amphibian appendages
has long been known (pp. 390-95). Apparently there is in these and other
similar cases a relation between the two members tending to make the
pair a symmetrical whole. If the asymmetry of the amphibian appendage
results from a physiological differential, these cases of mirror-imaging re-
ceive a simple interpretation ; the high or the low side of the differential
is common to both, or the differential in one may induce a differential
symmetrical to it in the other.

Both morphological and physiological evidences of asymmetry appear
in the chick blastoderm of the head-process stage. The anterior end of the
primitive streak bends slightly to the left (Fig. 167); and in chorio-allan-
toic grafts from this stage pieces from the left side show greater develop-
mental capacity than those from the right, suggesting a differential in
physiological condition from left to right (Rawles, 1936). In later stages
the embryo undergoes torsion progressively from the head region pos-
teriorly, so that it finally comes to lie on the left side. This suggests a
growth differential, higher on the left side during torsion, and dye reduc-
tion gives evidence of a physiological asymmetry at these stages (p. 160
and Fig. 55). Reversal of direction of this torsion (heterotaxia) occurs
occasionally under natural conditions. In a brief abstract Gray, Dodds,
and Worthing (1940, Anat. Rec, 78, Suppl., p. 77) report 6 per cent hetero-
taxia in some nine hundred embryos. They also find that various optically
active d- and /-substances have opposite effects on frequency of hetero-
taxia and suggest that the lateral reversal of differential growth is due to
interference with normal metabolism or to an asymmetrical utilization
of, or sensitivity to, the optically active substances.

Of the four embryos developing from a single armadillo egg, one pair
is connected with the right, the other with the left, placental disk. In a
study of symmetry relations, as indicated by anomalies in the integumen-
tary bands of scutes, it was found that members of a pair are more nearly
identical and exhibit interindividual mirror-imaging more often than indi-
viduals of different pairs (Newman, 19156, 1916). Mirror-imaging in
individuals of opposite pairs is interpreted by Newman as indicating the
primary bilaterality of the original embryo; mirror-imaging in individuals
of a pair, as indicating the secondary bilaterality of the primary bud,

694 PATTERNS AND PROBLEMS OF DEVELOPMENT

which is superimposed on the primary bilaterahty and more or less obht-
erates it; the tertiary bilaterahty of each individual is superimposed on
the secondary bilaterahty and tends to obliterate both it and the primary.
It seems possible, however, that mirror-imaging in members of a pair may
be, at least in part, the expression of action on each other, like inhibition
or other modification of adjoining sides in various other forms.

STRUCTURAL CONCEPTS OF DEVELOPMENTAL PATTERN

The "intimate structure" of some sort, so often postulated as the basis
of organismic developmental pattern, has been regarded by some as in-
herent in the protoplasm concerned, by others as originating in relation
to some factor external to the protoplasm. Resemblances or analogies
between organismic structures, cells, and even whole organisms and crys-
tals have been pointed out repeatedly; and hypotheses of an organismic
pattern essentially crystalline, or in some way resembling crystalline
structure, have appeared again and again. The discovery of fluid crystals
and supposed resemblances of their behavior to that of living protoplasms
seemed to give support to these views. ■'^ Observations with polarized light
have given evidence of optical anisotropy in many plant and animal struc-
tures, and X-ray analysis is advancing still further our knowledge of the
ultramicroscopic structure of biological materials. ^^ However, many of
the structures in which evidence of a definite molecular or micellar pat-
tern is found are nonliving products of protoplasmic activities, such as
cuticular substances, cellulose, and other nonprotoplasmic membranes —
shells, skeletal structures, fibers of cellulose or protein constitution, hair,
silk, etc. Evidence of ultramicroscopic orientation also appears in certain
protoplasmic structures, muscle, nerve, fibrillar connective tissue, vari-
ous other fibrillar structures of cells, and in the highly condensed or
structurized substance of sperm heads. It is perhaps a point of consider-
able importance that this evidence concerns nonprotoplasmic structural
products and protoplasms with a high degree of structurization. In some
of these structural patterns — for example, connective tissue and bone — ■
orientation of ultramicroscopic particles apparently represents primarily
a result of mechanical environmental factors rather than a crystalline
space lattice, and pressure and tension are probably concerned in many
other such patterns. Evidence of change in pattern of proteins in relation

3" See, e.g., O. Lehmann, 1911. Resemblances and differences between fluid crystals and
organismic structures are discussed in many other papers.

39 Seifriz, 1935; W. J. Schmidt, 1937; F. O. Schmitt, 1939 and citations.

ORIGINS OF EMBRYONIC PATTERNS 695

to tension has been obtained by X-ray analysis.''" There is also evidence
of molecular orientation with respect to surfaces and interfaces. These
mechanical and surface orientations, however, may be in all possible di-
rections in a cell or a multicellular organism; consequently, it does not
appear possible that they can constitute the basis of axiate developmental
pattern. Protoplasms in general have not yet been shown to possess any
such structure that might serve as a basis for developmental pattern.
Evidence of the presence in eggs or other reproductive cells of a space
lattice related to developmental pattern is at present lacking. The pe-
ripheral cytoplasm of the egg of a sea urchin {Strongylocentrotus pur-
puratus) and, in the centrifuged egg, the zone of clear cytoplasm become
birefringent on fertilization, according to Moore and Miller (1937); but
this pattern can be only secondarily associated with developmental pat-
tern, for the egg possesses a polar, and probably a ventrodorsal, pattern
before fertilization. The cytoplasm of Amoeba is said to be birefringent
(W. J. Schmidt, 1937), but Amoeba has no persistent axiate pattern.
Evidence of orientation of elongated molecules in relation to direction of
protoplasmic flow has also been obtained in many cases.

The hypothesis that the organism is fundamentally crystalline in char-
acter has been advanced in one form or another by various authors, and
similarities between organism and crystal have been pointed out repeat-
edly. Interpretation of the Bruchdreijachbildungen (see pp. 387-95) and
other features of development in terms of a hypothetical space lattice
has been attempted by Przibram."' Since the space lattice is entirely hy-
pothetical, it can, of course, be assumed to undergo the changes required
to account for the experimental results, and the required assumptions
are made for each case. Amphibian asymmetries and experimental rever-
sals have suggested analogies to crystals to certain authors.^' The ac-
cumulating evidence for existence of greatly elongated molecules and of
union of units, end to end, to form long chains — for example, in cellulose
and in various proteins — has been regarded by some as affording new
support for an essentially crystalline or paracrystalline basis for develop-
mental pattern. Seifriz, for example, in recent publications, holds that
there is in protoplasms a continuity of structure consisting of such elon-
gated molecules and that polarity and symmetry of organisms result from

■I" E.g., Astbury, 1937, 1939; Astbury and Bell, 1938.

•" Przibram, 1906, 1921, and other papers.

^^ Spemann und Falkenberg, 1919; Harrison, 1921(7.

696 PATTERNS AND PROBLEMS OF DEVELOPMENT

it.''^ However, it is a long jump from the molecular patterns of cellulose,
silk, wool, etc., or from those of muscle, nerve, or connective tissue, or
even those of proteins, to the polarity and symmetry of whole organisms;
and the categorical statement quoted in footnote 43 certainly goes far be-
yond the evidence.

A fundamental asymmetry of protoplasms has been postulated by some
because it has been found in many cases that only one of the two optical
isomeres of certain constituents of protoplasms, amino acids, various
lipoids, and sugars serves for protoplasmic synthesis. If such an asym-
metry is present, it evidently does not necessarily determine an asym-
metric organismic pattern, for many organisms show no evidence of asym-
metry. Moreover, in many of those that are asymmetric the asymmetry
becomes evident only relatively late in development, only in certain organ
systems or organs, and always in definite relation to more general features
of pattern that are not asymmetric. If molecular factors have anything
to do with organismic asymmetry, they seem to be effective only as locali-
zation of primordia and structurization progress.

If this kind of asymmetry is a general property of protoplasms and is
concerned in determining organismic asymmetry, what happens when the
organismic asymmetry is experimentally reversed — for example, the coe-
lomic asymmetry of the sea-urchin larva or vertebrate asymmetry? Has
the one isomere been transformed into the other? Does the individual
with reversed asymmetry use the other isomere in synthesis? Does ex-
perimental obliteration of asymmetry by differential inhibition or estab-
lishment of a new asymmetry by an environmental differential result
from a corresponding change in isomere pattern? These questions remain
unanswered, and there is still the more general question: How can a mo-
lecular asymmetry determine differences in metabolism, in rate and
amount of growth, and in differentiation on the two sides of an organism?
A general asymmetry of protoplasms has not yet been optically or other-
wise demonstrated; but, as pointed out in the preceding chapter, molecular
factors may be concerned in many of the asymmetries of highly differen-
tiated and structurized protoplasms.

A molecular hypothesis which attempts to account for specific regional
localizations has been advanced by Harrison.''-' Calling attention to the

« Seifriz, 1935; 1936, chap, xv; 1938. In the latest of these papers he says: "Polarity of cells
and symmetry of organisms exist in virtue of molecular patterns indicative of the crystalline
state."

■»^ 19366, 1937, and suggestions concerning a molecular basis of pattern in various earlier
papers.

ORIGINS OF EMBRYONIC PATTERNS 697

very general belief that the specific characteristics of protoplasms result
chiefly from their proteins and to the conception of protein molecules as
possessing definite spatial patterns, polarity, and symmetry or asym-
metry, he suggests that embryonic developmental pattern results from
the configuration of protein molecules. In consequence of their bipolar
character they tend to orient in the cell, "possibly with respect to the
point of attachment in the ovary." Different chemical properties at the
two poles bring about different reactions; and the resulting substances are
carried electrophoretically toward opposite poles of the cell, forming two
opposed material gradients, each decreasing in concentration from one
pole to the other. Differences in concentration along these gradients, to-
gether with substances of nuclear origin, initiate new reactions locally
and so begin the progressive localization and complication of develop-
ment. Symmetry and asymmetry are supposedly further expressions of
the molecular configuration

This hypothesis admits the possible relation of developmental pattern
to environment in suggesting that the molecular orientation may occur in
relation to attachment of the oocyte, but it does not suggest how the
orientation is brought about. If orientation does occur in this way, must
it not be because a difference in condition between point of attachment
and other regions produces a differential of some sort in the cell? If this
be granted, it is probable, since the cell consists of living, metabolizing
protoplasm, that this differential will involve a differential or gradient in
metabolism, rate of living, or whatever we may call the continuous physi-
cochemical change constituting life. In short, this hypothesis of Harri-
son's does not exclude, but seems to require, a gradient involving the es-
sential activities of the protoplasm to bring about the molecular orienta-
tion. It is perhaps questionable whether molecular orientation ever occurs
spontaneously, that is, entirely without any relation to factors outside
the molecules concerned. If such a gradient in the oocyte is a factor in
molecular orientation, it, not the orientation, is the primary axiate pat-
tern.

Origin of symmetry and asymmetry in terms of this hypothesis raises
further questions. If the molecules possess a dorsiventral and bilateral or
asymmetric pattern and different substances are produced on different
sides, they, like the polar substances, must be transported to opposite
sides of the cell, dorsally and ventrally or laterally; but how? Independ-
ent cataphoresis in two opposite directions along two or three different
axes does not appear possible. All such hypotheses require postulation

698 PATTERNS AND PROBLEMS OF DEVELOPMENT

of factors in addition to the molecular orientation to account for localiza-
tion on the molar or regional scale characteristic of organismic pattern.
In Harrison's hypothesis the essential point for developmental pattern
is not the orientation of the protein molecules but the transport and
localization of the substances produced toward opposite poles of an axis.
An electrophoretic factor originating from a metabolic differential in rela-
tion to attachment of the oocyte could bring about the transport whether
the protein molecules were oriented or not, or might determine produc-
tion of the substances in situ, and molecular orientation would be entirely
unnecessary. In this connection it is of particular interest that an attempt
to discover evidence of molecular orientation by application of X-ray
analysis to living amphibian neural plate, neural tube, ear, ectoderm,
myotomes, notochord, and yolk and various tissues of chick embryos is
entirely without positive result (Harrison, Astbury, and Rudall, 1940,
"An attempt at X-ray analysis of embryonic processes," Jour. Exp.
Zool., 85). The authors conclude that the negative results do not sup-
port the theory of molecular orientation but do not disprove it.

Not only the origin of spatial organismic pattern but many other fea-
tures of development are difficult to account for in terms of a primary
molecular pattern. For example, how is the reconstitution of apical partial
hydranths from short hydranth stem pieces under good conditions and of
complete individuals under inhibiting conditions to be accounted for?
How can molecular orientation determine that apical halves of early
sea-urchin cleavage stages produce only ectoderm in sea water but after
treatment with LiCl produce entoderm and mesenchyme? Ventrodorsal-
ity and even polarity can be obliterated in embryos or isolated pieces of
various organisms by exposure to certain concentrations of inhibiting
agents (chaps, v, vi). Moreover, ventrodorsality in echinoderm embryos
may be obliterated without obliterating polarity. Hydroid polarity can
be determined by difference in oxygen content of the water. In such de-
terminations orientation of molecules, if it occurs, must result from the
metabolic differential; also, it must be entirely independent of cell bound-
aries. In other words, any molecular orientation present in these and
many other cases must result secondarily from the metabolic differential
pattern.

The similarity between organism and crystal apparently consists chiefly
in appearance of a definite spatial pattern in both; but in the crystal the
pattern is geometric, while in the organism regional differences in metab-
olism and in concentration and kind of substances arise progressively.

ORIGINS OF EMBRYONIC PATTERNS 699

To the extent that chemical reaction takes place, the crystal pattern dis-
appears; chemical reaction is an essential feature of pattern in organisms.
That organismic pattern originates in a crystalline space lattice appears
on the basis of the evidence extremely improbable, but that crystalline
structures and molecular orientations appear in organisms is evident, and
the evidence indicates that they are secondary effects of the pattern.
Some of them are evidently determined by purely local conditions and
are only remotely related to the general pattern; others are more directly
related to that pattern. Probably molecular or micellar orientation is
very generally characteristic of fibrillar structures and of various particles,
inclusions, and metabolic products. Mechanical tension and pressure may
bring about orientation of molecules or particles, orientation may occur
in relation to interfaces and to flow, and some protoplasms may become
crystalline; but there is, at present, no evidence for, and much against,
the concept of organismic pattern as primarily a pattern of molecular
structure and orientation. Such concepts seem to put the cart before the
horse.

Some or all of the specific asymmetries of unicellular forms, spermato-
zoa, etc., discussed in the preceding chapter, and perhaps many of the
more minute structural features in multicellular forms, may be expressions
of a molecular or crystalline structure; but it may again be emphasized
that, in general, they appear to represent more nearly ends than begin-
nings of development and that they develop in definite relations to more
general patterns of earlier stages. If a structural factor of this sort is in-
volved in lateral asymmetries of multicellular forms, it also is apparently
a derivative of developmental pattern; but the evidence available sug-
gests that lateral asymmetry is a physiological gradient pattern.

More or less definite patterns on a molar scale appear in many inorganic
systems quite independently of any molecular pattern of the whole. For
example, any particular region of a flowing stream shows a definite pat-
tern of flow of water, erosion or deposition, and morphology of banks and
bed. This pattern develops from the activity of the water in relation to
gravity, an environmental factor, and to the banks and bed. The mo-
lecular pattern of the water undergoes continuous change; the banks and
bed may consist of various substances — sand, clay, and rock of different
kinds — with many different molecular patterns, but the general morphol-
ogy of the stream depends on the conliguration of banks and bed and
energy of flow rather than on molecular pattern of any of the stream
components. In "normal" environment, that is, within certain limits of

700 PATTERNS AND PROBLEMS OF DEVELOPMENT

change in conditions, the stream pattern, hke organismic pattern, is
highly persistent but does undergo a gradual development which is a
continuous equilibration. But under "abnormal" conditions the pattern
may change greatly — for example, in a flood or when dammed. That the
living organism is more nearly analogous to a system of this kind than to
one resulting from a particular molecular patterns seems probable.

THE VARIOUS GRADIENT CONCEPTS

The frequency with which the suggestion that physiological polarity
is a gradation or gradient pattern of some sort or a stratification of sub-
stances has appeared in biological literature is both interesting and signif-
icant. A polar pattern of this sort was suggested for the sea-urchin egg
and later for the egg of Ascaris by Boveri (19016; 1910a, b). In many
papers on regeneration Morgan postulated gradations of formative sub-
stances and also suggested gradation of tension. More recently Runn-
strom and his co-workers have interpreted their experimental results on
sea-urchin development in terms of two opposed substance gradients in the
polar axis, another, ventrodorsal gradient, and a left-right gradient pat-
tern (pp. 140, 241). A somewhat similar hypothesis has been advanced by
von Ubisch (1936a). Even in Harrison's hypothesis discussed in the pre-
ceding section molecular orientation serves merely to produce concentra-
tion gradients of substance. The hypothesis advanced by Ludwig (1932,
pp. 422-27) postulates different R- and L-agents in opposed concentra-
tion gradients as a basis for bilaterality and asymmetry, but how these
gradients originate and how they are maintained is not considered. A
maximum concentration of one of these agents higher than that of the other
determines asymmetry; and if the one of higher concentration is injured,
the one that was previously inferior becomes predominant, and reversal
of asymmetry results. '•^ The inheritance of asymmetry Ludwig regards
as analogous to sex inheritance.'*^ It may be questioned whether the data
on symmetry require different R- and L-agents and opposed gradients.
Bilaterality apparently results from a ventrodorsal or dorsiventral differ-
ential, and many right-left differences may result from a single quantita-
tive differential between the two sides of the body. If this is the case,
resemblance to sex inheritance is less evident.

In all attempts to account for experimental alterations of pattern in

45"Kommt durch Schadigung des pravalierenden das bisher inferiore Agens zum tjber-
wiegen, so tritt eine Umkehr der Asymmetrie ein" (Ludwig, 1932, p. 424.)
46 Ludwig, 1932, pp. 427-30; 1935-

ORIGINS OF EMBRYONIC PATTERNS 701

terms of concentration gradients it becomes necessary to assume that
they undergo change in position and concentration relative to each other
in orderly and definite ways; one must "suppress" the other, steepness
must change, and concentration must decrease or increase. The assump-
tions made often seem to be concerned primarily with activity gradients
and their changes, though stated in terms of concentration gradients. In
any case, metabolism seems to be necessary to bring about the postulated
changes in the concentration gradients. If a single metabolic gradient
may be associated with two opposed substance gradients, it, rather than
the substance gradients, is the effective factor in developmental pattern.
Concentration gradients cannot accomplish development without metabo-
lism. Also, differences in concentration of substances in different regions
may result from differences in metabolic activity. If we substitute rate
or intensity for concentration, many experimental modifications of de-
velopment present less difficulty to interpretation. In these terms many
lateral asymmetries and their reversals appear rather similar to cases of
physiological dominance of one side of the body over the other; something
of this sort seems to be what Ludwig has in mind when he suggests that
injury of the previously predominant agent may result in predominance
of the other and so bring about reversal of asymmetry.

At first glance the asymmetries, particularly those with more or less
constant laterality, may seem to require an underlying molecular or other
spatial structural pattern, genetically determined. But molecular orienta-
tion alone cannot determine asymmetry on an organismic scale. As Har-
rison has recognized, there must be another factor bringing about locali-
zation of different developmental potencies on a larger than molecular
scale. However, many lateral asymmetries become evident at different
developmental stages in different parts. Moreover, they behave experi-
mentally like gradients, not like specific localizations; they show differ-
ential susceptibility to external agents, and some of them can be obliter-
ated and reversed in the same ways as other features of axiate pattern.
Is it perhaps possible that the genetic factor in these asymmetries con-
sists not in determining an asymmetric structure in the egg but in a defi-
nite and constant relation of the developing oocyte to some inequality or
shift in conditions in the parent body determining ventrodorsality or dor-
siventrality? Or does presence of polar and ventrodorsal gradient in some
way determine pattern in a third dimension? Determination of a third
axis or differential by passage of an electric current in a conductor through
a magnetic field oriented at right angles to the conductor is cited by Hux-

702 PATTERNS AND PROBLEMS OF DEVELOPMENT

ley and De Beer as an interesting analogy in a physical system to deter-
mination of an asymmetry in relation to other axes already present in
organisms. That it is anything more than an analogy they do not main-
tain.^'

It is sufficiently evident from preceding chapters that gradients of ac-
tivity of various kinds — metabolism, rate of cell division, growth, differ-
entiation, etc. — -are very generally characteristic of developmental pat-
terns. Whatever our assumptions concerning "intimate structure" of one
kind or another, or concentration gradients, the earliest distinguishable
and most general form of axiate pattern appears to be a gradient or gra-
dient system on a molar scale, involving the essential activities of the
living protoplasm in which it appears. Within the more general systems
of earlier stages new, more restricted gradient systems of different kinds
arise as activities at different levels of the earlier systems become increas-
ingly different. These activity patterns may bring about definite and per-
sistent patterns of molecular or other structure or orientation, crystalline
or micellar. In short, patterns of many kinds may result from the orderly
and integrated activities of living protoplasms.

It is evident from earlier chapters that a persistent axiate pattern can
originate in a local activation or from a quantitative environmental differ-
ential and that at different levels of a primarily quantitative pattern dif-
ferentiations and new patterns may arise. There is also considerable evi-
dence indicating that ventrodorsality and dorsiventrality are primarily
gradient patterns, becoming effective at a later developmental stage than
polarity, and after the polar pattern has brought about more or less ma-
terial alteration along its course. According to this concept, regional dif-
ferentiations, concentration gradients of substance, specific chemical re-
lations of regions by means of hormones and other substances, and what-
ever molecular orientation or crystalline structure may be present are all
developmental expressions, not primary features of pattern.

The most general and primitive type of pattern on an organismic scale
is an excitation-transmission gradient in, or on the surface of, a proto-
plasm, resulting from transmission with an intensity decrement from a
region primarily excited by an external energy. This pattern can arise

'" As they point out, when a conductor carrying an electric current and a magnetic field of
independent origin are oriented at right angles to each other, the conductor is subjected to
a force acting in a direction at right angles both to it and to the magnetic field (Huxley and
De Beer, 1934, p. 79). With the magnetic field vertical, the north pole upward, and the con-
ductor carrying current horizontally away from the observer the conductor, tends to be dis-
placed to the left.

ORIGINS OF EMBRYONIC PATTERNS 703

without any pre-existing regional differentiation, and it represents the
most primitive pattern of organismic behavior in reaction to an external
factor. If a region of primary activation persists, the gradient may per-
sist and become a physiological axis, that is, a pattern in which axiate
development is possible. Many lines of evidence indicate that axiate or-
ganismic pattern in its simplest terms is such a gradient, but it does not
follow that individual development always begins with pattern in its
simplest terms. Many eggs at the beginning of embryonic development
are far beyond this stage.

The gradient pattern represents the "organism as a whole" in its sim-
plest terms. It is the primary and fundamental correlating and integrating
factor on an organismic scale. That the organism is more than the sum
of its parts is undoubtedly true, for it consists not only of the parts but
of the ordering and integrating factors of the gradient system. The unity
or wholeness, factor of wholeness, morphe, etc., of the organism, is pri-
marily the gradient system and the relation of dominance and subordina-
tion resulting from it; this makes possible the spatial and chronological
order of development.''^

Again it must be emphasized that embryonic development constitutes
only a small part of the problem of developmental physiology and, so far
as beginnings of developmental pattern are concerned, not the most im-
portant part. An adequate theory of developmental pattern must include
all types of development and must be based on the simpler, not on the
most highly specialized, types. There must be a fundamental identity of
developmental pattern underlying development of an individual from an
isolated piece of another individual, from a bud, from an aggregate of dis-
sociated cells, and from an egg. The concept of developmental pattern
as primarily a dynamic gradient system or pattern is an attempt, based
on much evidence, to formulate that identity in general physiological
terms. Can we conceive all the reconstitutions of wholes from parts —
either parts of an egg or embryo or parts of a mature individual — except
in terms of an essentially quantitative gradient system, unless we are
willing to follow Driesch in postulating a metaphysical ordering and inte-
grating principle, a sort of god in the machine, which we may name
"entelechy" or something else, as we please? Driesch argued that an or-
ganism is not a "machine," that is, a physicochemical system, because
isolated parts of a machine cannot make a whole machine, but isolated
parts of many organisms can become wholes. But Driesch was thinking

■''* Cf. Ritter, 1919; von Bertalanffy, 1932.

704 PATTERNS AND PROBLEMS OF DEVELOPMENT

of machines with specific locahzed parts. There are quantitative "ma-
chines" which can be divided into parts indefinitely, and each part pos-
sesses the characteristics of the whole, only on a smaller scale. A flowing
stream, an electric current in a conductor, and, to a considerable extent,
a flame are such machines. If axiate organismic pattern is primarily a
quantitative gradient pattern, the organism in its simplest terms is a
quantitative machine in a specific protoplasm. Parts capable of becoming
wholes when isolated differ from the wholes only quantitatively, as far
as the factors essential to reconstitution are concerned; if qualitative re-
gional differences are present, they are not essential. As a matter of fact,
we know that they are often obliterated. The pattern in an isolated part
may differ in scale from that of the whole, according to physiological and
external conditions; scale of organization may be larger than the piece, so
that a partial axiate pattern results, or smaller than the piece, so that a
single axiate pattern does not involve the whole piece. These and many
other experimental data are diflicult to interpret on any other basis than
a primarily quantitative gradient pattern.

Objection to a theory of origin of organismic pattern in relation or reac-
tion to conditions external to the protoplasm concerned has been made on
the ground that features of pattern so constant and fundamental as polar-
ity and symmetry cannot be determined by external factors, though they
may be modified by such factors. To the theory that metabolism is the
essential factor in the establishment of pattern the objection has also
been raised that a fixed persistent pattern becoming the basis of morpho-
logical development cannot originate in activity, in metabolism, alone.
These objections ignore certain important points. First, external factors
in their action on living protoplasms only initiate changes; the results
depend on the specific constitution and physiological condition of the
protoplasm concerned. The physiological condition of the protoplasm is
not independent of environment, but within the "normal" range an ap-
proach to a steady state is possible. If local action of an external factor
determines a local activation and a gradient results, or if an external dif-
ferential determines a gradient directly, the characteristics of the gra-
dient, its length, its steepness, the reactions occurring in it, and the spe-
cific or qualitative differences which develop at its different levels will all
depend primarily on the constitution of the protoplasm acted upon, rather
than on, the acting agent. The egg of Fucus is an excellent illustration.
Polarity may be determined in it by differential illumination, by electric
current, by a differential in hydrogen-ion concentration, by a stratification

ORIGINS OF EMBRYONIC PATTERNS 705

of protoplasmic substances by centrifugal force, and by deformation of
the egg (pp. 423-25); but however the polarity is determined, the same
sort of individual develops, provided the environment is "normal." In a
piece of hydroid stem polarity may be determined by the activation at a
cut surface, by electric current, or by an oxygen differential; but the indi-
viduals resulting have the same polar pattern, though they may differ in
scale of organization.

Second, developmental pattern is constant and uniform for the species
only within a certain range of so-called "normal environment," and even
within this range uniformity is by no means complete. Alteration in the
essential factors of environment beyond a certain relatively narrow range
alters developmental pattern in definite and, to a high degree, predictable
ways and may even obliterate it. Differential susceptibility of different
gradient-levels may bring about alterations so great that without knowl-
edge of the origin of the material its species could not be recognized (see
chaps, v-vii). The belief that developmental pattern is highly stable or
even autonomous in origin was largely a result of the observational study
of embryonic development. With application of experimental methods it
became evident that pattern is much less stable and less independent of
environment than had been supposed. It is true that the polarity of most
animal eggs appears to be relatively stable, but that has been altered
experimentally in both invertebrates and vertebrates and even obHter-
ated. Probably if earlier stages of ovarian development of the oocyte were
accessible to experiment, pattern would be found less stable. Embryonic
ventrodorsality and dorsiventrality and asymmetry have been experimen-
tally altered, reversed, or obliterated in various species and by various
methods.'^''

Third, metabolism, of course, requires a substrate; but a local activa-
tion and establishment of a gradient or gradient system may occur with-
out pre-existing regional differentiation in that substrate, and the metabo-
hsm may bring about regional differences. Many cases of reconstitution
permit little doubt on this point. The effective factor in development ap-
pears beyond question to be metabolism; concentration gradients, mo-
lecular arrangements, and morphological pattern are apparently results
of earlier metabolic patterns, though all of them, when present, may be-
come factors in modifying those patterns.

19 See chaps, vi, vii, xii, xiii.

CHAPTER XVII

PHYSIOLOGICAL INTEGRATION, DIFFERENTIATION

AND GROWTH IN THE PROGRESS OF

DEVELOPMENT

THE course of development is, in general, a progressive physiologi-
cal and morphological alteration and complication of pattern
with appearance of new patterns within those previously present
and often with regression and disappearance of certain features of earlier
stages, such as larval organs. That development represents the reactions
of a protoplasmic or cell system of a certain specific constitution to a
spatial pattern seems evident. The primary spatial pattern represents
the primary ordering and integrating factor, but the character of the
spatial and chronological order in development depends on the specific
constitution. This is true not only for the whole organism but for par-
ticular organ systems. Spatial pattern permits realization in development
of protoplasmic potentialities which cannot be realized in its absence.
With approach to a dynamic equilibrium or a steady state which repre-
sents the limit of development of which the particular protoplasm is
capable in a given environment, development comes virtually to an end.
To deal with details of later development is entirely beyond the present
purpose. Only questions of the relations of parts and some of the prob-
lems of growth are touched upon. As regards relations of parts, we find,
on the one hand, certain ordering or integrating factors, on the other,
a capacity for independent differentiation or self-difTerentiation of cer-
tain parts appearing at certain developmental stages in some forms. These
two factors are, in some measure, mutually exclusive or antagonistic:
ordering and integrating factors represent relations of a part to other
parts; self-differentiation of a part represents independence of integrating
factors. Growth is an important factor in development for morphological
form, and proportions are largely results of differential growth.

INTEGRATING FACTORS

That any form of organismic development is an orderly sequence of
events in space and time is evident. Experimental analysis has shown by

706

PATTERNS OF DEVELOPMENTAL PROGRESS 707

reconstitution of isolated parts, by development of parts following trans-
plantation or explantation, by relation of development of various parts
to innervation, and by discovery of the role of chemical correlative fac-
tors, hormones, etc., that ordering, integrating, or correlating factors play
an essential part in development, that a physiological unity is essential
to initiation of orderly development.

It was noted in chapter i that two groups of factors are concerned in
physiological integration — the transmissive or dynamic, consisting in
transmission of energy changes without mass transport of substance from
region of origin to region of effect, and the material, transportative, or
chemical factors, consisting in production by certain parts of the organism
of chemical substances and their mass transport by one means or another
and action on other parts. These factors constitute the physiological basis
of the unity of the organism, of the organism as a whole.

Transmission of mechanical, thermal, or electrical energies is possible in
living protoplasms; but the transmissive factor most important physio-
logically is transmission of the physiological change known as ''excita-
tion." Difference in electric potential between regions or parts of an or-
ganism results from many differences in physiological condition; but the
fact that characteristic potential differences are maintained, even when
regions concerned are connected by a conductor, indicates that the chief
ultimate source of the differences is metabolism. There are grounds for
believing that electrical transmission resulting from local activation or
from regional differences in activity is the most general and primitive in-
tegrating factor on an organismic scale. According to current theory, elec-
tric transmission is an essential factor in conduction of the nerve impulse.^
Granting this, it is doubtless also essential in the more primitive transmis-
sions of excitations or activations in protoplasms generally. The simplest,
most primitive sort of organismic or developmental pattern appears to
be the spread or transmission, usually or always with a decrement, of the
effect of a local activation or excitation of a protoplasm. Such an excita-
tion-transmission gradient is possible without any pre-existing local dif-
ferentiation of the regions concerned. In consequence of its activation the
region of primary excitation or activation becomes, for the time, a domi-
nant region. Various lines of experiment on determination of polarity
discussed in earlier chapters show that, if the activation persists, the

' See, e.g., R. S. Lillie, 1922, 1923, 1936; Adrian, 1932; and citations by these authors. The
action currents associated with nervous conduction have been studied by a host of investi-
gators for many years.

7o8 PATTERNS AND PROBLEMS OF DEVELOPMENT

protoplasmic substrate within the range of the transmitted effect may be
so altered that a persistent gradient may be established and become an
axiate developmental pattern.

The early differentiation of the central nervous system and the general
localization of the chief aggregations of nervous tissue at higher levels, not
only of the polar but of the ventrodorsal or dorsiventral gradient, suggest
that it represents the most direct physiological and morphological expres-
sion in later development of the dynamic factors of earlier stages. More-
over, it evidently represents the highest development in each species-proto-
plasm of the dynamic integrating factors. This remains true for the verte-
brates, even though an inductor is apparently necessary for neural de-
velopment. The region known as the "neural inductor" in amphibian de-
velopment is a secondary feature of developmental pattern; and direc-
tion of its invagination is evidently not independent of the primary pat-
tern, nor is the localization of anterior regions of the central nervous sys-
tem determined solely by the inductor in development under natural con-
ditions. And finally, it is not yet demonstrated that the natural inductor
is a substance; but whether it is or not, the presumptive chorda-mesoderm
is itself an expression of earlier developmental pattern rather than the
basis of pattern.

The relation to the general gradient pattern of localization of the cen-
tral nervous system and of nervous dominance does not exclude the possi-
bility that various factors, specific chemical, electrical, or mechan-
ical, may be concerned in determining distribution and direction of
growth of axons. The suggestion that axons grow up the gradients, den-
drites down,^ does not seem entirely in accord with the facts. Certainly
in the vertebrate embryo the earlier axons appear to grow down the gra-
dients; and in the final pattern of the spinal cord the long axons are those
which grow down, while the afferent, upward-growing axons are, in gen-
eral, short, as if able to grow only a short distance against some opposing
factor. In general, the chief aggregations of nervous tissue are at higher
gradient-levels, and the nerves grow down. Local conditions and their
changes during development are doubtless concerned in determining the
progressive increase in complexity of directions of growth of axons, par-
ticularly within the central nervous system. It has been maintained, on
the basis of tissue culture experiments, that a physical ultra-structure de-
termines the course of growth of nerve fibers (Weiss, 1934)- Whether
such structure is the only directive factor is still an open question.

2 Huxley and De Beer, 1934, p. 380.

PATTERNS OF DEVELOPMENTAL PROGRESS 709

The integrating action of the nervous system in the functional life of
higher animals is sufficiently evident from a great variety of experiment
and the resulting literature and also from daily life. Moreover, although
nervous function is not essential for differentiation of certain parts — for
example, vertebrate striated muscle — it is necessary for maintenance of
muscle structure after differentiation. Apparently the level of metabo-
lism in the developing muscle is high enough for its differentiation without
nervous stimuli, but in the differentiated muscle the intrinsic level sooner
or later falls below that necessary for maintenance of its structure. Re-
constitution and grafting experiments suggest that in planarians, prob-
ably in hydroids, and perhaps in annelids, the nervous system is the chief
integrating factor in adult life.

The fact that nervous stimuli do not appear to be important or essen-
tial as integrating factors during most of embryonic development may
appear, at first glance, to conflict with the view just advanced. Actually,
however, nervous function is merely the highest development of a func-
tion which we believe to be common to all living protoplasms, that of
excitability. Even though transmission of excitation by differentiated
nerves plays little or no part in embryonic and some other forms of de-
velopment, more primitive transmissions of effects of activations or ex-
citations undoubtedly do occur and play an essential part in early physio-
logical integration. After the general pattern of an organ system is es-
tablished, it may become temporarily more or less independent of pattern
in other parts and may undergo more or less self-differentiation, but it is
later integrated into the whole, at least in part by nervous factors.

Production of a specific substance by one part of an organism and its
mass transportation to, and action upon, another part is obviously pos-
sible only when some degree of difference is present in the parts concerned.
As long as they remain qualitatively alike, they produce the same sub-
stances, though perhaps at different rates. Differences in concentration
of metabolic products at different gradient-levels, resulting from different
rates of production, may be concerned in determining further differences;
but such effects do not constitute specific chemical relations between
parts. Theoretically the specific chemical effect of a product of one part
on another becomes possible as soon as specific differentiation of the parts
and production of different substances begins. Apparently, however, it
does not play a very important part in integration until a considerable
degree of differentiation has been attained. The work of recent years in
endocrinology and on hormones and other chemical products of metabo-

7IO PATTERNS AND PROBLEMS OF DEVELOPMENT

lism has demonstrated their importance in both animals and plants and
has also shown that their effects and interrelations become increasingly-
definite and complex with increasing differentiation and are therefore most
conspicuous and most essential in the higher animals and man. Neverthe-
less, since these specific chemical relations are possible only after the parts
concerned have become at least chemically, and perhaps morphologically,
different, they cannot initiate differentiation or developmental pattern;
they are secondary, not primary, factors in development — expressions of
pattern already present. They may, of course, play important parts in
influencing the further course of development and the character of func-
tion.

INTERRELATIONS OF FACTORS IN INTEGRATION

Actual physiological control and integration may result from the com-
bined action of dynamic and material factors, and the interrelations of the
two groups become exceedingly complex and varied in the higher verte-
brates and man. It is now a familiar fact that in these organisms hormones
may affect the nervous system and alter nervous reactions and general
behavior in many ways. Effects of the sex hormones, adrenalin, etc., on
behavior are cases in point. And it is no less true that nervous stimuli may
influence hormone production or liberation, adrenalin again being an ex-
ample. The chemical mediation of nerve impulses by formation or libera-
tion at nerve endings of acetylcholine oi sympathin represents a still more
intimate relation of dynamic and material factors.-^ Possibly it is not with-
out significance in this connection that the vertebrate hypophysis, which
appears to exercise, in some measure, a dominance in endocrine relations,
is localized, Hke the chief aggregations of central nervous tissue, in the
high region of the polar gradient.

These interrelations, however, have to do with the mature life of higher
animals and undoubtedly originate late in development. We have less
information concerning interrelations of dynamic and chemical factors in
earlier developmental stages, though they may exist after differentiation
has taken place. It is reported that presence of the larval cephalic nerv-
ous system is necessary in certain insects for metamorphosis from larva
to pupa, supposedly because of a substance, a hormone, produced by it.^
The phytohormone, auxin, is generally produced in larger amount in the
apical region than elsewhere in the plant axis and is transported more

•> W. B. Cannon and Rosenblueth, 1937; W. B. Cannon, 1939.
■t Kopec, 1922; Kiihn und Piepho, 1938; Plagge, 1938.

PATTERNS OF DEVELOPMENTAL PROGRESS 711

readily basipetally than acropetally. The axial differential in production
and the directional differential in transport, whether electrical, as at one
time suggested, or dependent on some other factor, indicate relation to
gradient pattern.

Interrelations within the group of dynamic factors attain a delicacy
and complexity far beyond our present understanding in the functional
activities of the central nervous system. The variety and complexity of
interrelation within the chemical group, particularly in the higher verte-
brates, are coming to light with the progress of investigation of hormone
action.

It is evident that the specificity and complexity of integrative inter-
relations increase during development. Transplantation experiments show
that, in general, tissue-specificities, individual-specificities, and even
species-specificities are less evident in earlier than in later stages. But
whatever the conditions in later stages, it is evident that an orderly or-
ganismic development to an integrated whole is not possible without an
underlying general ordering and integrating pattern of some sort — not
merely a structural substrate but a physiologically effective activity pat-
tern.

INDEPENDENT DIFFERENTIATION

In contrast to the ordering and integrating features of developmental
pattern and the dominance and subordination associated with it, experi-
mental isolations and transplantations of parts show that at certain de-
velopmental stages some parts become capable of continuing differentia-
tion for a time in other than the normal organismic environment or in
complete isolation from the rest of the organism. Such parts are said to
be determined and to undergo self-differentiation or independent differ-
entiation, as distinguished from dependent or correlative differentiation.
Also, removal of certain parts at certain stages results in permanent de-
fects, other parts being unable to reconstitute them. As far as parts show
this capacity for self-differentiation, development, or the stage in which
the capacity is present, has been regarded by many as a mosaic of inde-
pendent parts. Huxley and De Beer (1934) head their chapter vii: "The
Mosaic Stage of Differentiation," and give many examples of independent
differentiation, in large part from amphibian development. Many inter-
esting questions center about this capacity for self-differentiation. All
our knowledge of the progress of determination and differentiation and
their stability or irreversibihty results from experiments for the purpose

712 PATTERNS AND PROBLEMS OF DEVELOPMENT

of discovering whether, or to what extent, this capacity is present in par-
ticular primordia at particular stages. We know that in vertebrate de-
velopment different primordia attain the condition permitting independ-
ent differentiation at very different stages of development of the embryo —
some earUer, some later. Moreover, a part or organ system may be de-
pendent in certain respects, independent in others; an amphibian limb
bud, for example, is determined as a limb before its dorsiventrality is
fixed (pp. 285-88). The concept of mosaic development has in most, if
not in all, cases only relative significance. It was shown in chapter xiv
that even in organisms with determinate cleavage development is far
from being the complete mosaic that it was formerly supposed by some to
be. Even in amphibian development there is no single stage of develop-
ment of the whole organism at which it becomes a mosaic of independent
parts. Moreover, even though an optic primordium or a limb bud becomes
capable at a certain stage of more or less self -differentiation, the differen-
tiation of parts of these primordia is not independent of other parts and
in the urodele limb never becomes so.

Capacity for independent differentiation does not necessarily involve
any visible differentiation; but in complex organ primordia, such as the
limb bud or the eye, a definite developmental pattern must be present
and must persist unaltered after isolation from the normal organismic en-
vironment. Considering, again, the amphibian limb primordium, it un-
doubtedly has become different in some way from other regions at a cer-
tain stage, whether we call this difference "chemodifferentiation," with
Huxley and De Beer, or "invisible differentiation" (Gilchrist, 1937) or
merely "determination." But for development as a limb, not only this
difference but an axiate pattern is necessary. Histological differentiation
of a single cell may result from chemodifferentiation of the cell, but for
orderly differentiation of an organ system independently of other parts a
developmental pattern is obviously necessary. In the limb primordium
this is apparently primarily a gradient pattern in the chemodifferentiated,
invisibly differentiated, or determined limb region.

Independent differentiation of transplanted or isolated parts is further
illustrated by many other data discussed in earlier chapters. Amphibian
notochord, otic, branchial, and various other primordia acquire more or
less capacity for independent differentiation at a certain stage of their
development but at different stages of the embryo. Chorio-allantoic grafts
of portions of chick blastoderms show considerable differentiation in the
altered environment, but many of them give rise to more tissues or organs

PATTERNS OF DEVELOPMENTAL PROGRESS 713

than in their normal environment; that is, more or less reconstitution
occurs, indicating that their differentiation is not wholly independent
(pp. 529-35). The insect embryo is not a mosaic in early stages but, at
least as regards ability of larger regions to continue differentiation inde-
pendently for a time, approaches mosaic condition rather early, though
integrating factors are essential in metamorphosis.

According to the earlier predeterministic theories of development, every
differentiation is essentially a self-differentiation; but experiment has
made it evident that this is far from true. At present the capacity for self-
differentiation of a part is generally regarded as a condition appearing in
the course of development, earlier in some forms and in some parts than
in others. Weiss has called it an "autonomization" (1926a) or "emanci-
pation" of parts from the whole (1935, 1939)- The emancipation is rarely,
if ever, complete. Transplanted or isolated self-differentiating parts very
commonly show some departure from the normal; they may show some
reconstitution, developing more than they would normally; they may
develop less than normal; their development may continue for only a
short time, and differentiation is often less complete than normally; even
if histological differentiation is normal, departure from normal morphol-
ogy is frequent. Moreover, whatever the degree of emancipation at a
certain stage, it is not permanent.

Perhaps of even greater significance for an adequate conception of de-
velopment than occurrence of independent differentiation is its absence,
except in the dominant region, in some of the simpler animals. Reconsti-
tution of wholes from isolated blastomeres, parts of blastulae and of
planulae, and also from pieces of the mature individual proximal to the
hydranth takes place in many coelenterates. The hydranth of Tubularia
or Corymorpha and of various other forms, or even its extreme apical re-
gion, can differentiate quite independently of other parts; but no other
region of the hydroid body is capable of independent differentiation at
any stage of life. Medusa buds and probably some gonophores, after
attaining a certain stage, are capable of more or less independent differen-
tiation; but they, like the hydranth, are dominant regions. As regards
potentialities of parts of the planarian embryo, nothing is known ; but in
postembryonic reconstitution the head can develop quite independently
of other parts, but no other part of the body is capable of independent
differentiation. A hydranth or part of a hydranth or a planarian head,
when isolated, does not give rise to other parts; but other parts recon-
stitute hydranth or head, and these induce development of other parts.

714 PATTERNS AND PROBLEMS OF DEVELOPMENT

Annelid reconstitution shows essentially similar conditions. In all these
forms only the dominant region at the high end of the polar gradient is
capable of independent differentiation, and every body-level will develop
as dominant region unless it is subordinated to a dominant region or in-
hibited by external conditions (chaps, ix-xi).

Within the hydranth primordium there is apicobasal dominance. Re-
moval of the apical region during hydranth reconstitution may result
either in regression of the remaining basal portion and reconstitution of a
new hydranth or in regeneration of the apical region, according to stage
and level of removal. In the planarian head the cephalic ganglia evidently
constitute the dominant region of the mature individual, and this domi-
nance is essential to persistence of the individual. The differentiation of
explanted planarian parenchyma into cells and fibers identical in appear-
ance with explanted ganglion cells and their outgrowths (Murray, 1927,
1 931) suggests that these ganglia are the self-differentiating regions of the
head in normal development. Even if this is true, however, the inde-
pendence is probably not absolute, for in grafts of the ganglionic region
into the ganglionic region of a host polarity of the graft may be altered
or reversed (p. 382). The mosaic characteristics of annelid and mollusk
development, as far as determined, concern larval parts which represent
early differentiations of anterior regions. There is considerable evidence
that the presumptive trunk region, at least in annehds, is not a mosaic
(p. 558). Even in these forms the difference between embryonic and adult
stages, as regards independent differentiation, is perhaps not as great as
has been believed. In short, in hydroids and planarians and apparently
in many annehds, subordinate parts are never emancipated; their per-
sistence as subordinate parts always remains dependent on dominance of
higher levels. Emancipation of parts or attainment of mosaic condition
is by no means a universal characteristic of development. Completely
dominant regions do not become emancipated but are independent from
their initiation. Other parts may, in some animals, attain capacity for
independent differentiation, either when they have undergone a certain
degree of determination or differentiation in relation to other parts or
when a local developmental pattern involving dominance and subordina-
tion has been established in them. Ability of a part to continue differen-
tiation for a time independently of its normal relations to other parts does
not necessarily prove that it is independent in the intact animal. Self-
differentiating parts may be accomplishing more in the way of develop-
ment than normally. Whether there is complete isolation or emancipa-

PATTERNS OF DEVELOPMENTAL PROGRESS 715

tion of any other parts in the intact animal than those completely domi-
nant from the beginning, such as the hydranth region of a hydroid, may
at least be questioned.

GROWTH : THE QUESTION OF DEFINITION

In any consideration of the many problems involved in the changes that
biologists have called "growth" in living organisms, the question of defini-
tion must be raised. Many definitions of "growth" have been given, and
the term has often been used without definition for various features of
development. There is still no general agreement as to what constitutes
growth. For some biologists growth is increase in size or weight. Such
increase may occur in many ways — by increase of living protoplasm ; by
increase of fat, or in plants of starch; by swelling of protoplasm or other
cell constituents in consequence of absorption of water; by increase in
size of cell vacuoles; and by deposition of products of metabolism — chitin,
keratin, cellulose, mineral salts in various structural forms, skeletons,
shells, etc. Some of the plant physiologists have maintained that growth
involves change of form. For example, Sachs (1887) defined growth as
increase in volume intimately associated with change of form. According
to Pfeffer (1901), growth is permanent change of form in the protoplasmic
body, and increases in volume or mass are not correct criteria of growth.
Thompson (1917, p. 52) says: "The transference of portions of matter
into the system from without and from one widely distant part of the
organism to another" is what is usually regarded as growth. He points
out, further, that on the basis of this concept of growth the modifications
of form in organisms depend essentially on difference in rates of growth,
in different directions, except in so far as purely molecular forces are con-
cerned. In his study of relative growth Huxley (1932, p. 6) says: "One
essential fact about growth is that it is a process of self-multiplication of
living substance." Elsewhere (p. 149), however, he distinguishes the "mul-
tiplicative, intussusceptive or compound interest type of growth," in
which the increment of material is alive and contributes to further growth
and "the additive, accretionary or simple interest type," in which the in-
crement consists of nonliving material that makes no further contribution
to growth. It is obvious that the additions of different kinds of materials
take place under different conditions and involve different activities. If
growth includes all additions, the physiology of growth presents not one
but many problems.

There is still another aspect of protoplasmic activity involving change

7i6 PATTERNS AND PROBLEMS OF DEVELOPMENT

in volume or weight, to which comparatively little attention has been
given in studies of growth : that is, negative growth or reduction. Growth
as increase in size or weight is not irreversible. In the absence or inade-
quacy of nutritive material or other substances contributing to increase,
or under physiological or pathological conditions that increase catabolism,
reduction in size and in weight of the organism may occur, and this
may also be differential as regards parts of the body. In the warm-
blooded animals reduction soon results in death, but cold-blooded verte-
brates can undergo much greater reduction; and some of the lower
invertebrates — for example, planarians — can undergo reduction by
starvation to a minute fraction of the original size and weight and still
remain active and in good condition, and even capable of reconstitution.
When feeding is resumed, positive growth begins, and the original size
may be again attained. In planarian starvation the digestive tract under-
goes reduction more rapidly than other organs, with progressive disappear-
ance of its branches from their tips basipetally. In absence of digestive
function this organ serves, to a greater extent than any other, as a source
of nutrition. Even in vertebrates absence of function in a differentiated
organ — muscle, for example — results in its reduction, often in its complete
disappearance.

Perhaps, then, growth in living organisms should be regarded as in-
cluding not merely increment but both increment and decrement of sub-
stance, as transfer of substance either to or from an organismic system
as a consequence of protoplasmic constitution and activity.

GROWTH AND DEVELOPMENT

The relation of growth to development is by no means uniform, either
spatially or chronologically. An organism may undergo extensive develop-
ment without growth increment of the whole, or even during reduction,
as in planarian reconstitution during starvation and in early stages of
embryonic development of most animals. Certain regions of the reconsti-
tuting individual grow at the expense of others, and in the embryo proto-
plasm grows at the expense of nutritive material; but the total result in
both cases is reduction. The histological differentiation of cells and
growth, associated with cell division and synthesis of protoplasm, appear
to be more or less mutually exclusive. When cells undergo visible histo-
logical differentiation, they usually cease to divide and very commonly
cease to grow or grow but little or in an entirely different way from the
dividing cell, perhaps by swelling or by increase in size of vacuole, as in

PATTERNS OF DEVELOPMENTAL PROGRESS 717

many plant cells. If, for any reason, they resume division, they lose their
visible differentiation, though not necessarily their specificity; and the
earlier type of growth by synthesis of protoplasm may again appear. In
regeneration of striated muscle, for example, cytoplasm without visible
differentiation accumulates about the muscle nuclei, and spherical divid-
ing cells with protoplasmic growth result; later they cease dividing and
differentiate into muscle. Many other similar cases appear in reconstitu-
tions.

In all except spherical organisms growth differs in rate and amount
in different directions and is the chief factor in determining the specific
forms of organisms and their parts. Different kinds of growth may occur
in different parts: some may grow by synthesis of protoplasm, others by
swelling, deposition of skeletal material, etc. Changes in rate and char-
acter of growth may also occur in a region or organ in the course of de-
velopment. In general, growth rate decreases with progress of develop-
ment, at least after early embryonic stages; but the decrease may be modi-
fied by growth cycles, by appearance of new growth centers, by meta-
morphosis, and by other factors. Differential growth of certain parts, the
hydroid tentacle, the sea-urchin archenteron, and the amphibian optic
primordium, becomes evident at certain developmental stages; that
of others, hydroid stolons, arms of the sea-urchin pluteus, and the am-
phibian limb, at other stages. In regenerative reconstitution growth rate
of regenerating parts is usually much above that of others. In starv-
ing planarians growth of the regenerating head and posterior end acceler-
ates reduction of the whole. Growth patterns of reconstitution by re-
organization without regeneration, as in Tuhularia and in Corymorpha,
do not show these extreme differences in growth but approach more nearly
embryonic patterns. Growth of protoplasm is very generally characteris-
tic of earlier embryonic stages and earlier stages of particular organs; as
development progresses, other kinds of growth appear. The plant axis
provides an interesting example of both spatial and chronological differ-
ences in growth. Growth by increase of protoplasm is characteristic of
earlier embryonic and bud stages and persists in the vegetative tip
throughout its growing life ; but as cells become farther removed from the
tip by its continued growth, protoplasmic growth decreases or ceases,
vacuolization begins, and most of the increase in size of the plant results
from enlargement of the vacuole, from formation of substances that swell
by uptake of water, from formation of cellulose, and from deposition in
storage organs of starch. But whatever the changes in rate and kind of

7i8 PATTERNS AND PROBLEMS OF DEVELOPMENT

growth during development, spatial and chronological growth patterns
are orderly and definite in axiate organisms.

Most, perhaps all, growth patterns can be altered experimentally in
one way or another: some are altered by autoplastic, homoplastic, or
heteroplastic transplantation to other regions, though some growth pat-
terns tend to persist under these conditions. Changes in external environ-
ment of the whole organism are highly effective in altering growth pattern,
as the differential modifications of development show (chaps, v-vii). Al-
terations in nutrition, mineral salts, etc., may also be effective. Growth
patterns are evidently related in some way to the gradient patterns of
development, but the relation is not necessarily direct and simple.

A characteristic of malignant growths, and apparently an essential
factor in malignancy, is lack of a definite organismic pattern of growth and
differentiation. More or less differentiation of cells, according to the organ
of origin of the neoplasm may take place, but orderly pattern on a multi-
cellular organismic scale does not appear.

SPECIES-SPECIFIC GROWTH RATES AND SIZE FACTORS

Transplantation experiments, particularly those with amphibian em-
bryonic materials, have brought to light certain interesting and significant
facts concerning growth rate and size of certain organs in organismic en-
vironment of another species, as a few examples will show. When sufii-
ciently fed, Amhlystoma tigriniim grows much more rapidly than A. pimc-
tatum and attains about double the latter 's size. With feeding up to the
maximum intake, optic and limb primordia transplanted from one to the
other species show, during most of the larval development, the same
growth rate as the normal organs on the donor species and, consequently,
become very different in size from the host organs. Before metamorpho-
sis, however, the eye and the limb of A . punctatum on A . tigrinum become
slightly larger than the normal organs of ^. punctatum.^ In earlier experi-
ments it had been found that under ordinary nutritive conditions A . ti-
grinum eyes and limbs on A . punctatum hosts grew not only more rapidly
than the host organs but more rapidly than those of the donor-species
control (Harrison, 1924; 1929a, b). With A. tigrinum as host, growth of
A . punctatum organs is retarded. The work of Twitty and Schwind showed
that these alterations in growth rate were due, at least in large part, to
differences in "nutritive level," that is, to underfeeding of the A. tigrinum
larvae which have the higher nutritive requirement. Twitty and Schwind

5 Twitty and Schwind, 1928, 1931; Schwind, 1931.

PATTERNS OF DEVELOPMENTAL PROGRESS 719

suggested that the experiments indicate a similar organismic environment
regulating growth rate in the two species, rather than fixity of the intrinsic
rate. With this possibility in mind, Twitty made heteroplastic transplan-
tations of optic primordia, using individuals of the same two species but
of different age or growth stage and found that growth of the eye, rela-
tive to that of the host, could be retarded or accelerated. With older
A. pundatum as host, younger eyes of A. tigrinum grow very rapidly,
while growth of host eye and normal eye is prevented by light feeding.
With younger A . tigrinum as host, growth of transplanted eye is inhibited,
even with liberal feeding and rapid growth of host. "Expression of the
growth capacity of the eye is primarily a function of the relation between
the physiological conditions in the organ and in the environment provided
by the host, rather than of mere growth in size of the latter" (Twitty,

1930).

When the optic vesicle alone, without lens or corneal ectoderm, is
transplanted from one species to the other, the developing lens apparently
influences growth of the transplanted eye. The optic vesicle of A . tigrinum
transplanted to A . pundatum grows more rapidly than the eye of the
host species but more slowly than an A. tigrinum transplant of vesicle
with lens ectoderm. This retardation is attributed to the host lens, which
is "too small" for the transplanted optic cup. Conversely, the A. punc-
tatum vesicle without lens ectoderm, transplanted to A . tigrinum, grows
more slowly than that of the normal host but is accelerated by the
A . tigrinum lens, which is "too large." Transplants of lens ectoderm alone
from A . tigrinum to A . pundatum give rise to lenses too large for the eye.
Their growth is retarded, but growth of the host optic bulb is increased.
In reciprocal transplants of lens ectoderm the lenses are too small, but
their growth is accelerated and that of the bulb retarded.^ Evidently the
species-specific growth rates of these parts of the eye, which normally
develop in a definite relation to each other, may be mutually altered when
the parts originate from different species. This seems not to be true for
all parts of the embryo, for heteroplastic transplants of portions of the
shoulder girdle between the same two species do not influence growth of
host parts but may themselves be altered in form. On the other hand,
transplants of the larger limb primordium of A . tigrinum to A . pundatum

^ Harrison, 1929a, b; Rotmann, 1939, in experiments with other species. It has also been
shown that heteroplastic eye transplants between the same two species influence growth of
various parts of the brain, of the cartilaginous capsule developing about the eye, and of the
eye muscles (Twitty, 1932).

720 PATTERNS AND PROBLEMS OF DEVELOPMENT

which become articulated to the shoulder girdle of the host induce hyper-
trophy of the latter. Also, the A tigrinum limb on A. pundatum usually
induces hyperplasia, and the reciprocal transplantation, hypoplasia of
the spinal ganglia, innervating the limb (Schwind, 1931, 1932). Hetero-
plastic grafts of heart primordia between the same two species give hearts
functionally donor but with growth rate altered toward that of the host
(Copenhaver, 1939). AmUystoma tigrinum somites grafted in place of
A . punctaturn somites become much larger than those of the host (Det-
wiler, 1938).

According to these data, species-specific growth rates and amounts show
more or less tendency to persist with heteroplastic grafting but can be
altered by relations to the host. That different basal metabolic rates of
different species and stages may be factors in the experimental results
seems not improbable. The "nutritive level" may be largely metabolic
level. Amhlystoma tigrinum eyes and limbs on A. punctatum under ordi-
nary nutritive conditions grow even more rapidly than in the donor spe-
cies, perhaps because with the low metabolism of the host under sub-
maximal feeding they are able, in consequence of their high metabolism,
to obtain even a larger relative amount of available nutrition than under
normal conditions. Older A. punctatum eyes transplanted to younger
A. tigrinum are retarded, perhaps because of their low metabolism, com-
pared with that of the host.

Final size of the individual and of its parts is dependent not only on
intrinsic growth rate but on initial size and length of growth period, and
also, of course, on amount and adequacy of nutritive material available
and on other environmental conditions. Primordia of various organs,
eye, limb, balancer, etc., in A. tigrinum are initially formed on a larger
scale than those of A. punctatum. In "normal" environment most animal
species attain more or less definite and characteristic size and proportion
because both intrinsic genetic factors and environmental conditions do
not differ greatly for different individuals. With alteration in organismic
or external environment, size and proportions may be altered, as is evi-
dent from the above experiments and from the differential modifications
of development discussed in chapters v-vii.

GROWTH GRADIENTS

Gradients in rate or amount of growth are very generally characteristic
of all stages of development. In early stages growth is generally not spa-
tially localized by sharply defined boundaries but decreases radially,

PATTERNS OF DEVELOPMENTAL PROGRESS 721

asymmetrically, or axially from regions of highest rate of growth, the
growth centers. Adventitious buds of plants and buds in many animals
are, in their earliest stages, radial growth-gradient systems (see Figs. 1-4)
and become axial, heteropolar gradient systems in consequence of differ-
ential growth. Many organs — tentacles of most coelenterates, bryozoa,
and other forms; arthropod appendages; vertebrate limbs; etc.— begin
their growth as budlike systems, but many of them are not strictly radial.
The hydra bud shows a differential in tentacle development, apparently
in relation to the longitudinal gradient of the parent body (pp. 634-35).
In the amphibian limb an anteroposterior, and later a dorsiventral, differ-
ential, both influencing later growth, are determined in relation to the
general body axes; but the longitudinal axis of the limb is the result of
differential growth in the asymmetrically radial primordium. Even if
there is no growth of the whole in early embryonic development, for ex-
ample, during cleavage, there is usually more or less differential growth,
the regions about the apical pole growing more rapidly than, or at the
expense of, other regions. The growth-gradient systems of early stages
usually, if not always, coincide with the gradient systems otherwise in-
dicated or demonstrated. Primordia of organ systems and organs usually
become evident to the eye as growth-gradient systems. Various organ
primordia of the chick embryo in very early stages are characterized not
only by growth-gradient systems but by dye-reduction gradients in low
oxygen after staining, the two being coincident, as far as can be deter-
mined. The growth-gradient system of the anal arm of the sea-urchin
pluteus can be completely obliterated by inhibiting agents; and when it
is so obliterated, neither dye-reduction nor susceptibility gradients ap-
pear.

In the course of development the boundaries of organs and organ sys-
tems usually become more definite, and growth of these organs and sys-
tems is correspondingly limited; but within those limits growth gradients
may still be present, and the body as a whole may also show growth
gradients in these later stages. These are by no means always simple and
often change in form with time. Data and graphs for many such gradients
have been given by Huxley (1932, chap. iii). In some crustacean append-
ages there is a decrease in both directions from a growth center at a cer-
tain level. Also, the steepness of the gradients and the position of the
center may change, and new centers may appear in some of these ap-
pendages. As development progresses, the growth gradients become in-
creasingly species-specific in their spatial characteristics, though they may

722 PATTERNS AND PROBLEMS OF DEVELOPMENT

differ in certain organs of the two sexes, male and female chelae of cer-
tain decapods, for example, or even on the two sides of the individual
body, in decapods with unequal chelae. Although the growth gradients
of later stages may differ widely from the gradient patterns of early de-
velopment, it appears, beyond question, that they are in some way re-
lated to, and developmental consequences of, the earlier pattern and the
specific constitution of the protoplasm in which it appears.

SOME OTHER ASPECTS OF GROWTH PROBLEMS

Certain inorganic systems which show a sort of growth and develop-
ment have often been compared with growth and form in organisms.
The analogy of crystal growth and reconstitution of form to organismic
growth has been pointed out repeatedly, and a crystalline structural pat-
tern has been postulated by some biologists as the basis of developmental
pattern."^ Various physicochemical experiments have been devised in
which diffusion and precipitation of inorganic substances in a colloid
substrate result in growth, often definitely directed, of a precipitate and
development of form. Inorganic "cells" have been made, and various
simulations or models of cell form have been produced, with inorganic
substances. The Liesegang rings, patterns resulting from diffusion and
precipitation of inorganic salts in a gel substrate under certain conditions,
have been regarded by some as highly significant in relation to organismic
form. Under certain conditions surface tension may be a factor in pro-
ducing definite form. Tension and pressure influence growth and are es-
sential factors in determining form of various parts of organisms. The
book Growth and Form by D'Arcy Thompson (191 7) presents a most in-
teresting and valuable discussion of these data and of the physical char-
acteristics and problems of growth.^

Mathematical analyses of various aspects of growth and form in organ-
isms have produced many interesting results which cannot be discussed
here. Certain of them, particularly those having to do with organismic
form, are presented in Thompson's book. A recent analysis of differential
or relative growth, that is, of gjowth rates of parts or organs in relation
to rate of the whole organism, formulates a law of constant differential

7 See pp. 5, 296, 394, 629, 694.

8 See also Liesegang, 1907, and other papers for periodic patterns resulting from diffusion
of salts in a colloid substrate, and the summary by Zeiger, 1939, of papers concerned with pat-
terns of this type. Various inorganic growth patterns and sunulations of organismic growth
and form are described by Leduc (1910).

PATTERNS OF DEVELOPMENTAL PROGRESS 723

growth ratio. This law states that the ratio of growth rate of certain
parts to growth rate of the whole body is constant for considerable pe-
riods.^ The data on which the law is based concern growth in stages
following histological differentiation and are chiefly from arthropods and
vertebrates; but in these stages the law has been found, according to
Huxley (chap, i), to hold for a considerable number of organs of both
animals and plants. This growth of parts at a different rate from the
whole is heterogenic or allometric growth." It results in a continuous
change of proportion of the part to the whole, which may be either posi-
tive or negative. However general the validity of the law may prove to
be for growth ratios of parts to wholes in later developmental stages, the
question of its significance for many growth patterns of embryonic stages
remains open.

GRADIENT PATTERNS AND EVOLUTION OF MORPHOLOGICAL FORM

Quite apart from the question whether a gradient pattern is the pri-
mary pattern of development, it is evident that evolution of morphological
form is, to a considerable degree, a matter of change in gradient pattern.
Although the characteristics of a gradient pattern depend on the constitu-
tion of the protoplasm in which it appears, not on the initiating factor, it
and the resulting development can be altered experimentally in a single
species-protoplasm by altering physiological condition. The differential
modifications of development are cases in point (chaps, v-vii). How ex-
treme such modifications may be is indicated by the differential modifica-
tions of echinoderm embryonic development (chap. vi). Alteration of form
and proportions of the pluteus, exogastrulation, and decrease or oblitera-
tion of differences along polar and ventrodorsal axes by differential inhi-
bition, changes in form and proportion in the opposite direction by the
secondary modifications of differential tolerance, conditioning, or recovery
— all these in their various degrees result from differential physiological
effects of external agents on gradient patterns. Many of these modifica-
tions are highly suggestive as regards evolution of echinoderm larval
form. For example, certain degrees of differential inhibition of asteroid

' The law is expressed by the equation y = bx^, in which x is the magnitude of the animal,
as determined by standard linear measurement or by its weight minus the weight of the
organ concerned, y is the magnitude of the differentially growing organ, and b and k are con-
stants (Huxley, 1932).

" Originally called "heterogonic" by Huxley, 1932; the term "allometric" was suggested
by Huxley and Teissier, 1936.

724 PATTERNS AND PROBLEMS OF DEVELOPMENT

development with an agent having high differential action, such as LiCl,
give forms resembling sea-urchin gastrulae. With less extreme differential
action starfish larvae approaching crinoid larvae in form can be pro-
duced. With secondary modification of gradient pattern by tolerance,
conditioning, or recovery, echinid larvae showing some approach to the
asteroid larval form result. These experimental modifications suggest that
evolution of echinoderm larval form has resulted in part from changes in
gradient pattern in consequence of change in specific constitution of the
protoplasm.

The various head forms resulting from mediolateral differential inhi-
bition in planarian reconstitution, constituting a continuous series from
normal to acephalic, with obliteration of parts progressing laterally from
the median region (pp. 177-96), and the essentially similar results of
differential inhibition in vertebrate development (chap, vii) provide fur-
ther illustrations of the dependence of forms and proportions on physio-
logical condition of the protoplasm. The experimental alterations of scale
of organization (chap, x) show another aspect of this dependence. In all
these and many other developmental modifications the gradient pattern
has been altered differentially, with resulting alterations in rates and
amounts of growth and in localization, degree, or even presence or ab-
sence of certain differentiations. Since these differential modifications of
growth and form appear to be results of primarily quantitative differential
alterations of gradient pattern, they suggest the possibility that at least
some of the characteristic differences of growth and form in related species
may also be results of similar alterations, primarily quantitative, rather
than specific in character, so far as gradient patterns are concerned. Length
and steepness of gradients, rates and amounts of growth, and localizations
of new growth centers may be altered by genetic changes in protoplasmic
constitution, as well as by environmental conditions.

By the use of Cartesian co-ordinates Thompson (191 7, chap, xvii) has
made evident graphically some of the changes in distribution of growth
in the bodies and in various parts of related species. The procedure con-
sists in applying a network or grid of rectangular co-ordinates to the out-
line of the body or part concerned in one species and deforming or trans-
forming the co-ordinate grid in a regular manner so that it becomes or
approximates a corresponding co-ordinate system for the body or part
to be compared with the first. The deformation or transformation re-
quired in the second case indicates, in a general way, when compared with
the first, the differences in growth distribution between the two in certain

PATTERNS OF DEVELOPMENTAL PROGRESS

725

dimensions. The method can, of course, also be used for different develop-
mental stages of the same species. Values and limitations of the method
are discussed by Huxley (1932, chap. iv).

A few of Thompson's figures are reproduced here. In Figure 222 the
method is applied to carapaces of several genera of crabs. In Figure 223

Fig. 222, A-F. — Transformations of Cartesian co-ordinates applied to carapaces of different
genera of crabs. A, Geryon; B, Coristes; C, Scyramathia; D, Paralomis; E, Lupa; F, Chorinus
(from Thompson, Growth and Form, 19 17).

it is applied to the whole bodies, viewed laterally, of two related teleosts.
In this case transformation of the rectangular co-ordinates, applied in
Figure 223, A, to Diodon, into a system approximately hyperbolic and
transferring the outline of A to corresponding points in it gives, in Fig-
ure 223, 5, essentially the outline of Orthagoriscus, a genus closely related
to Diodon but very different in form. Evidently, difTerential growth is
much greater in the dorsiventral dimension of the posterior region in
Orthagoriscus than in Diodon. In Figure 224 the transformations required

01 2 3 4 5 6 0^

Fig. 223, A, B. — Transformation of Cartesian co-ordinates from Diodon, A, to the closely
related Orthagoriscns, B (from Thompson, Growth and Form, 1917)-

1 J

""^

"^

\

A

/

Ji

//A

\ "^

^ — ( ^

-/

's^

0

\ J

I t

I 4 5

B

Fig. 224, A-C. — Transformations from human skull, .1, to skull of chimpanzee, B, and
baboon, C (from Thompson, Growth and Form, igij).

PATTERNS OF DEVELOPMENTAL PROGRESS 727

for the skull of the chimpanzee (B) and the baboon (C), as compared with
the human skull (A), are shown. The transformations in B and C are
of the same order, differing only in degree. That alterations of gradient
patterns are concerned in the differences in form indicated in the figures
seems evident.

If gradient characteristics, length, steepness, etc., are concerned in
determining localization of particular differentiations, it is evident that
genetically determined changes in these characteristics may alter locali-
zations of organ systems and organs relatively to others. The kind of
differentiation that takes place at any particular gradient-level may be
altered by other genetically determined changes in constitution of the
protoplasmic substrate. If gradient patterns are essential factors of de-
velopment, it is to be expected that genetic changes altering gradient
patterns may often involve features apparently only slightly related or
quite unrelated. The genetic change may alter whole gradient systems,
either the general systems of the whole body or the systems of particular
organs, and so alter localization, differential growth, and differentiation
of many parts. The possibility that a single mutation may determine
many such alterations can scarcely be denied. In so far as genetic changes
permitting survival and reproduction have effects of this sort on the more
general gradient patterns, organisms or organ systems evolve more or
less as wholes with orderly relations of their parts.

APPENDIX I

Some further consideration of methods of determining respiration, as used in
attempts to discover whether regional respiratory differences are present along the
polar axis and of objections and criticisms advanced and the grounds on which they
are based, will perhaps serve to show more clearly some of the difficulties involved and
the precautions necessary in the use of the methods in this way.

In an attempt to determine whether respiratory gradients exist, Shearer (1924)
stated that there are serious objections to the Winkler method but did not say what
they are. Parker (1929, p. 422) agrees with Shearer but also fails to state what the
objections are. In a later paper Shearer (1930, p. 264) says that the error resulting
from the discharge of slime and tissue fluids into the water is always very great where
a large number of animals is employed, but he presents no data to confirm his asser-
tion or to show how great the error may be. In fact, there is no evidence in these papers
to show that either Shearer or Parker has ever used the Winkler method. Needham
(1931, p. 586) also questions the value of data obtained by Hyman with the Winkler
method on the basis of a personal statement of opinion by Shearer, but again without
presenting any evidence in support of his view. These somewhat dogmatic expressions
of opinion without evidence to support them can scarcely be regarded as valid scientific
criticism and are certainly an inadequate basis for discarding a large body of highly
consistent positive data. Moreover, in the use of the Winkler method in connection
with the gradient problem the relative, rather than the absolute, rates of oxygen con-
sumption of similar pieces from different body-levels, or of the same pieces at different
times, are the important data. Since equal, or approximately equal, numbers of pieces
and equal test periods have been used in lots to be compared, any possible error due
to slime production or other substances from the pieces should be approximately the
same in different lots. Actually, however, pieces of Dugesia dorotocephala used in
many determinations do not produce large quantities of shme when undisturbed in
normal environment, as they are during respiratory periods for Winkler determina-
tions; and any slime that is produced does not immediately spread through the water
but remains on the wall of the container, or, in the case of small pieces which do not
move about, it may form an envelope about the piece resembling a cyst, and the piece
may remain in it for several days. Samples of water taken from the container for analy-
sis contain httle or none of this slime. Cut surfaces of planarian pieces, hydroids, and
other forms contract rapidly after section, so that the discharge of tissue fluids occurs
only during a short time. With planarian and hydroid pieces the contraction occurs
before the pieces can be made ready for the respiratory period; moreover, determina-
tions were made not only soon after section but also several hours later. Attention
was particularly called to these points by Hyman (19166) on first using the Winkler
method, and determinations have been made repeatedly, showing that error from these
sources is negligible (Hyman, 193207). The criticisms of Shearer and Needham are with-
out foundation. As regards the question of possible nitrite error in Winkler determina-

729

730 PATTERNS AND PROBLEMS OF DEVELOPMENT

tions, a question raised in connection with other work, with this method, Allee and
Oesting (1934) have shown that Hyman's data on planarian pieces from different
body-levels are not open to criticism. A recent comparison of the Winkler method
with respirometer and Van Slyke methods has shown the Winkler method to be as ac-
curate as the others (J. Wilder, 1937)-

As a matter of fact, except for determinations with very small amounts of material,
the Winkler method has certain very definite advantages over most respirometer meth-
ods, particularly with aquatic material easily excited to movement. The material is
in normal environment and entirely undisturbed during the respiratory period, and
motor activity is reduced to a minimum. With most respirometer methods the shak-
ing and the starting and stopping tend to induce movement; and even if they do not
induce visible movement, their effects on respiration are not known.

The assertion, unsupported by proof, was also made by Shearer (1930) that the
differences in oxygen consumption found by Hyman in planarian pieces are too small
to be significant; that they are statistically significant was shown by Hyman (1932(7).
The high degree of consistency of Hyman's data on pieces from different levels, on
age differences, on effects of feeding and starvation, and on reconstitution in Dugesia
with repeated determinations constitutes strong evidence that they represent real
respiratory differences. Before they can be discarded, equally consistent and con-
clusive evidence must be presented to show that they are incorrect. Moreover, it is
also of some interest to note that these differences in oxygen consumption are closely
paralleled by differences in susceptibility to cyanide, and the differences at different
body -levels, by dye-reduction gradients.

Averages of determinations of oxygen consumption by a respirometer method made
by Shearer (1930) on two planarian species show a respiratory gradient; but, because
of the great variations and because of observation of greater motor activity in anterior
than in posterior pieces, the author holds the opinion that the respiratory gradient
observed results from the differences in motor activity. Apparently he made no
attempt to decrease motor activity, though it is possible, at least with most planarian
species, to eliminate it almost entirely by using short pieces instead of large fractions
of the body length. There is no evidence in his paper that the question whether a
temporary increase of respiration followed section was considered, nor are there any
data concerning the effect on respiration at different body-levels of presence or absence
of food and digestive activity in the gut; it is not stated whether the heads of the an-
terior pieces were removed; motor activity is greater when the head is present. The
data are not given in full but only as averages with a few of the determinations to
show the variation. Determinations for the two species are not given separately,
although the forms which he calls ''Planaria nigra" and "P. ladea" (incorrectly) sup-
posedly belong to different families and genera. The possibility that the great varia-
tion found may be due to the method and lack of proper precautions is not considered;
and because of this variation, the consistency and uniformity of Hyman's data are
criticized. He is apparently also ignorant of the fact that both Hyman and Parker
used species with a posterior zooid {D. dorotocephda and D. tigrina), while the forms
which he used, if they can be properly identified, have no such zooid. The posterior
zooid has a respiration about as high as that of anterior regions of the anterior zooid.
Concerning determinations made on Thysanozoon, a polyclad, Shearer (p. 263)

APPENDIXES

731

makes the remarkable statement that the ratio of oxygen consumption in different
pieces "was just about the same as the relative amount of movement of these parts
during the course of the measurements." In a determination on anterior and posterior
halves of Thysanozoon the pieces remained 48 hours in the respiratory chamber before
determinations and had apparently come to rest, but "some movement invariably
took place at long intervals." In this case the oxygen consumption of the posterior
half was 26 per cent less than that of the anterior half. The question whether some
movement at long intervals is sufficient to account for so great a difference is not con-
sidered. In view of the obvious lack of the precautions necessary in determinations of
this sort, Shearer's data cannot be regarded as having any significance whatever, and
certainly not as valid evidence against the highly consistent data which have been
obtained in repeated experiments performed with proper care.

The method of comparative estimation of CO2 production used with planarians
and with Corymorpha has been criticized by Parker (1929) and by Needham (1931,
p. 586) on the ground that precautions were not taken "to ensure that the acidity
measured was due to carbon dioxide and not to other acids." Whether other nongase-
ous or nonvolatile acids or acid-producing substances than CO2 are produced is
easily determined by bubbling air through the indicator solution or by allowing it to
stand exposed to air in a thin layer after change in color by, and removal of, animals
or pieces. In one of the earlier papers concerned with this method attention was called
to this point, and it was stated that with this procedure the color does return (Child,
igigb). It was not considered necessary to repeat this statement in later papers; but
the test was consistently made, and consequently the criticisms are not justified.
With this method the material is undisturbed during the respiratory period, but the
method used by Parker involves continuous movement of the apparatus, and a small
amount of water is desirable. With sluggish forms like Stylochus (Watanabe and Child,
1933) consistent results can be obtained. Parker says, regarding the planarian pieces
used in his experiments: "Ordinarily they remained quietly in the bottom of the
scoop," and, as regards Nereis pieces, that they "remained in relative quiescence."
Here, as in the respirometer methods, the question of the efifect of motor activity on
respiration arises. There can be no doubt that with the Winkler method motor activity
is less than with the other methods used in work of the kind under consideration;
also, the use of a considerable number of pieces eliminates or decreases the elTect of
individual differences. At best, however, respiratory determinations on pieces from
different body-levels of adult animals constitute only contributory evidence on the
gradient problem, but their close agreement with other lines of evidence increases
their value.

For further critical discussion see Watanabe and Child, 1933.

APPENDIX II

Methylene blue and Janus green have been found most useful in work on differ-
ential dye reduction. The former has been used both as an oxidized dye and as a
leucobase, the latter in oxidized form. A considerable number of samples from differ-
ent sources have given essentially similar results. With most forms tested, methylene
blue is less toxic than other dyes used, except, perhaps, brilliant cresyl blue. Janus
green has proved particularly useful because of the change in color from the blue green
of the oxidized form to brilliant red with partial reduction. This makes it possible
to distinguish slight differences and slight degrees of reduction more clearly than does
the mere loss of color on reduction of methylene blue and various other dyes. With
some forms the further reduction of Janus green to the colorless leucobase has been
observed. In all samples used, Janus green has been found much more toxic than
methylene blue, and great care has been necessary in its use in order to avoid altera-
tion of the normal reduction pattern by injury of the more susceptible regions and
consequent retardation of reduction or complete failure to reduce in them. Thionine
and cresyl echtviolett have been used for comparative purposes with results similar
to those obtained with Janus green and methylene blue. In fact, in every case thus
far, reduction gradients have been the same with different dyes.

The procedure followed with small forms — protozoa, and embryonic and larval
stages of echinoderms and other animals — consists in sealing a considerable number of
individuals in a cell of small volume, usually made by drawing a ring of melted vaseline
of the desired size on a slide and covering the material in water without air bubbles.
The oxygen uptake of the animals brings about gradual oxygen decrease; consequent-
ly, reduction occurs gradually, and slight differences are distinguishable. With oxi-
dized dyes the material is either stained before sealing or a concentration of dye
determined by preHminary experiment to be nontoxic within the period preceding and
during reduction is used as medium in the cell. With this latter procedure staining
and oxygen decrease are occurring at the same time; and with certain concentrations
of dye and rate of oxygen decrease, reduction may occur in one region while another
is becoming more deeply stained. For elongated hydroids of considerable size, such
as Corymorpha, elongated cells may be made; or for these and for earthworms and
other large annelids glass tubes large enough to admit the animal may be used. Mineral
oil has been used in some cases to exclude air. With many forms rapid staining with a
relatively high concentration of oxidized dye and rapid reduction following have been
found less likely to produce toxic effects than slow staining with a lower concentra-
tion. Toxic effect seems to depend on accumulation of the dye on certain cell constitu-
ents which appear as deeply stained granules, rather than on its mere presence in the
cell. When the deeply stained granules become visible, reduction is retarded. With
slow staining the granular accumulation may occur before reduction, and the reduc-
tion picture may be altered; with rapid staining and quick reduction this effect may
be avoided. That this holds for all material is not asserted; it appears to hold for

732

APPENDIXES 733

embryonic and larval stages of echinoderms and for planarians and more or less trans-
parent oligochetes. The method, properly used, is extremely delicate and of great
value as a means of making directly visible certain characteristics of developmental
pattern in intact embryos and larval stages of small size, as well as many larger de-
velopmental stages. It also serves to show an oxidation-reduction pattern in the ecto-
plasm of ciliate protozoa and in adult individuals of many metazoa. In order to avoid
misleading results, however, it is necessary to use a wide range of concentrations and
staining periods; only in this way is it possible to make certain that an observed differ-
ential or absence of a differential is not due to differential toxic action of the dye and
consequent retardation of reduction in the injured regions (see pp. 67-70; also Child,
i936<7, b).

APPENDIX III

The following papers are more or less directly concerned with differential sus-
ceptibility, chiefly with differential death and toxic effect. Material and agents used
are given. Algae: Child, 1916^, e, 191 7<^ b, 1919/, KCN, ethyl alcohol, HgCL, CuSO^,
H ion (HCl), OH ion (KOH, NaOH), neutral red, KMn04, stahng water, also un-
published data on a number of species exposed to ultra-violet radiation and to visible
light with sensitization by eosin which show the same susceptibility gradients as with
chemical agents. Protozoa: Hyman, 1917, Amoeba, KCN; Bovie and Barr, 1924,
Amoeba, ultra-violet; Bills, 1924, Paramecium, alcohols; Child and Deviney, 1926,
Paramecium and other ciliates, KCN, NH4OH, NH4CI, NaOH, NaHC03, NaHC03-|-
C0„ CH3COOH, HCl, H.SO^, neutral red, methylene blue, ultra-violet radiation,
visible light after sensitization by eosin, lack of oxygen; Merton, 1929, Vorticella,
formol, acetic acid, pilocarpine, alcohol; Monod, 1933, ciliates, ultra-violet. Coelen-
teras Child and Hyman, 1919, three species of hydra, KCN, ethyl alcohol, ethyl ether,
neutral red, methylene blue, Janus green; Weimer, 1928, hydra, KCN; Hyman,
1920ft, Tubularia, KCN, ethyl ether; Child, 1919^, hydroids, KCN, ethyl alcohol,
ethyl ether, ethyl urethane, HCl, MgS04 LiCl, neutral red, methylene blue, ig2sa,
blastulae and planulae of hydromedusae, Phialidium, Gonothyraea, Stomatoca, KCN,
HgCL, neutral red, methylene blue, 1926a, Corymorpha, KCN, ethyl alcohol, ethyl
ether, ethyl urethane, chloretone, HCl, NaOH, NH^OH, NH^Cl, LiCl, strychnine
sulphate, nicotine, caffein, neutral red, methylene blue, hypotonic sea water, sunlight
after sensitization by eosin; E. J. Lund, 1931, Obelia, KCN. Ctenophora: Child,
1917c, KCN, 1933a, KCN, ethyl alcohol, ethyl ether, chloretone, HCl, NaOH,
KMnO^, neutral red, methylene blue. Platyhelminthes: various species of planarians;
Child, 1913ft, 1919c, 1919c, 1930, 1933c, KCN, alkahne, neutral, and acid solutions,
ethyl alcohol, ethyl ether, chloretone, NH4OH, hypotonic and hypertonic salt solu-
tions, distilled water, lack of oxygen; Behre, 1918, KCN at different temperatures;
J. W. MacArthur, 1920, acids and bases, 1921, dyes; Buchanan, 1923ft, 1926ft, effects
of anesthetics on susceptibility to KCN, 1930, distilled water and hypertonic salt solu-
tions; Buchanan and Levengood, 1939, serum antibodies; Sivickis, 1923, KCN; Hin-
richs, 1924a, caffein; Strandskov, 1934, i937, X-rays; Wiercinski and Child, 1936;
Wiercinski, 1939, supersonic vibrations; differential susceptibility of planarians to
formahn, nicotine, strychnine sulphate, ultra-violet, and radium has also been deter-
mined and is similar to that with other agents; Stenostomum, Child, 1924ft, Fig. 42,
KCN, methylene blue; Stylochus, a poly clad, developmental stages and adults; Wata-
nabe and Child, 1933, KCN, alkaline and neutral, chloretone, hypotonic and hyper-
tonic solutions, methylene blue, a-naphthol. Annulata: Hyman, 1916a, microdrilous
oligochetes, KCN; Child, igi^d, developmental stages of polychetes, KCN, HgCL,
NaOH; Hyman and Galigher, 1921, Lumbriculus, KCN; Parker, 1929, Nereis, KCN;
Castelnuovo, 1932a, Limnodrilus, KCN; Kawaguti, 1932, Branchiura, methylene blue.
Echinoderms: Child, 1915a, eggs and developmental stages of starfish, KCN, 1916a,

734

APPENDIXES 735

developmental stages of sea urchin, KCN; Galigher, 1921(7, developmental stages of
the sand dollar Dendraster, KCN, NH4OH. Vertebrates: Hyman, 1921, teleost
embryos, KCN, NH4OH, 19266, cyclostome embryos, NH4OH, CH3COOH; Bellamy,
1919, Bellamy and Child, 1924, frog embryos, KCN, HgCL^, NH4OH, ethyl alco-
hol, low temperature; Cannon, 1923, frog embryos, KCN, HgCLj; Buchanan,
1926c, chick embryos, HCN; Hyman, 19270, h, chick embryos and hearts, KCN,
NH4OH, NaOH; Hinrichs, 1927, chick embryos, ultra-violet. Two papers are com-
parative studies of susceptibility of animals from different groups to a series of related
agents or to a single agent: J. W. MacArthur, 192 1, susceptibility to basic and some
acid dyes of Paramecium, Dileptus, Hydra, Diigesia, several rhabdocoels, and micro-
drilous oligochetes; Hinrichs, 19246, ciliate protozoa. Hydra, Dugesia, microdrilous
oligochetes, ultra-violet and visible light after sensitization by eosin. For general
discussions of differential susceptibility see Child, 1914/, 19206, 19236, 19246, pp.
76-80, i928(/; Hyman, 19266, pp. 112-15; Watanabe and Child, 1933; Buchanan,
1930, 1935. Unpublished data on Volvox show the same death gradient with KCN,
ethyl alcohol, HCl, hypertonic solutions, NaCl, sea water, ultra-violet, visible light
with sensitization by eosin, and lack of oxygen in darkness. Numerous other publica-
tions concerned with action of external agents on living organisms give incidental
evidence of differential susceptibility. This seems to be particularly the case with work
on irradiation by ultra-violet. X-rays, and radium. Often, however, the use of only
one or a few concentrations or intensities which are found to produce an effect, and of
temporary exposure, make it difficult to determine whether, or to what extent, the
results represent differential inhibition, differential tolerance, or differential recovery.
The work directly concerned with the problem of differential susceptibility began
with the discovery of differences in survival time of physiologically young and old
planarian individuals and of death gradients and tolerance gradients within the indi-
viduals, resulting from exposure to certain concentrations of ethyl alcohol and KCN
(Child, igiie; 1913a, 6, d; 19146). Further investigation indicated that these differ-
ences in susceptibility were related to real dift'erences in physiological condition; and it
began to appear probable that study of susceptibility to cyanide in different organisms,
particularly in unicellular forms — forms consisting of a single cell series, such as many
algae, eggs, embryos, and other small organisms — might afford data of interest and per-
haps serve to indicate regional physiological differences which could not otherwise be
determined in small organisms. In view of the fact that cyanides and KCN had been
shown to be powerful inhibitors of many physiological oxidations, a general relation
between the differences in susceptibility to cyanide and differences in rate of respira-
tion or oxidation seemed highly probable. For this reason KCN was used extensively
in the earHer experiments, with the result that an axial differential susceptibility, as
indicated by the progress of cytolysis, disintegration, and death along the axis was
found to occur in the forms examined, both plants and animals. In the attempt to
throw more light on the problem data concerning respiratory metabolism and the
effect of KCN upon it were obtained for different regions of the body and for indi-
viduals in different physiological condition, which had been found differently sus-
ceptible to KCN. The earliest respiratory data on planarians were comparative
estimations of CO2 production on pieces from different body-levels by means of the
Tashiro biometer (Child, 1913a; Behre, 191 8). The Tashiro apparatus has been super- VIlX

736 PATTERNS AND PROBLEMS OF DEVELOPMENT

seded by other methods, but it should not be forgotten that it represented one of the
pioneer methods in the field. Later CO2 estimations were made colorimetrically (Child,
igigb, c; Robbins and Child, 1920). Differential susceptibiHty to lack of oxygen was
found to parallel cyanide susceptibility in the forms tested. Comparison of the data
appeared to indicate that differential susceptibility to KCN may serve as a rough
indicator of quantitative differences in respiratory or oxidative metabolism in many
organisms. The determinations by Hyman of the action of KCN on oxygen consump-
tion (Hyman, 191 6i; 1919a, e; 1920c) and on oxygen consumption in Dugesia, fed and
starved individuals (1919^, 1920(7), pieces and animals after reconstitution (1919c,
19236), and young and old individuals (i9i9(/) provided a more adequate basis for
comparison of respiration and susceptibility to KCN and further evidence of parallel
differentials and individual differences.

Some of the results of these investigations and the conclusions drawn from them
aroused criticism of the hypothesis that differential susceptibility to cyanide might
indicate differentials in respiratory or oxidative metabolism. Certain of these criti-
cisms were due to misunderstanding. On the basis of his findings that oxygen con-
sumption is not decreased by KCN in starved Paramecium, even in gradually lethal
concentrations, E. J. Lund (1918a, b; 1921a) held that conclusions drawn from differ-
ential susceptibility were invalid, and B. L. Lund (1918) maintained that differences
in survival time in KCN of Paramecium and Didinium individuals of different age and
nutritive condition are due to differences in permeability. The criticisms are, in large
part, misdirected because based on the assumption that Child and Hyman had main-
tained that the toxic action of KCN is exerted only on the ectoplasm of Paramecium
and on the body wall only in planarians. No such view was ever advanced by these
authors. They merely pointed out that the susceptibility gradient observed in Para-
mecium occurred only in the ectoplasm and that the gradient in planarians could be
observed with certainty only in the body wall because the concentration of KCN
which reached the internal organs and the time when a lethal concentration reached
them might depend to some extent on the susceptibility and time of disintegration of
the body wall. That the internal organs of planarians were susceptible was shown
clearly enough by their complete disintegration and particularly by the much later
disintegration of the digestive tract in starved, than in fed, animals, as compared
with the body wall; but whether definite susceptibility gradients were present in the
gut or other internal organs, and, if so, whether they were in the same direction as the
gradient in the body wall, could not be determined with any certainty.

Lund's conclusion that KCN has little or no effect on respiration in Paramecium has
been in part confirmed by later work on Colpidium (Pitts, 1931), in which a decrease
of 25 per cent was observed, and on Paramecium (Shoup and Boykin, 1931; Gerard
and Hyman, 1931), showing httle effect. According to earlier unpublished data
obtained by Hyman, which are mentioned here with her permission, a considerable
decrease in oxygen consumption of Paramecium sometimes occurs in KCN, but thus
far the factors which determine the occurrence or absence of decrease are not known.
In connection with this question it is important to note that the cyanide death gradi-
ent in Paramecium and other ciliates appears only in the ectoplasm. This, of course,
does not mean that the entoplasm is not susceptible to cyanide but only that it shows
no definite gradient, as might be expected, since it is in continuous circulation. It seems

APPENDIXES 737

to be evident that there are oxidation systems in Paramecium httle or not at all affected
by cyanide, but the data on the death gradient in KCN and on the effect of KCN on
respiration are not comparable, because the former concern the ectoplasm only, the
latter the whole organism. It is still possible that cyanide may decrease respiration
in the ectoplasm of Paramecium; but, since the ectoplasm is, to a considerable extent,
differentiated into structural constituents, some of which may not respire actively,
the total ectoplasmic respiration may be only a small fraction of the total respiration,
and its decrease by cyanide may not be evident with the methods available. It is cer-
tain that the ectoplasmic gradient of differential dye reduction in low oxygen (Child,
19346) is the same as the susceptibility gradient to cyanide and to lack of oxygen.

G. D. Allen (1919(7), Hyman (igigc), and Buchanan (19260) found that oxygen
consumption in Diigcsia is decreased by KCN, and Lund (1921/)) showed that it is also
decreased by decrease in oxygen concentration. Determinations of CO2 production in
pieces of D. agilis after certain starvation periods led Allen {igigb, 1920) to question the
data on differential susceptibility in D. dorotocephala and the inferences drawn from
them. It was shown by Hyman (1919/), 19200), however, that with a longer starvation
period than that in AUen's experiments the respiratory changes in D. agilis were
similar to those in D. dorotocephala. It was also found by Hyman (i9i9r, d; 19236)
that the differences in oxygen consumption in relation to reconstitution, physiological
age, and size and level of origin of piece were parallel to the observed differences in
susceptibility. Comparative colorimetric CO2 estimations showed differences in CO2
production parallel, in general, to the differences in oxygen consumption found by
Hyman; differential susceptibility to lack of oxygen was found to be parallel to the
cyanide differential; and it was shown that KCN and lack of oxygen are additive as
regards their effect on survival time (Child, i9i9(:). Robbins and Child (1920) showed
that differences in CO2 production, as colorimetrically estimated, paralleled the differ-
ences in oxygen consumption and susceptibility. The action of KCN on oxygen con-
sumption in various other invertebrates was also investigated (Hyman, 19166, 1919a,
e, igioc; Galigher, 19216; Allee, 1923); in all cases there was a decrease with the
higher concentrations, but in some forms a primary increase occurred with low con-
centrations.

In planarians, as in Paramecium, data on susceptibility and on respiration are not
comparable in all respects. As noted above, observations on differential susceptibility
concern the body wall because susceptibility of internal organs cannot be determined
independently of that of the body wall. In well-fed animals the gut disintegrates as
early as, or even earlier than, the body wall; but after long starvation, later than the
body wall. This difference is suggestive of a considerable difference in physiological
condition between the actively functioning and the long-starved gut. Respiratory
determinations include gut and body wall and other internal organs. Pieces from
different body-levels of well-fed planarians usually do not show any consistent evi-
dence of a longitudinal respiratory gradient. In order to obtain evidence of this gradi-
ent, as well as in most other determinations of respiration, it has been found necessary
to keep the animals without food for at least a week or two before work with them, in
order that the digestive tract may be empty and quiescent. When starved animals are
fed, respiration increases very greatly. This brief survey of the question of cyanide
susceptibility in relation to respiration in Paramecium and planarians indicates that

738 PATTERNS AND PROBLEMS OF DEVELOPMENT

with proper care differential susceptibility may be a valuable, though a rough and
imperfect, means of indicating quantitative differences in physiological condition in
naked, aquatic organisms. As will appear, particularly in chapter iv, there is a general
parallehsm not only between cyanide susceptibihty and respiration but, as far as data
are available, between susceptibihty and quantitative differences indicated by other
methods, the indophenol blue reaction, differential dye reduction in low oxygen, etc.
With further investigation the question of the physiological significance of cyanide
susceptibility has become only one aspect of the problem of differential susceptibility—
at least as far as physiological gradients are concerned— for it has been found that
axial differential susceptibility to many other agents, both physical and chemical, in
certain ranges of concentration or intensity show the same relation to the axis con-
cerned as does cyanide susceptibility. Moreover, this relation appears not only in
differential lethal action but in differential effect on development. In the light of
these facts the problem of the physiological basis of differential susceptibility has be-
come a much more general problem. It appears now as the problem of the nature of the
factors in living protoplasms that are concerned in determining that differences in
susceptibility along physiological axes in early developmental stages and in many of
the simpler organisms throughout Hfe are so largely nonspecific for so many different
agents which act on protoplasms in different ways. Some consideration of this prob-
lem is undertaken in chapter iii.

APPENDIX IV

The general coincidence of susceptibility and dye-reduction gradients and the
similarity of these gradients observed in related species raises certain questions con-
cerning the apparent absence of a secondary acropetal susceptibility gradient in the
basal region of the sea urchin Arbacia and the starfish Asterias. A re-examination of
these forms as regards both susceptibility and dye reduction is desirable; but, lacking
this, certain possibilities may be noted. The change in condition in the basal region
may conceivably occur later in Arbacia and Asterias than in the othei species studied,
though this appears rather improbable in view of the general similarity in all, except
as regards absence of mesenchyme formation preceding gastrulation in asteroids.
The laboratory notes on which the earlier papers on susceptibility of Arbacia and
Asterias (Child, i9i5«, 1916a) were based give some evidence of increased suscepti-
bility about the blastopore in both forms after gastrulation has begun. At the time
the papers were written, this was thought to be associated, at least in Asterias, with
contraction of the blastopore observed in this form. In the observations on differ-
ential death in Arbacia it was noted that in late blastulae cells were sometimes given
off from the midbasal region, the region of primary mesenchyme; but the possible
significance of this fact was not realized. Some of the differential modifications of
Arbacia development by external agents give evidence that susceptibility of the ento-
derm does increase at some stage of development; but, since they result from exposure
to the agent during the whole, or a considerable, period of development, it is not pos-
sible to determine from them at what stage the change occurs. Irregularities in differ-
ential death, that is, cytolysis of a cell or a few cells earlier than others at that body-
level, are not infrequent and have been thought perhaps to indicate that these cells
were at a more susceptible stage of the division cycle than those about them. The
cytolysis of basal cells in Arbacia was assumed to be due to this or some other incidental
factor. As regards the differential modifications of development, attention was direct-
ed chiefly to the modifications of external form and proportions; and while the condi-
tion and degree of entodermal development were recorded and drawn, the question of
relative inhibition of entoderm and ectoderm received little consideration. Re-exami-
nation of these notes on the early studies of susceptibility indicates that attention was
so sharply focused on the primary basipetal gradient and the developmental modifi-
cations resulting from it that the significance of evidences of a secondary acropetal
gradient was not realized. They seem to be an excellent example of failure to grasp the
meaning of what is actually seen and recorded because attention is directed elsewhere.
It appears possible, however, that the change in condition in the basal region of
Arbacia is less extreme than in Strongylocentrotus. The invaginated entoderm of
Arbacia is larger in relation to size of the blastocoel, and its wall is thicker than in
Strongylocentrotus; also, it does not undergo so much elongation in reaching the stomo-
deal region, and its cells do not change in shape so greatly during these stages as in the
latter form. These differences may signify that the change in condition in the basal

739

740 PATTERNS AND PROBLEMS OF DEVELOPMENT

region and the entoderm, perhaps also the mesenchyme, is more gradual and reaches
full development at a later stage in Arbacia than in Strongylocentrotus. In any case,
a reinvestigation of differential susceptibility and differential dye reduction in Arbacia
is necessary for further light on these points.

Attention must also be called to the possibility, noted elsewhere (Child, 1936a,
p. 449), that the dye -reduction picture exaggerates more or less the change in con-
dition in the basal region preceding gastrulation. The greater thickness of the wall
in the basal region of the late blastula may bring about a more rapid oxygen de-
crease, and consequently an earlier reduction in the cells of that region, than in the
thinner ectoderm. If that occurs, the differential reduction picture does not repre-
sent correctly the difference in physiological condition of ectoderm and entoderm
mesenchyme. However, this factor of thickness of wall and volume of cells does not
account wholly, if at all, for the apparent change in condition in the basal region. It
was noted in the text that reduction at all levels is most rapid at the inner ends or
surfaces of the cells, that is, the parts bounding the blastocoel. As immigration of
mesenchyme occurs, the cells which lie in the blastocoel reduce more rapidly than the
inner surfaces of any other cells of the blastula, and even before invagination the
inner surface of the entoderm reduces more rapidly than the inner surfaces of ecto-
derm cells, except perhaps those in the apical region. Since this was not the case in
earlier stages, a real change in physiological condition of mesenchyme and entoderm
has evidently occurred.

It is also an interesting question whether this change is induced in the entoderm
by the mesenchyme or is independent of it. The evidence of a similar, though appar-
ently less extreme, change in the basal region of the Patina blastula in which no pri-
mary mesenchyme is formed shows that in that form it appears in the entoderm with-
out induction.

Even though evidence of a change in condition in the basal region seems to be con-
clusive, it may perhaps still be questioned whether this region actually attains before
gastrulation a higher oxidation-level than the apical region. If the dye-reduction pic-
ture does exaggerate the change, the level may not actually be higher. On the other
hand, cytolysis and death by chemical agents may be relatively somewhat retarded
in the basal region by the greater thickness of the cell wall. If there is such retardation,
the real condition as regards these gradient patterns in the later blastula may be
intermediate between that indicated by differential death and differential dye reduc-
tion. But whatever the final conclusions as regards degree and stage, there seems to
be conclusive evidence that a change in condition does occur in the basal region,
beginning in the later blastula or at gastrulation in StrongyloccntroiHS, Dcndrastcr, and
Patina and probably at this stage or later in Arbacia, Eckinarachiiius, and Asterias.
The form of the Arbacia gastrula suggests that the change in condition in the basal
region, or at least in the entoderm, preceding gastrulation is not so great as in Strongy-
locentrotus and Dcndrastcr. The apicobasal axis of the Arbacia gastrula is relatively
shorter, its walls are thicker, and the archenteron remains thick walled during in-
vagination, most of its elongation and decrease in thickness occurring later than in the
other forms.

APPENDIXES 741

These conclusions do not exclude the possibility that the gradient pattern, either
from the beginning of development or from a later stage, involves specifically different
substances differently distributed, as Runnstrom and his co-workers maintain; but
even if this is the case, differential susceptibility, which is not regionally specific with
different agents, and differential dye reduction may be expected to indicate similar
gradient directions, as they do very generally.

APPENDIX V

Runnstrom regards his experiments with potassium-free sea water (19250) as indi-
cating that basal and ventral regions lose potassium more rapidly than other parts in
absence of external potassium and undergo more decrease in colloid dispersion. He con-
cludes that they are more permeable to potassium. His criteria of a ventrodorsal differ-
ential effect of lack of potassium are apparently based largely on individual cases of
somewhat modified development.

The polar susceptibility gradient to lack of potassium is, according to Runnstrom,
acropetal in early cleavage. This is opposite in direction to the lethal gradients ob-
served by Child with various agents and to the dye reduction gradient, but if his
identification of ventral and dorsal regions is correct, the ventrodorsal susceptibility
gradient to lack of potassium is in the same direction as the lethal gradients and the
dye reduction gradient.

Differential susceptibility to absence of potassium may perhaps involve other fac-
tors than those considered by Runnstrom. For example, there may be axial differ-
entials in distribution of potassium in the egg; or in its absence externally its internal
distribution may be altered by physiological factors, or there may conceivably be a
differential tolerance to loss of it and a differential recovery from its absence on return
to normal sea water. Withdrawal of an essential element such as potassium may have
effects very different from those resulting from addition to the medium of a toxic
agent. A susceptibility gradient to lack of potassium, whether it depends on differ-
ential permeability or on some other factor in the protoplasm, may have little relation
to the gradients indicated by nonspecific susceptibility to positive action of chemical
and physical agents or by dye reduction, and the effect may be still further compli-
cated by differential recovery after temporary exposure to potassium-free sea water,
the procedure in Runnstrom's experiments. Further experiment with a wide range of
potassium content, both above and below that of sea water and with different ex-
posure periods, appears desirable.

742

APPENDIX VI

In his experiments on differential susceptibility with Raiia temporaria as material,
Cannon (1923) used only HgCU, and the only concentration mentioned in his paper is
m/i,ooo. According to his brief statement of results, disintegration showed no definite
relation to axes or other particular features of development. Entirely aside from the
question of differential susceptibility to different agents, it is highly improbable that
regional differences in susceptibility to a particular toxic agent in a developing embryo
can have no definite relation. In the light of the great body of definite positive evi-
dence of such relation these negative results must at least raise the question whether
the procedure employed is adequate, but it does not appear that this possibility was
considered by Cannon. In a re-examination of the matter Bellamy and Child (1924)
found it necessary to separate individual eggs from the mass, to remove the jelly,
and to agitate the solution. If these precautions were not taken, differences in suscepti-
bility might depend on differential exposure rather than on differences in physiological
condition. In the use of HgCl^ it must also be borne in mind that even rather low
concentrations act as fixing or coagulating agents and kill without any cytolysis.
Except for a surface effect on more advanced stages, apparently an action on the cilia
or the ciliated cells, Bellamy and Child found differential susceptibility to HgCL- es-
sentially similar to that observed with other agents.

Cannon also calls attention to the fact, previously noted by Bellamy, that alcohol
produces changes in the basal yolk-laden region earlier than elsewhere. It was pointed
out by Bellamy and Child that this is evidently an effect of a lipoid-soluble and lipoid-
solvent agent on the yolk and has no relation to the differential susceptibility of the
protoplasm. In the amphibian egg the yolk content of the basal region is so great that
susceptibility of that region is more or less specific for fat-soluble agents.

The argument advanced by Cannon that the disintegrating tissue is toxic and that
disintegration, once initiated at any point, will continue because of this toxicity is
shown by many facts to be without foundation. First, disintegration may be ar-
rested by return of the organism to the natural environment; and the parts which have
not disintegrated may recover and develop further, or, as in planarians, undergo re-
constitution. Second, when a locahzed area or gradient arises in some region other than
the end of an axis, for example, in an organ primordium, it may remain localized
or may progress in one direction and not in others, according to the pattern of the
organ concerned. Third, in cases of partial differential tolerance with death and dis-
integration of certain regions cessation of disintegration and finally reconstitution of
the parts which disintegrated may occur with continued exposure to the agent which
originally produced the disintegration (see pp. 113-17).

In experiments on differential modification of cleavage by KCN Cannon found that
in Biifo toxicity decreased, and in Ra)ia increased, with decrease in concentration.
The frog eggs were removed from the ovary and fertilized by addition of sperm, while
the toad eggs were laid and fertilized naturally. The possibiHty that the higher con-

743

744 PATTERNS AND PROBLEMS OF DEVELOPMENT

centrations of cyanide interfered with the swelling of the jelly in the frog eggs and
decreased permeability in the membrane is the only basis that suggests itself for in-
creasing toxicity of cyanide with decreasing concentration. This seems to be the
only case known to physiology in which toxicity of KCN increases with decreasing
concentration. In the case of the naturally laid and fertiHzed toad egg the membrane
and jelly had attained their characteristic condition before the egg was subjected to
cyanide. At any rate, it appears highly improbable that opposite relations as regards
toxicity and concentration of KCN should exist in two so closely related protoplasms
as those of frog and toad. That some factor external to the egg itself is probably re-
sponsible for the results with the frog seems certain.

APPENDIX VII

The following papers are concerned with planarian head frequency: Dugesia doro-
toccphala: Child, igixb; Child and Watanabe, 1935a, length of piece, body-level,
effect of delay of anterior and posterior section; Child, 191 1/, 1920a, effect of nutritive
conditions, size of animal (physiological age), stimulation; Child, igidb, alteration
of frequency by KCN; Behre, 1918, alteration by temperature; Buchanan, 1922,
alteration by anesthetics; Hinrichs, 1924, alteration by caffein; Rulon, 1936a, 1937,
alteration by CO2, H ion, organic acids, and calcium, effects of previous conditioning
to CO2 and to H ion; F. S. Miller, 1937, alteration by strychnine. Other species:
D. tigrina ( = maculata) : Watanabe, 1935^, length of piece, body-level, effect of delay of
anterior and posterior section and section of nerve cords; Planaria lata, now also re-
garded as D. tigrina: Sivickis, 1923, length of piece and body-level; P. gonocephala:
Abeloos, 1928, length of piece and body-level; Phagocata gracilis: Buchanan, 1933.

In a lot of pieces as nearly alike as possible as regards length, level of origin, size
and physiological condition of parent animals, and reconstitutional environment
there may or may not be uniformity of head form, according to conditions. In pieces
representing a large fraction of body length and developing in normal environment
the uniformity is high — almost or quite 100 per cent normal heads. In pieces repre-
senting very small fractions of body length in normal environment the uniformity is
also high — almost or quite 100 per cent acephalic. In lots intermediate between these
extremes and in pieces subjected to experimental conditions which alter head fre-
quencies, a certain range of head forms usually appears in percentages characteristic
for the physiological and environmental conditions. Such differences undoubtedly
result in part from slight differences in physiological condition and susceptibility of
different individuals and in part from unavoidable variations in length of pieces and
level of origin. Frequencies of the different head forms can be tabulated numerically
or as percentages, and a general comparison made from these data; but for graphic
presentation and comparison it is convenient to have a value for the head frequency of
a lot rather than the individual frequencies. Such values, adequate for the compara-
tive purposes for which they are used, are readily obtained as follows: numerical values
are assigned to each of the head forms — for example, normal, 100; teratophthalmic,
80; teratomorphic, 60; anophthalmic, 40; acephalic, 20. From these values an index
for the lot is obtained by multiplying the number of cases of each head form by the
numerical value for that form and dividing the sum of these products by the number of
pieces in the lot; or, if the frequencies are expressed in percentages, multiplication of
the percentages by the numerical values and division of the sum of the products by 100
gives, of course, the same result. The procedure is clear from the formula

_ ioo»5 + 8o«4 + 6o»3 -j- 40/72 + 2o;7i
~ N '

745

746 PATTERNS AND PROBLEMS OF DEVELOPMENT

in which //5-;/i are the numbers or percentages of the different head forms (7/5,
normal, to «,, acephahc) and N is the total number of pieces in the lot or of living
pieces, or, in percentages, 100. Since it is not certain that the different groups of
head forms represent corresponding amounts of physiological difference or change,
the values assigned are arbitrary, except that they indicate the order of degree of differ-
ential inhibition. The value, /, therefore is not a true mean head frequency but merely
an index.

The recent application by W. A. Castle (1940, "Methods for evaluation of head
types in planarians," Physiol. ZooL, 13) is of particular interest as providing a basis for
more accurate general expression of head frequency data. However, the head fre-
quency curves obtained in this way differ little in most cases from those graphed
from the "head frequency indices" obtained by the method described above. The
following quotation from Castle's paper summarizes the situation:

It is not to be inferred that the originators and users of the old scale of indices intended it
to be more than an approximation for convenience. Indeed, Buchanan has stated 11923a,
p. 410] that "no claim is made that the values 5, 4, 3, 2, i, represent any exact mathematical
valuation of regenerated tissue; this is simply a convenient and consistent way of showing the
effects of the agent on head frequency." For the purpose designed, the old scale of equal in-
tervals appears in most cases to have been entirely adequate for outlining major differences
in head frequency along the axis of the planarian body or for distinguishing between the
head frequencies of experimental and control groups where the differences are not obscure.
Where such differences are slight, however, the use of a scale in which the indices or class
ranges do closely approach the values based on observed frequencies of occurrence of the re-
generated structures probably comes closer to revealing the real relationships between control
and experimental groups.

APPENDIX VIII

Most of the experimental modifications of echinoderm development recorded else-
where have resulted from action of single or a few concentrations or intensities of action
of external agents. Many are effects of artificial sea waters differing from natural sea
water in absence or increase in amount of certain components or substitution of
certain components by others not present or present only in very small amounts in
natural sea water. In spite of the great volume of experiment on echinoderm develop-
ment there have been relatively few attempts to determine for single agents the effects
of a wide range of concentrations, intensities, and exposure periods. It seems often to
have been taken for granted that the modification produced by a particular concen-
tration or intensity is specific. Also, the possibility of secondary modifications resulting
from differential tolerance, conditioning, or recovery, and opposite in direction as re-
gards form and proportions from primary differential inhibitions, has been largely
ignored. In view of this situation a few further data concerning effective ranges of
concentrations, exposure periods, etc., supplementing those in the figure legends, are
given here.

With any agent in any effective concentration or intensity a considerable range of
modifications occurs. These depend in part on differences in susceptibility of individual
eggs and eggs of different females, but difference in conditions to which individuals are
subjected at different stages of development in the container may also be a factor.
Animals remaining on the bottom instead of swimming, and particularly aggregations,
if allowed to persist, show more extreme differential inhibition with certain agents
(e.g., KCN, LiCl) than animals swimming free. On the other hand, concentration of
an inhibiting agent may become lower about an aggregation of animals because the
agent is taken up by the cells. If there is a free surface exposed to air, animals reaching
that surface may be much less inhibited than those that do not reach it. These effects
of local differences in conditions in the container can be decreased by frequent gentle
agitation, by decreasing surface exposed to air, and by use of large volume of medium.
Undoubtedly, very slight physiological or environmental differences may determine
whether the final result in a particular individual is a modification representing differ-
ential inhibition or one with some degree of differential tolerance, conditioning, or re-
covery. It is true, however, that differential inhibition is universal or predominant with
higher concentrations or intensities and longer exposure periods; with decrease in con-
centration differential tolerance and conditioning become increasingly evident; and
with certain ranges of concentration or intensity the modifications of differential re-
covery become predominant following return to water. Degrees of secondary modifica-
tion differ widely with different agents: with KCN, for example, differential tolerance
or conditioning is very slight or absent in the brief period of development, but more or
less differential recovery may occur after return to water from low concentrations. With
LiCl the secondary modifications occur to a much greater degree; with alcohol and
various other agents they occur rapidly and to extreme degrees. The differential in

747

748 PATTERNS AND PROBLEMS OF DEVELOPMENT

effect at different levels of a gradient also differs with agent. With KCN it is relatively
slight; with CUSO4 and HgClz, considerable, at least for Arbacia; with LiCl, relatively
great for all forms used. These differences are doubtless related to the particular
way in which a given agent brings about its effect.

KCN has been used in many concentrations from m/ioo with brief temporary ex-
posures to m/ 200,000, m/3 00,000, and m/400,000 with continuous exposure. Ethyl
alcohol, 3-4 per cent, is gradually lethal for Arbacia, but with temporary exposure
differential inhibition is followed by rapid and extreme differential recovery; dif-
ferential tolerance and conditioning following inhibition appear in 2 and i per
cent {Arbacia). Rapid and extreme secondary modifications follow the initial inhi-
bition in sea water to which certain concentrations of acetic or hydrochloric acid are
added, the effective agent undoubtedly being CO2 set free by the acid. In earlier experi-
ments with Arbacia the pH was not determined, but it was determined later that the
concentrations of acid added were not high enough to make H-ion concentration effec-
tive. Sodium hydroxide from m/i,ooo to m/5,000 inhibits differentially, but with
secondary conditioning in the lower concentrations and recovery in water {Arbacia).
Primary differential inhibition and secondary modifications occur in hypotonic sea
water (go and 75 per cent. CUSO4 from m/ioo,ooo to m/2, 500,000 inhibits dif-
ferentially; rather extreme forms of differential recovery follow returns to water from
even the higher concentrations; and with continuous exposure to the lower concen-
trations forms showing marked degrees of differential tolerance or conditioning ap-
pear. HgClj m/i, 000,000 is strongly inhibiting differentially and finally lethal for
Arbacia; in m/5,000,000 and m/io,ooo,ooo there is primary differential inhibition
with slight secondary modification in the higher concentration in 10-20 per cent
of the culture, in the lower concentration in 80-90 per cent, forms with large oral lobe
and wide brachial angle resulting.

Lithium chloride has been used in a large number of concentrations from m/200 to
m/io and with exposure periods ranging from i or 2 hours to continuous throughout
development and beginning at different developmental stages; experiments have
been repeated again and again with different lots of eggs, with the same concentrations
and exposures and with slight variations. Arbacia punctiilata, Strongylocentrotiis
purpuratus, S. franciscanus, Echinarachnius parma, Dendraster cxccntricns, Asterias
forbesii, and Patiria miniata have all served as material for LiCl experiments. The
special interest in this agent results primarily from its high effectiveness in produc-
ing exogastrulation; this has led many to regard it as regionally specific in action. In
the writer's experiments LiCl has been used not only in the analysis of exogastrula-
tion but for production of other modifications. With many concentrations effects of
approximately isotonic solutions and of sea- water solutions which are somewhat hyper-
tonic have been compared on the same material. In the high concentrations, which
sooner or later stop development or are lethal, the hypertonic solutions are somewhat
more effective in modifying development differentially; with temporary exposure they
usually give a higher percentage or more extreme type of exogastrulation than the
isotonic solutions; with lower concentrations no distinct difference in effect has been
observed.

According to MacArthur (1924), m/i6o is about the optimum for exogastrulation in
Echinarachnius. Higher concentrations are required for Dendraster; 100 per cent or

APPENDIXES 749

nearly of exogastrulae usually results from i6 to i8 hours' exposure to m/so, m/40, or
m/30 from the two-cell stage. Lower percentages occur with lower concentrations, in-
creasing frequency of dissociation of the entodermal region, cessation of development
and death, with higher concentrations. The higher the concentration with a certain
exposure period the farther apically entodermization extends and the more frequent
is dissociation of the original prospective entoderm. Strongylocentrotus franciscanus
and S. purpuratus are apparently somewhat less susceptible to LiCl than Dcndraster,
and the first seems to be slightly less susceptible than the second.

Patina is less susceptible than the echinoids. In m/30 for 14 hours or longer from
early cleavage only 5-10 per cent exogastrulae appear: in m/20 for 18-25 hours from
early cleavage 80-90 per cent are exogastrulae. With longer exposure to these concen-
trations more or less dissociation of the original prospective entoderm occurs, as in
echinoids. With exposure beginning in later stages the frequency of exogastrulation
and degree of entodermization decrease. Inhibition and more or less dissociation of
invaginating entoderm, particularly of the enlarging apical region, occurs in m/40 with
continuous exposure.

With concentrations of LiCl which produce exogastrulation in Patina there is usually
little differentiation of ectoderm, but with lower concentrations the same differential
modifications of form and proportion appear as with other agents, primarily differ-
ential inhibition, secondarily the opposed modifications of differential tolerance, con-
ditioning, or recovery. Extensive entodermization of prospective ectoderm has been
obtained in Patina in Janus green 1/2,000,000. In the writer's experiments the sus-
ceptibility of Asterias forbesii was about the same as that of Patiria, but MacArthur
has reported exogastrulation in that species and in Orthasterias with LiCl m/ioo to
m/i6o.

Differential inhibition from crowding probably results from lack of oxygen or from
some toxic metabolite rather than from CO2, since it may be extreme with pH no lower
than 7.3-7.4. Concentrations of the dyes used to produce differential modification have
little significance, provided they are low enough, for they are accumulated within the
cells even from extremely low concentrations in the water.

BIBLIOGRAPHY

Abeloos, M. 1928. Sur la dysharmonie de croissance chez Planaria gonocephala
Duges et sa reversibilite au cours du jeune. C. R. Soc. biol., 98. 1930. Recherches ex-
perimentales sur la croissance et la regeneration chez les planaries. Bull. biol. France
Belg., 64. — Abercrombie, M. 1937. The behavior of epiblast grafts beneath the primi-
tive streak of the chick. Jour. Exp. Biol., 14. Abercrombie, M., and Waddington,
C. H. 1937. The behavior of grafts of primitive streak beneath the primitive streak of
the chick. Jour. Exp. Biol., 14. — Adelmann, H. B. 1928. The formation of lenses from
the margin of the optic cup in eyes implanted in the belly wall of Triton, etc. Arch.
Entw'mech., 113. 1929a. Experimental studies on the development of the eye. I.
Jour. Exp. ZooL, 54. 19296. Studies, etc. II. Ihid. 1930. Studies, etc. III. Ibid., 57.
1934. A study of cyclopia in AmUystoma pundatum with special reference to the meso-
derm. Ibid., 67. — Adrian, E. D. 1932. The mechanism of nervous action. Phila. —
Agassiz, A., and Whitman, C. O. 1885. The development of osseous fishes. Mem.
Mus. Comp. Zool. Harvard., 14.— Ahlfeld, F. 1882. Die Missbildungen des Mensch-
en. Leipzig.— Albaum, H. G., and Nestler, H. A. 1937. Xenoplastic ear induction be-
tween Rana pipiens and Amblystoma punctatum. Jour. Exp. Zool., 75. — Alexander,
L. E. 1937. An experimental study of the role of optic cup and overlying ectoderm in
lens formation in the chick embryo. Jour. Exp. Zool., 75. — Allee, W. C. 1923. The
effect of potassium cyanide on metabolism in two fresh water arthropods. Amer. Jour.
Physiol., 63. Allee, W. C, and Oesting, R. 1934- A critical examination of Winkler's
method for determining dissolved oxygen in respiration studies with aquatic animals.
Physiol. Zool., 7.— Allen, G. D. 19190. Quantitative studies on the rate of respiratory
metabohsm in Planaria. I. Amer. Jour. Physiol., 48. 1919b. Quantitative studies
etc. II. Ibid., 49. 1920. The rate of carbon dioxide production in pieces of Planaria
in relation to the theory of metabolic gradients. Abstr. Amer. Soc. Zool., Anat. Rec,
17.— Alsop, Florence M. 1919. The effect of abnormal temperatures upon the develop-
ment of the nervous system in chick embryos. Anat. Rec, 15. — Alvarez, W. C. 1928.
The mechanics of the digestive tract. 2d ed. New York. — Ancel, P., et Vintemberger,
P. 1933. Sur la soi-disant rotation de fecondation dans I'ffuf des Amphibiens. C. R.
Soc. Biol., 1 14. 1935. Sur les marques electrolytiques dans I'ccuf de Rana fusca. Ibid.,
118. 1938. Sur le determinisme de la symetrie bilaterale dans I'oeuf de Rana fusca.
Ibid., 128, pp. 95, 98, loi, 412, 414, 417, 1212.— Appleman, C. O. 1918. Physiological
basis for the preparation of potatoes for seed. Maryland State Coll. Agric. Exp. Sta.
Bull. 212. 1924. Potato sprouts as an index of seed value. Ibid., 265. — Arndt, A.
1937. Rhizopodenstudien. III. Arch. Entw'mech., 136.— Assheton, R. 1896. An ex-
perimental examination into the growth of the blastoderm of the chick. Proc. Roy.
Soc, 60.— Astbury, W. T. 1937. Protein structure from the viewpoint of X-ray anal-
ysis. C. R. Lab. Carlsberg, Ser. chim., 22. 1939. X-ray studies of the structure of
compounds of biological interest. Annual Rev. Biochem., 8. Astbury, W. T., and Bell,
Florence O. 1938. Some recent experiments in the X-ray study of proteins and related

75°

BIBLIOGRAPHY

751

structures. Cold Spring Harbor Symp. Quant. Biol., 6. — Avel, M. 1932. Sur une ex-
perience permettant d'obtenir la regeneration de la tete en absence certaine de la chaine
nerveuse ventrale ancienne chez les Lombriciens. C. R. Acad. Sci., Paris, 194. — ^Awer-
inzew, S. 1909. Studien iiber parasitische Protozoen. I. Arch. F. Protistenk., 14.

Babicka, J. 1934. Die Wuchsstoffe. Beihefte Bot. Centralbl. A, 52. — ^Baer, K. E.
von. 1828. tjber Entwickelungsgeschichte der Thiere. Entwickelungsgeschichte des
Hiihnchens im Eie. Konigsberg. — Baschlin, K. 1932. Histologische Untersuchungen
iiber Riickbildungserscheinungen an Siisswasser- und Meereshydroiden. Zool. Jahrb.
Abt. allg. Zool., 52. — Bagg, H. G. 1922. Disturbances in mammalian development
produced by radium emanation. Amer. Jour. Anat., 30. — -Bagini, M. 1923. Alcuni
effetti della centrifugazione sulle uova segmentate di Bujo vulgaris. Monit. Zool. ital.,
34- — Bailey, P. L. 1930. The influence of the nervous system in the regeneration of
Eisenia foetida Savigny. Jour. Exp. Zool., 57. 1939. Anterior regeneration in the
earthworm, Eisenia, in the certain absence of central nervous tissue at the wound re-
gion. Ibid., 80. — Balbiani, E. G. 1873. Memoire sur le developpement des Araneides.
Ann. Sci. nat., Ser. 5, 18. — Baldwin, C. R. 1915. The action of ultraviolet rays upon the
frog's egg. Anat. Rec, 9. 1919. The artificial production of monsters conforming to a
definite type by means of X-rays. Ibid., 16. — Balinsky, B. I. 1925. Transplantation
des Ohrblaschen bei Triton. Arch. Entw'mech., 105. 1926. Weiteres zur Frage der
experimentellen Induktion einer Extremitiitenanlage. Ibid., 107. 19270. Xenoplas-
tische Ohrblaschentransplantation zur Frage der Induktion einer Extremitiitenanlage.
Ibid., 110. 19276. tJber experimentelle Induktion der Extremitatenanlage bei Triton
mit besonderer Beriicksichtigung der Innervation und Symmetrieverhaltnisse dersel-
ben. Ibid. 1929. tJber die Mesodermverschiebungen bei Extremitateninduktion. Ibid.,
116. 1933. Das Extremitatenfeld, seine Ausdehnung und BeschafTenheit. Ibid., 130.
1935- Selbstdifferenzierung des Extremitatenmesoderms im Interplantat. Zool. Jahrb.
.\bt. Allg. Zool, 54. 1937a. Zur Frage der Natur der Extremitateninduzierenden Wir-
kung. Ibid., is^- I937&- tJber die zeithche Verhaltnisse bei der Extremitiitinduktion.
Ibid. — Ballantyne, J. W. 1904. Manual of antenatal pathology and hygiene, II.
Edinburgh.— Ballowitz, E. 1890a. Untersuchungen iiber die Struktur der Sperma-
tozoen. Vogel. Arch. mikr. Anat., 32. 18906. Untersuchungen etc. Insekten. Zeit-
schr. wiss. Zool., 50. 1890c. Untersuchungen etc. Arch. mikr. Anat., 36. — Bandier, J-
1936. Histologische Untersuchungen iiber die Regeneration von Landplanarien. Arch.
Entw'mech., 135. — Banki, 0. 1927. Die Lagebeziehung der Spermieneintrittsstelle
zur Medianebene und zur ersten Furche etc. Verb. anat. Ges., 1927. — Bardeen, C. R.
1902. Embryonic and regenerative development in planarians. Biol. Bull., 3. 1903.
Factors in heteromorphosis in planarians. Arch. Entw'mech., 16. 1907. Abnormal de-
velopment of toad ova fertilized by spermatozoa ex-posed to Roentgen rays. Jour. Exp.
Zool., 4. 1909. Variation in susceptibility of amphibian ova to X-rays at different
stages of development. Anat. Rec, 3.— Bartelmez, G. W. 1912. The bilaterality of
the pigeon's egg. Jour. Morphol., 23. 1918. The relation of the embryo to the prin-
cipal axis of symmetry in the bird's egg. Biol. Bull., 35.— Barth, L. G. 1934a. The
effect of constant electric current on the regeneration of certain hydroids. Physiol.
Zool., 7. 19346. The direction and magnitude of potential differences in certain hy-
droids. Ibid. 1934c. The chemical nature of the amphibian organizer. I. Biol. Bull.,
72. 1934c?. The chemical nature, etc. II. Anat. Rec, 61, Suppl. 1938a. Quantitative

752 PATTERNS AND PROBLEMS OF DEVELOPMENT

studies of the factors governing the rate of regeneration in Tubular ia. Biol. Bull., 74.
1938 b. Oxygen as a controlling factor in the regeneration of Tubular ia. Physiol. Zool.,
II. 1939a. The chemical nature of the amphibian organizer. IIL Ibid., 12. 1939b.
The oxygen consumption of the amphibian organizer. Abstr. Amer. Soc. Zool., Anat.
Rec, 75, 4, Suppl. 1939c. Neural differentiation without organizer. Ibid. 1939c?.
Oxygen consumption of the parts of the amphibian gastrula. Proc. Soc. Exp. Biol.
Med., 42. Barth, L. G., and Graff, S. 1938. The chemical nature of the amphibian
organizer. Cold Spring Harbor Symp. Quant. Biol., 6. — Bataillon, J. E. 1900a. Blas-
totomie spontanee et larves jumelles chez Petromyzon planeri. C. R. Acad. Sci., Paris,
130. 1900 b. Pression osmotique de I'oeuf et polyembryonie experimentale. Ibid.
1901. Etudes experimentales sur revolution des Amphibiens. Arch. Entw'mech., 12.
1904. Les agents dits "specifiques" en teratogenese et en parthenogenese experimen-
tales. Ibid., 18. — ^Bateson, W. 1894. Materials for the study of variation. London and
New York. — Bautzmann, H. 1926. Experimentelle Untersuchungen zur Abgrenzung
des Organisationszentrums bei Triton taeniatus etc. Arch. Entw'mech., 108. 1928. Ex-
perimentelle Untersuchungen iiber die Induktionsfahigkeit von Chorda und Meso-
derm bei rr/Zt*;?. Ibid., 11^. 1929a. iiber bedeutungsfremde Selbstdifferenzierung aus
Teilstiicken des Amphibienkeimes. Naturwiss., 17. 1929b. tjber Induktion durch
vordere und hintere Chorda in verschiedenen Regionen des Wirtes. Arch. Entw'mech.,
1 19- 1933' Uber Determinationsgrad und Wirkungsbeziehungen der Randzonenteilan-
lagen (Chorda, Ursegmente, Seitenplatten und Kopfdarmanlage) bei Urodelen und Anu-
ren. Ibid., 128. Bautzmann, H.; Holtfreter, J.; Spemann, H.; und Mangold, O. 1932.
Versuche zur Analyse der Induktionsmittel in der Embryonalentwicklung. Naturwiss.,
20. — Beadle, L. C, and Booth, F. A. 1938. The reorganization of tissue masses of Cor-
dylophora lacustris, etc. Jour. Exp. Biol., 15. — Beams, H. W. 1937. The air turbine ul-
tracentrifuge, together with some results upon centrifuging eggs of Fucus serratus.
Jour. Marine Biol. Assoc, 21.— Beatty, R. A.; De Jong, S.; and Zielinski, M. A. 1939.
Experiments on the effect of dyes in induction and respiration in the amphibian gas-
trula. Jour. Exp. Biol., 16.— Beckwith, C. J. 1914. The genesis of the plasma-struc-
ture in the egg of Hydractinea echinata. Jour. Morphol., 25. 1927. The effect of the
extirpation of the lens rudiment on the development of the eye in Amblystoma puncla-
tum, with special reference to the choroid fissure. Jour. Exp. Zool., 49. — Behre, Ellinor
H. 1918. An experimental study of acclimation to temperature in Planaria doroto-
cephala. Biol. Bull., 35. — Beissenhirtz, H. 1928. Experimentelle Erzeugung von
Mehrfachbildungen bei Planarien. Zeitschr. wiss. Zool., 132. — Belajeff, W. 1894.
iiber Bau und Entwickelung der Spermatozoiden der Pflanzen. Flora, 79. — Bellamy,
A. W. 1919. Differential susceptibility as a basis for modification and control of de-
velopment in the frog. L Biol. Bull., 37. 1921. Note concerning the origin of polarity
in the frog's egg. A correction. Ibid., 41. 1922. Differential susceptibility as a basis,
etc. IL Amer. Jour. Anat., 30. Bellamy, A. W., and Child, C. M. 1924. Susceptibility
in amphibian development. Proc. Roy. Soc, B, 96. — Berrill, N. J. 1931- Regenera-
tion in Sabella and other sabellid worms. Jour. Exp. Zool., 58. 1932. The mosaic de-
velopment of the ascidian egg. Biol. Bull., 63. 1935. Studies in tunicate development.
IV. Phil.Trans.,Roy. Soc.,B, 225. Berrill, N. J., and Mees, Doris. 1936a. Reorgani-
zation and regeneration in Sabella. I. Jour. Exp. Zool., 73. 1936b, Reorganization,
etc. n. Ibid., 74. — Bertalanffy, L. von. 1928. Kritische Theorie der Formbildung.

BIBLIOGRAPHY 753

Berlin. 1932. Theoretische Biologic. Leipzig.— Berthold, G. 1882. Beitrage zur Mor-
phologic der Meercsalgen. Jahrb. wiss. Bot., 13.— Beyer, Kathe, and Child, C. M.
1930. Reconstitution of lateral pieces of Planaria dorotoccphala and P. maculata.
Physiol. Zool., 3. — Bickford, Eliz. 1894. Notes on regeneration and heteromorphosis
of tubularian hydroids. Jour. Morphol., 9.— Bierens de Haan, J. A. 19130. Uber
homogene und heterogene Keimverschmelzung bie Echiniden. Arch. Entw'mech., 36.
19136. tJber die Entwicklung heterogenen Verschmelzungen bei Echiniden. Ibid., 37.
— Bigelow, M. A. 1902. The early development of Lepas. Bull. Mus. Comp. Zool.
Harvard, 40. — Bijtel, J. H. 193 1. Uber die Entwicklung des Schwanzes bei Amphi-
bien. Arch. Entw'mech., 125. Bijtel, J. H., en Woerdeman, M. W. 1928. Over de
ontwikkeling van den staart bij Amphibielarven. Versl. Kon. Akad. Wetensch. Afd.
Natuurkunde, 37.— Billard, A. 1904. Contribution a I'etude des Hydroides. Ann. Sci.
nat., ser. 8, 20. — Bills, C. E. 1924. Some effects of the lower alcohols on Parame-
cium. Biol. Bull., 47.— Bischler, V. 1926. L'influence du squelette dans la regenera-
tion et de potentialites des divers territoires du membre chez Trito7i cristatus. Rev.
Suisse de Zool., 33.— Blochmann, F. 1882. tJber die Entwicklung von Neritina fluvia-
tilis. Zeitschr. wiss. Zool., 36. 1883. Beitrage zur Kenntnis der Entwicklung der Gas-
teropoden. Ihid., 38.— Blount, Mary. 1909. The early development of the pigeon's
egg, etc. Jour. Morphol., 20. — Bodenstein, D. 1938. Untersuchungen zum Meta-
morphoseproblem. I, 11. Arch. Entw'mech., 137. — Bodine, J. H., and Boell, E. J.
1937. The action of certain stimulating and inhibiting substances on the respiration of
active and blocked eggs and isolated embryos. Physiol. Zool., 10. — Bohmig, L. 1891.
Untersuchungen liber rhabdocoelen Turbellarien. II. Zeitschr. wiss. Zool., 51. — Boell,
E. J. 1938. Metabolic properties of the regions of the amphibian gastrula. Abstr.
Amer. Soc. Zool., Anat. Rec, 72, 4, Suppl. Boell, E. J.; Koch, H.; and Needham, J,
1939. Morphogenesis and metabohsm: studies with the Cartesian diver ultramicro-
manometer. IV. Proc. Roy. Soc, B, 127. Boell, E. J., and Needham, J. 1939. Mor-
phogenesis and metabolism. II, III. Ibid. Boell, E. J.; Needham, J.; and Rogers, V.
1939. Morphogenesis and metabohsm. I. Ibid. — Bosenberg, H. 1905. Beitrage zur
Kenntnis der Spermatogenese bei den Arachnoiden. Zool. Jahrb. Abt. Anat. Ont., 21.
— Bohn, G. 1903a. Influence des rayons du radium sur les animaux en voie de crois-
sance. C. R. Acad. Sci., Paris, 136. 1903b. Influence des rayons du radium sur les oeufs
vierges et fecondes et sur les premiers stades du developpement. Ibid. — Bonfig, R.
1925. Die Determination der Hauptrichtungen des Embryos von Ascaris megalo-
cephala. Zeitschr. wiss. Zool., 124. — Borchardt, Eva. 1927. Beitrage zur heteromor-
pher Regeneration bei Dixippus morosus. Arch. Entw'mech., no. — Born, G. 1885.
Biologische Untersuchungen. I. Arch. mikr. Anat., 24. — Boveri, T. 1887. UberDiffer-
enzierung der Zellkerne wahrend der Furchung der Eier von Ascaris mcgaloccphala.
Anat. Anzeiger, 2. 1890. Zellenstudien. III. Jen. Zeitschr., 24. 1892. Uber die Ent-
stehung des Gegensatzes zwischen den Geschlechtszellen und somatischen Zellen bei
Ascaris megalocephala etc. Sitzber. Ges. Morph. Physiol. Miinchen, 8. 1899. Die Ent-
wicklung von Ascaris megalocephala mit besonderer Riicksicht auf die Kernverhalt-
nisse. Festschr. 70 Geburtstag von C. von Kupffer. Jena. 1901a. Uber die Polaritat
des Seeigeleies. Verb, phys.-med. Ges. Wiirzburg, N.F., 31. 1901b. Die Polaritat von
Ovocyte, Ei und Larve des Strongylocentrotus lividus. Zool. Jahrb. Abt. Anat. Ont., 14.
1902. Uber mehrpolige Mitosen als Mittel zur Analyse des Zellkerns. Verb, phys.-med.

754 PATTERNS AND PROBLEMS OF DEVELOPMENT

Ges. Wiirzburg, N.F., 35. 1904. Ergebnisse iiber die Konstitution der chromatischen
Substanz des Zellkerns. Jena. 1905. tJber Doppelbefruchtung. Sitzber. phys.-med.
Ges. Wiirzburg. Jahrg. 1907. Zellenstudien. VL Jen. Zeitschr., N.F., 43. 1910a.
tJber die Teilung zentrifugierter Eier von Ascaris megalocephala. Arch. Entw'mech.,
30. 1910b. Die Potenzen der Ascaris-Blastomeren bei abgeanderter Furchung.
Festschr. 60 Geburtstag R. Hertwigs. III. Jena. — Bovie, W. T., and Barr, C. E., 1924.
Photocytolysis as a measure of metabolic activity. Science, 59. — Boycott, A. E., and
Diver, C. 1923. On the inheritance of sinistrality in Lim»oea /'eregm. Proc. Roy. Soc,

B, 95. Boycott, A. E.; Diver, C; Garstang, S. L.; and Turner, F. M. 1930. The in-
heritance of sinistrality in Limnaea peregra. Phil. Trans. Roy. Soc, B, 219. Boycott,
A. E.; Diver, C; Hardy, S.; and Turner, F. M. 1929. The inheritance of sinistraHty in
Limnaea peregra. Proc. Roy. Soc, B, 104. — Boysen- Jensen, P. 1936. Growth hor-
mones in plants. Trans, and rev. by G. S. Avery and P. R. Burkholder. New York. —
Bracket, A. 1903. Recherches sur I'ontogenese des Amphibiens Urodeles et Anoures.
Arch. Biol., 19. 1904. Recherches experimentales sur I'ceuf de Ranafusca. Ibid., 21.
1910a. La polyspermie experimental comme moyen d'analyse de la fecondation.
Arch. Entw'mech., 30, I. 1910&. Recherches sur I'iniiuence de la polyspermie experi-
mentale dans le developpement de I'oeuf de Rana fusca. Arch. Zool. exp., ser. 5, 6.
1911a. Etudes sur les locaHsations germinales et leur potentiahte reelle dans I'ocuf
parthenogenetique de Rana fusca. Arch. Biol., 26. 19116. L'oeuf et les facteurs de
I'ontogenese. Paris, 191 1. 1912. La polyspermie experimentale dans I'ocuf de Rana
fusca. Arch. mikr. Anat., 79. 1927. Etude comparative des localisations germinales
dans I'oeuf des amphibiens urodeles et anoures. Arch. Entw'mech., in. — Bracket, J.
1934a. Metabolisme respiratoire et centre organisateur de la gastrula des Amphibiens.

C. R. Soc. Biol., 116. 1934&. Etude du metabolisme de I'oeuf de grenouille {Rana fusca)
au cours du developpement. i. Arch. Biol., 45. 1934c. Etude du metabolisme etc. 3.
Ibid., 46. 1936. Le metabolisme respiratoire du centre organisateur de I'oeuf de gre-
nouille. C. R. Soc. Biol., 122. 1938. La locaHsation des proteines sulfhydrilees pendant
le developpement des Amphibiens. Bull. Acad. roy. Belg. CI. des Sci., 1938. 1939.
Etude du metabolisme etc. 5. Arch. Biol., 50. Brachet, J., and Shapiro, H. 1937. The
relative oxygen consumption of dorsal and ventral regions of intact amphibian gastru-
lae, including observations on unfertilized eggs. Jour. Cell. Comp. Physiol., 10. —
Braem, F. 1893. Das Prinzip der organbildenden Keimbezirke und die entwicklungs-
mechanischen Studien von H. Driesch. Biol. Zbl. 13. 1894. Zur Entwicklungsge-
schichte von Ophryotrocha puerilis Clprd. Mecz. Zeitschr. wiss. Zool., 57. 1908. Die
Knospung der Margeliden etc. Biol. Zbl., 28. — Brandt, W. 1924a. Extremitatentrans-
plantation an Triton taeniatus. I. Arch. mikr. Anat. Entw'mech., 103. 1924b. Ex-
tremitatentransplantation etc. IL Anat. Anz., 58. 1925. Experimentell erzeugte GHed-
massenverdoppelungen bei Triton. Arch. Entw'mech., 106. — Brasil, L. 1905. Re-
cherches sur la reproduction des Gregarines monocystidees. Arch. Zool. exp., ser. 14, 3.
— Brauer, A. 1938. Modifications of development in relation to differential suscepti-
bility of the bruchid (Coleoptera) egg to KCN during different metaboHc phases.
Physiol. Zool., 11. Brauer, A., and Taylor, A. C. 1936. Experiments to determine the
time and method of organization in bruchid (Coleoptera) eggs. Jour. Exp. Zool., 73. —
Braus, H. 1914. tJber die Entstehung der Kiemen, ein Beitrag zur Homologiefrage.
Zeitschr. Morphol. Anthropol. 18. — Brecher, Leonore. 1924. Die Bedingungen fiir

BIBLIOGRAPHY 755

Fiihlerfiisse bei Dixippus {Caransius) morosus Br. et Realt. Arch. mikr. Anat. Ent-
w'mech., 102. — Brefeld, O. 1869. Dictyostdiummticor aides. Abh. senckenberg. naturf.
Ges., 7. 1876. Die Entwicklungsgeschichte der Basidiomyceten. Bot. Ztg., 34. 1884.
Polvspondylium violaceum und Dictyostelium mucaroides. Unters. aus dem Gesamtgeb.
d. Mycol., H. 6. — Brien, P. 1930. Contribution a I'etude de la regeneration naturelle et
experimentale chez les Clavelinidae. Ann. Soc. roy. Zool. Belg., 61. 1936. Contribution
a I'etude de la reproduction asexuee des Phylactolemates. Melanges P. Pelseneer. Mem.
Mus. roy. d'Hist. nat. Belg., ser. 2, fasc. 3. Brien, P., et Huysmans, G. 1937. La crois-
sance et le bourgeonnement du stolon chez les Stolonifera. Ann. Soc. roy. Zool. Belg.,
68. — Broer, W., und Ubisch, L. von. 1934. Uber die Bildung von Linsen aus Bauch-
haut von Bomhinator pachypus. Arch. Entw'mech., 132. — Broman, I. 1900. tjber
Bau und Entwicklung der Spermien von Bomhinator igneiis. Anat. Anz., 17. 1902.
tJber gesetzmassige Bewegungs- und Wachstumserscheinungen (Taxis und Tropismen-
formen) der Spermatiden, ihrer Centralkorper, Idiozomen und Kerne. Arch. mikr.
Anat., 59. Br0ndsted, H. V. 1939. Regeneration in planarians investigated with a
new transplantation technique. I-III. Kgl. Danske Videnskabernes Selskab. Bio-
logiske Meddelelser, 16. — Brooks, Matilda M. 1932. The effect of methylene blue on
HCN and CO poisoning. Amer. Jour. Physiol., 102. 1935. Methylene blue and hemo-
globin derivatives in asphyxial poisoning. Ibid., 1 14. — Brooks, W. K. 1893. The genus
Salpa. Mem. Biol. Lab. Johns Hopkins Univ., 2. — Browne, Ethel. 1909. The produc-
tion of new hydranths in Hydra by the insertion of small grafts. Jour. Exp. Zool., 7.
1910. Effects of pressure on Cumingia eggs. Arch. Entw'mech., 29. — Bruchmann, H.
1909. Vom Prothallium der grossen Spore und von der Keimesentwicklung einiger
Selaginella-Arten. Flora, 99. — Bruns, Emmy. 1931. Experimente iiber das Regula-
tionsvermogen der Blastula von Triton tacniatus und Bomhinator pachyPus. Arch.
Entw'mech., 123. — Brunst, V. 1927. Zur Frage nach dem Einiluss des Nervensystems
auf die Regeneration. Ihid., 109. — Buchanan, J. W. 1922. The control of head forma-
tion in Planaria by means of anesthetics. Jour. Exp. Zool., 35. 1923a. On the nature
of the determining factors in regeneration. Ihid.,2,']. 1923b. Weight changes and oxy-
gen consumption during long exposure to dilute anesthetics. Ibid., 38. 1926a. Depres-
sion of oxidative metabolism and recovery from dilute potassium cyanide. Ibid., 44.
1926b. Some antagonistic and additive effects of anesthetics and potassium cyanide.
Ihid., 1926c. Regional differences in rate of oxidation in the chick blastoderm, as
shown by susceptibiHty to hydrocyanic acid. Ibid., 45. 1927. The spatial relations
between developing structures. Ibid., 49. 1930. The nature of disintegration gra-
dients. Ibid.,$'j. 1933. Regeneration in P/mgoca/o gmaVw (Leidy) . Physiol. Zool., 6.
1935. An analysis of physiological states responsible for anteroposterior disintegration
in Planaria dorotocepliala. Protoplasma, 22. Buchanan, J. W., and Levengood, C. A.
1939. Axial susceptibility to serum antibodies in Euplanaria. Physiol. Zool., 12. —
Buchner, P. 1909. Das accessorische Chromosom in Spermatogenese und Ovogenese
der Orthopteren. Arch. Zellforsch., 3.— Btitschli, O. 1883-89. Protozoa. Bronn's
Klassen und Ordnungen des Tierreichs, Bd. I, Abtl. II, III.— Burr, H. S. 1932. An
electrodynamic theory of development suggested by studies of proliferation rates in the
brain of Amhlystoma. Jour. Comp. Neurol., 56. Burr, H.S., and Hovland, C. I. 19370-
Bioelectric potential gradients in the chick. Yale Jour. Biol. Med., 9. 1937^- Bioelec-
tric correlates of development in Amhlystoma. Ihid. Burr, H. S., and Lane, C. T.

7S6 PATTERNS AND PROBLEMS OF DEVELOPMENT

1935. Electrical characteristics of living systems. Yale Jour. Biol. Med., 8. — Busch,
W. 1851. Beobachtungen iiber Anatomie und Entwicklung einiger wirbellosen See-
thiere. Berlin. — Buschkiel, N. 1922. CaidlcrycUa pipicntis n. sp. Eine neue Schizo-
gregarine aus den Larven der Culcx pipiens. Zool. Jahrb., Abt. Anat. Ont., 43. — But-
ler, Eliz. 1935. The developmental capacity of regions of the unincubated chick blas-
toderm, as tested in chorio-allantoic grafts. Jour. Exp. Zool., 70. — Byerly, T. C. 1926.
Studies in growth. L Anat. Rec, 32. — Bytinski-Salz, H. 1929a. Untersuchungen iiber
das Verhalten des prasumptiven Gastrulaectoderms der Amphibien bei heteroplas-
tischer und xenoplastischer Transplantation ins Gastrocoel. Arch. Entw'mech., 114.
1929b. Die Wirkung von xenoplastischen Implantaten und Embryonalextrakten auf
die Entwicklung junger Amphibienkeime. Ibid. 1929c. Untersuchungen iiber die De-
termination und die Induktionsfahigkeit einiger Keimbezirke der Anuren. Ibid., 118.
193 1. Untersuchungen iiber die Induktionsfahigkeit der hinteren MeduUarplattenbe-
zirke. Ibid., 123.

Calkins, G. N. 1895. The spermatogenesis of Lumbricus. Jour. Morphol., 11. 1911.
Effects produced by cutting Paramecium cells. Biol. Bull., 21. 1933. The biology of
the Protozoa. 2d ed. Phila. — Camp, W. H. 1929. Glutathione in plants. Science, 69.
— Campbell, D. H. 1918. The structure and development of mosses and ferns. 3d ed.
New York. — Cannon, H. G. 1923. On the metabolic gradient of the frog's egg. Proc.
Roy. Soc, B, 94. — Cannon, W. B. 1939. The argument forchemical mediation of nerve
impulses. Science, 90. Cannon, W. B., and Rosenblueth, A. 1937. Autonomic neuro-
effector systems. New York. — Cary, L. R. 1916. The influence of the marginal sense
organs on the rate of regeneration in Cassiopea zamachana. Jour. Exp. Zool., 21. —
Castelnuovo, Gina. 1932a. Sulla rigenerazione e suscettibilita differenziale (KCN) di
Lumbricidus hojjmeisteri. Arch. Zool. ital., 17. 19326. Azione differenziale della tem-
peratura suUe uova di axolotl. Boll, di Zool., 3. — Castle, W. A. 1928. An experimental
and histological study of the life cycle of Planaria velata. Jour. Exp. Zool., 51. — Castle,
W. E. 1896. The early embryology of Ciona intestinalis Flemming (L). Bull. Mus.
Comp. Zool. Harvard, 27. — Chabry, L. 1887. Contribution a I'embryologie normale et
teratologique des Ascidiens simples. Jour. Anat. physiol., 23. — Chen, K. K.; Rose,
C. L.; and Clowes, G. H. A. 1933. Methylene blue, nitrites and sodium thiosulphate
against cyanide poisoning. Proc. Soc. Exp. Biol. Med., 31. — Chiarugi, G. 1898. Pro-
duzione sperimentale di duplicita embrionali in uova di Salamandrina perspicillata.
Monit. Zool. ital., 9. — Child, C. M. 1898. The maturation and fertilization of the egg of
Arenicola marina. Trans. N.Y. Acad. Sci., 16. 1900. The early development of .4 /r/zi-
cola sixid Sternaspis. Arch. Entw'mech., 9. 1902. Studies on regulation. I. Ibid., 1$.
1903a. Form regulation in Cc;7'a;f///z/5. II. Biol. Bull., 5, 6. 19036. Studies on regula-
tion. III. Arch. Entw'mech., 17. 1904a. Form regulation in Ccrianthus. VI. Biol.
Bull., 7. 19046. Studies on regulation. V, VI. Jour. Exp. Zool., i. 1905a. Form regu-
lation in Cerio»/m5. VIII. Biol. Bull., 8. 19056. Studies on regulation. IX, X. Arch.
Entw'mech., 20. 1906. The relation between regulation and fission in Planaria. Biol.
Bull., II. 1907a, 6, c. An analysis of form regulation in Tubularia. I, II, III. Arch.
Entw'mech., 33. 19076?, e,/. An analysis etc. IV, V, VI. Ibid., ^4.. 1908. Form regula-
tion in Ccrianthus aestuarii. Biol. Bull., 15. 1909a. Factors of form regulation in
Harenactis attenuata. I. Jour. Exp. Zool., 6. 19096. Factors, etc. II. Ihid.,j. 1909c.
Factors, etc. III. Ibid. 1909c?. The regulation of mutilated primordia in Tubularia.

BIBLIOGRAPHY 757

Arch. Entw'mech., 27. 1910a. Physiological isolation of parts and fission in Planaria.
Ibid., 30, Teil II. 1910b. The central nervous system as a factor in the regeneration of
polyclad Turbellaria. Biol. Bull., iq. 1910c. Further experiments in adventitious re-
production and polarity in //are/far//^. Ibid., 20. 1911a. Die physiologische Isolation
von Teilen des Organismus. Vortrage u. Aufsatze ii. Entw'mech. XI. 1911b. Studies
on the dynamics of morphogenesis, etc. I. Jour. Exp. Zool., 10. 1911c. Studies, etc.
II. Ibid., II. 191 irf. Studies, etc. III. Ibid. 19116. A study of senescence and re-
juvenescence based on experiments with Planaria dorotocephala. Arch. Entw'mech.,
31. 1911/. Experimental control of morphogenesis in the regulation of Planaria. Biol.
Bull., 20. 1912. Studies on the dynamics, etc. IV. Jour. Exp. Zool., 13. 1913a.
Studies, etc. V. Ibid., 14. 1913b. Studies, etc. VI. x\rch. Entw'mech., 37. 1913c. The
asexual cycle of Planaria velata in relation to senescence and rejuvenescence. Biol.
Bull., 25. 1913d. Certain dynamic factors in experimental reproduction and their sig-
nificance for the problems of reproduction and development. Arch. Entw'mech., 35.
1914a. The axial gradients in ciliate Infusoria. Biol. Bull., 26. 1914b. Starvation, re-
juvenescence and acchmation in Planaria dorotocephala. Arch. Entw'mech., 38.
1914c. Studies on the dynamics, etc. VTI. Jour. Exp. Zool., 16. 1914c/. Asexual breed-
ing and prevention of senescence in Planaria velata. Biol. Bull., 26. 1914^. Studies on
the dynamics, etc. VIII. Jour. Exp. Zool., 17. 1914/. Susceptibility gradients in ani-
mals. Science, 39. 1915a. Axial gradients in the early development of the starfish.
Amer. Jour. Physiol., 37. 1915b. Senescence and rejuvenescence. Chicago. 1915c.
Individuality in organisms. Chicago. 1916a. Axial susceptibility gradients in the
early development of the sea urchin. Biol. Bull., 30. 1916b. Studies on the dynamics,
etc. IX. Jour. Exp. Zool., 21. 1916c. Axial susceptibility gradients in algae. Bot.
Gaz., 62. 19160?. Experimental control and modification of larval development in the
sea urchin in relation to the axial gradients. Jour. Morphol.,28. 19166. Further ob-
servations on axial susceptibility gradients in algae. Biol. Bull., 31. 1917a. Suscepti-
bility gradients in the hairs of certain marine algae. Ibid., 32. 1917b. Experimental
alteration of the axial gradient in the alga, Griffithsia bornetiana. Ibid. 1917c. The
gradient in susceptibility to cyanide in the meridional conducting path of the cteno-
phort, Mnemiop sis. Amer. Jour. Physiol., 43. i9i7d. Differential susceptibility and
differential inhibition in the development of polychete annelids. Jour. Morphol.,30.
1919a. Demonstration of the axial gradients by means of potassium permanganate.
Biol. Bull., 36. 1919b. A comparative study of carbon dioxide production during star-
vation in Planaria. Amer. Jour. Physiol., 48. 1919c. The effect of cyanides on carbon
dioxide production and on susceptibility to lack of oxygen in Planaria dorotocephala.
Ibid. 1919c/. The axial gradients in Hydrozoa. II. Biol. Bull., 37. 19196. Susceptibil-
ity to lack of oxygen during starvation in Planaria. Amer. Jour. Physiol., 49. 1919/.
A study of susceptibility in some Puget Sound algae. Publ. Puget Sound Biol. Sta., 2.
1920a. Studies on the dynamics, etc. X. Jour. Exp. Zool., 30. 1920b. Some considera-
tions concerning the nature and origin of physiological gradients. Biol. Bull., 39.
1921a. The origin and development of the nervous system. Chicago. 1921b. Le prob-
leme de I'integration physiologique. Scientia, ann. 1921. 1921c. Studies on the dy-
namics, etc. XL Jour. Exp. Zool., 33. 1921c/. The axial gradients in Hydrozoa. IV.
Biol. Bull., 41. 1923a. The axial gradients in Hydrozoa. V. /&/</., 45. 1923b. The
general relation between susceptibility and physiologic condition. Arch. Int. Med., 32.

758 PATTERNS AND PROBLEMS OE DEVELOPMENT

1924a. Modification of development in relation to ditTerential susceptibility. Amer.
Naturalist. 58. 1924b. Physiological foundations of behavior. New York. 1925a. The
axial gradients in Hydrozoa. \I. Biol. Bull., 48. 1925&. The axial gradients in Hydro-
zoa. \'ll. Ibid. 1925c. The physiological significance of the cephalocaudal ditYerential
in vertebrate development. Anat. Rec. 31. 1926a. Studies on the axial gradients in
Corymorpha palma. IL Biol. Gen.. :;. 1926b. Studies, etc. IIL Ibid, igi^a. Modifi-
cation of polarity and symmetry in Cory?>wrpha palma, etc. I. Jour. Exp. Zool., 47.
1927b. Modification, etc. IL //)/(/., 48. 1927c. Experimental locaHzation of new axes
in Coryitiorp/ia without obliteration of the original polarity. Biol. Bull.. 53. igi^d.
Developmental modification and elimination of the lar\'al stage in the ascidian, Corclla
'd-illmcriaiia. Jour. Morphol. Physiol., 44. 1928a. Experimental transformations of bi-
polar forms in Corymorpha palmu. Biol. Bull., 54. 1928b. Physiological polarity and
dominance in the holdfast system of Corymorpha. Ibid. 1928c. Axial development of ag-
gregates of dissociated cells from Corymorpha palma. Physiol. Zool.. i. 1928^. The
physiological gradients. Protoplasma. 5. 1929a. Lateral grafts and incisions as organiz-
ers in the hydroid, Corymorpha. Physiol. Zool, 2. 1929b. Physiological dominance and
physiological isolation in development and reconstitution. Arch. Entw'mech., 117,
Teil IL 1930. The susceptibility of Planaria to potassium cyanide in relation to hy-
drion concentration and to certain changes in salt content of the medium. Physiol.
Zool., 3. 1931. Experimental modification of the scale of organization in the reconsti-
tution of r«/>w/ana. Ibid., 4. 1932a. Determination of new axes by lateral grafts and
incisions in Corymorpha, etc. Ibid., 5. 1932b. Experimental studies on a Japanese
planarian. I. Sci. Repts. Tohoku Imp. Univ., Ser. 4, Biol., 6. 1933a. The swimming
plate rows of the ctenophore, Pleurobraihia as gradients: with comparative data on
other forms. Jour. Comp. Neurol., 57. 1933b. Reconstitution in fl'a/Zc/y.y/w.y a«nV«/a
Clark. Sci. Repts. Tohoku Imp. Univ., Ser. 4, Biol.. 8. 1934a. Differential reduction of
methylene blue by living organisms. Proc. Soc. Exp. Biol. Med., 32. 1934b. The dif-
ferential reduction of methylene blue by Paramecium and some other ciliates. Proto-
plasma, 22. 1935. Dominance of hydranths induced by grafts in Corymorpha. Jour.
Exp. Zool.. 71. 1936a. Differential reduction of vital dyes in the early development of
echinoderms. Arch. Entw'mech., 135. 1936b. A contribution to the physiology of ex-
ogastrulation in echinoderms. Ibid. 1938. Certain differential modifications of asteroid
and echinoid development. Annot. Zool. Japon., 1 7. Child, C. M., and Bellamy, A. W.
1919. Physiological isolation by low temperature in BryophyUum and other plants.
Science, 50. 1920. Physiological isolation by low temperature in BryophyUum. Bot.
Gaz., 70. Child, C. M., and Deviney, Ezda. 1926. Contributions to the physiology of
Paramecium caudatum. Jour. Exp. Zool., 43. Child, C. M., and Hyman, L. H. 1919.
The axial gradients in Hydrozoa. I. Biol. Bull., 36. 1926. Studies on the axial gradients
m Corymorpha. I. Biol. Gen., 2. Child, C. M., and McKie, E. V. M. 1911. The central
nervous system in teratophthalmic and teratomorphic forms of Planaria dorotocephala.
Biol. Bull., 22. Child, C. M., and Rulon, O. 1936. Difi'erential reduction of vital dyes in
certain oligochetes. Jour. Exp. Zool., 74. Child, C. M., and Watanabe, Y. 1935a. The
head frequency gradient in Euplauaria dorotocephala. Physiol. Zool.. S. 1935b. Differ-
ential reduction of methylene blue by Corymorpha palma. Ibid. Child, C. M., and
Young, A.N. 1903. Regeneration of the appendages in nymphs of the .\grionidae. Arch.
Entw'mech., 15. — Chrisler, G. 1935. Methylene blue and anoxemia. .\mer. Jour.

BIBLIOGRAPHY 759

Physiol., no. — Chuang, Hsiao-Hui. 1938. Spezifische Induktionsleistungen von Le-
ber und Xiere. Biol. Zbl., 58. 1940. Induktionsleistungen von frischen und gekochten
Organteilen (Xiere, Leber) etc. Arch. Entw'mech. 139.— Chun, C. 1880. Die Cteno-
phoren des Golfes von Xeapel. Fauna und Flora d. Golfes v. Xeapel. I. 1888. Die
pelagische Tierwelt in grosseren Meerestiefen. Bibl. Zool., H. i, 1888. 1892. Die Dis-
sogonie, eine neue Form der geschlechtlichen Zeugung. Festschr. 70 Geburtstag R.
Leuckarts. 1895. Bemerkungen iiber den Aufsatz von H. Driesch und T. H. Morgan,
"\'on der Entwicklung einzelner Ctenophorenblastomeren." Arch. Entw'mech., 2.
1896. Atlantis. Bibl. Zool., 7. — Clark, W. G. 1937- Electrical rx)larity and auxin
transport. Plant Physiol., 12. 1938. Electrical polarity and auxin transport. Ihid.,
13. — Clarke, L. F. 1936. Regional differences in eye-forming capacity of the early
chick blastoderm as studied in chorio-allantoic grafts. Physiol. Zool., 9. — Clement,
A. C. 1938. The structure and development of centrifuged eggs and egg fragments of
Physa heteroslropha. Jour. Exp. Zool., 7Q. — Cleveland, L. R. 1938- Longitudinal and
tranverse division in two closely related flagellates. Biol. Bull., 74. — Coe, W. R. 1929.
Regeneration in nemerteans. Jour. Exp. Zool., 54. 1930- Regeneration, etc. II. Ibid.,
57. 1932. Regeneration, etc. III. Ihid., 61. 19340. Regeneration, etc. IV. Ihid., 6-;.
1934b. .\nalysis of the regenerative processes in nemerteans. Biol. Bull., 66. — Coghill,
G. E. 1929. Anatomy and the problem of behavior. Xew York. — Cohen, A. 1938a.
Induction by cauterization in the amphibian egg. Collecting Xet, 13. 1938b. Myo-
tome fusion in the embr\'o of Amblystoma after treatment with lithium and other
agents. Jour. Exp. Zool., 79. Cohen, A., and Berrill, N.J. 1936. The development of
isolated blastomeres of the ascidian egg. Ihid., 74. — Collin, B. 191 1. Etude mono-
graphique sur les .\cinetiens. I. .\rch. Zool. exp., ser. 5, 8. 1912. Etude, etc. 11. Ibid.,
51. — Collings, C. K. 1932. Cellular changes during the regeneration of small pieces of
Planaria. Abstr. .\mer Soc. Zool., Anat. Rec, 54, Suppl. — Colucci, V. S. 1891. .Sulla
rigenerazione parciale dell'occhio nei Tritoni. ^lem. .A.ccad. Sci. Bologna, i. — Conklin,
E. G. 1897. The embr>-ology of Crepidnla. Jour. Morphol., 13. 1898. Cleavage and
differentiation. Woods Hole Biol. Lect. 1898. Karyokinesis and c>'tokinesis in the
maturation, fertilization and cleavage of Crepidula and other gasteropods. Jour. Acad.
X'at. Sci. Phila., 12. 1903a. The cause of inverse symmetr\'. .\nat. Anz., 23. 1903b.
The cause of inverse symmetr\'. Proc. Acad. X^at. Sci. Phila., 55. 19050. Organization
and cell lineage of the ascidian egg. Jour. Acad. X'at. Sci. Phila., 13. 1905b. Organ-
forming substances in the eggs of ascidians. Biol. Bull., 8. 1905c. Mosaic development
in ascidian eggs. Jour. Exp. Zool., 2. 1906. Does half an ascidian egg give rise to a
whole larva? Arch. Entw'mech., 21. 1907. The embr>olog\' of Fulgiir. Proc. Acad.
Xat. Sci. Phila., July, 1907. 1908. The habits and early development of Linerges
mercurius. Papers Tortugas Lab. Carnegie Inst., 2. 1910. The effects of centrifugal
force upon the organization and development of the eggs of fresh water pulmonates.
Jour. Exp. Zool., 9. 1911. The organization of the egg and the development of single
hla-stomeres oi Phallusia mammilata. Ibid., 10. 1912. Experimental studies in nuclear
and cell di\'ision in the eggs of Crepidula plana. Jour. Acad. X'at. Sci. Phila., 15. 1917.
Effects of centrifugal force on the structure and development of the eggs of Crepidula.
Jour. Exp. Zool., 22. 1931. The development of centrifuged eggs of ascidians. Ibid.,
60. 1932. The embrv"olog>- of Amphioxus. Jour. Morphol., 54. 1933. The develop-
ment of isolated and partially separated blastomeres of Amphioxus. Ibid., 64. — Coon-

76o PATTERNS AND PROBLEMS OF DEVELOPMENT

field, B. R. 1936a. Apical dominance and polarity in Mncmiopsis leidyi Agassiz. Biol.
Bull., 70. 1936b. Regeneration in M«cm/o/>5/5 /e/(fyi Agassiz. Ibid., ji. 1937a. The
regeneration of plate rows in Mnemiopsis leidyi Agassiz. Proc. Nat. Acad. Sci., 23.
1937b. Symmetry and regulation in Mnemiopsis leidyi. Biol. Bull., 72. Coonfield,
B. R., and Goldin, A. 1937. The problem of a physiological gradient in Mncmiopsis
during regeneration. Biol. Bull., 73. — Cooper, W. C. 1936. Transport of root-forming
hormone in woody cuttings. Plant Physiol., 11. — Copenhaver, W. M. 1939a. Initia-
tion of beat and intrinsic contraction rates in the different parts of the Amhlystoma
heart. Jour. Exp. ZooL, 80. 1939b. Some observations on the growth and function of
heteroplastic heart grafts. Ibid.,S2. — Cotronei, G. 1921. I processi di inibizione differ-
enziale nei vestibulo buccale degh Anfibi anuri. Riv. di Biol., 3. 1922. Correlations et
differentiations. Essai de morphologie causale sur la tete des Amphibiens. Arch. ital.
Biol., 71. — Coulter, J. M.; Barnes, C. R.; and Cowles, H. C. 1910. A textbook of bot-
any. New York. Coulter, J. M., and Chamberlain, C. J. 1903. Morphology of angio-
sperms. New York. 191 0. Morphology of gymnosperms. Chicago. — Crabb, E. D.
1927. Genetic experiments with pond snails, Limnaea and Physa. Amer. Naturalist,
61. — Crampton, H. E. 1894. Reversal of cleavage in a sinistral gastropod. Ann. N.Y.
Acad. Sci., 8. 1896. E.xperimental studies on gastropod development. Arch. Ent-
w'mech., 3. 1897. The ascidian half embryo. Ann. N.Y. Acad. Sci., 10. 1916. Studies
on the variation, distribution and evolution of the genus Partula. Carnegie Inst.
Wash. Publ. 228. 1924. The coincident production of dextral and sinistral young in
the land gastropod Partula. Science, 59. — Crooks, D. M. 1933. Histological and re-
generative studies on the flax seedling. Bot. Gaz., 95. — Crowell, P. S. 1937. Factors
affecting regeneration in the earthworm. Jour. Exp. ZooL, 76. — Curtis, W. C. 1900.
On the reproductive system of Planaria simplissima, a new species. Zool. Jahrb. Abt.
Anat. Ont., 13. 1928. Old problems and new technique. Science, 67. Curtis, W. C,
and Hickman, Jane. 1926. The effects of X-rays on regeneration in planarians. Abstr.
Nat. Acad. Sci., Science, 63. Curtis, W. C, and Schuize, Lula M. 1924. Formative
cells of planarians. Abstr. Amer. Soc. ZooL, Anat. Rec, 29, Suppl.

Dalcq, A. 1925. Recherches experimentales et cytologiques sur la maturation et
I'activation de I'oeuf d'Asterias glacialis. Arch Biol., 34. 1932a. Experiences de mero-
gonie sur I'ocuf dWscidiella aspersa. C. R. Soc. Biol., 119. 1932b. Les locahsations
germinates de I'oeuf vierge d'ascidie analysee par des experiences de merogonie. C. R.
Assoc, anat., 1932. 1932c. Etude des localisations germinales dans roeuf vierge d'as-
cidie par des experiences de merogonie. Arch. Anat. micr., 2S. 1935. L'organisation
de I'ccuf chez les Chordes. Paris, 1935. 1938a. Le concept "champ-gradient-seuil" dans
I'interpretation de la morphogenese embryonnaire. C. R. Assoc, anat. 1938. 1938b.
Etude micrographique et quantitative de la merogonie double chez Ascidiella scabra.
Arch. Biol., 49. Dalcq, A., et Pasteels, J. 1938. Potentiel morphogenetique, regula-
tion et "axial gradients" de Child. Bull. Acad. roy. Med. Belg., ser. 6, 3. Dalcq, A., and
Vandebroek, G. 1937. On the significance of the polar spot in ripe unfertihzed and in
fertilized ascidian eggs. Biol. Bull., 72. — Dalton, A. J. 1935. The potencies of por-
tions of young chick blastoderms as tested in chorio-allantoic grafts. Jour. Exp. ZooL,
71. — Dalton, J. C. 1881. Position of the embryo in the fowl's egg. Med. Rec, 20. —
Danchakoff, Vera. 1924. Wachstum transplantierter embryonaler Gewebe in der
Allantois. Zeitschr. Anat. Entw'ges., 74. 1926. Lens ectoderm and optic vesicles in

BIBLIOGRAPHY 761

allantois grafts. Carnegie Inst. Wash. Publ. 363. — Dareste, C. 1891. Recherches sur
la production artificielle des monstruosites. 2d. ed. Paris. — Davydov, K. 1915. Resti-
tution bei den Nemertinen. Trav. Lab. Zool. et Sta. Biol. Sebastopol pres de I'Acad.
Sci. Petrograd, ser. 2, ann. 1915.— Dawydoff, C. 1901. Beitrage zur Kenntnis der Re-
generationserscheinungen bei den Ophiuren. Zeitschr. wiss. Zool., 69. — Debaisieux, P.
1909. Les debuts de I'ovogenese dans le ZJywaf.y. Cellule, 25. — De Beer, G. R. 1927.
The mechanics of vertebrate development. Biol. Rev., 2. — Dederer, P. H. 1910. Pres-
sure experiments on the egg of Ccrehratulus. Arch. Entw'mech., 29. De Garis, C. F.
1927. The effects of anterior and posterior selections on fission rate in pure lines of
Paramecium caudatum. Jour. Exp. Zool., 50. — Dehorne, A. 1932. Nouvelle observa-
tions sur la multiplication asexuee d'une Annelide du genre Dodecaceria. C. R. Acad.
Sci., Paris, 195. — De Laubenfels, M. W. 1932. Physiology and morphology of Porif-
era, exemplified by lotrochota birolulaia. Carnegie Inst. Wash. Publ. 435. — Delia
Valle, P. 191 1. La doppia rigenerazione inversa nelle fratture delle zampe di Triton.
Boll. Soc. di. Naturalisti, 25. 1913. Studi sui rapporti fra differenziazione e rigenera-
zione nelle fratture delle zampe di Triton. Ibid. — Dembowska, W. S. 1925. Studien
liber die Regeneration von Stylonychia mytilus. Arch. mikr. Anat. Entw'mech., 104.
1926. Studies on the regeneration of Protozoa. II. Jour. Exp. Zool., 43. — Detwiler,
S. R. 1918. Experiments on the development of the shoulder girdle and the anterior
limb of AmUystoma pmictatum. Ibid., 25. 1921. Experiments on the hyperplasia of
nerve centers. China Med. Jour., 35. 1922. Experiments on the transplantation of
hmbsin Amblystoma. Jour. Exp. Zool., 35. 1923. Experiments on the transplantation
of the spinal cord in Amblystoma, tic. Ibid., ^7. 1926. The effect of reduction of skin
and muscle on the development of spinal ganglia. Ibid., 45. 1927. The effects of ex-
tensive muscle loss upon the development of spinal ganglia in Amblystoma. Ibid., 48.
1930. Some observations upon the growth, innervation and function of heteroplastic
limbs. Ibid., 57. 1936. Neuroembryology. New York. 1938. Heteroplastic trans-
plantation of somites. Jour. Exp. Zool., 79. Detwiler, S. R., and Lewis, Ruth. 1925.
Size changes in primary brachial motor neurons following limb excision in Amblystoma
embryos. Jour. Comp. Neurol., 39.— Dimpker, A. M. 1917. Die Eifurchung von
Hcrpobdella atomaria Carena {Nephelis vulgaris Moqu. Tand.). Zool. Jahrb. Abt.
Anat. Ont., 40. — Diver, C. 1925. The inheritance of inverse symmetry in Limnaea
peregra. Jour. Genetics, 15.— Dodel, A. 1876. Ulothrix zonata, ihre geschlechtliche
und ungeschlechtliche Fortpflanzung. Jahrb. wiss. Bot., 10.— Doflein, F., und Reiche-
now, E. 1929. Lehrbuch der Protozoenkunde, Teil II. 5 Aufl. Jena.— Doms, H.
1915. Einfluss der Temperatur auf das Wachstum und Differenzierung der Organe
wiihrend der Entwicklung von Rana esculenta. Arch. mikr. Anat., 87. — Dragomirow,
N. 1929. ijber die Faktoren der embryonalen Entwicklung der Linse bei Amphibien.
Arch. Entw'mech., 116. 1930. tjber die Dauer der determinierenden Einwirkung des
Augenbechers auf die Linse. Ibid., 12;^. 1932. tJber Entwicklung von Augenbechern
aus transplantierten Stuckchen des embryonalen Tapetums. Ibid., 126. 1933. Uber
Koordination der Teilprozesse in der embryonalen Morphogenese des Augenbechers.
Ibid., 129. 1935. Die Determination des Augenkeimes bei Amphibien (Russian with
German abstr.). Acad. Sci. Ukraine. Trav. Inst. Zool. et Biol., 8. 1937. Experimentelle
Induktion der Retina bei Amphibienembryonen. C. R. (Doklady) Acad. Sci. URSS,
15— Drew, G. A. 1899. Yoldia limatula. Mem. Biol. Lab. Johns Hopkins Univ., 4.—

762 PATTERNS AND PROBLEMS OF DEVELOPMENT

Driesch, H. 1892. Entwicklungsmechanische Studien. IV. Zeitschr. wiss. Zool., 55.
1893. Entwicklungsmechanische Studien. VII-X. Mitt. zool. Stat. Neapel, 11. 1895.
Von der Entwicklung einzelner Ascidienblastomeren. Arch. Entw'mech., i. 1897. Stu-
dien iiber das Regulationsvermogen der Organismen. I. Ibid., S- 1899a. DieLokahsa-
tion morphogenetische Vorgange. Ibid., 8. 1899b. Studien iiber das Regulations-
vermogen der Organismen. IL Ibid.,g. 1900. Studien, etc. IV. Ibid., 10. 1901. Dieor-
ganischen Regulationen. Leipzig. 1902a. Neue Erganzungen zur Entwicklungsphysi-
ologie des Echinidenkeimes. Arch. Entw'mech., 14. 1902&. Studien, etc. Ibid. 1903.
Drei Aphorismen zur Entwicklungsgeschichte jiingster Stadien. Ibid., ly. 1906. Stu-
dien zur Entwicklungsphysiologie der Bilateralitat. Ibid., 21. 1908. The science and
philosophy of the organism. London. 1910. Neue Versuche iiber die Entwicklung
verschmolzener Echinidenkeime. Arch. Entw'mech., 30. Driesch, H., und Morgan,
T. H. 1895. Zur Analyse der ersten Entwicklungsstadien des Ctenophoreneies. I, II.
Ibid., 2. — Du Buy, H. G., and Olson, R. A. 1937. The presence of growth regulators
during the early development of FM(;M.y. Amer. Jour. Bot., 24. — Diirken, B. 1916. Das
Verhalten transplantierter Beinknospen von Ratia fusca und die Vertretbarkeit der
Quelle des formativen Reizes. Zeitschr. wiss. Zool., 115. 1928. Lehrbuch der Experi-
mentalzoologie. Berlin. — Duesberg, J. 1913. Plastosomes et "organ-forming sub-
stances" dans I'oeuf des Ascidiens. Bull. Acad. Sci. Belg., ann. 1913. 1926. Etude
cytologique des oeufs centrifuges de Ciona intestinalis. Arch. Biol., 36. — Duval, M.
1884. De la formation du blastoderme dans I'oeuf d'oiseau. Ann. Sci. nat. Zool., 18.
Eakin, R. M. 1933. Regulatory development in Triturns torosiis (Rathke). Univ.
Calif. Publ. Zool., 39. 1939a. Further studies in regulatory development of Triturus
torosus. Ibid., 43. 19396. Regional determination in the development of the trout.
Arch. Entw'mech., 139. — Eckert, F. 1927. Experimentelle Untersuchungen iiber die
Lage der Teilungszone von Stylaria lacustris. Zeitschr. wiss. Zool., 129. — Eddy, N. B.
1 93 1. Regulation of respiration. The antagonism between methylene blue and sodium
cyanide. Jour. Pharm. Exp. Ther.,41.^ — Eisig, H. 1898. Zur Entwicklungsgeschichte
der Capitelliden. Mitt. zool. Stat. Neapel, 13. — Eismond, J. 1910. tjber Regulations-
erscheinungen in der Entwicklung der in Teilstiicke zerlegten Rochenkeimscheiben.
Arch. Entw'mech., 30, Teil II. — Ekman, G. 1913a. Experimentelle Untersuchungen
iiber die Entwicklung der Kiemenregion (Kiemenfaden und Kiemenspalten einiger
anuren Amphibien). Morph. Jahrb., 47. 1913b. Uber die Entstehung von Kiemen-
faden und Kiemenspalten aus transplantiertem, ortsfremdem Ectoderm bei Bombina-
tor. Ibid. 1914. Experimentelle BeitragezumLinsenbildungsproblem bei den Anuren.
Arch. Entw'mech., 39. 1922. Neue experimentelle Beitrage zur friihesten Entwick-
lung der Kiemenregion und Vorderextremitat der Anuren. Soc. sci. Fennica. Comm.
Biol., I. 1925. Experimentelle Beitrage zur Herzentwicklung der Amphibien. Ibid.,
106. — Ellis, M. M. 1907. The influence of the amount of injury on the rate and amount
of regeneration in M ancasellus macrourus Carman. Biol. Bull., 13. 1909. The relation
of the amount of tail regenerated to the amount removed in tadpoles of Rana clami-
tans. Jour. Exp. Zool., 7. — Emerson, H. S. 1939. Embryonic induction in regenerat-
ing tissue of Rana pipiens and Rana clamitans. Abstr. Amer. Soc. Zool., Anat. Rec,
75, 4, Suppl. — ^Emmel, V. E. 1906. The regeneration of two "crusher claws" following
the amputation of the normal asymmetrical chelae of the lobster {Homarus anie-
ricanus). Arch. Entw'mech., 22. 1908. The experimental control of asymmetry at dif-

BIBLIOGRAPHY 763

ferent stages in the development of the lobster. Jour. Exp. Zool., 4. — Endres, H. 1895.
tjber Anstich- und Schnurversuche an Eiern von Triton taeniatus. Jahresber. schles.
Ges. vaterland. Kultur, 73. — Erdmann, W. 1931. tJber das Selbstdifferenzierungs-
vermogen von Amphibienkeimteilen bekannter prospektiver Bedeutung im Explantat.
Arch. Entw'mech., 24. — Ewert, Alice. 1937. Struktur und erste Differenzierung im
Ei des Mehlkafers, Tenebrio molitor. Ibid., 135.

Fankhauser, G. 1930a. Zytologische Untersuchungen an geschniirten Tritoneiern.
Arch. Entw'mech., 122. 1930b. Die Entwicklungspotenzen diploidkerniger Halften
des ungefurchten Tritoneies. Ibid. — Farmer, J. B., and "Williams, J. L. 1898. Con-
tributions to our knowledge of the Fucaceae. Phil. Trans. Roy. Soc, igo. — Faulkner,
G. H. 1932. The histology of posterior regeneration in the polychete Chaetopterus
variopedalus. Jour. Morphol., 58. — Faure-Fremiet, E. 1913. Le cycle germinative
chezV A scaris me galocephala. Arch. Anat. micr., 15. 1930. Growth and differentiation
of the colonies of Zoothamnium alternans (Clap, and Lachm.). Biol. Bull., 58. 1932.
Morphogenese experimentale (reconstitution) chez Faculina Jicus L. Involution ex-
perimentale et fusion de structure dans les cultures de Faculina Jicus. Arch. Anat.
micr., 28. — Fere, C. 1900. Note sur la multiplicite des causes des variations de I'em-
bryon du poulet. Jour. Anat. Physiol., 36. — Fernandez, M. 1909. Beitrage zur Em-
bryologie der Gurteltiere. Morphol. Jahrb., 39. — Filatow, D. 1925a. tJber die un-
abhiingige Entstehung (Selbstdifferenzierung) der Linse bei Rana esculenta. Arch,
mikr. Anat. Entw'mech., 104. 1925&. Ersatz des linsenbildenden Epithels von Rana
esculenta durch Bauchepithel von Bujo vulgaris. Arch. Entw'mech., 105. 1925c. Die
Transplantation der Linse bei Larven von Triton taeniatus und das Verhalten des
Epithels im Bereich des Transplants. Rev. Zool. russe, ann. 1925. 1927. Aktivierung
des Mesenchyms durch eine Ohrblase und einen Fremdkorper bei Amphibien. Arch.
Entw'mech., no. 1930a. Uber die Wechselbeziehungen des Epithels und des Mesen-
chyms einer vorderen Extremitatenknospe beim Axolotl. Ibid., 121. 1930b. Entwick-
lungsmechanische Untersuchungen an Embryonen von Acipenser Giildenstddtii und
Acipenser stellatus. Ibid., 122. 1934. Wie lange muss bei Rana temporaria die Augen-
becheranlage auf das Korperteil, wirken damit sich nach Entfernung des Bechers eine
freie Linse entwickelt? Zool. Jahrb. Abt. allg. Zool., 59.— Fischel, A. 1897. Experi-
mentelle Untersuchungen am Ctenophorenei. I. Arch. Entw'mech., 6. 1898. Ex-
perimen telle Untersuchungen etc. II-IV. Ibid., 7. 1900a. tJber die Regeneration der
Linse. Anat. Hefte, 14. 1900b. Zur Histologie der Urodelencornea und des Fhmmer-
epithels. Ibid., 15. 1902. Weitere Mitteilungen liber die Regeneration der Linse.
Arch. Entw'mech., 15. 1903. Entwicklung und OrgandifTerenzierung. Ibid. 1917.
tJber rucklaufige Entwicklung. I. Ibid., 42. 1919. Uber den Einfiuss des Auges auf
die Entwicklung und Erhaltung der Hornhaut. Klin. Monatsbl. Augenheilk., 62.
1921. iJber normale und abnorme Entwicklung des Auges. I. Arch. Entw'mech., 49.
—Fischer, F. G., und Hartwig, H. 1936. Die Vitalfarbung von Amphibienkeimen zur
Untersuchung ihrer Oxydations-Reduktions-Vorgiinge. Zeitschr. vergl. Physiol., 24.
1938. Vergleichende Messungen der Atmung des Amphibienkeimes und seiner Telle
wiihrend der Entwicklung. Biol. Zbl., 58. Fischer, F. G., und Wehmeier, E. 1933a.
Zur Kenntnis der Induktionsmittel in der Embryonalentwicklung. Naturwiss., 21.
I933&- Zur Kenntnis etc. Nachrichten ges. Wiss. Gottingen. Fischer, F. G., Weh-
meier, E.; Lehmann, H.; Juhling, L.; und Hultzsch, K. 1935. Zur Kenntnis etc. Ber.

764 PATTERNS AND PROBLEMS OF DEVELOPMENT

deutsch. chem. Ges., 68. — Fitting, H. 1935. Untersuchungen iiber die Induktion der
Dorsiventralitat bei den keimenden Brutkorpern von M archanlia und Limularia. I,
IL Jahrb. wiss. Bot., 82. 1937. Untersuchungen etc. Ill, IV. Ibid., 85. — Flexner, S.
1898. The regeneration of the nervous system of Planar ia torva and the anatomy of
double-headed forms. Jour. Morphol., 14. — Foerster, Margot, and Orstrom. 1933.
Observations sur la predetermination de la partie ventrale dans I'ocuf d'oursin. Trav.
Sta. biol. Roscoff, fasc. 11, 1933 (not available). — Fraps, M. 1930. Studies on respira-
tion and glycolysis in Planaria. I. Physiol. Zool., 3. — Friza, F., und Przibram, H.
1910. Johnston'sche Sinnesorgane in den Fiihlerfiissen der Sphodromantis und Droso-
phila. Biol. Zbl., 53. — Fuchs, K. 1914. Die Keimblatterbildung von Cyclops viridis
Jurine. Zool. Jahrb. Abt. Anat. Ont., 38. — Fuhrmann, O. 1914. Turbellaries d'eau
douce de Columbie. Mem. Soc. Neuchatel Sci. nat.

Galigher, A. E. 1921a. Axial metabolic gradients in the early development of the
sand dollar, Dendraster excentricus. Publ. Puget Sound Biol. Sta., 3. 1921&. On the ac-
tion of certain substances on oxygen consumption. Amer. Jour. Physiol., 58. — Galtsoff,
P. S. 1925. Regeneration after dissociation : an experimental study on sponges. I, II.
Jour. Exp. Zool., 42. — Garbowski, T. 1904. Uber Blastomerentransplantation bei
Seeigeln. Bull. Acad. Sci. Cracovie, x\nn. 1904. — Garcia-Banus, M. 1918. Is the the-
ory of axial gradients in the regeneration of Tubularia supported by facts? Jour. Exp.
Zool., 26. — Gardiner, E. G. 1895. Early development of Polychoenis candalus. Jour.
Morphol., II. — Gast, R., und Godlewski, E. 1903. Die Regulationserscheinungen bei
Pennaria cavoUnii. Arch. Entw'mech., 16. — Gatenby, J. B. 1918. Polyembryony in
parasitic Hymenoptera: a review. Quart. Jour. Micr. Sci., N.S., 63. — Gates, G. E.
1927. Regeneration in a tropical earthworm, Perionyx excavatus E. Perr. Biol. Bull.,
53. — Gebhardt, H. 1926. Untersuchungen iiber die Determination bei Planarienre-
generation. Arch. Entw'mech., 107. — Gee, W. 1916. Effects of acute alcoholization
onXhtgtrmctWsoi Fundulusheteroditus. Biol. Bull., 31. — Geigy, R. 1931. Erzeugung
rein imaginaler Defekte durch ultraviolette Eibestrahlung bei Drosophila melano-
gaster. Arch. Entw'mech., 125. — Geinitz, B. 1925a. Embryonale Transplantation
zwischen Urodelen und Anuren. Arch. Entw'mech., 106. 1925b. Embryonale Trans-
plantation etc. Ibid. 1925c. Zur weiteren Analyse des Organisationszentrums. Ver-
sammlungsber. ges. Vererbungsforsch., 37. — Gemmill, J. F. 1906a. Cyclopia in osse-
ous fishes. Proc. Zool. Soc. London, i. 1906b. Notes on supernumerary eyes and
local deficiency and reduphcation of the notochord in trout embryos. Ibid. 1912. The
teratology of fishes. Glasgow. — Gerard, R. W. 1932. Nerve metabolism. Physiol.
Rev., 12. Gerard, R. W., and Hyman, L. H. 1931. The cyanide insensitivity of
Paramecium. Abstr. Amer. Physiol. Soc, Amer. Jour. Physiol., 97. — Gerould, J. H.
1896. The anatomy and histology of Caudina arenata. Gould. Bull. Mus. Comp.
Zool. Harvard, 29. 1906. The development of Phascolosoma. Zool. Jahrb. Abt. Anat.
Ont., 23. 1931. Premature reversal of heart beat in Bombyx. Science, 73. i933- Or-
ders of insects with heart beat reversed. Biol. Bull., 44. — Gersch, M. 1937- Vital-
farbung als Mittel zur Analyse physiologischer Prozesse. Protoplasma, 27. Gersch,
M., und Ries, E. 1937. Vergleichende Vitalfiirbungsstudien. Arch. Entw'mech., 136.
— Gianferrari, L. 1921. L'azione di LiCl nello sviluppo embrionale di Salmo lacustris
e irideus. Natura, 12. — Giardina, A. 1901. Origine dell'oocyti e delle cellule nutrici
nel Dvtiscus. Internal. Monatschr. Anat. Physiol, 18. — Giesbrecht, W. 1892. Bei-

BIBLIOGRAPHY 765

trage zur Kenntnis einiger Notodelphyiden. Mitt. zool. Stat. Neapel, 3. — Gilchrist,
F. G. 1928. The effect of a horizontal temperature gradient on the development of the
egg oi the urodele, Triturus torosus. Physiol. Zool., i. 1929a. The progress of deter-
mination in the newt egg analyzed by means of thermal gradients. Abstr. Amer. Soc.
Zool., Anat. Rec, 44, 3. 1929&. Modification of polarity and cleavage by means of a
thermal gradient. Ibid. 1929c. The determination of the neural plate in urodeles.
Quart. Rev. Biol., 4. 1932. The development of bilaterality in an amphibian (Triturus)
egg. Abstr. Amer. Soc. Zool., Anat. Rec, 54, Suppl. 1933. The time relations of de-
terminations in early amphibian development. Jour. Exp. Zool., 66. 1937a. The hy-
droid polyp Corymorpha palina as Gestalt and as history. Amer. Naturalist, 71. 1937b.
The nature of organic wholeness. Quart. Rev. Biol., 12. 1937c. Budding and locomo-
tion in the scyphistoma of Aurelia. Biol. Bull., 72. Gilchrist, F. G., and Schmidt, L. K.
1932. Regeneration in Corymorpha as affected by a thermal gradient. Abstr. Amer.
Soc. Zool., Anat. Rec, 54, Suppl. — Gilman, P. K., and Baetjer, F. H. 1904. Some
effects of the Roentgen rays on the development of embryos. Amer. Jour. Physiol., 10.
— Girgolaff, S. S. 1911. Kompressionsversuche am befruchteter Ei des Ascaris
megalocephala. Arch. mikr. Anat., 76. — Giroud, A., et Bulliard, H. 1933. Reaction
des substances a fonction sulfhydryle. Protoplasma, 19. — Goebel, K. 1930, 1933. Or-
ganographie der Pflanzen. 3 Aufl. II, 1930; III, 1933. Jena.— Goerttler, K. 1925. Die
Formbildung der Medullaranlage bei Urodelen. Arch. Entw'mech., 106. 1926. Ex-
perimentell erzeugte "Spina bifida" and "Ringembryonenbildungen" und ihre Bedeu-
tung fiir die Entwicklungsphysiologie der Urodeleneier. Zeitschr. Anat., 80. 1927. Die
Bedeutung gestaltender Bewegungsvorgange beim Differenzierungsgeschehen. Arch.
Entw'mech., 1927. — Goetsch, W. 1921. Regeneration und Transplantation bei
Planarien. I. Arch. Entw'mech., 49. 1922. Regeneration, etc., II. Ihid., 51. 1929.
Das Regenerationsmaterial und seine experimentelle Behandlung. Ihid., 117. Goetsch,
W., und Hellmich, W. 193 1. Feldtheorie Gradientenlehre und relative Determina-
tion. Biol. Zbl., 51. — Goldfarb, A. J. 1906. Experimental study of light as a factor in
the regeneration of hydroids. Jour. Exp. Zool., 3. 1907. Factors in the regeneration
of a compound hydroid, Eudendrium ramosum. Ibid., 4. 1909. The influence of the
nervous system in regeneration. Ibid., 7. 1910. Light as a factor in the regeneration of
hydroids. 2. Ibid., 8. 1914a. Regeneration in the annelid worm, Aniphrnoma pacifica
after removal of the central nervous system. Papers Tortugas Lab. Carnegie Inst., 6.
1914b. Experimentally fused larvae of echinoderms with special reference to their
skeletons. Carnegie Inst. Wash. Publ. 183. 1915. Experimentally fused larvae, etc.
II. Arch. Entw'mech., 41. 1917. The symmetry of grafted eggs in relation to giant
larvae formation in Arhacia punctulata. Biol. Bull., 32. — Goldschmidt, R. 1927. Phy-
siologische Theorie der Vererbung. Berlin. 1938. Physiological genetics. New York
and London. — Goldsmith, E. D. 1932. Abnormal outgrowths induced in planarians
by cautery. Abstr. Amer. Soc Zool., Anat. Rec, 54, Suppl. 1933. Further observa-
tions on supernumerary structures in planarians induced by cautery. Abstr. Amer.
Soc. Zool., Anat. Rec, 57. Suppl. 1934. Correlation in planarian regeneration. Proc
Nat. Acad. Sci., 20. 1937. Production of supernumerary outgrowths in planarians by
the application of ethyl alcohol, aniline oil and coal tar. Abstr. Amer. Soc. Zool., Anat.
Rec, 70, Suppl. I. 1939. Regeneration and accessory outgrowths in planarians. I.
Growth, 3. 1940. Regeneration, etc. II. Physiol. Zool., 13.— Goldstein, K. 1904.

766 PATTERNS AND PROBLEMS OF DEVELOPMENT

Kritische und experimentelle Beitriige zur Frage nach dem Einfluss des Zentralnerven-
systems auf die embryonale Entwicklung und die Regeneration. Arch. Entw'mech.,
1 8. — Gostejewa, M. 1935. Artbesonderheiten der Linsenbildung aus dem Rumpf-
epithel bei Rana escidenta, R. ridibunda und R. temporaria. Jour. Biol., 4. — Gowan-
lock, J. N. 1923. Reversal of vertebrate heart beat. Abstr. Amer. Soc. ZooL, Anat.
Rec, 24. Graper, L. 1922a. Extremitatentransplantation an Anuren. I, IE Arch.
Entw'mech., 51. 1922b. Extremitatentransplantation an Anuren. Anat. Anz., 55,
Erganzungsh., Verh. anat. Ges. 1923. Determination und Differenzierung. Arch.
Entw'mech., 48. 1924. Extremitatentransplantation, etc. IV. Ibid., 102. 1925. Ex-
tremitatentransplantation, etc. V. Ibid., 105. 1926a. Zur Genese der Polydactylie.
Ibid.,io-j. 1926b. Die Potenzen langsgespaltener Beinknospen von 5«/o OT>/^ij-. Ibid.
1929. Die Primitiventwicklung des Hiihnchens nach stereokinematischen Untersuch-
ungen, kontroUiert durch vitale Farbmarkierung und verglichen mit der Entwicklung
anderer Wirbeltiere. Ibid., 116. — Griffin, B. B. 1899. Studies on the maturation, fer-
tilization and cleavage of Thalassema and Zirphaea. Jour. Morphol., 15. — Grobben,
K. 1879. Die Entwicklungsgeschichte von Moiiia rcctirostris. Arb. zool. Inst. Wien,
II. — Grodzinski, Z. tjber die Entwicklung von unterkiihlten Hiihnereiers. Arch.
Entw'mech., 129. i934- Weitere Untersuchungen iiber den Einfluss der Unterklihlung
auf die Entwicklung der Hiihnereier. Ibid., i^o. — GroU, O. 1924. tJber Transplanta-
tion von Ruckenhaut an Stelle der Conjunctiva bei Larven von Rana fusca (Rosel).
Arch. mikr. Anat. Entw'mech., 100. — Gross, F., and Huxley, J. S. 1935. Regenera-
tion and reorganization in Sabclla. Arch. Entw'mech. 133. — Giinthert, T. 1910. Die
Eibildung der Dytisciden. Zool. Jahrb., Abt. Anat. Ont. — Gurwitsch, A. 1895. Uber
die Einwirkung des Lithiumchlorids auf die Entwicklung der Frosch- und Kroteneier.
Anat. Anz., 11. 1896. Uber die formative Wirkung des veriinderten chemischen Medi-
ums auf die embryonale Entwicklung. Arch. Entw'mech., 3. 1922. Uber den Begriff
des embryonalen Feldes. Ibid., st^- 1923- Versuch einer synthetischen Biologic. Abh.
z. theoret. Biol., H. 17. 1927. Weiterbildung und Verallgemeinerung des Feldbe-
griffes. Arch. Entw'mech., 114. — Guyenot, E. 1927a. La perte du pouvoir regenera-
teur des Anoures etudiee par les heterogreffes, et la notion de territoires. Rev. suisse
Zool., 34. 1927b. La probleme morphogenetique dans la regeneration des Urodeles.
Ibid. 1928. Territoire de regeneration chez le lezard {Lacerta muraiis). C. R. Soc.
Biol., 99. Guyenot, E., et Ponse, K. 1930. Territoires de regeneration et transplanta-
tion. Bull. biol. France Belg., 64. Guyenot, E., et Schotte, O. 1926a. Demonstration
de I'existence de territoires specifiques de regeneration par le methode de la deviation
des troncs nerveux. C. R. Soc. Biol., 94. 1926b. Le role du systeme nerveux dans
I'edification des regenerats des pattes chez les Urodeles. C. R. Soc. phys. hist. nat.
Geneve, 43.

Haeckel, E. 1869. Zur Entwicklungsgeschichte der Siphonophroen. Utrecht. 188 1.
Metagenesis und Hypogenesis von Aurelia aurita. Jena. — Hammerling, J. 1924a.
Uber dauernd teilungsfiihige Korperzellen bei Aeolosoma hemprichi (Ehrb.). Biol.
Zbl., 44. 1924b. Die ungeschlechthche Fortpflanzung und Regeneration bei Aeolo-
soma hemprichi. Zool. Jahrb. Abt. allg. Zool., 49. 1927. Die Umkehrung der Polaritat
des ungefurchten Lies von Rana fusca und ihre Folgeerscheinungen. Arch. Ent-
w'mech., 1 10. 1932. Zur Frage des natiirlichen Todes, besonders der vielzelligen Tiere.
Naturwiss., 6. 1934a. Uber formbildende Substanzen bei Acctabidaria mcditciranca

BIBLIOGRAPHY 767

etc. Arch. Entw'mech., 131. 1934&. tJber Genomwirkungen und Formbildungsfahig-
keit bei Acetahularia. Ibid., 132. 1934c. Regenerationsversuche an kernhaltigen und
kernlosen Zellteilen von Acetabularia weUsteini. Biol. Zbl., 54. 1936. Studien zum
Polaritatsproblem. Zool. Jahrb. Abt. allg. Zool., 56. — Haffner, K. von. 1931. Die
iiberzahligen Bildungen des Korperstammes von Lumbriculus variegatus etc. Arch.
Entw'mech., 123. — Hall, E. K. 1937. The determination of the axes of the embryonic
ear. Jour. Exp. Zool., 75. 1939. On the duration of the polarization process in the ear
primordium of embryos of Ambl ystoma punctatum (L'mn.). Ibid.,%2. — Hallez, P. 1879.
Contribution a I'histoire naturelle des Turbellaries. Lille. — Hamburger, V. 1928. Die
Entwicklung experimentell erzeugter nervenloser und schwach innervierter Extremi-
taten von Anuren. Arch. Entw'mech., 114. 1939. The development and innervation of
transplanted limb primordia of chick embryos. Jour. Exp. Zool., 80. — Hanaoka, K.
1935. Experiments on the polarity of a stalked medusa Thaumatoscyphus distinctus
Kishinouye. Jour. Fac. Sci. Hokkaido Imp. Univ., Ser. VI, Zool., 4. — Hargitt, C. W.
1915. Regenerative potencies of dissociated cells of hydromedusae. Biol. Bull., 28. —
Harmer, S. F. 1893. On the occurrence of embryonic fission in cyclostomatous Poly-
zoa. Quart. Jour. Micr. Sci., 34. 1896. On the development of Lichcnospora vernuaria
Fabr. Ibid., 39. 1898. On the development of Tubulipora, etc. Ibid., 41. — Hamly,
M. H. 1926. Localization of the micromere material in the cytoplasm of the egg of
Arbacia. Jour. Exp. Zool., 45. — Harper, E. H. 1904. Notes on regulation in 5/y/arm
lacustris. Biol. Bull., 6. — Harper, R. A. 1926. Morphogenesis in Didyostclium. Bull.
Torrey Bot. Club. 53. 1929. Morphogenesis in Po/y5^o»Jy/fww. Ibid., $6. 1932. Or-
ganization and light relations in Polyspondylium. Ibid., 59. — Harrison, R. G. 1920.
Experiments on the lens in Amblystoma. Proc. Soc. Exp. Biol. Med., 17. 1921a. On
relations of symmetry in transplanted limbs. Jour. Exp. Zool., 32. 1921b. Experiments
on the development of the gills in the amphibian embryo. Biol. Bull., 41. 1924, Some
unexpected results of the heteroplastic transplantation of hmbs. Proc. Nat. Acad. Sci.,
10. 1925a. The effect of reversing the mediolateral or transverse axis of the fore-limb
bud in the salamander embryo. Arch. Entw'mech., 106. 1925b. The development of
the balancer in Amblystoma, etc. Jour. Exp. Zool., 41. 1929a. Correlation in the
development and growth of the eye, etc. Arch. Entw'mech., 120. 1929b. Hetero-
plastic transplantation in amphibian embryos. X^ Congres internat. zool., 1933. Some
difficulties of the determination problem. Amer. NaturaHst, 67. 1935. Factors con-
cerned in the development of the ear in Amblystoma punctatum. Abstr. Amer. Assoc.
Anat., Anat. Rec, 64, Suppl. i. 1936a. Relations of symmetry in the developing ear
oi Amblystoma punctatum. Proc. Nat. Acad. Sci., 22. 1936b. Relations of symmetry
in the developing embryo. Collecting Net, 11. 1937. Embryology and its relations.
Science, 85. — Harvey, Ethel. 1932. The development of half and quarter eggs of
Arbacia punctulata and of strongly centrifuged whole eggs. Biol. Bull., 62. 1936.
Parthenogenetic merogony or cleavage without nuclei in Arbacia punctulata. Ibid., 71.
1939. A comparison of the development of Arbacia half -eggs activated before and
after centrifugation. Abstr. Amer. Soc. Zool., Anat. Rec, 75, Suppl. — Hasama, B.
1935- Uber die bioelektrischen Phanomene am Ei von Hynobius nebulosus bei kiinst-
licher Befruchtung. Protoplasma, 22. — Haseman, J. D. 1907. The direction of dif-
ferentiation in regenerating crustacean appendages. Arch. Entw'mech., 24. — Hatai, S.
1924a. Contributions to the physiology of the earthworm. I. Sci. Repts. Tohoku Imp.

768 PATTERNS AND PROBLEMS OF DEVELOPMENT

Univ., Ser. 4, i. 1924b. Reply to the remarks of Prof. Wilhelm Michaelson concerning
the Perichaeta mcgascolidoides Goto et Hatai, etc. Ibid. — Hatschek, B. 1881. Uber
Entwicklungsgeschichte von Teredo. Arb. zool. Inst. Wien, 3. 1886. Entwicklung der
Trochophora von Eupomatus uncinatus Philippi. Ibid., 6. — Hatt, P. 1931. La fusion
experimentale d'ocufs de "Sabellaria alveolata L." Arch. Biol., 42. 1932. Essais ex-
perimentaux sur les localisations germinales dans I'oeuf d'un annelide {Sabellaria al-
veolata L.). Arch. Anat. micr., 28. 1934. Experiences d'induction sur la gastrula de
Triton au moyen de parties de blastoderme jeune de poulet. Ibid., 30. — Heath, H.
1899. The development ol Ischnochilon. Zool. Jahrb. Abt. Anat. Ont., 12. — Heatley,
N. G., and Lindahl, P. E. 1937. Studies on the nature of the amphibian organization
centre. \. Proc. Roy. Soc, B, 122. Heatley, N. G.; Waddington, C. H.; and Need-
ham, J. 1937. Studies, etc. \L Proc. Roy. Soc, B, 122. — Hegner, R. W. 1908.
Effects of removing the germ cell determinants from the eggs of some chrysomelid
beetles. Biol. Bull., 16. 1909. The effects of centrifugal force upon the eggs of some
chrysomeHd beetles. Jour. Exp. Zool., 6. 1911. Experiments with chrysomelid beetles.
in. Biol. Bull., 20. — Helff, O. M. 1928. Studies on amphibian metamorphosis. III.
Physiol. Zool., i. 1934a. Studies, etc. XII. Jour. Exp. Zool., 68. 1934b. Studies,
etc. XIII. Biol. Bull., 66. 1934c. Studies, etc. XIV. Anat. Rec, 59.— Hellmich, W.
1930. Untersuchungen iiber Herkunft und Determination des regenerativen Materials
bei Amphibian. Arch. Entw'mech., 121. — Herbst, C. 1892. Experimen telle Unter-
suchungen liber den Einfluss der veranderten chemischen Zusammensetzung des um-
gebenden Mediums auf die Entwicklung der Tiere. I. Zeitschr. wiss. Zool., 55. 1895.
Experimentelle Untersuchungen etc. II. Mitt. zool. Stat. Neapel, 11. 1896a. Experi-
mentelle Untersuchungen etc. III-VI. Arch. Entw'mech., 2. 1896b. IJber die Re-
generation von antennenahnlichen Organen an Stelle von Augen. I. Ibid., 2. 1896c.
tjber die Regeneration etc. II. Vierteljahrschr. naturforsch. Ges. Zurich. Jhg., i8g6.
1897. iJber die zur Entwicklung der Seeigellarven notwendigen anorganischen Stoffe,
ihre RoUe und ihre Vertretbarkeit. I. Arch. Entw'mech., 5. 1899. Uber die Regenera-
tion etc. Ill, IV. Ibid., 4. 1900. tJber das Auseinandergehen von Furchungs- und
Gewebszellen im kalkfreien Medium. Ibid., 9. 1901a. Uber die zur Entwicklung der
Seeigellarven etc. II. Ibid., 11. 1901b. Uber die Regeneration etc. V. Ibid., 12,. 1904.
Uber die zur Entwicklung etc. III. Ibid.,!-]. 1910. Uber die Regeneration etc. VI.
Ibid. 30. 1916. Uber die Regeneration etc. VII. Ibid., ^2. 1917. Uber die Regenera-
tion der Schwanzbeine von Palacnion nach Entfernung der Schwanzganglien. Ibid.,
43. — Hering, E. 1897. Theory of the functions of Hving matter. Brain, 20. — Herla, V.
1893. Etude des variations de la mitose chez 1' Ascaris megalocephala. Arch. Biol., 13.
— Herlant, M. 191 1. Recherches sur les oeufs di- et trispermiques de grenouille. Arch.
Biol., 26. — Herlitzka, A. 1896. SuUo sviluppo di embrioni completi da blastomeri iso-
lati di uova di tritone {Molge cristata). Arch. Entw'mech., 4. — Herrick, C. J. 1924.
Neurological foundations of animal behavior. New York. — Hertwig, G. 191 1. Radi-
umbestrahlung unbefruchteter Froscheier und ihre Entwicklung nach Befruchtung
mit normalem Semen. Arch. mikr. Anat., 77, Abt. 2. Hertwig, G. und P. 1913- Be-
einflussung der mannlichen Keimzelle durch chemische Stoffe. Arch. mikr. Anat., 83,
Abt. 2. — Hertwig, O. 1892. Urmund und Spina bifida. Arch. mikr. Anat., 39. 1895.
Beitrage zur experimentellen Morphologie und Entwicklungsgeschichte. Ibid., 44.
1898a. Uber den Einfluss der Temperatur auf die Entwicklung von Rana fusca und R.

BIBLIOGRAPHY 769

esculenta. Ibid., 51. 1898b. Beitrage zur experimentellen Morphologic und Entwick-
lungsgeschichte. IV. Ibid., si- 1904. Weitere Versuche iiber den Einfluss der Zentrif-
ugalkraft auf die Entwicklung tierischer Eier. Ibid., 67,. 1911. Die Radiumkrankheit
tierischerZellen. Ibid., yj, Aht. 2. 1913. Versuche an Tritoneiern iiber die Einwirkung
bestrahlter Samenfaden. Ibid., 82, Aht. 2. Hertwig, O. und R. 1880. Der Organismus
der Medusen und seine Stellungzur Keimblatttheorie. Denkschr. med. naturwiss. Ges.
Jena. Jhg. 1880. — Hescheler, K. 1896. tJber Regenerationsvorgange bei Lumbriciden.
Jen. Zeitschr., N.F., 23. — Heymons, R. 1893. Zur Entwicklungsgeschichte von Um-
brella meditenanea. Zeitschr. wiss. Zool., 56. — Higgins, G. M., and Sheard, C. 1926.
Effects of ultraviolet radiation on the early larval development of Rana pipiens. Jour.
Exp. Zool., 46. — Hinrichs, Marie. 1924a. A study of the physiological effects of caf-
fein upon Planaria dorotocephala. Jour. Exp. Zool., 40. 1924b. A demonstration of
the a.xial gradient by means of photolysis. Ibid., 41. 1925. Modification of develop-
ment on the basis of differential susceptibility to radiation. I. Jour. MorphoL, 11.
1926a. Modification of development, etc. II. Biol. Bull., 50. 1926b. Modification of
development, etc. III. Ibid. 1927. Modification of development. IV. Jour. Exp.
Zool., 47. 1928. Ultraviolet radiation and division in Paramecium caudatum. Physiol.
Zool., I. 1929. Modification of development in Arbacia punctidata on the basis of dif-
ferential susceptibility to certain alkaloids. Proc. Soc. Exp. Biol. Med., 27. Hinrichs,
Marie, and Genther, Ida. 193 1. Ultraviolet radiation and the production of twins and
double monsters. Physiol. Zool., 4. — His, W. 1874. Unsere Korperform und das phys-
iologische Problem ihrer Entstehung. Leipzig. — Hoadley, L. 1926a. Developmental
potencies of parts of the early blastoderms of the chick. I, II, III. Jour. Exp. Zool.,
43. 1926b. The /;/ situ development of sectioned chick blastoderms. Arch. Biol., 36.
1927. Concerning the organization of potential areas in the chick blastoderm. Jour.
Exp. Zool., 48. 1928. On the localization of developmental potencies in the embryo of
Fundulus heteroclitus. Ibid., 52. 1938. The effect of supramaximum temperatures on
the development of Rana pipiens. Growth, 2. — Horstadius, S. 1927. Studien uber
die Determination bei Paracentrotus lividus Lk. Arch. Entw'mech., 112. 1928a.
Transplantationsversuche am Keim von Paracentrotus lividus Lk. Ibid., 113. 1928b.
IJber die Determination des Keimes bei Echinodermen. Acta zool., 9. 1931. Uber die
Potenzverteilung im Verlaufe der Eiachse bei Paracentrotus lividus Lk. Ark. Zool., 23.
1935- Uber die Determination im \'erlaufe der Eiachse bei Seeigeln. Pubbl. Staz. zool.
Napoli, 14. 1936a. Uber die zeitliche Determination im Keim von Paracentrotus
lividus Lk. Arch. Entw'mech., 135. 1936b. Weitere Studien iiber die Determination
im Verlaufe der Eiachse bei Seeigeln. Ibid. 1936c. Studien iiber heterosperme See-
igelmerogone etc. Mem. Mus. roy. d'Hist. nat. Belg., ser. 2, fasc. 3, 1936. 1937a. In-
vestigations as to the localization of the micromere-, skeleton- and entoderm-forming
material in the unfertilized egg of Arbacia. Biol. Bull., 73. 1937b. Experiments on
determination in the early development of Cerebratulus lacteus. Ibid. 1938. Schniir-
ungsversuche an Seeigelkeimen. Arch. Entw'mech., 138. Horstadius, S., und Wolsky,
A. 1936. Studien iiber die Determination der Bilateralsymmetrie des jungen Seeigel-
keimes. Arch. Entw'mech., 135.— Hogue, Mary. 1910. Uber die Wirkung der Cen-
trifugalkraft auf die Eier von Ascaris megalocephala. Arch. Entw'mech., 29. — Holm-
dahl, D. E. 1925. Die erste Entwicklung des Korpers bei den Vogeln und Saugetieren.
Morphol. Jahrb., 54. 1926. Die erste Entwicklung etc. Ibid., SS- 1933. Die zweifache

77© PATTERNS AND PROBLEMS OF DEVELOPMENT

Bildungsweise des zentralen Nervensystems bei den Wirbeltieren. Arch. Entw'mech.,
129. 1935. Primitivstreifen bzw. Rumpfschwanzknospe im Verhiiltnis zur Korperent-
wicklung. Zeitschr. mikr. Anat. Forsch., 38. — Holmes, Gladys. 1931. The influence
of the nervous system on regeneration in Nereis virens Sars. Jour. Exp. Zool., 60.—
Holmes, S. J. 1899. Reversal of cleavage in Ancylus. Amer. Naturalist, ^2>- 1900.
The early development of Planorbis. Jour. Morphol., 16. 191 1. Minimal size reduc-
tion in planarians through successive regenerations. Ibid., 22. — Holtfreter, J. 1929.
tJber die Aufzucht isolierter Teile des Amphibienkeimes. Arch. Entw'mech., 117.
1931a. Potenzpriifung am Amphibienkeim mit Hilfe der Isolationsmethode. Verb,
deutsch. zool. Ges., 1931. 1931b. Uber die Aufzucht etc. II. Arch. Entw'mech., 124.
19330. Nicht typische Gestaltungsbewegungen sondern Induktionsvorgange bedingen
medullare Entwicklung von Gastrulaketoderm. Ibid., 127. 1933b. Der Einfluss von
Wirtsalter und verschiedenen Organbezirken auf die Differenzierung von angelagertem
Gastrulaektoderm. Ibid. 1933c. Nachweis der Induktionsfiihigkeit abgetoter Keim-
teile. Ibid., 128. 1933d. Die totale Exogastrulation, eine Selbstablosung des Ekto-
derms vom Mesoderm etc. Ibid., 129. 19336. Organisierungsstufen nach regionaler
Kombination von Entomesoderm mit Ektoderm. Biol. Zbl., 53. 1934a. Der Einfluss
thermischer, mechanischer und chemischer Eingriffe auf die Induzierfahigkeit von
J"r/to«-Keimteilen. Arch. Entw'mech., 132. 1934b. tJber die Verbreitung induzieren-
den Substanzen und ihre Leistungen in Tr/tow-Keim. Ibid. 19350. Morphologische
Beeinflussung von Urodelenektoderm bei xenoplastische Transplantation. Ibid., 133.
1935b. ijber das Verhalten von Anurenektoderm in Urodelenkeimen. Ibid. 1936. Re-
gionale Induktionen in xenoplastisch zusammengesetzten Explantaten. Ibid., 134.
1938a. Veranderungen der Reaktionsweise im alternden isolierten Gastrulaektoderm.
Ibid., 138. 1938b. Difi'erenzierungspotenzen isolierter Teile der Urodelengastrula.
Ibid. 1939a. Studien zur Ermittlung der Gestaltungs faktoren in der Organentwick-
lung der Amphibien. I. Ibid., i^g. 1939b. Studien etc. II. Ibid. — Howland, R. B.,
and Child, G. P. 1935. Experimental studies on development in Drosophila melano-
gaster. Jour. Exp. Zool., 70. Howland, R. B., and Sonnenblick, B. P. 1936. Studies,
etc. II. Jour. Exp. Zool., 73. — Hunt, H. R. 19 15. Regeneration posteriorly in Enchy-
traeus albidus. Amer. Naturalist, 49. — Hunt, T. E. 1929. An experimental study of
Hensen's node in the chick embryo. Proc. Soc. Exp. Biol. Med., 27. 1931. An experi-
mental study of the independent differentiation of the isolated Hensen's node and its
relation to the formation of axial and non-axial parts in the chick embryo. Jour. Exp.
Zool., 59. 1932. Potencies of transverse levels of the chick blastoderm in the definitive
streak stage. Anat. Rec, 55. — Hurd, A. M. 1919. Some orientating elYects of mono-
chromatic lights of equal intensities on Fucus spores and rhizoids. Proc. Nat. Acad.
Sci., 5. 1920. Effect of unilateral monochromatic light and group orientation on the
polarity of germinating Fucus spores. Bot. Gaz., 70. — Huxley, J. S. 1921a. Studies
in dedifferentiation. Quart. Jour. Micr. Sci., 65. 1921b. Further studies on restitution
bodies and free tissue culture in 5yco«. Ibid. 1922. Dedifferentiation in £f//z7M<5 larvae
and its relation to metamorphosis. Biol. Bull., 43. 1926. Studies in dedifferentiation.
VI. Pubbl. Staz. zool.NapoIi, 7. 1932. Problems of relative growth. New York. Huxley,
J. S., and De Beer, G. R. 1923. Studies in dedifferentiation. IV. Quart. Jour. Micr. Sci., 67
1934. Elements of experimental embryology. Cambridge. Huxley, J. S. and Teissier, G.
1936. Terminology of relative growth. Nature, 137. — Hyde, Ida. 1904. Differences in

BIBLIOGRAPHY 771

electric potential in developing eggs. Amer. Jour. Physiol., 12. — Hyman, L. H. 1916a.
An analysis of the process of regeneration in certain microdrilous oligochaetes. Jour.
Exp. Zool., 20. 1916b. On the action of certain substances on oxygen consumption.
Amer. Jour. Physiol., 40. 1917. The metabolic gradients in .4 wof^a. Jour. Exp. Zool.,
24. 1918. Suggestions regarding the causes of bioelectric phenomena. Science, 48.
1919a. On the action, etc. II. Amer. Jour. Physiol., 48. 1919b. Physiological studies
on Planar ia. Ibid., 4g. 1919c. Physiological studies, etc. II. Ibid., $0. 1919c?. Physi-
ological studies, etc. III. Biol. Bull., 37. 19196. On the action, etc. III. Ibid. 19200.
Physiological studies, etc. IV. Amer. Jour. Physiol., 53. 1920b. The axial gradients in
Hydrozoa. III. Biol. Bull., 38. 1920c. On the action, etc. IV. Publ. Puget Sound
Biol. Sta., 2. 1921. The metabolic gradients of vertebrate embryos. I. Biol. Bull., 40.
1923a. Oxygen consumption with respect to level, size and regeneration, and electric
polarity in Corymorpha. Abstr. Amer. Soc. Zool., Anat. Rec, 24. 1923b. Physiological
studies, etc. V. Jour. Exp. Zool., 37. 1925. Respiratory differences along the axis of
the sponge, Grantia. Biol. Bull., 48. 1926a. The axial gradients in Hydrozoa. VIII.
Ibid., 50. 1926b. The metabolic gradients of vertebrate embryos. II. Jour. Morphol.
Physiol., 42. 1927a. The metabolic gradients, etc. III. Biol. Bull., 52. 1927b. The
metabolic gradients, etc. IV. Ibid. 1928. Miscellaneous observations on Hydra with
special reference to reproduction. Ibid., ^4- i93i- Studies on the morphology, taxono-
my and distribution of North American triclad Turbellaria. IV. Trans. Amer. Micr.
Soc, 50. 1932a. The axial respiratory gradient: experimental and critical. Physiol.
Zool., 5. 1932b. Studies on the correlation between metabolic gradients, electric gradi-
ents and galvanotaxis. II. Ibid. 1939. North American triclad Turbellaria. IX.
Trans. Amer. Micr. Soc, 58. Hyman, L. H., and Bellamy, A. W. 1922. Studies on the
correlation between metabolic gradients, electric gradients and galvanotaxis. Biol.
Bull., 43. Hyman, L. H., and Galigher, A. E. 1921. Direct demonstration of the ex-
istence of a metaboHc gradient in annelids. Jour. Exp. Zool., 34.

Ichikawa, M. 1934. Experimental analysis of the gill formation in the Japanese
urodele, Triturus pyrrhogaster (Boie). Proc. Imp. Acad. Tokyo, 10. 1938. Induction
of gills in Triturus pyrrhogaster. Ibid., 14. — Ikeda, Y. 1937a. Beitrage zur ent-
wicklungsmechanischen Stiitze der Kupfferschen Theorie der Sinnesplacoden. Arch.
Entw'mech., 136. 1937b. tJber die Bildung akzessorischer Retina aus dem Tapetum
bei Hynobiiis. Ibid. — Ikeno, S. 1898. Untersuchungen iiber die Entwicklung der
Geschlechtsorgane und den Vorgang der Befruchtung bei Cycas revoluta. Jahrb. wiss.
Bot., ^2. — Inaba, D. 1930. Notes on the development of a holothurian, Caudina chil-
ensis. Sci. Repts. Tohoku Imp. Univ., Ser. 4, 5. — Ivanov, P. 1908. Die Regeneration
des vorderen und des hinteren Korperendes bei Spirographis spallanzanii. Zeitschr.
wiss. Zool., 91.

Jackson, C. M. (ed.). 1914. Morris, human anatomy. 5th ed. New York. — Jacobs,
Maria. 1925. Entwicklungsphysiologische Untersuchungen am Copepodenei {Cyclops
viridis Jurine). Zeitschr. wiss. Zool., 124. — Jacobson, W. 1938. The early develop-
ment of the avian embryo. Jour. Morphol., 62. — Janda, V. 1912a. Die Regeneration
der Geschlechtsorgane bei Cr/oJ/'/Zz^.y /acwzfm Hoffm. I. Arch. Entw'mech., 33. 1912b.
Die Regeneration etc. II. Ibid., t,4. 1913. Fiihlerahnliche Heteromorphosen an Stelle
von Augen bei Stylopyga orientalis und Tenebrio molitor. Ibid., 36. — Jenkinson, J. W.
1906a. On the relation between the symmetry of the egg and the symmetry of the em-

772 PATTERNS AND PROBLEMS OF DEVELOPMENT

bryo in the frog (Raiia tempoi'aria) . Biometrika, 5. 1906b. On the effect of certain so-
lutions upon the development of the frog egg. Arch. Entw'mech., 21. 1909. On the
relation between the symmetry of the egg, the symmetry of segmentation and the sym-
metry of the embryo in the frog. Biometrika, 7. 191 la. On the effect of certain isotonic
solutions on the development of the frog. Ibid., ^2. 1911b. On the origin of the polar
and bilateral structure of the egg of the sea urchin. Ibid. — Jennings, H. S. 1896. The
early development of Asplanchna herrickii. Bull. Mus. Comp. Zool. Harvard, 30. —
Jensen, O. S. 1883. Recherches sur la spermatogenese. Arch. Biol., 4. — Johnson,
H. P. 1901. The polychetes of the Puget Sound region. Proc. Boston Soc. Nat.
Hist., 29. 1902. Collateral budding in annelids of the genus Trypanosyllis. Amer.
Naturalist, 36. — Johnson, Myrtle. 1910. A quantitative study of the development of
the chain in Salpa Jusijormis-uncinata. Univ. Calif. Publ. Zool., 6. — Just, E. E. 1912.
The relation of the first cleavage plane to the entrance point of the sperm. Biol. Bull.,
22. 1933. Observations on effects of ultraviolet rays upon living eggs of Nereis limbata
exposed before insemination. Arch. Entw'mech., 130. 1939. The biology of the cell
surface. Philadelphia.

Kaan, H. W. 1926. Experiments on the development of the ear of Amblystoma
pHHctatiim. Jour. E.xp. Zool., 46. 1930. The relation of the developing auditory vesicle
to the formation of cartilage capsule in A mbly stoma punctatum. Ibid., 55. — Kalmus, H.
1928. Untersuchungen iiber die Atmung von Paramecium caudatum. Zeitschr. vergl.
Physiol., 7. — Kappers, C. U. A. 1934. Differences in the effect of various impulses on
the structure of the nervous system. Irish Jour. Med. Sci., 1934. — Kautsch, G. 1912.
Studien uber EntwicklungsanomaKen bei Ascaris. L Arch. Zellforsch., 8. 1913.
Studien etc. H. Arch. Entw'mech., 35. — Kawaguti, S. 1932. Studies on the disinte-
gration of Branchiura sp. by methylene blue. Mem. Fac. Sci. Taihoku Imp. Univ., 7.
1934. Studies, etc. II. Ibid. — Keil, E. 1924. Studien iiber Regulationserscheinungen
an Polycelis nigra. Arch. Entw'mech., 102. — Keiller, V. K. 1910. A histological study
of regeneration in short head pieces of Planar ia simplicissima. Arch. Entw'mech., 31.
— Kellicott, W. M. 1916. The effects of low temperature upon the development of
Fundulus. Amer. Jour. Anat., 20. — Kellogg, V. L. 1904. Restorative regeneration in
nature of the starfish Linckia diplax. Jour. E.xp. Zool., i. — Kenk, R. 1922. Devel-
oppement et regeneration de I'appareil copulateur chez les triclades paludicoles. Arch.
Zool. exp., 61. 1935a. A morphological proof of the existence of zooids in Euplanaria
dorotocephala. Physiol. Zool., 8. 1935b. Studies on Virginian triclads. Jour. Elisha
Mitchell Sci. Soc, 51. — Kennel, J. 1883. Biologische und faunistische Notizen aus
Trinidad. Arb. zool.-zootom. Inst. Wiirzburg, 6. 1888. Untersuchungen an neuen
Turbellarien. Zool. Jahrb. Abt. Anat. Ont., 3. — Kent, S. 1880-82. A manual of In-
fusoria.— Kerschner, L. 1879. Uber zwei neue Notodelphyiden etc. Denkschr. Acad.
Wiss. Wien, math.-nat. CI., 41. — Kesselyak, A. 1936. Beitriige zur Kenntnis der
.A.uslosungsfaktoren der Wolft'schen Linsenregeneration. Arch. Entw'mech., 134. —
King, Helen. 1898. Regeneration in Asterias vulgaris. Arch. Entw'mech., 7. 1900.
Further studies on regeneration in Asterias vulgaris. Ibid.,q. 1901. Observations and
experiments on regeneration in Hydra viridis. Ibid., 12. 1905. Experimental studies
on the eye of the frog embryo. Ibid., 19. — King, R. L., and Beams, H. W. 1938. An
experimental study of chromatin diminution in Ascaris. Jour. Exp. Zool., 77. — Kings-
bury, B. F. 1924. The significance of the so-called law of cephalocaudal differential

BIBLIOGRAPHY

773

growth. Anat. Rec, 27. 1926. On the so-called law of anteroposterior development.
Ibid., 2)i- 1932. The "law" of cephalocaudal differential growth in its application to
the nervous system. Jour. Comp. Neurol., 56. — Klausner, F. 1890. Mehrfachbildung-
en bei Wirbeltieren. Miinchen. — Kleinenberg, N. 1879. The development of the
earthworm, Lumbricus trapczoides. Quart. Jour. Micr. Sci., 19. — Kniep, H. 1907.
Beitrage zur Keimungsphysiologie und Biologie von Fuciis. Jahrb. wiss. Bot., 44. —
Kohler, P. 1907. Beitrage zur Kenntnis der Reproduktions- und Regenerationsvor-
gange bei Pilzen etc. Flora, 97. — Koelitz, W. 1910. Morphologische und experimen-
telle Untersuchungen an //_v(/;-a. Arch. Entw'mech., 31. — Kofoid, C. A. 1895. On the
early development of Lhnax. Bull. Mus. Comp. Zool. Harvard, 27. Kofoid, C. A., and
Swezy, Olive. 1915. Mitosis and multiple fission in trichomonad flagellates. Proc.
Amer. Acad. Arts Sci., 51. — Koltzoff, N. K. 1906. Studien uber die Gestalt der Zelle.
I. Arch. mikr. Anat., 67. 1909. Studien etc. II. Arch. Zellforsch., 2. — Komori, S.
1933. Effect of pH on the regeneration and polarity of the tubularian stem. Annot.
Zool. Japon., 14. — Kopec, S. 1922. Studies on the necessity of the brain for insect
metamorphosis. Biol. Zbl., 42. — Kopenhaver, Mary. 1937. Axial differences in water-
absorbing properties of oligochete tissues. Physiol. Zool., 10. — Kopsch, F. 1896. Ex-
perimen telle Untersuchungen iiber den Keimhautrand der Salmoniden. Verh. anat.
Ges., 1896. 1904. Untersuchungen fiber Gastrulation und Embryobildung bei Chor-
daten. Leipzig. 1926a. Die Lage des Primitivstreifens im Hiihnerei. Zeitschr. mikr.
anat. Forsch., 8. 1926b. Primitivstreifen und organbildende Keimbezirke beim Hiihn-
chen. Ibid. 1929. Ortliche Vitalfarbung bei ^fv/Z/z/w-Embryonen. Ibid., 10. 19340.
Die Lage des Materials fiir Kopf, Primitivstreifen und Gefasshof in der Keimscheibe
der unbebriiteten Hiihnereies etc. Ibid., 35. 19346. Die Beziehungen der aus dem
Primitivstreifen hervorgehenden Ursegmente zu den Korperabschnitten des Huhnes.
Ibid., 36.— Korschelt, E. 1891. Beitrage zur Anatomie und Physiologie des Zellkerns.
Zool. Jahrb. Abt. Anat. Ont., 4. 1894. tJherOphryotrocha pucriUs Clap.-Metschn. und
die polytrochen Larven eines anderen Anneliden {Harpochaeta cingidata nov. gen.,
nov. sp.). Zeitschr. wiss. Zool., 57. 1904. Uber Doppelbildungen bei Lumbriciden.
Zool. Jahrb., Suppl. 7, Weismann Festschr. 1919. Uber die natiirliche und kiinstliche
Texhing dt?, Ctenodrilus monostylosZt]i\)tY\n. Arch. Entw'mech., 45. 1929. Zur Frage
der morphogenetischen Induktion nach Transplantation. Ibid., 117. 1927-31. Re-
generation und Transplantation. Berlin. Korschelt, E., und Heider, K. 1902. Lehr-
buch der vergleichenden Entwicklungsgeschichte der wirbellosen Tiere, AUg. Teil,
Lief. I. Jena. 1909. Lehrbuch etc., AUg. Teil, Lief. 3.— Kowalewsky, M. von. 1886.
Uber die ersten Entwicklungsprozesse der Knochenfische. Zeitschr. wiss. Zool., 63. —
Krause, G. 1934. Analyse erster Differenzierungsprozesse im Keim der Gewachshaus-
heuschrecke durch kiinstlich erzeugte ZwilHngs-, Doppel- und Mehrfachbildungen.
Arch. Entw'mech., 132. 1938. Einzelbeobachtungen und typische Gesamtbilder der
Entwicklung von Blastoderm und Keimanlage in Ei der Gewachshausheuschrecke
Tachycines asynamorus kdthxng. Zeitschr. Morphol. Okol., 34. — Kfizenecky, J. 1913.
Uber Restitutionserscheinungen an Stelle von Augen bei Tenebrio-Larven nach Zer-
storung der optischen Ganglien. Arch. Entw'mech., 37. — Kropp, B. 1933. Brain
transplantation in regenerating earthworms. Jour. Exp. Zool., 65. — Kriiger, P. 1922.
Die Embryonalentwicklung von Scalpellmn L. Arch. mikr. Anat., 96.— Kuhn, A. 191 2.
Die Sonderung der Keimesbezirke in der Entwicklung der Sommereier von Poly-

774 PATTERNS AND PROBLEMS OF DEVELOPMENT

plicmus pcdicidus De Geer. Zool. Jahrb. Abt. Anat. Ont., 35. Kiihn, A., und Piepho, H.
1938. Die Reaktion der Hypodermis und der versonschen Dlisen auf das \'erpup-
pungshormon bei £/»//f.?//(7 H//w/e/7a Z. Biol. ZbL, 58. — Kiister, E. 1906. Normale und
abnorme Keimungen bei Fucus. Ber. deutsch. bot. Ges., 24. — Kurz, O. 1922. Ver-
suche liber Polaritatsumkehr am Tritonenbein. Arch. Entw'mech., 50. — Kusche, W.
1929. Interplantation umschriebener Zellbezirke aus der Blastula und Gastrula der
Amphibien. I. Arch. Entw'mech., 120.

Lang, A. 1884. Die Polycladen. Fauna und Flora des Golfes von Neapel, 11. —
Lang, P. 1912. Uber Regeneration bei Planarien. Arch. mikr. Anat., 79, Abt. i.
1913a. Experimentelle und histologische Studien an Turbellarien. I. Ibid., 82. 19136.
Beitrage zur Anatomie und Physiologie von Planaria polychroa. Zeitschr. wiss. Zool.,
105. — Le Cron, W. L. 1907. Experiments on the origin and differentiation of the lens
in Amphibia. Amer. Jour. Anat., 6. — Leduc, S. 1910. Theorie physico-chimique de la
vie. Paris. (1911. The mechanism of life.) — Le Fanu, B. 1936. Auxin and correlative
inhibition. New Phytol., 35. — Leger, L. 1909. Les schizogregarines des Tracheates.
2. Arch. Protistenk., 18. — Lehmann, F. E. 1926. Entwicklungsstorungen in der
Medullaranlage von Tn'tow erzeugt durch Unterlagerungsdefekt. Arch. Entw'mech.,
108. 1928. Die Bedeutung der Unterlagerung fiir die Entwicklung der Medullar-
platte. Ibid., Hi- 1929. Die Entwicklung des Anlagemusters im Ektoderm der Triton-
gastrula. Ibid., 117. 1932. Die Beteiligung von Implantats- und Wirtsgewebe bei der
Gastrulation und Neurulation induzierter Embryonalanlagen. Ibid., 12^. 1933. Tier-
ische ISIissbildungen, ihre experimentelle Erzeugung mit chemischen Mitteln und ihre
theoretische Bedeutung. Sitzber. naturforsch. Ges. Bern, 1933. 1934. Die Erzeugung
chordaloser Tritonenlarven durch chemische Behandlung des Gastrulastadiums. Verb,
schweiz. naturforsch. Ges., 1934. 1936a. Stehen die Erscheinungen der Otocephalie
und der Zyklopie bei Triton mit Axialgradienten oder mit Storungen bestimmter Or-
ganisatorregionen in Zusammenhang? Rev. Suisse Zool., 43. 1936b. Selektive Beein-
flussung friihembryonaler Entwicklungsvorgange bei Wirbeltieren. Naturwiss., 24.
1937a. Mesodermisierung des prasumptiven Chordamaterials durch Einwirkung von
Lithium chlorid auf die Gastrula von Triton alpcstris. Arch. Entw'mech., 136. 1937b.
Hemmung der Chordabildung beim Froschembryo durch chemische Mittel. Naturwiss.,
25. 1937c. Die Wirkungsweise chemischer Faktoren in der Embryonalentwicklung.
Rev. Suisse Zool., 44. 1938. Regionale Verschiedenheiten des Organisators von Triton
etc. Arch. Entw'mech., 138. — Lehmann, O. 1911. Die neue Welt der fliissigen Kris-
talle. Leipzig. — Leiby, R. W. 1922. The polyembryonic development of Copidosoma
gclcchiae with notes on its biology. Jour. Morphol., 37. — Lenhossek, M. von. 1898.
Untersuchungen iiber Spermatogenese. Arch. mikr. Anat., 51. — Leplat, G. 1920. Ac-
tion du milieu sur le developpement des larves d'amphibiens etc. Arch. Biol., 30. —
Levy, O. 1906 . Mikroskopische Untersuchung zu Experimenten iiber den Einfluss der
Radiumstrahlen auf embryonale und regenerative Entwicklung. Arch. Entw'mech.,
21. — Lewis, W. H. 1904. Experimental studies on the development of the eye in Am-
phibia. I. Amer. Jour. Anat., 3. 1905. Experimental studies, etc. IL Jour. Exp.
Zool., 2. 1907a. Experimental studies, IIL Ibid., 6. 19076. Lens-formation from
strange ectoderm in Rana sylvatica. Ibid., 7. 1909. The experimental production of
cyclopia in the fish embryo. Anat. Rec, 3. 1912. Experiments on localization and re-
generation in the embryonic shield and germ ring of a teleost fish {Fiindulus hdcro-

BIBLIOGRAPHY 775

clitus). Ibid., 6. — Leypoldt, H. 1910. Transplantationsversuche an Lumbriciden etc.
Arch. Entw'mech., 31. — Li, J., and Shen, S. 1934. Some experiments on the rate of
regeneration of Planaria. Peking Nat. Hist. Bull., 9. — Lieber, A. 1931. Zur Oogenese
einiger Dlopairaarten. Zeitschr. wiss. Zool., 138. — Liesegang, R. E. 1907. Uber die
Schichtungen bei Diffusionen. Leipzig. — Lillie, F. R. 1895. The embr>'ology of the
Unionidae. Jour. Morphol., 10., 1896, On the smallest parts of Slentor capable of re-
generation. Ibid., 12. 1901. Notes on regeneration and regulation in planarians. II-
IV. Amer. Jour. Physiol., 6. 1902. Differentiation without cleavage in the egg of the
annelid, Chaetopterus pergamentaceus. Arch. Entw'mech., 14. 1906. Observations and
experiments concerning the elementary phenomena of embryonic development in
Chaetopterus. Jour. Exp. Zool., 3. 1909. Polarity and bilaterality of the annelid egg:
experiments with centrifugal force. Biol. Bull., 16. 1912. Studies of fertilization in
Nereis. III. Jour. Exp. Zool., 12. 1927. The gene and the ontogenetic process. Sci-
ence, 66. 1929. Embryonic segregation and its role in the life history. Arch. Ent-
w'mech., 118. — Lillie, R. S. 1922. Transmission of physiological influences in proto-
plasmic systems, especially nerve. Physiol. Rev., 2. 1923. Protoplasmic action and
nervous action. Chicago. 1936. The passive iron wire model of protoplasmic and
nervous transmission and its physiological analogies. Biol. Rev., 11. — Lindahl, P. E.
1932a. Zur Kenntnis des Ovarialeies bei dem Seeigel. Arch. Entw'mech., 126. 19326.
Zur experimentellen Analyse der Determination der Dorsoventralachse beim Seeigel-
keim. I, II. Ibid.,i2-j. 1933. Uber "animalisierte" und "vegetativisierte" Seeigellarv-
en. Ibid., 128. 1935. Uber die Rolle des SO4 Ions in der Entwicklung des Seeigelkeims.
Ark. Zool., 28. 1936. Zur Kenntnis der physiologischen Grundlagen der Determination
im Seeigelkeim. Acta zool., 17. Lindahl, P. E., und Ohman, L. O. 1938. Weitere Stu-
dien iiber Stoffwechsel und Determination im Seeigelkeim. Biol.Zbl., 58. Lindahl, P. E.,
und Stordal, A. 1937. Zur Kenntnis des vegetativen Stoflfwechsels in Seeigelei. Arch.
Entw'mech., 136. — Liosner, L. D. 1938. Untersuchungen iiber die Eigenschaften der
Regenerationsknospe des Amphibienschwanzes. Bull. Biol. Med. exp., 6. Liosner,
L. D., et Woronzowa, M. A. 1937. Recherches sur la determination du processus
regeneratif chez les Amphibiens. Arch. Anat. micr., ^s- Liosner, L. D.; Woronzowa,
M. A.; und Kusmina, N. A. 1936. Regenerationspotenz der knochenlosen Extremitat.
Arch. Entw'mech., 134. — Little, C. C. and Bagg, H. J. 1924. The occurrence of four
inheritable morphological variations in mice and their possible relations to treatment
with X-rays. Jour. Exp. Zool., 41. — Litwiller, R. 1939. Mitotic index and size in re-
generating amphibian limbs. Jour. Exp. Zool., 82. — Locatelli, P. 1925. Nuovi esperi-
menti suUa funzione del sistema nervoso suUa rigenerazione. Arch. Sci. biol., 7. 1929.
Der Einfluss des Nervensystems auf die Regeneration. Arch. Entw'mech., 14. — Loeb,
J. 1891. Untersuchungen zur physiologischen Morphologie der Tiere, Bd. I. Wiirz-
burg. 1892. Ibid.,Bd.lI. 1895. The influence of light on the development of organs
in animals. Arch. ges. Physiol., 63. 1912. Heredity in heterogeneous hybrids. Jour.
Morphol., 23. 1915. The bHndness of the cave fauna and the artificial production of
Wind fish embryos by heterogeneous hybridization and low temperatures. Biol. BuU.,
29. — Lopaschov, G. V. 1935. Uber die Ausbildung von regionalen \'erschiedenheiten
im Mesoderm der AmphibiengastrvJa (Russian with German summar>0- Jour. Biol., 4.
— Lovell, H. B. 1937. Determination of the dorsoventral axis of the fore limb in
Amblystotna microstomiim. Abstr. .Amer. Soc. Zool., Anat. Rec, 70, Suppl. i. — Low-

776 PATTERNS AND PROBLEMS OF DEVELOPMENT

ranee, E.W. 1937a. Determination of polarity in Fiicus eggs by temperature gradients.
Proc. Soc. Exp. Biol. Med., 36. 1937&. Effect of temperature gradients upon polarity
in eggs of Fucus furcatus. Jour. Cell. Comp. Physiol., 10. — Ludwig, W. 1932. Das
Rechts-Links-Problem im Tierreich. Berlin. 1935. Analogien zwischen Bestimmung
des Geschlechtes und Bestimmung der Asymmetrieform (Rechts-Links-Problem).
Biol. Zbl., 55. — Lund, B. L. 1918. The toxic action of KCN and its relation to the
state of nutrition and age of the cell as shown by Paramecium and Didinium. Biol.
Bull., 35. — Lund, E. J. 1917. Reversibihty of morphogenetic processes in Bursaria.
Jour. Exp. Zool., 24. 1918a. Quantitative studies on intracellular respiration. I, IL
Amer. Jour. Physiol., 45. 19186. Quantitative studies, etc. III. Ibid., 47. 1921a.
Quantitative studies, etc. V. Ibid., 57. 192 16. Oxygen concentration as a limiting
factor in the respiratory metabolism of Planar ia agilis. Biol. Bull., 41. 1921c. Ex-
perimental control of organic polarity by the electric current. I. Jour. Exp. Zool.,
34. 1922. Experimental control, etc. II. Ibid., 36. 1923a. Experimental control, etc.
III. Ibid., 37. 1923b. Electrical control of organic polarity in the egg of Fucus. Bot.
Gaz., 76. 1924a. The possible role of constant bioelectric currents in growth. Proc.
Soc. Exp. Biol. Med., 21. 1924b. Experimental control, etc. IV. Jour. Exp. Zool., 39.

1925. Experimental control, etc. V. Ibid., 41. 1926. The electric polarity of Obelia
and frog's skin and its reversible inhibition by cyanides, ether and chloroform. Ibid.,
44. 1928a. Relation between continuous bioelectric currents and cell respiration. II.
Ibid., 51. 1928b. Relation, etc. III. Ibid. 1928c. Relation, etc. V. Ibid. 1929a.
Electric polarity in the Douglas fir. Publ. Puget Sound Biol. Sta., 7. 1929b. Relative
dominance of growing points in the Douglas fir. Ibid. 1930. Internal distribution of
the electric correlation potentials in the Douglas fir. Ibid. 1931a. The unequal effect
of O2 concentration on the velocity of oxidation in loci of different electric potential and
glutathione content. Protoplasma, 13. 1931b. External polarity potentials in the apex
of the Douglas fir before and after mechanical stimulation. Plant Physiol., 6. 1931c.
Electrical correlation between living cells in cortex and wood in the Douglas fir. Ibid.
1932a. Control of the flux equilibrium of electrochemical processes and electric polarity
in the Douglas fir by temperature. Ibid., 7. 1932b. Comparison of the effects of tem-
perature on the radial and longitudinal electrical polarities in wood and cortex of the
Douglas fir. Ibid. 1938. The distribution of electric potential on the external surface
of single cells. Science, 87. Lund, E. J., and Bush, M. 1930. Electric correlation po-
tentials in the leaf of Bryophyllum. Plant Physiol., 5. Lund, E. J., and Hanszen, A. H.
193 1. Antagonism between the effects of cyanide and oxygen on the basal oxidations in
Planaria agilis. Abstr. Amer. Soc. Zool., Anat. Rec, 51. Suppl. Lund, E. J., and Ken-
yon, W. A. 1927. Relation between continuous bioelectric currents and cell respiration.
Jour. Exp. Zool., 48. Lund, E. J., and Moorman, J. B. 1931. Electric polarity and
velocity of cell oxidation as functions of temperature. Jour. Exp. Zool., 40. — Lus, J.

1926. Regenerationsversuche an marinen Tricladen. Arch. Entw'mech., 108. — Luther,
A. 1904. Die Eumesostominen. Zeitschr. wiss. Zool., 77. 1927. Entwicklungsmecha-
nische Untersuchungen am Labyrinth einiger Anuren. Soc. sci. Fenn. comm. biol., 2.
Luther, W. 1935. Entwicklungsphysiologische Untersuchungen am Forellenkeim. Biol.
Zbl., 55. 1936a. Potenzpriifung an isolierten Teilstiicken der Forellenkeimscheibe.
Arch. Entw'mech., 135. 1936b. Austausch von priisumptiver Epidermis und Medul-
larplatte beim Forellenkeim. Ibid. 1937a. Transplantations- und Defektversuche am

BIBLIOGRAPHY 777

Organisationszentrum der Forellenkeimscheibe. Ibid., 137. 1937&. tJber einen Fall
atypischer Organbildung bei ForeUenkeim. Ibid. — Lynch, V. 1919. The function of
the nucleus of the living cell. Amer. Jour. Physiol., 48. — Lyon, E. P. 1906. Some re-
sults of centrifuging the eggs of Arbacia. Amer. Jour. Physiol., 15. 1907. Results of
centrifuging eggs. Arch. Entw'mech., 23. — Lyon, Florence. 1901. A study of the
sporangia of gametophytes of Selaginella a pus and Selaginclla rupestris. Bot. Gaz.,
T)2. — Lyon, H. L. 1904. The embryogeny of Ginkgo. Minnesota Bot. Stud., 3.

Maas, O. 1901. Experi men telle Untersuchungen iiber die Eifurchung. Sitzber.
ges. Morph. Physiol., Miinchen, 1901. 1905. Experimentelle Beitrage zur Entwick-
lungsgeschichte der Medusen. Zeitschr. wiss. Zool., 82. 1908. tJber den Bau des
Meduseneies. Verh. deutsch. zool. Ges., 1908. — MacArthur, J. W. 1920. Changes in
acid and alkah tolerance with age in planarians. Amer. Jour. Physiol., 54. 1921.
Gradients of vital staining and susceptibiHty in Planaria and other forms. Ibid., 57.
1924. An experimental study and a physiological interpretation of exogastrulation and
related modifications in echinoderm embryos. Biol. Bull., 46. — McClelland, C. K.
1916. On the regularity of blooming in the cotton plant. Science, 44. — McClendon,
J. F. 1906. On the development of parasitic copepods. I. Biol. Bull., 12. 1907. On
the development, etc. II. Ibid. 1910. The development of isolated blastomeres of the
frog's egg. Amer. Jour. Anat., ID. 1912a. An attempt toward the physical chemistry
of the production of one-eyed monstrosities. Amer. Jour. Physiol., 29. 1912&. The
effects of alkaloids on the development of fish (Fundulus) eggs. Ibid., 21. 191 7. The
direct and indirect calorimetry of Cassiopca xamachana. Jour. Biol. Chem., 32. — Mc-
Cutcheon, M., and Lucke, B. 1924. The mechanism of staining with basic dyes. Jour.
Gen. Physiol., 6. — Magnus, W. 1906. tJber Formbildung der Hutpilze. Arch. Bion-
tol., I. — Maloeuf, N. S. R. 1936. The "metabolic gradient" in the adult earthworm.
Biol. Zbl., 56. — Malowan, S. L. 1934. Wuchsstoffe und Pflanzenwachstum. Proto-
plasma, 21. — Mangold, O. 1920. Fragen der Regulation und Determination an um-
geordneten Furchungsstadien und verschmolzenen Keimen. Arch. Entw'mech., 48.
1921. Situs inversus bei Triton. Ibid. 1923. Transplantationsversuche zur Frage des
Spezifitat und der Bildung der Keimblatter bei Triton. Arch. mikr. Anat. Entw'mech.,
100. 19280. Das Determinationsproblem. I. Ergebn. Biol., 3. 19286. Neue Experi-
mente zur Analyse der friihen Embryonalentwicklung des Amphibienkeims. Natur-
wiss., 16. 1929a. Das Determinationsproblem. II. Ergebn. Biol., 5. 1929b. Organi-
satortransplantationen in verschiedenen Kombinationen bei Urodelen. Arch. Ent-
w'mech., 117. 1929c. Experimente zur Analyse der Determination und Induktion der
MeduUarplatte. Ibid. 1931a. Das Determinationsproblem. III. Ergebn. Biol. , 7.
1 93 1 6. Versuche zur Analyse der Entwicklung des Haftfadens bei Urodelen. Natur-
wiss., 19. 1931C. Transplantations- und Isolationsversuche bei Forellen. Ibid. 1932a.
Autonome und komplementare Induktionen bei x\mphibien. Ibid., 20. 1932b. Ver-
suche zur Analyse der Induktionsmittel in der Embryonalentwicklung. Ibid., 21.
1933- Uber die Induktionsfahigkeit der verschiedenen Bezirke der Neurula von Uro-
delen. Ibid. Mangold, O., und Seidel, F. 1927. Homoplastische und heteroplastische
Verschmelzung ganzer Tritonkeime. Arch. Entw'mech., in. Mangold, O., und Spe-
mann, H. 1927. Uber Induktion von MeduUarplatte durch MeduUarplatte im jiinger-
en Keim etc. Arch. Entw'mech., in. — Marchal, P. 1904. Recherches sur la biologie
et le developpement des Hymenopteres parasites. I. Arch. Zool. exp., ser. 4, 2. — Mar-

778 PATTERNS AND PROBLEMS OF DEVELOPMENT

gen, S., and Schechtman, A. M. 1939. Effect of localized increased temperatures on a
hog egg {H via regilla). Proc. Soc. Exp. Biol. Med., 41. — Marsh, G. 1928. Relation
between continuous bioelectric currents and cell respiration. IV. Jour. Exp. Zool., 51.
1930. The effect of applied electric current on inherent cellular E.M.F. and its possible
significance in cell correlation. Protoplasma, 11. 1934. Antagonism of methylene blue
and KCN for E.M.F. and respiration of frog's skin. Abstr. Amer. Soc. Zool., Anat.
Rec, 60, Suppl. 4. — Martin, E. A. 1933. Polymorphism and methods of asexual re-
production in the annelid, Dodecaceria of Vineyard Sound. Biol. Bull., 65. — Marx, A.
1925. Experimen telle Untersuchungen zur Frage der Determination der MeduUar-
platte. Arch. Entw'mech., 105. 1930. tjber Induktion durch narkotisierte Organisa-
toren. Ibid., 122,. — Marx, W. 1931. Zum Problem der Determination der BilateraHtat
im Seeigelkeim. Arch. Entw'mech., 125. — Maschlanka, Hildegard, 1938. Physiolo-
gische Untersuchungen am Ei der Mehlmotte, Ephestia kuehniella. Arch. Entw'mech.,
137. — Mathews, A. P. 1903. Electric polarity in hydroids. Amer. Jour. Physiol., 8. —
May, R. M. 1927. Modifications des centres nerveux dues a la transplantation de I'ceil
et de I'organe olfactif chez les embryons d'Anoures. Arch. Biol., 37. 1933. Reactions
neurogeniques de la mcelle a la greffe en surnombre ou a I'ablation d'une ebauche de
patte posterieure chez I'embryon de I'Anoure. Bull. biol. France Belg., 67. May,
R. M., and Frank, AHce. 1936. Replacement of the anterior limb by a grafted medulla
in the embryo of the anuran Discoglossus, etc. Abstr. Amer. Soc. Zool., Anat. Rec, 67,
Suppl. I. — Mayer, B. Uber das Regulations- und Induktionsvermogen der halbseitigen
oberen Urmundlippe von Triton. Arch. Entw'mech., 133. — Mead, A. D. 1897. The
early development of marine annelids. Jour. MorphoL, 13. — Meisenheimer, J. 1896.
Entwicklungsgeschichte von Limax maximus. Zeitschr. wiss. Zool., 62. 1901. Ent-
wicklungsgeschichte von Dreissensia polymorpha. Ibid., 69. — Mencl, E. 1903. Ein
Fall von beiderseitiger Augenlinsenausbildung wahrend der Abwesenheit von Augen-
blasen. Arch. Entw'mech., 16. 1908. Neue Tatsachen zur Selbstdifferenzierung der
Augenlinse. Ibid., 25. — Merrell, W. D. 1900. A contribution to the life history of
Silphium. Bot. Gaz., 29. — Merton, H. 1929. Die Wirkungen verschiedener Formol-
konzentrationen auf Vorticella. Protoplasma, 8. — Metschnikoff, E. 1886. Embry-
ologische Studien an Medusen. Wien.— Meves, F, 1903. tJber oligopyrene und apy-
rene Spermien und liber ihre Entstehung etc. Arch. mikr. Anat., 61. — Meyer, R. 1913.
Die ursachlichen Beziehungen zwischen dem Situs viscerum und Situs cordis. Arch.
Entw'mech., 37. — Miller, Faith. 1937. Some effects of strychnine on reconstitution in
Euplanaria dorotoccphala. Physiol. Zool., 10. — Miller, J. A. 1937. Some effects of
oxygen on polarity in Tubularia. Abstr. Marine Biol. Lab., Biol. Bull., 73. 1938.
Studies on heteroplastic transplantation in triclads. Physiol. Zool., 11. 1939. Experi-
ments on polarity determination in Tubularia regenerates. Abstr. Amer. Soc. Zool.,
Anat. Rec, 75, 4, Suppl. — Milojevic, B. D. 1924. Beitrage zur Frage iiber die Deter-
mination der Regenerate. Arch. mikr. Anat. Entw'mech., 103. Milojevic, B. D., et
Burian, H. 1926. Regeneration heteropolaire de la queue chez les Tritons adultes.
C. R. Soc. Biol., 95. Milojevic, B. D,, et Grbic, N. 1925. La regeneration et I'inversion
de la polarite des extremites chez les Tritons adultes a la suite d'une transplantation
heterotope. C. R. Soc. Biol., 93. — Monod, J. 1933. Mise en evidence du gradient axial
chez les infusoires cilies par photolyse a I'aide des rayons ultravioltes. C. R. Acad. Sci.
Paris, 196. — Monroy, A. 1939. Ricerche sulla capacita lentogena dell'iride degh An-

BIBLIOGRAPHY 779

fibi. Arch. Entw'mech., 139. — Montalenti, G. 1933. L'ontogenesi degli ibridi fra
Bufo vulgaris e Bufo viridis. Physiol. Zo5l., 6. Montalenti, G., e Maccagno, A. 1935.
Analisi della potenza dei primi blastomeri dell'uovo di Lampreda, Lampetra (pctromy-
zon) flimatilis. Arch. ital. Anat. embry., 35. — Moore, A. R. 1923. Galvanotropism in
the earthworm. Jour. Gen. Physiol., 5. Moore, A. R., and Kellogg, F. M. 1918, Note
on the galvanotropic response of the earthworm. Biol. Bull., 30. Moore, A. R., and
Miller, W. A. 1937. Occurrence of birefringence in the fertilized egg of the sea urchin.
Proc. Soc. Exp. Biol. Med., 36. — Moretti, G. 191 1. Sulla trasposizione delle varie
parti del corpo nella Planaria torva (Miiller). Arch. ital. Anat. embry., 10. — Morgan,
L. V. 1905. Incomplete anterior regeneration in the absence of the brain in Leptoplana
littoralis. Biol. Bull., 9. 1906. Regeneration of grafted pieces of planarians. Jour.
Exp. Zool.,3. — Morgan, T. H. 1894. Experimental studies on echinoderm eggs. Anat.
Anz., 9. 1895c. The formation of the fish embryo. Jour. Morphol., 10. 1895&. Half-
embryos and whole embryos from one of the first two blastomeres of the frog's egg.
Anat. Anz., 10. 1895c. The formation of one embryo from two blastulae. Arch.
Entw'mech., 2. 18956?. A study of variation in cleavage. Ibid. 1897. Regeneration in
Allolobophora foetida. Ibid., 5. 1898. Experimental studies of the regeneration of
Planaria maadata. Ibid.,"]. 1900. Regeneration in planarians. Ibid., 10. 1901a. The
factors that determine regeneration in Antennularia. Biol. Bull., 2. 1901b. Regenera-
tion in r«ii//arm. Arch. Entw'mech., II. 1901c. Growth and regeneration in P/a«or/a
lugubris. Ibid., 13. 1901c?. Regeneration of proportionate structures in Stentor. Biol.
Bull., 2. 19016. Regeneration. New York. 1902a. Further experiments on the regen-
eration of Tubidaria. Arch. Entw'mech., 11. 1902b. Experimental studies of the in-
ternal factors of regeneration in the earthworm. Ibid., 14. 1903a. Some factors in the
regeneration of Tubularia. Ibid., 16. 1903b. The relation between normal and abnor-
mal development of the embryo of the frog, etc. Ibid. 1904a. Regeneration of hetero-
morphic tails in posterior pieces of Planaria simplicissima. Jour. Exp. Zool., i. 1904b.
The control of heteromorphosis in Planaria maculata. Arch. Entw'mech., 17. 1904c.
Notes on regeneration. Biol. Bull, 6. 1905. "Polarity" considered as a phenomenon
of gradation of materials. Jour. Exp. Zool., 2. 1906a. The physiology of regeneration.
Ibid., 3. 1906b. Hydranth formation and polarity in Tubularia. Ibid. 1906c. The
influence of a strong centrifugal force on the frog's egg. Arch. Entw'mech., 22. 1908.
Some further records concerning the physiology of regeneration in Tubidaria. Biol.
Bull., 14. 1910a. The effects of altering the position of the cleavage planes in eggs with
precocious specification. Arch. Entw'mech., 29. 1910b. Cytological studies of centri-
fuged eggs. Jour. Exp. Zool., 9. 1919. The physical basis of heredity. Philadel-
phia and London. 1924. The artificial induction of symmetrical claws in male
fiddler crabs. Amer. Naturalist, 58. 1927. Experimental embryology. New York.
1933. The formation of the antipolar lobe in Ilyanassa. Jour. Exp. Zool., 64. 1935.
Centrifuging the eggs of Ilyanassa in reverse. Biol. Bull., 68. 1936. Further experi-
ments on the formation of the antipolar lobe of Ilyanassa. Jour. Exp. Zool., 74. 1938.
A reconsideration of the evidence concerning a dorsoventral preorganization of the egg
of Chaetopterus. Biol. Bull., 74. Morgan, T. H., and Dimon, A. C. 1904. An examina-
tion of the problem of physiological "polarity" and electrical polarity in the earth-
worm. Jour. Exp. Zool., I. Morgan, T. H., and Lyon, E. P. 1907. The relation of the
substances of the egg separated by a strong centrifugal force to the location of the em-

78o PATTERNS AND PROBLEMS OF DEVELOPMENT

bryo. Arch. Entw'mech., 24. Morgan, T. H., and Spooner, G. B. 1909. The polarity
of the centrifuged egg. Ibid., 28. Morgan, T. H., and Stevens, N. M. 1904. Experi-
ments on polarity in Tubularia. Jour. Exp. Zool., i. Morgan, T. H.; Sturtevant, A. H.;
Muller, H. J.; and Bridges, C. B. 1923. The mechanism of Mendelian heredity. Rev.
ed. New York. Morgan, T. H., and Tyler, A. 1935. Effects of centrifuging eggs of
Urechis caupo before and after fertihzation. Jour. Exp. Zool., 70. 1938. The relation
between entrance point of the spermatozoon and bilaterality of the egg of Chaetopterus.
Biol. Bull., 74. — Morgulis, S. 1907. Observations and experiments on regeneration in
Lumbriciilus. Jour. Exp. Zool, 4. 1909. Contributions to the physiology of regenera-
tion. L Ibid.,']. 1912. Beitrage zur Regenerationsphysiologie. VI. Arch. ges. Physi-
ol., 143. — Moroff, T. 1908. Die bei den Cephalopoden vorkommenden Aggregatasiritn
als Grundlage einer kritischen Studie iiber die Physiologie des Zellkerns. Arch. Protis-
tenk., II. — Mortensen, T. 191 3. Regeneration in ctenophores. Vidensk. Meddel.
Dansk. Naturh. Foren. 66. — Motomura, I. 193 1. Notes on the effect of centrifugal
force on the frog's egg. Sci. Repts. Tohoku Imp. Univ., Ser. 4, Biol., 6. 1934. Uber die
Exogastrulabildung beim Seeigelkeim durch Auxin, Glycogen und KCIO3. Ibid., 9. —
Mottier, D. M. 1904. The development of the spermatozoid in Chara. Ann. Bot., 18.
— Muller, C. 1908. Regenerationsversuche an Lumbriculus variegatus und Tubifex
rivulorum. Arch. Entw'mech., 26. — Muller, E. 1896. tJber die Regeneration der Aug-
enhnse nach Exstirpation derselben bei Triton. Arch. Entw'mech., 47. — Muller, H.
1903. Beitrag zur Embryonalentwicklung der Ascaris megalocephala. Zoologica, H.
17. — Muller, K. 1911. Das Regenerationsvermogen der Siisswasserschwamme etc.
Arch. Entw'mech., 32. — Miiller-Cale, C. 1913. tJber die Entwicklung von Cypris in-
congruens. Zool. Jahrb. Abt. Anat. Ont., 36. — Munson, J. P. 1898. The ovarian egg
of Limulus. Jour. Morphol., 15. — Murray, Margaret. 1927. The cultivation of pla-
narian tissues //; D/Zro. Jour. Exp. Zool., 47. 1931. /« fZ/ro studies of planarian paren-
chyma. Arch. exp. Zellforsch., 11. — Murray, P. D. F., and Selby, Doris. 1930.
Chorio-allantoic grafts of entire and fragmented blastoderms of the chick. Jour. Exp.
Biol., 7.

Nawaschin, S. 1900. Uber die Befruchtungsvorgange bei einigen Dicotyledoneen.
Ber. deutsch. bot. Ges., 18. — Needham, J. 1931. Chemical embryology. New York.
Needham, J., and Boell, E. J. 1938. Metabolic properties of the regions of the am-
phibian gastrula. Proc. Soc. Exp. Biol. Med., 39. Needham, J.; Rogers, V.; and
Shen, Shih -Chang. 1939. Morphogenesis and metabohsm, etc. V. Proc. Roy. Soc, B,
127. Needham, J.; Waddington, C.H.; and Needham, D. 1934. Physico-chemical ex-
periments on the amphibian organizer. Ibid., 12^. — Nemec, B. 1908. Einige Regener-
ationsversuche an Taraxacum-W wrzt\r\. Wiesner Festschr. Wien. — Newman, H. H.
1908. The process of heredity as exhibited by the development of Fundidus hybrids.
Jour. Exp. Zool, 5. 1914. Modes of inheritance in teleost hybrids. Ibid., 16. 1915a.
Development and heredity in heterogenic teleost hybrids. Ibid., 18. 1915b. Heredity
and organic symmetry in armadillo quadruplets. I. Biol. Bull., 29. 1916. Heredity
and organic symmetry, etc. II. Ibid., 30. 19170. The biology of twins. Chicago.
1917b. On the production of monsters by hybridization. Biol. Bull., 32. 1918. Hy-
brids between Fwj(^«/?<5 and mackerel. Jour. Exp. Zool., 26. 1921a. The experimental
production of twins and double monsters in larvae of the starfish, Patiria miniata, etc.
Ibid., :i,T). 1921b. On the occurrence of paired madreporic pores and pore canals in the

BIBLIOGRAPHY 781

advanced bipinnaria of .l5/<';-/;/a (P<j//>/a) ;«/;?/a/(7, etc. Biol. Bull., 40. 1922. Normal
versus subnormal development in Patiria miniata. Ibid., 43. 1923. The physiology of
twinning. Chicago. 1925. An experimental analysis of asymmetry in the starfish,
Patiria miniata. Ibid., 49. Newman, H. H., and Patterson, J. T. 1909. A case of nor-
mal identical quadruplets in the nine-banded armadillo, etc. Ibid., 17. 1910. De-
velopment of the nine-banded armadillo from the primitive streak stage to birth: with
special reference to the question of specific polyembryony. Jour. MorphoL, 21. 1911.
The Hmits of hereditary control in armadillo quadruplets, etc. Ibid., 22. — Newport, G.;
Ellis, G. v.; and Forbes, J. 1854. Researches on the impregnation of the ovum in the
Amphibia and the early stages of development of the embryo. Phil. Trans. Roy. Soc,
B, 144. — Nicholas, J. S. 1924. Regulation of posture in the fore limb of Amblystoma
pundatum. Jour. Exp. Zool., 40. 1925. The response of the developing limb of Am-
blystoma punctatum to variations in the orientation of the surrounding tissue. Abstr.
Amer. Soc. Zool., Anat. Rec, 29. 1926. The reactions of the developing Hmb girdle in
response to rotation of embryonic rudiments in Amblystoma punctatum. Abstr. Amer.
Assoc. Anat., Anat. Rec, 32. Nicholas, J. S., and Rudnick, Dorothea. i933- The de-
velopment of embryonic rat tissue upon the chick chorio-allantois. Jour. Exp. Zool.,
66. — Noll, F. 1900. liber die Umkehrungsversuche mit Bryopsis. Ber. deutsch. bot.
Ges., 18. — Novikoff, A. 1938. Embryonic determination in the annelid, Sabellaria
vulgaris. I, II. Biol. Bull., 74. 1939. Regulation in mosaic eggs. Abstr. Marine Biol.
Lab., Ibid., 77. — Nusbaum, J. 1912. Die entwicklungsmechanisch-metaplastischen
Potenzen der tierischen Gewebe. Vortr. u. Aufs. ii. Entw'mech., H. 17. Nusbaum, J.
und Oxner, M. 1910. Studien iiber die Regeneration der Nemertinen. I. Arch. Ent-
w'mech., 30. 1911. Weitere Studien etc. Ibid., 7,2. 1912. Fortgesetzte Studien etc.
Ibid., 35. 1913. Die Diovogonie oder die Entwicklung eines Embryo aus zwei Eiern
bei der Nemertine Lineus ruber Miill. Ibid., 36. Nusbaum-Hilarowicz, J. 1918.
liber das Verhalten des Chondrioms wahrend der Eibildung bei Dytiscus marginalis
L. Zeitschr. wiss. Zool., 117. — Nuzum, M., and Rand, H. W. 1924. Can the earth-
worm pharynx epithelium produce nervous tissue? Biol. Bull., 47.

Oellacher, J. 1872. Beitrage zur Entwicklungsgeschichte der Knochenfische etc.
Zeitschr. wiss. Zool., 22. — Ogawa, C. 1921. Experiments on the regeneration of the
Itns'm Diemydylus. Jour. Exp. Zool., 37. — Oka, H. 1934. Zur Analyse experimentell
erzeugter Doppelbildungen der Extremitat etc. Jour. Fac. Sci. Tokyo Imp. Univ., 3.
— Okada, T. 1929. Respiration of Branc/iiura. Dobutsugaku-Zasshi, 38. (In Japa-
nese.)— Okada, Y. K. 1934. Regeneration de la tete chez le polychete Myxicola aes-
thetica (Clap.) . Bull. biol. France Belg., 68. 1935. Some notes on Musculium heterodon
(Pilsbry), a fresh-water bivalve. Sci. Repts. Tohoku Imp. Univ., Ser. 4, Biol., 9. 1938.
Neural induction by inorganic matters, etc. Mem. Coll. Sci. Kyoto Imp. Univ., Ser. B,
14. Okada, Y. K., and Mikami, Y. 1937. Inductive effect of tissues other than retina
on the presumptive lens-epithelium. Proc. Imp. Acad. Tokyo, 13. Okada, Y. K., and
Sugino, H. 1934. Transplantation experiments in Planaria gonocephala. I, II. Ibid.,
10. 1937. Transplantation experiments, etc. Jap. Jour. Zool., 7. — Olive, E. W. 1902.
A monograph of the Acrasieae. Proc. Boston Soc. Nat. Hist., 30. — Olmsted, J. M. D.
1918. The regeneration of triangular pieces of Planaria maculata. Jour. Exp. Zool.,
25. — Olson, R. A., and Du Buy, H. G. 1937. The role of growth substance in the polar-
ity and morphogenesis of Fucns. Amer. Jour. Bot., 24. — Oppenheimer, Jane M. 1934.

782 PATTERNS AND PROBLEMS OF DEVELOPMENT

Experimental studies on the developing perch {Pcrca Jlavescens Mitchell). Proc. Soc.
Exp. Biol. Med., 31. 1936a. Structures developed in amphibians by implantation of
living fish organizer. Ibid., 34. 1936b. The development of isolated blastoderms of
Fundulus heteroclihis. Jour. Exp. Zool., 72. 1936c. Transplantation experiments on
developing teleosts {Fundulus and Perca). Ibid. ig36d. Processes of localization in
developing FiDididus. Ibid., 73. 1938. Potencies for differentiation in the teleostean
germ ring. Ibid., jg. — Opperman, K. 1913. Die Entwicklung von Forelleneiern nach
Befruchtung mit radiumbestrahlten Samenfaden. Arch. mikr. Anat., 83. — Oshima, H.
1921. On the development of Cucumaria echinata. Quart. Jour. Micr. Sci., 65. —
Otto, F. 1921. Studien iiber das Regulationsvermogen einiger Siisswasserbryozoen.
Arch. Entw'mech., 47.

Paff, G. H. 1936. Transplantation of sino-atrium to conus in the embryonic heart
in vitro. Amer. Jour. Physiol., 117. — Painter, B. T. 1938. Grafting experiments di-
rected toward ascertaining the location of factors of head regeneration in the earth-
worm. Abstr. Amer. Soc. Zool., Anat. Rec, 72, Suppl. 4. — Painter, T. S. 1915. The
effects of carbon dioxide on the eggs of Ascaris. Jour. Exp. Zool., 19. — Parke, O.
1929. The differential reduction of osmic acid in the cortex of Paramecium and its
bearing on the metabolic gradient conception. Physiol. Zool., 2. — Parker, G. H. 1905.
The movements of the swimming plates in ctenophores with reference to theories of
ciliary metachronism. Jour. Exp. Zool., 2. 1925. Production of carbon dioxide by
nerve. Jour. Gen. Physiol., 6. 1929. The metabolic gradient and its appHcations.
Brit. Jour. Exp. Biol., 6. — Parseval, M. von. 1922. Die Entwicklung zentrifugierter
Eier von Tubifex rivulorum Lam. Arch. Entw'mech., 50. — Pasquini, P. 1927. Tra-
pianti omeoplastici degli abozzi oculari negli embrioni di Plcurodcks waltli. Atti
Accad. naz. Lincei, ser. 6, 5. 1931. Sul differenziamento correlative della lente cristal-
lina e della cornea nello sviluppo di Anuri ed Urodeli. Rend. Accad. naz. Lincei, 14.
1932. SuUa determinazione e sul differenziamento del cristallino in Rana catesbiana
(Shaw). Jour. Exp. Zool, 61. 1933. L'eteromorfosi sperimentale nelle Ascidie (Clave-
liiia lepadiformis O. F. MuUer). Arch. Zool. ital., 18. — Pasteels, J. 1932. Etude des
localisations germinales de I'oeuf insegmente des Amphibiens anoures. Arch. Biol., 43.
1936a. Etudes sur la gastrulation des Vertebres meroblastiques. I. Ibid., 47. 19366.
Analyse des mouvements morphogenetiques de gastrulation chez les oiseaux. Bull.
Acad. roy. Belg., CI. des sci., ser. 5, 22. 1936c. Centre organisateur et glycogenolyse.
Arch. Anat. micr., 32. 1937a. Etudes sur la gastrulation des Vertebres meroblastiques.
II, III, IV. Arch. Biol., 48. 1937b. Sur I'origine de la symetrie bilaterale des Amphib-
iens anoures. Arch. Anat. micr., t,2>- 1938. Resultats experimentaux concernant le de-
terminisme precoce chez les Amphibiens. C. R. Assoc. Anat., 1938. 1939. Recherches
sur les facteurs initiaux de la morphogenese chez les Amphibiens anoures. II. Arch.
Biol., 50. — Patterson, J. T. 1909. Gastrulation in the pigeon egg. Jour. Morphol., 20.
191 2. A preliminary report on the demonstration of polyembryonic development of the
armadillo. Anat. Anz., 41. 1913. Polyembryonic development in r^/w^/a ;?oi)ewza»c/a.
Jour. Morphol., 24. 1915. Observations on the development of Copidosoma gelechiae.
Biol. Bull., 29. 1921. The development of Paracopidosomopsis. Jour. Morphol., 36. —
Paulcke, W. 1901. Uber die Differenzierung der Zellelemente im Ovarium des Bienen-
konigin. Zool. Jahrb. Abt. Anat. Ont., 14. — Pauli, Margarete. 1927. Die Entwicklung
geschniirter und centrifugierter Eier von Calliphora erythrocephala und Musca domes-

BIBLIOGRAPHY 783

tica. Zeitschr. wiss. Zool., 129. — Pease, D. C. 1938. The influence of centrifugal force
on the bilateral determination of the spirally cleaving egg of Urechis. Biol. Bull., 75.
1939. An analysis of the factors of bilateral determination in centrifuged echinoderm
embryos. Jour. Exp. Zool., 80. 1940. The effects of hydrostatic pressure upon the
polar lobe and cleavage pattern in the Chaetopterus egg. Biol. Bull., 78. — Peebles,
Florence. 1897. Experimental studies on Hydra. Arch. Entw'mech., 5. 1898. Some
experiments on the primitive streak of the chick. Ibid., 7. 1900. Experiments in re-
generation and in grafting of Hydrozoa. Ibid., 10. 1904. The location of the chick em-
bryo upon the blastoderm. Jour. Exp. Zool., i. 1912. Regeneration and regulation in
Paramecium. Biol. Bull., 23. — Penners, A. 1922. Die Furchung von Tubifex rivu-
lorum Lam. Zool. Jahrb. Abt. Anat. Ont., 43. 1923. Die Entwicklung des Keimstreifs
\ind (X\t OrgSinh\\d\xnghQ\ Tiibifex rivulorum'Lsim. Ibid., 4s. 1924a. Doppelbildungen
bei Tubifex rivulorum Lam. Zool. Jahrb. Abt. allg. Zool, 41. 19246. Experimentelle
Untersuchungen zum Determinationsproblem am Keim von Tubifex rivulorum Lam. I.
Arch. mikr. Anat. Entw'mech., 102. 1926. Experimentelle Untersuchungen. II.
Zeitschr. wiss. Zool., 127. 1929. Schultzesches Umdrehungsversuch an ungefurchten
Froscheiern. Arch. Entw'mech., 116. Penners, A., und Schleip, W. 1928. Die Ent-
wicklung der Schultzeschen Doppelbildungen aus dem Ei von Rana fusca. I-VI.
Zeitschr. wiss. Zool., 130, 131. — Perkins, M. 1929. Growth gradients and the axial
relations of the animal body. Nature, 124.— Perri, T. 1933. Espianto dell'abozzo
dell'occhio di Anfibi anuri. Ricerca scientif., II. — Petersen, H. 1924. Berichte iiber
Entwicklungsmechanik. I. Ergebn. Anat. Entw'ges., 24. — Petersen, Walburga. 1927.
The axial gradient in ParawmMm. Science, 66. — Pfeffer, W. 1901. Pflanzenphysio-
logie, Bd. II. Leipzig. — Pfeiffer, Wanda. 1907. Differentiation of sporocarps in Azol-
la. Bot. Gaz., 44. — Pickford, Grace. 1930. The distribution of pigment and other
morphological concomitants of the metabolic gradient in oligochetes. Biol. Bull., 58. —
Piepho, H. 1938. tjber Oxidations-Reduktionsvorgange im Amphibienkeim. Biol.
ZbL, 58. — Pitotti, M. 1939. Catahzzatori e determinazione negli embrioni degli echino-
dermi. Pubbl. Staz. zool. Napoli, 17.— Pitts, R. F. 193 1. The effects of cyanide on the
respiration of Colpidium campylum. Abstr. x\mer. Soc. Zool., Anat. Rec, 51, Suppl. i.
— Plagge, E. 1938. Weitere Untersuchungen iiber das Verpuppungshormon bei
Schmetterlingen. Biol. ZbL, 58.— Plough, H. H. 1927. Defective pluteus larvae from
isolated blastomeres of Arbacia and Echinarachnius. Biol. Bull., 52. — Politzer, G.
1936. Zur Kausalanalyse der Linsenregeneration. Arch. Entw'mech., 135. 1937. Zur
Frage der Induktion von Linse durch Pigmentepithel. Ibid. — Ponomarewa, W. N.
1936. tJber die Dauer der Induktionswirkung bei der Bildung der Horblase. Bull.
Biol. Med. exp., 2.— Potts, G. 1902. Zur Physiologie des Dictyostelium mucoroides.
Flora, 91. — Pourbaix, Nelly. 1931. La polarite dans la regeneration des Planaires.
Ann. Soc. roy. zool. Belg. 62.— Pratje, A. 1921. Noctiluca miliaris Sar. Beitriige zur
Morphologie, Physiologie und Cytologie. I. Arch. Protistenk., 42.— Pressler, K. 191 1.
Beobachtungen und Versuche iiber normalen und inversen Situs viscerum et cordis bei
Anurenlarven. Arch. Entw'mech., 32.— Preston, R. D., and Astbury, W. T. 1937. The
structure of the wall of the green alga Valonia ventricosa. Proc. Roy. Soc, B, 122.—
Prielgauskiene, A. 1933. Reservezellen bei der Planarienregeneration. Kosmos, 14.
— Pringsheim, N. 1858. Beitrage zur Morphologie und Systematik der Algen. I.
Jahrb. wiss. Bot., i.— Probst, G. 1931. Beitrage zur Regeneration der Anneliden. I.

784 PATTERNS AND PROBLEMS OF DEVELOPMENT

Arch. Entw'mech., 124. 1932. Studien iiber die Regeneration. IL Ihid., 127. — Przi-
bram, H. 1902. Experimentelle Studien iiber Regeneration. IL Arch. Entw'mech., 13.
1905. Die "Heterochelie" bei decapoden Crustaceen. Ibid., ig. 1906. Kristallanalo-
gien zur Entwicklungsmechanik der Organismen. Ibid., 22. 1907. Die "Scherenum-
kehr" bei decapoden Crustaceen. Ibid., 25. 1908. Versuche an den Scheren der Win-
kerkrabbe {Gelasimus). Zbl. Physiol., 22. 1910. Die Homoeosis bei Arthropoden.
Arch. Entw'mech., 29. 1917a. Fiihlerregeneration halberwachsener Sphodromantis-
Larven. Ibid., 43. 1917b. Transitare Scherenformen der Winkerkrabbe, Gelasimus
pugnax Smith. Ibid. 1919a. Fangbeine als Regenerate. Ibid., 4$. 1919&. Fussglieder
an Kaferfuhlern. Regeneration beim Hautflugler, Cimbex axillaris Panz. Ibid. 1920.
Teratologic und Teratogenie. Vortr. u. Aufs. ii. Entw'mech., H. 25. 1921. Die Bruch-
Dreifachbildung im Tierreiche. Arch. Entw'mech., 48. 1924. Achsenverhaltnisse und
Entwicklungspotenzen der Urodelenextremita-ten an ModeUen zu Harrison's Trans-
plantationsversuchen. Arch. mikr. Anat. Entw'mech., 102. 1925. Berichtigung zur
Abhandlung "Achsenverhaltnisse" etc. Ibid., 104. 1927. Deutung spiegelbildlicher
Lurchenarme. Arch. Entw'mech., log.

Rebaud, E. 1901-2. Recherches embryologiques sur les Cyclocephaliens. Jour.
Anat. Physiol., 27, 28. 1908. La position et I'orientation de I'embryon de poule sur le
jaune. Arch. Zoo), exp., ser. 4, 9. — Rabl, C. 1879. tjber die Entwicklung der TeUer-
schnecke. Morphol. Jahrb., 5. 1885. tJber Zellteilung. Ibid., 10. 1898. tJber den Bau
und die Entwicklung der Linse. Zeitschr. wiss. Zool., 63. 1900. Homologie und Eigen-
art. Verb, deutsch. zool. Ges., 1900. — Ragosina, M. N. 1936. Die Induktionswirkung
pflanzlicher Gewebe auf das Ektoderm der Gastrula. Bull. Biol. Med. exp., 2. 1937.
Die Induktionswirkung pflanzhcher Gewebe etc. Arch. Entw'mech., 112. — Rand,
H. W. 191 1. The problem of form in //y(/ra. Science, 33. 1912. The problem of or-
ganization. Ibid., 36. Rand, H. W., and Boyden, E. C. 1913. Inequahty of the two
eyes in regenerating planarians. Zool. Jahrb. Abt. allg. Zool., 34. — Ranzi, S. 1926.
Ricerchi di morfologia sperimentale nei cefalopodi. Boll. Soc. ital. Biol, sper., i.
1927. Inibizione difTerenziale nello sviluppo dei cefalopodi e considerazioni sul cosi
detto gradiente assiale. Rend. Accad. naz. Lincei, CI. sci., ser. 6a, 6. 1928. Suscetti-
bilita differenziale nello sviluppo dei cefalopodi. Pubbl. Staz. zool. Napoli., 9. 1929a.
Contributo al problema della cyclopia. Atti pont. Accad. Sci. nuovi Lincei, 82. 1929b.
Embriogenesie gradienti assiali. Mem. pont. Accad. Sci. nuovi Lincei, 12. 1931. I
gradienti assiali nello sviluppo embrionale. Boll. Soc. ital. Biol, sper., 6. 1932. Esistono
i gradienti fisiologici di Child? Scientia, 1932. 1938. La determinazione nell embrione
degh echinodermi. Boll. Soc. ital. Biol, sper., 13. Ranzi, S., e Falkenheim, Margarete.
1937. Ricerche sulle basi fisiologiche della determinazione neU'embrione degli echino-
dermi. Pubbl. Staz. zool. Napoh, 16. 1938. Lithiumwirkung und Determination des
Seeigelkeims. Naturwiss., 26. — Ratcliffe, H. L. 1927. Mitosis and cell division in
Euglena spirogyra Ehrenberg. Biol. Bull., 53. — Raven, C. P. 1931. Die Induktions-
fahigkeit des Ganglienleistenmaterials von Rana fusca. Proc. kon. Akad. Wetensch.
Amsterdam, 34. 1933a. Experimentelle Untersuchungen iiber den GlykogenstofT-
wechsel des Organisationszentrums in der Amphibiengastrula. I. Ibid., 36. 1933b.
Zur Entwicklung der Ganglienleiste. III. Arch. Entw'mech., 130. 1935a. Experimen-
telle Untersuchungen iiber den GlykogenstotTwechsel etc. II. Proc. kon. Akad.
Wetensch. Amsterdam, 38. 1935b. Uber assimilatorische Induktion in der dorsalen

BIBLIOGRAPHY 785

Urmundlippe der Amphibiengastrula. Ibid. — Rawles, Mary, 1936. A study in the
localization of organ-forming areas in the chick blastoderm of the head process stage.
Jour. Exp. Zool., 72. — Reed, H. S. 1928, The growth of cognate shoots. Amer.
Naturalist, 62. — Reese, R. M. 1912. The effect of narcotics upon the development of
the hen's egg. Science, 35. — Regel, F. 1876. Die Vermehrung der Begoniaceen aus
ihren Blattern. Jen. Zeitschr., 10 — Reith, F. 1925. Die Entwicklung des MM5ca-Eies
nach Ausschaltung verschiedener Eibereiche. Zeitschr. wiss. Zool., 126. 1931. Ver-
suche iiber die Determination der Keimesanlage bei Camponotus ligniperda. Ibid., 139.
1932. tjber die Lokalisation der Entwicklungsfaktoren im Insektenkeim. I. Arch.
Entw'mech., 127. 1935. tJber die Determination der Keimesanlage bei Insekten.
Zeitschr. wiss. Zool., 147. — Retzius, G. 1881. Zur Kenntnis der Spermatozoen. Biol.
Untersuch., 1881. — Reverberi, G. 1929. Risultati di esperimenti di asportazione par-
tiale e totale della vescicola ottica nel embrione di polio. Boll. 1st. zool. Univ. Roma, 7.
1931. Studi sperimentali suU'uovo di ascidie. I. Pubbl. Staz. zool. Napoli, 11. 1936.
La segmentazione dei frammenti dell'uovo non fecondato di ascidie. Ibid., 15. 1937.
Ricerche sperimentali suUa struttura dell'uovo fecondato delle ascidie. Commenta-
tiones, i. Reverberi, G., e Pitotti, M. 1939. Differenziazioni fisiologiche nell'uovo
delle ascidie. Commentationes, 3. — Rhumbler, L. 1897. Stemmen die Strahlen der
Astrosphare oder ziehen sie? Arch. Entw'mech., 4. 1899. Die Furchung des Cteno-
phoreneies nach Ziegler und deren Mechanik. Ibid., 8. — Riddle, O. 1923. On the
cause of twinning and abnormal development in birds. Amer. Jour. Anat., 32. — Ries,
E. 1936. Zelldifferenzierungen und vitale Farbung. Verb, deutsch. zool. Ges., 1936.
1937. Die Verteilung von Vitamin C, Glutathione, Benzidinperoxidase, Phenolase
(Indophenolblauoxidase) und Leucolethyleneblauoxidoredukatse wahrend der friihen
Embryonalentwicklung verschiedener wirbelloser Tiere. Pubbl. Staz. zool. Napoli, 16.
1939. Versuche iiber die Bedeutung des Substanzmosaiks fiir die embryonale Gewebe-
differenzierung bei Ascidien. Arch. exp. Zellforsch., 23. Ries, E., und Gersch, M.
1936. Die Zelldifferenzierung und Zellspezialisierung wahrend der Embryonalent-
wicklung von Aplysia limacina L. Pubbl. Staz. zool. Napoli, 15. Ries, E., und Schol-
zel, G. 1934. Beitrage zu Problemen der vitalen Farbung. L Zeitschr. Zellforsch., 20.
— Ritter, W. E. 1919. The unity of the organism. Boston. Ritter, W. E., and John-
son, Myrtle. 1911. The growth and differentiation of the chain of Cyclosalpa ajfinis
Chamisso. Jour. Morphol., 22. — Robbins, Harriet, and Child, C. M. 1920. Carbon
dioxide production in relation to regeneration in Plnnaria dorotocephala. Biol. Bull.,
38. — Robert, A. 1903. Recherches sur le developpement des Troques. Arch. Zool.
exp., 10. — Robertson, Alice. 1903. Embryology and embryonic fission in the genus
Crisia. Univ. Calif. Publ. Zool., i. — Rose, S. M. 1939. Embryonic induction in the
ascidia. Biol. Bull., 77. — Rosene, Hilda. 1934. Dependence of continuous bioelectric
currents upon cell oxidation. Proc. Soc. Exp. Biol. Med., 31. Rosene, Hilda, and Lund,
E. J. 1930. Evidence from the effects of KNC for linkage between polar growth, elec-
tric potentials and cell oxidation in Obelia. Publ. Puget Sound Biol. Sta., 7. 1935.
Linkage between output of electric energy by polar tissues and cell oxidation. Plant
Physiol., 10. — Roskin, G., und Semenoff, W. 1933. Studium der Oxido-Reduktions-
prozesse der Zelle. Zeitschr. Zellforsch., 19. — Rotmann, E. 1931. Die RoUe des Ekto-
derms und Mesoderms bei der Formbildung der Kiemen und Extremitiiten von Triton.
Arch. Entw'mech., 124. 1934. Heteroplastische Austausch einiger induktiv entstehen-

786 PATTERNS AND PROBLEMS OF DEVELOPMENT

der Organanlagen. Ibid., 131. 1935a. Reiz und Reizbeantwortung in der Amphibien-
entwicklung. \'erh. deutsch. zool. Ges., 1935. 1935b. Der Anteil von Induktor und
reagierendem Gewebe an der Entwicklung des Haftfadens. Arch. Entw'mech., 133.
1935c. Der Anteil von Induktor und reagierendem Gewebe an der Entwicklung der
Kiemen und ihrer Gefasse. Ibid. 1939. Der Anteil von Induktor und reagierendem
Gewebe an der Entwicklung der Amphibienlinse. Ibid., 139. — Roux, W. 1885. Bei-
trage zur Entwicklungsmechanik des Embryo. Einleitung. Zeitschr. Biol., 21. 1887.
Beitrage zur Entwicklungsmechanik des Embryo. IV. Arch. mikr. Anat., 29. 1888.
Beitrage etc. V. Virchow's Arch., 114. — Rudnick, Dorothea. 1938a. Differentiation
in culture of pieces of the early chick blastoderm. Anat. Rec, 70. 1938b. Differentia-
tion, etc. Jour. Exp. Zool., 79. — Riickert, J. 1892. Die Entwicklungsgeschichte des
Ovarialeies bei Selachiern. Anat. Anz., 7. — Rulon, O. 1935. Differential reduction of
Janus green during development of the chick. Protoplasma, 24. 1936a. The effects of
carbon dioxide, the hydrogen ion, calcium and experimental conditioning on reconsti-
tution in Euplanaria dorotocephala. Physiol. Zool., 9. 1936b. Experimental asym-
metries of the head of Euplanaria dorotocephala. Ibid. 1937. The effects of certain
organic acids on reconstitution in £?//>/a;zaria Jo^otoce/'/m/a. Ibid., 10. 1938. Modifica-
tion of development in the sand dollar with NaCNS and Ca-free sea water. Abstr.
Amer. Soc. Zool., Anat. Rec, 72, Suppl. 4.' Rulon, O., and Child, C. M. 1937a. Ob-
servations and experiments on developmental pattern in Pclmatohydra oligactis. Physi-
ol. Zool., ID. 1937b. Experiments on scale of organization. Ibid. — Runnstrom, J.
1914. Analytische Studien iiber Seeigelentwicklung. I. Arch. Entw'mech., 40. 1915.
Analytische Studien etc. II. Ibid., 41. 1917. Analytische Studien. III. Ibid., 43.
1920. Entwicklungsmechanische Studien an Henricia sanguinolenta Forbes und So-
laster. sp. Ibid., 46. 1925a. Uber den Einfluss des Kaliummangels auf das Seeigelei.
Pubbl. Staz. zool. NapoH, 6. 1925b. Regulatorische Bildung von Colomanlagen bei
Seeigelkeimen mit gehemmter Urdarmbildung. Arch. Entw'mech., 105. 1925c. Zur
experimentellen Beeinflussung der Asymmetric bei dem Seeigelkeim. Ibid., 17B. 1926a.
ExperimenteUe Bestimmung der Dorso-Ventralachse bei dem Seeigelkeim. Ark. Zool.,
18A. 1926b. tjber die \'erteilung der Potenzen der Urdarmbildung bei dem Seeigel-
keim. Acta, zool., 7. 1928a. Plasmabau und Determination bei dem Ei von Paraccn-
trotiis lividns Lk. Arch. Entw'mech., 113. 1928b. Zur experimentellen Analyse der
Wirkung des Lithiums auf den Seeigelkeim. Acta zool., 9. 1928c. Struktur und At-
mung bei der Entwicklungserregung des Seeigeleies. Ibid. 1928^. Die Veranderungen
der PlasmakoUoide bei der Entwicklungserregung des Seeigeleies. I. Protoplasma, 4.
1929a. Die \'eranderungen etc. II. Ibid., 5. 1929b. Uber Selbstdifferenzierung und
Induktion bei Seeigelkeim. Arch. Entw'mech., 117. 1931. Zur Entwicklungsmechanik
des Skelettmusters bei dem Seeigelkeim. Ibid. ,12^. 1933. Kurze Mitteilung zur Physi-
ologie der Determination des Seeigelkeimes. Ibid., 129. 1935a. An analysis of the ac-
tion of lithium on sea urchin development. Biol. Bull., 68. 1935b. Influence of iodo-
acetate on activation and development of the eggs of Arbacia piinctidata. Ibid., 69.
Runnstrom, J. und S. 1920. Uber die Entwicklung von Cucumaria frondosa Gunnerus
und Psolus phantapiis Strussenfelt. Bergens Mus. Aarbook, 1918-19. — Rustia, C. P.
1925. The control of biaxial development in the reconstitution of pieces of Planaria.

' Also Rulon, O. 1040. Modifications in development of the sand dollar by NaCNS. Proc.
Soc. E.xp. Biol. Med., 40.

BIBLIOGRAPHY 787

Jour. Exp. Zool., 42. — Ruttloff, C. 1908. Transplantationsversuche an Lumbriciden.
Arch. Entw'mech., 25. — Ruud, G. 1925. Die Entwicklung isolierter Keimfragmente
friihester Stadien von Triton taeniatiis. Ibid., 105. Ruud, G., und Spemann, H. 1922.
Die Entwicklung isolierter dorsaler und lateraler Gastrulahalften von Triton tacniatus
und alpestris, ihre Regulation and Postgeneration. Ibid., 52.

Sachs, J. 1887. Vorlesungen iiber Pflanzenphysiologie. Leipzig. — St. Hilaire, I. G.
1932-37. Histoire generate et particuliere des anomalies de I'organisation chez Thomme
et les animaux. Bruxelles. — Sala, L. 1895. ExperimenteUe Untersuchungen iiber die
Reifung und Befruchtung der Eier bei Ascaris. Arch. mikr. Anat., 44. — Salensky, W.
1882-83. Etude sur le developpement des AnneHdes. Arch. Biol., 3, 4. 1885. Etude
etc. II. Ibid., 6. — Samassa, P. 1893. Die Keimblatterentwicklung bei den Cladocer-
en. Arch. mikr. Anat., 41. — Santos, F. 1929. Studies on transplantation in P/a;/a/-/a.
Biol. Bull., 57. 1931. Studies, etc. Physiol. Zool., 4. — Sato, T. 1930. Beitrrge zur
Analyse der Wolffschen Linsenregeneration. I. Arch. Entw'mech., 122. 1933a. tJber
die Determination des fetalen Augenspalts bei TrZ/o;/. Ibid., 128. 1933b. Beitrage zur
Analyse etc. II. Ibid., 130. 1935. Beitrage zur Analyse etc. III. Ibid., 133. — Sayles,
L. P. 1927. Origin of the mesoderm and behavior of the nucleolus in regeneration in
Lumbriculus. Biol. Bull., 52. 1934. Regeneration in the polychaete Clymenella tor-
quata. I. Physiol. Zool., 7. 1936. Regeneration, etc. III. Biol. Bull., 70. 1937. New
structures induced by implants of adult nerve cord in the polychaete, Clymenella
torquata. Ibid., 73. 1939. Buds induced from implants of nerve cord and neighboring
tissues in the polycheate, Clymenella torquata. Ibid., 76.— Scammon, R. E., and Calk-
ins, L. A. 1929. The development and growth of the external dimensions of the human
body in the fetal period. Minneapolis.— Schaffner, J. H. 1897. Contribution to the
life history of Sagittaria variabilis. Bot. Gaz., 23.— Schaper, A. 1904. ExperimenteUe
Untersuchungen iiber die Wirkung des Radiums auf embryonale und regenerative
Entwicklungsvorgiinge. Deutsch. med. Wochenschr., 30.— Schapiro, J. 1914. tJber
die Regenerationserscheinungen verschiedener Seesterne. Arch. Entw'mech., 38.—
Schaudinn, F, 1902. Studien iiber krankheitserregende Protozoen. II. Arb. kais.
Gesundheitsamt, 19.— Schaxel, J. 1914. \'ersuch einer cytologischen x\nalysis der
Entwicklungsvorgiinge. III. Zool. Jahrb. Abt. Anat. Ont., 37. 1915. Die Leistungen
der Zellen bei der Entwicklung der Metazoen. Jena. 1921. Untersuchungen uber die
Formbildung der Tiere. Arb. Geb. exp. Biol., i. 1922. Uber die Natur der Formbild-
ungsvorgiinge in der tierischen Entwicklung. Arch. Entw'mech., 50.— Scheben, L.
1905. Beitrage zur Kenntnis des Spermatozoons von Ascaris megalocephala. Zeitschr.
wiss. Zool., 79.— Schechter, V. 1934. Electrical control of rhizoid formation in the red
alga. Griff it/isia bornctiana. Jour. Gen. Physiol., 18. 1935. The effect of centrifuging
on the polarity of an alga, Griffithsia bornetiana. Biol. Bull., 68.— Schechtman, A. M.
1938. Localization of the neural inductor and tail mesoderm in a frog egg {Hyla regil-
la). Proc. Exp. Biol. Med., 39. Schechtman, A. M., and Smale, L. E. 1939. Experi-
ments on the mechanism of gastrulation in a frog's egg. Proc. Exp. Biol. Med., 40.—
Schleip, W. 1914c. Die Furchung des Lies der Russelegel. Zool. Jahrb. Abt. Anat.
Ont., 37. 1914b. Die Entwicklung zentrifugierter Eier von f/c/).yme 5exoa//a/a. Verb,
deutsch. zool. Ges., 1914. 1924. Die Herkunft der Polaritiit des Lies von Ascaris
megalocephala. Arch. mikr. Anat. Entw'mech., 100. 1925. Die Furchung dispermer
Dentalium-Eier. Arch. Entw'mech., 106. 1929. Die Determination der Primitivent-

788 PATTERNS AND PROBLEMS OF DEVELOPMENT

wicklung. Leipzig. Schleip, W., und Penners, A. 1925. tjber die Duplicitas cruciata
bei dem O. Schultzeschen Doppelbildungen von Rana fusca. \^erh. phys.-med. Ges.
Wiirzburg, N.F., 50. 1926. Weitere Untersuchungen iiber die Entstehung der Schultz-
eschen Doppelbildungen beim braunen Frosch. Ibid., 51. — Schmidt, F. J. 1936. \'er-
gleichende Untersuchungen an Chimaren und Bastarden von Seeigelplutei. Arch.
Entw'mech., 135. — Schmidt, G. A. 1930. Schniirungsversuche bei Anuren. Arch.
Entw'mech., 122. 1931. Die Entwicklung der Palpen bei Ascidienhalblarven. Arch.
Zool. ital., 16. 1933. Schniirungs- und Durchschneidungsversuche am Anurenkeim.
Arch. Entw'mech., 29. 1936a. Uber die Unterschiede in der Induktionsfahigkeiten
des Organisationszentrums der Urodelen und Anuren. Zool. Anz., 116. 1936b. Die
laterale \'eranderung der induzierenden Fahigkeiten im Organisationszentrum der
schwanzlosen Amphibien. Biol. Zhurnal, 5. 1936c. Die Induktion des Hororgans aus
dem Ektoderm der Gastrula im xenoplastischen Experiment. Bull. Eiol. Med. exp., 2.
19370. Bildung des Haftnapfes und der Mundbewaffnung von Anuren im Tritonkeim.
Zool. Anz., 117. 1937b. Entwicklung der ausseren Kiemen aus dem Ektoderm der
Anuren und dem IMesoderm von Triton. Ibid. — Schmidt, W. J. 1928. Der submikro-
skopische Bau des Chromatins. L Zool. Jahrb. Abt. allg. Zool, 45. 1937. Die Doppel-
brechung von Karyoplasma, Zytoplasma und Metaplasma. Berlin. — Schmitt, F. 1901.
Systematische Darstellung der Doppelembryonen der Salmoniden. Arch. Entw'mech.,
13. 1902. tJber die Gastrulation der Doppelbildungen der ForeUe etc. Verb,
deutsch. zool. Ges., 1902. — Schmitt, F. O. 1939. The ultrastructure of protoplasmic
constituents. Physiol. Rev., lo. — Schneider, C. C. 1891. Untersuchungen iiber die
Zelle. Arb. zool. Inst. Wien, 9. — Schnetter, M. 1934a. Morphologische Untersuch-
ungen iiber das Differenzierungszentrum in der Embryonalentwicklung der Honig-
biene. Zeitschr. Morphol. Okol., 29. 1934b. Physiologische Untersuchungen iiber
Dififerenzierungszentrum in der Embryonalentwicklung der Honigbiene. Arch. Ent-
w'mech., 131. 1936. Die Entwicklung von Zwerglarven in geschniirten Bieneneiern.
\'erh. deutsch. zool. Ges., 1936. — Schotte, O. 1926a. La regeneration de la queue
d'Urodeles est liee a I'integrite du territoire caudal. C. R. Soc. phys. hist. nat. Geneve,
43. 1926b. L'importance du systeme nerveux sympathetique dans la regeneration des
pattes chez le Triton. C. R. Soc. Biol., 94. 1926c. Systeme nerveux et regeneration
chez le Triton. Rev. Suisse Zool., ^^. 1930. Der Determinationszustand der Anuren-
gastrula im Transplantations-Experiment. Arch. Entw'mech., 122. 1932. Xenoplas-
tic inductions. Abstr. Amer. Soc. Zool., Anat. Rec, 54. Suppl. i. 1933. Organizers and
inherent potentialities in the embryonic development of amphibians. Abstr. Marine
Biol. Lab., Collecting Net, 8. 1937. Embryonic induction in regenerating tissue.
Abstr. Nat. Acad. Sci., Science, 85. 1938. Induction of embryonic organs in regener-
ates and neoplasms. Abstr. Marine Biol. Lab., Collecting Net, 13. 1940. Recupera-
tion of lens competence in the epidermis of hatched larvae of Rana pipiens. Abstr.
Amer. Assoc. Anat., Anat. Rec, 76, Suppl. 2. Schotte, O., and Hummel, Katharine.
1939. Lens induction at the expense of regenerating tissues of amphibians. Jour. Exp.
Zool., 80. — Schuckmann, W. 1925. Zur Morphologic und Biologie von Dictyostelium
mucoroides Bref. Arch. Protistenk., 51. — Schultz, E. 1907. Uber Reductionen. III.
Arch. Entw'mech., 24. 1908. Uber umkehrbare Entwicklungsprozesse. Vortr. u. Aufs.
ii. Entw'mech., H. 4. — Schultze, O. 1894. Die ktinstliche Erzeugung von Doppelbild-
ungen bei Froschlarven mit Hilfe abnormer Gravitationswirkung. Arch. Entw'mech.,

BIBLIOGRAPHY 789

I. — Schwalbe, E. (ed.). 1907. Die Morphologic der Missbildungen des Menschen
und der Tiere. II. Die Doppelbildungen. Jena. 1913. Die Morphologie etc. III.
Abt. I, Kapitel 5, 6; Abt 2, Kapitel i, 5, 6. — Schwind, J. L. 1931. Heteroplastic ex-
periments on the limb and shoulder girdle of Amhlystoma. Jour. Exp. Zool., 59. 1932.
Further experiments on the limb and shoulder girdle of Amhlystoma. Ibid., 63. — Scott,
J. W. 1903. Periods of susceptibility in the differentiation of unfertilized eggs of Am-
phitrite. Biol. Bull., 5. 1906. Morphology of the parthenogenetic development of
Amphitrite. Jour. Exp. Zool., 3. — Seevers, C. H. 1932. Potencies of the end bud and
other caudal levels of the early chick embryo, etc. Anat. Rec, 54. — Seidel, F. 1926.
Die Determinierung der Keimanlage bei Insekten. I. Biol. Zbl., 46. 1928. Die De-
terminierungetc.il. Ibid.,^2>. 1929a. Die Determinierung etc. III. Ibid., 4g. 1929b.
Untersuchungen liber das Bildungsprinzip der Keimanlage im Ei der Libelle, Platyc-
nemis pennipes. I-V. Arch. Entw'mech., 119. 1931. Die Reaktionsfolge im Determina-
tionsgeschehen des Libellenkeims. Verb, deutsch. zool. Ges., 1931. 1934. Das Differ-
enzierungszentrum im Libellenkeim. I. Arch. Entw'mech., 131. 1935. Der Anlageplan
im Libellenkeim. Ibid., 132. — Seifriz, W. 1935. The structure of protoplasm. Biol.
Rev., I. 1936. Protoplasm. New York. 1938. Recent contributions to the theory of pro-
toplasmic structure. Science, 88. — Severinghaus, A. E. 1930. Gill development in
Amblystoma punctatiini. Jour. Exp. Zool., 56. — Sharp, L. W. 1912. Spermatogenesis
in Equisetum. Bot. Gaz., 54. — Shearer, C. 1924. On the oxygen consumption rate of
parts of the chick embryo and of segments of the earthworm. Proc. Roy. Acad., B, 96.
1930. A reinvestigation of metabolic gradients. Jour. Exp. Biol., 7. — Shen, S. C. 1939.
A quantitative study of amphibian neural tube induction with a water soluble hydro-
carbon. Jour. Exp. Biol., 16. — Shorey, M. Louise. 1909. The effect of the destruction
of peripheral areas on the differentiation of the neuroblasts. Jour. Exp. Zool., 7. —
Shoup, C. S., and Boykin, J. T., 1931. The insensitivity of Paramecium to cyanide and
effects of iron on respiration. Jour. Gen. Physiol., 15. — Siegmund, G. 1928. Die Be-
deutung des Nervensystems bei der Regeneration untersucht an Eisedia. Biol, gen., 4.
— Silber, R. H., and Hamburger, V. 1939. The production of duplicilas cruciata and
multiple heads by regeneration in Euplanaria tigrina. Physiol. Zool., 12. — Silvestri, F.
1906. Contribuzioni alia conoscenza biologica degli imenotteri parasiti. I. Ann. roy.
Scuola agric. Portici, ser. 2, 6. 1908. Contribuzioni etc. II. Ibid., 8.— Sivickis, P. B.
1923. Studies on the physiology of reconstitution in Planaria lata, with a description
of the species. Biol. Bull., 44. 1930. Distribution of setae in the earthworm, Phere-
tima benguetensis Beddard. Ibid., 58. 1931a. A quantitative study of regeneration in
Dendrocoelum lacteum. Arb. II. Abt. ungarisch. Biol. Forschungsinstituts, 4. 1931b.
A quantitative study of regeneration along the main axis of the triclad body. Arch.
Zool. ital., 16. 1933. Studies on the physiology of regeneration in triclads. Vytauto
didziojo Universiteto, matematikos-gamtos Fakulteto Darbai, 7. — Skoog, F., and
Thimann, K. V. 1934. Further experiments on the inhibition of the development of
lateral buds by growth hormone. Proc. Nat. Acad. Sci., 20. — Smith, G. M, 1938.
Cryptogamic botany. New York and London. — Snow, Mary and R. 1931. Experi-
ments on phyllotaxis. Phil. Trans. Roy. Soc, B, 221. 1934. The interpretation of
phyllotaxis. Biol. Rev., 9. — Snow, R. 1929. The young leaf as the inhibiting organ.
New Phytol., 28. 1931. Experiments on growth and inhibition. II. Proc. Roy. Soc,
B, 108. 1932a. Experiments, etc. III. Ibid., 111. 1932b. Growth-regulators in plants.

790 PATTERNS AND PROBLEMS OF DEVELOPMENT

New Phytol., 31. 1937. On the nature of correlative inhibition. Ibid., 36. — So-
landt, O. M.; Solandt, D. Y.; Ross, E. D.; and Gerard, R. W. 1934. Methemoglobin
and methylene blue as cyanide antagonists. Proc. Soc. Exp. Biol. Med., 31. — Sonne-
born, T. M. 1932. Experimental production of chains and its genetic consequences in
the cihate protozoan, Colpidium campylum (Stokes). Biol. Bull., 63. — Spek, J. 1918.
Differenzen im Quellungszustand der Plasmakolloide als eine Ursache der Gastrulain-
vagination sowie der Einstulpungen und Faltungen von Zellplatten iiberhaupt. Kol-
loidchem. Zeitschr., 9. 1919. Studien iiber den Mechanismus der Gastrulainvagina-
tion. Biol. Zbl., 39. 1926. tjber gesetzmassige Substanzverteilung bei der Furchung
des Ctenophoreneies und ihre Beziehung zu den Determinationsproblemen. Arch.
Entw'mech., 107. 1927. tJber die Winterknospenentwicklung, Regeneration und Re-
duktion bei Clavellina lepadiformis und die Bedeutung besonderer "omnipotente"
Zellelemente fiir diese Vorgiinge. Ibid., iii. 1930. Zustandsanderungen der Plasma-
kolloide bei Befruchtung und Entwicklung des Nereis-Eies. Protoplasma, 9. 1933.
Die bipolare Dififerenzierung des Protoplasmas des Teleosteer-Eies und ihre Entsteh-
ung. Ibid., 18. 1934a. Die bipolare Differenzierung des Cephalopoden- und des Proso-
branchiereies (Vitalfarbungsversuche mit Indikatoren etc.). Arch. Entw'mech., 131.
1934b. ijber die bipolare Differenzierung der Eizellen von Nereis limbata und Chaetop-
terns pergamentaceus. Protoplasma, 21. 1934c. Die Reaktion der Protoplasmakompo-
nenten des Asteriastxt?,. Ibid. — Spemann, H. 1901a. Uber Correlationen in der Ent-
wicklung des Auges. Verb. anat. Ges., 1901. 1901b. Entwicklungsphysiologische Stu-
dien am rr//o«ei. Arch. Entw'mech., 12. 1902. Entwicklungsphysiologische Studien
etc. II. Ibid., IS- 1903a. tJber Linsenbildung bei defekter Augenblase. Anat. Anz.,
23. 1903b. Entwicklungsphysiologische Studien etc. III. Arch. Entw'mech., 16. 1904.
tJber experimentell erzeugte Doppelbildungen mit cyclopischem Defekt. Zool. Jahrb.,
Suppl., 7. 1905. tJber Linsenbildung nach experimentellen Entfernungen der primaren
Linsenbildungszellen. Zool. Anz., 28. 1906. Uber embryonale Transplantation. Verh.
Ges. deutsch. Naturforsch. Arzte, 1906. 1907. Neue Tatsachen zum Linsenproblem.
Zool. Anz., 31. 1912a. Zur Entwicklung des Wirbeltierauges. Zool. Jahrg. Abt. allg.
Zool., 32. 1912b. liber die Entwicklung umgedrehter Hirnteile bei Amphibienembry-
onen. Zool. Jahrb. Suppl., 15. 1914. tJber verzogerte Kernversorgung von Keim-
teilen. Verh. deutsch. zool. Ges., 1914. 1918. tJber die Determination der ersten Or-
gananlagedesAmphibienembryo. I-VI. Arch. Entw'mech., 43. 1921. Die Erzeugung
tierischer Chimaren durch heteroplastische embryonale Transplantation zwischen
Triton cristatus und taeniatus. 1927. Organizers in animal development. Proc. Roy.
Soc, B, 102. 1928. Die Entwicklung seithcher und dorsoventraler Keimhalften bei
verzogerter Kernversorgung. Zeitschr. wiss. Zool., 132. 1931. Uber den Anteil von
Implantat und Wirtskeim an der Orientierung und Beschaffenheit der induzierten Em-
bryonalanlage. Arch. Entw'mech., 123. 1936. Experimentelle Beitrage zu einer The-
orie der Entwicklung. Berhn. 1938. Embryonic development and induction. New
Haven. Spemann, H., und Bautzmann, E. 1927. tJber Regulation von Tritonkeimen
mit iiberschiissigem und fehlendem medianem Material. Arch. Entw'mech., no.
Spemann, H., und Falkenberg, H. 1919. tJber asymmetrische Entwicklung und Situs
inversus Viscerum bei Zwillingen und Doppelbildungen. Ibid., 45. Spemann, H.;
Fischer, F.; und Wehmeier, E. 1933. Fortgesetzte X'ersuche zur Analyse der Induk-
tionsmittelinderEmbryonalentwicklung. Naturwiss.,21. Spemann, H., und Geinitz,B.

BIBLIOGRAPHY 791

1927. tjber Weckung organisatorischer Fahigkeiten durch Verpflanzung in organisato-
rische Umgebung. Arch. Entw'mech., 109. Spemann, H., und Mangold, O. 1924. tJber
Induktion von Embryonalanlagen durch Implantation artfremder Organisatoren.
Arch. mikr. Anat. Entw'mech., 100. Spemann, H., und Schotte, O. 1932. Uber xeno-
plastische Transplantation als Mittel zur Analyse der embryonalen Induktion. Natur-
wiss., 20. — Spengel, J. W. 1879. Beitrage zur Kenntnis der Gephyreen. I. Mitt. zool.
Stat. Neapel, i. — Spirito, A. 1935. Innesti e processi organizzatori in Planaria torva.
Arch. Zool. ital., 21. — Spooner, G. B. 1911. Embryological studies with the centrifuge.
Jour. Exp. Zool., 10. — Stachowitz, W. 1914. \'eranderungen in der Entwicklung von
Amphibienembryonen die auf dem Stadium der MeduUarplatte mit Radium bestrahlt
werden. Arch. mikr. Anat., 85. — Stahl, E. 1885. Uber den Einfluss der Beleuchtungs-
richtung auf die Teilung der Equisclumsporen. Ber. deutsch. bot. Ges., 1885. — Stauf-
facher, H. 1893. Eibildung und Furchung bei C^/c/a^ con?ca. Jen. Zeitschr., N.F., 21.
— Stein, Kathryn. 1933. The location and differentiation of the presumptive ectoderm
of the forebrain and hypophysis as shown by chorio-allantoic grafts. Physiol. Zool., 6.
— Steinmann, P. 1926. Prospektive Analyse von Restitutionsvorgangen. I. Arch.
Entw'mech., 108. — Stevens, N. M. 1902. Regeneration in Antcnnularia raniosa.
Arch. Entw'mech., 15. 1904. Further studies on the oogenesis of 5ag///a. Zool. Jahrb.
Abt. Anat. Ont., 21. 1907. A histological study of regeneration in Planaria simplicis-
sima, Planaria maculata and Planaria morgani. Arch. Entw'mech., 24. 1909. The
effect of ultra-violet light upon the developing eggs of Ascaris megaloccphala. Ibid., 27.
1910. Regeneration in Antennularia. Ibid., 30, I. — Stockard, C. R. 1907a. The arti-
ficial production of a single median cyclopean eye in the fish embryo by means of sea
water solutions of magnesium chloride. Ibid., 2^. 1907b. The influence of external fac-
tors, chemical and physical, on the development of Fiindulus hcteroditus. Jour. Exp.
Zool., 4. 1909. The development of artificially produced cyclopean fish. The mag-
nesium embryo. Ibid., 6. 1910a. The influence of alcohol and other anesthetics on
embryonic development. Amer. Jour. Anat., 10. 1910b. The independent origin and
development of the crystalline lens. Ibid. 1910c. The experimental production of vari-
ous eye abnormalities and an analysis of the development of the primary parts of the
eye. Arch, vergl. Ophthalmol., i. 1913. An experimental study of the position of the
optic anlage in Amblystoma punctatum with a discussion of certain eye defects.
Amer. Jour. Anat., 15. 1914. The artificial production of eye abnormalities in the chick
embryo. Anat. Rec, 8. 1921. Developmental rate and structural expression. Amer.
Jour. Anat., 28. — Stolte, H. A. 1936. Die Herkunft des Zellmaterials bei regenerativen
\'organgen der wirbellosen Tiere. Biol. Rev., ir. — Stone, L. S. 1926. Further experi-
ments on the extirpation and transplantation of mesectoderm in Amblystoma punc-
tatum. Jour. Exp. Zool., 44. 1931. Induction of the ear by the medulla and its relation
to experiments on the lateralis system in Amphibia. Abstr. Nat. Acad. Sci., Science,
74- — Strandskov, H. H. 1934. Certain physiological effects of X-rays on Euplanaria
dorotocephala. I. Physiol. Zool., 7. 1937. Certain effects, etc. II. Ibid., 10.— Stroer,
W. F. H. 1933. Experimentelle Untersuchungen iiber die Mundentwicklung bei den
Urodelen. Arch. Entw'mech., 130. — Stultz, W. A. 1936. Relations of asymmetry in
the hind limb of Amblystoma punctatum. Jour. Exp. Zool., 72. — Sturtevant, A. H.
1923. Inheritance of direction of coiling in Limnaea. Science, 58. — Sugino, H. 1938.
Miscellany on Planaria transplantation. Annot. Zool. Japon., 17. — Summers, F. M.

792 PATTERNS AND PROBLEMS OF DEVELOPMENT

19380. Some aspects of normal development in the colonial ciliate, Zoothamnium al-
tcrnans. Biol. Bull., 74. 1938b. Form regulation in Zoothamnium alternans. Ibid. —
Sumner, F. B. 1903. A study on early fish development. Arch. Entw'mech., 17. —
Sun, K. H., and Pratt, K. C. 1931. Do earthworms grow by adding new segments?
Amer. Naturalist, 65. — Surface, F. M. 1907. The early development of a polyclad,
Planocerainquilina. Proc. Acad. Nat. Sci. Phila., 59. — Suster, P. 1933. Fiihlerregen-
eration nach Ganglienexstirpation bei Sphodromantis hioculata Burm. Zool. Jahrb. Abt.
allg. Zool., 53. — Svetlov, P. 1934. tjber die Regeneration wahrend der Entwicklung.
Arch. Entw'mech., 131. — Swett, F. H. 1923. The prospective significance of the cells
contained in the four quadrants of the primitive limb disc of Ambly stoma. Jour. Exp.
Zool., 38. 1924. Exceptions to Bateson's rules of minor asymmetry. Anat. Rec, 28.
1926. On the production of double Hmbs in amphibians. Jour. Exp. Zool., 44. 1927.
Differentiation of the amphibian limb. Ibid., 47. 1928a. Further experiments on the
determination of the mediolateral axis of the fore limb of Amblystoma punctatum
(Linn.). Ibid., $0. 1928b. Transplantation of divided limb rudiments in .4 w?Wy5/w«a
punctatum (hinn.) . Ibid., s^. 1928c. Relations of symmetry in double limbs. 76/(f. 1930.
The permanence of limb-axis polarity. Ibid., S5- 1932. Reduplication in heteroplastic
limb grafts. Ibid., 61. 1937a. Experiments upon delayed determination of the dorso-
ventral limb-axis in Amblystoma punctatum (Linn.). Ibid.,-]$. 1937b. Determination
of limb-axes. Quart. Rev. Biol., 12. 1938a. Experiments designed to hasten polariza-
tion of the dorsoventral limb-axis in Amblystoma punctatum. Ibid.,'j8. 1938b. Experi-
ments upon the relationship of the surrounding areas to polarization of the dorsoven-
tral limb-axis \n Amblystoma punctatum. Ibid., 78. 1938c. Further experiments upon
axial polarization in the embryonic limb rudiment of the salamander. Abstr. Amer.
Soc. Zool., Anat. Rec, 72, Suppl. 4. 1939. Further experiments upon the establish-
ment and the reversal of prospective dorsoventral limb-axis polarity. Jour. Exp. Zool.,
82.

Takaya, H. 1936. Combination of two limb rudiments in urodele, Triturus with
special reference to their symmetry relations. Mem. Coll. Sci. Kyoto Imp. Univ., B,
12. 1938. Transplantation of limb-discs of varying size. Ibid., 14. — Tanaka, S. 1934.
Glycogen distribution in amphibian embryos. Proc. Imp. Acad. Tokyo, 10. — Tann-
reuther, G. W. 1907. History of the germ cells and early embryology of certain aphids.
Zool. Jahrb. Abt. Anat. Ont., 24. — Taylor, C. V. 1928. Protoplasmic reorganization in
Uronychia uncinata sp. nov. during binary fission and regeneration. Physiol. Zool., i.
1931. Polarity in normal and centrifuged eggs of Urechis caupo. Ibid., 4. Taylor, C. V.,
and Tennent, D. H. 1926. Experimental modifications of polarity in marine ova.
Science, 64. Taylor, C. V.; Tennent, D. H.; and Whitaker, D. M. 1926. An experimen-
tal study of organization in the egg. Abstr. Amer. Soc. Zool., Anat. Rec, 34, Suppl.
— Tennent, D. H. 1931. The maturation of the egg of the sea urchin, Mespilia globulus.
Abstr. Nat. Acad. Sci., Science, 74. Tennent, D. H.; Taylor, C. V.; and Whitaker,
D. M. 1929. An investigation on organization in a sea urchin egg. Carnegie Inst.
Publ. 391. — Tepliakova, M. J. 1937. Ontogeny of the cell and the problem of develop-
mental mechanics. VIII. Biol. Zhurnal, 6. — Theel, H. 1901. On a singular case of
hermaphroditism in holothuroids. Bihang K. Sv. Vet. Akad., Hdl. 27. — Thimann,
K. V. 1935. Growth substances in plants. Ann. Rev. Biochem., 4. Thimann, K. V.,
and Bonner, J. 1938. Plant growth hormones. Physiol. Rev., 18. Thimann, K. V.,

BIBLIOGRAPHY 793

and Skoog, F. 1933. Studies on the growth hormone of plants. III. Proc. Nat. Acad.
Sci., 19. 1934. On the inhibition of bud development and other functions of growth
substance in plants. Proc. Roy. Soc, B, 114. — Thompson, D. W. 1917. On growth
and form. Cambridge. — Tirala, L, 1912. Regeneration und Transplantation bei CnV
drilns. Arch. Entw'mech., 35. — Titlebaum, A. 1928. Artificial production of Janus
embryos. Proc. Nat. Acad. Sci., 14. — Tobler, F. 1902. Zerfall und Reproductions-
vermogen des Thallus einer Rhodomelacee. Bar. deutsch. bot. Ges., 20. 1904. tjber
Eigenwachstum der Zelle und Pflanzenform. Jahrb. wiss. Bot., 39. 1906. tJber Re-
generation und Polaritat, sowie verwandte Wachstumsvorgange bei Polysiphonia und
anderen Algen. Ibid., ^2. — Toro, E. 1932. Die Regeneration der Linse in friihen Ent-
wicklungsstadien bei Amblystoma mexkanum. Arch. Entw'mech., 126. — Tokin, B. P.,
und Gorbunowa, G. P. 1934. Untersuchungen liber die Ontogenie der Zelle. Zeitschr.
Biol., 3. (Russian with German summary.) — Tokura, R. 1925. Entwicklungsmecha-
nische Untersuchungen iiber das Horblaschen und das akustische, sowie faciale Ganglion
bei den Anuren. Folia anat. Japon., 3. — Tonkoff, W. 1900. Experimentelle Erzeugung
von Doppelbildungen bei Triton. Sitzber. k. preuss. Akad. wiss. phys.-math. Kl. Jhg.
1900. 1904. tJber die Entwicklung von Doppelbildungen aus dem normalen Ei. Trav.
Soc. imp. Nat. St. Petersbourg, 35. — Torrey, H. B. 1907. Biological studies on Cory-
morpha.ll. Univ. Calif. Publ. Zool., 3. 1912. Oxygen and polarity in Ti^te/ana. Ibid.,
9- — Torrey, J. C. 1903. The early embryology of Thalassema. Ann. N.Y. Acad. Sci.,
14- — Treadwell, A. L. 1901. The cytogeny of Podarke obscura. Jour. Morphol., 17.
1902. Notes on the nature of "artificial parthenogenesis" in the egg of Podarke ob-
scura. Biol. Bull., 3. — Tung, T. C. 1933. Recherches sur la determination du plan
median dans I'oeuf de y?a;w/«5rt7. Arch. Biol., 44. 1934. Recherches sur les potential-
ites des blastomeres chez Ascidiella scabra. Arch. Anat. micr., 30. — Tur, J. 1904. Mal-
formations embryonnaires obtenus par Taction du radium sur les oeufs de la poule.
C. R. Soc. Eiol. 57. — Twiesselmann, F. 1938. Experiences de scission precoce de I'aire
embryogene chez le poulet. Arch. Eiol., 49. — Twitty, V. C. 1930. Regulation in the
growth of transplanted eyes. Jour. Exp. Zool., 55. 1932. Influence of the eye on the
growth of its associated structures studied by means of heteroplastic transplantation.
Ibid., 61. 1939. Experiments on the determination of the fin ectoderm in salamander
embryos. Abstr. Amer. Soc. Zool., Anat. Rec, 75, Suppl. 4. Twitty, V. C, and
Schwind, J. L. 1928. Growth of heteroplastically transplanted eyes and limbs in
Amblystoma. Proc. Soc. Exp. Biol. Med., 25. 1931. The growth of eyes and hmbs
transplanted heteroplastically between two species of Amblystoma. Jour. Exp. Zool.,
59- — Tyler, A. 1930. Experimental production of double embryos in annelids and
mollusks. Ibid., 57. 1933. On the energetics of differentiation. Pubbl. Staz. zool.
Napoli, 13. 1935. On the energetics of differentiation. II. Biol. Bull., 68.

Ubisch, L. von. 1915. tJber den Einfluss von Gleichgewichtsstorungen auf die
Regenerationsgeschwindigkeit. Arch. Entw'mech., 41. 1923. Linsenbildung bei i?a«a
fusca trotz Entfernung des Augenbechers. Verh. deutsch. zool. Ges., 1923. 1924. tJber
den Einfluss verschieden hoher Temperatur auf der Bildung der Linse von Rana fusca,
R. esculenta und Bombinator pachypus. Zeitschr. wiss. Zool., 123. 1925a. Entwick-
lungsphysiologische Studien an Seeigelkeimen. I, II, III. Ibid., 124. 1925b. Uber die
Entodermizierung ektodermaler Bezirke des Echinoideen-Keimes und die Reversion
dieses Vorganges. \'erh. phys.-med. Ges. Wiirzburg, 50. 1925c. Uber die unabhiingige

794 PATTERNS AND PROBLEMS OF DEVELOPMENT

L'msenhildung hei Ranafusca. Arch. Entw'mech., 105. 1927. Beitriige zur Erforsch-
ung des Linsenproblems. Zeitschr. wiss. Zool., 129. 1929. Uber die Determination der
Larvalorgane und der Imaginalanlage bei Seeigeln. Arch. Entw'mech., 117. 1931.
Untersuchungen liber Formbildung mit Hilfe experimentell erzeugter Keimblattchi-
maren von Echinodermlarven. Ibid., 124. 1932a. Untersuchungenetc.il. Ibid., 126.
1932b. Untersuchungen etc. in. Ibid., i2j. 1933. Untersuchungen etc. IV, V. Ibid.,
129. 1934. Untersuchungen etc. VI. Ibid., 131. 1936a. Uber die Organisation des
Seeigelkeims. Ibid., 134. 1936b. Untersuchungen etc. VII. Ibid. 1938a. Eine ver-
gleichende Studie liber die Organisation des Keimes von Seeigeln, Amphibien und
Ascidien. Biol. Zbl., 58. 1938b. Uber Keimverschmelzungen an Ascidiclla aspersa.
Arch. Entw'mech., 138. 1940. Uber die Entwicklung von Ascidienlarven nach frlih-
zeitiger Entfernung der einzelnen organbildenden Keimbezirke. Ibid., 139. — Umanski,
E. 1932. tjber Induktion der MeduUarplatte bei Triton taeniatus durch Implantation
von Regenerationsblastem in die Blastula. Zool. Anz., 97. 1933. Uber das Induktions-
vermogen des Regenerationsblastems. Ibid., 104. — Unger, W. B. 1926. The relation
of rhythms to nutrition and excretion in Paramecium. Jour. Exp. Zool., 43.

Van Beneden, E. 1870. Recherches sur la composition et la signification de I'ceuf.
Mem. cour. Acad. roy. Eelg., 34. Van Beneden, E., et Julin, C. 1884. La segmentation
chez les Ascidiens dans ses rapports avec I'organisation de la larve. Arch, tiol., 5. —
Van Cleave, C. D. 1929. An experimental study of fission and reconstitution in Steno-
stomum. Physiol. Zool., 2. 1937. A study of the process of fission in the naid Pristina
longiseta. Ibid., 10. — Vandebroek, G. 1936a. Les mouvements morphogenetiques au
cours de la gastrulation chez ScyUium caiiicula Cuv. Arch. Biol., 46. 1936b. Plasma-
bewegingen tydens de bevruchting in het ei van Ascidia aspersa. Naturwetensch.
Tydschr. congres-nummer, 1936. — Van Duyne, J. 1896. Uber Heteromorphose bei
Planarien. Arch. ges. Physiol., 64. — Van Tieghem, P. 1880. Sur quelques myxo-
mycetes a plasmode aggrege. Bull. Soc. bot. France, 27. — Vejdovsky, F. 1878. tJber
die Eibildung und die Miinnchen von BoneUia viridis. Zeitschr. wiss. Zool., 30. 1883-
92. Entwicklungsgeschichtliche Untersuchungen. 1-4. Prag. — Vialleton, L. 1888,
Recherches sur les premieres phases du developpement de la Seiche {Sepia officinalis).
Ann. Sci. nat., ser. 7, 6. — Vochting, H. 1885. Uber die Regeneration der Marchantieen.
Jahrb. wiss. Bot., 16. — Vogt, W. 1922a. Operativ bewirkte "Exogastrulation" bei
Triton und ihre Bedeutung fiir die Theorie der Wirbeltiergastrulation. Verh. anat.
Ges., 1922. 1922b. Die EinroUung und Strechung der Urmundlippen bei Triton nach
Versuchen mit einer neuen Methode embryonaler Transplantation. Verh. deutsch.
zool. Ges., 27. 1923. Morphologische und physiologische Fragen der Primitiventwick-
lung, Versuche zu ihrer Losung mittels vitaler Farbmarkierung. Sitzber. ges. Morph.
Physiol., Mlinchen, 35. 1925. Gestaltungsanalyse am amphibienkeim mit ortlicher
V^italfarbung. I. Arch. Entw'mech., 106. 1926a. Uber Wachstum und Gestaltungs-
bewegungen am hinteren Korperende der Amphibien. Verh. anat. Ges., 1926. 1926b.
Die Beziehungen zwischen Furchung, Hauptachsen des Embryo und Ausgangsstruktur
im Amphibienei, nach Versuchen mit ortlicher Vitalfarbung. Sitzber. ges. Morph.
Physiol., Mlinchen, 38. 1927. Uber Hemmung der Formbildung an einer Halfte des
Keimes. Verh. anat. Ges., 1927. 1928a. Ablenkung der Symmetrie durch halbseitige
Beschleunigung der Frlihentwicklung. Ibid., ig28. 1928b. Mosaikcharakter und Reg-
ulation in der Frlihentwicklung des Amphibieneies. Verh. deutsch. zool. Ges., 32,

BIBLIOGRAPHY 795

1929. Gestaltungsanalyse am Amphibienkeim mit ortlicher Vitalfarbung. II. Arch.
Entw'mech., 120. 1931. tjber regeneratives und regulatives Wachstum nach Defekt-
versuchen am Schwanz und an Schwanzknospe der Amphibienkeime. Verb. anat.
Ges., 1931. Vogt, W., und Bruns, Emmy. 1930. Experimente liber das Regulations-
vermogen der Blastula von Triton taeniatus und Bomhinator igneiis. Arch. Ent-
w'mech., 122.

Wachs, H. 1 9 14. Neue Versuche zur Wolffschen Linsenregeneration. Arch. Ent-
w'mech., 39. 1919. ZurEntwicklungsphysiologiedes AugesderWirbeltiere. Naturwiss.,
7. 1920. Restitution des Auges nach Exstirpation von Retina und Linse bei Tritonen.
Arch. Entw'mech., 46. — Waddington, C.H. 1930. Developmental mechanics of chick-
en and duck embryos. Nature, 125. 1932. Experiments on the development of chick
and duck embryos cultivated in vitro. Phil. Trans. Roy. Soc, B, 221. 19330. Induc-
tion by the endoderm in birds. Arch. Entw'mech., 128. 1933b. Induction by the prim-
itive streak and its derivatives in the chick. Jour. Exp. Biol., 10. 1934. Experiments
on embryonic induction. I-III. Ibid., 11. 19350. The development of isolated parts of
the chick blastoderm. Jour. Exp. Zool., 71. 19356. Cancer and the theory of organiz-
ers. Nature, 135. 1936a. A failure of induction in normal development. Jour. Exp.
Biol., 13. 19366. Organizers in mammalian development. Nature, 138. 1937. Ex-
periments on determination in the rabbit embryo. Arch. Biol., 48. 19380. The dis-
tribution of the evocator in the unfertilized egg. Jour. Exp. Biol., 15. 19386. The na-
ture and mode of action of chemical evocators in embryonic induction in Amphibia.
Abstr. Amer. Soc. Zool., Anat. Rec, 72, 4, Suppl. Waddington, C. H., and Needham,
D. M. 1935. Induction by synthetic polycychc hydrocarbons. Proc. Roy. Soc, B, 117.
Waddington, C. H.; Needham, J.; and Brachet, J. 1936. Studies on the nature of the
amphibian organizer. III. Ibid., 120. Waddington, C. H.; Needham, J.; and Need-
ham, D. M. 1933a. Physico-chemical experiments on the amphibian organizer. Nature,
132. 19336. Beobachtungen iiber die physikalisch-chemische Natur des Organisators.
Naturwiss., 21. Waddington, C. H.; Needham, J.; Nowinski, W.; and Lemberg, R.

1935. Studies on the nature of the amphibian organization centre. Proc. Roy. Soc, B,
117. Waddington, C. H.; Needham, J.; Nowinski, W.; Lemberg, R.; and Cohen, A.

1936. Studies on the nature of the amphibian organizer. Ibid., 120. Waddington,
C.H.; Nowinski, W.; Needham, J.; and Needham, D. M. 1934. Active principle of am-
phibian organization centre. Nature, 134. Waddington, C. H., and Schmidt, G. A.
1933. Induction by heteroplastic grafts of the primitive streak in birds. Arch. Ent-
w'mech., 128. Waddington, C.H., and Taylor, J. 1937, Conversion of presumptive ecto-
derm to mesoderm in the chick. Jour. Exp. Biol., 14. Waddington, C. H., and Wolsky,
A. 1936. The occurrence of the evocator in organisms which possess no nerve cord.
Jour. Exp. Biol., 13.— Wallengren, H. 1901. Zur Kenntnis des Neubildungs- und
Resorptionsprocesses bei der Teilung der hypotrichen Infusorien. Zool. Jahrb. Abt.
Anat. Ont., 15.— Walter, F. K. 191 1. Welche Bedeutung hat das Nervensystem fiir
die Regeneration der Tritonextremitiiten? Arch. Entw'mech., 33.— Watanabe, Y.
1928. On the electrical polarity in the earthworm, Perichaeta communissima. Sci.
Repts. Tohoku Imp. Univ., Ser. 4, 3. 1930. On the physiological axial gradients of
chaetopod annehds. I. Ibid., 5. 1931. On the physiological axial gradients etc. II.
Ibid., 6. 1935a. Rate of head development as indicated by time of appearance of eyes
in the reconstitution of Eiiplanaria dorotocephala. Physiol. Zool, 8. 19356. Head fre-

796 PATTERNS AND PROBLEMS OF DEVELOPMENT

quency in Euplanaria maculata in relation to the nervous system. Ibid., 1935c. Physio-
logical dominance in Corymorpha palma in relation to reconstitution and methylene
blue reduction. Ibid. 1937. Dominance and axial differentials in indophenol blue re-
action during reconstitution of the stalked medusa, Halidystus auricula Clark. Sci.
Repts. Tohoku Imp. Univ., Ser. 4, 12. Watanabe, Y., and Child, C. M. 1933. The
longitudinal gradient in Stylochus ijimai: with a critical discussion. Physiol. Zool., 6.
— Watase, S. 1891. Studies on cephalopods. I. Jour. Morphol., 4. — Waterman, A. J.
1932. Effects of isotonic salt solutions upon the development of the blastula of the sea
urchin, Faracenirotus lividus. Arch. Biol., 43. 1934. Further studies of exogastrulation
in the sea urchin, Biol. Bull., 47. 1937. Effect of salts of heavy metals on development
of the sea urchin, Arbacia punctulata. Biol. Bull., 73. 1938. Respiratory stimulants
and gastrulation in Arbacia. Abstr. Marine Biol. Lab., Biol. Bull., 75. — Webber, H. J.
1901. Spermatogenesis and fecundation of Zamia. U.S. Dept. Agric, Bur. Plant In-
dustry, Bull. 2. — Weber, H. 1928. tjber Induktion von MeduUarplatte durch seitlich
angeheilte Keimhalften bei Triton taeniatus. Arch. Entw'mech., 113. — Weber, R. A.
1917. Observations on the structure of double monsters in the earthworm. Biol. Bull.,
7,1,. — Wehmeier, E. 1934. Versuche zur Analyse der Induktionsmittel bei der Medul-
larplatteninduktion von Urodelen. Arch. Entw'mech., 132. — Weigmann, R. 1926.
Zur Kenntnis des grauen Halbmondes und der Entstehung der Bilateralitiit im Ei von
Ranafiisca. Zool. Anz., 69. 1927. tJber die Bestimmung der Medianebene im Froschei.
Zeitschr. wiss. Zool., 129. — Weimer, B. R. 1928. The physiological gradients in Hy-
dra. I. Physiol. Zool., i. 1932. The physiological gradients, etc. II. Jour. Exp. Zool.,
62. 1934. The physiological gradients, etc. III. Physiol. Zool., 7. — Weismann, A.
1877. Die Eibildung bei den Daphniden. Zeitschr. wiss. Zool., 28. 1880. Beitriige zur
Naturgeschichte der Daphniden. V. Ibid., i:^,. — Weiss, P. 1924. Physiologic der
Formbildung. Jahresber. ges. Physiol., 1924. 1925a. Abhangigkeit der Regeneration
entwickelter Amphibienextremitaten vom Nervensystem. Arch. mikr. Anat. Ent-
w'mech., 104. 1925b. Unabhangigkeit der Extremitatenregeneration vom Skelett. Ibid.
1925c. Die seitliche Regeneration der Urodelenextremitat. Ibid. 1926a. Morpho-
dynamik. Abh. theoret. Biol., H. 23. Berhn. 1926b. Ganzregenerate aus halbem Ex-
tremitatenquerschnitt. Arch. Entw'mech., 107. 1927a. Die Herkunft der Haut im
Extremitatenregenerat. Ibid., 109. 1927b. Potenzpriifungam Regenerationsblastem. I.
Ibid., III. 1928. Morphodynamische Feldtheorie und Genetik. \'erh. V. internat.Kongr.
Vererbungswiss. 1930. Entwicklungsphysiologie. Dresden. 1934. /» I'/Zz-o experiments
on the factors determining the course of the outgrowing nerve fiber. Jour. Exp. Zool.,
68. 1935. The so-called organizer and the problem of organization in amphibian devel-
opment. Physiol. Rev., 15. 1939. Principles of development. New York. — Weitzmann,
W. R. 1927. Zur Kenntnis der Regenerationserscheinungen im Mesoderm der Oligo-
chaten. Arch. Entw'mech., no. — Welch, P. S. 1921. Bifurcation in the embryos of
Tubifex. Biol. Bull., 41. — Went, F. W. 1935. Auxin, the plant growth hormone. Bot.
Rev., I. 1936. AUgemeine Betrachtungen iiber das Auxin-Problem. Biol. Zbl., 56.

Went, F. W., and Thimann, K. V. 1937. Phytohormones. New York. Werber, E. I.

1 91 5. The influence of products of pathologic metabolism on the developing teleost
ovum. Biol. Bull., 28. 1916a. On the blastolytic origin of the "independent" lenses of
some teratophthalmic embryos and its significance for the normal development of the
lens in vertebrates. Jour. Exp. Zool., 21. 1916b. Experimental studies of the origin of

BIBLIOGRAPHY 797

monsters. I. Ibid. 1918. Critical notes on the present status of the lens problem. Biol.
Bull., 34. — Wetzel, G. 1895. tJber die Bedeutung der circuliiren Furche in der Ent-
wicklung der Schultzeschen Doppelbildungen von Rana fusca. Arch. mikr. Anat., 46.
1904. Zentrifugalversuche an unbefruchteten Eiern von i^a«a/M.jra. Ibid., 6;^. 1925a.
Untersuchungen am Hiihnchenkeim. Verh. deutsch. anat. Ges., 1925. 1925b. Unter-
suchungen am Huhnerkeim. I. Arch. Entw'mech., 106. 1929. Untersuchungen am
Hiihnchen. Ibid., iig. 1931. Urmund und Primitivstreifen. Ergebn. Anat., 29. 1936.
Primitivstreifen und Urkorper nach Storungsversuchen am 1-2 Tage bebriiteten
Hiihnchen. Arch. Entw'mech., 134. — Wheeler, W. M. 1896. The sexual phases of
Myzostoma. Mitt. zool. Stat. Neapel, 12. 1897. The maturation, fecundation and
early cleavage of Myzostoma glabnmi Leuckart. Arch. Biol., 15. — Whitaker, D. M.
1935. Induction of polarity in Fucus furcatus by a localized concentration of hydrogen
ions. Proc. Soc. Exp. Biol. Med., ^^. 1937a. The effect of hydrogen ion concentration
upon the induction of polarity in Fucus eggs. L Jour. Gen. Physiol., 20. 1937b. De-
termination of polarity by centrifuging eggs of Fucus furcatus. Biol. Bull., 73. 1938a.
The effect of hydrogen ion concentration, etc. IIL Jour. Gen. Physiol., 21. 1938b. The
effect of pH on the development of ultracentrifuged Fucus eggs. Proc. Nat. Acad. Sci.,
24. 1938c. The effect of shape on the polarity of the Fucus egg. Abstr. Amer. Soc.
Zool., Anat. Rec, 72, Suppl. 4. 1940. The effect of shape on the developmental axis
of the Fucus egg. Biol. Bull., 78. Whitaker, D. M., and Lowrance, E. W. 1936. On
the period of susceptibility in the eggs of Fucus furcatus when polarity is induced by
brief exposure to directed white Hght. Jour. Cell. Comp. Physiol., 7. 1937. The effect
of hydrogen ion concentration upon the induction of polarity in Fucus eggs. II. Jour.
Gen. Physiol., 21. — Whitman, C. O. 1878. The embryology of C/c/>5me. Quart. Jour.
Micr. Sci., 18. — Wiercinski, F. J. 1939. The effects of supersonic vibrations on recon-
stitution and head frequency in Euplanaria dorotocephala. Physiol. Zool., 12. Wier-
cinski, F. J., and Child, C. M. 1936. Differential susceptibility of living organisms to
supersonic vibrations. Science, 83. — Wiersejski, A. 1906. Embryologie von Physa
fontinalis. Zeitschr. wiss. Zool., 83. — Wilder, H. H. 1904. Duplicate twins and double
monsters. Amer. Jour. Anat., 3. 1908. The morphology of cosmobia. Ibid., 8. —
Wilder, Janet. 1937. A correlation of results on oxygen consumption obtained by the
Winkler method and by respirometers, using as a standard the methods of Van Slyke.
Physiol. Zool., 10. — Will, L. 1893. Beitrage zur Entwicklungsgeschichte der Rep-
tilien (Gecko). Zool. Jahrb. Abt. Anat. Ont. 6. — Willier, B. H., and Rawles, Mary.
1931. The relation of Hensen's node to the differentiating capacity of whole chick
blastoderms as studied in chorio-allantoic grafts. Jour. Exp. Zool., 59. — Wilson, C. B.
1897. Experiments on the early development of the amphibian embryo under the in-
fluence of Ringer and salt solutions. Arch. Entw'mech., 5. 1899. The habits and early
development of Cerebratulus lacteus. Quart. Jour. Micr. Sci., 63. — Wilson, E. B. 1892.
The cell Hneage of Nereis. Jour. Morphol., 6. 1893. Amphioxus and the mosaic theory
of development. Ibid., 8. 1894. The mosaic theory of development. Woods Hole Biol.
Lect., 2. 1896. Appendix. On cleavage and mosaic work. Arch. Entw'mech., 3. 1898.
Considerations on cell-Hneage and ancestral reminiscence. Ann. N. Y. Acad. Sci., 11.
1903a. Reversal of asymmetry in the regeneration of the chelae of Alpheus hetero-
clielis. Biol. Bull., 4. 1903b. Notes on merogony and regeneration in Renilla. Ibid.
1903c. Experiments on cleavage and locahzation in the nemertine egg. Arch. Ent-

798 PATTERNS AND PROBLEMS OF DEVELOPMENT

w'mech., i6. 1904. Experimental studies in germinal localization. I, IL Jour. Exp.
Zool., I. 1925. The cell in development and heredity. New York. Wilson, E. B., and
Mathews, A. P. 1895. IMaturation, fertilization and polarity in the echinoderm egg.
Jour. Morphol., 10. — Wilson, H. V. 1907. On some phenomena of coalescence and re-
generation in sponges. Jour. Exp. Zool., 5. 1911a. Development of sponges from dis-
sociated tissue cells. Bull. Bur. Fish., 30. 1911b. On the behavior of dissociated cells
in hydroids, Alcyonaria and Asterias. Jour. Exp. Zool., 11. — Wilson, J. W. 1926.
Regeneration of Planar ia maculatu in isotonic Ringer fluid. Abstr. Amer. Soc. Zool.,
Anat. Rec, 34, Suppl. — Winkler, H. 1900a. Uber Polaritiit, Regeneration und Hetero-
morphose bei Bryopsis. Jahrb. wiss. Bot., 35. 1900b. tjber den Einfluss ausserer Fak-
toren auf die Teilung des Eies von Cytosira harhata. Ber. deutsch. bot. Ges., 2^. —
Wintrebert, P. 1904. Sur la regeneration des membres posterieurs chez I'axolotl
adulte apres ablation de la mccUe lombosacree. C. R. Soc. Biol., 56. 1933a. La mecan-
ique embryonnaire des Amphibiens consideree d'une maniere epigenetique, comme un
enchainement de structures et de fonctions transitoires. C. R. Acad. Sci. Paris, 197.
1933b. Mosaique, regulation, epigenese. Ihid. 1934. Les lois de I'epigenese chez les
Amphibiens. Ibid., 198. — Wistinghausen, C. von. 1893. Untersuchungen liber die
Entwicklung von Nereis dumerilii. Mitt. zool. Stat. Neapel, 10. — Wittmann, C. 1929.
Untersuchungen an Schultzeschen Doppelbildungen von Rana fusca und Triton taenia-
tus. Zeitschr. wiss. Zool., 134. — Woerdeman, M. W. 1933a. Uber den Glykogenstoff-
wechsel des Organisationszentrums in der Amphibiengastrula. Proc. kon. Akad. Wet-
ensch. Amsterdam, 36. 1933b. Uber den Glykogenstofifwechsel tierischer Organisa-
toren. Ibid. 1933c. Embryonale Induktion durch Geschwulstgewebe. Ibid. 1933d.
Uber die chemischen Prozesse bei der embryonalen Induktion. Ibid. 1934. Uber die
Determination der Augenlinsenstruktur bei Amphibien. Zeitschr. mikr. Anat. Forsch.
36. — Wolf, E. 1932. Pulsation frequency of advisceral and abvisceral heart beat of
Ciona intcstinalis in relation to temperature. Jour. Gen. Physiol., 16. — Wolff, G.
1895. Entwicklungsphysiologische Studien. L Arch. Entw'mech., i. 1901. Entwick-
lungsphysiologische Studien. II. Ibid., 12. 1910. Regeneration und Nervensystem.
Festschr. 60 Geburtstag R. Hertwigs. Jena. 1913. Entwicklungsphysiologische Stu-
dien. III. Arch. mikr. Anat., 63. — Wood- Jones, F. 1912. Coral and atolls. London. —
Woodside, G. L. 1937. The influence of host age on induction in the chick blastoderm.
Jour. Exp. Zool., 75. — Woronzowa, M. A. 1938. Die Regenerationspotenzen der
Schultergurtelmuskulatur. Bull. tiol. Med. exp., 6. — Wright, S. 1934. On the genetics
of subnormal development of the head (otocephaly) in the guinea pig. Genetics, 19.
Wright, S., and Wagner, K. 1934. Types of subnormal development of the head from
inbred strains of guinea pigs and their bearing on the classification and interpretation
of vertebrate monsters. Amer. Jour. Anat., 54.

Yamanouchi, S. 1908. Spermatogenesis, oogenesis and fertilization in Ncphro-
dium. Bot. Gaz., 45. — Yatsu, N. 1910a. A note on the polarity of the primary oocyte
of Asterias forbesii. Annot. Zool. Japon., 7. 1910b. Experiments on cleavage in the
egg of Cerebratulus. Jour. Coll. Sci. Tokyo Imp. Univ., 27. 1910c. Experiments on
germinal localization in the egg of Cerebratulus. Ibid., 1911. Observations and experi-
ments on the ctenophore egg. II. Annot. Zool. Japon., 7. 1912. Observations, etc. I.
Jour. Coll. Sci. Tokyo Imp. Univ., 32. 1912b. Observations, etc. III. Annot. Zool.
Japon., 8. — Yntema, C. L. 1933. Experiments on the determination of the ear ecto-

BIBLIOGRAPHY 799

derm in the embryo of Amblystoma punctatum. Jour. Exp. Zool., 65. — Young, D. B.
1922. A contribution to the morphology and physiology of the genus Uronvchia. Ibid.,
36.

Zacharias, O. 1885. Uber die amoeboiden Bewegungen der Spermatozoen von
Polyphemus pediculus. Zeitschr. wiss. Zool., 61. — Zalpeter, Helene. 1927. The isopo-
tency of generally homologous parts of the body. Arch. Entw'mech., log. — Zanoni, G.
1934a. Simmetria morfologica e fisiologica degli elementi tepalici. Boll. Mus. Lab. Zool.
.\nat. Univ. Genova, 14. 1934b. Equivalenti fisiologici della simmetria. Ihid. 1934c.
Gradienti autonomi e correlazioni d'origine embrionale nel seme delle Leguminose.
Arch, bot., 10. — Zeiger, K. 1939. R. E. Liesegang's Dififusionshistologie. Kolloid.
Zeitschr., 89. — Zeleny, C. 1902. A case of compensatory regulation in the regeneration
oi Hydroides dianthus. Arch. Entw'mech., 13. 1904. E.xperiments on the localization
of developmental factors in the nemertine egg. Jour. Exp. Zool., i. 1905a. Compensa-
tory regulation. Ibid., 2. 1905b. The relation of the degree of injury to the rate of re-
generation. Ibid. 1907. The direction of differentiation in development. I. Arch.
Entw'mech., 23. 1909. The relation between degree of injury and rate of regeneration.
Jour. Exp. Zool., 7. 191 1. Experiments on the control of asymmetry in the develop-
ment of the serpulid, Hydroides dianthus. Jour. Morphol., 22. — Zelinka, C. 1891.
Studien iiber Raderthiere. 3. Zeitschr. wiss. Zool., 53. — Zhinkin, L. 1935. Uber den
determinierenden Einfluss des Nervensystems auf die Regeneration bei Lumbriculus
variegalus. Arch. russ. Anat. Histol. Embryol., 14. (Russian, German summary.) 1936.
The influence of the nervous system on regeneration in Rhynchelmis limosclla. Jour.
Exp. Zool., 73. — Ziegler, H. E. 1894. Uber Furchung unter Pressung. Verh. anat.
Ges., 1894. 1898. Experimentelle Studien uber die Zellteilung. III. Arch. Entw'mech.,
7. 1903. Experimentelle Studien. I\'. Ihid., 16. — Zimmermann, W. 1929. Experi-
mente zur Polaritat von Caulcrpa und zum allgemeinen Polaritiitsproblem. Ibid., 116.
— Zoja, R. 1895. SuUo sviluppo dei blastomeri isolati dalle uova di alcune meduse (e di
altriorganismi). Ibid., 1,2. 1896. Untersuchungen iiber die Entwicklung der Ascaris
megalocephala. Arch. mikr. Anat., 47. — Zur Strassen, O. 1896. Embryonalentwick-
lung von Ascaris megalocephala. Arch. Entw'mech., 3. 1898. Uber die Riesenbildung
bei Ascariseiern. Ibid., 7. 1906. Die Geschichte der T-Riesen von Ascaris megalo-
cephala. Zoologica, H. 40.— Zwilling, E. 1939. The effect of removal of perisarc on
regeneration in Tubularia crocca. Abstr. Marine Biol. Lab., Biol. Bull., 76.

INDEX

Except when it is desired to call particular attention to different points under the same
subject head on successive or closely following pages, page references give only number of
first page concerned; following pages may deal with the same subject.

Acceleration, differential: of planarian head,
192; of fish development, 257; of frog de-
velopment, 262

AccUmation, differential: to external agents,
72. See also Conditioning, differential;
Tolerance, differential

Acetabularia, reconstitution in, 31, 361

Acrasieae, development of axiate pattern in,
636

Allolobophora, physiological gradients of, 122

Amblysloma: somite modification by lithium
in, 264; eye field of, 282; hmb field of, 286;
otic development in, 288; limb trans-
plantation in, 390; exogastrulation in, 463;
lens development in, 489; otic induction
in, 496; balancer of, 498; species-specific
growth in, 718. See also Amphibian de-
velopment; Amphibian dorsal inductor;
Amphibian limb

Amphibian development : differential suscep-
tibiUty in, 151, 255; dorsiventrahty in,
151, 259, 286, 289, 390, 657, 684; respira-
tion in, 153; glycolysis in, 154; dye reduc-
tion in, 156; sulphydrylin, 157; embryonic
differential inhibition in, 258, 262; obliter-
ation of pattern in, 259; secondary modi-
fications of, 262; effects of ligature on, 266,
524; hybrid differential inhibition of, 267;
in relation to gravity, 428; dominance in,
43 S; formal patterns of, 447; cell move-
ments in, 448; neural induction in, 454;
normal embryonic, 456; exogastrulation
in, 463, 464; lens induction in, 472, 488,
491; induction of optic parts in, 496; in-
cluction of balancer in, 499; gill induction
in, 500; "double assurance" in, 500;
blastomere reconstitution in, 523; after
Ugature, 524; independent differentiation
in, 528; effect of embryonic fusions on,
542; effect of centrifugal force on, 584;
dispermy in, 595, 685; gray crescent in,
657, 684. See also Amphibian dorsal in-
ductor

Amphibian dorsal inductor: differential sus-
ceptibility in, 152, 263; respiration of , 153;
glycolysis in, 154; glycogen in, 156, 477;
dye reduction in, 156, 157; sulphydryl in,
157; differential inhibition of, 258; effect
of implants of, 455; not species-, genus-,
or order-specific, 457; change in capacity

for reaction to, 458; regional differentia-
tion in, 459, 460; regional specificity in,

461, 526; neural differentiation without,

462, 466; reconstitution of, 462; formation
of, from other ectoderm, 465; direction of
invagination of implanted, 468; origin of,
486; in ligatured eggs, 524; in centrifuged
eggs, 584. See also Amphibian develop-
ment; Induction

Amphibian Umb: early stages of, 19; field of,
225, 285, 293; dominance in field of, 286;
nervous influence in regeneration of, 341;
new polar pattern in regeneration of, 370,
390; reduplication and mirror-imaging of,
390; transplantation and dorsiventrahty
of, 391; rotation of transplants of, 394

Animalization, in echinoderm development,
243

Annelids: reconstitution in, 47, 48, 337, 355,
368; differential susceptibility of, 120, 126,
128; oxygen uptake of, 121, 123; CO2 pro-
duction of, 122; differential dye reduction
in, 124; other differentials in, 127; larval
differential inhibition in, 247; dedifferen-
tiation in, 301 ; cephaUc independence in
reconstitution of, 336, 339; bipolar forms
of, 368; twinning in, 537, 558; cleavage of,
547, 548; polar lobes of, 551; reconstitu-
tion in embryo and adult, 562; differentia-
tion without cleavage in, 596; forms of
development in, 597; oogenesis with ac-
cessory cells in, 665

Antennularia, gravity and polarity of, 422

Aplysia, developmental pattern of, 143

Arbacia: larval development of, 197; differ-
ential inhibition in, 199; differential toler-
ance, conditioning, and recovery in, 204;
effect of centrifugal force on development
of, 427; gradient pattern of, 739

Arenicola: embryonic differential suscepti-
biUty of , 120; cleavage of, 547, 553; oocyte
of, 658, 676; spiral aster in, 680

Armadillo: polyembryony in, 539, 690; mir-
ror-imaging in, 693

Arthropods: cleavage patterns of, 574; sper-
matozoa of, 622; oogenesis of, 662, 667

Ascaris: giant eggs of, 541 ; germ path in, 568;
cleavage pattern of, 570, 571; axiate pat-
tern of, 570, 572, 574, 681; centrifuged

801

802

PATTERNS AND PROBLEMS OF DEVELOPMENT

eggs of, 585; dispermic development of,
594; oogenesis of, 662

Ascidians: reconstitution in adult, 48; devel-
opmental pattern in, 145, 577, 580, 598;
differential inhibition in, 250; bilateral lar-
vae of, 253; buds of, 325, 635; bipolar
forms of, 369; induction in, 480; fusion of
first cleavage stages of, 541; cytoplasmic
movements in egg of, 577; organ-forming
regions in egg of, 577, 682; cleavage of,
577; embryonic reconstitution in, 577,
579; formative substances in, 579; pattern
in centrifuged eggs of, 587; pressure and
cleavage in, 592; dorsiventrality and
asymmetry in buds of, 635; dorsiventrality
in egg of, 681

Asymmetry: of chick embryo, 159, 532, 693;
experimental, of planarian heads, 193; ex-
perimental alteration of starfish, 219; in
inhibited fish embryos, 257; in amphibian
limb regeneration, 390, 395; compensatory
reversal of, 411, 412; and gasteropod
cleavage, 553, 680; of plant zoospores and
gametes, 601; spiral protozoan, 616; spiral
of plant spermatozoids, 618; of spermato-
zoa, 622; unicellular, in relation to po-
larity, 627; as secondary features of uni-
cellular patterns, 627, 629; in excystment
of Colpoda, 628; nature of unicellular,
629; of ascidian buds, 635; origin and
nature of embryonic, 672, 700, 701; in
echinoderms, 677; in relation to cleavage,
679; genetics of, 680, 691, 700, 701; of
vertebrates, 691; in reversed frog eggs,
692; in relation to optical isomeres, 696;
protein configuration as basis of, 697; as
concentration gradients, 700. See also
Dorsiventrahty; Symmetry; Ventrodor-
saHty

Auxin: in plant dominance, 309; direction of
transport of, 310; blocking of transport of,
311; in Fiiciis egg, 424

Avian embryo: asymmetry of, 159, 693; oxy-
gen uptake of, 159; difTerential dye reduc-
tion in, 159; differential susceptibiHty in,
162; differential inhibition in, 265; recon-
stitution in parts of, 528; potency fields
of, 530; origin of axiate pattern in, 688

Balancer: developmental field of, 289, 498;
induction of, 499

Beroe, egg cortex of, 563
Bildungszentrum: in insect egg, 515; sugges-
tions concerning, 520

Bryopsis, gradient in, 86, 89

Bryozoa: reconstitution in, 48; buds of, 325,

635; polyembryony in, 536; statoblast

pattern in, 635
Buds: as form of development, 14; gradients

and axes in, 16, 17, 18, 356; from plant

epidermal cells, 17, 18; of syllid anneUds,

20; amphibian limb, 20; "inverse" axes
from, 21; of hydroids, 104, 313; of bryo-
zoa, 325, 537, 635; of ascidians, 325, 635;
range of dominance in, 357; circumferen-
tial localization of, 431 ; in armadillo poly-
embryony, 539, 690; of Suctoria, 609, 610,
613, 614; in Noctiluca, 615; symmetry in,
633; asymmetry of ascidian, 636

Carbon dioxide production: methods of de-
termining or estimating, 60, 731; gradient
in Corymorpha, 98, 100; in Metridium, 106;
gradient in planarians, no, 731, 737; gra-
dient in Stylochus, 117, 731; gradient in
anneUds, 122, 731; in amphibian embryo,
153

Cell aggregates, origin of pattern in, 418, 419,
636, 640

Centrifugal force : and polarity of Fiicus eggs,
425; and patterns of animal eggs, 427, 428,
583; and amphibian development, 428

Cephalopod: differential developmental mod-
ification in, 248; cleavage of, 563

Cerebral uliis: early cleavage of, 546; embry-
onic reconstitution in, 555; alteration of
cleavage pattern in, 591, 593

Cerianthus: reconstitutional gradient in, 40;
ventrodorsal and radial pattern in, 675

Chaetopterus: embryonic differential suscep-
tibility of, 120; centrifugal force and po-
larity of, 427, 585; embryonic duplication
in, 559; differentiation without cleavage
in, 596

Chara: oospore and axiate pattern of, 613;
antherozoid development in, 618

Clavellina, bipolar reconstitution of, 369

Cleavage: of sea urchin, 133, 438, 589, 592;
differential modification of amphibian,
258; spiral, 544, 561, 591, 593, 594, 679;
cell homologies in, 552; bilaterahty and
spiral, 554; gradient in spiral, 561; of tur-
bellaria, 562; of rotifers, 563; of cteno-
phore, 564; of Ascaris, 570, 571; entomo-
stracan, 574; ascidian, 577; and centrifugal
force, 583; and mechanical pressure, 589;
alterations of spiral, 591; in Ca-free sea
water, 592; of dispermic and polyspermic
eggs, 593; differentiation without, 596

Clepsine { = Glossipho>ua): cleavage of, 548,

553; appearance of bilaterality in, 554
Coelom, differential modification of, in Pa-

tiria, 219
Colpoda, asymmetry in excystment of, 628
Conditioning, differential; in general, 72, 166;
in planarian, 114, 176, 195; in echinoderm
development, 204, 208
Corals, physiological dominance in, 319, 635
Corella: embryonic gradients of, 145; differ-
ential inhibition in, 250

INDEX

803

Corymorpha: reconstitutional gradient in, 38;
respiratory determinations in, 60; gradi-
ents in early stages of, 96; gradients in full-
grown, 98, 99, 100, 102; differential de-
velopmental modification in, 170; regres-
sion in, 170; reconstitution field in, 278,
279; dominance in reconstitution of, 316;
apical independence in reconstitution of,
Z^y, scale of organization in, 345, 357;
new patterns in reconstitution of, 359,
371, 414; multipolar forms of, 360, 416,
417; induction by implants in, 378; oxygen
and reconstitution of, 414; experimental
obliteration and determination of pattern
in, 416; experimental dorsiventrality in,
418, 693; cell aggregates of, 419; oocyte of,
659; origin of radial tentacle pattern in,
674

Crepidiila: early cleavage of, 549; effects of
centrifuging on, 584

Ctenophores: reconstitution in adult, 41;
transmission in plate row of, 56; differen-
tial susceptibility in, 106; functional domi-
nance and physiological isolation in, 327;
cleavage of, 563; egg cortex of, 563; egg
polarity of, 565; embryonic reconstitution
in, 566

Cumingia: axiate pattern in ultracentrifuged
eggs of, 427, 590; embryonic duplication
in, 558

Ciirtisia, reconstitution in, 45

Cyclopia: in planarians, 178; in squid em-
bryo, 248; in fish embryo, 256; in amphib-
ian embryo, 262; in chick embryo, 265;
in hybrid fishes and amphibia, 267; in ver-
tebrates in general, 270, 282; and lens de-
velopment, 489

Cytophore, and spermatozoan polarity, 624

Dedifferentiation: in adventitious plant
buds, 17; views concerning, 299, 301, 302;
of egg and sperm, 300; in alterations of
determination, 300; in epidermal plant
cells, 300; in protozoan fission and recon-
stitution, 300, 615, 618; in reconstitutions
in general, 301; of annelid neoblasts, 301;
in ascidian reconstitution, 302; in verte-
brate regenerations, 302; and modulation,
303; in lens regeneration, 396

Dendraster: larval development of, 197; dif-
ferential inhibition in, 202; secondary
modifications of, 205; exogastrulation of,
223; centrifuging and development of, 427

Dentalium: polar lobes of , 552, 56o;dispermic
eggs of, 594

Determination: nature of, 85, 291; relative
character of, 85, 291, 293; in relation to
gradient-level, 293; and dedifferentiation,
299; axial progress of, 332; in amphibian
presumptive neural region, 466; of amphib-
ian lens, 490; of amphibian ear, 496; in
relation to "double assurance," 500

Differentiation: and formative substances,
32, 292, 362, 437, 552, 558; independent
or self-, 212, i^y 336, 338, 3S7, 359, 559,
711, 713; concept of, 291, 294; cytoplasm
and nucleus in, 295; genes in relation to,
295; and molecular pattern, 296, 629; and
gradient pattern, 297; and chemical domi-
nance, 307;, and embryonic induction,
chap, xii; without cleavage, 596; proto-
zoan, 616; and unicellular asymmetries,
627; of spermatozoa, 630; chemo- or in-
visible, 712; of planarian parenchyma /;;
vitro, 714

Dileptus, dift'erential dye reduction in, 94
Dominance, physiological: as integrating fac-
tor, 8, 330, 432; transmissive and trans-
portative, 8, 304, 308, 330; in relation to
gradients, 11, 432; range of, 11, 305, 306,
344, 349, 351, 357; in fields, 278; in am-
phibian limb field, 286; of high region of
gradient, 304; in relation to rate of growth,
305, 324; gradient determination by, 307;
as inductor, 307; of plant vegetative tip,
308; blocking of, in plant, 308, 311; in
conifer, 30S; in relation to plant growth
forms, 309; role of auxins in, 309, 311; in
phyllotaxis, 311; of root systems and root
axis, 312; in lower plants, 313; in mush-
room reconstitution, 313; of adventitious
buds, 313; in Tubidaria, 314, 320; in Cory-
morpha, 316, 359, 371; and delayed sec-
tion, 316, 408; in branching hydroids, 319;
in corals, 319, 635; in planarians, 321; in
Stenostomum, 323; in annelids, 324; in
Halidystus reconstitution, 326; in seg-
ment formation, 327; in ctenophore plate
row, 327; reversal of functional, 328, 329,
330; in vertebrate heart, 328; in vertebrate
alimentary tract, 329; independent de-
velopment of, 2,2,2,, 2Z^, 357, 359, 362, 713;
and scale of organization, 344; induction
of regression by, 398; destruction of
Stenostomum zooids by altered, 399 ; effects
of other parts on, 406; in compensatory
asymmetry reversal, 411; by implanted
sea-urchin micromeres, 442; in amphibian
neural induction, 455; in homeogenetic
induction, 481; in avian induction, 483; in
embryonic sea-urchin fusions, 541; in
Schizocystis, 604; in Vorticella fission, 616;
in hydra budding, 635; in asymmetry, 701.
See also Amphibian dorsal inductor; In-
duction

Dorsiventrality: in Marchantia, 32; in ascidi-
an egg, 145, 577, 682; in lamprey embryo,
147; in teleost embryo, 149; in amphibian
embryo, 151, 684; experimental oblitera-
tion of, in amphibian, 259; determination
of, in amphibian limb, 286, 390; in amphib-
ian otic primordium, 289; in embryonic
reconstitution, 375; experimentally de-
termined in Corymorpha, 418; in relation

8o4

PATTERNS AND PROBLEMS OF DEVELOPMENT

to light in plants, 430; in relation to ex-
ternal factors in liverwort gemmae, 430;
in relation to cleavage, chap, xiv, 630,
679; of hypotrichous protozoa, 617; in
hydrozoan budding, 635; visible evidence
of, in animal eggs, 657; embryonic, 672;
origin and nature of, 675; in meroblastic
vertebrate embryos, 687; in mammalian
embryo, 689

Double assurance: definition of, 500; physio-
logical significance of, 501

Drosophila, difTerential ovarian dye reduction
in, 144, 671, 677

Dugesia, synonomy of, 41, footnote 7. See also
Planarians

Dyes: differential staining by, 63; toxic
effects of, 65, 91, III, 125, 732; differential
reduction and oxidation of, 67, 90, 94,
no, 119, 124, 133, 134, 137, 143, 144, 156,
159, 162, 563, 662, 671, 732, 740; local
staining by, 438, 447; induction by, 478

Ear: developmental field of amphibian, 288;
induction and determination of amphibi-
an, 496

Earthworm: physiological gradients in, 121,
1 23; cephalic regeneration and old nervous
system of, 339

Ectodermization: in echinoderm exogastrula,
227, 235, 239; by sodium thiocyanate, 243

Electric current, determination of pattern by,
421, 424

Electric potential : in relation to developmen-
tal pattern in general, 77; origin of, in or-
ganisms, 78; gradients in plants, 88; gra-
dient in sponge, 96; gradients in hydroids,
102; gradients in annelids, 127; gradients
in fish embryo, 151

Emancipation: of parts from whole, 713;
limitations of, 713, 714

Entelechy, 291, 645, 703

Entodermization: in echinoderm exogastrula,
221, 223, 225, 228, 232, 233, 238; by im-
planted sea-urchin micromeres, 440

Entomostraca: cleavage of, 574; centrifuged
eggs of, 586

Evocation: and individuation in induction,
475; as activation, 479

Exogastrula: of echinoderms as differential
modification, 221, 222, 233; agents pro-
ducing, 221, 232; entodermization in echi-
noderm, 221, 223, 225, 228, 232, 233, 238;
forms of echinoid, 223, 226; radial forms
of, 225, 228, 231; ectodermization in echi-
noderm, 227, 235, 239; development of
echinoderm, 227; of Patiria, 231; recon-
stitution in echinoderm, 236, 241; scale of
organization in echinoderm, 237; bipolar
forms of, 244, 443; amphibian, 463, 464;
annelid, 559

Explantation: definition of, 83; of parts of
amphibian embryo, 462, 471, 526; of parts
of chick embryo, 533, 535

Eye: in relation to field, 282, 531; induction
of parts of, 495; reconstitution in embry-
onic amphibian, 527; chick potency field
of, 531, 534

Field, developmental: value and implication
of concept of, 276, 281; relation of gra-
dients to, 277; potency and differentiation
in, 277, 288; as gradient system, 278; in
reconstitution, 278; dominance in, 278; in
Corymorpha, 279; relation of metabolism
to, 279; of vertebrate eye, 282, 531, 534; of
amphibian limb, 285, 293; of amphibian
ear, 288; of balancer, 289; boundaries of,
290; of host in relation to induction, 474,
476, 483; of different parts in chick em-
bryo, 528, 531, 532, 533, 534

Fir, electric-potential gradients in, 78

Fishes: embryonic differential susceptibility
of, 147, 149, 256; reversible heart beat in,
329; formal developmental pattern of tele-
ost, 450; embryonic cell movements in
teleost, 451; induction in, 481; extra-
embryonic inductors in, 482; reconstitu-
tion of isolated blastomeres of, 521; em-
bryonic dorsiventrality of, 688

Fission: in general, 13; in unicellular forms,
23, 24, 25; in metazoa, 26; in planarian,
26, 43, 321; in Stenosiomum, 27, 323, 399;
in annelids, 324; in Paramecium, 324; in
Fonticola, 325; regression of zones of, 398;
in Sporozoa, 604; asymmetry in protozo-
an, 617

Fonticola: reconstitution in, 45; fragmenta-
tion in, 325

Formative cells: in relation to differentiation
and dedifterentiation, 295, 301, 340; neo-
blasts as, 301

Formative substances: in Acetabitlaria, 32,
362; do they exist? 292, 597; in echinoderm
development, 437; in spiral cleavage, 552,
558, 561; in ascidian development, 579

Fragmentation: in planarians, 13, 325; in ne-
merteans and annelids, 13, 26, 326

Fiicus, egg and early stage: original polarity
of, 86, 423; gradients in, 87; experimental
determination of polarity in, 423

Fiindulus: embryonic differential susceptibili-
ty of, 149; embryonic differential inhibi-
tion in, 256; embryonic reconstitution of,
521, 523

Fusion: of sponge larvae, 540; of planulae,
540; of nemertean cleavage stages, 540; of
A scar is eggs, 541; in sea-urchin develop-
ment, 541; of ascidian cleavage stages,
541 ; of amphibian two-cell stages, 542

INDEX

805

Gadus, embryonic differential susceptibility

of, 149
Gastrulation: in echinoderms, 137, 236, 440,
443, 445; in Phialidiiim, 167; in differen-
tially inhibited amphibian embryos, 259;
induced by implanted sea-urchin micro-
meres, 440; regional migrations in verte-
brate, 447, 449, 451, 453; and induction in
amphibians, 454
Glycogen, in amphibian embryo, 156, 477
Glycolysis: and susceptibility, 76; in am-
phibian embryo, 154; in amphibian induc-
tion, 477
Gradients, concentration: in echinoderm de-
velopment, 129, 140, 227, 241, 243, 436,
507, 508, 509, 513; action of external
agents on, 227, 242, 243, 245; in relation
to activity gradients, 241, 272, 513; in re-
lation to protein configuration, 697; as
basis of symmetry and asymmetry, 700;
as results of physiological pattern, 702
Gradients, physiological: as features of de-
velopmental pattern, 7, 270, 272, 432, 485,
700, 703; of buds, 16; reconstitutional dif-
ferentials as evidence of, 31; embryonic
and functional differentials as evidence of,
53; in mammalian intestine, 56; experi-
mental methods of demonstrating and in-
dicating, chap, iii, 723; of algae, 86; of
higher plants, 88; of Paramecium, 90; of
other protozoa, 94; of sponges, 96; of hy-
droids, 96; of a siphonophore, 104; of
scyphozoa and actinozoa, 105; of cteno-
phores, 108; of planarians, 108; of poly-
clads, 117; of annelids, 119; in echinoderm
development, 129, chap, vi, 436, 723; in
beetle embryo, 144; in Drosophila ovary,
144; in ascidian embryo, 145, 250; of
lamprey embryo, 147; of teleost embryo,
150; of embryonic vertebrate heart, 150,
162; in amphibian development, 151, 163,
256, 258; of chick embryo, 159, 162, 265;
differential modifications of, chaps, v, vi,
vii; mediolateral, in planarian head, 180,
190, 724; experimental obliteration of,
168, 201, 202, 208, 215, 216, 225, 228, 231,
259, 416, 539; in insect development,
250, 518; experimental determination of,
273, 413; origin of specific differences
from, 274, 297; as co-ordinate system, 274;
relation of, to field, 277; in relation to
metabolism, 280; in amphibian limb field,
285, 293; in harmonious-equipotential
system, 290; dominance of high regions of,
304; localization of central nervous system
in high regions of, 305; auxin transport
and, 310; progressive extension of, 332;
in amphibian lens regeneration, 396; in
relation to implanted sea-urchin micro-
meres, 440, 444; in relation to embryonic
cell movements, 450, 454; in amphibian
dorsal inductor, 459, 460; in amphibian

presumptive neural region, 466 ; of amphib-
ian induced embryo, 468; in effects of
foreign inductors, 473; in amphibian lens
induction, 490; in otic induction, 497, 498;
and "double assurance," 501; and induc-
tion in general, 502; in embryonic sea-
urchin reconstitution, 506, 508, 509; in
polyembryony, 537, 538, 539; in relation
to egg cortex, 588; in Polys pondyliumfiig;
of diatomaceous pseudothallus, 641; in re-
lation to embryonic symmetry and asym-
metry, 673; in relation to protein con-
figuration, 697; hypotheses concerning,
700; most primitive type of organismic,
702; in relation to organism as a whole,
703; and "machines" of Driesch, 703; ini-
tiation and character of, 704, 727; changes
of, in evolution of form, 723; in relation to
genetic change, 727. See also Condition-
ing; Gradients, concentration; Gradients,
respiratory; Inhibition; Pattern, develop-
mental; Pattern, ultrastructural; Recon-
stitution; Recovery, differential; Tolerance
Gradients, respiratory: of Corymorpha, 98; of
Tubularia, 99; of Obelia, 99; of planarians,
108; of Stylochus, 117; of annelids, 121; of
amphibian embryo, 153; of chick embryo,
159; other data and discussion concerning,
729. See also Gradients, physiological

Gravity: in relation to plant pattern, 422,
430; and Antenmdaria polarity, 422; and
amphibian development, 428; and dorsi-
ventrality of liverwort gemmae, 430

Griffithsia: polarity determined by electric
current in, 421; polarity determined by
centrifuging in, 422

Growth: allometric, 55, 723; definitions of,
715; negative, 716; in relation to develop-
ment, 716; differential, 717; in earlier and
later stages, 717; species-specific, 718;
gradients of, 720, 724; inorganic, 722;
mathematical analyses of, 722; in evolu-
tion of form, 724

Guinea pig, inhibited head forms of, 268

Haliclysius: reconstitution of, 40, 50, 319,
326; biaxiate forms of, 335

Harenactis: reconstitution gradient in, 40; bi-
polar forms of, 41; symmetry in reconsti-
tution of, 336; new axiate pattern in rings
of, 372; mirror-imaging in, 387

Harmonious-equipotential system: concept
of, 290; early sea-urchin embryo as, 505

Head frequencies: in planarians, 181; factors
determining, 194, 400, 406, 408

Heart: gradient in embryonic, 150, 162, 163;
differential inhibition of embryonic fish,
257; differential inhibition of embryonic
chick, 266; dominance and physiological
isolation in, 328; reversal of dominance

8o6

PATTERNS AND PROBLEMS OF DEVELOPMENT

in, 329; developmental field of, in chick
embryo, 531

Heterochely, compensatory reversal of, 412

Heteromorphosis: axial, in Tubtdaria, 315;
of reconstituting hydranth or head, 335

Homeosis, in arthropods, 342

Hydra: tentacle reconstitution in, 36; gradi-
ents in, loi; budding in, 313, 431; induc-
tion by implants in, 378; bud symmetry
in, 634, 721

Hydro ides, compensatory reversal in, 41 1

Ilyanassa: polar lobes of, 559, 560; centri-
fuged eggs of, 585

Individuation, in amphibian induction, 475

Indophenol blue reaction: as method, 64; in
Paramecium, 92; in echinoderm develop-
ment, 139, 140

Induction: in general, 9, 485, 502; of amphib-
ian limb, 288; by dominant region, 307,
358; by lacerated partial section, 371; by
grafts in coelenterates, 378; by grafts in
planarians, 381; of regression by altered
dominance, 398; by implanted sea-urchin
micromeres, 440, 443 ; of amphibian neural
plate, 455, 459, 460, 462, 469, 485; ques-
tions of specificity of substance and effect
in, 461, 472, 473, 483, 485; amphibian neu-
ral development without, 462, 466; ex-
amples of homeogenetic, 469; by living
and dead amphibian tissues, 470; by tis-
sues and extracts in general, 471, 483;
effectiveness of, 472, 473; of amphibian
lens, 472, 488, 491 ; as activation, 472, 475,
478, 482, 485, 48^1, 496, 502, 503; influence
of host field in, 474, 483; in relation to or-
ganization, 475, 485; evocation and indi-
viduation in, 475; by chemical substances,
476, 478; glycogen and, 477; metabolism
and, 477; by methylene blue, 478; by fatty
acids, 479; in relation to host cytolysis,
479; in ascidians, 480; in teleosts, 481; in
birds, 483 ; by entoderm, 484; in mammals,
485; in normal development, 486; of chick
lens, 495; of optic parts, 495; of amphibian
ear, 496, 498; of otic capsule, 497; of tym-
panic membrane, 498; of amphibian bal-
ancer, 499; xenoplastic, of amphibian oral
region, 499; of amphibian gill, 500; and
"double assurance," 500. 6"ee a/^o Amphib-
ian dorsal inductor; Dominance, physio-
logical; Organizer; Pattern, developmen-
tal
Inductor, definition of, 9. See also Amphibian
dorsal inductor; Dominance, physiologi-
cal; Induction; Organizer; Pattern, devel-
opmental
Inhibition, differential: in general, 72, 166;
in Phialidiiim, 167, 425; in Corymorpha,
170; inplanarian, 175, 177; in echinoid de-
velopment, 199; in asteroid development,

212; echinoderm exogastrula as, 235; in
experiments of Herbst, 240; in squid em-
bryo, 248; in insect embryo, 250; in ascidi-
an embryo, 250; in fish embryo, 256; in
frog development, 258; of amphibian noto-
chord, 264; in chick embryo, 265; by in-
jury of spermatozoa, 266; in hybrids, 267;
in inbred guinea pigs, 268; in inbred mice,
269; in ctenophore plate row, 327; of
heart, 329. See also Conditioning; Recov-
ery; Tolerance; Susceptibility, differential

Insects : differential ovarian dye reduction in ,
144, 671, 677; differential developmental
modification in, 250, 518; embryonic re-
constitution in, 514, 517, 518; Bildiings-
zentrum of, 515, 519, 520; differentiation
center of, 517, 521; developmental gradi-
ents in, 518; polyembryony in, 537; cen-
trifuged eggs of, 587; oogenesis in, 668

Integration, physiological: in development in
general, i, 706; factors of, 8, 304, 330, 706,
709, 710. See also Dominance, physiologi-
cal; Induction; Pattern, developmental;
Reconstitution

Isolation, physiological: in general, 11, 82;
factors and results of, 306; of plant vege-
tative tips, 311; in Tubidaria, 314; in rela-
tion to delayed section, 316, 319; in corals,
320; in planarian, 321; in Stenosfomum,
323; in annelids, 324; in segment forma-
tion, 327; functional, in ctenophore plate
row, 327; in heart, 328; in apical and
cephalic reconstitution, 333, 338, 357, 359;
in development of Polys pondylium, 640

Lamprey: embryonic differential suscepti-
biUty of, 147; reconstitution of isolated
blastomeres of, 521

Lens: regeneration of amphibian, 395; induc-
tion of amphibian, 472, 488, 491; develop-
ment of amphibian, 487; specificity in in-
duction of, 492; polarity of, 494; free de-
velopment of teleost, 494; induction of
avian, 495

Lepto plana: reconstitution of, 46; dye-reduc-
tion gradient of, 119

Light: in plant reconstitution, 420; inhydroid
reconstitution, 420; polarity determina-
tion in Fhcus egg by, 424; and localization
of parts in plants, 430; and pattern of
liverwort gemmae, 430

Lithium, action of: on hydrozoan develop-
ment, 167; on echinoid development, 203,
206, 208; on asteroid development, 212,
216, 219; in echinoderm exogastrulation,
221, 223, 226, 227, 229, 230; locally spe-
cific or differential, 227, 233, 234, 235, 241,
243, 263; on frog development, 258, 263;
on amphibian notochord, 264; on scale of
sea-urchin organization, 357, 440, 445, 507

Loligo, difl'erential inhibition in, 248

INDEX

807

Liimhriciilus: oxygen uptake of, 121; differ-
ential susceptibility of, 126; differential in-
hibition in, 247, 368; fragmentation in,
326; reconstitution from partial trans-
verse section in, 337

Lytechinus, experimental alteration of polar-
ity in, 376

Macromeres: of sea urchin, 438; in spiral
cleavage, 545; of ctenophores, 564

Mammals: intestinal gradient of, 56, 164,
329; developmental pattern of, 163, 689;
polyembryony in, 539, 690

Maps, of presumptive regions and cell move-
ments and of potencies: amphibian, 447,
448, 449; teleost, 451; chick, 452, 453, 454;
of chick developmental potencies, 531, 532,
533. 534

Marchantia: reconstitution of, 32; external
factors and gemma pattern in, 430

Metabolism: methods of determining respir-
atory, 58; vital dyes in relation to, 67;
and differential susceptibility, 70, 736; and
electric potential, 78; and concentration
gradients, 164, 241, 272, 432, 513, 702,
704; and developmental pattern, 280, 281,
702, 704; and formative substances, 292;
in differentiation of gradient pattern, 297;
in amphibian induction, 472, 477; and
ultrastructural pattern, 696, 697, 698;
and substrate, 705. See also Carbon di-
oxide production; Gradients, physiologi-
cal; Gradients, respiratory; Oxygen up-
take; Pattern, developmiental; Suscepti-
bility, differential

Micromeres: formation of sea-urchin, 438; in-
duction by implanted sea urchin, 440, 443 ;
in spiral cleavage, 545; of ctenophores, 564,
566

Mirror-imaging: in bipolar forms, 387; in Ha-
renactis, 387; in Brnchdreifachbildungen,
388; in reduplicated amphibian limbs, 390;
in armadillo quadruplets, 693

Mollusks: developmental pattern of, 143; lar-
val differential inhibition in, 248; cleav-
age of, 549, 564; polar lobes of, 551; cleav-
age and asymmetry of, 553, 680; embryon-
ic duplication in, 558; centrifuged eggs of,
584; genetics of sinistrality in, 680

Mosaic development: and independent dif-
ferentiation, 212, ^1,2,, 336, 338, 357, 359,
544, 711, 713; and cleavage patterns, chap,
xiv; relative character of, 712; in relation
to dominant region, 713

Nemerteans: reconstitution in, 47, 355, 555;
fragmentation in, 326; fusions of cleavage
stages of, 540; early cleavage of, 546

Nereis: oxygen uptake in, 121; CO2 produc-
tion in, 122, 731; embryonic duplication
in, 558

Nervous system : in physiological dominance,
8, 304, 308, 320, 330, 709; gradients and
localization of, 305; in coelenterate domi-
nance, 320; new dominant region inde-
pendent of, 338, 339; and planarian head
regeneration, 340, 407; in starfish-arm re-
generation, 340; in statoblast develop-
ment, 341; in amphibian limb regenera-
tion, 341; and in vitro differentiation of
planarian parenchyma, 714

Obelia: reconstitution gradient in, 40, 99;
electric-potential gradient in, 78, 102;
electric current and polarity of, 421

Oogenesis: polar gradient in echinoderm, 130;
in relation to intraorganismic environ-
ment, 658; pedunculate, 660; of starfish,
661; of Ascaris, 662; in chordates, 662,
663; in arthropods, 662, 667; follicular
types of, 664; with nutritive cells, 665,
667; and symmetry patterns, 676; and
protein configuration, 697

Organizers: definitions of, 9, 455; dominant
regions as, 307, 344, 358, chap, xi; and
sca'e of organization, 343; in amphibian
inductions, chap, xii from p. 454; in anne-
lid development, 561. See also Amphibian
dorsal inductor; Dominance, physiologi-
cal; Induction; Inductor

Oxidation-reduction, indicators of, 63, 67

Oxygen uptake: methods of determining, 58,
729; gradients in flowers, 89; of Parameci-
um and KCN, 94, 736; gradient in
sponges, 96; gradient in Corymorpka, 98;
gradient in Tubularia and Obelia, 99; gra-
dient in planarians, 108, 730; gradient in
annelids, 121; in amphibian embryo, 153;
in chick embryo, 159; of Thysanozoon, 730

Paramecium: multiple monsters of, 36; dif-
ferential dye reduction in, 90; indophenol
blue reaction in, 92; differential suscepti-
bility of, 92; KCN and o.xygen uptake of,
94, 736

Patiria. See Starfish

Pattern, developmental: types of, 2, 12, 597;
protoplasmic and organismic, 3; surface-
interior, 4, 606, 615, 646; axiate, 5; in re-
lation to ultrastructural pattern, 5, 281,
296, T,ii, 389, 394, 629, 631, 694; as in-
herent in protoplasm, 6; as a reaction to
external factors, 6, chaps, xi, xv, xvi, 646;
Roux-Weismann hypothesis of, 6; genes
in relation to, 6, 29, 645; gradients as
features of, 7, chap, ii, 104, 270, 272, 297,
311, 332, 396, 702, 703; of buds, 16, 17, 18,
285, 356, 633, 634; bipolar and multipolar,

22, 23, 169, 309, 316, 319, 325; and fission,

23, 26, 27, 43, 321, 323, 398, 604, 617; re-
constitutional differentials as evidence
of, chap, ii; and formative substances, 32,

8o8 PATTERNS AND PROBLEMS OF DEVELOPMENT

292, 362, 437, 552, 558, S6i, 579> 597; visi-
ble evidence of, in embryonic develop-
ment, 53; experimental methods of analy-
sis of, chap, iii; differential modification
of, 72, chaps, V, vi, vii, 723; experimental
determination of Fucus, 86, 423 ; of echino-
derms, 129, chap. vi. 376, 583, 677, 739; of
moUusks, 143, 549, S5i; of ascidians, 145,
577, 580; of lamprey, 147; of teleost fishes,
149, 451, 452; features of amphibian, 151;
avian, 159, 43 i, 432, 433, 434, 688; of
mammalian alimentary tract, 164, 329;
experimental obliteration of, 168, 201,
202, 208, 215, 216, 225, 228, 231, 259, 540;
regression of, 170, 398, 399; and metabo-
lism, 280, 281, 446, 477, 702, 704; of am-
phibian limb, 285, 370, 390; of harmoni-
ous-equipotential systems, 290, 505; dy-
namic character of, 298; dominance as
feature of, chaps, ix, x, xi, 713; progressive
determination of, 332; scale of organiza-
tion of, 342, 644; initiation of, in recon-
stitution, 359, 371, 376, 390; of appendage
reconstitution, 369; induction of, m coe-
lenterates, 371, 378; mirror-imaging of,
387, 693; in relation to o.xygen, 413; in
cell aggregates, 418, 419, 636; determina-
tion of, by electric current, 421; in rela-
tion to gravity, 422; maps of, amphibian,
447, 448, 449, teleost, 451, chick, 452, 453,
454, 531, 532, 533, 534; effect of tempera-
ture gradient on, 463; in amphibian exo-
gastrula, 463, 464, 465; of amphibian host
and implanted inductor, 467; as organizer,
485; and neural inductor, 485; origin of
lens, 494; and double assurance, 500; and
induction in general, 502; of insects, 515;
in polyembryony, 536, 537, 539, 690; i"
armadillo quadruplets, 539, 690, 693; and
spiral cleavage, 544, 554, 562; polar lobes
in relation to, 551, 559; of ctenophore,
563; of Ascaris, 568, 572; of entomostra-
cans, 574; of Amphioxus, 582; in relation
to egg cortex, 588; effect of pressure on,
589; of dispermic and polyspermic eggs,
593; origin of, in plant spores, 599; of
sporozoa, 604, 605, 606, 607; suctorian,
609, 610, 613; of Noctiliica swarm spores,
615; asymmetrical protozoan, 617, 627;
of plant spermatozoids, 618; of animal
spermatozoa, 622 ; of bryozoan statoblasts,
63s; of Polys pondylium, 637; of diatoma-
ceous pseudothallus, 641 ; of pileus in fun-
gi, 641; of bryophytes, 647; of pterido-
phytes, 647; of gymnosperms, 651; of an-
giosperms, 652; of animal oocytes and
eggs, 657; symmetry and asymmetry in
embryonic, 672, 675, 677, 679; genetics of
asymmetric, 680, 691, 700, 701; and
crystalline pattern, 695; and optical iso-
meres, 696; and protein molecular con-
figuration, 697; of flowing stream, 699; as
primarily gradient pattern, 700, 703; most

primitive form of axiate, 702; fundamental
identities of, 703; and "inachines" of
Driesch, 703; initiation and reaction in,
704; susceptibility of, 705; integrating
factors of, 706; in relation to growth, 716;
in evolution of form, 724. See also Asym-
metry; Buds; Conditioning, differential;
Dominance, physiological; Dorsiventrali-
ty; Field, developmental; Gradients, con-
centration; Gradients, physiological; In-
duction; Inhibition, differential; Isolation,
physiological; Metabolism; Organizers;
Pattern, ultrastructural; Reconstitution;
Recovery, differential; Scale of organiza-
tion; Susceptibility, differential; Symme-
try; Tolerance, differential; Ventrodor-
sality
Pattern, ultrastructural: in relation to de-
velopmental pattern, 5, 281, 296, 333, 389,
389, 394, 629, 631, 694, 701, 702; as a
reaction, 629; in relation to differentiation,
630; of cell walls and surfaces, 631; in
protoplasmic products and structures,
694; crystalline, 695, 698; of protein con-
figuration, 697; and specific asymmetries,
699; of flowing stream, 699; and growth of
nerve fibers, 708
Pennaria: gradient pattern in, 38, 104; elec-
tric-potential difference in, 103
Permeability, differential, 76
Pheretima, physiological gradient of, 122
Phialidium: gradients in egg and early stages
of, 96; differential developmental modifi-
cation in, 167; obliteration and determina-
tion of polarity in, 425; oocytes of, 659
Pileus, axiate pattern in, 642
Planarians: fission in, 41, 321; reconstitution
and body-level in, 108, 368, 729; differen-
tial inhibition in, 175, 177, 181, 183, 190,
193; differential conditioning and recovery
in, 176, 190, 195; asymmetry of head in,
193; conditions for head reconstitution
in, 193, 400, 406, 408; reconstitution field
in' 278; dedifferentiation in, 301; bipolar
and multipolar forms of, 303, 366, 367;
original nervous system and head regener-
ation in, 339; scale of organization in re-
constitution of, 349, 357; reconstitution in
acephalic forms of, 362; apolar forms of,
364; reconstitution of lateral pieces of, 365 ;
induction by grafts in, 381 ; delayed section
and head regeneration in, 408
Pliimularia, oxygen and reconstitution of, 413
Polar lobes: of annelids and moUusks, 551;
in relation to developmental pattern, 559;
effect of removal of, 559
Polarity, physiological: as a serial order, 4;
progressive determination of, 332; of nu-
cleus, 645. See also Dominance, physio-
logical; Gradients, physiological; Pattern,

INDEX

809

developmental; Reconstitution; and other
subjects designated under these heads

Polyembryony: in bryozoa, 536; in insects,
537; in vertebrates in general, 538; in ar-
madillo, 539, 690; in spermatophytes, 654.
For experimental forms of, see Reconstitu-
tion

Polyspondylium, development of axiate pat-
tern in, 637

Potato, axiate pattern in tuber of, 88

Potencies, developmental: definition of, 81;
methods of analyzing, 81; and differentia-
tion in field, 277, 288; maps of chick, 531,
532, 533, 534; in relation to spiral cleav-
age, 555; of ctenophore egg and embryo,
565; of ascidian egg and embryo, 577. See
also chaps, v, vi, vii, xi, xii, xiii; Pattern,
developmental; Reconstitution

Potentiality, definition of, 81

Pseudo-exogastrula, of echinoderms, 232

Pseudothallus, diatomaceous, 641

Rana: embryonic differential susceptibility
of, 151, 743; embryonic metabolism of,
153; lens induction in, 488; otic induction
in, 496; gray crescent of, 657, 684; circu-
lation and egg axiate pattern of, 663; em-
bryonic dorsiventrality of, 684, 686; em-
bryonic asymmetry of, 692

Reconstitution: definition of, 28; types of, 30;
in Acetabularia, 31; in Marchantia, 32; in
relation to body-level, 34, 36, 41, 47; and
degree of injury, 49; after oblique section,
50, 326; differential modification of hy-
droid, 172; differential modification of
planarian, 175; dedifferentiation in, 301;
dominance and physiological isolation in
hydroid, 314; delayed section in, 316,
408; apical and cephalic independence
in, 333, 335, 33^, 339, 358, 359, 362;^ in
Lumbriculus, 337; scale of organization
in hydroid, 344; scale of organization
in planarian, 349; scale of organization
in annelid, 355; new patterns in, 359,
371-372, 376, 378, 381; bipolar and
multipolar, 359, 364; pattern in early
embryonic, 374; in embryo and adult, 376;
induction of, by grafts, 378, 381; mirror-
imaging in, 387; of amphibian lens, 395;
inhibition of, by altered dominance, 399;
in relation to oxygen, 413, 420; induced by
implanted sea-urchin micromeres, 440;
of amphibian dorsal inductor, 462; of
extraembryonic inductors in teleosts, 482;
of avian inductor, 483; of hydroid blasto-
meres, 504; of parts of sea-urchin embryo,
505; embryonic ventrodorsal in sea urchin,
512; in insect embryos, 514; of isolated
fish blastomeres, 521 ; of isolated amphibi-
an embryonic parts, 523, 526; absence of,
in amphibian tail bud, 527; of isolated

parts of chick embryo, 528; in polyembry-
ony, 536; in embryonic and larval fusions,
540; embryonic, in forms with spiral
cleavage, 555, 559; embryonic, in cteno-
phores, 566; after blastomere dislocation,
589; embryonic, after Ca-free sea water,
592; in relation to ultrastructure, 698

Recovery, differential: in general, 74, 166; in
planarian reconstitution, 190; after differ-
ential inhibition in sea urchin, 204, 507;
after differential inhibition in Patiria, 216,
221 ; in echinoderm exogastrulae, 230, 235,
238; in "animalized" Dendraster larvae,
245; in amphibian embryo, 262; in cteno-
phore plate row, 328. See also Condition-
ing, differential; Inhibition, differential;
Tolerance, differential

Redifferentiation, definition of, 30. See also
Dedifferentiation; Differentiation; Recon-
stitution

Reduction, differential: of KMn04, 63, 64,
96, 145; of vital dyes, 67, 90, 94, no, 119,
124, 133, 134, 137, i43> 144, 156, 159, 162

Regeneration, definition of, 31. See also Re-
constitution

Regression: in hydroids, 170, 398; of fission
zones, 398. See also Dedifferentiation

Respiration: methods of determining, and
complicating factors, 58, 729, 731; differ-
ential susceptibility in relation to, 75, 736;
significance of planarian data on, 730, 731.
See also Carbon dioxide production; Gra-
dients, respiratory; Metabolism; Oxygen
uptake; Susceptibility, differential

Sabella, reorganization in reconstitution of,
355- 368

Scale of organization: in planarian reconsti-
tution, 45, 349; in echinoderm exogastrula,
237; in "animalized" and "Vegetalized"
echinoderm larvae, 243; as spatial order
of magnitude of pattern, 343; in relation
to dominance, 344; in Tubularia reconsti-
tution, 344; in Corymorpha reconstitution,
345; in nemertean and annelid reconstitu-
tion, 355, 555; alteration of, by lithium,
357; in embryonic sea-urchin reconstitu-
tion, 357, 440, 445, 507, 513; and localiza-
tion of parts, 358

Scyphozoa. See Haliclystus

Sea urchin : physiological gradients in egg and
early development of, 130, 133, 134, 136,
137; differential developmental modifica-
tion in, chap, vi, 439, 506; apical partial
forms of, 357, 439; alteration of polarity
of, 376; cleavage of, 438; embryonic re-
constitution in, 438, 505, 508, 509, 511,
512, 587, 591; induction by implanted mi-
cromeres in, 440; early embryo of, as har-
monious-equipotential system, 505; re-
constitution of combined blastomeres in,

8io

PATTERNS AND PROBLEMS OF DEVELOPMENT

511; alteration of ventrodorsality in, 512;
fusions in development of, 541; centri-
fugedeggsof, 587; pressure, cleavage, and
development of, 590; development and
Ca-f ree sea water, 59 1 ; dispermy and poly-
spermy in, 593 ; symmetry and asymmetry
of, 677; embryonic skeleton-forming ma-
terial of, 677
Siphonophore, physiological gradients in, 104
Sodium thiocyanate, action of, on echinoderm
development, 243, 244

Spermatozoa: developmental patterns of,
621; origin of polarity in, 624, 625, 627;
differentiation and molecular pattern in,
630
Spermatozoids of plants, developmental pat-
terns of, 618
Spina bifida: as differential inhibition, 261;

in chick development, 266
Spirostomum, differential dye reduction and

susceptibility in, 94
Spirotrichonvnipha, spiral asymmetry in, 617,

628
Sponges: physiological gradients in, 96; cell
and larval aggregates of, 418; origin of
pattern in gemmules of, 636
Spores: as material for developmental physi-
ology, 12; origins of pattern in plant, 599;
asymmetry of plant, 601; patterns of cni-
dosporidian, 607; of Noctiluca, 615
Sporozoa, origins of pattern in, 604, 605, 606,

607
Starfish: physiological gradient in oocyte,
egg and early stages of, 130, 134, 739; dif-
ferential inhibition in, 212, 219; differen-
tial conditioning and recovery in, 216; dif-
ferential modification of coelom in, 219;
exogastrulation in, 230; egg section and
pattern in, 376; polar body position in,
661; symmetry and asymmetry in, 677
Stenostomum: fission in, 27, 323; differential
dye reduction and susceptibiUty in, 117;
destruction of zooids in, 399
Stolons: in differentially inhibited planulae,
169, 426; replacement of hydranths by,
172; of Tubularia, 314; of Corymorpha,
316; experimental localization of, in Cory-
morpha, 414
Strongylocentrotiis: differential developmental
modification in, 202; exogastrulation in,
226; egg section and pattern in, 376. See
also Sea urchin
Stylochus, physiological gradients in, 117
Substitution: definition of, 30; in plants, 32.
See also Reconstitution

Suctoria: budding and pattern of, 609, 610,
613; "embryos" or larvae of, 609, 610, 613;
larval and parental pattern in, 613

Sulphydryl : gradient in plant vegetative tips,
89; in amphibian embryo, 157

Susceptibility, differential: aspects and sig-
nificance of, 70, 166; patterns of, 71, 735,
737, 743; of algae, 86; of Amoeba, 90; of
Paramecium, 92; of other ciliate protozoa,
94; of Phialidium, 96, 166; of other hy-
droids, 103, 104, 170, 172; of a siphono-
phore, 104; of ctenophores, 106; of planar-
ians, 112, 175; of polychete annelids, 120;
of oligochete annelids, 121, 126; of echino-
derm eggs and embryos, 131, 134, 140,
chap, vi, 739, 742; of mollusk embryos,
248; of insect embryo, 144, 250, 518; of
lamprey embryo, 147; of teleost embryos,
149, 256; of frog embryo, 152, 258, 743;
of chick embryo, 162, 265; Ranzi's inter-
pretation of, 249; of ascidian embryo,
250; in vertebrate development, 254, 394,
743; of amphibian dorsal lip region, 263;
of mammalian embryo, 266; in diatoma-
ceous pseudothallus, 641. See also Condi-
tioning, differential; Gradients, physi-
ological; Inhibition, differential; Pattern,
developmental; Recovery, differential;
Tolerance, differential

Symmetry: radial in adventitious plant buds,
18; in buds in general, 20, 104, 633; of a
siphonophore, 104; radial, in differentially
inhibited echinoderm larvae, 208, 215;
radial in echinoderm exogastrulae, 225,
228, 231; bilateral of amphibian and chick
eye field, 282, 531, 534; in Harena^tis re-
constitution, 336; in reconstitution gener-
ally, 338; new radial, 359, 370; in relation
to light in plants, 430; in relation to cleav-
age, chap, xiv, 679; origin and nature of
embryonic, 672; different origins of radial,
674; patterns of echinoderm, 677; protein
configuration as basis of, 697; concentra-
tion gradients in, 700. See also Asym-
metry; Dorsiventrality; Pattern, develop-
mental; Ventrodorsality

Taittogolahnts, embryonic differential suscep-
tibility in, 149

Tolerance, differential: in general, 72, 166;
in planarians, 114; in Chaetoplerus, 120;
in echinoderm development, 204, 218. See
also Conditioning, differential

Triton: differential dye reduction in embryos
of, 156; ligatured eggs of, 266, 524; neural
induction in, 455; lens development in,
489, 492; otic induction in, 496; branchial
induction in, 500. See also Amphibian de-
velopment; Amphibian dorsal inductor

Triliirus: differential dye reduction in em-
bryos of, 157; temperature gradient and
development of, 463 ; lens induction in, 492

Tubifex: embryonic differential susceptibility

INDEX

8ii

of, 121 ; differential dye reduction in, 124;
early cleavage of, 548; polar plasms of,
550, 557; embryonic reconstitution in,
556; centrifuged eggs of, 585
Tubularia: reconstitution in, 36; respiratory
gradient in, 60; dominance and physiologi-
cal isolation in, 314; so-called "axial
heteromorphosis" in, 315; blocking of
dominance in, 320; apical independence in
reconstitution of, ^t,7,\ scale of organiza-
tion in, 344, 357; new patterns in reconsti-
tution of, 359; oxygen and reconstitution
of, 413; pattern and electric current in, 421

Valonia: pattern of cell wall in, 631; polarity
of, 632

Vegetalization, in echinoderm development,
243

Ventrodorsality: in planarian reconstitution,
44; of echinoderm embryos, 134, 137; of
Drosophila egg, 145, 671, 677; experi-
mental obliteration of, 201, 208, 215,
216, 225, 228, 231; in embryonic reconsti-
tution, 375, 377, 512; in centrifuged
eggs, 427, 585; spiral cleavage and,
554, 562, 585; Ascaris cleavage and, 571;
entomostracan cleavage and, 574; ascidian
cleavage and, 577; in animal eggs in gen-
eral, 657, 679; embryonic, in general, 672;
of Cerianthus, 675; origin and nature of,
675, 677. See also Dorsiventrality; Sym-
metry

[PRINTED 1[
IN USA J