1. "
Ilaajs
tit
Jhr^^
i^^TRFFl
CAMBRIDGE COMPARATIVE PHYSIOLOGY
General Editors :
J. BARCROFT, C.B.E., M.A., F.R.S.
Fellow of King's College and Professor of
Physiology in the University of Cambridge
and
J. T. SAUNDERS, M.A.
Fellow of Christ's College and Lecturer in
Zoology in the University of Cambridge
THE ELEMENTS OF
EXPERIMENTAL EMBRYOLOGY
LONDON
Cambridge University Press
FETTER LANE
NEW YORK • TORONTO
BOMBAY • CALCUTTA • MADRAS
Macmillan
TOKYO
Maruzen Company Ltd
All rights reser'ved
THE ELEMENTS OF
EXPERIMENTAL EMBRYOLOGY
BY
JULIAN S. HUXLEY, M.A.
Honorary Lecturer in Experimental Zoology,
King's College, London
AND
G. R. DE BEER, M.A., D.Sc.
Fellow of Merton College, and Jenkinson Lecturer
in Embryology, Oxford
E
i
New York - London
1963
First published in 1934
Reprinted by permission of
the Cambridge University Press
Printed and Pubhshed by
HAFNER PUBLISHING COxMPANY. INC.
31 East 10th Street
New York 3, N.Y.
Library of Congress Catalog Card Number: 63-14241
No part of this book may be reproduced
in any manner without written permission
from the publisher.
To
ROSS HARRISON and HANS SPEMANN
t#
CONTENTS
Preface page ix
Acknowledgments xiii
Chap. I Historical introduction to the problem of differ-
entiation I
II Early amphibian development: a descriptive
sketch 1 3
III Early amphibian development: a preliminary ex-
perimental analysis 35
IV The origin of polarity, symmetry, and asymmetry 60
V Cleavage and differentiation 83
VI Organisers : inducers of differentiation 134
VII The mosaic stage of differentiation 194
VIII Fields and gradients 271
IX Fields and gradients in normal ontogeny 312
X Gradient-fields in post-embryonic life 354
XI The further differentiation of the amphibian
nervous system 373
XII The hereditary factors and differentiation 397
XIII The prefunctional as contrasted with the func-
tional period of development 418
XIV Summary 438
Bibliography and index of authors 443
Appendix 481
Index of subjects 499
PREFACE
A few words are needed to explain the scope of this book. The study
of the developmental processes of animals is an enormous field, of
which only a small fraction can be dealt with in a volume of this
size. The observational and comparative study of embryology falls
outside the boundaries of this series ; in any case, it has already been
treated in numerous authoritative works. Even on the experi-
mental and physiological side, however, there remains the difficulty
of selection from the vast mass of somewhat heterogeneous material
which many lines of research have provided for consideration and
synthesis.
In the first place, development is not merely an affair of early
stages ; it continues, though usually at a diminishing rate, through-
out life. The processes of amphibian metamorphosis or of human
puberty ; the form-changes accompanying growth ; senescence and
natural death itself — these are all aspects of development ; and so,
of course, is regeneration.
We feel that it is impossible to treat the whole life-cycle in a
single volume, and have accordingly set an arbitrary limit to our
material. We have deliberately restricted ourselves to the early
period of development, from the undifferentiated condition up to
the stage at which the main organs are laid down and their tissues
histologically differentiated — in other words, to Wilhelm Roux's
" prefunctional period ". Growth, absolute and relative ; the effects
of function on structure and on size ; the morphogenetic effects of
hormones — the details of these and of other related topics we have
deliberately omitted, and we have contented ourselves with the
addition of a final chapter in which the main peculiarities of the
functional period are contrasted with those of the pre-functional
period of primary differentiation. Any satisfactory treatment of the
latter portion of the developmental cycle would require a separate
volume.
In the second place, within the period of early development, we
have exercised a further selection. In a new field of biology such as
X PREFACE
this, there are always two levels of approach. One of these is
broadly biological, while the other is physiological in the stricter
sense. The prime aim of the worker approaching the problem on
the physiological level will always be to analyse the processes in-
volved in terms of physics and chemistry. The worker on the
biological level will aim at discovering general rules and laws which
he is content to leave to his physiological colleague for future
analysis in more fundamental terms, but which, meanwhile, will
give coherence and a first degree of scientific explanation to his
facts. Both methods are necessary for progress; and while most
biologists hope and expect that one day their laws will, thanks to
the labours of their physiological colleagues, be made compre-
hensible in the most fundamental physico-chemical terms, they
can reflect that it is they who must first reveal the existence of these
laws before the pure physiologist can hope to begin his analysis.
The biologist can also remember that these laws have their own
validity on their own level, whether they be physico-chemically
analysed or not.
We may take a salient example from the contents of this book.
Spemann's discovery of '' organisers " in the process of gastrulation
of Amphibia, and the extension of the concept to other stages of
development and to other groups of organisms, have made it
possible to understand on the biological level many processes of
development which were previously obscure. At the moment we
can only throw out crude guesses as to the underlying physiology
of organisers and their eflFects, but the discovery opens a new field
of research to physiologists, which they themselves would not have
been likely to hit upon for many years. And even if and when the
physiological analysis has been made, the empirical biological laws
concerning organisers will not lose their validity or their interest ;
they will merely have been extended and deepened.
At the present moment, research into developmental problems
is being actively prosecuted on both the biological and the physio-
logical levels. Following up the early work of Roux, Hertwig,
Driesch, Herbst, Jenkinson, Delage, Brachet, Morgan, and Wilson,
a flourishing school of Entwicklimgsmechanik has grown up in
Germany, and another, no less successful, in the United States.
Meanwhile, on the physiological side, the advance has also been
PREFACE xi
Striking, and we may perhaps cite as particular examples such
works as Faure-Fremiet's Cinetique du Developpement ; Gray's
Experimental Cytology ; Dalcq's Bases Physiologiques de la Feconda-
tion ; and Needham's classic book on Chemical Embryology.
So far, however, little progress has been made in equating the
results of the two lines of approach, and it seems clear that a con-
siderable time must elapse before it will be possible to do so satis-
factorily. At the moment the two fields are almost as unrelated as
were, through most of the nineteenth century, the cytological and
the experimental-genetic approaches to the problem of heredity,
which are now inseparable.
That being so, we have not attempted to include the results of
the purely physiological study of development in this survey. This
means that we have deliberately excluded such topics as the
physiology of fertilisation, the mechanics of cleavage, and the bio-
chemistry of the egg and embryo, save where they have a specific
bearing on the biological problems involved.
In other words, what we have attempted to do is to give some
account of the results of the experimental attack on the problem
of the biology of differentiation — the production of an organised
whole with differentiated parts out of an entirely or relatively un-
differentiated portion of living material. Almost the only short
books on this subject since Jenkinson's Experimental Embryology
and his (posthumous) Lectures are Brachet's L'CEuf et les Facteiirs
de VOntogenese, Diirken's Grundriss der Entwicklungsmechanik^
Weiss' Entwicklungsphysiologie der Tiere, and de Beer's Introduction
to Experimental Embryology \ and each of these treats the subject
along rather different lines. Among larger works, Wilson's The
Celly Morgan's Experimental Embryology, Diirken's Lehrbuch der
Experimentalzoologie, and Schleip's Determination der Primitivent-
wicklung are the most important which have appeared since the
pioneer works on the subject. A perusal of them will suffice to
show the extreme diversity of their lines of approach. What we have
felt is that at present there exists in the subject a vast body of facts
and a relative paucity of general principles. We have accordingly
aimed at marshalling the facts under the banner of general prin-
ciples wherever possible, even when the principle seemed to be
only provisional.
Xll PREFACE
Many of the illustrations have been drawn specially for this book
by Miss B. Phillipson, to whose care and skill we wish here to
make acknowledgments. Particular thanks are due to Miss P.
Coombs for her help with typing and many other details of pre-
paring the book for press. Acknowledgment is hereby made to
those authors and publishers of the journals whose names appear
in the legends to the figures, by whose courtesy they are here
reproduced.
We wish to express our thanks to Prof. E. S. Goodrich, F.R.S.,
to Mr and Mrs R. Snow, and to Mr C. H. Waddington, who not
only read part or all of the manuscript but also made several helpful
suggestions. We are under a particular debt of gratitude to Dr Sven
Horstadius, and to Professor J. Runnstrom, who very kindly per-
mitted us to make use of some as yet unpublished results, and also
to Professor M. Hartmann who has kindly enabled us to reproduce
a figure from the (as yet unpublished) 2nd edition of his Allgemeine
Biologie.
In conclusion, we should like to acknowledge our debt to the
late Dr J. W. Jenkinson, an Oxford man, and the pioneer of
Experimental Embryology in this country, and to express our deep
appreciation of the care and skill which the Cambridge University
Press has expended on the production of this volume.
J.S.H.
G. R. DE B.
January, 1934
ACKNOWLEDGMENTS
Acknowledgment is due and hereby gratefully made to the
following, for permission to reproduce figures :
Akademische Verlagsgesellschaft m.b.H., Leipzig (Verh. d.
Deutsche!' Zool. Gesells., Zeitschrift f. wissenschaftl. Zoologie)\
Cambridge University Press {Anatomy and the Problem of Be-
haviour, Coghill; Biological Reviews) Experimental Cytology,
Gray); Chicago University Press {Botanical Gazette \ General
Cytology, Cowdry; Individuality in Organisms, Child; Physio-
logical Zoology); Columbia University Press {Experimerital
Embryology, Morgan); MM. Gaston Doin et Cie, Paris {Archives
de Morphologic)', Herren Gustav Fischer, Jena {Allgemeine
Biologic, Verh. d. Afiatomischen Gesells., Zoologische Jahrbikher);
Messrs Henry Holt & Co., U.S.A. {Physiological Foundations of
Behavior, Child); The Lancaster Press, Inc., U.S.A. {Biological
Bulletin); The Marine Biological Laboratory, Woods Hole,
Mass., U.S.A. {The Biological Bulletin); Messrs Methuen & Co.,
Ltd. {Problems of Relative Growth, Huxley); Le Museum d'His-
toire Naturelle, Geneve {Revue Suisse de Zoologie); National
Academy of Sciences, Washington, U.S.A. {Proceedings); Oxford
University Press {Introduction to Experimental Embryology,
de Beer; Experimental Embryology , Jenkinson; Quarterly Journal
of Microscopical Science); The Council of the Royal Society,
London {Proceedings, Philosophical Transactions); The Science
Press, U.S.A. {The American Naturalist); Herren Julius Springer,
Berlin {Archiv f. Entwicklungsmech. d. Organismen, Archiv f.
Mikros. Anat. u. Entwicklungsmech., Naturwissenschaften, Ergeb-
nisse d. Biologic, Ha?idb. d. norm. u. pathol. Physiol., Zeitschrift
f. vergleich. Physiol.); Herr Georg Thieme, heipzig {Biologisches
Zentralblatt); The Waverley Press {The Science of Life, Wells,
Huxley and Wells); The Waverley Press, Inc., U.S.A. {Journal
of Experimental Medicine); The WiHiams and Wilkins Company,
U.S.A. {Quarterly Review of Biology); The Wistar Institute of
Anatomy and Biology, U.S.A. {Journal of Experimental Zoology).
Acknowledgment to the authors of the works from which
these illustrations are taken is made in the legends.
Chapter I
HISTORICAL INTRODUCTION TO THE
PROBLEM OF DIFFERENTIATION
The production of the adult Uving organism with all its complexity
out of a simple egg (or its equivalent in the terminology of the
ancients) is a phenomenon and a problem which has attracted the
attention of philosophers as well as of scientists for over two thou-
sand years. To give a brief account of the history of ideas relating
to this problem is no easy matter, but the task is fortunately facili-
tated by the fact that Dr E. S. Russell and Prof. F. J. Cole, F.R.S.
have recently devoted volumes to certain aspects of this subject,
and to the reader who desires to become better acquainted with it,
no better advice can be given than to refer him to The Interpretation
of Development and Heredity, and to Early Theories of Sexual
Generation. The historical section of Dr Needham's Chemical Em-
bryology, and various works of Dr Charles Singer also provide much
valuable information.
Meanwhile, a brief attempt will be made in the following few
pages to outline the essential features of the chief schools of thought
concerning problems of development, in order to show how the
modern science of experimental embryology came into being, and
to present it in its proper historical setting.
The kernel of the problem is the appearance during individual
development of complexity of form and of function where
previously no such complexity existed. In the past, there have
existed two sharply contrasted sets of theories to account for it. One
view accepts the phenomenon as essentially a genesis of diversity,
a new creation, and attempts to understand it as such. This coming
into existence of new complexity. of form and function during
development is styled epigeftesis.
The difficuhies which other thinkers experienced in trying to
understand how epigenesis may be brought about led them to deny
that it exists: i.e. to say that there is no fresh creation of diversity
HEE ^
2 HISTORICAL INTRODUCTION TO THE
in development from the egg, but only a realisation, expansion, and
rendering visible of a pre-existing diversity. Preformation is the
fundamental assumption of views of this type, and they are classed
together as preformationist theories. But the doctrine of preforma-
tion, however, met with even graver obstacles, both logical and
empirical, than the opposite view, and biological opinion is now
united in maintaining the existence of a true epigenesis in develop-
ment. In recent years, however, the discoveries of genetics have
reintroduced certain elements of the preformationist theory, but
in more subtle form. As will be seen later, the modern view is
rigorously preformationist as regards the hereditary constitution
of an organism, but rigorously epigenetic as regards its embryo-
logical development.
To a large extent, the preformationist view assumes as already
given that which the epigenetic attempts to study and to explain ;
and the problem is complicated by the fact that notions of emi-
bryonic development have been confused with concepts of heredity.
This is evident in the attempt, on the part of the author of Peri
Gones in the Hippocratic corpus, to explain development by as-
suming a part-to-part correspondence between the parts of the body
of the parent and those of the offspring : the corresponding parts
being related to one another via the "semen", or, as would now be
said, via the germ-cells. By assuming that the embryo at its earliest
stage is a minute replica of the adult, its parts having been "pre-
formed " by representative particles coming from the corresponding
parts of the parent, the preformationist hypothesis attempts to solve
at one stroke both the problem of hereditary resemblance between
generations and the problem of development within each generation.
This view was in reality shattered by Aristotle's criticism, but it
was revived and widely held during the seventeenth and eighteenth
centuries, largely owing to the fact that mechanistic explanations
had come into vogue, and it seemed impossible to understand
epigenesis on mechanistic lines. One of the foremost exponents
of the preformationist hypothesis was Charles Bonnet. His views
were freed from the crude idea that the preformation in the egg
was spatially identical with the arrangement of parts in the adult
and fully developed animal, or that the ''homunculus" in the
sperm, with the head, trunk, arms and legs which it was supposed
PROBLEM OF DIFFERENTIATION 3
to have (and which certain over-enthusiastic observers claimed to
have seen through their microscopes ; see Cole's Early Theories of
Sexual Generation) only required to increase in size, as if inflated by
a pump, in order to produce development. Instead of regarding
the rudiments of the organs as being preformed in their definitive
adult positions, Bonnet imagined them as "organic points" which
subsequently had to undergo considerable translocation and re-
arrangement. He was thus able to reconcile his belief in preforma-
tion with the empirical fact that the germ or blastoderm of the early
chick showed no resemblance to a hen.
Bonnet's theories were ahead of his facts, and, indeed, he seems
to have been proud of it, for he refers to the preformationist view
as " the most striking victory of reason over the senses ". The hypo-
thesis of such an invisible and elastic preformation was perhaps
permissible in Bonnet's day, but later observational and experi-
mental evidence has rendered it utterly untenable. Further, a rigid
preformationist view which asserts that the &gg is a miniature and
preformed adult, necessarily implies that the egg must also contain
the eggs for the next generation ; the latter eggs must therefore also
contain miniature embryos and the eggs for their subsequent
generations. Bonnet realised that an emboitement or encasement of
this kind ad infinitum would be an absurdity. (Incidentally, it may
be noticed that if it were true, phylogenetic evolution — unless it too
were preformed and predetermined — would be an impossibility.)
But then, if all subsequent generations are not preformed in minia-
ture now% there must come a time when they are determined and
preformed. Before this time they were neither determined nor
preformed, and this making of a new determination, albeit pushed
into the future, is the antithesis of preformation.
If pushed to its extreme conception of infinite encasement, then
preformation is absurd ; if not pushed to this extreme, preforma-
tion will not account for the determination of ultimate future
generations ; and if it did apply, preformation would be an unsatis-
factory view in that it assumes that the diversity which is progres-
sively manifested in development is ready-made at the start, and
in no way attempts to explain it causally or to interpret it in simpler
terms.
4 HISTORICAL INTRODUCTION TO THE
§2
Logically, the preformationist view is associated with the notion
of separate particles being transmitted from parent to offspring,
though the converse does not hold. In preformationist theory, the
hypothetical particles establish the one-to-one link between the
corresponding organs and parts of parent and offspring, whereas
the modern view, which combines an epigenetic outlook on de-
velopment with the particulate theories of neo-Mendelism, denies
any such simple correspondence between hereditary germinal unit
and developed adult character. Darwin's theory of pangenesis re-
sembles that of the Hippocratic writer in this respect, the pangens
being supposed to come from all parts of the body of the parent and
to be transmitted, via the germ-cells or "semen", to the offspring
whose development they mould. Embryologically, however,
Darwin's theory is vague, and leaves the question of preformation
open. Weismann's theory of the germ-plasm, in which the deter-
minants are regarded as representing the predetermined but not
spatially preformed diversity of the future embryo, differs from that
of previous preformationists in that the particles are regarded as
coming, not from the corresponding parts of the body of the parent,
but from the germ-plasm, of which each generation of individual
organisms is held to be nothing but the life-custodian. Weismann
identified the determinants with the material in the nuclei of the
cells, which material he (wrongly) supposed was divided unequally
in the process of division or cleavage of the tgg, so as to form a
mosaic, the pieces (cells or regions) of which would then contain
different determinants and would therefore be predetermined to
develop in their respective different and definite directions.
According to the writer of the Hippocratic treatise Peri Gones
and to Darwin, therefore, offspring resembles parent because the
particles responsible for the development of the parts of the off-
spring come from the corresponding parts of the parent. According
to Weismann, however, offspring resembles parent because both
have derived similar particles (determinants) from a common
source — the germ-plasm.
The question of the origin of the particles or hereditary factors
and of their distribution from the parent to the offspring is one
PROBLEM OF DIFFERENTIATION 5
which principally concerns the science of genetics. The modern
tendency is to accept the principle of a germ-plasm while recog-
nising that it is not as inaccessible to the modifying action of ex-
ternal factors as Weismann contended. The question oi the function
of the particles or factors in converting the fertilised egg into the
body of the adult is the concern of that modern and rather special
branch of embryology usually called physiological genetics.
Before dealing with the conclusion derived directly from experi-
mental work, a moment's attention may be turned to philosophical
criticisms of the preformationist view that particles, determinants,
or any hereditarily transmitted units or factors, can ''explain"
development. First of all, Aristotle pointed out that certain features
in which offspring resembled parent could not be ascribed to
the transmission of particles from corresponding parts, for the
latter might be dead structures like nails or hair from which no
particles could be expected to come, or again they might be such
characters as timbre of voice or method of gait. He goes on to say,
by way of illustration, that if a son resembles his father, the shoes
he wears will be like his father's shoes, yet there can, of course, be
no question of particles here. In other cases, resemblance may refer
to structure, plan or configuration rather than to the material of
which it is composed, and it is hard to see how particles can repre-
sent such structure, plan or configuration. Again, how is the
eventual beard of a son to be explained if he was born to a beardless
father? To these objections might be added the insuperable diffi-
culty of accounting for the production of oflrspring structurally
different from the parent, as when the egg laid by a queen bee
develops into a worker, or, even more generally, when a mother
bears a son or a man fathers a daughter.
If, then, particles coming from corresponding parts are not re-
quired in some cases and cannot be resorted to in others in order
to explain development and hereditary resemblance, why should
they be postulated in any case? This, of course, concerns genetics
as much as embryology, but Aristotle came very close to the crucial
problem of the latter when he wrote : "either all the parts, as heart,
lung, liver, eye, and all the rest, come into being together, or in
succession.... That the former is not the fact is plain even to the
senses, for some of the parts are clearly visible as already existing
6 HISTORICAL INTRODUCTION TO THE
in the embryo while others are not ; that it is not because of their
being too small that they are not visible is clear, for the lung is of
greater size than the heart, and yet appears later than the heart
in the original development".^
Simple observation, therefore, had even in Aristotle's time given
the lie direct to the view that the embryo is a spatially preformed
miniature adult. Similar but more exhaustive and more crucial
observational evidence against the preformationist view was sup-
plied by William Harvey (who referred to development as '' epigene-
sin sive partium super additionem ") and, notably, by Caspar Fried-
rich Wolff. The conclusion to which the latter came is the same as
that of Aristotle. In the earliest stages of the development of the
fowl, the microscope reveals the presence of little globules heaped
together without coherence, and a miniature of the adult simply
does not exist. Further, no refuge can be taken in the assumption
that the miniature is too small to be seen, for its parts (globules) are
clearly visible, and, a fortiori, therefore, the whole. The plain fact
is that the miniature of the adult is not there.
The necessary epigenetic correlate of this fact has been admirably
put by Delage in the following words :*' latent or potential characters
are absent characters. . . . The egg contains nothing beyond the
special physico-chemical constitution that confers upon it its in-
dividual properties qua cell. It is evident that this constitution is
the condition of future characters, but this condition is in the Qgg
extremely incomplete, and to say that it is complete but latent is to
falsify the state of affairs. What is lacking to complete the conditions
does not exist in the egg in a state of inhibition, but outside the egg
altogether, and can equally well occur or not occur at the required
moment. Ontogeny is fiot completely determined in the tgg ".^ We
might sum up the position by saying that to maintain the full pre-
formationist view would partake of the nature of fraudulent book-
keeping.
There is no way of saving the view that the adult is preformed
in the egg as a diminutive replica. The more subtle idea of Bonnet's,
of preformed "organic points", or of determinants unequally dis-
tributed between the cells into which the tgg divides, also met its
doom a century ago, when Etienne Geoffroy St Hilaire (1826) experi-
^ Quoted from Russell, loc. cit.
PROBLEM OF DIFFERENTIATION 7
mentally produced developmental monsters out of chick embryos,
and rightly concluded that since there cannot have been any
preformation of these experimentally induced monstrosities, normal
embryos need not be preformed either. A better known death-knell
for the preformationist hypothesis is Driesch's demonstration that
in many forms, the parts (blastomeres or groups of blastomeres) of
the dividing egg could, if separated, develop into complete little
embryos. It is impossible to imagine any theory of preformation,
however elastic, which will explain the fact that an egg normally
develops into a single embryo, and yet can be made to give rise to
two or four whole embryos.
§ 3
The inevitable conclusion is that development involves a true in-
crease of diversity, a creation of differentiation where previously
none existed, and that the interpretation of embryonic development
must be sought along the lines of some epigenetic theory. The
problem is narrowed down to a search for a principle on which it is
possible to understand how the determinations of the future em-
bryo can arise out of a non-diversified egg. It is the great merit of
C. M. Child to have shown in theory how this is possible. Briefly,
his view (which will be considered in detail later) is that certain
external factors set up quantitative differentials in the egg and
embryo, as a result of which qualitative differences of structure
ultimately ensue. The egg contains a complex of inherent factors,
notably the genes of Mendelian theory, which have been trans-
mitted from its parents and ensure that it shall develop in a specific
fashion, and that if the environment is normal it shall develop so
as to resemble other members of its kind. However, these internal
inherent and transmitted factors of the egg, though genetically pre-
formed, cannot be regarded as a preformation in a spatial or em-
bryological sense. What they do is to confer upon the developing
organism the capacity to respond in a specific way to certain stimuli
which in the first instance are external to the organism. It is, as
Ray Lankester and Herbst first suggested, these responses of a
specific hereditary outfit to stimuli outside themselves, which
constitute development.
Differentiation is evoked out of the egg afresh in each and every
8 HISTORICAL INTRODUCTION TO THE
generation: every individual organism is created by epigenesis
during its own life-history. The environment is as important as are
the internal and transmitted hereditary factors, and both must be
normal for a normal embryo to be developed. If the environment
is abnormal, there will either be no development at all, or an ab-
normal and abortive development, and the same fate befalls an
abnormal hereditary constitution reacting with a normal environ-
ment. If both the environmental and hereditary factors are within
the bounds of normality, then development will follow the lines
which are characteristic for the particular species of organism in
question.
The origin of differentiation and of the epigenetic process are
therefore to be found in the processes by which in the first place
quantitative differentials are induced in the egg by external factors,
and in the second place qualitative structural diversities result from
the interaction of the quantitative differentials with the inherited
constitution. It is these problems which form the subject-matter
of this book.
§4
Meanwhile, it is necessary to pause, and to consider for a moment
how the causal postulate can be applied to development conceived
as an epigenesis. On the preformationist view, the causes of de-
velopment present no particular difficulty, for diflferentiation is then
supposed to be there all the time and to require nothing but ex-
pansion or unrolling (''evolution" in the eighteenth-century sense)
in order to become visible. Even after the discomfiture of the pre-
formationist view at the hands of Wolff and others, and the
acceptance in principle of an epigenetic theory of development,
the need for an application of the causal postulate was cloaked by
the unfortunate effects of Haeckel's theory of recapitulation. This
view, pushed to its ultimate conclusion, maintained that ontogeny
or embryonic development was inevitably a recapitulation of
phylogeny or racial evolutionary history, and that phylogeny was
the mechanical cause of ontogeny, whatever Haeckel may have
meant by such a statement. If this was true, then clearly there was no
need to look for other causes than the evolutionary history in order
to explain development. But, as Wilhelm His saw, it was not true.
PROBLEM OF DIFFERENTIATION 9
The Aristotelian view of the causes of epigenesis is complicated
and somewhat grotesque from the modern point of view, but it
introduces some notions which are very apposite in any discussion
of this problem. First of all, Aristotle realised the principle of
linked causes, which may be illustrated with reference to the inter-
dependence of meshed cogwheels in machinery. He wrote: "that
which made the semen sets up the movement in the embryo, and
makes the parts of it by having first touched something, though not
continuing to touch it".^ This is the principle on which a clock
works after it has been wound up, and many thinkers have imagined
development as the working of machinery originally wound up and
set going at conception, the continued working of which was due
to the progressive assumption of causal activity by the results of
previous causes.
But Aristotle did not regard this view as providing a sufficient
explanation; in addition, he held that the "soul" was active in
controlling the material forces and mechanical processes of de-
velopment. Kindred views have been expressed by von Baer and
by Driesch. The former held that each stage of development was a
necessary conditioji for the production of the following stage, but
was not in any full sense its cause, for in addition he regarded the
"essential nature" of the parent as responsible for controlling the
development of the offspring. Driesch has adapted Aristotle's view
of the functions of the "soul" in his theory of entelechies.
On the other hand, Wilhelm His, having overthrown Haeckel's
theory of recapitulation, regarded each stage of development as a
sufficient cause of the following stage, and so paved the way for a
new branch of science : Entwicklungsmechanik or causal embryo-
logy, the foundations of which were laid by Wilhelm Roux. In
what may be regarded as the "charter" of the new science, Roux
prescribes the analysis of development into so-called complex
components, such as assimilation, growth, cell-division, etc. Ulti-
mately he supposed these complex components to be reducible to
simple components, which in turn would be capable of interpreta-
tion in terms of physics and chemistry (Roux, 1885).
Whether future research will succeed in so reducing the complex
components of development as to render them susceptible of ex-
^ Quoted from Russell, loc, cit.
lO HISTORICAL INTRODUCTION TO THE
pression in fundamental physico-chemical terms is a question of
its own, and one which has been much obscured by the introduc-
tion of what are ambiguously called ''mechanistic" explanations.
As Woodger's (1928) analysis has shown, the term "mechanistic"
as applied to biological phenomena may mean : either
1. That the structure and function of living organisms is to be
completely explained in terms of "little bits of stuff pushing
one another about" in accordance with the classical laws of
mechanics; or
2. That all the phenomena presented by a living organism are
ultimately capable of analysis in terms of the laws of physics and
chemistry; or
3. That a living organism is in some sense analogous to a human-
made machine and that its processes are explicable in terms of
this analogy; or
4. That the causal postulate is perfectly applicable to living organ-
isms and can be satisfactorily applied to the biological order of
things, whether or no the phenomena of the biological order can
ultimately be brought into line with physico-chemical phenomena
and prove susceptible of analysis in physico-chemical terms.
The fourth of these alternatives is generally accepted, and, in-
deed, the whole science of causal embryology is based upon it. The
second alternative is also widely accepted, and is the only fruitful
working hypothesis for the biologist. It is clear, however, that it
may require modification, for further study, notably of the
phenomena of life, is likely to reveal new and hitherto unsuspected
physico-chemical properties of matter. Accordingly, it is necessary
to take physics and chemistry in the most extended sense. The ad-
vances made in physics itself have rendered the first alternative
untenable, and the third cannot pretend to have more value than
can ever be ascribed to processes of reasoning by analogy; thus,
what may be called the cruder mechanistic view embodied in
alternatives i and 3 may be excluded.
PROBLEM OF DIFFERENTIATION
II
§5
We are not concerned, here, with the construction of a philosophical
system, nor do we wish to prejudge the question of the relationship
to one another of phenomena of the physico-chemical and of the
biological order ; the reader to whom these matters are of interest
Normal
Nerve Tube"
Induced
--'' Nerve -Tube
Normal
Ear-rudiment
Normal
Nerve -Tube ^"-^.^
Induced
Ear-rudimont
Normal ^^"^
Muscle-Seements W^-
B wm^..
Eye - rudiment
I nduced
Nerve-Tube
I nduced
-* Muscle-Segments
Induced ''-'^^^^#8^
Tall-bud -^^^2^-
Fig. I
Normal Tail-bud
Induction of secondary embryo by grafted organiser in Triton. A, 3 days after
operation. B, Some days later. (From Wells, Huxley and Wells, The Scietice of
Life, London, 1929; after Bautzmann.)
may mcst profitably be referred to the recent work of Drs von
BertalanfTy and Woodger: Modem Theories of Development. But as
biologists we do believe that the phenomena which we study in
living organisms conform to a biological order, in which the causal
postulate is strictly applicable. The great value of the new science
of experimental embryology or developmental physiology (the term
12 THE PROBLEM OF DIFFERENTIATION
"Entwicklungsmechanik" is hardly translatable, and, now that its
birth has been described, may best be avoided in English writings)
is that it is enabling biologists to discover the complex components
of development, and so to explore new aspects of the biological order.
The dorsal lip of the amphibian blastopore (the so-called "organ-
iser") has been shown (see fig. i) to be capable of inducing
neighbouring tissues to give rise to all the essential structures of an
embryo ^ (brain, spinal cord, eyes, ears, muscles, kidney tubes, etc.).
The result of grafting an organiser into a suitable environment is
just as definitely causally determined and predictable as the result
of mixing two known reagents in a test-tube, although the pheno-
mena are in the one case of the biological and in the other of the
physico-chemical order. It may be confidently expected that in
time the physiological basis of the organiser's action will be dis-
covered and accurately analysed in physico-chemical terms. ^ Until
then, however, it is both desirable and necessary to push the analysis
as far as possible on the biological level.
It is as a contribution to the analysis of early development on the
biological level that the following pages have been written.
^ Spemann and Mangold, 1924.
- Already it is known that the organising action is due to a substance which
is almost certainly lipoidal and probably a sterol (Waddington, Needham and
Needham, 1933). See pp. 154 and 497.
Chapter II
EARLY AMPHIBIAN DEVELOPMENT:
A DESCRIPTIVE SKETCH
§1
It will be best to base the analytical treatment of development upon
a concrete example, and for this purpose the Amphibia are by far
the most suitable material, as analysis is much more complete in
them than in any other group of organisms. However, before em-
barking upon analysis, it will be desirable to give a brief descriptive
sketch of amphibian development in so far as it is relevant to sub-
sequent chapters ; to do this is the purpose of the present chapter.
The chief stages of amphibian development are as follows : the
changes associated with fertilisation; cleavage, leading to the
blastula stage ; gastrulation, leading to the gastrula ; the elongation
of the embryo and the formation of the neural folds and tube, con-
stituting the neurula stage ; the appearance of the tail, and of the
remaining organ-rudiments, leading to the fully formed embryo,
which then hatches as a young larva ; and then the period of growth
and of functional differentiation. These stages overlap somewhat,
especially the last two, but they provide a useful broad classification.
The typical amphibian egg is a spherical object of which one
hemisphere (known as the vegetative hemisphere) is loaded with
yolk, while the other hemisphere (the animal hemisphere) is freer
of yolk and contains the nucleus. There is, as a matter of fact, a
graded distribution of yolk from the animal to the vegetative pole.
In the Anura, the animal hemisphere is characterised by the posses-
sion of a layer of dark pigment at the surface, which distinguishes
it at a glance from the lighter-coloured vegetative hemisphere. A
similar distinction exists in the eggs of Urodela but is not so marked
because the pigment is less dark. Yolk being of a higher specific
gravity than the other constituents of the egg, it is found that after
the egg has been laid and fertilised and is free to rotate within its
membrane, the main egg-axis, or axis passing through the centres
or poles of both animal and vegetative hemispheres, is practically
H
EARLY AMPHIBIAN DEVELOPMENT
vertical. (This is the rule in the majority of the Amphibia, but it
should be mentioned that in Rana esculenta there are complications,
into which there is no need to go here, as a result of which the egg-
axis appears oblique. See Jenkinson, 1909 b.)
Even before development can be said to have begun, therefore,
the tgg possesses one mark of dissimilarity between its various
regions, one mark of differentiation, which is expressed by saying
Front i
^^^^WTI^^*'
^^<^re Hind ^^
Bac/f
Fig. 2
8cily
Polarity and bilaterality in the frog's egg. Above, in equatorial view ; below, seen
from the vegetative pole : left, before fertilisation ; right, after fertilisation. The
unfertilised egg possesses a single main axis (polarity) at fertilisation, bilaterality
is established through the formation of the grey crescent in or close to the future
mid-dorsal line. (From Wells, Huxley and Wells, The Science of Life, London,
1929.)
that the egg has polarity. This polarity is of great importance for
future development because the future front end of the animal will
be formed in proximity to the animal pole of the egg, and the hind
end of the animal close to the vegetative pole. Apart from this
polarity, which concerns the differential distribution of pigment,
yolk and cytoplasm, and the excentric position of the nucleus, the
egg is undifferentiated.
A DESCRIPTIVE SKETCH 15
As a rule among Amphibia, the ist polar body is given off before
fertihsation, and the 2nd polar body after that event. The fertilisa-
tion of the egg by the sperm has a threefold significance. In the
first place it activates the egg to begin its development; secondly,
it brings in to the resulting zygote its supply of paternal hereditary
factors ; and lastly, it is responsible for bringing about the next step
in differentiation, which is the determination of a plane of bilateral
symmetry.
In the frog it has been shown by experiment that the mid-ventral
line of the embryo will be formed close to the meridian on which the
sperm enters the egg.^ The only visible differentiation at this stage,
however, concerns the dorsal side, opposite the point of sperm
entry. A region of this, rather below the equator of the egg, is
marked soon after fertilisation by changes in the surface layer, lead-
ing in the case of the Anura to the formation of the so-called grey
crescent, due to the retreat of pigment into the interior of the egg.
Analogous, but less well-marked changes on the dorsal side of the
recently fertilised egg are observable in the Urodela.^
After the entry of the sperm, therefore, the developing organism,
although still a spherical object, has all three of its axes determined.
The antero-posterior axis and the dorso-ventral axis of the future
embryo lie in the plane of bilateral symmetry, which, in turn,
passes through the original egg-axis of polarity. At the same time,
the transverse, or left-right axis, is also necessarily fixed with the
determination of the other two axes. The symmetry relations of
the organism are thus completely and definitely fixed (fig. 2).
The grey crescent of the Anuran egg (or its equivalent in the egg
of Urodela) is the place at which the next marked step in differen-
tiation appears. The egg has meanwhile undergone cleavage, and
instead of being a single large cell, has come to consist of a large
number (over a thousand) of smaller cells or blastomeres, which
enclose a small cavity, the blastocoel. These blastomeres are smaller
in the animal hemisphere than in the vegetative. This is a result of
the prime differentiation of polarity, for yolk retards cell-division,
^ Roux, 1903; Jenkinson, 1907, 1909 A.
- Vogt, 1926 B.
l6 EARLY AMPHIBIAN DEVELOPMENT
and the cells containing more yolk (those of the vegetative hemi-
sphere) will necessarily divide less fast than the cells of the animal
hemisphere which are relatively free from yolk. Consequently, the
cells of the vegetative hemisphere will be larger than those of the
animal hemisphere at any given time during cleavage.
There is also a slight difference in the size of the blastomeres at
different positions on the same circle of latitude : a difference which
is already shown by the animal hemisphere cells at the 8-cell stage. ^
Though the cause of this size difference at this early stage is ob-
scure, at later stages of cleavage it is due to the fact that the cells on
the dorsal side divide slightly faster and therefore become a little
smaller than those on the ventral side.
The next stage in differentiation consists in the conversion of the
ball of cells — the blastula — into a double-layered sac or gastrula,
by means of the process of gastrulation. Owing to the large amount
of yolk present in the amphibian egg, this process is not as simple
as in other forms (such as Amphioxiis) where gastrulation is a simple
invagination of one side of the blastula into the other. In the am-
phibian, the same result is achieved by the spreading of the cells of
the animal hemisphere and their downgrowth over those of the
vegetative hemisphere, at the same time as they tuck in or invaginate
and then extend forwards beneath the surface of the outer layer.
This process of spreading and growing over (epiboly), and of tuck-
ing in (invagination), first takes place on the dorsal side of the
embryo, in the region of the grey crescent, and gives rise to a lip
known as the dorsal lip of the blastopore.
Eventually, this lip of overgrowth and tucking in forms a com-
plete ring by extending laterally, until the two sides of the lip meet
on the ventral side of the embryo. In this way the blastopore be-
comes a circular aperture leading into the cavity of the archenteron
or future gut. This gut-cavity is a new formation and the direct
result of gastrulation. Its lining is made up partly of the cells that
have been tucked in round the rim of the lip of the blastopore, and
partly of the yolk-laden cells which originally occupied the vege-
tative pole of the egg. The amount of these yolk-cells is too large
for them to be completely accommodated in the newly formed gut-
cavity, with the result that some of them protrude through the
^ Morgan and Boring, 1903.
A DESCRIPTIVE SKETCH
17
mouth of the blastopore forming the so-called yolk-plug. At the
same time, the original cavity of the blastula, the blastocoel, has
b
Fig. 3
Diagrams to show the directions of movement and displacement of the parts of
the blastula in the process of gastrulation in Amphibia, a. Seen from the vege-
tative pole, b, From the left side. The thick lines show the tracks followed on the
surface: the thin lines the tracks beneath the surface after invagination at the
blastopore rim. (From Vogt, Arch. Entzomech. cxx, 1929.)
been more or less obliterated by the formation of the new cavity,
the archenteron or gut (fig. 4).
Since it is on the dorsal side of the embryo that the overgrowth
l8 EARLY AMPHIBIAN DEVELOPMENT
and ingrowth begins and continues with the greatest activity, the
mass of heavy and inert yolk-cells becomes piled up on the opposite
or ventral side of the gut-cavity. This alters the position of the
embryo's centre of gravity, and as a result the entire embryo rotates
ventralwards through about ioo°, until the original egg-axis of
polarity is nearly horizontal, and the animal pole faces forwards
and a little downwards. The blastopore becomes smaller and
eventually closes by the apposition of its lateral lips to one another.
At closure, it is situated close to the original vegetative pole, which
in its turn, as a result of the embryo's rotation, is now facing back-
wards and slightly upwards.
Internally, meanwhile, the endoderm and the mesoderm are
becoming sorted out, so that gastrulation results in the delimitation
of the primary germ-layers, ectoderm on the outside, endoderm
lining the gut-cavity, and mesoderm in between.
§3
The details of the manner in which the mesoderm and endoderm
arise in Amphibia have only recently been made out and established,
thanks to the method of marking definite regions of the living
embryo with easily visible stains, and following them through
development.^
The following account applies to the Urodele type. The material
which becomes tucked and rolled in over the rim of the blastopore
on the dorsal side of the embryo, and thus forms the primitive gut-
roof, will ultimately give rise to the notochord and some of the
mesoderm. Meanwhile, the yolk-laden cells of the original vegeta-
tive pole are carried in under the lip of the blastopore by the pro-
cess of invagination and find themselves forming the anterior end,
floor, and sides of the gut-cavity. Later, these sides grow up be-
neath the primitive gut-roof and meet one another in the mid-
dorsal line, forming the definitive gut-roof. The remainder of the
mesoderm is formed from the material rolled in at the lateral and
ventral lips of the blastopore ; though continuous dorsally with the
primitive gut-roof, it is never in direct contact with the archenteric
cavity (fig. 4).
The mesoderm thus forms paired sheets of tissue (right and left
^ Vogt, 1929.
A DESCRIPTIVE SKETCH
19
Fig. 4
Diagrams showing the process of invagination and mesoderm-formation in
Urodeles. Each diagram is a median sagittal section on to which the mesoderm
of one side has been projected, a, Early stage ; the dorsal lip is well advanced, the
ventral lip barely indicated, the sheet of mesoderm is beginning to spread for-
wards from the dorsal and lateral lips. &,The tip of the notochord {Ch) is growing
forwards beneath the neural plate ; the edge of the mesoderm sheet {p) has ex-
tended further forward ; a small blastocoel (F) is still visible ; K, the front of the
neural plate, c, d, Further stages ; mesoderm is growing in at the ventral lip ; the
mesodermal sheet has extended forwards and downwards, and leaves only a small
area unoccupied. The paired rudiments of the4ieart are situated near the growing
edge of the mesodermal sheet on each side. (From Vogt, Arch. Entwmech. cxx,
1929.)
2-2
20 EARLY AMPHIBIAN DEVELOPMENT
of the notochord), continuous with one another posteriorly and
ventrally, round the rim of the blastopore. The lateral edge of each
sheet of mesoderm rests upon and is more or less confluent with
the outer surface of the endoderm of the floor and sides of the gut,
and this confluence follows a line passing diagonally forwards and
upwards from the ventral lip of the blastopore, on each side of the
embryo. The lateral edges of the mesodermal sheets then become
free from the endoderm, and gradually extend forwards and
ventrally. The mesoderm, which ultimately comes to be situated
in the mid-ventral line of the embryo in front of the blastopore, and
which among its derivatives will include the heart, is thus of paired
origin. That portion of each sheet of mesoderm which immedi-
ately flanks the notochord undergoes metameric segmentation to
form the somites and myotomes, while the remainder gives rise to
the unsegmented mesoderm of the lateral plate. The kidney tubules
arise from tissue on the margin between the segmented and un-
segmented portions of the mesoderm.
It will be noticed from this account that in the Urodele, the
mesoderm and endoderm are separate zones, more or less sharply
marked off from one another, from the very outset of and right
through gastrulation. The endoderm is soon fashioned into a cup
in the antero-ventral region of the embryo, with its concavity facing
backwards and upwards : the mesoderm forms another cup, in the
postero-dorsal region of the embryo, inverted over the endoderm
cup, and with its concavity facing forwards and downwards. Each
of these two cups then completes itself into a hollow sphere by the
growth of its margins. In this way, the endoderm undergrows the
mesoderm and notochord to form the definitive gut-roof, while the
mesoderm overgrows the endoderm until it eventually encircles it
almost completely.^
^ The detailed study of the processes of gastrulation and germ-layer formation
in Urodela and in Anura throws an important light on the distinction (based on
morphological considerations) between peristomial and gastral mesoderm. The
former is regarded as derived from the active tissue round the rim of the blasto-
pore, while the latter is supposed to be derived (by delamination or evagination)
from the wall of the gut. In both Urodela and Anura the mesoderm is derived
from a ring of tissue surrounding the blastopore, and is, strictly, peristomial. But
in Anura the conditions of invagination are such that the mesoderm is rolled in
as a mantle closely applied to the endoderm, and it is its subsequent delamination
from the latter which gives the mesoderm the appearance of being of gastral
origin.
A DESCRIPTIVE SKETCH 21
The conditions in the Anuran type during gastrulation are in the
main similar to those in the Urodele, except that, for reasons into
which we need not here enter, the gut-cavity possesses its definitive
endodermal gut-roof from the start. This definitive roof is com-
plete except for a thin longitudinal strip corresponding to the noto-
chord and to the cells immediately underlying it which will give rise
to the hypochordal rod. When the notochord and hypochordal rod
become lifted off from the gut-roof, a narrow gap is formed, but it
soon becomes closed by the approximation of the free edges of the
endoderm.^
§4
Since the cells that become tucked in during gastrulation were
originally on the outer surface of the blastula before the process of
gastrulation started, it is possible to outline on the surface of the
blastula the various regions which will, in normal development,
give rise to the various organs of the future embryo. By the
method alluded to above, of making stains intra vitam in particular
places on the surface of the blastula, and by following their changes
of position during gastrulation and subsequent development, it is
possible to discover the normal futures in store for all the regions
of the blastula, and in this manner to ascertain their normal
potencies. One may thus speak of the various regions of the blastula
as presumptive organs: one region is presumptive notochord,
another presumptive brain, and so forth.
By methods of this kind, and by making small injuries in definite
places w^ith the electric cautery, Vogt and his pupils have been able
to map the amphibian blastula completely in terms of presumptive
organ-rudiments. This has been accomplished both for a Urodele
and an Anuran type.-
For purposes of description, a system of notation similar to that
used in fixing the positions of places on the earth's surface will be
found convenient. The dorsal meridian of the egg or blastula,
which passes through the future dorsal lip of the blastopore, may
be taken as a standard meridian, corresponding to the meridian of
Greenwich in geography, and other meridians may be indicated
by degrees of longitude, right or left, from the dorsal meridian.
^ Mayer, 193 1. ^ Vogt, 1929; Suzuki, 1928.
22 EARLY AMPHIBIAN DEVELOPMENT
In the same way, the great circle at right angles to the egg-axis is
the equator of the egg. It coincides more or less with the line of
demarcation between the pigmented cells of the animal hemisphere
and the Hghter-coloured cells of the vegetative hemisphere; fre-
quently, however, the pigment extends well below the equator.
Latitudinal position is not so easy to define as longitudinal, since
the egg-equator is not clearly marked. In the meridian of sym-
metry, however, latitudinal position can be accurately defined as so
many degrees above or below the dorsal lip of the blastopore.
With this in mind, it is now possible to pass to a description of
the facts as found in the egg of the Urodele. Most of the cells of
the vegetative hemisphere of the blastula eventually get tucked in
or enclosed, and find themselves inside the embryo when gastru-
lation has been completed. A crescent-shaped region immediately
above the position of the dorsal lip of the blastopore, and extending
up some way above the equator, is presumptive notochord. On
each side of this is a strip which will give rise to mesodermal so-
mites and to the unsegmented mesoderm of the lateral plate. Below
the latitudinal level of the dorsal lip is a region which includes the
yolk-cells of the vegetative pole, and which will give rise to the front,
ventral, and lateral walls of the gut-cavity and, eventually, to its
definitive roof as well. Most of the ventral half (not to be confused
with vegetative half) of the blastula, composed of portions of the
vegetative as well as of the animal hemisphere, is presumptive
epidermis. This leaves only one region unaccounted for; this,
occupying most of the dorsal half of the animal hemisphere (minus
the presumptive notochord and mesoderm regions mentioned
above), is presumptive neural folds. This latter region may be
described in the blastula as a crescent of which the horns extend
down the sides of the embryo from the animal pole to the equator
along meridians rather more than 90° right and left from the dorsal
mid-line. The central part of the crescent extends from the animal
pole to the point on the dorsal meridian to which the presumptive
notochord region reaches, i.e. about 30° latitude above the equator.
It is important to notice that at this early stage, in the blastula,
the presumptive neural fold region occupies an elongated region
which hes at right angles to the plane of bilateral symmetry.^ While
1 Goerttler, 1925; Vogt, 1926 a.
A DESCRIPTIVE SKETCH
dorsal
23
ventral,
;;S^'''Eg
dorsal
Map of the presumptive regions of the Urodele embryo, projected on to the sur-
face of the blastula, as seen from the vegetative pole and from the left side.
Epidermis, sparse broken lines ; neural plate, dense broken lines ; notochord, dense
dots; mesoderm, fine dots; endoderm, white. The future mesoderm segments
are numbered. Eg-Eg, limit of invaginated region ; J, site of formation of first
invagination; K, gill-pouches; Sch, tail region; Spl, lateral plate mesoderm;
u, position of future blastopore lip ; uP, lowermost pole at this stage ; Vii.Ex, fore-
limb; vP, vegetative pole. (From Vogt, Arch. E?itzvviech. cxx, 1929.)
24
^^m
HK
k" dorsal
Fig. 6
Map of the presumptive regions of the Anuran embryo, projected onto the surface
of the blastula, as seen from the dorsal and left sides. Epidermis, sparse broken
lines; neural plate, dense broken lines; notochord, dense dots; mesoderm, fine
dots; endoderm, white. The future mesodermal segments are numbered.
A, eyes ; «P, animal pole ; Eg,r, limit of invaginated region ; Ex, epidermis of limb
region ; Hhl, ear vesicle ; Hd, ventral sucker ; J, site of formation of first invagina-
tion (dorsal lip); K, gill-pouches, and epidermis of gill region; Kgr, broken line
indicating limits of head ; L, lens ; Met', neural fold ; U, position of future blasto-
pore lip; VEx, forelimb; vP, vegetative pole. Kopjdarm, foregut; Mxmd, buccal
cavity; Riimpfdarvi, hindgut. (From Vogt, Arch. Entzanech. cxx, 1929.)
A DESCRIPTIVE SKETCH 2$
differing from the Urodele picn in certain details of relative life
of regions, the Anuran plan is fundamentally similar (figs. 5 and 6).
The process of gastrulation entails remarkable streaming move-
ments and displacements of the various regions of the embryo.
Presumptive notochord, mesodermal somites, and mesoderm of
the lateral plate become tucked in over the rim of the blastopore
lip as already described. Their places on the surface of the embryo
are taken by the presumptive neural fold region undergoing dis-
placement, its original position being occupied by the expanding
region of the presumptive epidermis. This displacement and ex-
pansion, however, takes place in a peculiar way. It must be re-
membered that at the start of gastrulation the lip of the blastopore
is present only in the dorsal meridian, and the lateral lips are formed
later. Consequently, the material on the dorsal meridian becomes
tucked in first and reaches further forward on the under-side of the
superficial layer of the embryo than does material which is situated
more laterally. One result of this state of affairs has already been
noted : the piling up of the yolk-cells on the ventral side of the gut-
cavity with the resultant rotation of the whole embryo to conform
to the new centre of gravity. There is another important result:
since the disappearance of the cells from the surface of the embryo
and their plunging in over the lip of the blastopore is more active
on the dorsal side, there is a consequent stretching of the regions
right and left of the dorsal meridian, and a movement towards that
meridian to take the place of the invaginated material. In this way,
the two horns of the crescent-shaped region of the presumptive
neural folds, which at the start of gastrulation were situated at the
sides of the embryo, now move nearer to the dorsal mid-line and
to one another, so that they form parallel strips which eventually
enclose the blastopore between their hindmost ends. Between
these parallel strips, the central part of the presumptive neural fold
region stretches backwards along the dorsal meridian to the dorsal
lip of the blastopore, which it reaches. Thus, instead of lying as a
transverse band across the embryo as at the blastula stage, the pre-
sumptive neural fold region after gastrulation occupies a position
extending longitudinally along the dorsal side of the embryo, where
the neural folds will in fact arise.^ It may be referred to at this
^ Goerttler. 1925.
26
EARLY AMPHIBIAN DEVELOPMENT
Stage as the neural plate. All the remainder of the surface of the
embryo is now occupied by presumptive epidermis (figs. 3, 7 and 8).
The movements which have brought about gastrulation are
therefore also responsible for bringing the presumptive neural fold
material into place in preparation for the formation of the neurula,
and this, in turn, as will shortly be seen, paves the way for the
changes which result in the formation of the tail.
Fig. 7
The process of gastrulation in Urodeles, revealed by the movement of intra
vitatn stain marks placed on the surface of the blastula, as in a. The marks stretch
and move towards the blastopore rim. In b mark i has become invaginated ; in
d only mark 4 is left on the surface; the others have become invaginated and
passed forwards, forming the gut-wall, and can be seen by transparency through
the epidermis. (After Goerttler, Arch. Entwmech. cvi, 1925, modified.)
Accompanying the processes of displacement and stretching
which have just been described, growth also takes place, which pro-
cess results in the elongation of the embryo along the line of the
original egg-axis, now the antero-posterior axis — in other words,
produces growth in length.
A DESCRIPTIVE SKETCH
27
§5
The neural folds now rise up as a pair of parallel ridges along the
dorsal side of the embryo, and come to enclose the blastopore,
which is now reduced to a mere slit, between their hind ends. As
soon as this has happened, the embryo may be termed a neurula.
The groove between the neural folds becomes converted into a tube
4 5 6
BO. p.
C d
Fig. 8
The process of neurulation in Urodeles, revealed by the movement of intra vitani
stain marks placed on the surface of the gastrula in a transverse line across the
animal hemisphere, a, Seen from the dorsal side, b, From the right side, c, With
the progress and completion of gastrulation, the band of stain marks becomes
U-shaped, the arms parallel with one another along the dorsal side, and marking
the site of formation of the neural folds {d). (After Goerttler, Arch. Entwmech.
cvi, 1925, modified.)
as the neural folds arch over to join one another and fuse above it,
and the blastopore is no longer at the surface of the embryo, but is
covered over by these folds. In this manner, a neurenteric canal
(actual or virtual according as to whether the blastopore is or is not
still open) is formed, connecting the cavities of the neural tube and
of the gut. After the fusion of the neural folds, epidermis covers
the entire surface of the embryo, and the rudiments of all the other
28 EARLY AMPHIBIAN DEVELOPMENT
Structures have come to lie beneath the surface (with the exception
of a few sense-organs and placodes).
Meanwhile, inside the embryo, the notochord has become an
elongated cylindrical rod above the roof of the gut in the mid-
dorsal line. A split within the substance of the mesoderm gives rise
to the coelomic cavity : this becomes restricted to the region of the
unsegmented lateral plate, and separates an outer somatic from an
inner splanchnic layer of coelomic epithelium.
The formation of the tail is closely bound up with the processes
of gastrulation and neurulation. Although there is still uncertainty
concerning one or two points, the following appears to be the
course of events. When the neural folds arch over towards one
another and fuse, there is formed a double arch or vault of tissue
over the original dorsal surface of the blastula. The outer arch is
the superficial epidermis, and the inner arch is the neural tube
itself. A backgrowth of the hindmost part of the outer arch of the
neural folds gives rise to the epidermis of the tail, which of course
becomes progressively longer. Beneath this epidermis, and in
consequence of the outgrowth of the tail, the inner arch of the neural
folds becomes bent into a J, the bottom of the J occupying the
region of the tip of the tail, and is so disposed that the anterior
four-fifths of the neural folds, from the brain to the tip of the tail,
form the long arm of the J . The other arm of the J is bent ventrally
and forwards, and reaches from the tip of the tail to the region of
the blastopore; it is formed from the posterior one-fifth of the
neural folds. The notochord grows and stretches back between the
arms of the J to the tip of the tail, and that part of the inner arch
of the neural folds that lies dorsal to it (the anterior four-fifths)
gives rise to the definitive neural tube ; while that part of the inner
arch of the neural folds that comes to lie ventral to the notochord
gives rise to the myotomes or muscle-segments of the tail ^ (fig. 9).
There is therefore no undiflFerentiated tail-bud from which the
structures of the tail arise: the neural tube and notochord are
present in the neurula, and their hind ends simply grow and stretch
backwards into the lengthening epidermal bag which forms the
tail, and the material for the muscles of the tail is also present in
the neurula in the hindmost part of the inner arch of the neural
^ Bijtel and Woerdeman, 1928; Bijtel, 193 1.
A DESCRIPTIVE SKETCH
29
folds. But it is to be noticed that these caudal muscles arise from
material that has never been invaginated.
This state of affairs need not give rise to undue astonishment,
for the region from which this presumptive caudal muscle material
Fig. 9
Four stages of development of an embryo of Afublystoma fnexicanum to which
intra vitam stain marks were appHed as shown in a ; mark 1-2 is later found in the
epidermis of the mid-dorsal line and in the hinder part of the neural tube ; mark
2-3 in the epidermis of the tip of the tail and in the hinder muscles of the tail;
mark 3-4 in the epidermis of the mid-ventral line and in the muscles of the base
of the tail and the hinder part of the trunk. Mark 3-4 has been invaginated in
part; the other marks have not been invaginated, but mesodermal muscles have
nevertheless been formed from the median part of mark 2-3. (From Bijtel,
Arch. Entzvmech. cxxv, 1931.)
arises lies immediately to the side of the blastopore at the moment
of the latter's closure, and at the blastula stage it lay touching the
presumptive regions of the hindmost mesodermal somites of the
trunk. It might be said that if the blastopore did not close so soon
but remained open for a little time longer, it would tuck in and
30 EARLY AMPHIBIAN DEVELOPMENT
invaginate this material, which would then differ in noway from the
presumptive mesoderm of the trunk. That the presumptive caudal
muscle material does not get invaginated is probably due simply to
the large amount of yolk present, which fills most of the interior
of the embryo and decreases the space available for material to be
invaginated.
However, the activities which lead to the uprising of the neural
folds, and their fusion, appear of necessity to take in the whole
region from anterior end to blastopore, and so this presumptive
caudal muscle material, through the mere fact of its being left on
the surface, is made to participate in this essentially alien process.
Thus in the Amphibia the embryonic structures known as the
neural folds do not represent a single ultimate morphological unit,
but are composite and represent, in addition to epidermis, two
distinct sets of structures, the neural tube and the muscles of the
tail. The earliest stages of development of these sets of structures
are merely bound up in a single morphogenetic process, the forma-
tion of the embryonic neural folds. The distinction between processes
involving form-change and those involving chemical predetermina-
tion, which it will be necessary to discuss at more length later, is
here very evident.
§6
The formation of the gut, the notochord, the neural tube, the meso-
derm and coelomic cavity, and the tail, together with the elongation
of the whole embryo along the antero-posterior axis, are examples
of morphological differentiation, as a result of which the main
organ-systems of the embryo become roughly blocked out as re-
gards their position and their form. As development proceeds, the
remaining organs become roughed out in the same way. Owing to
the greater width of the groove between the neural folds in the
anterior region, the neural tube is at its first formation already
differentiated into regions of brain and spinal cord, the diameter
of the cavity of the tube being greater in the region of the brain.
From the brain the optic vesicles are pushed out on each side, and
become converted into the optic cups by the invagination of their
outer sides. Opposite the mouth of each optic cup, the lens is
formed as a thickening of the overlying epidermis, and eventually
A DESCRIPTIVE SKETCH 31
becomes separated off from the epidermis to occupy a position in
the mouth of the optic cup. The hypophysis grows in towards the
ventral surface of the brain from the epidermis of the front of the
head. On the under-side of the head, folds of epidermis give rise
in Anura to the ventral sucker, while in many Urodela a finger-
shaped outgrowth beneath the eye forms the so-called balancer.
On either side of the brain, behind the eyes, epidermal pits sink
in to form the ear-vesicles. These pits arise from the deeper layers
of the epidermis, and so the invagination may or may not be covered
over by the superficial epidermal layer. At all events, the ear-
vesicles soon become closed if they were open, and their original
connexion with the epidermis and the exterior is reflected in the
endolymphatic duct (open to the exterior throughout life in the
Selachii). Another pair of pits, on the snout, gives rise to the nasal
sacs and nostrils, and a larger median depression beneath them
sinks in and breaks through into the anterior end of the endodermal
gut. This anterior ectodermal portion of the gut is known as the
stomodaeum, and its aperture of course constitutes the mouth-
opening. A posterior ectodermal portion of the gut, or procto-
daeum, is formed in a similar manner, close to the point at which
the blastopore closed. Its aperture constitutes the anus, and in-
ternally it fuses with and breaks through into the hinder end of the
endodermal gut.
The fusion of the neural folds has not only resulted in the
formation of the neural tube, but it has also led to the inclusion
beneath the epidermis of narrow strips of cells, situated along the
dorso-lateral sides of the neural tube, which constitute the neural
crests. From the neural crests arise the nerve-cells or neurons
which make up the ganglia or aggregations of nerve-cells situated
on the dorsal roots of the segmented cranial and spinal nerves.
Other cells derived from the neural crests give rise to the sheaths
in which various nerves are enclosed. In the head region, it appears
that the neural crests also give rise to parts of the visceral carti-
laginous skeleton. In various places on the surface of the head,
thickenings of the epidermis give rise to placodes, which form the
sense-organs of the lateral-line system, and also contribute some
nerve-cells to the ganglia of some of the cranial nerves. Outgrowths
from the sides of the head form the rudiments of the external gills,
32 EARLY AMPHIBIAN DEVELOPMENT
while the Hmbs arise (early in Urodela, much later in Anura) as
little thickenings which rapidly become conical and continue to
elongate by growth.
As regards the internal development, the dorsal portions of the
mesoderm of the trunk and the mesoderm of the tail become
metamerically segmented, and give rise to the myotomes or muscle
plates. These myotomes are at the outset connected with the meso-
dermal lining of the general coelomic cavity by short stalks, called
intermediate cell-masses or nephrotomes, which, like the myotomes,
are segmental in arrangement. From some of these stalks, out-
growths are formed, ultimately giving rise to the tubules of the
kidney, and from these tubules a duct (the pronephric duct) grows
back on each side into the proctodaeum, which from now on can
be styled the cloaca.
The heart arises beneath the anterior part of the gut in the mid-
ventral line, but the rudiments which form its muscular wall (parts
of the splanchnic layer of coelomic epithehum) are at first paired.
When they have fused together in the middle line, these rudiments
roll up along the longitudinal axis of the embryo to form a tube,
suspended by a mesentery (strictly, mesocardium) from the dorsal
wall of the coelomic cavity, w^hich in this region takes the name of
pericardial cavity. Within the tube thus formed, some cells are en-
closed which will give rise to the lining or endothelium of the heart.
Originally these cells lay scattered irregularly between the floor of
the gut and the splanchnic layer of coelomic epithelium, whence
they arose.
The gut-cavity still contains a considerable quantity of yolk-cells,
and these are heaped up and occupy most of the central and hinder
parts of the gut, being piled up high on the floor, and reducing the
actual free cavity to modest dimensions. Just behind the region of
the heart, and in front of this mass of yolk-cells, a downgrowth is
formed from the floor of the gut. This is the rudiment of the liver :
its cavity will eventually develop into the lumina of the liver
tubules and gall-bladder, while the connexion with the gut persists
as the bile-duct. A lengthening of the gut takes place in the region
immediately in front of the. cloaca, and this gives rise to the in-
testine, which later becomes coiled on itself like a watch spring.
By these processes of stretching, displacement, folding, and
A DESCRIPTIVE SKETCH 33
growth, morphological difTerentiation runs its course, and results
in the placing of material in particular geometrical relations, roughly
in the form and position of the various organs which are to arise.
These simple rudiments then undergo growth at particular rates,
which rates may be proportional to that of the whole embryo, or
faster, or slower. It is obvious that the rate of growth of any
particular rudiment relative to that of its neighbour, and any differ-
ence in the rates of growth of any one rudiment in the three dimen-
sions of space, contribute essential factors in determining the final
form of the organ and of the embryo as a whole (see pp. 225, 366).
§7
After the position and form of an organ has been roughly blocked
out, there follows the process of elaboration of the cells of the organ
for the function which they are to undertake in the organism. This
is the process of histological differentiation, or histo-differentiation^
as it may be more briefly styled. As a result of this process the cells
of the neural tube, for instance, become diversified into supporting
or ependyma cells and into neuroblasts, which latter produce axon-
fibres and give rise to the tracts of the central nervous system and
to the ventral nerve-roots. The dorsal nerve-roots are formed as
a result of the production of fibres by the cells of the neural crests.
In the eye, the various layers of the retina are very early differenti-
ated from one another. Similarly, the cells of the myotomes become
differentiated into fibres of striated muscle ; mesenchyme cells in
particular regions produce cartilage; others elsewhere produce
connective tissue, and others again eventually give rise to bone. The
cells of the hypophysis, which comes into relation with the floor of
the fore-brain or infundibulum to form the pituitary body, become
differentiated into the glandular elements characteristic of that body.
Thus, in every rudiment, the cells undergo specialisation to form
characteristic tissues, differing from one another and from the
simple undifferentiated blastomeres from which all the cells of the
embryo arose. When histo-differentiation of an organ has ap-
proached completion, the organ is able to enter on a new phase of
its development, viz. that of functional activity. Up to this point
development has proceeded without function of the organs : indeed,
they did not exist at the start and have had to be made. After this
HEE 3
34 EARLY AMPHIBIAN DEVELOPMENT
point (which does not occur at the same time for all the organs
of an organism) development can only proceed with function.
Function then perfects the results of the differentiation which
has been achieved without it, and is necessary for full and final
differentiation.^
The onset of function of the organs therefore marks an important
epoch in development, and, following Roux, it is possible to dis-
tinguish a prefunctional period during which morphological and
histological differentiation proceed to make the organs ready to
enter upon their functions, from a functional period during which
functional differentiation effects the final elaboration, interde-
pendence, and control of the rudiments, and converts them into
the perfected organs of the free-living organism.^ It will be neces-
sary to say more on this point in the final chapter.
This book concerns itself almost entirely with the prefunctional
period. As has been shown, this period is characterised by certain
remarkable sequences of morphological and histological processes
of differentiation. Complications of structure and texture appear
which had previously been absent. The next problem to be tackled,
therefore, is the origin of differentiation. This concerns the question
as to how developmental processes are causally related to one
another in the sequence of events, i.e. whether the development of
any given rudiment would take place as it normally does if it had
not been for the previous development of some other rudiment, and
also the question as to what are the factors, causes, or conditions
which are responsible for initiating these sequences of processes
of development and differentiation.
1 The term function is here used to denote function in the ordinary physio-
logical sense, as some specialised activity performed by the organ, normally for
the physiological benefit of the organism as a whole. The tissues are always
"functional" in the sense of being alive and working, and in addition they may
be performing special developmental functions even in that period which is here
denoted as the prefunctional period. Nevertheless, the distinction is an important
and useful one.
2 Roux, i88i.
Chapter III
EARLY AMPHIBIAN DEVELOPMENT:
A PRELIMINARY EXPERIMENTAL ANALYSIS
§1
It has been shown that even before the amphibian egg is fertihsed
it possesses one differentiation, in respect of its egg-axis, which
determines the future positions of the anterior and posterior ends
of the embryo. The factors determining this axis of polarity must
be looked for at a stage before the egg is laid, for, while it is still in
the ovary, the yolk is already concentrated into one hemisphere. It
is possible that the orientation of the blood-vessels with regard to
the follicles and developing oocytes in the ovary may be the deter-
mining factor. It has been asserted^ that these blood-vessels are
so distributed that the arterial blood reaches the oocyte from one
side while the venous blood leaves it at the opposite side. This
w^ould cause a gradient in oxidation, and this in its turn would
produce a gradient in the relative amounts of cytoplasm and yolk,
more yolk being deposited in the regions of low oxidation.
In this particular case, the matter cannot be regarded as certain,
since the same author has later qualified his assertion.'^ In other
organisms, however, it appears assured that the regions of the
oocyte where the rate of oxidation is highest will become the animal
pole of the egg and the anterior end of the embryo (see Chap. iv).
In the absence of evidence to the contrary, we are justified in
assuming that some causal agency of this type is operative in pro-
ducing the primary polarity of the amphibian egg.
Once the amphibian egg is fully formed, however, gravity will
determine that the vegetative hemisphere (containing the relatively
heavy yolk) shall be undermost. This is normally brought about by
rotation of the egg within its membranes after being laid and
fertilised. But if the egg is forcibly inverted and maintained in that
position, gravity will determine that the yolk shall flow down to the
new lower surface. It does this by means of streaming movements,
1 Bellamy, 19 19. ^ Bellamy, 1921.
3-2
36 EARLY AMPHIBIAN DEVELOPMENT
and except in a few cases, where the vegetative pole is almost
exactly uppermost, which condition must be expected to lead to
special difficulties in the way of rearrangement of the yolk, such in-
verted eggs give rise to normal embryos. The cells at what is now
the upper pole divide more rapidly than those at the lower pole,
regardless of whether they are pigmented or unpigmented, and the
dorsal lip of the blastopore appears at the proper level with regard
to the vertical axis.^
Gravity is therefore responsible for the fact that in many forms
the primary egg-axis is brought into a vertical position in normal
development, but it is not responsible for the initial formation of
the axis ; nor is gravity an essential factor in normal development,
for eggs withdrawn from the directive action of gravity by being
forced to roll about continually in a clinostat,^ or by being con-
stantly disturbed by a stream of air bubbles,^ nevertheless develop
into normal embryos.
The original determination of the egg-axis, therefore, appears to
be due to the development of a primary physiological gradient
within the oocyte, which finds visible expression in the graded
distribution of cytoplasm and yolk. And this in turn appears to be
brought about by factors operative in the ovary which are external
as regards the oocyte or egg itself. This point is of considerable
importance, for it shows that even this first step in differentiation
is externally determined, and is not due to an internal factor or
factors.* Cases will be met with where the main axis of the future
organism is normally not determined until after the egg is laid, and
where its direction can be experimentally controlled (p. 60).
§2
The next step in differentiation is the acquisition of bilateral sym-
metry. Localisation of the future median plane of the organism
has been shown to depend mainly upon the point of entry of the
sperm. This has been demonstrated experimentally in the frog by
making the sperm enter the egg on a selected meridian, either by
means of a fine pipette, or by laying a thread against one side of the
egg and allowing a drop of liquid containing sperm to creep along
1 Pfluger, 1883; Born, 1885. - Roux, 1884.
3 Kathariner, 1901. ^ Child, 1924, p. 133.
A PRELIMINARY EXPERIMENTAL ANALYSIS
37
the thread. The result of the experiment can be checked by cutting
the egg into sections, for the path of entry of the sperm is indicated
by a trail of pigment leading into the interior of the egg, and the
grey crescent which indicates the dorsal meridian can also be
identified by the retreat of pigment from the surface By this means
it can be proved that the grey crescent and therefore the mid-
dorsal line is normally opposite or nearly opposite to the point of
entry of the sperm. If, as sometimes happens, two sperms enter
an egg simultaneously, the grey crescent is determined relatively
Fig. lo
Diagrammatic equatorial sections through dispermic frogs' eggs, showing that
the grey crescent (position of which is indicated by thin outHne) is formed
opposite the midpoint between the two points of sperm-entry. The plane of
symmetry is indicated by a broken line. (From Herlant, Arch, de Biol, xxvi,
191 1, figs, ix, X, p. 250.)
to them both, and arises antipodally to the meridian half-way
between their two points of entry. ^ The second step in differentia-
tion, the acquisition of bilateral symmetry, is therefore also deter-
mined mainly in relation to a factor external to the egg (fig. 10).
But, as is very often found in the study of development, the main
determining factor is not the sole one capable of exerting an effect.
This conclusion is necessitated in this case by studying partheno-
genetic eggs. Artificial parthenogenesis may be induced in the egg
of the frog by pricking it with a needle dipped in blood or lymph.
There is then no point of sperm-entry, and yet the eggs develop
^ Roux, 1887; Jenkinson, 1909 a; Herlant, 191 1.
38 EARLY AMPHIBIAN DEVELOPMENT
bilateral symmetry. Furthermore, the plane of symmetry bears no
relation to the point of pricking. ^ It is necessary, therefore, to
assume that even in the unfertilised egg all the meridians are not
perfectly equivalent, and that one of them has some slight differ-
ential in respect of the others. This meridional differential, how-
ever, must also be supposed to be due to some unequal incidence
of external factors operating in the ovary. However this may be,
the egg must acquire and possess some feeble determination of a
plane of bilateral symmetry which becomes realised in the absence
I
i
Fig. II
Cortical localisation of dorsal lip region in frog's egg shown by forced rotation
of the egg. Thick line, original plane of symmetry ; chain line, new plane of sym-
metry, passing through centre of grey crescent region (stippled) and mass of yolk
which has streamed down to lower pole by gravity. (From Weigmann, Zeitschr.
f. Wiss. ZooL cxxix, 1927.)
of any more powerful stimulus, as in the case of artificial partheno-
genesis, but which may be overridden by such stimuli as the point
of entry of the sperm, or the direction of incident light, ^ or the
direction in which the yolk streams down when the egg has been
forcibly inverted. In the latter case, the plane of symmetry is
determined in such a way as to include the centre of the original
grey crescent and the centre of the mass of yolk which has
streamed down under the effect of gravity : the dorsal lip of the
blastopore therefore arises in the normal position, but the lateral
lips form a crescent the concave side of which is always turned
towards the mass of yolk, wherever it may be (fig. 11).^
^ Bataillon, 1910; Brachet, 1911. - Jenkinson, 1909 a.
^ Weigmann, 1927.
A PRELIMINARY EXPERIMENTAL ANALYSIS 39
It appears that once a differential is established, the plane of
symmetry will thereby be determined, and that it will be deter-
mined just as efficiently by a feeble differential as by a strong one.
The possibilities of realising normal bilaterality are thus inherent
in the egg ; but the factors which determine the fact of its realisation
and decide its localisation are external^.
§3
The next step in development is cleavage, the splitting up of the
egg by cell-division into a number of smaller cells, the blastomeres.
Here, one of the effects of the axes already determined (the antero-
posterior, and the dorso-ventral) manifests itself in a differential
rate of activity and cell-division, and therefore a gradient in cell
size, from the animal pole with its small, actively dividing cells, to
the vegetative pole with its more sluggish yolk-containing cells;
and, at any given circle of latitude, the cells on the dorsal side
divide faster and are therefore smaller than those on the ventral
side, at any given time. As will be pointed out in Chap, ix, the
main organisation of the developing egg at this stage consists of
these quantitative gradients, or, as we shall call them, gyadient-fields.
The rate of cleavage and subsequent differentiation can be
locally altered by subjecting the egg to differential temperature-
exposure: one pole or side hot, the other cold.^
The amount of yolk present in the vegetative hemisphere of the
amphibian egg, while responsible for the larger size of the vegetative
blastomeres, is not too great to prevent holoblastic cleavage of the
egg. It is possible, however, to make the cleavage of the frog's egg
conform to the meroblastic type characteristic of Selachians and
^ It might be supposed that the bilateral symmetry of the egg, once established,
is necessarily identical with that of the resultant embryo. However, Jenkinson
(1907, 1909 a) by means of an elaborate biometrical study has shown that the
correlation between the two, though high, is not perfect: in other words, the grey
crescent does not always lie exactly in the future mid-dorsal line. Thus both the
determination of the grey crescent in the meridian of sperm-entry, and that of the
axis of bilateral symmetry of the embryo in the meridian of the grey crescent are
imperfect. In spite, however, of the slight elasticity of the determination at these
two links in the causal chain, it is clear that in normal development the symmetry
of the embryo is mainly determined by the point of sperm-entry. See also Tung,
1933-
^ Huxley, 1927; Gilchrist, 1928, 1929; Vogt, 1928 b.
40 EARLY AMPHIBIAN DEVELOPMENT
Sauropsida, by means of centrifugalisation. The eggs orientate
themselves in the centrifuge tube in such a way that the animal pole
is directed centripetally, and the yolk is concentrated into an
abnormally dense mass at the vegetative pole. Cleavage then results
in the formation of a disc of cells or blastoderm resting upon an
undivided mass of yolk. The nuclei of some of the blastomeres
migrate into the yolk and become enlarged, irregular and
chromatic, and thus resemble the "yolk-nuclei " (bodies responsible
for the precocious digestion of the yolk) characteristic of selachian
development (fig. 12).^
The causes of cleavage concern the problem of cell-division,
which, as such, lies outside the scope of this book.
Modified cleavage of frog's egg, under the influence of centrifugal force. The
yolk (d) is concentrated in the vegetative hemisphere, and cleavage results in the
formation of a blastoderm, m, yolk-nuclei ; kh, blastocoel. (After Hertwig, from
Jenkinson, Experimental Embryology, 1909-)
§4
Following upon cleavage, the next step is gastrulation. This process,
which, of course, results in the conversion of a single-layered hollow
ball (the blastula) into a double-layered sac (the gastrula), is
heralded in Amphibia by the appearance of the dorsal lip of the
blastopore at a particular latitudinal level on the blastula, in the
dorsal meridian. The level at which the lip appears is under the
control of the primary physiological gradient along the egg-axis,
^ Hertwig, 1897, 1904; Jenkinson, 191 5.
A PRELIMINARY EXPERIMENTAL ANALYSIS
41
and can be altered by experimental means (see fig. 149 and
P- 320).
In Amphibia, it has been found that the act of gastrulation can be
analysed into a number of component processes. First, there is the
tendency on the part of the cells of the animal hemisphere to ex-
pand and cover a larger surface. Next, the cells which constitute
the marginal zone between the animal and vegetative hemispheres
tend to stretch downwards towards the vegetative pole. This is
accomplished by rearrangement of the cells, with the result that the
ring-shaped band, increasing in depth, attempts to decrease in
a b
Fig. 13
The expanding growth-tendency of the presumptive epidermis of the Urodele
embryo, a. Two ventral gastrula-halves grafted together, the epidermis of each
of which is thrown into ridges and folds in vainly trying to overgrow the other.
b. The same, 16 hours later, showing intensification of ridges and folds. (From
Spemann, Arch. Entzvmech. cxxiii, 193 1.)
diameter. Thirdly, the cells just beneath the marginal zone in the
dorsal meridian have the tendency to invaginate and form a pit-like
depression. Normally, of course, all these processes take place
together, with the result that the excess of material obtained by the
stretching of the marginal zone becomes tucked into the invagina-
tion round the rim of what may now be called the blastopore. New
material, as it arrives at the rim, becomes tucked in, and this
tendency to roll or tuck in is also an independent process. Mean-
while, the space vacated on the surface by the material which has
thus been invaginated, is occupied by the shifting and expanding
regions of the animal hemisphere.
42 EARLY AMPHIBIAN DEVELOPMENT
By a simple operation, the constituent processes of gastrulation
can be dissociated from one another. Removal of a portion of tissue
at the animal pole of a blastula leads to closure of the wound by
approximation of the cut edges. This results in raising the marginal
zone above the equator of the egg. Nevertheless, this zone soon
shows its characteristic stretching movements, and decreases its
diameter. Normally, of course, this decrease in diameter corre-
sponds to the curvature of the egg from the equator to the vegetative
pole. But as the marginal zone is now above the level of the equator,
it cannot simply grow down over the vegetative hemisphere: in-
stead, it constricts the embryo into the form of an hour-glass.
Meanwhile, an invagination appears in the lower half of the hour-
glass, at a place which the marginal zone would normally have
reached, but which it has been prevented from reaching by the
conditions of the experiment.^ That the process of rolling in or
diving beneath the surface is an autonomous one is shown by the
fact that isolated portions of the dorsal lip region, when grafted
into strange situations in another embryo, promptly proceed to
transfer themselves into the interior by this means.
The fact that all these processes should begin and take place
more actively at the dorsal meridian before extending to lateral
meridians and eventually all round the egg, is a consequence of the
gradient of activity from dorsal to ventral side, mentioned above.
While the marginal zone is stretching, overgrowing the vegetative
hemisphere, and being invaginated and tucked in round the lip of
the blastopore, and the presumptive neural fold region is being
stretched and displaced, thus taking the place of the presumptive
primitive gut-roof which is being invaginated, the presumptive
epidermis region expands and extends by growth so as to cover the
area vacated by the presumptive neural folds. This growth-tendency
on the part of presumptive epidermis is also shown by isolated
pieces when grafted,^ and by two ventral half-gastrulae grafted
together : the epidermis of each half tries in vain to overgrow the
other (fig. 13), with the result that it is thrown into numerous folds. ^
The harmonious co-operation of all these processes, which norm-
ally result in gastrulation, can be thrown out of gear by interference
with the gradients, and alteration of the relative rates of activity in
^ Vogt, 1922. ^ Mangold, 1924. ^ Spemann, 193 1.
A PRELIMINARY EXPERIMENTAL ANALYSIS 43
different parts of the embryo, ^ or by changes of shape, such as
those which are consequent on the release of the embryo from its
viteUine membrane'- (see Appendix, p. 481).
With regard to the actual paths of displacement followed by the
invaginated tissues during these "mass movements" which bring
about gastrulation, it may be said that the nearer any given piece of
tissue is to the dorsal lip of the blastopore at the outset of gastrula-
tion, the farther forward in the embryo will it find itself when that
process is completed. So, those cells which occupy the place where
the invagination first forms become the front wall of the fore-gut ;
those cells of the marginal zone in the mid-dorsal line which are the
first to be tucked in form the tip of the notochord (figs. 3 and 4).
§5
Attention may now be turned to the presumptive regions of the
future organs. As has been shown in Chap. 11, these regions can be
mapped out on the blastula, although there are no visible limits to
distinguish them. The question arises as to how these various
regions have their respective fates allotted to them.
The first point to make clear in any discussion of the origin of
differentiation is the fact that it is impossible to appeal to differ-
ences between the nuclei of the cells of the blastula in order to
account for the eventual differentiation of those cells. By making
eggs undergo cleavage under compression between glass plates,
the normal regular sequence of directions of cleavage can be dis-
turbed, so that the nuclei come to be situated in cells other than
those in which they would find themselves in normal unhindered
cleavage. Nevertheless, the development of embryos so treated and
then released from pressure is normal, and it is therefore clear that
it is quite immaterial whether any given nucleus finds itself in one
particular cell or in another.^ This is confirmed in other ways
and on other forms (see p. 85 and fig. 36).
This means that there is no inequality in nuclear division during
early cleavage, and it is therefore impossible to attribute any deter-
1 Huxley, 1927; Vogt, 1928 b; Gilchrist, 1928, 1929; Dean, Shaw, and
Tazelaar, 1928; Tazelaar, Huxley, and de Beer, 1930; Castelnuovo, 1932.
^ Spemann, 193 1.
^ Hertwig, 1893; Spemann, 1914, 1928.
44
EARLY AMPHIBIAN DEVELOPMENT
minative effect to differences between the nuclei. The position of
any given nucleus in one or another presumptive region is without
Fig. 14
The development of regions by dependent differentiation during the stage of
plasticity, before the onset of irreversible determination. At the early gastrula
stage, a piece of presumptive epidermis (gill region) of Triton cristatus is
exchanged for a piece of presumptive brain region of T. tae?iiatus. a, The dark
taeniatus embryo with the light cristatus graft, h. The light cristatus embryo
with the dark taeniatus graft, c, The taeniatus embryo at a later stage, with the
cristatus graft in the region of the left side of the brain, d. Transverse section
through the taeniatus embryo, showing part of the wall of the forebrain (between
X-X) formed from the grafted light-coloured cristatus tissue, which has under-
gone dependent differentiation according to its surroundings. (From Spemann,
Arch. Entwmech. XLViii, 1921.)
effect on the subsequent normal differentiation of that region. The
key to the origin of the differentiation of the various regions must
,i€li*»>>.
45
OJ
T!
fH
u
'^
C5
^
M
M
u
'C
x;
^
«
c
2
N
J2
Cv
C
o
i;
^
-T3
^
<u
f
>
• ^
c
J
£
0
X
i-(
Ch
<u
c«
~5;
-g
Vj
^
"a
5
^
; — ;
-C
"!;;
'u
t-t-(
t^
D
o
<4-(
0
c
£
a
-
c;
^
C
o
T'
C
V
^i
-o
QJ
-0
C
X
B
c
a
o
-a
£
(^
a
— '
^1
o
B
c
bJ: -S
.o
fe 'I
j::
^
b£
.-^
C
't£
OJ
^
j::
c
X
C/3
C
ja
.ti
oT
<i
^
X
^
cjc
-Ci
3
^
o
■^
C«
rC
fcJD
tC
05
c
o
C
£«
""
0
y
c
b
OJ
^
x:
u
B
OJ
w
o"
C
>>
w
;-i
a;
[-1
x:
^
flj
'-^
<u 12 -^
46 EARLY AMPHIBIAN DEVELOPMENT
therefore be looked for in some factor which affects the various
regions of the cytoplasm in general, and not of the nucleus alone.
By the method of transplantation it can be shown that, up to a
certain stage in the gastrulation of the newt, the fates of most of the
regions of the embryo are not irrevocably determined. A piece of
presumptive neural tube material removed from its embryo and
grafted into the side of another, may differentiate into the external
gills of its new host if it happened to be grafted into the presumptive
gill region of the latter. Conversely, a piece of presumptive epi-
dermis grafted into the appropriate region of the presumptive
neural tube of another embryo, will undergo differentiation into
part of the brain and the eye. Up to this stage of gastrulation, there-
fore, the regions develop according to their actual surroundings,
and regardless of their origin and former surroundings:^ they are
in fact still plastic as regards their final fate. Even the future germ-
layers are plastic up to this stage, for presumptive epidermis can
be made to differentiate into mesodermal structures such as muscle
fibres, and vice versa (figs. 14, 15 and 16).^
There comes a critical time, however, during the process of
gastrulation, after which the various presumptive regions are no
longer plastic. Their fates are then irrevocably determined, and,
whatever the position into which they may be grafted, pieces of
any given presumptive region will then undergo the differentiation
which is typical of that particular region in normal development.
Pari passu with the determination to differentiate in any given
direction goes the loss of power to differentiate in other directions.
In other words, the regions can then only develop towards their
presumptive fates. One can then, for instance, graft the presump-
tive eye region from one late gastrula into another, and obtain the
differentiation of a typical eye, facing into the body cavity (fig. 17)^.
Something invisible has happened to fix the prospective ^ates on
the various presumptive regions, and since this something must be
due, presumably, to chemical changes in the various regions, this
phase of development may be referred to as chemo-dijferentiation^
Through this process the organism has become a patchwork or
mosaic of separately determined regions. It is of some interest to
1 Spemann, 1918. - Mangold, 1924. ^ Spemann, 1919.
* Huxley, 1924; Goldschmidt, 1927; Bertalanffy, 1928.
x:
M
1
(U
o
1)
-n
c;
3
a,
M
J
5
•;:
3
D,
(3
c
o
c
^
^
o;
Dh
;_
(U
^
^§
(4-1
o
O
5ri
i+-<
q"
^-s
.a
"O
4
'Oc
(U
•
N
rt
4J
?3.
ON
bJD
>
^
•5 -^
•^
S*
VJ
o
^
o
i
t^
c
1
t3
a>
o
"
X
1
2
(U
"^
(U
'rt
^
"m
B
u
O
c
^
.
o
o'
~5;
fcJD gi
«
^
fe S
^
o
^
S
h;
CO
'^
2
.^•*«>,V
3§
-^-^
.22
a
s '^
s
2
l^
w
B'*-
'a
.
t, o
ID
.22
^
3
cc C
B
d
>. rt
CO
a;
^H
' — .
a
<u o
l^
s
B
•■^ — ^
^E
<u
"t:' c3
-a
Oh eo
TD
'a,
9
>
a
^2
QJ
a
c
3
>,.£
o
M
OJ
'o-^
o
CO tti
%
£
U
1
&«.w.i.J»^
48 EARLY AMPHIBIAN DEVELOPMENT
note that the existence of this mosaic phase, so different in its total
lack of plasticity and of power of regulation from anything known
in other stages of development, was only detected through experi-
mental analysis.
c
Fig. 17
The development of regions by self-differentiation after the stage of irreversible
chemo-differentiation. a, A piece of presumptive eye region from an early
neurula of Boyiibiiiator is grafted into the flank of another embryo of similar age, b.
c, Transverse section through the resulting embryo showing the eye-cup which
has developed by self-differentiation from the graft in its abnormal position;
oc. grafted eye-cup ; pron. pronephros. (From Mangold, Ergeb. der Biol, iii, 1928,
after Spemann.)
During the period of plasticity, chemo-differentiation sets in
progressively until irrevocable determination of the various regions
is achieved. But although the regions may still be plastic before
this critical time, in the sense that they can be made to undergo a
A PRELIMINARY EXPERIMENTAL ANALYSIS
49
^,
%fe^
^-^■
.X^-v.^^
-<i*;;.
,?fi..
..-.^^•i*.
I-
.;•?'
Fig. 1 8
The labile determination of regions, o, Morula of Triton, showing the cells
subsequently isolated indicated by lines. The remaining cells were destroyed
with needles, their contents forming a supporting and nutritive medium sur-
rounding the four living cells, b, i hour after operation, the four living cells have
divided to form eight, c, g days after operation, the cells have produced a com-
plicated structure, d, Section through c; nervous tissue («.), a lens (/.), muscle
segments (w.) and epidermis (e.) have been differentiated. It is to be noted that
no organiser was present in the explanted cells and that no gastrulation took
place ; the differentiations are therefore the effects of labile determinations of the
cells themselves. (From Holtfreter, Arch. Entwrnech. cxxiv, 193 1.)
HEE
50 EARLY AMPHIBIAN DEVELOPMENT
differentiation which they would not normally have carried out, this
plasticity does not mean that the regions are entirely indifferent.
On the contrary, experiments have shown that even at the start of
gastrulation in the newt, there is a feeble determination of the
presumptive neural tube region, in virtue of which it tends to
differentiate along the lines of its prospective fate,^ and the same
is true of other regions, as will be seen later (pp. 136, 203, figs. 18,
62,63, 64)-
§6
There is, however, one region of the amphibian embryo which
makes a very important exception to the statement that the tissues
at early stages are plastic. This is the region of the dorsal lip of the
blastopore, which has arisen from the grey crescent and is destined
to form the notochord and mesoderm (chorda-mesoderm). This is
determined from very early stages (possibly even in the fertilised
egg before cleavage has begun). When grafted into other embryos
h will differentiate in no direction other than that of its normal
presumptive fate.^
This presumptive notochord, gut-roof, and mesoderm region is
predetermined to invaginate beneath the surface. It has other
properties which are as remarkable as they are important. If a portion
of this region be grafted into another embryo in the blastula or early
gastrula stage and in any position, it will there pass below the sur-
face and proceed to induce the neighbouring host-tissues to under-
go differentiation into the main organs of an embryo, often including
neural tube and brain, eyes and ears, spinal cord, mesodermal
somites and pronephric tubules, quite regardless of what the pre-
sumptive fates of these host-tissues may have been. In other words,
the dorsal lip of the amphibian blastopore has the property of being
able to force other tissues (during their state of plasticity) to under-
go the organised differentiations and developments which lead to
the production of an embryo. For this reason, the dorsal lip of the
blastopore has received the name of organiser, as the German term
Organisator coined by Spemann may be translated^ (^gs- 1? i9> 65).
^ Goerttler, 1926; Holtfreter, 193 1 a.
" In certain conditions, as when cultivated in vitro, etc., it may give rise to
other organs, such as nervous system and gut (Holtfreter, 193 1 a).
^ Spemann and Mangold,. H., 1924.
Med. I
Q ; » (« * ^1>''^V sec. Med.
Fig. 19
The induction of secondary embryos by organiser grafts, a. Dorsal view showing
primary embryo of Triton taeniatiis, b. Side view showing secondary embryo
induced by grafting an organiser of T. cristatus (distinguishable by lack of
pigment) into the flank of a. c, Dorsal view of secondary embryo, and left side
view of primary embryo, at later stage ; note ear- vesicles of secondary embryo in
line with that of primary, d. Transverse section through c. Med. neural tube of
primary embryo; r. sec.Pron. pronephric duct; r.sec. Uw. mesodermal somite;
sec.Ch. grafted notochord; sec. D. gut; sec. A'led. neural tube; of secondary
embryo. Note that most of the structures of the secondary embryo have been in-
duced from host tissues, but that the graft has contributed to some (distinguishable
by lack of pigment). (From Spemann and Mangold, Arch. Mikr. Anat. u.
Ejitzvmech. c, 1924.)
4-2
The presence of the organiser region is essential for development, a, A newt's egg
constricted into two in the transverse plane, thus separating dorsal and ventral
halves, b, The result of isolation of a dorsal half (containing the organiser
region) : a perfect embryo, c, The result of isolating a ventral half (lacking the
organiser region) : a blastula-like ball of cells which develops no further, d, A
newt's egg constricted in the plane of symmetry, thus separating lateral halves,
each of which contains a portion of the organiser region, and, e, develops into a
perfect embryo. (From Spemann, Naturzviss. iv, 1924.)
A PRELIMINARY EXPERIMENTAL ANALYSIS 53
The vital importance of the organiser for development is shown
by the classical experiment of separating the first two blastomeres
of the newt's egg. If the plane of the first cleavage separates the
future right and left halves of the body, both blastomeres will re-
ceive a portion of the organiser region, and both will organise them-
selves and produce miniature but otherwise normal embryos.^
But if the first cleavage separates future dorsal and ventral halves,
only the dorsal half will produce an embryo ; the ventral half under-
goes cleavage and makes an abortive attempt to produce germ-
layers, but develops no further^ (fig. 20). The same is true in the
case of the frog.^
The action of the organiser raises a number of important
problems which will receive more detailed consideration in a
subsequent chapter. For the moment, attention may be focussed
on the light which these phenomena throw on the analytical study
of development.
§7
It has been seen that the newt's egg when fertilised has already had
two determinations imposed upon it : that of polarity and that of
bilateral symmetry. As a result of these determinations, one region,
the future organiser, is localised and apparently fully determined
at very early stages. Until a certain time, which is roughly half-way
through the process of gastrulation, the various other regions of the
embryo are still plastic, although they are presumably passing
through the preliminary stages of chemo-differentiation. But the
time comes when they, too, are irreversibly determined to follow
the course of differentiation which characterises each part in normal
development.
The terms mdepende?it or self-differentiation and dependent differ-
entiation were introduced by Roux to characterise these two types
or phases of diff"erentiation. In Amphibia before gastrulation, all
regions save that of the organiser show dependent differentiation :
their developmental fate is dependent upon and conditioned by
factors external to themselves — in this case the presence of an
organiser in a particular spatial relation with them. This is proved
^ Herlitzka, 1896; Spemann, 1903. 2 Schmidt, 1930, 1933.
54 EARLY AMPHIBIAN DEVELOPMENT
by the two types of experiment we have mentioned ; the grafting
of tissues into abnormal positions relative to an intact organiser,
and the grafting of an organiser in abnormal positions relative to
an otherwise intact host embryo.
However, after a certain critical time during gastrulation, the
various main regions develop, in respect of the type of tissue they
produce, by self- differentiation. A piece of tissue grafted into an
abnormal situation no longer has its fate determined by its position
in relation to other tissues ; the factors controlling its development
are now situated within itself.
Of course, all differentiation is in certain respects dependent, in
others independent. When grafts are made from one species to
another before gastrulation, the grafted piece shows dependent
differentiation as regards the organs and tissues which it forms, but
self-differentiation as regards various fundamental characters such
as cell-size and pigmentation (see p. 142). Conversely, in certain
respects the fate of a piece of tissue in the self-differentiating phase
is dependent on external conditions, for, as we shall see (p. 249),
the development of its shape is dependent on mechanical factors in
its new situation, whereas the type of tissue which it produces is
not.
In experimental embryology, the terms are generally used in
respect of dependence of type of tissue produced upon the activities
of other parts of the embryo. Dependence upon external agencies
is not usually discussed in this connexion (although some differ-
entiations such as that of polarity are dependent upon them), but
these are assumed to remain more or less constant, within the range
permitting of normal development; and form-differences due to
purely mechanical distortion are also usually omitted from con-
sideration. Within these limits, the terms will be found very useful.
Other examples of self-differentiation are to be found in the
development of the organiser region, of the eye-cup and of many
other organs mentioned in Chap, vii, and of particular types of
tumours and cancers irrespective of their site. Other examples of
dependent differentiation which will be met with are the depen-
dence of the lens and conjunctiva upon the eye-cup (pp. 178, 183),
of the ear-capsule upon the ear- vesicle (p. 175), the dependence
of amphibian metamorphosis upon a certain concentration of
''***(^P»-T**»>
'.^^^^^ ^7-> »• "I, ''"?'^"'' > .'■'
.^^
*■
. Pr.
J Li'*
Fig. 21
The dependence of lens-differentiation on the optic-cup in Triton, a, Larv^a into
which at the mid-gastrula stage a piece of presumptive brain region from another
embryo of the same age was grafted. The graft (g.) developed by self-differentia-
tion into parts of the brain, and an eye-cup which induced the formation of a lens
(/.) from the ventral trunk epidermis of the host, b, Section through the same
larva showing (v.) the vesicle formed from host tissue and containing the graft;
br. portion of grafted brain ; e. grafted tissue differentiated into eye-cup ; Nr. spinal
cord ; Pr. pronephric tubules, of host embryo. (From Mangold, Arch. Entwmech.
cxvii, 1929.)
56 EARLY AMPHIBIAN DEVELOPMENT
thyroid hormone (p. 427), the dependence of the fine structure of
bone upon the functional stresses to which it is exposed (p. 434).
The lateral line in tadpoles is independent as regards its histo-
logical differentiation and increase in size, but dependent in regard
to the position it comes to occupy (p. 355).
In first origin, each process of differentiation is dependent. As
we have seen, the differentiation of an axis of polarity is dependent
on factors in the ovary ; the differentiation of bilateral symmetry is
normally dependent on the point of entry of the sperm ; and the
localisation and determination of the organiser itself is dependent
on both the axis of polarity and the plane of bilateral symmetry, for
Fig. 22
Mosaic stage: localised determination of limb-potencies. Left: Amblystojna
embryo immediately after removal of the right fore-limb field. The pronephros is
seen in the wound-area. Right: a larva on which a similar operation has been
performed, but on the left fore-limb field, 3 months later. There is no trace of
a left fore-limb. (From Harrison, Proc. Nat. Acad. Sci. i, 191 5.)
it arises in the latter and at a particular level (or parallel of latitude)
with regard to the former. The differentiation of all other regions
is dependent on some presumably chemical action of the organiser
and on their position relative to the axis of polarity and the plane
of bilateral symmetry, though in a manner which is still obscure
(see Chap. ix).
The case of the neural folds raises a problem of particular in-
terest, for, as has been mentioned, not only can neural folds arise
by chemo-differentiation in situ even if the organiser is removed or
prevented from invaginating, but also the organiser is capable of
inducing the formation of neural folds wherever it is grafted. In
normal development, that tissue which the organiser normally in-
duced to become neural folds is also that which in the absence of
A PRELIMINARY EXPERIMENTAL ANALYSIS 57
the organiser can become neural folds by self-differentiation. There
seem, therefore, to be two methods by means of which neural folds
can arise; such a phenomenon is referred to as ''double assurance''.
Further discussion of this question is given in Chap, vi
(pp. 139, 187).
As soon as some organs have reached the stage of full self-
differentiation, they become able to induce other organs to arise by
dependent differentiation. In many forms, for instance, the eye-
cup induces the formation of a lens from the overlying epidermis
(see p. 183), in a manner analogous to that by means of which the
organiser induces the formation of neural folds (fig. 21). How
general such secondary induction may prove to be in development
is not as yet known.
However, we do know that in many cases what is first determined
is a large region or field, and that later this region becomes split up
into a further mosaic of independently determined subregions. For
instance, as set forth more in detail in Chap, vii, the limb area is early
determined as a region in the flank (fig. 22) : only later are the
various subregions, such as hand, forearm, upper arm, determined
within the main region.
During the period of self-differentiation, the embryo is thus a
patchwork or mosaic of developing regions, the differentiation or
localisation of all of them being originally dependent on something
else, ultimately on the axis of polarity and plane of bilateral sym-
metry. The differentiation, however, is progressive, the mosaic
coming to consist of more and smaller pieces, each of which eventu-
ally undergoes independent differentiation.
At this stage, almost the only integrating influences acting upon
the embryo appear to be the simple ones of mechanical construction.
Biological integration is almost absent : neither neural nor humoral
correlation is yet possible, and little trace has been detected of
influences analogous to that of the organiser or the optic-cup, or
of chemical influence by contact. The chief exception appears to be
that the polarity of the egg may persist to cause the polarisation of
some on all of the separate organ-rudiments (see Chaps, vii and x).
The embryo at this stage is like a multiple tissue-culture, the parts
58 EARLY AMPHIBIAN DEVELOPMENT
of which happen to cohere mechanically in a particular form : the
only correlations are mechanical ones.
This lack of co-ordination accounts for the fact that, whether by
regulation or regeneration, the making good of material or of parts
that have been lost appears to be impossible during this stage of
regional self- differentiation of the various organs,^ although regu-
lation was possible at the stage of the egg, blastula, and early
gastrula, and regeneration will become possible in the larva. The
loss of the earlier power of regulation seems to be due to the super-
position upon the original unitary gradient-field system of a patch-
work of independent chemo-differentiated regions (pp. 221, 350);
while the later appearance of the power of regeneration is in the
main due to the onset of growth, which in turn depends upon the
acquisition of function by the nervous and vascular systems. The
latter introduce the possibility of nervous and humoral correlation,
and further make possible the mutual interplay of the functions of
the various organs as soon as their histological differentiation has
proceeded far enough to enable their tissues to function and so
permit them to perfect their final development by functional
differentiation (see Chap. xiii).
§9
From the foregoing sketch it will be obvious that development in
Amphibia is epigenetic, and involves the creation of differentiation
afresh in each and every generation. There can be no question of
preformation, for the structures of the future organism are not
there, nor are their positions localised or determined in the un-
fertilised egg. This epigenetic character of development is based
on the capacity of the protoplasm of the egg to react in a particular
way to certain stimuli which in the first instance are external, as
when the egg-axis and plane of bilateral symmetry are induced, and
then later internal, as when the tissues are induced to differentiate
under the influence of an organiser. The whole of development is
a series of such reactions or responses to stimuli. It therefore
follows that no development can be normal in an abnormal en-
vironment, and, also, that the hereditary endowment of an organ-
^ Harrison, 1915; Spurling, 1923. (See also figs. 22, 94.)
A PRELIMINARY EXPERIMENTAL ANALYSIS 59
ism, represented by the inherited factors transmitted to it by its
parents, is by itself insufficient to account for development.
Development is always the product of an interaction between a
specific protoplasm and hereditary outfit on the one hand, and
a particular complex of environmental factors on the other. The
environmental factors operative with regard to any part of the
organism are partly those of the external world, partly those of
the internal environment provided by the rest of the organism.
Chapter IV
THE ORIGIN OF POLARITY, SYMMETRY,
AND ASYMMETRY
§1
It has been seen that when the amphibian egg is laid, all that can
be said about its future development is that the anterior end of the
animal will be formed near the animal pole, and the posterior end
near the vegetative pole. In all animals above the Protozoa, the
primary differentiation during their development is this axiation,
as Child calls the determination of the axis of polarity.
It is of great importance to realise that the factors invoked in
order to explain the determination of polarity are external to the
egg. In the sea-weed Fucus, it is found that the determination of
polarity is normally due to the direction of incident light. ^ But it
has been shown experimentally that the application of an electric
current is also capable of inducing the determination of the axis of
polarity in the egg of Fucus, in any direction, at will.^ Further, it
is found that when Fucus eggs are placed in groups very close to one
another, each egg develops a polarity in such a way that its apical
point faces away from the group. ^ Here it seems that a chemical
factor is responsible, for the COa-tension will be higher and the
oxgen-tension lower in the midst of the eggs in the group than in
the surrounding fluid (fig. 23).
One of the agencies capable of inducing polarity in the egg of
Fucus thus appears to be differential exposure to oxygen, and the
same is true of many animals. In the sea-urchin the oocyte develops
with one pole attached to the wall of the ovary, and the other
pole projecting freely into the cavity, and exposed to the ovarian
fluid and nutritive wandering cells. It appears that the centre of
this portion, where physiological exchange with the immediate en-
vironment is most active, will become the animal pole of the egg.^
Similar cases, where the attached and free surfaces of the
^ Hurd, 1920; Whitaker, 1931. ^ Lund, 1923 b.
^ Jenkinson, 1911; Lindahl, 1932.
ORIGIN OF POLARITY, SYMMETRY, AND ASYMMETRY 6]
Fig. 23
Electrical control of primary polarity in the eggs of the brown alga Fuciis. Two-
cell stages, a, Eggs subjected to appropriate current density. They are practically
all oriented with the smaller end towards the^anode. b, Eggs serving as control,
subjected to current density below the threshold requisite to secure orientation
(equivalent to a fall of 0-025 volt across the egg). The eggs point at random.
(From Lund, Bot. Gaz. lxxvi, 1923.)
62 ORIGIN OF POLARITY, SYMMETRY, AND ASYMMETRY
oocytes are exposed to different conditions, would appear to
account for the polarity of the unfertilised egg in many other forms,
e.g. Chaetopterus,^ Sternaspis,^ Cerebratulus,^ and Cyclas,^ where
Fig. 24
The primary gradient in oocytes, a, In the Annehd Sternaspi^ the oocyte is
attached by a narrow peduncle containing a vascular loop, and the nucleus is at
the opposite end, which protrudes into the ovarian cavity. The attached end
becomes the vegetative pole, the free end the animal pole, b, Gradient in
amount of yolk and size of yolk-spheres in the oocyte of the frog; p, pedicle of
attachment. (From Child, Physiological Foimdations of Behavior, New York, 1924.)
the exposure is to the ovarian fluid, and the Coelenterate Phiali-
dium,^ where the exposure is to the surrounding sea-water. In all
the above-mentioned cases the surrounding fluid, be it ovarian
C. B. Wilson, 1900.
54-
^ Lillie, 1906.
* Stauffacher, 1894.
- Child, 1915 B, p. 341. ^
^ Child, 1921 B, p.
ORIGIN OF POLARITY, SYMMETRY, AND ASYMMETRY 63
or sea-water, may be regarded as containing more oxygen than the
tissues of the ovary. But in other forms, such as vertebrates, not
only does the coeiomic fluid in such small enclosed spaces as the
ovarian cavity lose its respiratory function, but the ovary itself is
well supplied with blood-vessels, and there is therefore reason to
believe that the oxygen-supply of the tissues of the ovary is greater
than that of the fluid surrounding the ovary. It is consequently
of great interest to find that in birds the exposed side of the
oocyte becomes the vegetative pole of the egg, while the attached
side becomes the animal pole.^ The same is true in Amphioxus,
but here the attached side of the egg is turned towards the
secondary ovarian cavity which is close to the atrial cavity, from
which oxygen is probably derived. (For the frog, see p. 35.)
Asexual reproduction and regeneration phenomena also provide
a number of examples in which polarity is induced from the outside,
and such cases are, from the standpoint of general theory, as im-
portant as those concerning development from an egg. An axis
of polarity can be experimentally induced in regenerating frag-
ments of the Hydroid polyps, Obelia and Corymorpha. These
organisms are built on a radially symmetrical plan, with an axis
passing down from the oral end of the polyp along the stem. If a
piece of Obelia stem be isolated, it normally retains its polarity, as
shown by its regenerating a polyp at the original distal end earlier
than at the proximal end. But if such cut pieces are subjected to
the passage of an electric current of a certain strength through the
water in which they are lying, it is found that regardless of the
original polarity of the pieces, polyps are regenerated only at that
end which points towards the anode : while stems (or stolons) may
be formed from the end which is directed towards the kathode. This
shows that the original polarity can be overridden by external stimuli
such as an experimentally controlled electric current^ (fig. 25).
A piece of the stem of Corymorpha regenerating in normal sea-
water likewise retains its polarity, and regenerates a polyp at its
originally distal end. But if such a piece is placed in water contain-
ing weak poisons, it dedifl'erentiates and loses its form, becoming
converted into a banana-shaped mass lying on the bottom of the
vessel. Replacement in clean water will lead to regeneration of a
^ Conklin, 1932. ^ Lund, 1921, 1923 A, 1924.
64 ORIGIN OF POLARITY, SYMMETRY, AND ASYMMETRY
polyp, not, however, from either of the original ends, but from the
central portion of the piece which is farthest from the glass bottom
of the vessel and most freely exposed to the water and therefore to
oxygen.^ In this case, an original polarity-gradient has not merely
H-
g^^^^^^
Fig. 25
Control of polarity by external agencies in Hydroids. A series of internodes of
Obelia regenerating towards the anode when exposed to the passage of a weak
electric current. The control series at the same stage had all regenerated hydranths
at both cut ends. (From 'Lund, Jour n. Exp. Zool. xxxiv, 1921.)
been reversed, but the original polarity has been obliterated, and
a wholly new polarity induced (fig. 26).
The winter-buds of the social Ascidian Clavellina appear to be
irregular aggregations of cells with no relation to the polarity of
their parent. The polarity of the zooids which later arise from these
^ Child, 1925 B, 1927.
ORIGIN OF POLARITY, SYMMETRY, AND ASYMMETRY 65
buds must therefore be imposed on them from without.^ An even
more striking example is the masses produced by the joining up
of cells and cell-groups after the tissues of sponges and hydroids
have been strained through fine gauze, ground up with sand, or
otherwise dissociated. Here, clearly, all traces of the original
polarity must have been lost. However, the masses may later
develop into miniature sponges with polarity of their own. This
^i^
h
Experimental imposition of a new primary axis in Corymorpha. In fragments of
stem immediately after cutting {a) or after regeneration to form biaxial (c) or
single hydranths, immersion in dilute alcohol causes dedifferentiation {b, g). On
replacement in sea-water, redifferentiation occurs with a new axis at right angles
to the old, with apical region at the centre of the free surface {c, d, h, i). (From
Child, Physiological Foundations of Behavior, New York, 1924.)
polarity must have been induced by external factors.- Similar results
have been obtained with hydroids (figs. 27, 132; see also p. 281).
Thus, apart from the cogent theoretical reasons advanced by
Child, there is abundant evidence, experimental and circum-
stantial, for the view that the initial determination of an axis of
polarity, or axiation, is due to the action of factors external to the
developing organism.
^ Huxley, 1926; Brien, 1930.
2 H. V. Wilson, 1907, 191 1 ; Child, 1928 b; Huxley, 1911, 1921 a.
HEE 5
66 ORIGIN OF POLARITY, SYMMETRY, AND ASYMMETRY
Once the axis of polarity has been determined in an egg, it often
becomes manifested by a stratified and graded distribution of egg-
contents, some of which may be visibly distinct, such as pigment,
fat, yolk, etc. (e.g. Arbacia). It is to be noted, however, that this
stratification of materials is only an effect and not a cause of polarity.
iirS''
te
-02?
X'
/^l
^
Fig. 27
Differentiation after dissociation in the hydroid Pennaria. The dissociated cells
united to form rounded reconstitution-masses which surround themselves with
perisarcs {op) and later form outgrowths which give rise to stolon-like structures
{x) and normal hydranths. (After H. V. Wilson, from Gray, Expermiejital
Cytology, Cambridge, 193 1.)
A completely new restratification can be induced in any direction
by means of the centrifuge, but development continues to be
governed by the original axis of polarity.^ It is probable in these
cases (Arbacia) that the polar organisation of the egg, once it is
determined, resides in the cortex and an invisible internal frame-
work of more viscous cytoplasm which resists the displacing action
^ Morgan and Spooner, 1909.
ORIGIN OF POLARITY, SYMMETRY, AND ASYMMETRY 67
of centrifugalisation. In eggs which contain a large quantity of
yolk {Crepidula, Styela, Rand) on the other hand, it seems that the
viscous cytoplasmic framework can be permanently distorted and
changed by displacement of the egg-contents as a result of pro-
longed centrifugalisation or inversion^ (see also pp. 94, 218, 313).
§2
After the fixation of the axis of polarity, the most important deter-
mination in animals with bilateral symmetry is the determination of
the plane of the latter. In the frog, this is normally due to the point
of entry of the sperm. Before fertilisation, the ^gg is capable of
forming its plane of symmetry in any one of an infinite number of
possible planes passing through the egg-axis of polarity ; the actual
determination of a particular plane is fixed from the outside. The
matter has been considered in detail in Chap, iii (p. 36). We may
sum up our conclusions as follows. The machinery for realising full
normal bilateral symmetry is inherent in the egg ; even very slight
differential action of various external agencies can act as a trigger
permitting a particular plane of symmetry to realise itself: normally,
the entry of the sperm provides a strong diflPerential which readily
overrides the influence of other agencies.
The formation of the grey crescent in the amphibian egg ap-
pears to be bound up with the establishment of an activity-gradient
of some sort extending dorso-ventrally across the equator of the
egg. The existence of this gradient is revealed by various facts. In
the first place, cleavage in the animal hemisphere proceeds more
rapidly in the dorsal meridian, so that at the close of segmentation
there is a slight gradient in cell-size from dorsal to ventral along
each circle of latitude. In the second place, there is the fact that
gastrulation and invagination is initiated in the dorsal lip region,
and then spreads progressively round each side until it reaches the
ventral meridian, and the blastopore lip becomes circular.
Thirdly, there are the results of susceptibihty experiments.
These show that in anuran eggs exposed to lethal low temperatures
or lethal concentrations of KCN, NH4OH, HgCl2, and other toxic
agents, disintegration at any level begins at or near the dorsal
meridian, and extends thence round the egg towards the ventral
^ Conklin, 1924.
5-2
68 ORIGIN OF POLARITY, SYMMETRY, AND ASYMMETRY
side. Further, in sub-lethal concentrations, the dorsal regions are
the most inhibited in their differentiation.^
This last method allows us to make a further statement, namely
that the dorso-ventral activity-gradient becomes progressively more
intense (steeper) between fertilisation and gastrulation. In just-
fertilised eggs, disintegration in lethal concentrations begins at the
animal pole and then spreads along the dorsal side : in some cases
a second centre of disintegration appears in the region of the grey
Fig. 28
Differential susceptibility in a frog's egg exposed to KCN from the 2-cell
stage: disintegrated cells are shown light. The animal pole area (central) has
disintegrated ; also an area of cells, near the equator on one side, in the future
organiser region. (After Bellamy, Biol. Bull, xxxvii, 1919, modified.)
crescent before the primary disintegration has spread to this area.
During cleavage, the susceptibility of the dorsal lip region in-
creases, until in late blastulae this region begins to disintegrate
before or at the same time as the apical pole. In gastrula stages, the
dorsal lip region is always the first to disintegrate.^ It is probable
that this process is correlated with the acquisition of organiser
properties by the dorsal (grey crescent or dorsal lip) region (fig. 28).
The method by which bilateral symmetry is determined in the
egg of Echinoderms is still problematical- ; but the localisation of the
plane can be revealed by susceptibility experiments at a stage before
any bilateral symmetry is visible in the embryo.^
Further, a labile determination of bilateral symmetry has been
^ Bellamy, 1919; Bellamy and Child, 1924.
2 Horstadius, 1928. ^ Child, 1916A.
ORIGIN OF POLARITY, SYMMETRY, AND ASYMMETRY 69
discovered even in the egg of sea-urchins. If the egg is exposed to
certain anaerobic conditions, a pit is formed in a particular place,
but disappears on returning the egg to normal conditions. By means
of ifitra vitam stains, it has been shown that the site of the pit
coincides with the ventral side of the future larva.^
At the same time, as in the case of the amphibian tgg, this labile
determination of the plane of symmetry can be overriden by a
variety of factors, of which, however, the point of sperm-entry is
not one. 2 Artificial stretching and deformation of the tgg (in a
direction making some fairly large angle with the axis of polarity)
leads to the determination of the dorso-ventral axis along that of
tension. The primary axis of polarity is unaffected.^
Artificial rearrangement of the egg-contents has also been shown
to influence the localisation of the plane of bilateral symmetry. In
the sea-urchin, Psammechmtis miliaris, the presence in the ripe egg of
a subcortical layer of lipoid granules has been observed* and they
may be displaced by means of the centrifuge. Neglecting those
cases in which the granules are heaped up at either the animal or
vegetative poles, it is found that the meridian of the egg on which
the granules are accumulated becomes the ventral side of the larva. ^' ^
Similarly, the visible granules of the egg of Arbacia can be con-
centrated on any meridian, which then becomes the ventral side of
the larva^ (see also p. 218). The dorso-ventral axis, it seems, is
determined as a gradient with high point ventrally. The curious
fact that the ventral surface is associated in Psammechiniis with
centripetal lipoid granules, but in Arbacia with centrifugal
yolk-particles, can be explained if their concentration leads to
relatively higher metabolism. Similarly the ends of the stretched
egg appear to be in a peculiar labile active condition. Interesting
possibilities of analysis are here opened up.
In some Echinoderms, the dorso-ventral axis is visible in the
unfertilised egg (Psoitis phantappusy and marked by a particular
distribution of yolk, or (Asterma gibbosa)^ by an elongation of the
egg. The latter state of afl^airs is also found in some insects and
^ Orstrom, in the press. 2 Horstadius, 1928.
^ Boveri, 1901 ; Lindahl, 1933 a. ^ Runnstrom, 1924.
^ Runnstrom, 1925. ^ Lindahl, 1933 b.
' Runnstrom and Runnstrom, 1921. ^ Horstadius, 1925.
70 ORIGIN OF POLARITY, SYMMETRY, AND ASYMMETRY
Cephalopod Molluscs, where the egg is not only polarised, but
bilaterally symmetrical in shape while still in the ovary.
It is of considerable theoretical importance to find that one and
the same determination can occur either before or after fertilisation
in different forms. In Fucus, no axes of symmetry at all are deter-
mined until after fertilisation. In most animals, radial symmetry
and the primary axis are determined before fertilisation, bilateral
symmetry at or after fertilisation. In some insects and Cephalo-
pods, the determination of bilateral symmetry too has been shifted
back to the period before fertilisation, and takes place under the
influence of ovarian factors. It will be seen later that a similar
shift in time-relations has occurred as regards the processes of
chemo-differentiation in a number of forms, and that this shift
contributes to the difference, which long puzzled experimental
embryologists, between so-called "regulation-eggs" and ''mosaic-
eggs" (see Chap, v).^
§3
A further problem of symmetry is the determination of bilateral
asymmetry. There are certain animals in which nothing is known
as to the embryological determination of asymmetry, e.g. the skulls
of certain whales and owls with asymmetrical formation of some of
the bones ; the bill of the wrybill plover ; the various insects with
spiral torsion of the genitalia;^ the fish Anableps in which the
copulatory tube points either left or right in males, and the genital
aperture faces right or left in females ; ^ the flatfish, in which either
the left or the right side becomes uppermost when the fish is lying
on the sea bottom, and the structure of the head is modified ac-
cordingly ; or Amphioxus, in which the larval stages are markedly
asymmetrical.
^ In his large book on experimental embryology, Schleip (p. 842) argues at
some length against the idea that the primary axes of the egg are imposed from
without, and supposes that they arise by self-differentiation, though they may
be modified by external agencies. It is logically almost impossible to conceive
how a non-polarised fragment of living matter can acquire polarity by self-
differentiation; and further, the experimental evidence in certain cases strongly
supports the view that external differentials are responsible (Fiicus, egg; Cory-
viorpha, redifferentiation). The concept advanced above, that very slight external
differentials may serve to release the capacity of the egg to develop polarity, re-
conciles both views. The type of the polarity is predetermined and therefore
"self-differentiating"; the direction of the polarity is determined from without.
2 Richards, 1927. ^ Garman, 1895, 1896.
ORIGIN OF POLARITY, SYMMETRY, AND ASYMMETRY 71
It is clear that a fandamental difference must exist between the
eggs of bilaterally symmetrical and those of radially symmetrical
forms. In the former case, however, what is given by genetic con-
stitution cannot be bilateral symmetry per se, but the capacitv of
developing such symmetry in relation to various external agencies.
Harrison (1921 A, 1925 a) has suggested that the ultimate capacity
for developing symmetry-relations is linked up with the intimate
properties of the protoplasm and the "space-lattice " formed by its
constituent parts.
The asymmetry of the large chelae found in many Crustacea
either in one or both sexes, and also that of the opercula in certain
a
Fig. 29
Cleavage asymmetry in Molluscs. The position of the large mesoderm cell (^d) is
reversed in laeotropic and dexiotropic cleavage. (From Morgan, Experimental
Embryology, Columbia University Press, igzj.)
tubicolous Annelid worms, involves special problems of relative
growth-rate, which are discussed by Przibram (193 1 a).
The most marked asymmetry known is that of Gastropod Mol-
lusca, most of which manifest a marked torsion of the internal
anatomy together with unequal development of many paired organs.
In addition, a large number of forms have their shells twisted into
a spiral, which is usually dextral. Here it has been shown that the
dextral or sinistral type of structure is under the control of Men-
delian factors, whose action, however, is delayed for a generation^
(see Chap. xii). The asymmetry of the a^dult is determined not by its
own genetic constitution, but by that of the oocyte from which it
arose, before it underwent the reduction divisions. The cleavage
^ Boycott, Diver, Garstang, and Turner, 1930.
72 ORIGIN OF POLARITY, SYMMETRY, AND ASYMMETRY
of the Gastropod egg is of the spiral type, and it has been shown
that the direction of the initial spiral cleavage division is opposite
in dextral and sinistral races of snails.^ It is therefore probable that
the orientation of the spindles of the first spiral cleavage division
is responsible for the determination of the type of asymmetry which
<^^
Fig. 30
Above, a sinistral (left) and a dextral (right) Gastropod shell. Below, corre-
sponding asymmetry of cleavage. The obliquity of the spindles in the 2-cell
stage (centre) and the portion of the cross-furrows in the 4-cell stage (bottom)
are reversed in eggs with the left-handed (laeotropic) and the right-handed
(dexiotropic) type of cleavage. (After Morgan, Experimefttal Embryology (Figs. 80,
79, 78 c and c\ pp. 256-7), Columbia University Press, 1927, modified.)
the adult will exhibit,- and that the orientation of the spindles
is, in turn, controlled by the Mendelian factors present in the
oocyte (figs. 29, 30).
Reversed spiral cleavage has been observed exceptionally in the
development of the leech Clepsme,'^ but as the adult is apparently
perfectly symmetrical, no subsequent effects of the reversed cleav-
age can be detected. Occasionally, the leech egg gives rise to a
1 Crampton, 1894. - Conklin, 1897. ^ Miiller, 1932.
ORIGIN OF POLARITY, SYMMETRY, AND ASYMMETRY 73
double monster, apparently by the production of two D-cells in
place of one (see p. io8). In such cases the direction of spiral
cleavage is reversed in the right-hand D-cell and all other cells on
the right side of the plane of bilateral symmetry.
In the Echinoderms, most of the larval forms are asymmetrical,
in that the left, but not the right, coelomic pouch acquires a water-
pore placing it in communication with the exterior. Further, the
fates of the various right and left coelomic pouches are very differ-
ent. As a result, the hydrocoel and the rudiment of the body of
the adult Echinoderm are formed on the left side of the body of
the larva. It will be best to postpone the analysis of conditions
in this group until the state of affairs in Vertebrates has been
considered.
All Vertebrates are in reality asymmetrical. The stomach projects
to the left of the middle line, while the heart and intestine show
spiral twisting and are asymmetrical in other ways. The asymmetry
of the gut and heart of the newt and frog has been experimentally
shown to be dependent on a factor situated in the gut-roof. At the
stage when the neural folds are still open, a square piece of pre-
sumptive neural tube material, together with the underlying gut-
roof, is removed from the dorsal side, about half-way down the
length of the embryo. The square piece is rotated through i8o° and
grafted back into place again so that the antero-posterior axis of the
piece is reversed. The result of such an experiment is a normal
embryo, except that it shows situs inversus of the asymmetrical
organs, i.e. the stomach is on the right and the intestine and heart
are twisted in the direction opposite to the normal. Rotation of the
presumptive neural tube material alone, without the underlying
gut-roof, does not interfere with the development of the normal
asymmetry. The ventral regions of the embryo are not touched by
the operation, and therefore the asymmetry of the heart and gut
must be determined by some factor or agency differentially dis-
tributed across the gut-roof^ (fig. 31).
If, however, the square piece which in the previous experiments
was rotated, is simply removed, the embryo will show normal
asymmetry. This may mean either that the differential factor ex-
tends, though with diminished intensity, on either side of the gut-
^ Pressler, 191 1; Meyer, 1913; Spemann, 1918.
ORIGIN OF POLARITY, SYMMETRY, AND ASYMMETRY
.«
P. o
o
O
3
o o <U ^
"5 'S '^
111 1^ .'^ •
8 c ^ -^
S C '^
M
t-i i-i ~
4-> ra ^ .^
«J « & -2 k^
!^ -^ =^ 'o •
^ CS !<
•S > u o
■■+-1 ^Ln •
o ^ .ii
3 rt C3 M 5
fe
^, A
, rota
ated V
Ddm.
1
^ u .^
M
C+: :3 jj c^
=3
o o a 3
M
O X. o^
g
>
at-r
L cu
ryo
bul
&;■ S X5 .
M
<U r- ~0
3
^M<S<^
G
on t
has
um;
M
»
S "1::; <" "C
^
$ O -73 -M
5 O •- c«
;^ U 73
u
visce
gut-
tral
Atr.
13
§
ing
ven
us; .
CO
S >:• oj G
u^ i: -c rt
.■^
o «J -*^ .
^
ence ^
h und
from
le. A
CO
C ;> 5 M
o
<D > ^ ^
c«
a
s^
^^1 ^
u
g;
^
hS 8-5
I
ORIGIN OF POLARITY, SYMMETRY, AND ASYMMETRY 75
roof, or that other factors exist capable of controlling asymmetry,
but normally overruled by the gut-roof factor.
i\nother line of attack on this problem is provided by those ex-
periments in which a blastula of a newt is partially constricted by
tying a fine hair round it in the plane of bilateral symmetry. The
result is the production of double-headed monsters, and, while the
Fig. 32
Anterior doubling producing situs inversus viscerum et cordis in the right-hand
member. The doubhng was produced by partial constriction in the plane of
symmetry of an early cleavage stage of Triton. The heart, gut, and position of
liver (L.) and pancreas (P.) of the right-hand member (seen on the left in this
ventral view) are reversed. (After Spemann and Falkenberg, Arch. Entwmech.
XLV, 1919, simplified.)
left-hand member of such a pair always shows the normal asym-
metry, the right-hand member nearly always shows situs inversus }
Double-headed monsters also occur in trout, in wild conditions and
in hatcheries. When the two members are joined together only by
the hinder region of the trunk (behind the abdominal cavity), both
members have the usual vertebrate asymmetry. But when the join
between the two members is farther forward, so that the alimentary
1 Spemann and Falkenberg, 19 19.
76 ORIGIN OF POLARITY, SYMMETRY, AND ASYMMETRY
tract forks at a point between the stomach and the cloaca, the right-
hand member frequently shows situs inversus, while the left-hand
member is normal^ (fig. 32).
The rudiment of the heart can be divided in amphibian embryos
(at the tail-bud stage) by means of a longitudinal cut in the middle
line; each half rudiment will give rise to a heart, and while the
asymmetry of the left one is normal, that of the right one is
reversed^ (fig. 115).
The remarkable point about these experiments and observations
is the constancy of normal asymmetry in the left-hand member,
and the restriction of situs inversus to the right-hand member. This
fact emerges still more clearly from those experiments in which the
blastula of the newt is constricted by a hair in the plane of bilateral
symmetry, and the hair is pulled tight, thus resulting in the com-
plete separation of two half-blastulae, of which one represents the
right and the other the left half of the original embryo. The left
halves develop into perfect little newts with normal asymmetry ; of
the right halves, about equal numbers show normal asymmetry and
situs inversus respectively.^
Whatever the asymmetry factor may be, it cannot be regarded as
an absolute and localised producer of one specific type of asym-
metrical structure — at least, not during the earliest stages of de-
velopment— and for the following reasons. It is true that when
newt embryos are divided into left and right halves at the blastula
stage, about half of the right-hand portions show reversed asym-
metry. But if the left and right blastomeres are separated from one
another (likewise by constricting in the plane of bilateral symmetry
with a hair) at the 2-cell stage, the right-hand blastomeres do not
show any greater tendency to production of reversed asymmetry
than is found to be the case in normal development of newts' eggs
— 2 to 3 per cent.* At the 2-cell stage, therefore, the asymmetry
factor has not become predominant on the left side at the expense
of the right. The same conclusion emerges from the simple ex-
periment of reversing an egg and forcing it to continue its develop-
ment in that position. If the prepotent normal-asymmetry factor
were definitely located on the left side at this stage, then since the
1 Stockard, 1921 ; Morrill, 1919; Swett, 1921. - Ekman, 1924, 1925.
3 Ruud and Spemann, 1923. * Mangold, 1921 b.
ORIGIN OF POLARITY, SYMMETRY, AND ASYMMETRY 77
plane of bilateral symmetry is already determined at fertilisation,
rotation of the egg so that the vegetative pole becomes uppermost,
while the dorsal meridian remains unchanged, would cause the
original left side of the egg to become the right side of the embryo.
But as a matter of fact, the embryos arising from such reversed eggs
do not show situs inversus.^
The asymmetry factor must therefore be regarded as a factor
which results in a greater activity of the tissues on one side (the
left) of the body as compared with the other : such greater activity
developing progressively. It is interesting to note that a similar
progressive accentuation of a differential or gradient is to be ob-
served in the dorso-ventral axis of the amphibian egg, between the
time when the grey crescent is first formed and the establishment
of full organiser capacity in the dorsal lip (see p. 68). Further
research should be directed to discovering whether such accentua-
tion or steepening of activity-gradients is a regular feature in their
development.
The result of the action of this asymmetry factor is seen in the
more rapid growth and differentiation of the left side, as regards
certain organs. Experimental proof of this is provided by the meso-
dermal rudiments from which the muscular wall of the heart is
formed. These rudiments are at first situated on each side of the
body, and later on move towards the middle line. But if the rudi-
ments are removed from the embryo while they are still lateral in
position, and are made to develop in isolation (explantation in ecto-
dermal jackets, in suitable culture media), the remarkable thing is
that the rudiments from the left side show pulsations while those
from the right do not. Further, the histological differentiation of
which the rudiments from the right side are capable is inferior in
degree to that of the rudiments from the left.^ The difference be-
tween the left and right rudiments of the heart is a physiological
one, and appears to be quantitative rather than qualitative, and in
every case the left side is prepotent (fig. 33).
If the asymmetry factor is, as suggested, concerned with the
relatively greater activity of one side, it should be susceptible of
experimental control. One method of affecting its action is the
simple mechanical one of removing some of the material from the
^ Hammerling, 1927. " Goerttler, 1928.
78 ORIGIN OF POLARITY, SYMMETRY, AND ASYMMETRY
left side of an otherwise normal developing embryo. This experi-
ment has been performed on the blastula of the newt, and resulted
in the production of situs inversus.^ Another method is to subject
the developing embryo to physical factors which are calculated to
affect the rate of activities of the tissues, and to direct these physical
Fig. 33
a, Dorsal ; b, Ventral ; c, Left side views of an embryo of Urodele to show the
position of the paired rudiments of the heart in the mesoderm beneath the
surface, d, Embryo from which the dorsal surface has been cut off and the entire
gut peeled out, thus revealing the ventral mesoderm, with the position of the
paired heart rudiments indicated by circles. The left rudiment, when isolated,
develops further than the right. (From Goerttler, Verb. Anat. Ges. xxxvii, 1928.)
factors in such a way that one side of the embryo is affected more
than the other. This would appear to be why chick embryos show
situs inversus when they are locally damaged by overheating on the
left side during incubation. ^ Here, the intensity of action of the
tissues on the left side has been decreased, while in the previously
described experiment it is their amount which has been reduced.
^ Wilhelmi, 1920, 1921. ^ Warynsky and Fol, 1884.
ORIGIN OF POLARITY, SYMMETRY, AND ASYMMETRY 79
Now, since the embryos that develop from eggs in which the
plane of bilateral symmetry has been artificially selected (by con-
trolling the point of sperm-entry; see p. 36) show the normal
asymmetry, it follows that the determination of the left as the ulti-
mately prepotent side must be made at the same time as that of the
plane of bilateral symmetry. And if this determination were due
to an external factor, it would be impossible to understand why it
invariably acts so as to produce its effect on a meridian 90° right
of the meridian of sperm-entry, and convey a greater power of
activity to this eventual left side of the embryo. The conclusion is
therefore enforced that the determination of the left-right axis is
in some way connected with that of the plane of bilateral symmetry,
and is the result of some factor acting within the egg.
There is as yet no indication of how this factor acts, but it may
be pointed out that the determination of a third axis (in the case of
the egg, the left-right axis) as a consequence of the determination
of the other two axes of space (in the egg, the primary egg-axis,
i.e. antero-posterior, and the dorso-ventral), is a phenomenon not
without an analogy in the inanimate world. It is well known that
if a conductor carries an electric current through an independent
magnetic field of force which is orientated at right angles to the
conductor, then the conductor will be subjected to a force acting
at right angles both to the magnetic field and to the conductor. If
it be imagined that the magnetic field is vertical with the North
Pole uppermost, and a horizontal conductor carries an electric
current away from an observer, the force acting on the conductor
will tend to displace it to the observer's left. It is not pretended
that the egg-axis is the site of a magnetic field, nor that the dorso-
ventral axis is a simple conductor. But the physical analogy de-
scribed above does show how it is possible to obtain a determina-
tion of a third axis, and a polarisation in it, as a consequence of the
determination and polarisation of axes in the other two planes of
space.
In the larva of Amphioxus, asymmetry is very marked. In this
form^ double monsters can be artificially produced by disarranging
the 4-cell stage. In such cases, both components always show the
normal asymmetry : symmetry is never reversed. The diflPerence
^ Conklin, 1933.
8o ORIGIN OF POLARITY, SYMMETRY, AND ASYMMETRY
between these results and those on the newt is doubtless to be ex-
plained as a result of the precocious appearance of bilaterality in the
egg of Amphioxus, extending to chemo-diiTerentiated substances,
which is established immediately after fertilisation. As suggested
Fig. 34
Double monster of Amphioxus, produced by mechanical disarrangement and
partial separation of the blastomeres in the 2-cell stage. Note that both com-
ponents show normal asymmetry, i^, ist gill-cleft; eg, club-shaped gland;
m, mouth; p, preoral pit. (From Conklin, Journ. Exp. Zool. lxiv, 1933.)
in the preceding paragraph, this rigid bilaterality might establish
an equally rigid asymmetry-gradient (fig. 34).
§4
The conclusions arrived at from a consideration of the results
obtained from experiments on amphibian development are sup-
ported and extended by experiments on Echinoderm larvae. In
ORIGIN OF POLARITY, SYMMETRY, AND ASYMMETRY 8l
these, also, it is the left side of the body which is prepotent as com-
pared with the right, and this prepotency manifests itself in the fact
that the hydrocoel, the water-pore, and the rudiments of the adult
animal are formed on the left side of the body of the larva. In
Asterina, the gastrula can be divided into left and right halves by
section in the plane of bilateral symmetry. The left halves develop
into larvae with normal asymmetry : they have a hydrocoel on the
left. The right halves can do one of three things ; they may have a
hydrocoel on the left side only ; they may have a hydrocoel on the
right side only ; or they may have a hydrocoel both on the left and
on the right side.^ This last condition is sometimes found in other-
wise normal echinopluteus larvae,'^ and can be experimentally
produced by subjecting the larvae to hypertonic sea-water,^ while
it is the rule in ophioplutei.
The occurrence of halves produced from the right side of the
original larva, which develop a hydrocoel both on the left and on
the right, is of great interest, for it provides a situation which could
scarcely be realised in the amphibian embryo. There, the gut and
heart must be twisted either one way or the other, but cannot be
twisted in both ways at once in the same embryo.^
Both this result on Echinoderm larvae, however, and that ob-
tained by dividing newt blastulae into right and left halves (p. 76),
can be plausibly explained along similar lines. It has already been
found necessary to postulate a main activity-gradient, concerned
with asymmetry, and extending transversely across the body from
left to right. This is presumably superposed on minor activity-
gradients extending inwards from the surface towards the centre of
the embryo. In any case, when the developing tgg is cut in half, the
inner surfaces of each half are damaged or interfered with, and their
activity reduced. In the left-hand halves, the effect of this will
merely be to steepen the existing asymmetry-gradient ; all resulting
organisms will therefore be of normal asymmetry. In the right-hand
halves, however, the effect will be in the contrary direction to that
1 Horstadius, 1928. 2 MacBride, tgii. 3 MacBride, 1918.
* In the Gastropod Limncea (see p. 411) occasional specimens have been bred
in which the dextral and sinistral forces are so delicately balanced that the result
is an animal with a flat shell coiled in one plane, like that of Planurbis. Most of
these specimens are abnormal in their anatomy and die early (Boycott, Diver,
Garstang and Turner, 1930).
82 ORIGIN OF POLARITY, SYMMETRY, AND ASYMMETRY
of the main asymmetry-gradient. If the result is merely to flatten
this gradient, animals of normal asymmetry will still result. But if
the effect of the cut is strong enough to reverse the existing gradient,
animals of reversed asymmetry will arise. This applies both to
Echinoderms and to Amphibia: in addition, in Echinoderms the
almost complete balancing of the two lateral halves of the gradient
will give rise to bilaterally symmetrical forms (with both left and
right hydrocoels), whereas this result is impossible in Amphibia,
where the normal and reversed asymmetries are mutually exclusive
alternatives.
The main points of this chapter may be briefly summarised as
follows. In amphibian development, polarity or axiation and bi-
lateral symmetry are both established as the result of agencies
external to the egg. In both cases, an important effect of these
agencies is the production of activity-gradients extending through
the whole egg. In the production of bilateral asymmetry, an
activity-gradient is also involved. At the moment it is not possible
to state what is the originating cause of this asymmetry-gradient ;
we do know, however, that its establishment is in some way de-
pendent upon the establishment of the dorso-ventral gradient
which determines bilateral symmetry. Both these latter gradients
appear to become progressively accentuated during the period of
cleavage.
Chapter V
CLEAVAGE AND DIFFERENTIATION
§1
The most obvious visible change during the first phase of develop-
ment of the fertilised egg is its cleavage into a number of separate
cells. We must now ask whether other equally important but less
obvious changes may not be taking place at the same time, and en-
quire into the relation between cleavage and the processes leading
to morphological differentiation.
The pattern of cleavage is normally oriented in relation to the
existing major axis of the egg, e.g. the first two cleavage planes are,
in all known cases except one, meridional: the exception is pro-
vided by the Nematodes, where the plane of the first cleavage is
still oriented with reference to the axis, but at right angles to it,
and therefore latitudinally. In Cephalopods and Ascidians, the
cleavage pattern is oriented with reference to the secondary axis
of bilateral symmetry as well.
The orientation of cleavage-pattern can, however, be modified.
It may be modified in relation to a new, induced, axis of polarity.
For instance, in the sea-urchin Lytechinus and the star-fish Patina,
cut fragments of the unfertilised egg, subsequently fertilised, al-
ways have the first two cleavage planes at right angles to the cut
surface, which, as we shall see later (p. 313), has established a new
polarity.^
The cleavage-pattern may also be modified by mechanical means,
e.g. by a restratification of the egg-contents by the use of the centri-
fuge (see p. 218). In the sea-urchin Arbacia, for instance, the first
two cleavages are perpendicular to the stratification, whatever its
relation to the original axis.^ Or the cleavage-pattern maybe altered
by forcing eggs to undergo cleavage while compressed between
glass plates. The orientation of the division spindles in these cases
is governed by the principle known as " Hertwig's rule ", which lays
down that at mitosis the spindle will form with its long axis in the
^ Taylor and Tennent, 1924; Taylor, Tennent and Whitaker, 1925; Taylor
and Whitaker, 1926. 2 Morgan and Lyon, 1907.
6-2
84
CLEAVAGE AND DIFFERENTIATION
direction of the longest axis of the cytoplasm of the cell.^ The
distortion occasioned by the glass plates causes the third cleavage
plane to form meridionally instead of latitudinally^ (figs. 35, 36).
One of the consequences of the experiments of forcing eggs to
cleave under compression is that the normal distribution of the
Fig. 35
Disarrangement of cleavage by pressure in sea-urchin eggs, a-d, Normal
cleavage. e,f and i-k. Flat plates arising from cleavage under pressure, g, h, Sub-
sequent cleavage of/ when released from pressure. /, The same for k. The late
cleavage stages are drawn with the vegetative pole uppermost. The disarrangement
of the nuclei does not prevent the development of normal plutei. (After Driesch,
from Morgan, Experimental Embryology , Columbia University Press, 1927.)
cleavage nuclei is altered, but subsequent development is normal
in spite of the fact that a number of nuclei find themselves in
blastomeres other than those in which they would be situated in
^ Although this rule is of very general application, there are some notable ex-
ceptions to it. For instance, cleavage in the star-fish Patiria occurs in relation to
the polarity of the egg, whether original or induced by operation (see p. 313),
even when the egg is deformed by pressure (Taylor and Whitaker, 1926). Other
exceptions are found in the first cleavage of Ascaris eggs, and in the divisions of
the cells forming the germ-bands of Crustacea (see Jenkinson, 1909B, p. 34).
- Frog, Hertwig, 1893; sea-urchin, Driesch, 1893.
CLEAVAGE AND DIFFERENTIATION 85
normal cleavage. As already mentioned (p. 43), these facts prove
that during cleavage the nuclei divide in such a way that their
daughter-nuclei are quantitatively and qualitatively equal.
A more recent and very elegant demonstration of the equivalence
of the nuclei of the blastomeres was carried out as follows. By
X^
^^N.
rv^
^-\
U^
^V
\f^
'rB^
'^Z-
'--^f^
«^
XeA
JAy><,
■^ \
t1
7.\
y^2A
t1
l\
6A^
n54
u
\(^A
4/^N^
S^^
"^f^
^
7S
SB
Fig. 36
Diagram to show the altered distribution of nuclei in frogs' eggs made to segment
under pressure. A, Normal eggs. B, Eggs subjected to pressure. Left, 8-cell
stage; right, i6-cell stage. In each case a polar view is shown above, a side view
below. Cells produced by the division of corresponding cells are numbered alike.
(From Wells, Huxley and Wells, The Science of Life, London, 1929.)
means of a fine hair, the fertilised egg of the newt can be con-
stricted into the shape of a dumb-bell, in such a way that the zygote
nucleus is confined to one side. This side will then undergo cleavage
as the nucleus divides, while the other side of the dumb-bell will
remain uncleaved. By releasing the ligature, the constriction can
be relaxed, and one nucleus — any one, at random — may be allowed
86 CLEAVAGE AND DIFFERENTIATION
to pass across from one side to the other. This may be done at the
2-cell, 4-cell, 8-cell, i6-cell, or 32-cell stage of the cleaved side. If
the constriction lay in the plane of bilateral symmetry of the original
fertilised tgg, and if after the passage of one nucleus from the
cleaved side to the other the ligature is then drawn tight again so
as to separate the two halves completely, each half will develop into
a normal little newt. One of these little embryos will contain only
the nuclear material of one blastomere of the normal 2-, 4-, 8-, or
i6-cell stage, or as we may for brevity write it, a 1/2, 1/4, 1/8 or
1/16 nucleus, depending on the time when the two halves were
separated ; the other embryo will contain all the rest of the nuclear
material. This means that in normal development, the nuclei of the
blastomeres of the i6-cell stage contain material which is equivalent
to that of the nucleus of the fertilised egg. Nothing has been lost
by the nuclei in the process of cleavage, at least up to and including
the i6-cell stage. Further, all the 1/16 nuclei have retained this
equivalence, for in the numerous experiments performed it would
not have been possible for the nucleus which passed across from
one side to the other to be the same^ (fig. 37).
When cleavage in one-half of the dumb-bell has reached the 32-
cell stage, the passage of a nucleus into the other half is insufiicient
to enable the latter to undergo normal development. This is, how-
ever, probably not to be attributed to a qualitative insufficiency of
a 1/32 nucleus. It is more likely that the failure to develop is due
to some alteration of the cytoplasm of the uncleaved half, in turn
due to the length of time during which it has been deprived of a
nucleus, and therefore prevented from prosecuting its normal
physiological activities. This explanation follows from the fact that
a 1/16 nucleus is incapable of ensuring development beyond the
late gastrula, or, rarely, early neurula stage, if the constriction had
been placed in such a way as to separate dorsal and ventral halves
of the future embryo, and the zygote nucleus had been restricted
to the ventral half. A 1/16 nucleus is therefore unable to do in a
previously enucleate dorsal half what it can do perfectly well in a
previously enucleate lateral half. It would appear that the failure
in this case lies with the cytoplasm. The susceptibility of the cyto-
plasm of a dorsal half is greater than that of a lateral half (see
^ Spemann, 1928.
CLEAVAGE AXD DIFFERENTIATION
87
Fig. 37
The equality of nuclear division during cleavage. A fertilised egg of Triton
taeniatus was constricted by a ligature, restricting the nucleus to the right-hand
half, in which cleavage has reached the 8-cell stage, while the left-hand half is
still undivided. A, At the i6-cell stage one of the cleavage-planes coincides with
that of the ligature, and 1/16 nucleus has passed across into the as yet undivided
left half. B, The ligature was then drawn tight so as to effect complete separation
between the two (lateral) halves. C, Each developed into a perfect embrvo (one
slightly further advanced than the other) ; 140 days after the operation, they were
identical. A nucleus of the i /16 stage is therefore equivalent to that of the whole
egg. (From Spemann, Zeitschr. Wiss. Zool. cxxxii, 1928.)
88 CLEAVAGE AND DIFFERENTIATION
p. 68, and Chap, ix) ; accordingly the latter can survive absence of
a nucleus during the time required for the zygote nucleus to divide
four times, while the former loses its capacity for complete de-
velopment if it has remained enucleate for a longer time than that
required for three divisions of the zygote nucleus.^
The equivalence of nuclei at later stages of cleavage has been
established from experiments conducted on the eggs of insects. The
egg of the dragon-fly Platycnemis is an elongated structure in which
the nucleus is central and divides several times before its products
of division reach the surface of the egg and the cytoplasm is par-
titioned off into separate blastomeres. By focussing a pencil of
ultra-violet rays on a nucleus at the 2-nucleus stage (corresponding,
of course, to the 2-cell stage of forms with ordinary cleavage) it is
possible to kill it. But the remaining nucleus and its products of
division are sufficient to allow a normal embryo to be formed.
The insect egg is further peculiar in that it possesses near its hind
end a region which is essential for the subsequent differentiation
of the embryo (see Chap. vi). But the activities of this region are
not manifested unless some of the nuclei which have resulted from
cleavage migrate into it. This "population" of the hinder end, and
indeed of all the surface of the egg, by nuclei, normally takes place
after the 5th cleavage, corresponding to the 3 2-cell stage. Again,
by means of ultra-violet rays, it is possible to affect a zone of cyto-
plasm of the egg in such a way that the products of division of the
nuclei are delayed in passing through it, and instead of receiving
nuclei after the fifth cleavage, the hinder end only receives them
after the eighth cleavage, i.e. at the 256-cell stage. Nevertheless,
these nuclei are adequate to activate the region in question, and
normal embryos are produced. Here, then, is evidence that the
division of the nuclei is qualitatively equal as far as the 256-cell
stage.^
^ We have already noted that an isolated ventral half, since it does not contain
any of the organiser-region, is incapable of development beyond a stage roughly
equivalent to the late blastula. It might be supposed therefore that nuclei which
had been restricted to a ventral half had been in some way affected so as to be
unable to promote full development on passage into a dorsal half. There is,
however, no positive evidence for such a possibility, while the greater suscepti-
bility of the dorsal half of the egg is an established fact (Spemann, 1901B, 1902
1903, 1914, 1928; Ruud and Spemann, 1923).
2 Seidel, 1932.
CLEAVAGE AND DIFFERENTIATION 89
§2
It is clear from these experiments that whatever the first mani-
festation of differentiation in the embryo may be, it is not to be
found in the division of the nuclei of the blastomeres during
cleavage. Attention must therefore be turned to the cytoplasm, in
order to see whether it, too, is equivalent in the different blastomeres.
Considering first the case of the newt : the fact that a lateral or a
dorsal half of an egg at the 2-cell stage, a blastula or an early
gastrula will develop, but that a ventral half will not (p. 53), shows
that all the regions are not equivalent; and since this non-equiva-
lence cannot reside in the nuclei, it must concern the cytoplasm.
Actually, the importance of the orientation of the constriction
separating the halves in the experiments described above, has been
shown to lie solely in the fact that the presence of some of the
organiser area (grey crescent, dorsal lip region) is essential if
development beyond the blastula stage is to take place. When the
constriction coincides with the plane of bilateral symmetry, the
halves will be lateral and each will possess a portion of the region
of the grey crescent. But if the constriction is at right angles to the
plane of bilateral symmetry and separates a dorsal half from a ventral
half, the former will contain the w^hole
of the region of the grey crescent and --««^'
will develop, while the latter will not -rj
contain any portion of the region in
question and will not develop.
A ventral half of an embryo (blastula „. ^
or early gastrula) can be made to ^, ^ .■ ' c u
•^ , ° ^ . The lormation or an embryo
develop if the dorsal lip of the blastO- with neural tube, somites, and
pore of another embryo is grafted into notochord, out of a ventral half
: , , 1 • 1-11 (see fig. 20, c) of a Tn^o« embryo,
lt,l and this proves conclusively that by grafting an organiser into it.
the inabihtyof a ventral half to develop (From Bautzmann, H., Arch.
is due, not to lack of any nuclear ^'^^^^"'^^^'- ^^' ^927.)
material or factors, but to lack of a definite portion of cyto-
plasm— the organiser (fig. 38).
In the newt, therefore, there is already a differentiation of the
cytoplasm just after fertilisation and before the first cleavage, and
^ Bautzmann, 1927.
90
CLEAVAGE AND DIFFERENTIATION
this differentiation is of such a kind that a certain cytoplasmic
region is essential for development. There is no other qualitative
difference between the blastomeres of the 2- or 4-cell stage, as is
shown by the fact that a single blastomere of the 4-cell stage will —
provided that it contains a portion of the region of the grey crescent
— develop into a complete but miniature embryo, although in
normal development this blastomere would have furnished material
for only one-quarter of an embryo.^ Further, two embryos of the
-cr
Fig. 39
Diagram showing results of uniting pairs of Triton eggs in the 2-cell stage. The
future organiser-region (dorsal lip) is represented as a black crescent. Above, con-
dition when the first cleavage of both divides dorsal from ventral halves. Below,
condition when the first cleavage of both is median and divides right and left
halves. The result expected is a multiple monstrous form with three components.
(After Seidel, from Morgan, Experimental Embryology, Columbia University
Press, 1927, modified.)
newt placed together cross-wise over one another at the 2-cell stage
can undergo development to form one single large embryo, provided
that the grey crescent regions of both are adjacent. (If these regions
are not adjacent, a double monster is produced'^ (%s. 39, 40).)
The findings in the newt can be extended to other Urodela, and
in them it can be said that except for the determination of the region
of the grey crescent which will give rise to the dorsal lip of the
blastopore, or organiser, the cytoplasm of the ^^g is not unequally
distributed between the blastomeres up to the end of the 4-cell
stage inclusive. When animal and vegetative regions are divided, as
^ Ruud, 1925. ^ Mangold, 1921A; Mangold and Seidel, 1927.
CLEAVAGE AND DIFFERENTIATION 9I
happens at the 8-cell stage, it might be expected that this separa-
tion of different portions of the primary egg-axis would mean an
unequal distribution of potencies, and, as will be seen later, this
expectation is in fact realised.
The case of the frog is in principle similar to that of the newt.
But the experiments conducted from time to time on the frog have
suffered so much from unforeseen complications, that the con-
clusions drawn from them w^ere for a long time misleading. The
chief difficulty arises from the fact that the eggs of Anura have
long defied attempts to secure the constriction and separation of
blastomeres. Experimental technique has therefore been largely
Fig. 40
One embryo from two eggs. Left, two Triton eggs in the 2-cell stage are laid
across each other, so that their blastomeres alternate. Centre, each blastomere
has divided once. Right, a giant neurula resulting from such a fusion. (After
Seidel, from Morgan, Experimental Embryology, Columbia University Press,
1927, modified.)
restricted to injuring one of the blastomeres : this is usually accom-
plished with a hot needle. The result of such an experiment at the
2-cell stage is that the uninjured blastomere develops into a half-
embryo, and does not produce much more than it would have done
if its sister-blastomere had developed normally alongside it, for it
is a condition of the experiment that the injured blastomere re-
mains attached to the uninjured one.^ (For the present purpose,
the subsequent attempt of the half-embryo to complete itself by
" post-generation "'-^ may be passed over here as irrelevant (fig. 41).)
1 Roux, 1888.
^ In some of the cases originally described by Roux, the half-embryo obtained
by injuring one blastomere with a hot needhe appeared to be subsequently con-
verted into a whole embryo, by the utilisation of the materials of the injured
blastomere. To this restorative process, the name "post-generation" was given.
It was imagined that the reorganisation of the injured blastomere was brought
about either by belated cleavage of its nucleus, or by invasion of cells from the
uninjured half, or by overgrowth of the injured half by layers of tissue from the
92
CLEAVAGE AND DIFFERENTIATION
The result of this experiment was at first interpreted to mean that
the two blastomeres were already differentiated at the first cleavage
and were determined to give rise, each of them, to one-half of the
future embryo. But this conclusion was later shown to be errone-
ous in a number of ways. In the first place, it was noticed that the
half-embryo which developed might be a lateral half, or a dorsal
half, or an oblique half, according as to whether the plane of the
A, Lateral, and B, anterior, partial embryos of the frog produced from eggs in
which one of the first two blastomeres have been killed but allowed to remain in
place. (After Roux, from Morgan, Experimental Embryology, Columbia Uni-
versity Press, 1927.)
first cleavage coincided with, or was perpendicular, or oblique, to
the plane of bilateral symmetry. The alleged determination of the
blastomeres at the 2-cell stage was therefore not constant.^ Then
it was found that if a normal embryo at the 2-cell stage is inverted,
each of the two blastomeres will then develop into as much as it
can of a complete embryo. The limitations on completeness are due
uninjured half, or by a combination of these methods. Morgan (1895) was unable
to confirm these findings, and the position is still obscure. Discrepancies be-
tween various results seem to be due to the relative degrees of injury inflicted by
the hot needle. Where the coagulation of the protoplasm is extensive and cleavage
of the injured blastomere cannot proceed, it is unlikely that the half-embryo ever
becomes complete, although it may appear to be more complete than it really is,
as a result of the spreading of the epidermis from the uninjured half and con-
sequent concealment of the underlying defects. If, on the other hand, the cleavage
of the injured blastomere is only delayed but it nevertheless reaches the blastula
stage by the time that the uninjured half is ready to gastrulate, the rapid restora-
tion of the missing half would be possible. At all events, the theoretical arguments
originally based on the alleged phenomenon of post-generation have long ceased
to be important.
^ Hertwig, 1893; Brachet, 1903, 1905.
CLEAVAGE AND DIFFERENTIATION 93
to purely mechanical reasons ; and the result is the formation of a
double monster. This means that each blastomere at the 2-cell
stage of the frog is capable of giving rise to more than it would
produce in normal development, and therefore the various regions
of the egg cannot all be determined at this stage. ^
It is clear, therefore, that it is not the mere presence of the injured
blastomere, when the latter is pricked with a hot needle at the 2-cell
stage, which prevents the other blastomere from developing into
a complete embryo. This is still more evident from the experiment
in which one blastomere of the 2-cell stage is injured as before, and
then the embryo, injured and uninjured blastomeres together, is
inverted and maintained in that position. The uninjured blastomere
will then develop into a more or less complete embryo.'^ The in-
version results in a streaming of the contents of the uninjured
blastomere so that the yolk again becomes undermost, and it is to
this rearrangement that the power of developing into a whole
emibryo on the part of a single blastomere must be ascribed. It
must be because there is no such rearrangement in the case where
a blastomere is injured and the embryo is not inverted, that the
uninjured blastomere in such an experiment develops into a half.
The presence of the injured blastomere necessitates the retention
of the hemispherical shape on the part of the uninjured blastomere,
and no possibility is provided for the rearrangement of its contents,
which appears to be necessary if the half is to regulate into a whole.
Indeed, it is difficult to see what kind of stimulus other than in-
version could upset what in the uninjured blastomere is merely the
continuation of normal development. In one case, two frog's eggs
were found enclosed within one membrane, which deformed both
of them into a hemispherical shape. In the subsequent develop-
ment each embryo was deficient on the flattened side.^
In another anuran, Chorophilus, it has been found possible to
remove the injured blastomere altogether by sucking it out with a
fine pipette, and the uninjured blastomere then develops into a
whole embryo, presumably as a result of the rearrangement of its
contents, for after removal of its injured sister the uninjured blasto-
mere becomes spherical.^ Lastly, improved technique has made
1 Schultze, 1894; G. Wetzel, 1895. ^ Morgan, 1895.
3 Witschi, 1927. ^ McClendon, 1910.
94 CLEAVAGE AND DIFFERENTIATION
it possible to separate the blastomeres of the frog at the 2-cell stage,
and it has been found that each blastomere thus isolated can (pro-
vided that it contains a portion of the grey crescent region) develop
into a whole embryo.^
It will be remembered, as explained in Chap, iii, that grafting
experiments have shown in the newt that the various regions
(except that of the organiser) are plastic up to a certain stage in
gastrulation, and that tissue which was presumptive epidermis can
differentiate into part of the brain and eye. Recent improvements
in technique have permitted of analogous experiments on anuran
material, and it has been found that the tissues of the frog (again
with the exception of the organiser) are plastic up to a similar stage.^
Cleavage of the egg of Anura, then, does not result in the
separation of qualitatively unequal cytoplasm between the blasto-
meres, certainly of the 2-cell stage, and presumably of the 4-cell
stage, with the exception of the specialised region of the grey
crescent. In this respect, therefore, the anuran egg does not differ
from that of the Urodele.
A few more words may be added concerning the cause of the
production of the double monsters from embryos of the frog which
have been inverted at the 2-cell stage. It has been found possible
to obtain such monsters by inverting the undivided egg, and there-
fore the duplicity of the monsters is not due to the number of
blastomeres into which the egg has cleaved when it is inverted.
Triple monsters may also arise from inversion. These anomalies
have been shown to be due to the fact that when the streaming of
the yolk takes place, consequent on the inversion, a streak of inert
^ Schmidt, 1930, 1933.
2 Schotte, 1930; Schmidt, 1930.
The question of the existence at early stages (fertilised but unsegmented egg)
of anuran development of cytoplasmic regions possessing a determination has
been attacked by the method of making small injuries with a heated or unheated
needle. Loss of tissue (by damage in situ or by extra-ovation) at this early
stage leads to the development of imperfect larvae, and it has been held that even
the unsegmented egg possesses (labile) determinations (Brachet, 1905, 1906, 191 1,
1923, 1927 ; Pasteels, 1932). But it is necessary to point out that the eventual mal-
formation or non-appearance of an organ after injury to the egg is, by itself, no
logical justification for the view that the rudiment of the organ in question was
determined at the stage operated upon: the injury done to a particular part of an
egg persists, and may exert an inhibiting influence on the subsequent determination
and chemo-differentiation of whatever rudiment comes to occupy its site. On the
other hand, clean removal of pieces of blastulae (other than the organiser-region)
in Bombinator and Triton allows normal development to occur (Bruns, 193 1).
CLEAVAGE AND DIFFERENTIATION
95
yolk is left near the surface, and this interferes with the process of
gastrulation. The lip of the blastopore becomes as it were split on
this obstacle, and invagination takes place in opposite directions,
away from the streak of yolk. In other cases, the blastocoel is dis-
placed, and it seems that the pressure within it causes the cells which
form its wall to present an obstacle on which the blastopore lip
'X^/i
^'1as;«g(»^-i
VI^
Fig. 42
Double monsters of duplicitas critciata type, produced by inverting the 2-cell
stage of frogs' eggs. (After Schleip and Penners, from Morgan, Expemnental
Embryology , Columbia University Press, 1927, fig. 157, p. 393.)
becomes split, and likewise forks. Each portion of the blastopore lip
then invaginates on its own, and gives rise to the essential features
of an embryo, in so far as this is mechanically possible^ (fig. 42).
Double monsters have also been obtained in the frog simply by
fertilising over-ripe eggs. The cleavage of such eggs is abnormal in
that the blastomeres of the vegetative hemisphere are relatively
much too large. Presumably, the physiological condition of such
^ Penners and Schleip, 1928.
96
CLEAVAGE AND DIFFERENTIATION
eggs involves a decrease in the activities of the cytoplasm, or, in
other words, a relative increase in the inertia of the mass of yolk.
At all events, the splitting into two of the blastopore lip has been
observed in such eggs, at the onset of gastrulation.^ An additional
result obtained in these experiments is cases of disorganised growth,
leading to tumour-like proliferations, which increase at the expense
of the embryo itself, may give rise to metastases, and can be
propagated by grafting (fig. 43).
Fig. 43
The effect of delayed fertilisation in frogs' eggs. Duplication, teratological
monstrosities, and tumour-like growths in tadpoles derived from late-fertilised
eggs (over 3 days over-ripe). Top left, anterior duplication, the upper head im-
perfect, with single sucker. Top right, tadpole with irregular tail and rudimentary
secondary ("parasite") head. Below, larva with much reduced head and
tumorous growths ventrally. (Redrawn after Witschi, Verh. Naturfursch. Ges.
Basel, XXXIV, 1922.)
§3
Turning now to the experiments of isolating blastomeres in other
groups of animals, it was found that the results differ considerably
in the various groups. In some forms, the isolated blastomere de-
velops into a complete and normally proportioned larva, differing
from a normal larva merely in its small absolute size. In other
forms, the isolated blastomere is incapable of doing this, and gives
rise to a partial structure only. As extreme examples of these two
types we may take the Hydrozoa and the Ascidians, respectively.
^ Witschi, 1922, 1930.
CLEAVAGE AND DIFFERENTIATION
97
In the Hydrozoan Clytia, for example, if the blastomeres are iso-
lated at the 4-cell stage, all four of them can give rise to complete
little planula larvae which then settle down and develop into
hydroid polyps.^ To a certain extent, this totipotence of Hydroids
continues up to the i6-cell stage, at which isolated blastomeres can
still produce larvae, though apparently not polyps: whether this
Fig. 44
Sea-urchin gastrulae and plutei from a whole egg (left) and a 1/2 blastomere
(right). The latter are normal except in size. (From T. H. Morgan, Sci. Monthly,
XVIII, 1924, p. 532.)
is due to lack of material or to a real restriction of potency is
obscure. In the Ascidian Styela, on the other hand, even the first
two blastomeres, if isolated, will produce only half-embryos.^ It
is true that the ectoderm grows over the whole surface of the half-
embryo, that its notochord develops. to form a normally shaped but
half-sized rod, and that there is some rounding off of the general
form. But in its essentials, the organisation is that of a left or right
1 Zoja, 1895, 1896; Maas,
~ Conklin, 1905, 1906.
^905.
HEE
98 CLEAVAGE AND DIFFERENTIATION
half. The same mosaic development is seen in anterior and
posterior 2/4 halves (fig. 45).
In the early period of study of experimental embryology, these
two types were sharply distinguished from one another as "regula-
tion-eggs " and "mosaic-eggs " respectively. Later work has, how-
ever, shown first, that all forms do not fall into one or the other of
two sharply marked categories, but that the two extremes are con-
nected by a complete series of intermediate steps ; and secondly, that
at least two very distinct processes impeding complete regulation
maybe operative in "mosaic-eggs" (pp. 105, 108). Furthermore, it
appears that all developing organisms at some stage of their career
possess the power of regulation, but lose it at some later stage. Thus
the distinction between " regulation-eggs " and " mosaic-eggs " loses
a great deal of its theoretical importance, and if the terms are to
continue being used, it is best that they should be employed in a
purely descriptive sense with reference to their behaviour during
cleavage.
The most extreme case of regulation is that of the Hydrozoa,
already cited, in which single blastomeres from either the animal
or the vegetative regions of the egg will develop into larvae as if
they were whole eggs. But in a number of forms, the diflFerentiation
along the main axis of polarity of the egg is sufficiently fixed by the
time of fertilisation to render this impossible, while differentiation
round the main axis is still absent or so slight as to permit of regu-
lation in a fragment containing all levels of the egg along its main
axis.
In most eggs, latitudinal division does not occur until the third
cleavage (leading from the 4- to the 8-cell stage), and this means
that isolated 1/8 blastomeres, or isolated animal or vegetative
halves, will be unable to give rise to whole larvae, whereas 1/2 or
1/4 blastomeres, or isolated lateral halves, will be capable of com-
plete regulation. This is the case, for instance, in Echinoderms^
and to a certain extent Nemertines'^ (fig. 44).
The eggs of Amphibia approach this last type, but the capacity
of their blastomeres to achieve complete development is limited
by the restriction of organiser capacity to the dorsal side. The
organiser-region is determined at fertihsation, and therefore 1/2 or
1 Driesch, 1900. ^ E. B. Wilson, 1903; Zeleny, 1904.
CLEAVAGE AND DIFFERENTIATION
99
B
Fig. 45
Mosaic development in the Ascidian Styela. (Compare fig. 59.) Two blasto-
meres of the 4-cell stage have been killed in each case. In A and B, these
are the two left blastomeres, resulting in the formation of (A) a right half-
gastrula, (B) a right half-larva with one muscle rudiment {ms.) and mesenchyme-
rudiment (tnch.). In C and D, the two posterior blastomeres have been killed,
leading to the formation of anterior half-embryos with complete neural plate but
no muscles. In E and F, the two anterior blastomeres have been killed. E, The
segmentation is typical of a posterior half. F, A posterior half-embryo is pro-
duced with complete muscle-rudiments but no neural plate or notochord. The
only regulation is the overgrowth of the ectoderm and the form-regulation of the
notochord. (From Conklin, Chap, ix of Cowdry, General Cytology, Chicago,
1924.)
7-2
100 CLEAVAGE AND DIFFERENTIATION
1/4 blastomeres which lack a portion of the organiser will go no
further than the germ-layer stage (unless a foreign organiser is
grafted into them). Something similar to the conditions in Am-
phibia is found in Amphioxus, where it has been shown that the
blastomeres are totipotent at but not beyond the 2-cell stage. ^ This
restriction is due to the localisation of chemo-differentiated sub-
stances necessary for mesoderm formation in the ventral meridian,
and of other substances necessary for notochord and neural plate
formation in the dorsal meridian. The fertilised tgg is thus bilaterally
symmetrical with regard to those chemo-differentiated substances
it contains ; and, since the first cleavage always occurs in the plane
of bilateral symmetry, the 2-cell stage is therefore the latest at
which the blastomeres can contain all levels of the main axis, and
therefore all these various substances (see below, p. 123).
As already mentioned, the plastic stage of development, in which
regulation is still possible, comes to an end in Amphibia at about
the stage of mid-gastrulation. A similar state of affairs, though the
precise moment has not been so accurately determined, appears to
hold good for other vertebrates; e.g. in fish (Funduhis) defect-
experiments on stages prior to the formation of the germ-ring
(i.e. early gastrulation) give rise to defects in the size of the resultant
embryos.'^ Other experiments have shown that qualitative irre-
versible differentiation begins only when the embryonic shield has
reached a distinct size — i.e. some time after the beginning of gastru-
lation. In birds, it is known from experiments (see Chap, vi, p. 161)
in which an organiser is grafted beneath another blastoderm and
there induces the formation of neural folds that irreversible deter-
mination has not yet set in after 22 hours' incubation, but as no
interchange experiments have been performed, it is not known at
what stage determination of the various regions is definitely fixed.
In this connexion it should be mentioned that isolation experiments
demonstrate the "competence" (Waddington, 1932) to differen-
tiate into various structures, but they give no information as to
whether the power to differentiate into any other structures has
been lost. In mammals, nothing has as yet been experimentally
determined with regard to these points.
^ Conklin, 1924, 1933.
2 Lewis, 1912.
CLEAVAGE AND DIFFERENTIATION lOI
A Special case is found in the Nematoda (Ascaris). Here the first
cleavage division is latitudinal, at right angles to the main axis, and
separates animal and vegetative portions. The developmental
potencies of the blastomeres have been tested by killing unwanted
ones with ultra-violet light. It is then found that the surviving
blastomeres develop just as they would have done under normal
conditions,^ and produce anterior or posterior half-embryos. How-
ever, by means of the centrifuge, the first cleavage division in these
eggs may be made to pass meridionally, and then both of the first
two blastomeres will develop a set of reproductive organs, i.e. will
produce more than they would normally have produced.^ The
regulative capacity of the Nematode egg before cleavage is shown
in the fact that fusion may occur between two eggs, which can then
regulate to form a single giant embryo of normal proportions.^ The
inclusion of the Nematode egg among "mosaic-eggs" is therefore
merely a consequence of the fact that in this group the first cleavage
division is latitudinal (see p. 398, and figs. 192, 193).
The Echinoderms are of further interest in this respect. In some
forms, such as the star-fish Patiria and the sea-urchin Lytechinus,
at the earliest stage the apico-basal differentiation is absent, or at
least ineffective (see p. 313); both animal and vegetative halves of
unfertilised eggs, subsequently fertilised, are capable of giving rise
to normal miniature larvae.'* This is, however, not the case in
another sea-urchin, Paracentrotiis lividiis, for here, the apical organ
and stomodaeum-forming potency is restricted to the animal half,
and the gut-forming potency to the vegetative half of the un-
fertilised tgg. The Qgg can be cut into two equatorially, and then
both halves fertilised. The animal half will give rise to a blastula
with long cilia, the ciliation covering an abnormally large area and
thus forming a very diffuse apical organ, but such larvae have no
gut and no mesenchyme. The vegetative half will produce a larva
with a normally tripartite gut, mesenchyme and skeletal spicules
(the latter without any regular arrangement or orientation), but
without stomodaeum, cilia, apical organ, or arms. The same is true
for 4/8 animal and vegetative fragments. Animal and vegetative
^ Stevens, 1909. - Boveri and Hogue, 1909,
^ Zur Strassen, 1898.
* Taylor and Tennent, 1924; Taylor and Whitaker, 1926.
102 CLEAVAGE AND DIFFERENTIATION
halves isolated at later (blastula) stages show the same develop-
mental potencies, the only difference being^ that a stomodaeum is
formed in such animal halves (see p. i66, and fig. 46).
Already in the unfertilised egg of this sea-urchin {Paracentrotus),
therefore, the cytoplasm of an animal half, which represents only
presumptive epidermis and other epidermal structures, lacks the
potency to form an enteron, while that of a vegetative half is in-
capable of forming an apical organ or stomodaeum.^ There is there-
fore an important distribution of potencies along the primary egg-
a b
Fig. 46
Partial larvae from fragments of sea-urchin eggs, a, Blastula with abnormally
extensive apical organ, derived from animal half of unfertilised egg, fertilised egg,
or young blastula. Note absence of gut, mesenchyme, stomodaeum. b. Ovoid
larva, without stomodaeum, apical organ, ciliated band or arms, derived from
vegetative half of unfertilised egg, fertilised egg, or blastula. Note presence of
spicules and tripartite gut. (From Horstadius, Acta Zool. ix, 1928.)
axis, and it is because they include all the levels of this axis that
the blastomeres of the 2- and 4-cell stages of Paracentrotus and
meridional halves of gastrulae (see p. 81) are totipotent. This dis-
tribution of potencies along the egg-axis has been further analysed by
studying the developmental potencies of pieces smaller than halves.
At the 32-cell stage, the cells of the animal half of Paracentrotus
form two plates or discs of mesomeres, one above the other. They
may be designated as an. i and an. 2 (presumptive ectoderm). The
cells of the vegetative half (at the 64-cell stage) also form two discs
of macromeres, which may be referred to as veg. 1 (presumptive
ectoderm) and veg. 2 (presumptive endoderm). Lastly, at the ex-
treme vegetative pole of the egg, there are the micromeres (pre-
sumptive primary mesenchyme). Accordingly, the egg of Para-
centrotus can be divided latitudinally into five layers, each of which
^ Horstadius, 1928.
CLEAVAGE AND DIFFERENTIATION 103
is capable of being isolated, at the 32- or 64-cell stage, and studied
in respect of its developmental potency ^ (fig. 47).
An isolated an. i disc develops into a blastula covered all over
with the long stiff cilia characteristic of the apical organ. An ap-
parent regulation later occurs in that these sensory cilia are lost and
replaced by mobile short cilia, with which the larva swims about.
An isolated an. 2 disc develops into a blastula, three-quarters of
the surface of which are at first covered with the large stiff cilia. In
both of these two cases, a true pluteus larva is never formed.
An isolated veg. i disc develops into a larva which may or may
not possess an apical organ. Ordinary cilia are present, and a small
gut is invaginated.
An isolated veg. 2 disc produces a larva without an apical organ
but with cilia, and a gut is invaginated which may become tripartite
in the normal manner.
The micromere group when isolated produces a ball of cells
which soon falls apart. Disc veg. 2 together with the micromeres
produces a larva in which the gut is so disproportionately large that
it fails entirely to invaginate: instead it protrudes outwards and
forms a so-called exogastrula.
It is clear, therefore, that not only are the potencies of the animal
half different from those of the vegetative half, but that these
differences are graded along the main axis of the ^^g.
As a result of this differentiation, whereas two sea-urchin eggs or
blastulae can give rise to a single double-sized larva when united
with their primary axes parallel, union with divergent axes results
in a double monster. The same applies to the results of uniting
two previously separated 1/2 blastomeres- (fig. 48).
The Echinoderms present another curious phenomenon. Iso-
lated 1/2 or 1/4 blastomeres, though they give rise to whole larvae,
cleave as parts (see p. 128); e.g. a 1/2 blastomere will form four
mesomeres, two macromeres, and two micromeres, just as it would
have done if it had been left forming part of a whole ^gg : the early
blastula too is clearly a half and not a whole. If the consistency of
the cytoplasm in the developing Echinoderm tgg were so stiff as to
prevent a half or quarter blastula, produced in this way, from
rounding up into a sphere, the fragment could not have formed a
^ Horstadius, 1931. - References in von Ubisch, 1925.
104
Mesomeres
Macromeres
Micromeres
Sensory Cilia
of
Apical Organ
Invag-inated
Enter on
micromeres
\
Veg.2.
plus micromeres
oo
oo o
ooo
Exogastrula
Fig. 47
Diagram illustrating the developmental potencies of isolated fragments of the
sea-urchin embryo, representing different levels along the egg-axis. Note distri-
bution from animal to vegetative pole of potencies for the formation of apical
organ, ectoderm with cilia, and invaginated endoderm. The position of the
third (equatorial) cleavage plane varies; when nearer to the animal pole, an
isolated veg. i disc has more animal and less vegetative potencies, which it
shows by forming an apical organ and not forming a gut ; when the cleavage plane
is nearer to the vegetative pole, an isolated veg. I disc invaginates a small gut
and forms no apical organ. (Original, based on Horstadius.)
CLEAVAGE AND DIFFERENTIATION IO5
whole larva, but would have been forced to continue development
as a part.
It appears, in point of fact, that one of the reasons for mosaic
development from egg or blastula fragments is extreme viscosity
of the cytoplasm.
Il)||||lil(l|!l!;!lll!lllillllll
Fig. 48
Diagram to show the influence of the primary axial gradient in fusion- experi-
ments with sea-urchin eggs. Left, single egg and resultant pluteus. Centre, two
eggs united with their axes parallel produce a single pluteus. Right, two eggs
united with their axes at an angle produce double monsters. (From Przibram,
Handb. norm, and pathol. Physiol, xiv, 1925, fig. 411, p. 1099.)
A good example of this is found in Ctenophores. Here the adult
has eight swimming plates or costae. But although in these forms
the first two cleavage divisions are meridional, larvae developed
from 1/2 blastomeres have only four costae: i/8 and 1/4 blasto-
meres give larvae with one and two costae respectively.^ In the un-
cleaved tgg of Beroe, there is a complete and uniform peripheral
layer of a clear substance which appears green by dark-field illu-
mination. By an elaborate series of changes, due to streaming move-
ments of the peripheral zone, and to alternation of more viscous
and less viscous phases in the general cytoplasm, the end of cleavage
sees this green substance lodged in the micromeres and forming
their entire contents, while none of it remains in the macromeres.
The micromeres give rise to the ectoderm, including the costae, -
and contain some materials, precociously chemo-differentiated
in the green substance, needed for costa-formation (figs. 49, 50).
At the beginning of each cleavage division during the early
1 Fischel, 1898. 2 spek, 1926.
io6
CLEAVAGE AND DIFFERENTIATION
•Fig. 49
Mosaic development of Ctenophores. A, i6-cell stage divided into two equal
lateral halves. B, Partial larvae developed from these halves; each has four
costae. c, cilia; e, endoderm;/, fold in egg-case showing line of division of the
halves ; /?, egg-case ; tn, stomach ; ;/, nerve centre ; ot, otolith ; p, pigment ; r, costae.
(After Fischel, from Schleip, Determination der Primitivetitzvicklung, 1929, fig. 18,
P- 5I-)
CLEAVAGE AND DIFFERENTIATION 107
Stages, the green substance is largely localised at one end of the
blastomeres. During this period the cytoplasm is highly viscous :
it then becomes more fluid, and the green substance is redistributed
uniformly round each blastomere. After the 8-cell stage, however,
it remains localised near one pole, and is progressively separated
off by a series of unequal cell-divisions into the micromeres.
If the egg is cut at stages up to the 8-cell stage, the result will
depend on two factors : first, whether the distribution of the green
D
Fig. 50
Mosaic development of Ctenophores. A, B, Fragments of i6-cell stage, divided
unequally, so that A has five macromeres and five micromeres; B, three macro-
meres and three micromeres. C, D, Partial larvae developed from A and B,
respectively; C has five costae, D has three. (After Fischel, from Schleip, Z)£'/^r-
mination der Primitiventwicklung, 1929, fig. 19, p. 52.)
substance at the moment is uniform over the surface of the blasto-
meres, or if it is asymmetrically localised ; and secondly, whether
the egg is in a more fluid state when redistribution of the green
substance is easy, or in a very viscous state when redistribution
may be impossible before the next cleavage. These facts account
for the certain amount of regulation which has been obtained in
some experiments on Ctenophore eggs. Immediately after being
laid, the tgg of Bero'e is in a highly viscous state, but with the ap-
proach of the first cleavage division it becomes more fluid. If in
I08 CLEAVAGE AND DIFFERENTIATION
the former period portions of cytoplasm are removed from the egg,
some of the costae may be entirely absent ; if, on the other hand,
portions (even quite large) of cytoplasm are removed from the egg
in the latter period (which, it may be noted, is also later in time),
none of the costae are absent, although they may be small. ^
Another interesting example in which viscosity plays an im-
portant part is provided by the Ascidian egg. The unfertilised egg
is very fluid, and, indeed, as will be seen below (p. 119), extensive
internal rearrangements of the contents take place at fertilisation.
But 10 minutes after fertilisation, the cytoplasm takes on a high
degree of viscosity ; this is reduced for a short period at 40 minutes
after fertilisation, and then rises again.^
§4
In Beroe, in addition to a variable high viscosity, we find, as men-
tioned above, the precocious formation, prior to fertilisation, of
certain specific substances, which are apparently of an "organ-
forming " nature. As we shall see, precocious chemo-differentiation
of such substances is universal among so-called mosaic-eggs. As
a result of this precocity in their formation, the specific organ-
forming or morphogenetic substances are already formed in the
just-fertilised egg, instead of being produced only after gastrulation
as in Amphibia. If these morphogenetic substances are distributed
unevenly during cleavage, mosaic development is the result. One
of the classical illustrations of this is the Mollusc Dentalium.
Dentalium is an example of that group of animals which exhibit
the remarkable form of determinate segmentation known as spiral
cleavage, to be found in most Molluscs and many worms. It will
be advisable to give a brief general description of this type of cleav-
age before continuing our discussion of Dentalium. The first two
cleavages are meridional, and are often unequal, so that one of the
cells at the 4-cell stage (blastomere D) is larger than the other three
{A, B, C). The next cleavage is latitudinal but very unequal,^
1 Yatsu, 1912A, b; Spek, 1926. ' " Dalcq, 1932.
3 The inequality which characterises these cleavage divisions seems to depend
on a gradient of permeability extending through the cytoplasm of the dividing
cell. Ultra-violet ravs and MgCL render the permeability of the cytoplasm
uniform throughout the cell, and after exposure to these agencies cleavage divi-
sions (of the Lamellibranch Molluscs), which would normally be unequal, take
place equally (Pasteels, 193 1).
CLEAVAGE AND DIFFERENTIATION
109
separating four micromeres (lato id) from four macromeres {lA
to iD). At the next three cleavages, the micromeres divide sub-
equally, but the macromeres bud off three further quartets of small
cells or micromeres (2^ to 2^, 3 « to 3 ^, 4 « to ^d). After a perfectly
definite and fixed number of cleavage divisions, which difiters for
Fig- 51
The polar lobe in Dentalium. a. Fertilised egg with animal and vegetative clear
zones (pole plasms), b, Protrusion of first polar lobe. c,It passes to one of the
first two blastomeres. d, 2-cell stage, retraction of polar lobe, e, 2-cell stage, pro-
trusion of second polar lobe. /, End of second cleavage, second polar lobe passes
to the D blastomere. (After Wilson, from Morgan, Experimental Embryology ,
Columbia University Press, 1927.)
diflferent blastomeres, a larva with a fixed number of cells is pro-
duced.
In Dentalium, at the approach of the first cleavage, a portion of
the vegetative region is partially constricted off from the rest of the
egg as the so-called polar lobe. This passes to one of the first two
blastomeres and is then withdrawn into it. The blastomere with
the polar lobe is destined to be posterior, and is called CD in con-
tradistinction to the AB or anterior blastomere (figs. 51, 52).
no
CLEAVAGE AND DIFFERENTIATION
The AB blastomere, if isolated, produces a larva which lacks the
apical organ and the region of the body behind the main ring of
cilia (the post-trochal region), including the coelo-mesoderm. But
the apical organ and post-trochal region are present in larvae de-
veloped from isolated CD blastomeres : these structures are, how-
Fig. 52
A, B, 4-cell stage, later cleavage in Dentaliuni, vegetative views. A, The polar lobe
has been retracted. B, It is protruded again but of smaller size than earlier
(fig. 5 1) in preparation for the third cleavage. C, 8-cell stage : polar lobe retracted
into iD. D, fourth cleavage: iD divides into 2D and a 2^ cell (first somatoblast)
containing polar lobe material. E, i6-cell stage. F, 32-cell stage, showing
formation of third quartet of micromeres. (After Wilson, from Morgan, Ex-
perimental Effibryology, Columbia University Press, 1927.)
ever, of normal full size, and therefore disproportionately large for
the half-sized larva ^ (fig. 53).
At the approach of the second cleavage, the polar lobe is pro-
truded again from blastomere CD, and becomes incorporated into D.
If blastomeres A, B,or C are isolated, they resemble AB in that the
larvae into which they develop lack the apical organ and the post-
trochal region. These structures are present, but relatively much too
large, in the miniature larvae developed from isolated D blastomeres.
^ E. B. Wilson, 1904 a.
CLEAVAGE AND DIFFERENTIATION
III
The third cleavage separates the first quartet of micromeres
(la, lb, ic, id) from the four macromeres {lA, iB, i C, iD).
Fig. 53
Dentaliiim, development of isolated blastomeres. a. Larva resulting from an
isolated CD blastomere (therefore containing the first polar lobe) ; this larva is of
reduced size but normal in form except that the apical organ and post-trochal
region are proportionately too large, b, Twin larva to a, resulting from an isolated
AB blastomere from the same egg; this larva lacks the apical organ and post-
trochal region, c. Larva resulting from an isolated D blastomere (therefore con-
taining the second polar lobe) ; this larva is of reduced size but normal in form
except that the apical organ and post-trochal region are proportionately much too
large, d, Twin larva to c, resulting from an isolated C blastomere from the same
egg; this larva lacks the apical organ and post-trochal region, e, Larva resulting
from an isolated A or B blastomere ; this larva lacks the apical organ and post-
trochal region. /, Larva resulting from an isolated id blastomere; this larva
possesses the apical organ but lacks the post-trochal region, g, Twin larva to/,
resulting from an isolated i c blastomere from the same egg ; this larva lacks the
apical organ and post-trochal region. (From Jenkinson, Experimental Embryology,
Oxford, 1909, after Wilson.)
Isolated la, ib, or ic blastomeres give larvae which possess a ring
of cilia, but lack gut, apical organ, and post-trochal region. An
isolated i d blastomere gives a similar larva, except that it possesses
CLEAVAGE AND DIFFERENTIATION
an apical organ. It is clear, therefore, that in the egg of Dentaliiim
there is a particular portion of the cytoplasm which is precociously
chemo-difTerentiated, and essential for the formation of the apical
organ and post-trochal region. This portion is contained in the first
polar lobe, the contents of which are distributed in a definite and
unequal way between the various blastomeres (fig. 53).
These conclusions are confirmed by experiments (also on Denta-
liiim) of a rather diflferent kind, in which the polar lobe is simply
cut oiT, without separating the blastomeres. If the polar lobe is cut
off at the onset of the first cleavage, the larva (like that from isolated
a b
Fig. 54
Organ-forming substances in Dentalium. a, Normal larva, 24 hours old. b, Larva
lacking apical organ and post-trochal region , obtained after removal of the first polar
lobe. (From Jenkinson, Experimental Embryology, Oxford, 1909, after Wilson.)
AB, A,B,ov C blastomeres) lacks the apical organ and post-trochal
region. If the polar lobe is cut off at the onset of the second cleavage,
the larva possesses the apical organ and lacks the post-trochal region.
At the approach of the second cleavage, therefore, the organ-form-
ing materials for the post-trochal region become separated from
those for the apical organ, for the latter are not included in the polar
lobe at its second extrusion ; instead, they presumably migrate into
the animal end of the D blastomere where they are in a position to
become included in i ^ at the next cleavage (fig. 54).
Similar occurrences whereby chemo-differentiated substances
present in the uncleaved egg are restricted by specialisations of the
CLEAVAGE AND DIFFERENTIATION II3
cleavage mechanism to particular blastomeres, and therefore later
distributed to particular and circumscribed regions of the embryo,
are found in other Molluscs and Annelids. A polar lobe very
similar to that of Dentalium is found in the Gastropod Ilyanassa
and the Polychaetes Chactopterus and Myzostoma. In Ilyanassa, if
the polar lobe is removed, no mesoderm is produced and the larva
is abnormal in form. Isolated blastomeres give rise to incomplete
embryos which die before reaching the larval stage.^ No experi-
ments appear to have been performed on Chaetopterus and
Myzostoma, but there is every reason to think that the polar lobe
in them plays a similar part.
In the Oligochaete Tubifex, the Q^g possesses so-called pole-
plasms — clear areas of cytoplasm at the animal and vegetative poles.
By means of extremely unequal cleavage, these are entirely re-
stricted, first to blastomere CD, and then to blastomere D. The
two pole-plasms then unite near the centre of the cell. At the next
cleavage the united pole-plasms remain entirely within the macro-
mere (i/)), but a portion of them passes to the first (ectodermal)
somatoblast, 2 d. The remainder, which is left in 2-D, passes at the
fifth cleavage entirely into 3Z), and then at the sixth into the second
(mesodermal) somatoblast, 4^. The course of events in the leech
Clepsine appears to be the same.
In Tubifex, none of the blastomeres AB, A, B, or Cis capable by
itself (the unwanted blastomeres being killed by uhra-violet light)
of developing into anything approaching a complete embryo;
blastomere D, however, can develop into a complete and properly
proportioned embryo.-
Other experiments on Tubifex have confirmed and extended
these results. By certain methods (application of heat, or depriva-
tion of oxygen) the first cleavage can be made to take place equally
instead of unequally, and in this case both blastomeres of the 2-cell
stage possess equal amounts of the pole-plasms. From such eggs,
double monsters (of the cruciata type, see p. 156) are produced.
It would appear that when the time comes for the formation of
micromeres, each set of pole-plasms will give rise to a set of
somatoblasts (one ectodermal and one mesodermal), and these will
differentiate independently into the main organs of the trunk.^
1 Crampton, 1896. 2 Pgnners, 1925. ^ Pgnners, 1924.
HEE C
114 CLEAVAGE AND DIFFERENTIATION
Similarly, in Chaetopteriis, it has been found possible to make
the first cleavage take place equally instead of unequally, by tem-
porary compression exerted after extrusion of the second polar
body and released when the first cleavage plane has cut half-way
through the egg. In this case, both the blastomeres of the 2-cell
stage receive a half of the polar lobe, and the result is the formation
of double monsters (likewise of the cruciata type). If, however, the
1/2 blastomeres of such eggs are isolated, each can give rise to a
single whole embryo. ^ Double monsters have also been found in
the leech Clepsine^^ where they are probably due to equality of
cleavage divisions. It is interesting to note in this case that the
spiral cleavage of the right-hand member is reversed.
It further appears from experiments on Clepsine, in which the
method of damaging small areas of the unsegmented egg was
employed, that the animal pole-plasm is necessary for cleavage to
occur at all. When only the vegetative pole-plasm has been
destroyed, cleavage is more or less normal except that it is delayed
in the D quadrant. Both the somatoblasts (2^ and 4^) are formed,
though /[d contains more yolk than normal. However, although 4^
produces the rudiments of mesodermal germ-bands, these are in-
capable of differentiation, and the embryo dies. It is not known
whether the ectodermal germ-bands (derived from zd) could
differentiate, as this only occurs at a later stage.
The cleavage-pattern, even in the absence of vegetative pole-
plasm, is thus predetermined down to the formation of the
rudiments of mesodermal germ-bands of typical appearance. But
the chemo-differentiation of these to definitive mesoderm is
dependent on the presence within them of an organ-forming sub-
stance derived from the vegetative pole-plasm.^
We may here draw attention to the work on the limpet Patella,'^
which demonstrates the extraordinary restriction of potency shown
by the micromeres in forms with spiral cleavage. In Patella, no
polar lobe is formed, but, as in Tiihifex, the potentiality for pro-
ducing mesoderm is restricted to the D quadrant. An isolated
micromere of the first quartet (i.e. a 1/8 animal blastomere, la
to id) produces a purely ectoblastic structure with cells typical of
1 Titlebaum, 1928. '" Muller, 1932.
3 Leopoldseder, 193 1. * E. B. Wilson, 1904B.
CLEAVAGE AND DIFFERENTIATION
II
Fig. 55
Development of isolated blastomeres of Patella, a, An isolated i/8 blastomere
(micromere of the first quartet: i). b, c, Its division into two cells: i^ and i-.
i" and i^-, 1 2^ and i--; one of these (i^^)
d. Their division into four cells
is an apical rosette-cell, another (i^-) is a " Molluscan cross" cell, and the
remaining tw^o (i^^ and i^^) are trocho blasts (stippled), e, Resulting larva
24 hours old, containing an apical organ (formed from derivatives of i^^), inter-
mediate cells (derivatives of i^-), and four cilia-bearing trochoblasts, the results
of division of i^^ and i". These cells are exactly those to which 1/8 micromere
would give rise in normal development in a whole embryo. /, An isolated 1/16
blastomere (primary trochoblast, i^). g, h, Its division into two cells (i-^ and i--).
/, Their division into four cells (i-^^ and i-^- ; i^-^ and i"^).j, The same, 24 hours
later ; each of the four cells (trochoblasts) has put out cilia, but divides no further.
These cells are exactly those to which a 1/16 blastomere (primary trochoblast)
would give rise in normal development in a whole embryo, k, A pair of trocho-
blast cells, the only product of an isolated 1/32 blastomere (trochoblast: i-^ or
T--). /, ?n, Isolated 1/64 blastomere (trochoblast: i-^^ or i-^-, i--\ i-"); such a cell
puts out cilia but does not divide. Ji, A pair of " secondary trochoblast cells," the
product of an isolated 1/32 blastomere (i^-, " Molluscan cross " cell). o,p, Isolated
1/64 blastomere ("secondary trochoblast," i^'"); such a cell puts out cilia but
divides no further. (From Jenkinson, Experimental Embryology , Oxford 1909,
after Wilson.)
8-2
Il6 CLEAVAGE AND DIFFERENTIATION
the apical sense-organ at one end, and powerfully ciliated cells
characteristic of the prototroch at the other : these are separated by
non-ciliated epidermal cells. The types of cell and the number of
each type produced by the isolated i/8 micromere are the same
as it would have produced in the young swimming trochophore
larva if it had been left in place in the developing egg (fig. 55).
Descendants of a 1/8 micromere, if isolated later, continue to
divide just as often and to produce just the same kind and number
of cells as would have happened in the whole intact embryo. For
instance, the vegetative member of the first product of division of
a micromere of the first quartet (i «2 to i d.^ in normal development
produces four ciliated prototroch cells. It does the same if isolated ;
while either of the products of its division divides once only to
produce two ciliated cells.
Isolated cells of the second quartet {2 a to zd) produce certain
ciliated cells which contribute to the prototroch, certain others of
a different type which belong to the pre-anal ciliated band, non-
ciliated epidermal cells, and larval mesenchyme in the interior.
These types of cells in the same numbers are produced by the
micromeres of the second quartet in normal development.
The exact meaning of these facts has not been determined. Pre-
sumably, two agencies are at work. First, certain chemo-differ-
entiated substances are probably restricted to particular micro-
meres ; secondly, it appears that the number of cleavage divisions
possible to any isolated blastomere is fixed. This may perhaps be
correlated with the fact (described below, see p. 132) that nuclear
synthesis during cleavage takes place at the expense of certain
materials in the cytoplasm, present in finite amount. When
these materials are exhausted in an isolated cell (deprived of
contact with the yolk of its Qgg^ or other food-supply), cleavage
stops.
It is probable that in all Annelids and Molluscs (other than
Cephalopods), even when no differentiated substances can be de-
tected in the uncleaved egg, they do in fact exist, and are distributed
during cleavage in a similar way. Only on these lines can such facts
be explained as the almost universal restriction of the potentiality
of forming mesoderm bands to 4^, and of forming ectoderm bands
to zd. The very general fact that the Z) 1/4 blastomere is larger than
CLEAVAGE AND DIFFERENTIATION II7
its three sisters is doubtless to be explained by the presence in it
alone of such specific organ-forming stuffs.
Summing up the evidence, we may say that the animal pole-
plasm, and its presumable homologue, the slightly thickened cap of
cytoplasm at the animal pole of the egg of De?italmm, are in some
way responsible for normal cleavage, though this has only been
demonstrated for Clepsine (see above). On the other hand, organ-
forming substances for apical organ (where present), and mesodermal
and ectodermal germ-bands, appear to be located in the vegetative
pole-plasm or polar lobe. This has been demonstrated for the apical
organ in Dentalium, for the mesodermal germ-bands in all forms
investigated, and for the ectodermal germ-bands in Dentaliiim. The
migration in the direction of the animal pole of the material for the
apical organ occurs before the second cleavage {Dentaliiim), for the
ectodermal and mesodermal germ-bands only later, in some cases
(Annelida) after a union of the two pole-plasms within the D
macromere, to be segregated at the fourth and sixth cleavages
respectively.
In general we may say that determinate spiral cleavage provides
an effective method of distributing precociously differentiated
substances to particular regions of the embryo, and that special
advantage of this has been taken by the Annelids and Molluscs,
though in varying degrees by different forms.
§5
The conditions found in Bero'e and Dentalium introduce us to
another principle of considerable importance. In Beroe, the forma-
tion by chemo-differentiation of the green ectoderm-producing
substance and the uncoloured endoderm-producing material is
effected prior to fertilisation, but the localisation of these substances
in their definitive positions is only brought about during cleavage.
In regard to the tgg, these substances are preformed, but not
prelocalised.
The same is true, though the details are even more elaborate,
concerning the distribution of the materials contained in the pole-
plasms and polar lobes of Mollusca and Annelida.
When the distinction between mosaic- and regulation-eggs was
regarded as fundamental, this distinction between the preformation
i8
CLEAVAGE AND DIFFERENTIATION
and the prelocalisation of organ-forming substances appeared to
be of considerable theoretical importance. From the point of view
here adopted, the absence of prelocalisation in such cases is seen
to be a frequent (though not universal) consequence of precocious
>»^'-'^P
C
Oy-i-
Fig. 56
Localisation of organ-forming substances in Ascidians. Views of eggs of Styela.
Yellow cytoplasm containing mitochondria {yp) small circles. Yolk {Gy)
stippled. Clear cytoplasm {cp) white. A, Before fertilisation, showing germinal
vesicle {GV), yellow cytoplasm evenly distributed over the surface. B, Imme-
diately after fertilisation, showing clear region {kp) derived from germinal
vesicle at animal pole, yellow cytoplasm streaming down to vegetative pole.
C, The yellow cytoplasm forms a cap at the vegetative pole {yz) containing the
male pronucleus. The clear cytoplasm forms a layer just above it. D, Left side
view of egg just before first cleavage, showing yellow crescent {yc) and clear
crescent {cp) posterior, and grey crescent (Gc) anterior. (From Conklin,
Chap. IX of Cowdry, General Cytology, Chicago, 1924.)
chemo-differentiation. When chemo-differentiation occurs prior to
fertilisation, the differentiated substances thus produced are able to
shift their relative positions, either in the uncleaved tgg, or as a
result of manoeuvres effected during cleavage. If, on the other hand,
CLEAVAGE AND DIFFERENTIATION IIQ
it does not occur until after the end of the cleavage period (as in
Amphibia) , the substances are precluded from this type of movement
through their being confined within cell-membranes, and redistri-
bution can only be effected by the movements of cell-regions.
The examples so far given concern the migration of organ-form-
ing substances during cleavage. Other forms show striking localisa-
tion phenomena in the uncleaved egg, usually initiated by polar
body formation, as the result of fertilisation. The classical ex-
ample of this is the Ascidian Styela {Cynthia). Before fertilisation,
the egg contains a cap of clear cytoplasm at its animal pole, a central
mass of yolk, and a superficial layer of yellow cytoplasm laden with
mitochondria. The clear cytoplasm is chiefly derived from the
breaking down of the large germinal vesicle. Almost immediately
after the entry of the sperm, the polar bodies are given off, and the
clear cytoplasm and the yellow cytoplasm flow down to the vege-
tative pole, leaving the animal pole occupied by the yolk, except for
a very small cap of clear cytoplasm. Next, the sperm moves to-
wards the centre of the &gg, along an apparently predetermined
path (indicating that a plane of bilateral symmetry already exists
in the tgg), and another rearrangement of the cytoplasmic regions
ensues. The sperm appears to drag much of the yellow cytoplasm
with it into the interior of the egg, and this yellow cytoplasm now
forms a crescent on the surface, beneath the equator, with its horns
extending a quarter of the way round the egg on each side. The
clear cytoplasm forms a crescent immediately above the yellow
cytoplasm, and the centre of these crescents marks the future
ventro-posterior side of the embryo. After the first cleavage,
another crescent, light grey in colour, is formed opposite the yellow
crescent (fig. 56). A pattern almost precisely similar is found in
Amphioxus.^ Thus, in these forms, both radial and bilaterally
symmetrical localisation are effected prior to cleavage.
Other examples of such rearrangements are afforded by other
Ascidians (e.g. Ciona) and by Myzostoma, in which a green vegeta-
tive area is formed in the oocyte, while fertilisation results in the
withdrawal of a red substance to the animal pole, leaving a clear
equatorial zone.- In the leech Clepsine, the pole-plasms, or areas
of clear cytoplasm at the two poles, only form after the polar bodies
^ Conklin, 1933. - Driesch, 1897.
I20 CLEAVAGE AND DIFFERENTIATION
have been extruded. Here, and in various other Annehds, the
material for the pole-plasms and polar lobes appears to have been
previously distributed over the whole surface of the tgg, and to
some extent in the interior. This follows from the fact that all and
sundry unfertilised egg-fragments of Chaetoptenis subsequently
fertilised are capable of developing into normally formed miniature
larvae,^ whereas the polar lobe in later stages is sharply localised.
Further evidence of a rearrangement of materials as a result of
fertilisation is provided by other experiments in which the develop-
ment of egg-fragments is studied. In the Nemertine Cerehratiilus,
for instance, such experiments show that there is a progressive in-
crease, from before fertihsation, to the onset of the first cleavage,
in the restriction of the potencies of animal and vegetative regions.
The animal region becomes progressively less able to produce
digestive tract and larval lappets, while the proportion of vegetative
fragments which produce an apical organ becomes smaller, during
the period in question.'^
A remarkable fact, whose precise interpretation is not clear, is
found in De7italmm. If the unfertilised tgg is cut across, latitu-
dinally or obliquely, and then fertilised, the vegetative portion seg-
ments as a whole, with a polar lobe usually of correct proportional
size.^ The resulting larva also has a correctly proportioned apical
organ and trunk. It will be remembered that when a CD blasto-
mere is isolated, it forms a polar lobe as large as in the whole tgg,
and the larva is disproportionate. A further remarkable fact is that
when an already fertilised egg is cut so as to produce an enucleate
vegetative fragment, though this does not cleave, it will protrude
its polar lobe synchronously with the first division of the nucleated
portion : the polar lobe is of the same size as in an intact Qgg. Some
irreversible change concerning the quantity of material in the polar
lobe must take place at fertilisation.
Further light on the mechanism of formation of polar lobes is
thrown by experiments on t\\^^lo\\u?,c Ilyanassa. Here, a polar lobe
is protruded four times : when the first and second polar bodies are
forming, and at the first and second cleavages. At its first ap-
pearance, the degree of protrusion is extremely slight ; at its second,
moderate ; while at its last two appearances, it is very marked, and
^ E. B. Wilson, 1Q29. ^ Yatsu, 1910. ^ E. B. Wilson, 1904 a.
CLEAVAGE AND DIFFERENTIATION
121
the polar lobe is at one moment only connected with the rest of the
egg by a narrow stalk (figs. 57, 58).
/
Fig. 57
Polar lobe formation in the normal cleavage of the Mollusc Ilyanassa. The small
circles represent yolk-spheres, a, Protrusion of the lobe by the uncleaved egg.
b, c, Its passage to the CD blastomere. d-f, Second cleavage; the polar lobe
passes to the D blastomere. g, Formation of first quartet of micromeres by
unequal dexiotropic division, h, i, Formation of second quartet of micromeres
by an unequal laeotropic division. (From- Morgan, Experimental Embryology,
Columbia University Press, 1927, fig. 135, p. 360.)
Normally, the lobe is composed of very yolky material. Centrifuge
experiments, however, show that its protrusion occurs irrespective
122
CLEAVAGE AND DIFFERENTIATION
of the materials which it contains, for it may contain only oil and
cytoplasm, or these on one side and yolk on the other. It was further
found that on centrifuging the egg at the time of the first cleavage,
the nuclear spindle may be disarranged so as to divide the egg
equatorially with reference to the original axis. In such cases the lobe
appears with reference to the original axis, i.e. at the end opposite
the polar bodies, and not with reference to the cleavage plane.^
Fig. 58
Effects of centrifuging the uncleaved eggs of the Mollusc Ilyatiassa. A (modified)
polar lobe is formed at the first cleavage ; it forms at the region opposite to the
polar bodies — i.e. in its normal relation to the original egg-axis — irrespective of
the direction of the first cleavage furrow or of the materials it contains. (From
Morgan, youni. Exp. Zool. lxiv, 1933.)
It appears that during mitosis, most of the egg undergoes some
degree of gelation, but that the region of the polar lobe is not
involved in this, and that the superficial layer of the lobe-region
is predetermined during the oocyte stage to behave as it does
during maturation and early cleavage. It is interesting to note that
if the polar lobe is detached before the first cleavage, it undergoes
spontaneous changes of form, apparently synchronised with the
cleavages of the egg.^
• §6
For mosaic development to occur, some degree of precocious
chemo-difTerentiation must have been effected, prior to the onset
of cleavage. The organ-forming materials thus available may be
1 Morgan, 1933.
CLEAVAGE AND DIFFERENTIATION 123
differentially distributed by a specialised cleavage, or regulation
may be prevented by a high degree of viscosity in the egg, at any
given stage. Bero'e affords an example of the latter method, but the
best illustration is provided by the Ascidians.
It will be remembered that the Ascidian Qgg is in a liuid state
before fertilisation, but that after this event its viscosity is enor-
mously increased. Further, we have already mentioned the localisa-
tion of different organ-forming substances which takes place at
fertilisation. Elaborate experiments on the effects of killing blasto-
meres and of centrifugalisation have shown that the fertilised egg
is already a mosaic of chemo-differentiated regions. 1/2, 1/4, and
3/4 blastomeres all develop into those parts of a larva to which they
would have given rise in normal development. The partial embryos
round themselves off, and this process in some forms (e.g. Phallusia)
goes much further than in others (e.g. Styela, Ciona), so that the
products of single blastomeres may appear superficially to be whole
larvae, but sections invariably show that they are only parts, halves
or quarters (e.g. with mesoderm only on one side; see p. 97 and
figs. 45, 59). The egg of Amphioxus behaves in an extremely similar
manner.^ In the experiments on Amphioxus, the blastomeres of the
2- and 4-cell stages were frequently disarranged without being totally
separated. In such cases, they always preserved their inherent
polarity, though there was complete fusion between their products.
The result was the formation of double monsters in various orienta-
tions (fig. 34).
In Styela, the fertilised egg contains yolk in the animal hemi-
sphere, cytoplasm with yellow mitochondrial granules at the
vegetative pole, and clear cytoplasm in between. In the centrifuge
tube the eggs tend to orientate themselves in such a way that the
animal pole with the relatively heavy yolk is centrifugal, and the
stratification of the egg is then increased by the centrifugalisation.
But if the eggs are slightly compressed, either by mutual pressure,
or by being placed in fine tubes so that they cannot rotate, centri-
fugalisation can restratify the egg-contents in such a way that, for
instance, all the yellow granular cytoplasm is confined to one of
the blastomeres of the 2-cell stage. The resulting embryo then
possesses muscle-fibres only on one side of the body.^ The mosaic
^ Conklin, 1933. - Conklin, 1924, 193 1.
124 CLEAVAGE AND DIFFERENTIATION
has been forcibly disarranged by the centrifuge, taking advantage
of the different specific gravities of the various egg-contents, but
each part of the mosaic continues its predetermined course. In this
way, organ-forming substances have been shown to be present in
the fertiHsed egg, and respectively responsible for the formation
of ectoderm ("ectoplasm"), endoderm ("entoplasm"), neural
plate ("neuroplasm"), notochord ("chordoplasm"), muscle fibres
("myoplasm"), and mesenchyme ("chymoplasm")^ (%s. 56, 59).
It is thus clear that the fertilised egg of the Ascidian is already
a highly complex mosaic of chemo-differentiated stuffs, and we may
now turn to the experiments in which the developmental potencies
of fragments of the unfertilised egg have been tested.
Latitudinal halves of unfertilised eggs of Ascidiella, subsequently
fertilised, show that there is already at this stage a differential re-
partition of potencies along the egg-axis. The larvae obtained may
be deficient in one or more kinds of tissue according to the level of
the cut: myoplasm can be separated from chymoplasm, neuro-
plasm from chordoplasm, the former in each case being situated
nearer to the animal pole. The various substances must, therefore,
occupy different levels.^
In view of the rigid mosaic behaviour of isolated blastomeres, and
of the definite localisation of substances at fertilisation (as tested
by the centrifuge experiments), the further result may seem sur-
prising that meridional halves of unfertilised eggs, subsequently
fertilised, may give rise either to apparently normal and sym-
metrical larvae, or to lateral half-larvae. The former type appear to
provide a case of regulation, which would be remarkable in such a
form as an Ascidian. These results can, however, be explained on
the view that the various organ-forming substances in the unfer-
tilised egg occupy circular zones at particular levels surrounding
the egg (or possibly crescentic zones, the horns of which quite or
^ Careful analysis has shown that the visible prelocalised substances in Styela,
such as the mitochondria, which impart the yellow colour to the region destined
to give rise to muscles, are not themselves morphogenetic substances. Muscles
can develop without mitochondria. The various regions differ in the consistency
of their cytoplasm, and it is these sharply marked off differentiated regions which
appear to constitute the true organ-forming substances. The mitochondria and
other gross differences are symptoms, not causes (see Duesberg, 1928). This
question of the relation of organ-forming substances to raw materials will be
discussed in Chap. vii. 2 Dalcq, 1932.
CLEAVAGE AND DIFFERENTIATION
125
Fig. 59
D
Mosaic development and prelocalisation in the egg of Ascidians (Styela). A, B,
Normal development, dorsal surface views : A, Late gastrula. The neural plate
(n.p.) overlies the notochord rudiment (not seen) ; and the muscle-rudiments (tns.)
border the blastopore laterally. B, Neurula. The neural tube (n.t.) has formed,
and the mesenchyme (m'ch.) is visible. C, D, Sections of abnormal neurula
stages derived from eggs centrifuged before the first cleavage. The disarrange-
ment of organ-forming substances by centrifuging had led to the disarrangement
of organ-rudiments. Endoderm (end.) and notochord (ch.) appear on the outside,
ectoderm (ect.) and neural plate substance (-9ts.) on the inside. Eye spots (E.) and
muscle-rudiments are also ectopic. (From Conklin, Chap, ix of Cowdry, General
Cytology, Chicago, 1924.)
126 CLEAVAGE AND DIFFERENTIATION
almost meet). Any meridional half will thus contain a portion of
all the necessary substances. However, the cytoplasm of the un-
fertilised egg appears to be already endowed with a plane of bilateral
symmetry, and if the cut through the egg is made at right angles to
this plane, the resulting half-egg will be able to form a complete
and symmetrical larva. But if the cut coincides with the plane of
bilateral symmetry, the half-egg will give rise to an asymmetrical
half-larva.^
At this stage, therefore, regulation is possible in some cases, owing
to the fact that the organ-forming substances are localised in such
a way that egg-fragments may contain portions of all of them. Sub-
sequently, however, at fertilisation, the localisation becomes more
restricted, the circular bands or crescents become reduced to
smaller crescents, the horns of which do not extend more than a
quarter of the way round the egg on each side, and this, together
with the high viscosity, effectively prevents regulation.
It is worth stressing that in Dentalium, the CD blastomere pro-
duces a larva which, while showing disproportion in regard to the
organs derived from the polar lobe, appears to have undergone
regulation round the major axis, thus becoming bilaterally sym-
metrical. Similarly, in Amphioxus a lateral 1/2 blastomere produces
a bilaterally symmetrical larva. Both in Bero'e and the Ascidians,
however, 1/2 larvae preserve the laterality of the blastomere from
which they arose. Here again, it must presumably be the high
viscosity of these eggs which has prevented the rearrangement round
the main polar axis of materials needed for regulation.
§7
Returning to the question of the relation of cleavage to differentia-
tion, it may then be said that the part which cleavage plays is only
indirect. Cleavage is a process whereby the single-celled fertilised
egg is split up into a number of separate cells whose differing
qualities depend upon factors which are originally independent of
cleavage, and concern the viscosity of the Qgg and the time of
chemo-differentiation of its cytoplasm.
In this connexion, we may refer to the very interesting case of
the insect egg. Here, cleavage of the nucleus begins and continues
^ Dalcq, 1932.
CLEAVAGE AND DIFFERENTIATION I27
for a long time in the interior of the egg, while the peripheral
cytoplasm or blastema remains undivided. It is only later that the
nuclei, now very numerous, migrate to the surface of the egg, and
the cytoplasm becomes partitioned off into blastomeres, forming the
blastoderm (see also p. 88).
Experiments on the regulatory capacity of the insect egg have
given different results in the various groups. In the house-fly
Musca domestica, the nuclei have already begun to divide when the
Qgg is laid, but the cytoplasm is still quite undivided. Nevertheless,
all the parts of the cytoplasm are already determined and chemo-
differentiated ; damage done to any part of the cytoplasm results in
damage to or absence of some definite structure in the developed
organism, and no regulation is possible. Here, then, is a clear case
of precocious chemo-differentiation of the cytoplasm and mosaic
development in which cleavage plays no part at all.^
In the ant Campojiotus ligniperda, it has been possible to deter-
mine the time at which chemo-differentiation sets in. This is found
to coincide with the start of the visible differentiation of the blastema
into various regions, such as those of the future embryonic shield,
extra-embryonic blastoderm, etc., which takes place before the
nucleus has begun to divide at all. Prior to this time, the egg is
undetermined and capable of regulation : after this time the cyto-
plasm is chemo-differentiated, and development strictly mosaic.'-
In the dragon-fly Platycnemis pennipes, the time of onset of
chemo-differentiation is relatively later, during the blastoderm
stage, and the early egg is therefore capable of regulation. It has
been possible to obtain a normally proportioned diminutive insect
from one (posterior) half of an tgg constricted transversely into
two; duplications and triplications of structures after making
longitudinal slits in the blastoderm; and two insects from one egg,
the blastoderm of which was divided transversely.^ Later on, how-
ever, constrictions and injuries result in the development of
partial embryos only. In this case, as in that of Camponotus, it has
been possible to establish the very interesting fact that the process
of chemo-differentiation emanates as a stream from an activating
centre, situated near the hinder end of the egg (figs. 60, 84 and
122; see also pp. 170, 252).
1 Reith, 1925; Pauli, 1927. - Reith, 193 1. ^ Seidel, 1926, 1928, 1929.
28
CLEAVAGE AND DIFFERENTIATION
The relative unimportance from the point of view of differentia-
tion of the way in which the egg cleaves is revealed by the following
experiments.
When a blastomere of a sea-urchin is isolated at the 2- or 4-cell
stage, it develops, as already mentioned, into a whole larva, but the
6d)w.H
Fig. 60
Regulation in the insect egg. a. Normal embryo of the dragon-fly Platycnemis
pennipes, seen from the left side, b, Dwarf embryo, obtained by partial constric-
tion of the egg at the 4-nucleus stage ; the dwarf is normally proportioned and
developed and its organs have arisen from regions the presumptive fate of which
was quite different ; their fates were therefore not irreversibly determined at the
stage operated upon, and regulation has been possible. At. antenna; An. eye;
Epf. hindgut; M. mandible; M.Ch. chitinous muscle-attachments; Md. midgut;
M.Vi, first maxilla; Mx.^, second maxilla; O. labrum; Pr. proctodaeum; Sch.Ch.
apical chitin; St. stomodaeum; Stg. spiracle; Thi_^, first to third thoracic legs;
Schw.K. gills. (From Seidel, Biol. Zentralbl. xlix, 1929.)
cleavage which it undergoes is the same as that which it would have
undergone if it had been left in contact with its sister-blastomeres.
In normal development in these forms, the first and second cleavages
are meridional and equal : the third cleavage is latitudinal and equal ;
the fourth cleavage in the animal hemisphere is meridional and
equal, in the vegetative hemisphere it is latitudinal and unequal.
Each cleavage division is therefore recognisably distinct. Now, in
CLEAVAGE AND DIFFERENTIATION 129
a blastomere isolated at the 2-cell stage, the first cleavage which it
undergoes after isolation is meridional and equal (corresponding
to the second normal cleavage), and its next cleavage is latitudinal
and equal (corresponding to the third normal cleavage), and so
forth. The first cleavage of a blastomere isolated at the 4-cell stage
is latitudinal and equal (corresponding to the third normal cleavage).
In other words, the isolated blastomeres cleave as if they were still
parts of a whole, but they develop into whole larvae. Here, clearly,
the method of cleavage is without effect on the subsequent develop-
ment and differentiation.
The system of cleavage in the sea-urchin egg has been shown to
depend on a number of factors. First, there is the control which
the cytoplasm exerts on the orientation of the division spindles;
this is of such a kind that for a certain period of time (normally
corresponding to that between fertilisation and the attainment of
the 4-cell stage) any nuclear spindles that there may happen to be
are restricted to a latitudinal plane so that division will be meri-
dional; after this period, the spindles are rotated into the longitu-
dinal axis so that division will be latitudinal. From now onwards
there will be two sets of division spindles ; one in the animal and
one in the vegetative half of the egg. Those in the former set
revert to the latitudinal plane (meridional division of meso-
meres), while those in the latter remain longitudinal (latitudinal
division of macromeres from micromeres). Experiments of cutting
eggs at varying times after fertilisation have shown that the
fixation of a division spindle to a given axis is progressively
determined: a 1/2 egg cut meridionally within a quarter of an
hour of fertilisation can as it were start again with the deter-
mination of its spindle axis, and the 1/2 will cleave as a whole egg;
a similar 1/2 egg cut meridionally three-quarters of an hour after
fertilisation has its spindle axis set and fixed, and it cleaves as
a 1/2 blastomere.
Secondly, there is localised at or near the vegetative pole a special
region of cytoplasm which determines a marked inequality of
cleavage, leading to the formation of tiny micromeres split off from
the large macromeres. Thirdly, there is the fact that this special
region of the cytoplasm at the vegetative pole does not acquire its
property of causing unequal division until after a certain definite
HEE 9
130
CLEAVAGE AND DIFFERENTIATION
Fig. 6 1
Cleavage of the sea-urchin egg. Column A, normal cleavage as far as the
i6-cell stage (eight mesomeres, four macromeres, four micromeres), serving as
time-scale (read from top to bottom) for the other columns. By treatment with
hypotonic sea-water or shaking, the formation of the mitotic spindles can be
delayed : the other columns show the effects of increasing retardation of spindle-
formation. Column B, the first two cleavage spindles latitudinal, the third vertical
but so delayed that it falls within the period of micromere-formation : result,
four micromeres at the 8-cell stage. Column C, first cleavage spindle latitudinal,
the second fall within the period during which the spindles are rotated into the
vertical position ; they have not quite achieved it here and are oblique ; the third
cleavage spindles, at right angles to the second, are also oblique : result, two meso-
meres and two micromeres at the 8-cell stage. Column D, the first cleavage spindle
latitudinal, the second vertical, the third similar to the fourth of normal cleavage,
i.e. latitudinal in animal, vertical in vegetative cells: result, four mesomeres, two
macromeres, two micromeres at 8-cell stage, Blastomeres isolated at the 2-cell
stage cleave according to this pattern. Column E, cleavage of blastomeres isolated
at the 4-cell stage or of eggs cut into meridional halves (in which the mitotic
apparatus is so delayed that the first cleavage spindle coincides with the third
of normal cleavage and is vertical) ; the second (like the fourth normal) cleavage
spindles are latitudinal in animal, vertical in vegetative cells. (From Horstadius,
Acta Zool. IX, 1928, slightly modified.)
CLEAVAGE AND DIFFERENTIATION 131
period of time. In normal cleavage, this time corresponds to the
attainment of the 8-cell stage.
By various methods (use of dilute sea- water, shaking, and cutting
the egg into halves), it is possible to alter the time-relations of
mitosis relatively to these three factors. By delaying the rate of
cell-division, it is possible to make the second, or even the first
cleavage of an egg fall into the period when the nuclear spindles are
forced into the longitudinal axis. The result will be latitudinal
division at the 2-cell and i-cell stages respectively, whereas it
normally happens at the 4-cell stage. Very instructive are the cases
in which the cleavage division falls during the change of position of
the nuclear spindles, i.e. when the latter are oblique. One more
cleavage division in eggs whose mitoses are thus delayed w^U lead
to formation of micromeres precociously (fig. 61).
It will thus be seen that it is possible to make a whole egg cleave
as if it were an isolated blastomere of the 2-cell or 4-cell stage. When
a blastomere is isolated from a normal egg, the mitotic speed of
which has not been interfered with, the subsequent cleavage
divisions continue to be governed by the same factors as in the
normal tgg, with the result, therefore, that the blastomere cleaves
as a part.i
The second example of the relative unimportance of cleavage as
regards differentiation is provided by those cases in which a frog's
egg has been penetrated by several sperms. One sperm-nucleus
fuses with the egg-nucleus, but the other sperm-nuclei remain
isolated in the cytoplasm of the egg. When the egg begins to under-
go cleavage, not only does the zygote-nucleus divide and induce the
division of the cytoplasm into blastomeres, but each of the isolated
sperm-nuclei has a portion of cytoplasm allotted to it, and this be-
comes separated off as a little blastomere and subsequently divides.
Cleavage is therefore very irregular, and the embryo is composed
of an indiscriminate mixture of blastomeres, some containing the
products of division of the zygote-nucleus and representing the
normal blastomeres of typical cleavage, and some representing
blastomeres which would normally never have come into existence.
The two kinds of blastomeres can be recognised without difficulty,
for those derived from the zygote-nucleus are of course diploid,
^ Driesch, igoo; Horstadius, 1928.
9-2
132 CLEAVAGE AND DIFFERENTIATION
while the others are haploid. Since the volume of the cell is pro-
portional to the quantity of nuclear material which it contains, it is
easy to recognise the descendants of the two kinds of blastomeres in
the tissues to which they give rise. In spite of their abnormal
cleavage, such polyspermic frogs' eggs can sometimes develop
normally, the stage ultimately reached depending on the number
of supernumerary sperms present. A pentaspermic egg can produce
a free-swimming tadpole which lives for 10 days after hatching:^ a
dispermic egg can produce a tadpole which lives for three months.^
Lastly, it has been shown in the case of Chaetopterus and Nereis
that a certain amount of differentiation can take place even if cleav-
age is totally suppressed, by treatment of the egg with KCl.^ Cilia
are put out and internal rearrangements occur, the most interesting
of which is the assumption by certain granules of the position in
the Qgg which corresponds to that of the cells of the prototroch,
which cells in normal cleavage come to contain these granules.'*
§8
But besides splitting up the cytoplasm of the egg into smaller units,
cleavage has one very important effect, though its bearing on differ-
entiation is indirect, and this concerns the adjustment of the ratio
between amount of nuclear matter and amount of cytoplasm present
in the cell.
In the oocyte of the sea-urchin {Echinus niicrotuberciilatus) it has
been shown that the ratio between the volume of the cytoplasm
and that of the nucleus is 7 : i . Maturation results in a certain in-
crease in cytoplasmic volume and a reduction in nuclear volume,
so that the ratio of cytoplasm to nucleus in the ripe egg is 400 : i .
But the volume of the cytoplasm has been only about doubled, so
that the explanation of the high ratio in the ripe egg must be looked
for to a small extent in the extrusion of nuclear material in the polar
bodies, and to a large extent in the passage of nuclear material into
the cytoplasm. Now the total amount of nucleic acid in the egg
and in subsequent stages of cleavage up to the blastula is constant.^
^ Brachet, 1910. - Herlant, 191 1. ^ Lillie, 1902; Spek, 1930,
^ What is in some ways a complementary experiment has been carried out by
removing the zygote nucleus from uncleaved axolotl eggs by means of a micro-
pipette. In spite of the absence of nuclei, the cytoplasm makes an attempt to
carry out cleavage, though this is partial and irregular. Jollos and Peterfi, 1923.
^ Masing, 1910.
CLEAVAGE AND DIFFERENTIATION 133
But at the start of cleavage, most of this nucleic acid is in the
cytoplasm. At each cleavage division, the nuclei of the daughter-
blastomeres are slightly larger than half the nucleus of the blasto-
mere that gave rise to them. There is consequently a gradual return
of nuclear material from the cytoplasm into the nuclei of the blasto-
meres, and this is shown by the drop in the ratio of total volume of
cytoplasm to total volume of nuclei at successive stages of cleavage.
At the 4-cell stage the ratio is about i8 : i, at the 64-cell stage it is
12 : I, while in the blastula the ratio has returned to the original
value of 7 : i.^
These results are of considerable interest, and for two reasons.
In the first place, the return of the cytoplasmo-nuclear ratio to the
original value occurs in the blastula, when cleavage has ended, and
when the hereditary effects of the nuclear material can begin to
manifest themselves, as will be shown in Chap. xii. It is not im-
probable that these two sets of events are causally related. In the
second place, the recognition of the existence in the cytoplasm of
the ripe egg of a finite amount of nuclear material accounts for the
termination of cleavage. It is well known that eggs which are made
to develop in the haploid condition (as by artificial partheno-
genesis) go on cleaving until their cells are half the volume of
normal diploid cells. ^ The haploid nuclei of the blastomeres re-
quiring only half the amount of nuclear material from the cyto-
plasm, the supply in the cytoplasm will last longer than is the case
with diploid nuclei ; cleavage will therefore go on for a longer time,
and the cells will be smaller. Conversely, it is known that if half
an egg, containing a nucleus, is fertilised (that is to say, diploid
nuclei but only half the normal quantity of cytoplasm is present),
the resulting larva has cells of normal (diploid) volume but is itself
of half size. It follows that it has half the number of cells that the
normal has, and this is what would be expected since it had only
half the reserves of nuclear material in the cytoplasm. Lastly, it is
possible in some cases to obtain fertilised eggs with tetraploid
nuclei. The size of the embryos which these produce is normal, but
their cells are twice as large and half as numerous as normal. The
quantity of nuclear reserve materials in the cytoplasm has given
out sooner than during normal cleavage, with the result that the
division of the blastomeres has not proceeded so far.
^ Godlevvski, 1925. " Boveri, 1905.
Chapter VI
ORGANISERS: INDUCERS OF DIFFERENTIATION
§1
The remarkable organising properties of the dorsal Up of the blas-
topore of amphibian embryos were discovered in the following
manner. In the experiments with newts' eggs of grafting pieces
of the presumptive neural fold region into other positions, in order
to discover the time at which they became irrevocably determined
to develop by self- differentiation, it was observed that the deter-
mination of the posterior part of the presumptive neural fold region
(i.e. that portion which lies near the dorsal lip of the blastopore)
was effected sooner than that of the anterior part (i.e. farther away
from the dorsal lip). It looked as if some agency emanated from
the dorsal lip of the blastopore like a "flow of determination", and
either streamed or was carried forwards^ (see also p. 173).
This suspicion was confirmed when it was found that if the
animal hemisphere is cut off from an early gastrula of the newt,
rotated through any angle about the egg-axis, and then stuck on to
the vegetative hemisphere again, the neural folds arise in line with
the dorsal lip of the blastopore, which, of course, is situated in the
vegetative hemisphere. The neural folds therefore arise from tissue
which would normally not have formed them, and neural folds are
not formed from the presumptive neural fold material which has
been rotated away from the meridian of the dorsal lip of the blasto-
pore.2 Something of the nature of what Herbst (1901) called a
"formative stimulus" appears thus to be associated with the dorsal
lip of the blastopore.
As to the time when the dorsal lip region exerts its organising
action, there are two possibilities. The first is to imagine a trans-
mission of stimuli through the tissues from the region of the organ-
iser before gastrulation ; the second possibility is to attribute its
action to the transmission of stimuli from underneath the surface
1 Spemann, 1916. ^ Spemann, 1906B, 1918.
organisers: inducers of differentiation 135
layer after gastriilation, at which time the organiser has been in-
vaginated, and forms the primitive gut-roof, i.e. notochord and
axial mesoderm (future myotomes). In both cases, the hinder part
of the presumptive neural fold region will be affected before the
front part.
It appears that the organiser acts in both these ways. That it can
exert its inducing action from below, after gastrulation, is demon-
strated by the fact that when a graft is made from the dorsal lip of
the blastopore of one gastrula into the flank of another blastula or
gastrula, it brings about the formation of the essential structures
(so-called axial structures) of an embryo. This embryo is called the
secondary embryo in order to distinguish it from the primary
embryo formed from the tissues of the host in the ordinary way. ^
The secondary embryo arises from tissue which had very differ-
ent prospective fates. The grafted organiser invaginates beneath the
surface of the tissues of the host and itself gives rise to part or all of
the notochord and axial mesoderm of the secondary embryo. The
other structures of the secondary embryo are usually formed of
host tissue only, but may contain an admixture of graft tissue also.
These can be easily distinguished by performing the experiment
with material derived from two species of Triton, T. cristatus and
T. taeniatiis, which differ in the pigmentation of their tissues.
There is therefore no doubt that the organiser can bring about
the determination of tissues by the transmission of stimuli from
underneath after gastrulation. This is further proved by the fact
that pieces of the primitive gut-roof (notochord and mesoderm,
which of course are derived from invaginated organiser material)
are capable of inducing the formation of axial structures.- \s we
shall see later, the main activity of the organiser in normal develop-
ment is to induce the formation of the neural plate and tube. This
^ In all cases, portions of central nervous system, notochord, and axial meso-
derm (somites) are formed ; in addition, brain and spinal cord, eyes, ears, kidneys,
peripheral mesoderm (lateral plate), gut-roof and heart may be produced. Such
embryos have not been kept beyond the tail-bud stage. Whether certain organs
of the secondary embryo are formed or not<iepends on several factors: (i) the
level of the host's main axis at which the graft is made ; (2) the region of the
organiser which is used as a graft; (3) the distance of the primary from the
secondary embryo, resulting in a greater or lesser degree of mechanical inter-
ference. (Spemann and Mangold, H., 1924.)
^ Marx, 1925; Bautzmann, 1926.
136 organisers: inducers of differentiation
it does by contact. All of the gastrular ectoderm underlain by noto-
chord and axial mesoderm will become neural plate (see p. 155).
But these facts, however, do not preclude the possibility of the
organiser exerting some effect in earlier stages also. And, as a
C
Fig. 62
Labile determination of neural folds in Urodeles. A, Early gastrula of Pleurodeles,
from which, B, the entire dorsal lip region is extirpated. C, Resulting embryo
showing spina bifida and neural folds prevented from reaching mid-dorsal line.
(From Goerttler, Zeitschr.f. Anat. 11. Entzuick. lxxx, 1926.)
matter of fact, other experiments have shown that during the period
before the irrevocable deterniination of the presumptive neural
fold material, it is nevertheless not wholly indifferent, and possesses
a labile determination ^ to develop into neural folds. This can be
tested in situ in an embryo by removing small portions of the
^ "Bahnung", Vogt, 1928 a; "competence", Waddington, 1932.
organisers: inducers of differentiation 137
organiser before gastrulation,i by preventing the organiser from in-
vaginating, which can be effected either by removing it entirely
by killing part of it and so preventing invagination on one side;'-
or by reducing its activities by means of exposure of the organiser
region to cold or deprivation of oxygen.^ In spite of the absence of
an organiser or of any invagination, distinctive but somewhat
.^C^\\
-J
Fig. 63
a. Embryo of Pleiirodeles in which gastrulation has been prevented by reducing
oxygen-access to the region of the dorsal Hp; neural folds are nevertheless
formed, b, Transverse section through the same embryo, showing neural tube,
but absence of notochord ; the lining of mesoderm and endoderm has been de-
rived from the floor of the blastocoel, which, here, is the large central cavity.
(From Vogt, Verh. deiitsch. Zool. Ges. xxxii, 1928.)
imperfect neural folds and tubes are developed. It is of interest
to note that in the absence of an underlying organism, the brain
achieves a more perfect differentiation than the spinal cord.^
In experiments of a different nature, in w^hich developing Uro-
dele eggs are subjected to a lateral temperature-gradient (seep. 342),
it is found that on the warmed side, structures appear in the ecto-
derm resembling neural material in cell structure, but may differ
considerably from neural folds in form.* These structures arise in
^ Lehmann, 1926, 1928 a.
3 Vogt, 1928A.
2 Goerttler, 1925, 1926.
* Gilchrist, 1929.
138
organisers: inducers of differentiation
positions where they are not underlain by mesoderm. If, on the
other hand, they are situated in regions where mesoderm does
underHe them, they become typical neural folds (see fig. 64).
Fig. 64
Effects of a lateral temperature differential on development in Urodeles. A-F,
Triturus torosus, treatment by temperature-gradient in blastula stage, with 5° C.
temperature difference between the two sides of each egg. A, B, Warmed on left ;
A, dorsal view ; B, anterior view ; the warmed neural fold is much larger and more
differentiated. C-E, Dorsal views. F, Lateral view, showing secondary neural
structures on the previously warmed side, either connected with the main neural
folds, or F, isolated from them. In C, D, E, the secondary formations appear to
be underlain by mesoderm, and have differentiated into structures of neural fold
type. In F they are not underlain, and do not show typical morphogenesis.
(From Gilchrist, Quart. Rev. Biol, iv, 1929.)
Another method of testing this labile determination is by inter-
plantation, i.e. the grafting of portions of blastulae (i.e. portions of
tissue taken from an embryo before the invagination of the
organiser) into the eye-sockets or coelomic cavities of other larvae.^
The differentiation of various structures can be obtained in this
way (fig. 148, p. 316). The fact that interests us here is that neural
tube may be differentiated in these circumstances from tissue which
^ Diirken, 1926.
organisers: inducers of differentiation 139
has never been acted upon by an invaginated organiser. Or we may
adopt the method known as explantation, in which the pieces of
blastulae, after being enclosed in epidermal jackets, are grown in
vitro in suitable media. DiiTerentiation of neural tube and of noto-
chord can be obtained in this way also, from tissue which has never
been acted upon by an invaginated organiser^ (see fig. 18, p. 49).
There is therefore some determinative agency at work in ad-
dition to the invaginated organiser. The labile determinations thus
induced are presumably due to the transmission of stimuli from
the organiser before gastrulation, in relation to the main axes of the
tggy in a manner which will be considered below in connexion with
gradient-fields^ (see p. 310).
In any case, it is clear that the labile determination of the
blastula stands in some relation to the bilateral symmetry imposed
upon the tgg at the moment of fertilisation.
The action of the organiser, then, must be considered as taking
place in two phases. First, working as part of the gradient-field,
the organiser may be figuratively said to sketch out the presumptive
regions in pencil, and then, after invagination, the organiser goes
over the same lines with indelible ink. At the same time, the organ-
iser is capable of roughing out the sketch straightway in ink, with-
out any previous pencil work, as in those experiments in which
the organiser is grafted into the flank of another embryo. Neural
folds can arise from the pencilling alone, and from the inking alone,
and this duplicity of methods whereby neural folds can be formed
is another example of the principle of " double assurance ".
But there is another point to notice here. When an organiser
is grafted into the flank of another embryo, the host-tissues are
^ Bautzmann, 19296,0; Holtfreter, 1929 A, b.
^ These examples have been mentioned in order to show that determination
and differentiation can take place in the absence of an invaginated organiser. But
several of these experiments introduce a new complication, since the tissue which
is differentiated in interplantation and in explantation frequently is of a nature
quite different from the presumptive fate of the region from which the piece was
taken. Presumptive neural tube material, for instance, has been found to differ-
entiate into notochord, muscle, mesenchyme^ and glandular epithelium (Kusche,
1929; Holtfreter, 1931A; Erdmann, 1931); presumptive epidermis can give rise
to neural plate, especially, for some unknown reason, when interplanted into
the coelomic cavity. Pieces of tissue from any part of the blastula have been
seen to differentiate into notochord and muscle (Bautzmann, 1929B) (see
P- 317)-
140 organisers: inducers of differentiation
induced to differentiate in particular ways under its influence, and
the labile determinations of these host tissues, whatever they may
have been, are obliterated^ and overridden. A cell-region which
possesses a labile determination to become epidermis may be made
to become neural folds. The organiser can, as previously mentioned
(p. 46), even override the presumptive distinction between the
germ-layers. For instance, a piece of presumptive ectoderm
(epidermis) implanted just below the dorsal lip will be carried into
the interior of the embryo, and there may give rise to a portion of
any of the following organs : vertebral centrum, myotome, lateral
plate, pronephros (mesodermal), notochord, or gut- wall (endo-
dermal).^ Presumptive neural folds can also form myotomes and
pronephros. Similarly, pieces of presumptive mesoderm grafted
into the region of presumptive ectoderm will (provided of course
that they are taken at the stage prior to chemo-differentiation) form
epidermis. The determination of epidermis, however, appears to
be less rigorous, and already differentiated epidermis can be made
to form conjunctiva (p. 178).
§2
It must be remembered that in the production of an end-result,
such as a differentiated structure, two sets of factors are involved :
first, the causal agent, in this case the organiser ; second the material
acted upon, the tissues. Examples of this resultant effect will be
given in the following paragraphs.
The action of the amphibian organiser is not species-specific,
i.e. it can induce the formation of axial structures when grafted into
^ Another example of the overriding of a previous labile determination is
provided by the Gephyrean v^^orm Bonellia. This form shows extreme sexual
dimorphism, the female being about the size of a plum with a proboscis a yard
long, while the male is only a few millimetres in length, and lives parasitically in
the uterus of the female. The larvae which hatch from the eggs all pass through
an indifferent stage. If such larvae do not come into contact with an adult female,
they themselves undergo development into females, by means of processes for
which the larva must presumably possess some sort of determination. But this
determination can be overridden if the larva comes into contact with an adult
female and settles on her proboscis. The proboscis secretes a substance which
induces in the larva the development of the male characters, involving reduction
of the anterior end of the body, and differentiation of the male reproductive
organs (Baltzer, 193 1).
^ Mangold, 1924.
r" horb
A • ^*» ir^I^TJ — ^ r
"•» ; **? , 4' •?-• -^' J
Fig. 65
Anuran organisers in Urodele hosts. A piece from the dorsal lip of the blastopore
of a gastrula of Bombinator pachypus grafted into a young gastrula of Triton
taeniatus induces the formation of a secondary embryo, a, b, Two stages of
development of an embryo thus obtained, c, Transverse section through b. Capital
letters refer to structures of primary embryo, small letters to secondary embryo.
Au, optic- vesicle; Ch, notochord; Horb, ear- vesicle; M, neural folds; Md, neural
tube; Urw, mesodermal somites; of primary embryo, ch, grafted notochord;
horb, ear- vesicle; m, neural folds; md, neural tube formed from graft tissue;
md', neural tube induced from host-tissue ; ms, undifferentiated mesoderm of graft
tissue ; urw', mesodermal somites induced from host-tissue. (From Geinitz, Arch.
Entiumech. cvi, 1925.)
142 organisers: inducers of differentiation
an embryo of a species different from its own. We can go further,
and say that its action seems singularly non-specific. Not only can
an organiser from Triton cristatus function in Triton taeniatus^ but
also organisers from Pleurodeles waltli, Amblystoma mexicanum,
and even the Anuran Bombinator pachypus^ can induce the forma-
tion of secondary embryos in Triton taeniatus.^ It is therefore
established that the inducing action of the organiser is not impeded
by a taxonomic difference of the order of value of a sub-class be-
tween its own tissue and that on which it works (fig. 65).
These experiments of heteroplastic and xenoplastic organiser
grafts between different species demonstrate the fact that the action
of the organiser is specific as to the general type of organs and
structures produced by induction, but non-specific as to the details
of these structures ; these latter are governed by local and intrinsic
properties and determinations of the tissues themselves, over which
the organiser has no control. For instance, a piece of presumptive
neural fold tissue of Triton taeniatus grafted on to the side of the
head of an embryo of Triton cristatus will differentiate into gills in
its new position. But, gills though they are, they preserve their
taeniatus character in being larger than the normal cristatus gills
on the other side of the embryo.^ Conversely (fig. 15), cristatus
tissue on Triton taeniatus gives rise to gills which are smaller than
the normal taeniatus gills.^ The retention of specific characters in
spite of induced determination to develop into structures other
than those which a piece of tissue would normally have produced,
is shown even more strikingly in those experiments in which a piece
of Anuran presumptive epidermis (from the ventral side of the
trunk) is grafted over the future mouth- region of a Urodele embryo.
In its new and strange position, the Anuran tissue differentiates into
mouth-parts, and it also gives rise to a ventral sucker of Anuran
pattern* which is functional and secretes an adhesive substance. It
also appears that horny teeth can be formed as well. No Urodele
normally possesses a sucker or horny teeth (fig. 66).
As a further illustration, we may take the results of experiments
in which a Urodele organiser (from Triton alpestris) is grafted into
an Anuran embryo {Bufo vulgaris). The induced secondary embryo
^ Geinitz, 1925 b; Schott^, 1930. ^ Spemann, 1921. ^ Rotmann, 193 1.
* Spemann, 1932, 1933; Spemann and Schotte, 1932.
organisers: inducers of differentiation
H3
possesses a ventral sucker, although the organiser which induced it
comes from a species which does not possess one.^ As the matter
has been figuratively put : the organiser disposes of the fates of the
h.
Fig. 66
The preservation of specific characters by a tissue, in spite of its having been in-
duced to undergo differentiation into structures other than those representing its
presumptive fate, a, A piece of ventral epidermis from a gastrula of the frog Rana
esculenta is grafted into the mouth- region of an embryo of the newt Triton
tae?iiatus, where it differentiates into mouth-parts in accordance with its position,
but, in addition, gives rise to ventral suckers (h., h.). b, Section through such an
embryo, showing: b. basal membrane of grafted epidermis; sc. typical secreting
cells of ventral sucker; 5. functional secretion. A sucker is never formed by a
newt. (From Spemann and Schotte, Natiiriuiss. xx, 1932.)
tissues in a general way, but as regards the details of their differen-
tiation, the tissues already possess thieir instructions.-
The age at which an organiser first acquires its power of in-
duction is not known, but constriction experiments on the egg of
^ Spemann, 1932, 1933; Spemann and Schotte, 1932. - Spemann, 1921.
144
Head-Organiser ^n- Head Level
Heeid-Organiser w Trunk Level
R. sec.
a.v. sec. eye I.
Trunk-Organiser in Head Level
Trunk-Organiser ztt Trunk Level
Fig. 67
Diagram showing the results of experiments testing the inductive capacities of
head-organiser (invaginated early: the anterior region of the primitive gut-roof)
and of trunk-organiser (invaginated late: the posterior region of the gut-roof),
and the reacting capacities of the host-tissues at head level and trunk level (see
also fig. 68). Head-organiser at head level forms only the head of a secondary
embryo with eyes and ear- vesicles ; head-organiser at trunk level may form a
complete secondary embryo, and the cephalic structures may arise at a level
considerably behind those of the primary embryo ; trunk-organiser at head level
may form a complete secondary embryo with cephalic structures at levels more
or less corresponding to those of the primary embryo ; trunk-organiser at trunk
level produces the trunk of a secondary embryo ; ear- vesicles are formed if the
secondary embryo reaches to the level of those of the primary embryo. Head-
organiser can thus form a head in both head and trunk levels, but trunk-organiser
can only form a head in head level ; the reaction to trunk-organiser of the host-
tissues at head level is to form a head, and at trunk level to form a trunk.
Pr.o.v. eye; Pr.a.v. ear, or primary embryo; sec.o.v. eye; sec.a.v. ear (left, L. or
right, i?.); sec.cycl.o.v. cyclopic eye, of secondary embryo. (Original, based on
Spemann.)
organisers: inducers of differentiation 145
the newt show that its site is already determined and locaUsed ten
minutes after fertiUsation.^ Rather later, portions of the blastula
in the region of the grey crescent have been found to possess
the inductive property.^ As to the time at which this property
is lost, it has been shown that the notochord, which of course
is formed from the invaginated organiser, retains for a consider-
able period the power of inducing the formation of neural
folds.3
It has also been shown that in the neurula stage, myotome material,
which of course w^as originally derived from the organiser region,
still retains the capacity of inducing neural tube formation from
presumptive epidermis when grafted into an early gastrula. How-
ever, slightly more lateral mesoderm material, which had dif-
ferentiated into pronephros, in similar experiments only induced
other pronephric tubules.*
This is known as '* homoiogenetic induction", to contrast it with
the heterogenetic power of the organiser, which induces the forma-
tion of structures different from itself. It is found that the neural
plate, once underlain by the organiser, possesses and retains for
a very long time — certainly up to the free-swimming larva — this
power of inducing the formation of structures of its own type. This
is proved by grafting portions of neural tube into blastulae, where
secondary neural folds are induced.^ It is of interest that the hind-
most portion of the neural fold region of the neurula induces the
formation of mesoderm, which agrees with the fact that this region
gives rise to the muscles of the tail in normal development (Chap. 11,
p. 28).® Accordingly, this induction also is homoiogenetic. Lens
rudiments implanted into blastulae have no power of induction,
either hetero- or homoio-genetic.^
Spatially, the region of the blastula and early gastrula which has
organising capacities appears to coincide with the region which will
become invaginated at gastrulation, i.e. the presumptive notochord
and axial mesoderm regions.^
This is a large area, and it might be expected that there would be
^ Fankhauser, 1930. 2 Bautzmann, 1926.
^ Bautzmann, 1928, 1929 A. * Holtfreter, 1933 b.
^ Mangold and Spemann, 1927; Mangold, 1929 b.
•^ Bytinski-Salz, 193 1. ' Kruger, 1930.
^ Bautzmann, 1926.
HEE 10
146 organisers: inducers of differentiation
•.e.v.- — ' ^
■ p.e.v.
#
p.e.v.., — ►
i'
M
B
D
Fig. 68
The regional inductive properties of the organiser and the regional reactive
properties of different levels of host- tissue in Urodele embryos. Photographs of
the embryos on which the diagram fig. 67 is based. A, Head-organiser grafted
at head level ; the secondary embryo (on the right) consisting only of a head with
ear- vesicles, and eyes fused with those of the primary embryo (f.ov.). B, Head-
organiser grafted at trunk level ; the secondary embryo (on the right) is nearly
complete but its anterior end is imperfect, it lacks eyes, and its ear-vesicles (s.e.v.)
are at a lower level than those of the primary embryo (p.e.v.). C, Trunk-organiser
at head level ; the secondary embryo (on the left) is complete, its cephalic struc-
tures (s.o.v. eyes, s.e.v. ear-vesicles) on a level with those of the primary embryo
(p.o.v., p.e.v.). D, Trunk-organiser at trunk level ; the secondary embryo (on the
right) consists only of a trunk, ending anteriorly with ear- vesicles on a level with
those of the primary embryo. (From Spemann, Arch. Entzumech. cxxiii, 1931.)
organisers: inducers of differentiation 147
regional differences in different portions of it. It will be realised
that that portion of the organiser area which is the first to become
invaginated at the rim of the dorsal lip of the blastopore will reach
furthest forward and come to underlie the head, while that portion
which becomes invaginated later will come to underlie the trunk.
It has in point of fact been found that these two portions of
the organiser show a regional difference as regards their power of
induction. For instance, "head-organiser" (invaginated early),
grafted at head level in the host, will form the cephalic axial
structures (brain, eyes, ears) as might be expected, and the
secondary embryo so formed may lack the trunk region. On the
other hand, "trunk-organiser" (invaginated late), grafted at trunk
level in the host, will form the axial structures characteristic of the
trunk, and such secondary embryos will lack brain and eyes, and
in many cases ears as well (figs. 67, 68; see also Appendix).^
Similarly, with regard to homoiogenetic induction by the neural
tube, it is found that anterior portions induce the formation of
anterior cephalic structures (e.g. eye), middle portions induce
posterior cephalic structures (e.g. ear), while posterior portions
induce structures characteristic of the trunk and tail.'^
These facts make it clear that there exists a regional differentia-
tion within the organiser area itself. The result of induction, how-
ever, is also dependent on the level along the main axis of the host
of the tissues upon which the organiser exerts its action. This is
shown by the following experiments. Head-organiser grafted at
trunk level in the host will induce the somewhat imperfect forma-
tion of cephalic axial structures, including brain, eyes, and ears.
On the other hand, trunk- organiser grafted at head level in the host
can also produce these cephalic structures, but eyes will only be
formed if the anterior end of the neural tube of the secondary em-
bryo reaches forward as far as the level of the eyes of the primary
embryo.^
Thus, as noted above (p. 140), the host-tissues are not without
influence on the formation of the secondary embryo. As a general
rule, it is found that the secondary embryo is arranged with its long
axis roughly parallel with that of the primary embryo, or, in other
^ Spemann, 1927, 1931; Bautzmann, 1929 A,
^ Mangold, 1929 b; 1932.
Ent.
d
Fig. 69
Homoiogenetic induction of neural folds by brain tissue. « A free-swirnming
larva of Triton taeniatus, with limbs {Extr.) and balancer (T. , from which a
portion of brain tissue was grafted into b, a young gastrula of the same species ;
U blastopore, c, The same embryo, 68 hours after the operation from the lett
side, showing the graft {Impl.H.). d, Section through the graft and the induced
neural tube (Med.ind.); Impl.Fas. nerve fibres, and Impl.Gangl grey matter, ot
the highly differentiated graft; Ent. endoderm, V.D. foregut of host embryo.
(From Mangold, Ergehnisse der Biol, in, 1928.)
organisers: inducers of differentiation
149
words, meridional with reference to the host, and with its head
facing in the same direction as that of the host.^
The axis of the secondary embryo is determined by the direction
taken by the mass of material which is invaginated beneath the
surface in relation to the grafted organiser
fragment. It appears that the direction in
which this invagination occurs is deter-
mined in part by the orientation of the
grafted organiser,^ but in part also by the
activities of the host-tissues: the in-
vaginated mass tends to bend round
towards the animal pole of the host. This
has been discovered by grafting portions
of organiser with their original polarity
rotated 90° or 180° relative to that of the
host, so as to lie either transversely or
reversed. In almost all cases, some in-
fluence of the host is to be observed,
but the precise degree varies a great deal
in individual instances. In some cases,
the axis of an embryo derived from a re-
versed organiser may be completely de-
flected so as to coincide with the main host shown in this figure was
, . , , induced by an organiser
axis, but m other cases it may be almost grafted with reversed orien-
precisely opposed to that of the host.^ ^^^^°^ ^"^^ the host; its
The orientating influence of the host is tli oT 'th^ p'r;'; e*!
greatest in the region surrounding the bryo; its anterior end is
blastopore, and least at the opposite pole. ""^ 1^ T""^ '"^ '^ "\^'
^ /^^ ' ^t^ pv^iv.. ^j^^ jjgg transversely to the
Un the other hand, what we may call the host. (From Spemann,
invaginating power of organisers varies, ^^''^^'- Euuumech. cxxm,
and is greater in organisers from old than
in those from young gastrulae. Consequently, reversed orientation
of the secondary embryo is most often to be observed when
^ Geinitz, 1925 a.
2 The determination of the organiser to become invaginated is an instance of
what has been called " dynamic determination" (Vogt, 1923), leading to form-
changes which in turn result in the processes of gastrulation and neurulation (see
p. 26). The possible relation between dvnamic and chemo-differentiation is
discussed below (pp. 163, 250, 301). s Spemann, 1931.
Fig. 70
The orientation of the
secondary embryo is depen-
dent partly on the polarity
of the host-tissues, and
partly on that of the grafted
organiser and the direction
in which it is implanted.
The secondary embryo
150
organisers: inducers of differentiation
an old organiser is grafted, reversed, into the antero-ventral region
of the host.^
Those cases in which the secondary embryo fails to adapt itself
to the polarity of the primary embryo are of interest because
certain of the paired structures of the secondary embryo, such as
ear-vesicles, lie at different levels in the host. In these cases it is
Fig. 71
Section through an organiser- graft in Triton, in which the anterior end of the
secondary embryo lay at right angles to the long axis of the primary embryo. The
left ear- vesicle of the secondary embryo, l.sec.a.v., which lies nearer the anterior
end of the host embryo, is larger than the right, r. sec. a. v. pr.br. brain of primary
embryo; g.c. gut cavity; sec.br. brain of secondary embryo. (After Spemann,
Arch. Entzomech. cxxiii, 193 1, simplified.)
found that the vesicle nearer to the anterior end (animal pole) of
the host is larger than the other, and this shows that there is in the
tissues of the host a stratification of capacities to react to the
organiser (figs. 70, 71 ; and see pp. 147, 319).
In addition to the regional difference between head-organiser
and trunk-organiser, it seems, however, that (contrary to previous
^ Lehmann, 1932.
organisers: inducers of differentiation 151
indications^) the organiser region is not divisible into right and
left portions possessing predetermined laterality; for a lateral
piece of primitive gut-roof, taken well to the left of the middle
V
C D
Fig. 72
The "infective" properties of the organiser region in Urodeles. A, A piece of
presumptive ectoderm from the roof of a blastula of Triton crista tus is 'grafted into
the dorsal lip of the blastopore of a gastrula of T. alpestris, where it is plainly
visible on account of its light colour. B, The graft participates in the normal
gastrulation process of the host and becomes invaginated. C, When gastrulation
is completed, the embryo is cut open and the graft is found forming^'part of the
gut-roof in the mid-dorsal line, in the position of the notochord. D, The graft
is cut out and grafted a second time into a gastrula of T. taeniatus, where
it induces the formation of neural folds. (From Spemann and Geinitz, Arch.
Entzoniech. cix, 1927.)
line, can induce the formation of a bilaterally symmetrical
secondary embryo when grafted^. One organiser region can thus
induce several embryos (see also p. 310).
1 Goerttler, 1927. 2 Spemann, 1931.
152 organisers: inducers of differentiation
The facts also permit of the interpretation that the quantitative
potency of inductive capacity falls off in a graded way from the
dorsal lip region, although this gradient appears to be steep. ^
In birds (p. 161) there appears to be a definite gradation of in-
ductive potency along the organiser (primitive streak), this being
highest anteriorly and lowest posteriorly.
The properties of the organiser are not intimately associated with
any particular type of cell. If ordinary presumptive epidermis is
grafted into the region of the organiser before gastrulation has
started, it becomes '* infected" with the power to organise. This
has been proved by heteroplastic grafting of a piece of epidermis
from Tritofi cristatiis into the organiser region of T. taeniatus. Such
a piece of tissue, originally presumptive epidermis, treated in this
way, is found when grafted into another embryo to possess all the
qualities of a normal organiser."
Thus, the properties of the organiser seem to be attached to a
certain region of the embryo, regardless of the identity of the cells
which occupy it. This region, which owes its localisation to the egg-
axis and the plane of bilateral symmetry, must be determined in
the outermost or cortical layer of cytoplasm of the ^gg. For even
when an tgg is forcibly inverted and its contents stream about in-
side, the dorsal lip of the blastopore appears in the region of thq
grey crescent, i.e. where it would normally have appeared on the
surface of the egg.^ Since, however, the cells of this region divide
more rapidly (see p. 39), it seems that some physiological activity
is set up in this region of the cortex which later affects the dividing
cells of this region, to a considerable depth below the surface.
In passing, it is of interest to note that in certain experiments,
e.g. those in which myotome and pronephros material from a
neurula were grafted into an early gastrula (p. 145), and those
referred to on p. 191, show that the morphogenesis of artificially
induced structures may differ considerably from that shown by the
same structures in normal development. Thus the epidermis may
be induced to form pronephric tubules without passing through
a nephrotome-like stage (see p. 32) : portions of brain-like structures
may be induced to form from the epidermis by thickening and
subsequent delamination without the formation of neural folds.* A
^ Bautzmann, 1926, 1933. ^ Spemann and Geinitz, 1927.
^ Weigmann, 1927. * Holtfreter, 1933 b.
organisers: inducers of differentiation 153
corresponding set of facts is known from the study of normal
events in Ascidians, where the same organ may be formed by quite
different morphogenetic processes and sometimes even from dif-
ferent germ-layers, in development from the egg and development
by budding. Similar cases are also known in regeneration.
§3
Concerning the physico-chemical aspect of the method of action
of the organiser, little can be said, although the results so far ob-
tained are of the greatest interest. In the first place, it is clear that
the inducing tissue does not require to be alive in order to exert its
effects. After an organiser has been subjected to a narcotic (tri-
chlorbutyl alcohol) for a certain length of time, the tissues of the
organiser may be so heavily damaged that they disintegrate after
being grafted, but a secondary embryo is nevertheless induced.^
Even more drastic treatment, such as desiccation, or killing with
high temperatures, or immersion for 3I minutes in 96 per cent,
alcohol, does not destroy the inductive capacity of the amphibian
organiser region.^ (See also p. 497.)
It would seem therefore that the inductive effects of the organiser
are due to some chemical substance which is elaborated by it, and
support for this view is provided by the fact that pieces of agar jelly,
or of gelatine, after being in contact with inductive tissue (neural
folds) are themselves capable of inducing.^
The question next arises as to whether the initiation of the in-
ducing effect, and therefore the productionof the necessary chemical
substance, is in any way dependent on the intimate structure of
^ Marx, 1930.
^ Here a new complication is introduced by the fact that certain tissues which
possess no inductive capacities when alive, such as epidermis and endoderm,
are able to act a organisers when killed. While the detailed significance of this
fact is still obscure, it is of interest in suggesting that the normal living
organiser differs only in some physical degree, and not in kind, from the tissues
of the remainder of the embryo. (Spemann, 1929; Bautzmann. Holtfreter,
Spemann and Mangold, 1932; Holtfreter, 1933c.)
It may here be noted that living regenerating amphibian tissue (adult newts
12-day limb regeneration-buds) is capable of inducing neural folds in blastulae
of the same species when introduced into the blastocoel (Umanski, 1932 b).
Similar results have been obtained with insertions of mammalian and avian
malignant tumour tissues (Woerdeman, 1933 c). No control experiments have
yet been made with non-malignant tissues of the same species.
^ Bautzmann, Holtfreter, Spemann and Mangold, 1932.
154 organisers: inducers of differentiation
the organiser. If three extra organisers are grafted into the close
vicinity of an organiser in an intact embryo so that their polarities
all converge to a point in the centre of the host-organiser, there is
no inductive effect of any kind.^ This annihilation of the inductive
effect is difficult to understand. It can scarcely be that an intact
structure, or an unimpeded gastrulation-process, are essential pre-
liminaries to the production of the chemical substance responsible
for the organising effect; for even if a piece of the organiser is
made to wait for some time before it is grafted, when it rolls up into
a ball, and the arrangement of its cells is markedly altered, its
organising power is not affected 'or reduced.^
The possibility that the organiser effect in birds is in some way
dependent on the normal tissue-movements which take place in
gastrulation, i.e. on so-called "dynamic determination " (Vogt), will
be discussed below (pp. 163, 250).
Recently, the decisive discovery has been made that cell-free
fractions of a liquid extract of whole neurulae can exert an organising
action, as evidenced by neural tube induction. The liquid is coagu-
lated by heat and portions of the resultant solid material implanted
into the blastocoele. The active substance is certainly ether-
soluble, and probably lipoidal.^
Meanwhile, some interesting results have emerged from investi-
gations into the glycogen-content of the cells of the amphibian
embryo. This is high in the cells of the animal hemisphere ; low in
those of the vegetative hemisphere, and intermediate around the
equator. But as soon as the cells of the organiser have become in-
vaginated, they immediately lose what glycogen they contained. It
is not improbable that this sudden disappearance of glycogen con-
notes an expenditure of energy connected with the physiological
activities characteristic of the organiser.*
§4
The fact that the organiser, in the form of the primitive gut-roof,
is capable of organising the epidermis overlying it so as to induce
it to give rise to neural folds, explains a number of phenomena
^ Goerttler, 193 1. ^ Holtfreter, 1933 b.
^ Waddington, Needham and Needham, 1933.
* Woerdeman, 1933 a; Raven, 1933 b.
organisers: inducers of differentiation
155
which would otherwise be obscure. As regards the ordinary data of
comparative embryology, this property of the organiser makes it
possible to understand why there is a correlation between the width
Fig. 73
A, Two dorsal gastrula-halves of Triton grafted together so that the directions of
invagination of their blastopores are directly opposed. B, The resulting embryo,
showing crossed doubling, or duplicitas cniciata ; each half-gastrula has produced
a posterior trunk region with spinal cord, but two heads and brains are formed, at
right angles to the axis of the trunks, each formed partly from both half-gastrulae.
(Redrawn from Morgan, Experimental Embryology, New York, 1927, after
Spemann.)
of the neural plate and the width of the primitive gut-roof in
different groups of Vertebrates:^ the former is dependent on the
latter.
Turning to experimental results, the production by operative
1 Marx, 1925.
156 organisers: inducers of differentiation
treatment of monstrosities which conform to the teratological types
known as anterior, posterior, and crossed doubhng {duplicitas an-
terior, duplicitas posterior, and duplicitas cruciata), is expUcable only
in terms of these functions of the organiser.
^o
— sek. Med. I
sek. Med. II
prim. Schw.
Fig. 74
Duplicitas cruciata, obtained by grafting together two gastrula-halves (see fig. 73) ;
nearly the whole extent of each embryonic rudiment {sek. Med. I, II) is composite
and derived partly from each of the half-gastrulae ; only the tips of the tails (prifn.
Schw.) are uncrossed, i.e. each formed from one of the half-gastrulae. One of
the trunks {sek. Med. II) is less well developed than the other, and ends
anteriorly in a knob(*). (From Wessel, Arch. Entwmech. cvii, 1926.)
If the tgg of a newt is partially constricted in the plane of bi-
lateral symmetry during the period of gastrulation, the resulting
embryo will show anterior doubling, i.e. it will have two more or
less perfectly formed anterior ends joining on to a single posterior
end.^ The explanation is that when the primitive gut-roof becomes
invaginated, it finds an obstacle in the constriction and has to fork,
one portion going forward on .each side of the constriction. The
organiser or primitive gut-roof is therefore Y-shaped, and its an-
terior prongs underlie tissue which would normally not have given
rise to neural folds. But the action of the organiser induces the
^ Spemann, 1903; Hey, 191 1.
organisers: inducers of differentiation
SI
formation of neural folds in these strange positions, with the result
that two perfect heads and anterior trunk regions are formed.
Further, it may be noted that it is impossible by the method of
partially constricting gastrulae to obtain duplicitas posterior, or
■'^T^<.
Fig- 75
Duplicitas cruciata, obtained by grafting together two gastrula-halves (see fig. 73) ;
the heads and anterior regions of the trunk have a plane of symmetry which is at
right angles to that of the posterior regions of the trunk ; the former are seen in
ventral, the latter in side view. One of the heads has a cyclopia eye. (From
Wessel, Arch. Entwmech. cvii, 1926.)
doubling of the hind end. This is obviously because the constric-
tion forces the anterior but not the posterior part of the primitive
gut-roof to fork (fig. 169).
On the other hand, both anterior and posterior doubling can be
obtained by grafting together halves of gastrulae in such a way that
their original planes of symmetry (and therefore, directions of
158 organisers: inducers of differentiation
organiser-invagination) either diverge or converge anteriorly.^
In the former case, the compound embryo will have a Y-shaped
primitive gut-roof with the divergence anterior, and will develop
double heads; in the latter case the divergence will be posterior,
and there will be double hind ends.
Perhaps the most remarkable cases of teratological development
induced experimentally are those producing crossed doubling
Herz
Med.
Herz
Fig. 76
Transverse section through a duplicitas cruciata embryo of Triton, such as that
shown in fig. 75. The hearts {Herz) are formed partly from each embryonic
rudiment, and are therefore situated laterally. Med., neural tube. (From Wessel,
Arch. Entwmech. cvii, 1926.)
(duplicitas cruciata). These result from the grafting together of two
gastrula halves each containing the dorsal lip, in such a way that the
directions of organiser-invagination are directly opposed to one
another. Invagination takes place in each half, and the primitive
gut-roofs meet one another, head on. Being unable to make any
further progress forwards, they move to each side. The result is
that the primitive gut-roofs together form a cross, two (opposite)
branches of which are formed each from one of the two invagina-
tions, and the other two branches are composite, half of each being
^ Spemann, 1916, 1918; Koether, 1927.
organisers: inducers of differentiation 159
formed from each of the two invaginations. ^ The former two
branches represent the posterior portions of the primitive gut-roof:
the latter two branches represent the anterior portions (figs. 73-76).
Overlying the cross-shaped gut-roof, neural folds arise, and a
monstrous double embryo is formed, the hinder portions of which
have each been induced by a single organiser, while the anterior
portions have been induced by tissue derived from two organisers.
Furthermore, these anterior portions are formed from induced
tissues w^hich had very different normal presumptive fates. The
relative lengths of the arms of the cross, or of the composite an-
terior and of the simple posterior portions of the double embryo,
can be controlled by varying the distance which separates the two
blastopore lips at the start of gastrulation. If they are far apart, the
primitive gut-roofs will travel a long way forwards before they
meet and cross, and the anterior composite portions will be short :
if they are close together, the gut-roofs will meet and cross very
soon, and continue their invagination as parts of the composite
anterior ends. Crossed doubling can also be obtained by grafting
an organiser into a normal embryo in such a way that the anterior
ends of the primary and secondary embryos meet and obstruct one
another.'"^
§5
Experiments on the blastoderm of the chick and duck have pro-
duced results of the greatest interest. They have shown that the
primitive streak has organising powers similar to those of the am-
phibian dorsal lip of the blastopore (with which it is morpho-
logically homologous), and they have confirmed and extended the
results obtained from experiments with amphibian material.^
In these experiments, the method of tissue culture has been used.
The embryonic rudiment of the bird at a very early stage consists
of an upper layer (ecto-mesoderm), and a lower layer (endoderm).
These layers can be separated from one another, and cultured
separately in vitro. The upper layer will differentiate into neural
folds, notochord, and mesodermal somites, but the lower layer will
not differentiate at all. This is due to the fact that the lower layer
^ Wessel, 1926. 2 Bautzmann, 1926.
^ R, Wetzel, 1929; Hunt, 1929; Waddington, 1930, 1932, 1933 a, b, c.
i6o organisers: inducers of differentiation
lacks the primitive streak which the upper layer possesses. The
lower layer is therefore in the same case as a ventral half of an
embryo of an amphibian. The organising action of the primitive
streak on the lower layer is shown by the fact that the upper layer is
capable of inducing the lower layer to give rise to the fore-gut in
the correct position with regard to the notochord, from tissue
which would normally not have given rise to fore-gut at all. This
is shown by experiments in which an upper and a lower layer are
cultivated together in such a way that the primitive streak overlies
a region of the lower layer other than that which represents the
presumptive fore-gut.
i.n.s- ^^ u.n.g.
Fig. 77
Induction by organiser in birds. Two blastoderms of the chick grafted together.
u.n.g. normal neural plate of upper blastoderm ; i.n.g. secondary induced neural
plate in upper blastoderm, formed in relation to Iji.g. normal neural plate in
lower blastoderm. (From Waddington, Phil. Trans. Roy. Soc. B, ccxxi, 1932.)
It is clear, therefore, that the primitive streak is an organiser. It
has further been found that it possesses regional differences of
potency, both as regards self-differentiating capacities and in-
ductive power. Anterior pieces of the primitive streak differentiate
into neural tube, notochord, and mesodermal somites; middle
pieces produce mesoderm with or without neural tube; posterior
pieces never produce neural tube. In other words, there is a
gradient in developmental potencies along the primitive streak.^
It should, however, be noted that when portions of primitive
streak are cultivated in isolation, they give rise to considerably
more than their presumptive fate'^ (fig. 78).
* See also Hunt, 1932.
- Waddington and Schmidt, 1933.
organisers: inducers of differentiation i6i
Anterior and middle pieces of primitive streak (corresponding
to the dorsal and lateral lips of the amphibian blastopore) grafted
beneath an upper layer induce the formation of neural folds from
host-tissue, but posterior pieces seem to be unable to do this. Thus,
in the primitive streak, there appears to be a graded distribution of
organising power. Since the induced neural tube is usually situated
immediately above the mesodermal tissue of the graft (corre-
sponding to the primitive gut-roof of the amphibian organiser),
the latter is probably responsible for the inductive effect. The
notochord in the bird is apparently unable to induce.^
When a graft which in the normal course of development would
have formed trunk mesoderm is implanted into the head region, it
produces only head mesoderm there, whether or not it succeeds in
inducing the formation of a secondary embryo. This shows that
there must be some influence of the host-tissues on the fate of the
grafted organiser.'^
The homoiogenetic power of the neural tube has been demon-
strated in birds, for a grafted portion of neural tube will induce the
formation of neural tube^ (fig. 77).
The organising action of the avian primitive streak is not species
specific, for the primitive streak of the duck is functional when
grafted into the blastoderm of the chick, and vice versa}
The orientation of the avian embryo is found to be dependent
on the polarities of both the primitive streak and of the lower layer.
The influences of the upper and of the lower layer are tested by
rotating the one relatively to the other through 90° or 180°, and
culturing them together. As in the comparable experiments in
Amphibia (p. 149), in which rotated or reversed organisers are
grafted, the results vary considerably in different individual cases.
In some, the orientation of the primitive streak, and therefore of
the upper layer, determines that of the embryo. In other cases,
however, the embryo is developed in relation to the polarity of the
lower layer or endoderm. The polarity of the upper layer is then
either deflected or obliterated. This is very remarkable, for, as
already stated, the lower layer lacks the primitive streak from which
all the axial structures of the embryo are formed.^
1 Waddington, 1933 b. See also Umanski 1932 a. '^ Waddington, i933 b.
- Waddington and Schmidt, 1933. ^ Waddington, 1933 C
1 62
prospective
fate
actually
' obtained
reacts by forming a
head neural plate
I
induced head
neural plate
Induced head
The influence of the host
overcomes that of the graft
Fig. 78
Diagrams illustrating some of the properties of the organising centre in birds.
A, The developmental potencies of a portion of the organiser region are greater than
its prospective fate. B, Analysis of the problem presented by the fact that when
a piece of the organiser region, the prospective fate of which is trunk mesoderm,
is grafted into the head region of another blastoderm, it itself gives rise to head
mesoderm, while at the same time inducing the formation of neural folds (B i,
B 3). The conversion of the graft into head mesoderm may be explained by
assuming either: B 2, that after the graft has induced the formation of a head
neural plate the latter in turn acts upon the graft and determines it to give rise to
head mesoderm; or B 2^, that the conversion of the graft into head mesoderm
is due to a process of interaction between the graft and the host's own organising
centre, to which latter the property must be ascribed of exerting an influence
over an area of given extent, termed an " individuation-field ", in which the whole
complex of tissues are controlled in such a way as to lead to the formation of a
complete individual. It is, further, an effect of the host's individuation-field that
the neural plate which trunk mesoderm induces out of the host- tissues in the head
region is head neural plate. That alternative B 2^ is the correct interpretation
follows from the cases, C i-C 3, in which the grafted trunk mesoderm in the
head region of the host becomes converted into head mesoderm without inducing
the formation of a neural plate at all: here, the graft can only be under the in-
fluence of the host organising centre. (From Waddington and Schmidt, Arch.
Entwmech. cxxviii, 1933.)
organisers: inducers of differentiation 163
It seems that the endoderm can determine the polarity of the
embryo by determining the locahsation and polarity of the primi-
tive streak itself, in the blastoderm overlying it. We are here con-
fronted with a phenomenon which seems to be nothing less than
the determination of the organiser itself. The primitive streak is
dependent in some way ultimately on the endoderm, and it would
seem that we have to look for the morphogenetic expression of this
determination in certain streaming movements which take place in
the blastoderm. The direction of these movements is backwards
along the periphery on each side, and forwards along the central
line, immediately along which line the primitive streak is formed.^
In some as yet undetermined way, the endoderm seems to control
these movements.
If this should turn out to be correct, we have here an example
of the effects of "dynamic determination" referred to on p. 154.
From the theoretical point of view, the interest and importance of
these facts lies in the question whether dynamic determination can
be regarded as the causal antecedent of '* material" (chemical and
histological) determination. The answer to this question appears
to differ in different groups of vertebrates. In the amphibian
embryo the early stages are characterised by well-marked move-
ments (dynamic effects) of tissues ; and attempts made to test the
power of chemo-differentiation of other tissues which have been
prevented from undergoing such movements ^ have yielded results
which can only be regarded as negative (see Chap, vii, p. 250). For
the moment, therefore, the general significance of dynamic deter-
mination in birds must remain an open question.
With regard to the physico-chemical nature of the action of the
organiser in the bird, it is interesting to note that it retains its
organising capacity although coagulated as a result of having been
dipped in a thin glass tube into boiling water for 30 seconds.^
From all these results, it is abundantly clear that the dorsal lip
of the amphibian blastopore, and its homologue the avian primi-
tive streak, possess the function of an organiser, and it is probable
that these structures will be found to have similar properties in
other groups of Vertebrates.
^ R.Wetzel, 1925, 1929; Graper, 1929.
^ Goerttler, 1927 ; Holtfreter, 1933 a. ^ Waddington, 1933 a.
164 organisers: inducers of differentiation
§6
Attention may now be turned to Invertebrates, and the question
naturally arises whether regions with similar or comparable organ-
ising capacities xist among them. This is found to be the case,
although the details, not unnaturally, vary considerably.
In Hydro, the hypostome, or region surrounding the mouth, is
an organiser of simple type. When grafted into the proximal end
of another polyp it induces the formation of tentacles. It further
Fig. 79
Organiser grafts in Hydra. Induction of a bud by grafting an oral end of one
individual {a) into the flank of another. The polarity of the bud is the reverse of
that of the graft. (From Mutz, Arch. Entiomech. cxxi, 1930.)
causes an outgrowth of host-tissues in which the original polarity
is overridden, and a new polarity established in relation to that of
the graft. ^ Although grafts of organisers in Hydra between different
species rarely succeed, an organiser from Pehnatohydra has been
found to produce an inductive effect in Hydra (figs. 79, 80).
In another Coelenterate, Cbrymorpha, pieces of stem have the
power of inducing the formation of new polyps when grafted into
other stems." This case is particularly interesting, for the facts
^ Browne, 1909; Rand, Bovard and Minnich, 1926; Mutz, 1930.
^ Child, 1929 B, 1932.
organisers: inducers oe differentiation
65
indicate that the organiser in Coiymorpha is not a specific tissue or
structure, but any level of the stem will act as an organiser, although
pieces from distal levels are more
potent. The bearing of these facts on
the theory of gradient-fields and the
interpretation of the mode of action of
organisers will be discussed at greater
length in Chap, viii (fig. 138).
In Platiaria, the head of one worm
grafted into the posterior region of the
body of another induces the formation
of a pharynx and brings about the re-
organisation of the host-tissues so as to
make them conform to the new polarity
set up by the graft. Here again, the
effect is not species-specific, for a head
of Planaria dorotocephala will act as an
organiser in the tissues of Planaria
maciilata^ (figs. 81, 82).
These last experiments merely extend
previous w^ork on regeneration in Plan-
arians and various worms. In Planaria,
for instance. Child had shown ^ that the
reorganisation of the old tissues of a
posterior fragment, of which the most
obvious eflFect is the production of a
new pharynx, only occurs if a head is
regenerated. He further showed, in ex-
periments where the size of the re-
generated head was varied and controlled
by the use of anaesthetics in varying (From Mutz, Arch
concentrations, that the size of the new '"^^^'- ^xxi, 1930.)
pharynx and its distance from the anterior end of the piece were
correlated with the size of the regenerated head (see Chap, viii,
pp. 287, 290).
As noted in Chap, viii (p. 288) the head segments of the Poly-
chaete worm Sabella act during regeneration as an organiser capable
^ Santos, 1929. 2 See Child, 1915 a, pp. 102, 138.
Fig. 80
Organiser grafts in Hydra.
Bud (k) induced from stock
{b) by grafting an oral end of
another individual {a) on to
the aboral end of the stock.
Entw-
1 66
organisers: inducers of differentiation
of transforming more posterior segments from the abdominal to the
thoracic type.^ Similar facts have been noted for the Oligochaete
Stylaria} These results will be further considered in connexion
with gradient-fields (fig. 137). They are of great importance in
showing that the processes at work in the organiser phenomena in
the early stages of Vertebrate development are similar in essentials
to those operating throughout life in regeneration and grafting
experiments in lower forms.^ There are, however, certain differences,
in that the vertebrate organiser works mainly by contact, whereas
3 5
Fig. 81
Head-grafts in Planaria. 1, 2, 3, showing portions employed as grafts; 4, 5,
isolated fragments of the type of i and 3 respectively, 12 days after operation.
(From Santos, Biol. Bull. LVii, 1929-)
these invertebrate organisms can effect a reorganisation of tissues
at a distance. For a further discussion of this point, see Chap, viii,
p. 310.
Organiser phenomena in normal ontogeny, though again of a less
specialised type than in Amphibia, have been found in Echino-
derms. A curious result (referred to in Chap, v, p. 102) of the
isolation of animal halves of eggs and blastulae of the sea-urchin
Paracentrotus, is that such halves do not develop a stomodaeum if
they have been isolated from their vegetative counterpart at a
stage earher than 20 hours after fertilisation. A stomodaeum is,
1 Berrill, 1931. " Harper, 1904. ^ See Child, 1928 c, 1929 a.
organisers: inducers of differentiation
Fig. 82
Organising action of engrafted heads in Planaria.
■ a, A lateral post-pharyngeal graft has induced a
lateral outgrowth and a secondary pharynx, c, A
sub-terminal graft has caused a reversal of polarity
in the terminal portion of the host, and has induced
two secondary pharynges. d, A terminal graft has
induced a single secondary pharynx with reversed
polarity ; tlie region of reversed polarity is capable
of considerable autonomy of movement (dotted
outline), b, A sub-terminal graft has induced two
secondary pharynges, and a marked reversal of
polarity in the host's terminal region, a, b, c,
heteroplastic grafts, d, homoplastic graft. (From
Santos, Biol. Bull, lvii, 1929.)
i68 organisers: inducers of differentiation
however, developed in animal halves which have been isolated
later than this. This result can only be understood on the view
that the vegetative half of the egg and embryo contains a factor
whose presence and action for a certain minimum period of time
is essential for the production of a stomodaeum in the animal
half.i
Further experiments suggest that this factor is situated at the
vegetative pole of the egg, where invagination takes place and the
blastopore arises. Recent improvements in technique have made
it possible to assemble certain definite blastomeres, or groups of
blastomeres, of the sea-urchin, at will. At the i6-cell stage, there
are normally eight mesomeres (presumptive ectoderm) : four macro-
meres (the animal half of each of which is presumptive ectoderm,
the vegetative half, endoderm) : and four micromeres (presumptive
mesenchyme). Embryos artificially assembled and consisting of
sixteen mesomeres, four macromeres, and four micromeres; or of
the even more abnormal combination of twelve mesomeres, two
macromeres, and two micromeres (in each case, therefore, contain-
ing too much presumptive ectoderm), develop into normal pluteus
larvae. There is present, therefore, a regulating agent which organ-
ises the available material to form a harmoniously proportioned
larva. That this agent is situated at the vegetative pole of the egg is
probable from the facts that vegetative tissue must be present if
gastrulation is to take place at all, and that the micromeres (which
occupy the vegetative pole) are predetermined to initiate invagina-
tion, and do so wherever they may be grafted. Further, if an embryo
at the i6-cell stage is divided meridionally and the two halves are
stuck together again so that their axes of polarity are reversed in
respect of one another, invagination takes place at each end, where
the micromeres are situated, and the resulting larva is a double
monster, with two guts, skeletons, etc. This can be understood if
the micromeres act as organisers^ (fig. 83).
But, as in the case of Cory?norpha, this sea-urchin organiser is not
specifically located in or restricted to the micromeres, for if they
are removed, the next most vegetative material can function as an
organiser, and induce the formation of a fairly normal pluteus larva.
But if no material from the vegetative hemisphere is present, there
^ Horstadius, 1928.
o a
■M 3
3J
"rt
QJ
c S
■^
£
■-
o tc
OJ
'^
-o X
bx)
a; o
OJ
■'-' C3
f )
^
13
C
C3
'"c
OJ
2
OJ
2^
T3
O
s-^
C
<u
p
H
OJ
't,
0)
>
bio;
-co
M
^5
1 70
organisers: inducers of differentiation
is no regulation and nothing resembling a pluteus is formed. Further
consideration of this and related phenomena is reserved for Chap. ix.
Lastly, a region of cytoplasm essential for subsequent differentia-
tion of the embryo has been discovered in Insects. As mentioned
Fig. 84
The activating centre in early development of the dragon-fly Platycnemis.
a, The egg was constricted tightly (some nuclei already present in posterior
portion). No development anterior to constriction, owing to inability of the
activating substance to reach it, b. The egg was constricted loosely, and the
activating substance was able to diffuse, and development occurs in the anterior
portion, c. Constriction behind the activating centre does not interfere with
development, d. Normal embryo resulting from c. e, f, Constriction and ex-
clusion of nuclei from the activating centre: no development. (From Morgan,
Experimental Embryology , Columbia University Press, 1927, after Seidel.)
in Chap, v (p. 128), the hinder end of the tgg of the dragon-fly con-
tains a region, whose destruction, or isolation by means of a con-
striction, prevents development of the embryo. Before this region
can exert its activity, it is necessary that nuclei should reach it ; if
the nuclei are prevented from doing so (by a constriction drawn
organisers: inducers of differentiation 171
only just tight enough to prevent their passage), there is no de-
velopment. After the nuclei have reached this region, a substance
appears to be given off from the activating region, and to diffuse
through the egg in an anterior direction. As time goes on, increas-
ingly large portions of the hinder part of the egg can be destroyed
without interfering with development, which shows that as soon
as a region has received the diffused substance, it is no longer de-
pendent on the activating region^ (fig. 84).
This activating region differs, however, from an organiser in that
it is not concerned with the differentiation of this or that structure
in any particular position : it is merely a starter or activator, con-
ferring on the remaining regions of the Ggg the power to undergo
development.
Farther forward in the insect egg, in the region which will
normally give rise to the thorax, the existence of a differentiating
centre has been established. For the differentiation of the regions
anterior to this, it is necessary not only that the activating substance
from the activating centre should have reached the differentiating
centre, but there must be cellular continuity between the differ-
entiating centre and the regions of the blastoderm anterior to it.
The activating substance, on the other hand, diffuses freely through
the egg, whether the cells are in continuity or not. It follows, either
that the differentiating centre absorbs this substance in its cells and
distributes it from cell to cell, or that it initiates a new chain of
reactions. At all events, the differentiating centre is responsible for
the localisation and determination of the various regions of the
embryo, and its presence is necessary if a properly and harmoniously
proportioned embryo is to result from an egg in which an anterior
portion has been isolated by constriction, or by a discontinuity
between the cells of the blastoderm.^
Further results must be awaited before the question can be
answered as to whether the mode of action of the Insect differ-
entiating centre is comparable with that of the organiser in other
groups.
It has been mentioned (Chap, v, p. 113) that in Chaetopterus and
in Tubifex, when the polar lobe or pole-plasms are equally dis-
tributed to the first two blastomeres, instead of being restricted to
^ Seidel, 1929, 193 1. ^ Seidel, 193 1.
172 organisers: inducers of differentiation
the blastomere CD, double monsters, each member of which
possesses a complete set of trunk organs, are produced. It might
therefore seem as though the somatoblasts (for the formation of
which the polar lobe or pole-plasms are essential) of Annelids and
Molluscs deserve the title of organiser.^ It is, however, unlikely
that these exert an effect similar to that of the amphibian dorsal
lip, or of an engrafted Planaria head : it is more probable that the
growth processes initiated by a single somatoblast automatically lead
to the production of a more or less complete set of trunk organs.
A hitherto unique type of determination is found in the wings of
moths. As is well known, in Lymantrta, intersexual types can be
produced by appropriate crosses of local races (see Chap, xii, p. 409).
The normal wing pigmentation is white in females, dark in males.
In male intersexes (i.e. animals which begin adult differentiation
as males but continue it as females) the wing shows a mosaic of
white (female) and dark (male) pigmentation. The quantity (total
area) of female-type pigmentation is directly proportional to the
degree of intersexuality as measured by other secondary sexual
characters, but the pattern is irregular and varies from specimen to
specimen. Careful observation shows that the limits of the male
and female areas are defined in reference to the course of the veins.
The appearance is as if there had been a flow of a certain quantity of
dark pigment through the veins. ^ However, from other work we
know that pigment deposition occurs in relation to the determina-
tion of the scales. The scales, if determined as female, develop
quickly ; if as male, develop slowly. The visible determination of
sexual type can be seen to occur long before the wings become pig-
mented. Meanwhile the processes leading to the deposition of
white pigment occur some time before the end of pupal develop-
ment, and those leading to the deposition of melanin occur later.
Pigment can only be deposited during a certain stage in the develop-
ment of a scale. Thus in the female the white pigment-precursors
find the female-determined scales at the right stage, while later,
when the processes leading to melanin-production occur, no scales
are available in which it can be deposited. The reverse is true in
males. The result is brought about by interaction of two indepen-
dent sets of processes.
^ E. B. Wilson, 1929. ^ Goldschmidt, 1923, 1927.
organisers: inducers of differentiation 173
The intersexual males demonstrate that the determination of the
scales must in them occur as the resuh of the streaming out of some
chemical agent, responsible for initiating male-type determination,
from the body over the wings along the course of the veins. There
exists what Goldschmidt calls a " stream of determination ". Slight
variation in the resistance of the various veins will lead to large
individual variation in the precise course of the flow\ Normally
after a time the flow reaches every part of the wing. But in the
intersexes, if the switch-over from male to female metabolism
occurs during the time occupied by the outstreaming of this sub-
stance, all the parts which it has not yet reached will develop as
female, and the male-determined areas, later becoming coloured
with melanin, remain as a record of the early course of the flow.
We may presume that there is some passage of a determining
substance from the organiser to other regions in the pre-gastrulation
stage of the amphibian egg (see p. 139); but this is the only case
known where one must postulate a flow of such a substance along
anatomically diflFerentiated channels. Much remains to be cleared
up as regards this phenomenon. For instance, it manifests itself in
certain rare cases among female intersexes, but the wings of these
are usually whole-coloured and of male (dark) type.
§7
Organiser phenomena are clearly special cases of what Roux
termed dependent differentiation. As noted in Chap, in, p. 54, we
will use the term in its restricted sense for cases in which the diflFer-
entiation of one part depends, in one way or another, upon the
presence and previous differentiation of another part. The factors
concerning dependent differentiation fall into several rather difTerent
categories.
It will be convenient first to give brief consideration to those
effects which commonly begin to operate after the functional
period of development has started. These are of various distinct
types. First, there are the morphological effects of hormones, such
as the influence of the gonad hormones upon secondary sexual
characters in vertebrates. Another example, this time from inver-
tebrates, concerns the differentiating capacities of rudiments of
insect organs, which have been tested by means of explantation
174 organisers: inducers of differentiation
experiments. If the leg imaginal discs of mature blow-fly larvae are
cultured in vitro in media of inorganic salts or of larval body-fluid,
they will remain healthy for several days, but will not develop. If
now the larval body-fluid is replaced by pupal body-fluid, or if the
cultures had been put up in this medium straightaway, the leg
imaginal discs become evaginated and grow into segmented limbs.^
They do not, however, develop beyond the stage corresponding to
the fifth day of pupal life, and this is possibly due to the absence
from the culture medium of some substances necessary for further
development. At all events, it is clear that the differentiation of the
leg-rudiments is dependent on changes which occur in the body-
fluid at the onset of pupal life. Other experiments have shown that
the process of moulting, so characteristic of insect development, is
a reaction of the epidermis to a substance in the body-fluid,
amounting to a hormone.^
Next, there are the effects of other substances carried in the
blood stream. The classical example of this concerns the pig-
mentary pattern of the late embryo of the fish Fundulus. The
pattern is due to the pigment-cells arranging themselves along
the blood-vessels of the yolk-sac, i.e. in situations where the
maximum amount of oxygen is available.^
Then, there are the trophic effects of nerves, such as the de-
pendence of the differentiation of taste-buds in the fish Amiurus
upon contact with the nerve endings of the facial nerve* (see p. 430).
Finally, there are the moulding effects of pressure and tension upon
the form, size, and intimate structure of such organs as sinews and
blood-vessels (see p. 432).
With regard to differentiation of the type seen in the prefunc-
tional period of development, there are, at the opposite extreme
from the organiser phenomena, effects primarily mechanical in
nature. An example of these is seen in the dependence of the arms
of the pluteus larvae of sea-urchins upon the growth of the larval
skeleton. In the absence of the skeleton, no arms are produced : if
an abnormal number of skeletal spicules are formed, a correspond-
ing number of arms are produced : if the spicules are abnormal in
^ Frew, 1928. 2 von Buddenbrock, 1930; Bodenstein, 1933.
^ J. Loeb, 1912.
* Olmsted, 1920; G. H. Parker, 1932 A, b.
organisers: inducers of differentiation 175
position, so are the arms.^ It appears that the formative stimulus
consists in the continuous pressure exerted on the epidermis by
the growing tips of the skeletal spicules. But, as we shall shortly
see, the position of the skeletal spicules is itself under the control
of the epidermis, and therefore arms and spicules are, in a measure,
mutually dependent."
Other examples of dependent differentiation are seen in the ad-
justment of the skeleton of Vertebrates to the underlying organs.
For example, if the rudiment of the optic-cup is extirpated in early
amphibian embryos, when the cartilaginous cranium comes to be
formed, the skeleton of the orbital region is markedly smaller on
the operated side, and, in certain respects, irregular."^ When foreign
structures, e.g. mesonephros, are grafted in place of the mid-brain,
the cartilaginous cranium is distorted by the increased intracranial
pressure due to the graft. ** When the rudiment of the nasal sac is
extirpated, the cartilages of the nasal region arise by self-differentia-
tion, but the nasal capsule is completely collapsed : the normal form
of the nasal capsule is attained through the cartilage adjusting its
growth to the form of the nasal sac.^
Of a rather different nature, however, is the relation of the carti-
laginous auditory capsule to the primary ear-vesicle. In this case,
the cartilaginous capsule wholly fails to develop if the vesicle has
been extirpated at an earlier stage. Conversely, a grafted ear-
vesicle may induce the formation of a cartilaginous capsule around
it. The dependence has been shown to obtain both in amphibian^
and in avian^ embryos. Here it would appear that a chemical
stimulus from the ear-vesicle is necessary to initiate cartilage pro-
duction by the neighbouring mesenchyme, though, doubtless,
mechanical factors play a part in the later growth of the capsule.
The effect is not species-specific, for an ear-vesicle of Rana can
induce the formation of a cartilaginous capsule from tissues of
Amhly stoma, when grafted into an embryo of that animal.'^
In the fish Acipenser, the relations between the ear-vesicle and
the cartilaginous capsule are slightly different, but resemble those
^ Herbst, 19 12. - Runnstrom, 1929.
^ Steinitz, 1906. * Nicholas, 1930.
^ Burr, 1916.
•^ Filatow, 1916; Luther, 1925; Guareschi, 1928.
' Reagan, 1917. ^ Lewis, 1906.
176
'•^
rf"^.
Fig. 85
The formation or non-formation of a given structure depends not only on the
presence of an inductive or formative stimulus (organiser), but also on local
specific factors, intrinsic to the fields. Triton taeniatiis normally develops a
balancer, the axolotl does not. Nevertheless, a piece of gut-roof of axolotl
grafted into an embryo of Triton can induce the balancer field of the latter to
develop supernumerary balancers, A. Conversely, B, a piece of trunk epidermis
of Triton grafted on to the head of an axolotl embryo, gives rise to a balancer (left
side of photograph), while no balancer is formed from the normal axolotl
epidermis on the other side. The axolotl therefore possesses the necessary
formative stimulus for balancer formation, but its epidermis fails to react to it.
g. grafted axolotl tissue; w. normal balancer; s. supernumerary balancer. (From
Mangold, Natunuiss. xix, 193 1.)
organisers: inducers of differentiation 177
between the nasal sac and olfactory capsule described above. After
removal of the ear-vesicle in Acipe?iser, no auditory capsule is
formed, but a shapeless chondrification appears in its place. Here,
then, the actual formation of cartilage is independent of the ear-
vesicle, but the differentiation of the cartilage into an auditory
capsule is dependent.^
Of a different nature again is the response of the uterine mucosa
to the presence of foreign bodies in the uterus.'^ Any foreign body —
glass, platinum wire, paraffin, etc. — causes a proliferation of the
mucosa essentially similar to that w^hich it shows as a result of the
implantation of the embryo. In both cases, the proliferation will
only occur provided that certain of the ovarian hormones are present
in the blood-stream. Here, the response is not the direct result of
mechanical forces, as with the arms of the pluteus. A somewhat
similar case from the early stages of development is the effect of
grafting foreign objects under the flank ectoderm of Urodele em-
bryos : these in some cases induce the formation of supernumerary
limbs. These experiments w^ere first performed with ear-vesicles,^
but it has since been found that inorganic objects, such as celloidin
beads, have the same effecf* (see also Chap, x, p. 362, for the
effect of nerve- endings on limb- induction). The type of structure
induced thus appears to be determined by local regional factors,
regardless of the specific nature of the graft, which acts as a
releasing mechanism (see Chap, vii, p. 231).
Another case which appears to be comparable is the induction of
supernumerary balancers in Triton within the balancer field (see
p. 236) by means of grafts of neural crest cells of Rana,'' or of
anterior neural plate cells, or even of fore-gut wall-cells of Ambly-
stoma tigri?tum, which donors possess no balancer.^ These cases
serve as a further illustration of the fact noted above (p. 140), that
local regional properties of the tissues acted upon, as well as the
properties of the releasing mechanism, do play a part in deter-
mining the quality of the induced structure. The relative import-
ance of these two sets of factors varies in different cases: the
amphibian organiser is capable of overriding nearly all the local
^ Filatow, 1930. 2 L Loeb, 1908.
^ Balinsky, 1925-6; Filatow, 1927. ^ Balinsky, 1927.
'" Raven, 1931A. ^ Mangold, 1931c.
H F E 12
178 organisers: inducers of differentiation
properties of the tissues acted upon : the various grafts mentioned
in the previous paragraphs do not override the local regional
potencies, but merely evoke them (hg. 85).
We may now return to cases in which the dependent differentia-
tion appears definitely to be due to chemical effects arising from
proximity with some other organ. A classical example is the de-
pendence of the conjunctiva^ upon the presence of the eye.
In the absence of contact with an optic vesicle, the epidermis of
the presumptive conjunctiva region remains pigmented and opaque.
If, however, contact is established, it loses its pigment and becomes
transparent.^ This effect is exerted not only by the whole optic
vesicle, but also by portions of the retina, by the lens alone, and
even by disorganised fragments of the optic vesicle grafted under
the skin. It even appears that an engrafted limb occupying the place
of an eye is capable of inducing the differentiation of the conjunctiva.^
Pieces of already differentiated epidermis from other regions
grafted over the eye, or when eye or lens is grafted under them,
can be induced to undergo modification into conjunctiva.*
A case which is in many ways comparable with that of the con-
junctiva is provided by the Anuran tympanic membrane. This
structure is differentiated out of the epidermis at metamorphosis
by means of processes involving histolysis and reconstructions of
certain layers. Here, the annular tympanic cartilage is the structure
on which the differentiation is dependent. Epidermis from other
regions of the body will differentiate into tympanic membrane if
grafted over the tympanic cartilage, and if the cartilage is extir-
pated, no membrane forms. Tympanic cartilage grafted under the
skin of the back induces the formation of a tympanic membrane
in that place.^ A similar case is the differentiation of an articular
cup on the palato-quadrate to fit the base of the balancer. This is
dependent on the presence of the balancer, and its formation can
^ The conjunctiva is of course the epidermal, and the cornea the mesodermal
layer of tissue overlying the pupil and lens. These terms have often been used
very carelessly, the conjunctiva being called the cornea, and vice versa. In most
cases, the experiments have not been carried on long enough for the cornea to
become properly differentiated.
2 Spemann, 1901 a; Lewis, 1905.
^ Diirken, 1916.
^ Fischel, 1917; W. H. Cole, 1922; Groll, 1924,
^ Helff, 19^8.
organisers: inducers of differentiation 179
be induced by balancer grafts even in species which normally have
no balancer.^
In Anura, the nervous portion of the pituitary (infundibulum
and pars nervosa) is dependent for its full differentiation and growth
upon contact with the epithelial or hypophysial invagination,
which originates from the epidermis of the front of the head, and
later gives rise to the pars anterior, intermedia, and tuberalis.'- If
the hypophysial rudiment be extirpated or destroyed, the infundi-
bulum and pars nervosa fail to develop normally, both as regards
size and qualitative differentiation.
In chick embryos, it appears probable that contact of the heart
rudiment with the endodermal gut-floor is necessary for the latter
to undergo differentiation into a liver.^ (In Amphibia, however, the
liver appears to possess marked powers of self-differentiation : see
Chap. VII, p. 203.)
In tissue cultures, it has been conclusively shown that the
differentiation of kidney-epithelium into characteristic tubules is
dependent on the presence of connective tissue. When cultivated
alone, the kidney-tissue merely forms an undifferentiated sheet.*
Similarly, tissue-cultures of mammary gland carcinoma may be
induced to redifferentiate into structures resembling the acini
of m^ammary gland by addition of connective tissue. Again,^
epithelial tissues grown in culture tend to dedifferentiate unless
connective tissue is present also,^ and cultures of chick-epithelium
can be induced to differentiate into structures resembling salivary
glands by the addition of fibroblasts."
The perforation of the mouth in Urodele embryos is preceded by
a reduction of the ectoderm from a two-layered to a one-layered
condition, and by the sinking in of the stomodaeal depression. It
has been found that these processes, and the consequent perfora-
tion of the mouth aperture, is dependent on the establishment of
contact between the ectoderm and the underlying endoderm of the
fore-gut. The latter is capable of inducing these changes even when
ectoderm from other regions is grafted in place of the normal
' Harrison, 1925 b. - Smith, 1920.
^ Willier and Rawles, 193 1 a. * A.H.Drew, 1923 ; see also Rienhoff, 1922.
^ A. H. Drew, 1923. '^ Champy, 1914.
^ Ebeling and Fischer, 1922.
i8o organisers: inducers of differentiation
stomodaeal ectoderm.^ [See also p. 498.] The perforation of the
choanae, on the other hand, is dependent on the estabhshment of
contact between the nasal rudiment and the endodermal roof of
the mouth. Even a rudimentary nasal pit is capable, provided it
establishes contact with the endoderm, of inducing the latter to give
rise to a typically normal choana.^
A curious case is that of the perforation of the operculum in
Anuran larvae at metamorphosis. This occurs on the right-hand
side, allowing the right fore-limb to emerge (the left fore-limb
emerges through the spiracle, an aperture which has been present
since the first formation of the operculum). It was at first supposed
that this was due to mechanical pressure exerted by the growing
limb. Then it was discovered that perforation took place even if
the rudiment of the fore-limb had previously beeen extirpated.^
Finally, it has been established that the perforation is caused by a
substance produced by the gills as they degenerate during meta-
morphosis.* The degenerating gills will cause perforation of the
skin in any region if grafted beneath the surface (see p. 429 and
figs. 208, 209).
In Echinoderms, it has been shown that the formation of the
amnion and of large portions of the rudiment of the adult sea-
urchin are dependent on the presence of the hydrocoel. This follows
from the cases in which the abnormal presence of a right hydrocoel
is accompanied by the formation of a right amnion and echinoid
rudiment, with dental sacs, perihaemal rudiments, oesophagus, and
mouth. ^ The size of this echinoid rudiment is correlated with that
of the hydrocoel;^ and in those cases in which by experimental
treatment the position of the left (normal) hydrocoel is altered, it
is found that the amnion and adult echinoid rudiment arise im-
mediately over the hydrocoel wherever it happens to be, and not
from their presumptive tissues.^ At the same time, it seems that
the presence of the amnion is necessary for the complete diflFerentia-
tion of the hydrocoel, so that we are here confronted with a case of
mutual dependence.^
Another example of this, also from Echinoderms, concerns the
^ Adams, 1924, 193 1. " Ekman, 1923.
3 Braus, 1906. ^ Helff, 1924, 1926.
^ MacBride, 1911, 1918. ^ von Ubisch, 1913.
^ Runnstrom, 1918. ^ Runnstrom, 1929.
organisers: inducers of differentiation i8i
location of the skeleton in the larva. This skeleton arises from groups
of primary mesenchyme cells, which are normally to be found on
either side of the blastopore at the close of invagination. If these
cells are scattered through the blastocoel by shaking, they return
to their original position.^ It would appear that the ectoderm near
the line of its junction with the endoderm exercises a specific
attraction on these mesenchyme cells, and this view is further sup-
ported by the following experiments. When sea-urchin larvae are
made to develop in water to which lithium salts have been added,
the proportion between the relative amounts of tissue devoted to
the formation of endoderm and ectoderm is altered, to the ad-
vantage of the endoderm and at the expense of the ectoderm, with
increasing concentrations^ (see p. 334). The ectoderm may be re-
duced to a tiny region occupying the animal pole, and in such
larvae, the skeleton-forming cells are to be found there, and not in
their normal position near the vegetative pole.
The ectoderm is thus responsible for the localisation of the
skeleton-forming cells, and, in addition, it appears to control certain
details of the growth of the skeleton itself. The mesenchyme cells
secrete a triradiate spicule, apparently as an act of self-differentia-
tion. The type of the spicule is also a result of self-differentiation,
as is clearly seen in those experiments in which micromeres (pre-
sumptive skeletogenous mesenchyme cells) of Echinocyamus (which
normally possesses complex spicules) are grafted into the animal
half of a blastula (presumptive ectoderm) of Echinus (which normally
possesses simple spicules). The spicules ultimately developed in
such larvae are of the complex type.^ But the growth of the various
spicules and struts characteristic of the pluteus skeleton is depen-
dent on the ectoderm. This has been shown by experiments
similar to those described above, in which the relative proportions
of ectoderm and endoderm are varied. If the ectoderm is very
deficient, skeleton production goes no further than the triradiate
stage, in spite of the fact that the mesenchyme cells are present in
ample quantity. With increasing development of ectoderm, and
particularly of the ciliated band, there is progressive development
of the skeletal arms.* As we have already seen (p. 174), the pressure
1 Driesch, 1896. - Herbst, 1895.
^ von Ubisch, 193 1. * Runnstrom, 1929.
1 82
organisers: inducers of differentiation
of the tips of the skeletal spicules against the ectoderm in the region
of the ciliated band is necessary for the formation of the arms.
The ectoderm and skeleton are therefore mutually dependent in
the formation of the arms.
To'p.
Fig. 86
Eye development in Amphibia. Above, early neurula showing neural plate
(Mpl.) and limits of presumptive eye-rudiments (Au.). Below, left, section of
early optic cup, with tapetal (Tap.) and retinal (Ret.) layers, and epidermis
proliferating to form the lens rudiment (L.). Below, right, eye at onset of
functional stage. C. cornea. The central portion of the lens has differentiated
into lens fibres. (From Mangold, Naturwiss. xvi, 1928, figs, b, e,f.)
It should be noted that under the influence of abnormal environ-
mental agencies, the course of local differentiation may be markedly
modified. One example is the formation by frog embryos, markedly
retarded by being kept in solution of KCl, of an almost solid
neural tube, recalling that normally found in the development
of PetrotJiyzon and Teleosts. Another, of extreme interest, is the
organisers: inducers of differentiation 183
development, in tadpoles arising from eggs kept in urea solutions,
of patches of tissue, within the nerve-cord or the gut, whose histo-
logical structure is identical with that of the notochord. These
patches of ectopic notochordal tissue are always adjacent to the
true notochord. It would appear that there has been some spread
of the factors responsible for this particular histo-differentiation,
possibly by the diffusion of specific substances from the notochord-
rudiment. As Lehmann {Naturwiss. xxi, 737) has recently shown,
lithium treatment results in differential reduction of the trunk-
notochord in Triton.
Fig. 87
Spread of notochordal type of histo-differentiation to neighbouring organs in
frog tadpoles reared in 1-5 per cent. urea. Left, notochordal differentiation in the
gut-roof. Right, notochordal differentiation in the nerve cord. Below the noto-
chord in each case is the sub-notochordal rod. (Redrawn after Jenkinson, Arch.
Entiumech. xxi, 1906.)
§8
We have left to the last what is the most celebrated example of de-
pendent differentiation — the formation of the lens of the vertebrate
eye from the epidermis under the influence of the eye-cup (fig. 86).
The matter, however, is not simple, and is worth going into at
some length.
In Rana temporaria {fusca) the lens is dependent for its develop-
ment on contact with the eye-cup. If the latter is removed (at
the tail-bud stage), the lens is not formed. ^ Further, the eye-cup in
this species is capable of inducing the formation of a lens out of
^ Spemann, 1901A, 1905.
184 organisers: inducers of differentiation
epidermal tissue which would normally not have given rise to a
lens at all. This can be effected either by grafting the eye-cup under
y f
h
Fig. 88
Self-differentiation of the lens in Rana esculent a. a, Extirpation of the pre-
sumptive eye-rudiment at the early neurula stage, b. Transverse section through
resulting larva 14 days after operation; in spite of the absence of an eye-cup, a
lens (L.) has developed by self-differentiation, c, This lens, at the same scale as
d, normal eye and lens, for comparison. (From Mangold, Ergehn. der Biol, vii,
1 93 1, after Spemann.)
organisers: inducers of differentiation
8s
the skin in an abnormal position, or by grafting a piece of foreign
epidermis over the eye-cup in situ. In both cases, a lens is formed.^
L.n.
Fig. 89
Development of the lens in Bomhinator pachypus. a, Transverse section through
larva from which the presumptive eye-rudiment was removed at the early
neurula stage: result, no lens, b. Transverse section through larva from which
the optic vesicle was removed at the early tail-bud stage : result, small lentoid
thickening of epidermis (L.f.). L.n. normal lens of unoperated side. (From
Mangold, Ergehn. der Biol, vii, 193 1, after Spemann.) c, Enlarged view of
lentoid (also seen at Fr.L.) developed after eye-cup removal and rearing at 23° C.
N. nose; H. brain. (From von Ubisch, Zeitschr. Wiss. Zool. cxxiii, 1924.)
^ Filatow, 1924, 1926.
i86 organisers: inducers of differentiation
Bufo, Triton, and the chicks agree with Rana temporaria in the
conditions of formation of the lens (see fig. 21, p. 55). In the
chick, the interesting observation has been made that the optic
vesicle, as well as the optic cup, is capable of inducing lens-forma-
tion.^ This means that the degree of histological differentiation of
the eye is immaterial for the inductive effect.
In Rana esculenta, however, removal of the eye-rudiment, even
at the early neural fold stage, does not prevent the formation of a
lens, which latter structure is therefore self-differentiating in this
species at a stage even earlier than that at which it is dependent-
differentiating m Rana temporaria.^ The lens, however, is sometimes
subnormal in size. Bomhinator pachypus is intermediate between
the two species of Rana in this respect, for after removal of the
eye-cup a small lens-like structure develops. This occasionally
happens in Rana temporaria^ (figs. 88, 89).
Although this experiment shows that the lens of Rana esculenta
is self-differentiating, it gives no information concerning the power
of the eye-cup of this species to induce the formation of a lens by
dependent differentiation. This can be tested by grafting foreign
epidermis of the same species, from various regions of the body,
over the eye-cup. The resuhs obtained differ according to the age
of the epidermis used. At the late tail-bud stage in Rana temporaria
and in Hyla arborea, epidermis from any region is capable of forming
a lens when in contact with an eye-cup, while in Bomhinator, lens-
forming potencies are restricted to the epidermis of the head. In
Rana esculenta at the late tail-bud stage, no epidermis other than
that of the presumptive lens region can be made to form a lens,^
though at the early tail-bud stage, epidermis from any other regions
can do so.^
The inducing power of the eye-cup of Rana esculenta may be
further tested by grafting over it some epidermis from another
species, in which the lens is normally dependent in its differentia-
tion, such as Bufo vulgaris, and such experiments invariably result
1 Danchakoff, 1924. - Hoadley, 1926 b.
"^ It appears that the lens in Rana esculeiita is not invariably self-differentiating,
especially at low temperatures. Further experiments on the modifiability of lens
induction and lens differentiation in different species are much to be desired. See
von Ubisch, 1924.
4 von Ubisch, 1927. ^ Spemann, 1912B. ^ von Ubisch, 1927.
organisers: inducers of differentiation 187
in the induction of a lens.^ Thus at this stage, the lens of Rana
esciilenta is self-differentiating, but the eye-cup also possesses the
inductive power of forming a lens, so that there is here another
example of the principle of "double assurance."
This state of affairs can be interpreted as follows. In Rana
esculenta, the lens is already determined irrevocably at a stage (early
tail-bud) when in Rana tenipovaria it is usually still plastic. We
may conjecture, therefore, that the determination of the lens occurs
precociously in Rana esciilenta. In this form, the lens is presumably
determined by the presumptive eye-rudiment while this is still
an invisibly determined region of the neural plate.
We may also assume, however, that, just as with the presumptive
neural plate before gastrulation, there has been a preliminary labile
determination of the lens, so that the lens-forming potencies will
be more easily called forth at a certain spot, viz. the presumptive
lens region." When definitive determination occurs, we must as-
sume that some influence, presumably of a chemical nature, diffuses
from the eye-area, and affects the region of optimum lens-forming
potency. In a similar way (as will be seen in Chap, vii, p. 223) we
may note the limb is actually formed at a region of maximum limb-
forming potency, in a much more extensive potential limb-area.
However, when the eye-rudiment of Rana esciilenta has become
converted into an optic cup, it still retains its lens-inducing power.
Indeed, it would seem that in some forms this power is retained
throughout life, for in many Urodeles it has been shown that the
adult eye can regenerate a new lens from its own margin if the lens
has been removed^ (see p. 237). It is interesting in this connexion
to note that the eye can resort to this method of lens-formation in
embryonic development and form a lens from its own margin if it
is deprived of contact with epidermis"^ (fig. 90).
The apparent "double assurance" found in Rana esciilenta thus
apparently means {a) that there exists a region of optimum lens-
forming potency in the epidermis of the neurula, and {b) that the
power of the eye-cup to induce a lens persists after the lens has
differentiated, and after the remaining epidermis has been deter-
mined to form epidermis and has ceased to be capable of responding
^ Filatow, 1925. - Spemann, 1912 b, ^ Colucci, 1891 ; G.Wolff, 1895.
* Spemann, 1905; Beckwith, 1927; Adelmann, 1928.
i88
organisers: inducers of differentiation
to induction. There is thus an overlap in time between these two
phases of differentiation ordinarily spoken of as dependent differ-
entiation and self-differentiation. There seems to be little reason
to doubt, however, that both the methods concerned in '* double
assurance " are ultimately referable to one and the same causative
agent: in this case presumably situated in the rudiment of the
eye-cup and in the fully formed cup which later arises from it.
Fig. 90
Lens-formation from the margin of the optic cup in ontogenetic development.
The presumptive eye-rudiment of an embryo of Triton was grafted into the side
of the body of another embryo, and developed by self-differentiation, deep
beneath the epidermis. Under these circumstances it has given rise to a lens from
the margin of its own cup, in themanner characteristic of regeneration experi-
ments. Br. portion of grafted brain tissue ; I.e. wall of intestine ; L. lens ; S. epi-
dermis of ventral side. (From Adelmann, Arch. Entwmech. cxiii, 1928.)
The divergent results obtained with different species are ap-
parently to be accounted for by differences in the rates at which
the two processes, of capacity of the eye-cup to induce and of the
epidermis to differentiate, occur.
The proliferation of cells from the epidermis is not, however,
the only process involved in lens-formation: the cells require to
become converted into the characteristic lens-fibres. While the
organisers: inducers of differentiation 189
proliferative effect may, as we have seen above, be more or less
independent, the subsequent differentiation of lens-fibres appears
to be always dependent, usually on the eye-cup. But the action of
the latter in this case does not appear to be specific, for experiments
in which lens-rudiments are allowed to develop in proximity with
portions of brain or nose tissue show that the latter are also capable
of inducing the formation of lens fibres.^
Recent work on the American bull-frog, Rana cateshiana^ has
given additional results. This species shows an extreme of depen-
dent differentiation for the lens, rivalling or exceeding Rana tem-
poraria in this respect. Of greatest interest is the fact that here
the continued presence of the optic vesicle or eye is necessary for
the lens to achieve full differentiation and full size, even after it
has been initially determined. The lens-rudiment, however, once
determined, has a certain power of self-differentiation. After
determination but before visible differentiation, the lens-rudiment,
by itself, is only capable of producing lentoid structures without
differentiation of fibres. After visible thickening has occurred,
however, the rudiment left in situ after removal of the underlying
eye-cup will produce a true lens, but this is small and slightly
abnormal. There is thus a complementary action of inherent
potencies and external induction (see p. 264 for what may be a
similar effect with the avian gonad). Another interesting point is
that if the visible lens-rudiment at the same stage is separated from
the eye-cup and grafted heterotopically, it undergoes a certain
amount of regression and never reaches the same degree of dif-
ferentiation as if left in situ, though in both cases it is removed from
the inductive influence of the eye. Thus in this species, although
epidermis from any region can be made to form a lens, potencies
favourable to lens-differentiation are highest in the area of the
normal lens-field.
The crystalline fibres of the lens in Amphibia are oriented in a
definite manner, normally converging to a sutural line which is
dorso-ventral on the outer surface, and antero-posterior on the
inner surface of the lens. It is to be noted that the plane of the ex-
ternal sutural line coincides with that plane of the eye-cup in which
the choroid fissure is situated, for this structure occupies the most
^ Balinsky, 1930. ^ Pasquini, 1933.
iQo organisers: inducers of differentiation
ventral region of the cup. It has been found that the orientation
of the fibres of the lens is also dependent on the eye-cup^, and, in
particular, on the position of the choroid fissure. This is proved by
experiments on embryos of Rana escidenta at the early neurula stage,
in which the presumptive lens epidermis is rotated through 90° : the
lens-fibres are nevertheless normally oriented. On the other hand,
if the eye-rudiment is rotated so that the choroid fissure comes
to occupy an abnormal position, the lens-fibres are also abnormally
oriented.'^ Therefore, while the lens is normally self-differentiating
as regards its general formation in Rana escidenta at this stage,
the determination of the orientation of its fibres is still dependent
on the eye-cup. At later stages, rotation of the lens-area of epidermis
shows that this orientation becomes self-differentiating also.
Lastly, it may be noticed from those cases in which a lens can be
induced by an eye-cup of a different species, that the lens-inducing
capacity of the eye-cup, like the organising capacity of the am-
phibian dorsal lip, the avian primitive streak, the hypostome of
Hydra, the head of Planaria, and the capsule-inducing capacity of
the amphibian ear-vesicle, is not species-specific.^ In this respect,
the action of organisers and inducing structures has much in
common with that of hormones. Many if not all hormones are
similar or identical in widely separated groups: thyroxin from a
mammal will metamorphose amphibian larvae; testis hormones
from a bull will cause the comb of capons to grow ; adrenalin from
a fish will excite vaso-constrictor effects in man. On the other hand,
the precise effect produced depends on the reacting tissues, just as
it does with the organiser effects during development. The tail and
limbs of Anuran larvae react to thyroxin, while those of Urodela do
not; the larval epidermis of most Amphibia reacts to thyroxin, while
that of the adult never does. The relation of organisers to induced
organs, as of hormones to reacting tissues, is thus much less specific
than the interaction of hereditary outfits (genomes) from different
species, where a difference of generic degree is usually more than
sufficient to prevent co-operation.
^ Dragomirow, 1930. " Woerdeman, 1932.
^ Woerdeman (1933 b) finds marked changes in glycogen content in the
eye-rudiment before and during the period of lens-differentiation. The precise
meaning of these, as of similar changes in the organiser region (p. 154), remains a
subject for future investigation.
organisers: inducers of differentiation 191
§9
In reviewing the various aspects of dependent differentiation it
is clear that the organiser phenomena occupy a special place. The
part which organisers play is of supreme importance. From the
theoretical point of view, they present a biological property of the
first order, and had Roux known of their existence he w^ould un-
doubtedly have classified them among the ''complex components"
of development (see p. 9).
However, the precise mode of action of organisers cannot be
understood except in relation to the properties of gradient-fields :
this problem will be considered in some detail in Chaps, viii and ix.
Meanwhile, attention may be turned to the general result of the
presence in an embryo, such as an amphibian, of an organiser and
other structures, exerting effects of hetero- and homoio-genetic
induction, and some of them showing local regional differentiation.
The main result is that almost everywhere in the body formative
stimuli are found capable of inducing plastic tissues to undergo this
or that type of differentiation, according to their position. Nor-
mally, of course, the tissues cease to be plastic as soon as they have
undergone the inductive action of their organiser. But the exis-
tence, distribution, and local regional characters of the various in-
ductive influences in the amphibian embryo can be studied by
grafting portions of plastic early gastrula tissue into older hosts (at
the neurula stage), thanks to the fact that the inductive effects
persist for a longer time than is necessary for the normal determina-
tion of the embryo's own tissues.
It has been found, using pieces of presumptive epidermis or
neural fold tissue as grafts, implanted into the dorso-lateral region
of neurulae, that the quality of the differentiations which the grafts
then undergo is dependent on their position in the host embryo. In
the head, grafts may differentiate into portions of brain wdth epi-
physis, nasal sacs, and eyes: in the gill region, into portions of
hind-brain : in the trunk region, into portions of spinal cord. The
grafts may be induced to form ear-vesicles, sense-organs, visceral
cartilages, and ganglion cells in the head; gills in the gill region;
fore-limbs in the fore-limb region ; pronephric tubules in the prone-
phric region, etc.^ (fig. 91).
^ Holtfreter, 1933 b.
192
organisers: inducers of differentiation
A fact of great importance is that the various determinations are
regional, but the regions are not very closely circumscribed. There
is as it were a certain amount of latitude as to exactly where a
particular structure will arise, although it is bound to be within
V.:.>^' ^mJ"
^^S^Z^ZSrrrsi--^ -'
"--::;,
Nose, eye, fore- and mid-brain
Balancer
Ear, hind-brain
Frontal epidermis
Head neural crest
Gills
Fore-limb
Pronephros
Dorsal crest
>
— .
<->
Myotomes
Spinal cord, tail
Epidermis, connective tissue
>•
Fig. 91
Diagram showing the results of grafting portions of presumptive epidermis or
presumptive neural plate from a gastrula into various regions of the flank of a
neurula (Triton). The grafts are induced to undergo differentiation into the
structures enumerated in the left-hand column ; the effect of the position of the
graft in the host upon the type of structure resulting is indicated by the extent
of the lines in the right-hand column, imagined as projected on to the larva shown
above. (From Holtfreter, Arch. Entwrtiech. cxxvii, 1933.)
a certain region. The following table will show the frequency with
which particular structures are induced from ectodermal grafts in
different parts of the body :
Ear ,
region
Gill
region
Fore-limb
region
Phrone-
phros
region
Balancer
Ear-vesicle
Pronephros
10
72
6
15
2
^
28
organisers: inducers of differentiation 193
Each of the various inductive effects accordingly covers a wide
area, or field, and the intensity of the induction decreases with
increasing distance from a sub-central point in each field (see
Chap, vii, p. 223).
If, now, w^e stop to inquire which structures are responsible for
the inductive effects, the answer appears to be in most cases that
each field is dependent, not on one but on several other structures.
The organiser for the neural tube induction in the trunk region
appears to be the segmented mesoderm; this, which of course is
derived from the invaginated organiser, is known to induce neural
tube when grafted beneath strange epidermis. Here, the converse
experiment has been performed, and strange epidermis has been
grafted over the derivative of the organiser. The inductive action
which produces portions of brain, etc., in the head, appears to
proceed from the neural crest, which is also capable of inducing
cartilage, ganglion-cells, sense-organs, and ear-vesicles.^ At the
same time, the induction of ear-vesicles can be performed homoio-
genetically, by ear-vesicles, just as fore-limb and pronephros can
induce fore-limb and pronephros respectively.
The formation of a tail is the combined and coordinated
result of a number of inductive influences. The elongation and
stretching of the notochord and musculature, and the metameric
arrangement of the latter, are dependent on the presence of
neural crest mesenchyme; dorsal and ventral fins are formed
when neural tube is present ; in the absence of neural crest
mesenchyme, the initial elongation of the tail-bud stops, and
regression sets in.
We see, in general, that as a result of the inductive capacities of
the organiser and of certain other structures (themselves the result
of induction by the organiser), the amphibian embryo at the neurula
stage is already what may figuratively be called a physiological
mosaic of formative stimuli, leading to the demarcation of fields,
each of which represents the sphere of action of a particular type of
inductive effect. We shall see in the next chapter that these fields
constitute one of the most important features of the next or mosaic
stage of development.
^ In these experiments the grafts do not appear to have come into close contact
with the brain itself of the host-embryo. The homoiogenetic inductive capacity
of the brain has, however, been established by other work (see above, p. 147).
HEE 13
Chapter VII
THE MOSAIC STAGE OF DIFFERENTIATION
§1
It has been seen in previous chapters that, after a certain stage, the
various regions of the early embryo are irrevocably determined to
^^
'C
-^
^^^^1^^
B
Fig. 92
Mosaic development in explanted tissues. A, Isolated head, containing brain,
eyes, nasal pits, tip of notochord, cartilage, and functional jaw-muscles,
differentiated in inorganic culture medium from the dorso-anterior portion of
an early neurula of Ambly stoma: 25 days after explantation. B, Isolated trunk,
containing spinal cord, notochord, muscles, pronephric tubules, gut, fore-limbs,
and tail with dorsal fin, differentiated in inorganic culture medium from the
postero- ventral portion of an early neurula of Triton : 25 days after explantation.
(From Holtfreter, Arch. Entwmech. cxxiv, 193 1.)
undergo some particular type of development, although at the time
that they are thus determined there is no visible differentiation
of any kind. This determination is presumably to be ascribed to
the local elaboration of specific chemical substances, and may be
THE MOSAIC STAGE OF DIFFERENTIATION 195
referred to as chemo-differentiation (see p. 46). ^ It now becomes
necessary to consider this phase of development in greater detail.
Experiments and operations on early stages (early tail-bud) of
embryos of Urodeles (Amblystoma) have now shown that, beyond
mechanical wound healing, no regeneration or regulation occurs."^
If, for instance, the embryo is cut into two by a transverse section,
the two portions continue their prospective development, the front
portion forming a head and neck region, the posterior portion a
trunk and tail. The number of external gills on the one or the other
portion depends upon the precise position of the cut (fig. 92).
Similarly, it has been found that the anterior third of a 24-hour
blastoderm of a chick embryo grafted on to the chorio-allantois of
another egg gives rise to just those organs which it would have
produced in normal development^ (fig. 93). Two half-embryos
of frogs grafted together will develop into a single frog, even if the
halves belong to diiTerent species. Each half retains its specific
characteristics (see p. 406, and fig. 196).
In Amphibians, such fragments of course cannot develop far
beyond hatching. If, however, in the early tail-bud stage, the tip
of the tail is cut oflF, the organism develops into a healthy larva, but
with a permanently shortened tail.^ Removal of rudiments of eyes,
gills, limbs, heads, or snouts at this stage results in permanent
absence of these structures in the later embryo and larva. Similarly,
experiments on the fish Fundulus have shown that removal of
portions of the embryonic shield results in permanent absence of
the structures whose rudiments have thus been affected^ (fig. 94).
More recent and detailed work has shown that in Triton at the
stage when the tail-bud is hemispherical, complete or almost com-
plete removal of the mesodermal contents of the bud results in
completely tailless larvae, whereas in only slightly later stages
complete regeneration can and does occur.^ Further, by appro-
priate operations, more localised defects can be obtained, e.g.
absence of ventral fin membrane, of mesodermal somites, of noto-
chord, or of nerve tube. When regeneration experiments are
carried out on such partially defective tails in the larval stage, it is
^ Huxley, 1924; Goldschmidt, 1927; Bertalanffy, 1928.
2 Schaxel, 1922 b. ^ Murray and Huxley, 1925.
* Eycleshymer, 1914; Nicholas, 1927; Hoadley, 1928. ^ Vogt, 1931.
13-2
196
rhomb
iF.F.
opt. st. opt. c
Fig. 93
Self-differentiation in fragments of the vertebrate embryo. The head-region of
a 24-hour chick embryo was grafted on to the chorio-allantois of another egg and
allowed to develop for 4 days. Above, eye-region enlarged. Below, longitudinal
section. It differentiated the main regions of the brain {rhomb . rhombencephalon ;
Twj;^/. myelencephalon ; m.6.mid-brain;/.6. fore-brain; e^. epiphysis ; inf.f.inixxn-
dibulum), together with an eye showing optic stalk (opt.st.), optic cup (opt.c), and
lens, /; bl.v. blood-vessel. The histogenesis was normal, but the form-differen-
tiation, notably of the eye, and also of the brain (note b. bar across cavity of fore-
brain), abnormal in many respects. (From Murray and Huxley, Brit.Journ. Exp.
Biol. Ill, 1925.)
THE MOSAIC STAGE OF DIFFERENTIATION
197
found that, although they have the capacity for regeneration, the
regenerate still shows the defective organisation (e.g. with regard
to the ventral fin membrane). It appears in these cases that the
organism cannot regenerate a structure which has never been
formed in its own ontogeny : a fact of great interest in itself and
.>■'■- \^ ' ', '
.^-
^^.
Fig. 94
Effect of extirpation of tail-rudiment at early stage, in Triton ; left, the operation ;
right, the resulting larva, with total absence of tail. (From Schaxel, Arch.
Entwmech. l, 1922.)
with an important bearing on the problem of gradient-fields, to be
discussed in Chap. x. It is still, however, uncertain whether this
limitation is universal. ^
1 Cases are known in which an abnormal limb (which owes its abnormality to
the fact that its rudiment was grafted at an early embryonic stage and failed to
develop normally) can after amputation regenerate a normal limb (Swett, 1924).
Clearly, the conditions here are different from those in which a structure is ab-
normal, imperfect, or absent as a result of retnoval of its rudiment. The abnor-
mality of the grafted limb is a consequence of some local conditions due to the
experiment, and does not reflect any intrinsic restrictions of potency in the limb-
rudiment. Consequently, when a new set of conditions supervenes as a result
of amputation, these potencies are present and able to control the regeneration
of a normal limb.
198 THE MOSAIC STAGE OF DIFFERENTIATION
A similar total and permanent absence of a whole organ has been
obtained by extirpation of the presumptive limb-area, both in
Urodela^ and in the chick. ^ In these cases, the presumptive limb-
area is a discoid region of mesoderm and ectoderm, with no visible
differentiations. As will be seen later (p. 420), in adult Urodela,
regeneration of a limb will occur even when the whole limb and its
skeleton, including the girdle, is extirpated, provided that the
sympathetic nervous system is left intact.^ No experiments seem to
have been carried out to discover whether any regeneration would
occur in an animal lacking a limb owing to early embryonic
extirpation of the limb-area, if the region on the flank where the
limb ought to be were removed at the adult stage ; we may presume,
however, that there would be no regeneration. (See fig. 22, p. 56.)
The hypophysis arises from a rudiment of ectoderm on the front
of the head. This rudiment can be extirpated from Anuran larvae
at the tail-bud stage, and it is found that the larvae^ which ulti-
mately develop are normal except that they lack the pituitary gland. ^
Even the blood in the Anuran embryo has a definite and localised
rudiment, situated in the mesoderm of the splanchnopleur, in the
mid-ventral line, anterior to the heart. If this rudiment is extir-
pated completely from embryos of Rana temporaria at the early
tail-bud stage, no erythrocytes are formed, and in cases of partial
extirpation the quantity of erythrocytes produced is proportional
to the amount of the rudiment which is left.^
The Ascidians provide another case of animals which in the adult
state are capable of extensive and far-reaching regeneration and
reorganisation, but which in the early stages of embryonic de-
velopment are unable to make good any loss which the various
determined regions may sustain.'^
^ Harrison, 1915. ^ Spurling, 1923.
^ Bischler, 1926.
^ Incidentally, it may be mentioned that such larvae are of great interest also
from another point of view, for they are incapable of producing the pituitary
hormones, and are therefore permanently light in colour, and incapable of normal
metamorphosis .
^ Smith, 1920. ^ Frederici, 1926.
' Conklin, 1905, 1906; Huxley, 1926.
THE MOSAIC STAGE OF DIFFERENTIATION 199
§2
This mosaic predetermination of various regions in the chemo-
differentiated stage in development is also demonstrated by
numerous experiments in which a region continues its presumptive
development even after grafting into an abnormal position. As
previously mentioned, the presumptive eye-region of amphibian
embryos has the power of self-differentiation at the early neurula
stage (p. 46). Limb-discs of Amhlystoma grafted on to the flank
or into other abnormal situations will still continue to form limbs. ^
The presumptive ear- region will differentiate semi-circular
canals, etc., when grafted into abnormal situations.'- The mosaic
nature of this power of self-differentiation is further shown by the
fact that if a neurula of Rana esculenta is divided transversely by a
cut passing through the presumptive ear-region, it is found that
both halves develop auditory vesicles, but they are incomplete, the
details varying with the precise position of the cut. On each side
of the body, there is only one ductus endolymphaticus developed
on each side, and this may be either in the anterior or the posterior
half.=^
Other experiments of grafting and extirpation have shown that
the gill-region, the balancer, nerve placodes, portions of the neural
crest, and various other amphibian organ-rudiments possess this
capacity for self-differentiation (for references, see later sections).
The outgrowth of the glomerulus from the aorta has been shown
to be due to self-differentiation, independent of the presence or
absence of the pronephros with which it normally comes into
functional relation.*
In the chick, grafting of embryonic rudiments on to the chorio-
allantois of another embryo has been the main method employed,
e.g. with the ear-region, eye-region, complete limb-rudiments,
fractions of limb-rudiments, presumptive thyroid- region,^ meso-
nephros (see also below), metanephros, adrenal,^ spleen, portions
of brain and spinal cord,^ lung^, etc. (figs. 95, 96).
^ Harrison, 19 18; Detwiler, 191 8.
^ Streeter, 1906, 1907; Sternberg, 1924.
^ Spemann, 1910. ^ Howland, 1916.
^ Rudnick, 1932. ^ Willier, 1930.
' Rienhoff, 1922; Danchakoff, 1924; Hoadley, 1924, 1925, 1926 a, 1929:
Murray and Huxley, 1925. ^ Rudnick, 1933.
200
THE MOSAIC STAGE OF DIFFERENTIATION
i t-l
B
Fig. 95
Self-differentiation of grafted chick metanephros. A, Metanephric rudiment at
the time of grafting (5-day chick). B, Differentiation after 5 days on the chorio-
allantois of another egg. (From Danchakoff, Zeitschr.f. Anat. u. Entzogesch. Lxxiv,
1924; B, after Atterbury.)
THE MOSAIC STAGE OF DIFFERENTIATION 201
In fish, optic cups and other organ-rudiments grafted into the
yolk-sac of other embryos show self-differentiation.^
The capacity for self-differentiation in mammalian embryos has
been tested in rabbits by grafting portions of the embr>'onic area
on to the omentum of other rabbits, where they show a degree of
differentiation comparable to that of normal embryos of the same
age."
The self-differentiating capacity of mammalian tissues has also
been tested by grafting thirds of ii-day rat embryos on to the
Fig. 96
^Mosaic development and self-differentiation of the eye of the chick, grafted on to
the chorio-allantoic membrane. The rudiment was removed from an embryo
incubated for 48 hours, and grafted for 7 days. (From Hoadley, Biol. Bull.
XLVi, 1924.)
chorio-allantoic membrane of the chick, where, in spite of the wide
taxonomic difference between donor and host (involving as it does,
among others, the difference between the temperatures of normal
uterine and incubatory development), they are able to differentiate.
In these conditions, different structures vary greatly as regards
their capacity for self-differentiation ; endoderm and nervous tissue
show hardly any differentiation, but epidermis with its included
hair-follicles, cartilage, and bone, possess it to a high degree, and
reach a stage comparable to that of the corresponding structures
1 Mangold, 1931B. - Waterman, 1932.
202
THE MOSAIC STAGE OF DIFFERENTIATION
40'
^^m?^'""^
-$e-«r-* ■'"''«»'
e^f:
"-^|-_i^
C.J
^>-^
^
^-^
Fig. 97
Mosaic development and self-differentiation of isolated regions of the neurula
(Urodele) explanted in inorganic culture medium. A, Epidermal vesicle, g days
after explantation. B, Section through A; note ridges and folds (cf. fig. 13).
C, Notochords, 11 days after explantation. D, Endodermal vesicle, representing
an everted gut with the endothelial cells directed outwards. E, Section through
D showing the endodermal cells secreting outwards into the medium, 13 days
after explantation. (From Holtfreter, Arch. Entwmech. cxxiv, 193 1.)
THE MOSAIC STAGE OF DIFFERENTIATION
203
in a normal rat of similar age.^ Nasal sacs and mesonephros
achieve a less but considerable degree of differentiation.
Explantation methods have also been applied to the problem.
Presumptive rudiments of organs, as yet without any visible differ-
entiation, are removed from the body and allowed to develop in
culture media. In some cases they are enclosed within jackets of
epidermis, but this is not an essential condition. In addition to
Fig. 98
Self-differentiation of median heart-
rudiments in vitro. The heart-rudi-
ments together with some of the
neighbouring ento-mesoderm were
removed from early tail-bud stages
of Bornhinator , and cultivated as ex-
plants in epidermal jackets. Above,
part of a micro-cinema film of an
explanted heart, 11 days after opera-
tion; a. systole; ^.diastole. Below,
longitudinal section through a similar
explant showing differentiation into
a auricle; v, ventricle. In addition,
/, liver; d, yolk-sac; £), gut, also pre-
sent in explant. (From Stohr, Arch.
Mikr. Anat. u. Entwmech. cii, 1924.)
notochord, neural tube, muscle-segments, epidermis and kidney
tubes, 2 auditory vesicles, gut, liver, and pancreas-rudiments of
amphibian embryos treated in this manner develop for considerable
periods of time, and produce their appropriate structures, including
functional ciliated epithelium and secretory tissue with actual secre-
tion : portions of gut thus differentiated may even show peristaltic
action.^
Paired heart-rudiments of Urodele embryos at the neurula stage,
Hiraiwa, 1927; Nicholas and Rudnick, 193 1,
2 Erdmann, 1931.
Holtfreter, 1931A, b.
204
THE MOSAIC STAGE OF DIFFERENTIATION
before they have united in the middle line, will, when explanted
singly, form vesicles of heart-tissue, and some (those from the left
side, see p. 77) may show pulsations.^ Heart-rudiments taken at
later stages, when they have united in the mid-ventral line, give
still more elaborate self-differentiation, showing sinus, auricle,
ventricle, and bulbus^ (fig. 98).
be
/I
rc.l on I
nl
■,.,-1
ami.
A
B
Fig. 99
Mosaic development and self-differentiation of the chick eye-rudiment cultivated
in vitro. The eye-cup and lens were removed from an embryo incubated for
66 hours and cultured for 8 days in plasma with embryo extract. Histological
differentiation has proceeded at almost the normal rate, in spite of the fact that
the morphological differentiation of the structures is highly abnormal and that
they are subnormal in size. Histological differentiation is therefore independent
of morphological differentiation, and of the normal rate of cell multiplication.
A, Sectionof the whole explant. B,Section through the retina of a 17-day explant,
in which all the layers are normally developed, a.c. amacrine cells; b.c. bipolar
cells; ect. ectoderm; e.l.m. external limiting membrane; g.c. ganglion cell;
i.l.77i. internal limiting membrane; im.l. inner molecular layer; i.n.l. inner
nuclear layer; /. lens; M.f. Muller's fibres; o.ni.l. outer molecular layer; o.n.l.
outer nuclear layer ; p.cil. pars ciliaris. retinae •,p.l. pigment layer ip.op. pars optica
retinae. (From Strangeways and Fell, Proc. Roy. Sac. B, c, 1926.)
In vitro cultivation of rudiments of presumptive regions has also
been practised with chick material. The optic cup (fig. 99), por-
^ Goerttler, 1928.
^ Stohr, 1924.
THE MOSAIC STAGE OF DIFFERENTIATION 205
tionsof limb,^ ear,^ metanephros^ and other rudiments^ thus treated
have shown successful histological self-differentiation. Interesting
examples of chemical self-differentiation are found in isolated
portions of the skeleton. The cartilages of the palato-quadrate and
of the femur normally undergo ossification, whereas the distal
portion of Meckel's cartilage does not. The future histological
structure is already determined by the sixth day of incubation, al-
though there is then no visible distinction. The difference between
these two types of cartilage is revealed by cultivation in vitro, where
rudiments of the palato-quadrate and of the femur show a marked
synthesis of phosphatase, while that of the distal portion of Meckel's
cartilage does not : phosphatase activity is correlated with ossifica-
tion.^ If cultured long enough, ossification of a normal type super-
venes in the rudiments. It is worth mentioning that even in the
abnormal conditions provided by tissue-culture, in which the organs
are without blood supply, the volume of a chick femur will increase
up to about thirty times.
Even extra-embryonic regions, such as the presumptive blood-
islands, develop histologically differentiated blood when cultivated
in vitro, ^
In some cases, at least, the determination imposed upon regions
in the mosaic stage of development concerns even the duration of
progressive differentiation and growth. The mesonephros of the
chick embryo normally undergoes regression at about the tenth day
of incubation, and if its rudiment is grafted on to the chorio-
allantoic membrane of another tgg, it will first differentiate the
typical mesonephric tissue, and then proceed to regress at about
the same time as regression would normally have occurred if it had
been left in place in the embryo." The time of regression in these
cases is, of course, in no way determined by the age of the host-egg
on to the chorio-allantois of which it is grafted (fig. lOo).
The specific growth-capacities of the rudiments may also be
determined. In the intact bird, the right ovary is rudimentary and
the left is well developed. Four-day rudiments of the ovaries
^ Strangeways and Fell, 1926. " Fell, 1928.
3 Rienhoff, 1922. * Hoadley, 1924.
'" Fell and Robison, 1929, 1930. ^ Murray, 1932.
' Danchakoff, 1924.
206 THE MOSAIC STAGE OF DIFFERENTIATION
grafted on to the chorio-allantoic membrane show the same specific
differences between the growth- capacities of the right and left
sides. ^ In the mammal, also, the tissue culture of embryonic
material provides evidence of self-differentiation. ^ Portions of
rabbit embryos 9 to 12 days old, cultured in vitro, reveal the
mosaic character of development : the various rudiments differen-
tiate independently,^ just as similar fragments do when grafted.
The growth-partition coefficients of Urodele limbs are inherently
determined.*
A striking case of independent differentiation is provided by the
silkworm. The wing rudiments in Lepidoptera are protruded from
the surface during pupation, and the pupal case has pockets
into which the wings fit snugly. In the silkworm, a mutation has
been found which results in the animal being wingless. Neverthe-
less, the pupal cases of such mutants possess the characteristic
pockets, although no wings project into them.^ Similar occur-
rences have been observed in Papilio dardanus where in the female
the wings may have no tails, but pockets for them are provided
in the pupal wing-cases.^
This case is of considerable theoretical interest, for, in general,
when two structures are closely associated topographically, it is
found that the differentiation of the one is often dependent on the
other. Numerous examples have been given in Chap, vi : we may
recall the eye-cup and the lens of Rana fiisca ; the eye and the con-
junctiva ; the tympanic cartilage and the tympanic membrane ; the
skeleton and the arms of the pluteus ; the hydrocoel and the am-
niotic cavity of the echinoid rudiment. In Rana esculenta, however,
as we have seen (p. 186), the eye-cup and the lens are independent
from a very early stage, and in this they resemble the wing and
wing-case of the silkworm.
Further evidence of the self- differentiating capacity of the wing-
rudiment in Lepidoptera is provided by the experiments of grafting
the wing-rudiments of caterpillars from one sex into the other. The
fully developed wing is markedly different in the two sexes, and
it is found that regardless of the sex of the host into which it has
^ Willier, 1927. ^ Waddington and Waterman, 1933.
^ Maximow, 1925. ^ See Huxley, 1932, Chap. vi.
^ Goldschmidt, 1927, p. 203. ^ Lamborn, 1914.
207
B
Fig. loo
Self-determination of degenerative development in the chick mesonephros.
A, After 5 days as a chorio-allantoic graft, the mesonephros-rudiment shows
marked progressive differentiation. The figure is of a grafted short section of the
trunk; similar differentiation is obtained with isolated mesonephros -rudiments.
B, After 7 or more days, the graft shows regression. All secreting tubules have
disappeared. The malpighian capsules persist, as in normal development. (From
Dstnchakofi, Zeitschr.f. Anat. u. Entwgesch. LXXiv, 1924.)
2o8 THE MOSAIC STAGE OF DIFFERENTIATION
been grafted, the wing-rudiment differentiates according to the sex
of its donor. ^ Recent work on various Insects indicates that after
a certain stage the embryo is a mosaic of chemo-differentiated re-
gions, although the details of the determination-process differ
considerably from those found in Amphibia.^
In Cephalopods it has been shown that fragments of embryos
cultivated by explantation methods continue their differentiation
as if they formed part of the whole organism.^ These experiments
were undertaken after visible differentiation had appeared ; others,
however, indicate that the embryo passes into the mosaic chemo-
differentiated stage just before visible differentiation occurs.* This
would, in general, be similar to the state of affairs in Amphibia.
Another remarkable case of self-differentiation during the mosaic
stage of development concerns the self-orientating properties of
the auditory vesicle in Amphibia. If at the stage when it is a simple
vesicle, the auditory sac is turned upside down and left in situ, it
often rights itself by rotation, so that its dorso-ventral axis con-
forms to that of the whole animal.^ The suggestion that the ear-
vesicle rights itself because it only fits properly into the neigh-
bouring structures when it is in its normal position must be dis-
carded, because a right ear-vesicle, grafted upside down in the space
vacated by an extirpated left vesicle, rotates and becomes right way
up and right way out, but as the vesicle retains its laterality, it
develops with its normally anterior side pointing backwards in the
animal. It thus rights itself in respect of its dorso-ventral axis in
spite of the evident misfit which results. Further, an inverted
vesicle of Rana will right itself in Amblystoma, and vice versa.^
The rotation of the ear-vesicle may be impeded by special local
conditions of the experiment, but when it occurs it takes place
gradually, and, to all appearances, in relation to gravitational
stimuli. The ear is, of course, an organ whose function it is to detect
the direction of maximum gravitational attraction, and, should the
supposition be verified that the righting effect is directed by gravi-
tation, the ear- vesicle in Amphibia may be regarded as determining
its orientation independently of the rest of the organism. Un-
^ Kopec, 1911, 1913. " Seidel, 1929, 1931; Reith, 1932; Pauli, 1927-
3 Ranzi, 193 1. * Ranzi, 1928.
^ Streeter, 1906, 1914; Spemann, 1910. *' Ogawa, 1921.
THE MOSAIC STAGE OF DIFFERENTIATION 209
fortunately, it has not yet been found possible to test the directional
eifects of gravity on the developing ear- vesicle by forcing the em-
bryo to adopt abnormal positions, for the embryo invariably rights
itself also, and explantation methods have not been applied to this
interesting problem. ^
§3
The principle of self-differentiation is further illustrated by ex-
periments of tissue- culture, from which it emerges clearly that the
cells of any particular tissue are permanently determined (except
in so far as metaplasia may occur: see below). Mesenchyme,
smooth muscle, heart-muscle, striped muscle, epithelium, endo-
thelium, kidney-epithelium, and blood-corpuscles of adult birds
and Mammals have been shown to preserve their specific character
in a wide range of media, and experiments have now been conducted
long enough to show that they can preserve them indefinitely.
Fibroblasts of the fowl have been cultured in vitro for over 20 years
(a much longer period than the maximum length of life of the fowl)
and show unchanged characters and an unchanged rate of growth.
In many cases, particular characteristics assumed by a cell are
a function of the environment or medium in which it finds itself.
Epidermis which, like that of the chorio-allantoic membrane of the
avian embryo, does not normally show keratinisation, may do so
as a reaction to grafts of tissue placed upon it.^ Under certain con-
ditions of the medium, an apparent loss of specific characters, or
dedifferentiation, may occur, and the tissue reverts to an undiffer-
entiated type. Such dedifferentiation is, however, a reversible
phenomenon. Cartilage-cells^ or kidney-epithelium* may undergo
dedifferentiation and grow as sheets of embryonic cells, but on
restoration of the original conditions, the cells readopt the differ-
entiated character typical of the tissue to which they belong. This
may take place in vitro, or after interplantation subcutaneously
under the wing of a young chick. Cartilage-cells, epithelial cells,
^ A further interesting fact is that in those cases in which the ear-vesicle has
been inverted and has failed to rectify its position completely, the resulting tad-
poles have an altered sense of balance, which they show by swimming in abnormal
attitudes and upside down (Spemann, 1906 a).
^ Huxley and Murray, 1924.
3 Strangeways, 1924. ^ A.H.Drew, 1923.
HEE 14
210 THE MOSAIC STAGE OF DIFFERENTIATION
and intestinal endothelial cells which had completely dediffer-
entiated in vitro were found to possess equally complete powers of
redifferentiation.
The various strains of cells differ not only in their structural
characters, but determined physiological differences may also be
observed between cells which are morphologically indistinguish-
able. Thus, strains of fibroblasts have been found differing from
one another in their nutritional requirements, and differing also
from epithelial cells and macrophages.^ The differences show them-
selves in the rate of proliferation of the cells in any given medium,
and by specific reactions, such as cytolysis,^ to certain induced
pathological conditions.
Tissue-culture methods have also thrown certain new light upon
the problem of differentiation. It has been found in the first place
that fibroblasts, isolated from different organs of the same embryo,
exhibit different growth-rates and other physiological characteristics
such as resistance to acidity and capacity to digest fibrin: these
differences appear to be persistent. For instance, fibroblasts
isolated from the skeletal muscle of a 17-day chick embryo have a
growth-rate nearly three times as high as that of fibroblasts from
the thyroid of the same embryo, and nearly ten times as high as those
from the heart. ^ A further and more surprising result is that com-
parable physiological differences exist between fibroblasts isolated
from the same organ of embryos of different age. For example,
fibroblasts from the skeletal muscle of the leg of the 17-day chick
embryo have a growth-rate about 60 per cent, higher than those
from the same tissue of 8-day embryos.^ These differences con-
tinue to be shown even when the strains have been subjected to
marked environmental changes, and are returned to standard
conditions.
There seems no escape from the conclusion that the primitive
mesenchymatous tissue, from which the fibroblasts of the body are
derived, receives some impress affecting its physiological charac-
teristics from the regions in which it happens to find itself, and this
impress changes with age. As regards the regions, the process is
doubtless an aspect of the self-differentiation which we have been
^ Carrel, 193 1. ^ Horning, 1932.
^ R. C. Parker, 1932 A. * R. C. Parker, 1932 b.
THE MOSAIC STAGE OF DIFFERENTIATION 211
considering : but the reactions of the fibroblasts to this process are
purely passive ; and if removed from the local influence, they simply
retain the characteristics impressed up to the time of isolation. This
we may regard as a new type of dependent differentiation : mesen-
chyme is predetermined to differentiate into fibroblast tissue, but
the detailed characteristics of the fibroblasts are impressed from
without.
It is of interest to note that many of the characteristics of tissues
are dependent on specific physiological characters of the cells them-
selves. In tissue-cultures, fibroblasts form an irregular matted
tissue; epithelial cells associate with one another in an orderly
manner; amoebocytes remain separate and never form a compact
tissue. These determined types of cell behaviour persist indefinitely
in vitro.
Cells which normally form part of a more highly differentiated
tissue possess and retain the type of behaviour which leads to the
formation of such tissue. Thus, kidney cells can redifferentiate into
kidney tubules,^ and capillary cells can redifferentiate into capillaries
in vitro ^^
Under certain circumstances, however, it appears to be possible
for cells to undergo a permanent and irreversible change in type
and characters, comparable in its way to the changes observed in
somatic mutations in vivo. This phenomenon, known as metaplasia^
has from time to time been claimed to occur in many cases of re-
generation, when it has been asserted that certain structures have
been formed from cells of a different tissue. It is, however, often
difficult if not impossible to be sure that undifferentiated and em-
bryonic cells were not present, and that the differentiation of the
structure in question did not proceed from them. This possibihty
seems to be excluded in the regeneration experiments performed
on Nemertines. In these animals, there is a certain region at the
anterior end of the body in front of the mouth, which contains no
endodermal tissue at all. If such a piece be isolated, it will recon-
stitute itself into a complete worm, with an alimentary canal which
quite certainly, therefore, is derived from cells of an entirely
different tissue.^
^ A.H.Drew, 1923. ^ Lewis, 1931.
^ Nussbaum and Oxner, 19 10.
14-2
212 THE MOSAIC STAGE OF DIFFERENTIATION
■W^ - ' - . "— .^ - "' ^'"^ — msl.fbr.
^
'/--
msl.fbr.
^=-fbl.lyr,
■ — deg. ov.
i-':.-: %
t) f >
_ — msl.fbr.
— -msl. nuc,
>fbl.lyr.
— cil. ep.
.^^deg.ov.
— b.c.
Fig. loi
Metaplasia of fibroblasts of Pecten into ciliated epithelium. A piece of ovarian
tissue was grafted into the adductor muscle where a cyst was formed round it,
lined by an epithelium formed of fibroblasts. 1-4, stages in the transformation
of the fibroblasts into ciliated epithelium ; i , after 23 days ; 2, 26 days ; 3, 30 days ;
4, 98 days. b.c. blood corpuscles; cil.ep. ciliated epithelium; deg.vo. degene-
rating ovarian grafted tissue; fhl.lyr. fibroblast layer; msl.fbr. muscle fibres;
msl. nuc. nuclei of muscle cells. (From Gray, Experime?ital Cytology, Cambridge,
193 1, after G. H. Drew.)
THE MOSAIC STAGE OF DIFFERENTIATION 213
Metaplasia has been observed to occur as a result of certain graft-
ing experiments. If a small piece of the ripe ovary of the Mollusc
Pecten is grafted into the adductor muscle of another individual,
the implant rapidly becomes surrounded by a layer of fibroblasts.
The grafted tissue degenerates and is destroyed by phagocytosis,
but the fibroblasts remain, forming the lining of a cyst containing
the debris. After three weeks, the fibroblasts begin to take on the
appearance of columnar epithelium, which eventually becomes
ciHated.i It is almost impossible to believe that undifferentiated
ciliated cells were originally present in the muscle, and we are ac-
cordingly forced to regard this case as one of true metaplasia (fig. i oi ) .
Tissue-culture experiments likewise provide evidence for meta-
plasia. Monocytes which have been treated with filtered extracts of
a particular type of tumour (the Rous sarcoma) become transformed
into fibroblasts.^ The crowding of the cells in the culture often
produces the same effect, whereas various modifications of the
medium fail to do so. The change into fibroblasts is of an adaptive
nature, occurring when conditions are becoming impossible for the
continued existence of monocytes.^ This transformation may be
permanent.^ On the other hand, fibroblasts treated with plasma
containing liver extract may become transformed into macrophages
with all their physiological characteristics, which they now keep
indefinitely.^
Cultivation of fibroblasts in a plasma medium which only permits
of their slow growth may also induce metaplasia into macrophages.
Here again, the change to the outwandering macrophage type is
probably adaptive. Even Carrel's 20-year old strain of fibroblasts
has been made to produce daughter-strains of macrophages in this
way (fig. 102).
The rate of growth of the macrophages is markedly superior to
that of their parent fibroblasts ; they appear to retain their charac-
teristics indefinitely^ (fig. 103).
Lastly, it may be mentioned that the obscure changes which
tissues undergo when tumours and_^cancers arise are of the nature
of metaplasia. The morphological characters of the cells are lost to
^ G, H. Drew, 191 1. 2 Carrel, 193 1.
^ Carrel and Ebeling, 1926. * Fischer, 1925.
^ R. C. Parker, 1932 c.
214 THE MOSAIC STAGE OF DIFFERENTIATION
a greater or lesser extent, and these transformations are accom-
panied by irreversible physiological changes, as a result of which
the tumour cell becomes capable of glycolysis (or fermentative in-
t
\M^^:^
'^i.'\
' '^ ••
^ t
p
c
k^:'
-%.
c>
Fig. 1 02
Microphotographs of living cultures of chick fibroblast tissue. Above, typical
fibroblasts (after 103 days' cultivation and 13 passages). Below, macrophages
derived by metaplasia from a pure culture of fibroblasts (12 passages as pure
fibroblast culture in optimum medium, then 29 days' growth in an unfavourable
medium containing no embryonic tissue juice). (From R. C. Parker, Joiirn. Exp.
Med. LViii, 1932.)
tramolecular respiration) and is less dependent on normal aerobic
respiration.^
These examples will be sufficient to demonstrate the real exist-
^ See Warburg, 1926.
THE MOSAIC STAGE OF DIFFERENTIATION
215
ence of what has here been called the mosaic stage of differentiation
and development. This is of great theoretical interest, since it shows
that the capacity for regulation, which has been regarded by some
authors ^ as a universal property of life, does not hold at all for an
important stage of development, universally passed through by all
higher animals.
'0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54
Days
Fig. 103
Physiological changes accompanying metaplasia. Solid line throughout (401-3),
growth-curve of a flask culture of fibroblasts from embryo chick muscle ; although
grown in an unfavourable medium (lacking embryo extract), the culture showed
no metaplasia. 104-1 , growth-curve of sister-culture under similar conditions, in
which metaplasia fibroblasts to macrophages occurred on the eighth day. The
subsequent growth (dotted line) was much more rapid. (From R. C. Parker,
Journ. Exp. Med. lviii, 1932.)
§4
Turning now to the question of the time of onset of the mosaic
stage of development, we must refer to the classical experiments on
so-called mosaic-eggs, referred to in Chap. v. They serve as a
further illustration of the principle which is here under discussion ;
in their case, the onset of chemo-differentiation has merely been
transferred to an earlier stage of development.
^ Driesch, 1921 ; J. S. Haldane, 1929.
2l6 THE MOSAIC STAGE OF DIFFERENTIATION
It will be remembered that in Dentalium (p. no), chemo-difFer-
entiated substances are present in the polar lobe and become in-
corporated in blastomere Z), with the result that of the cells of the
4-cell stage, only blastomere D is able to produce a complete larva,
but that the structures to which the polar lobe gives rise (apical
organ and post-trochal region with mesoderm) are full-sized, and
therefore disproportionately large.
On the other hand, in Tubifex (p. 113), chemo-differentiated
substances are present in the pole-plasms, and likewise become
incorporated in blastomere Z), which, alone of the blastomeres of
the 4-cell stage, is capable of giving rise to a complete larva. But
this larva produced from D in Tiihifex is properly proportioned.
Blastomeres AB, A, B^ or C of Dentalium and Tuhifex are in-
capable of giving rise to a complete larva, not because of any
positive determination to differentiate along the lines of their pro-
spective fates, but because of the negative fact that they lack the
essential ingredients for forming structures to which they do not
give rise in normal development. Further, it is clear that in Den-
talitim, regulation in blastomere D occurs in some respects but not
in others. The larvae thus formed have regulated as regards their
external form and show no trace of asymmetry, but the charac-
teristics dependent upon the polar lobe (apical organ and post-
trochal region) are disproportionately large. In Tubifex, on the
other hand, regulation in blastomere D seems to be complete. If
we are to make a conjecture as to the meaning of this distinction,
it wouM be that chemo- differentiation is more precocious in
Dentalium and results in a complete determination, quantitative as
well as qualitative, of the organ-forming substances contained in
the polar lobe. The state of affairs in Tubifex, on the other hand, is
more like that of the early half-gastrula of Triton, in which quanti-
tative regulation of the neural folds is still possible (see p. 239).^
The most striking demonstration of the presence of organ-
forming substances is that of the Ascidians, already referred to
(pp. 119, 123). The fertilised egg possesses a yellow crescent and
^ In the absence of experiments involving the removal of the pole-plasms of
Tubifex, comparable to those in which the polar lobe of Dentaliwn is cut off, it is
impossible to rule out the suggestion made by Morgan (1927, p. 379) that the
pole-plasms of Tuhifex may be indices of some underlying peculiarity of organi-
sation, rather than organ-forming substances.
THE MOSAIC STAGE OF DIFFERENTIATION 217
a clear crescent, and a large amount of yolk. The first cleavage takes
place in a plane passing through the centre of the yellow and clear
crescents, and in each blastomere the clear cytoplasm displaces the
yolk from the animal hemisphere so that the latter now occupies
the vegetative hemisphere. Immediately opposite the yellow
crescent, and therefore marking the antero-dorsal side, a third
region termed the grey crescent makes its appearance, containing
slaty-blue coloured yolk. Eventually, the yellow crescent shows a
darker and a lighter coloured region, and there are then at least six
different organ-forming substances, which become sorted out
between the various blastomeres during cleavage. The determina-
tions which these substances represent are the following :^
Clear cytoplasm . . . Ectoderm
Dark yellow cytoplasm Muscle
Light yellow cytoplasm Mesenchyme
Yolk region Endoderm
Slaty-blue Notochord and neural plate
This distribution can be seen in normal cleavage; the causal
connexion between the substances and the organs to whose rudi-
ments they are distributed has been proved by the experiments
involving killing and disarranging of blastomeres, referred to in
Chap, v (pp. 97, 123).
As mentioned in Chap, v (p. 124), the visible inclusions in the
various regions of cytoplasm, such as mitochondria, yolk, etc., are
not themselves organ-forming substances, but merely cytological
indices of the organisation of the egg.^ In many other cases par-
ticular regions of the cytoplasm may be distinguished by their
pigmentation, but it can in most cases be shown that the visible or
coloured elements do not represent any qualitative determination.
The egg of the sea-urchin Arbacia contains fat, yolk, pink granules,
and clear cytoplasm, and as these materials differ in their specific
gravities, they can be disarranged by the centrifuge. When the eggs
of Arbacia are centrifugalised for 5 minutes at 10,000 revolutions
per minute, the contents are stratified into four zones, quite regard-
less of the original egg-axis, which, it is found, may come to occupy
any position in the centrifuge tube. Centripetally (with reference
1 Conklin, 1905, 1906, 1924, I93i- ^ Conklin, 193 1.
2l8
THE MOSAIC STAGE OF DIFFERENTIATION
to the centrifuge), the fat forms a layer, and beneath this, in suc-
cession, there are layers of clear cytoplasm, yolk, and pink granules,
the latter occupying the centrifugal pole^ (fig. 104).
But in spite of this complete restratification of the visible egg-
contents along a new axis, the original axis of polarity has not been
affected. The position of the original axis is indicated by the funnel
in the jelly which marks the position of the original animal pole of
the egg : the micromeres are formed and invagination begins at the
opposite (original vegetative) pole, regardless of the visible contents
Fig. 104
Persistence of the primary axis in sea-urchin (Arbacta) eggs in spite of the re-
arrangement of visible substances in the cytoplasm. After centrifuging, the egg
becomes stratified with fat at the centripetal pole, then clear cytoplasm, then
yolk with increasing amounts of pigment. The first cleavage (left top) is always
at right angles to the stratification, but the micromeres are always formed at the
vegetative end of the original axis, whether this coincides with the centripetal
pole of the centrifuged egg (top right), its centrifugal pole (bottom left) or its
side (bottom right). (Redrawn after Morgan, Experimental Embryology, Columbia
University Press, 1927.)
which happen to be situated there.'^ Development continues along
the lines of the original axis and is normal, from which it follows
that the various substances which have been disarranged are not
organ-forming. (See also p. 69 as regards determination of
bilaterality by centrifuging.)
Similar results have been obtained from centrifuge experiments
on eggs of other animals. The egg of the Lamellibranch Mollusc
Cumingia after centrifuging shows a stratification into four zones :
Lyon, 1906; Morgan and Lyon, 1907.
Morgan and Spooner, 1909.
THE MOSAIC STAGE OF DIFFERENTIATION
219
an oil cap, a clear zone, a yolk field, and a zone of pigment; this
stratification may bear any relation to the original axis of polarity.
Nevertheless, normal larvae develop, regardless of the distribution
of the visible contents. ^ The same is true of the egg of Chaetopterus,
and the polar lobe may contain any of the visible materials without
influencing normal development.^
A B
Fig. 105
Section through a frog tadpole (external gill stage) developed from an egg centri-
fuged for 5 minutes at about 1500 revolutions a minute. A, Through the head
region. The brain (b.) is represented by a degenerate mass of pigment cells. The
cranium (cr.) is rudimentary; hy. hyoid; br.c. branchial cleft. B, Through the
trunk and the spinal cord (sp.c); the distribution of cells is abnormal, and the
spinal ganglia (sp.g.) are fused below it; int., intestine. In both, the myotomes
(my.) are fused in the middle line. (After Jenkinson, Quart. Journ. Micr. Sci.
LX, IQ15.)
By way of contrast with specific materials of the organ-forming
type as seen in Dentaliiim and Styela, the preformed substances
such as yolk and fat to be found in many eggs thus appear to play
the part of raw materials only. Their importance as regards normal
development is perhaps best illustrated by the experiments of
centrifuging the eggs of the frog. When a frog's tgg is thus treated,
int.
Morgan, 1910.
Lillie, 1906.
220 THE MOSAIC STAGE OF DIFFERENTIATION
its ability to rotate within its membranes and the greater weight of
its yolk cause it to orientate itself in the centrifuge tube in such a
way that the animal pole is centripetal and the vegetative centri-
fugal. The result of centrifugalisation is therefore an intensification
of the stratification normally found along the primary egg-axis. The
yolk is concentrated more densely than ever at the vegetative pole ;
above it is a layer of clear cytoplasm, and the animal pole is occu-
pied by a layer of fat. If the centrifugalisation is heavy, develop-
ment proceeds a certain way and then stops, largely owing to
mechanical difficulties arising from the inertia of the abnormally
dense mass of yolk. But if the centrifugalisation is light, develop-
ment is normal except for the fact that the structures of the head
contain an abnormally large amount of fat. The cells of the brain
may contain many times the normal quantity of fat, but neverthe-
less the differentiation of the brain and the development of its form
are normal. Similarly, some of the regions of the trunk can develop
normally although their cells contain less than the normal quantity
of yolk. It is obvious, therefore, that yolk and fat are only raw
materials. 1 When, however, the amount of fat at the animal pole
exceeds a certain proportion, normal development is impossible.
Vacuolisation is the first sign, but in more extreme cases the
brain and other head-structures are reduced to a small degenerated
mass of cells ^ (fig. 105).
We may illustrate the part played by the yolk and fat in the frog's
egg with the help of an analogy. The construction of a conservatory
is of course conditioned by the availability of the necessary raw
materials — wood and glass. There is an optimum proportion in
which these materials should be present in order to give the best
results, but this proportion may be altered in either direction up to
^ Jenkinson, 1915.
- With slightly heavier centrifugaHsation, curious malformations appear in the
trunk region. The myotomes are frequently fused together beneath the nerve tube,
with consequent absence of the notochord. The spinal gangHa may also be fused
ventrally beneath the nerve tube. The latter has an abnormally thick floor and
thin roof, with the white matter concentrated ventrally. From other experiments
(see Chap, xi, p. 375), it is known that these effects are associated with notochord
absence, and it is probable therefore that here absence of the notochord is the
cause of the other observed changes, but the cause of this primary change re-
mains for the present obscure. Possibly the centrifugalisation has resulted in an
alteration in the composition of the organiser region : further research is needed
on the question.
THE MOSAIC STAGE OF DIFFERENTIATION 221
a certain point without preventing the construction. It may have
too much wood and not enough glass, or too much glass and not
enough wood, but, provided that the disproportion does not exceed
a certain degree, it will still be a conservatory. But if the amount
of glass be too great for the wood, the construction is mechanically
impossible. The yolk and fat in the frog's egg may fancifully be
compared with the wood and glass in the conservatory.
Other evidence of a similar nature is provided by centrifuge
experiments on the eggs of echinoderms, in which centrifugation
has been continued until the tgg has separated into two or even
four (unequal) portions along the direction of centrifugal force
(which of course may bear any relation to the original polarity of
the egg). The fragm.ents differ considerably in colour and the type
of their contained granules. We may call these halves A and B, and
the quarters A^, A^, B^, B2, in order from centripetal to centrifugal
region. In Sphaer echinus gramdatus, a centrifugal half {B) or either
of the two centrifugal quarters {B^ or ^2) is capable of producing
plutei. A centripetal half {A) on the other hand never goes further
than the blastula, and the same is true for the most centripetal
quarter (^1). The other quarter {A^, however, may in some cases
produce a pluteus. We may provisionally assume that the fragments
incapable of pluteus-formation contain an excess or defect of certain
raw materials, as in the frog experiments described above. Pre-
sumably the excess substance responsible for failure to develop in
the A halves was all contained in the A^ quarters, thus permitting
the A2 quarters to develop. Results similar in principle have been
obtained for several other genera : in one case ( Tripneiistes esculentes)
the conditions are reversed, the A (centripetal) pieces being capable
of fuller development.^
§5
The determination and localisation of organ-rudiments is revealed
sooner or later by the presence of chemo-differentiated material or
morphogenetic substances in certain places which constitute what
may be called fields, or areas of differentiation of organs. Within
the fields the presumptive rudiments become determined by pro-
gressive chemo- differentiation. As an illustration of this important
1 Harvey, 1933.
222
THE MOSAIC STAGE OF DIFFERENTIATION
principle, we may turn to the phenomena presented by some of the
presumptive organ-regions in Amphibia, beginning with the Hmbs.
C C
Fig. 1 06
Diagram of the fate of the four quadrants of the fore-Hmb field in Anihly stoma.
The Hmb-disc is shown at A in each case, with one quadrant stippled. The dotted
lines intersect at the point of maximum limb-forming potency. At C in each case
is shown the young limb, viewed dorsally (above) and laterally (below), showing
the portions derived from the stippled quadrant. (From Swett, Journ. Exp.
Zool. xxxvn, 1923.)
In Amblystoma, the presumptive fore-Hmb area or field occupies
a discoid zone on the side of the body, extending from the anterior
margin of the third trunk segment to the middle of the sixth. The
THE MOSAIC STAGE OF DIFFERENTIATION
223
limb potencies are restricted to the mesoderm of this region,
the ectoderm not being predetermined in any way.^ The first
important point to notice is that within this limb-disc there is
no definite spot or area which is necessarily destined to form a limb
in normal development : all regions of the field have the power of
forming a limb, and the extent of the field is greater than the region
which actually does form the limb in normal development.^ The
Hmb-field is already determined at the middle gastrula stage.^
The limb-forming potencies are highest in a subcentral region
of the field, situated near to its anterior and dorsal margins, and
grade away from this.^ A normal limb can be formed from half the
Fig. 107
Polarisation of the limb-field. Axolotl in which a limb-disc from the right side of
the body has been grafted on to the same side, a little way behind the normal
limb, the correct side out but with the antero-posterior and dorso- ventral axes
reversed. It has developed into a limb (TR) with correct dorso- ventral relations,
but with the preaxial border facing the tail of the larva ; consequently it possesses
left-handed asymmetry. (From Harrison, Jowrw. Exp. Zool. xxxii, 1921.)
limb-field, either from what is left in situ after removal of half, or
from a half grafted elsewhere : a single field may therefore give rise
to two perfect limbs. Conversely, a single perfect limb can be
formed from two half-rudiments grafted together (provided only
that their antero-posterior axes are coincident, see below, p. 224
and also pp. 357, 418). From a very early stage, therefore, the
limb-field is irreversibly determined as a whole to give rise to
limb-tissue, but there is as yet no regional determination within
the field, of the constituent parts of the future limb. In addition
to the fact that a limb will arise somewhere within the limits of
the field, there is only one additional determination, and that is
^ Harrison, 1918; Detwiler, 1918.
3 Swett, 1923.
Detwiler, 1929 a; 1933 a.
224
THE MOSAIC STAGE OF DIFFERENTIATION
that the preaxial border of the Hmb (marked by the first digit and
radius in the fully developed limb) will arise from the anterior
portion of the limb-disc. Although the limb-field is regionally still
undetermined, it is polarised along an antero-posterior axis from
the first moment at which its existence can be detected: this is
proved by grafting limb-discs in abnormal orientations^ (figs- 107,
108, 173).
A B
Fig. 108
Limb determination in Amblystoma. A, Middle gastrula stage showing pre-
sumptive limb area which was removed and grafted with reversed orientation into
a neurula (B), where it developed into a limb with left asymmetry, though on
the right side, like that shown in fig. 107. This proves not only that the limb is
determined at the middle gastrula stage, but also that its antero-posterior axis is
already determined. (From Detwiler, Jo^rw. Exp. Zool. lxiv, 1933, figs. 2, 3.)
The hind-limb field in Amblystoma extends from the level of the
sixteenth to the eighteenth trunk segments inclusive, and shows
properties similar to those of the fore-limb field. ^ Its determina-
tion and differentiation takes place later than that of the fore-limb
field.
^ A common occurrence when portions of limb-discs are grafted is the fact
that they give rise to reduplications, i.e. monstrous double or even treble limbs
are formed connected with one another at some point along their length. This
in itself is merely another example of the fact that the limb-area is as yet only a
field and not a regionally determined rudiment. But these reduplications are of
interest from another point of view, for the reduplicated member is as a rule a
mirror image of the original member. They therefore supply an illustration of
Bateson's rule, which may be forrriulated as follows: (i) the long axes of re-
duplicated structures lie in the same plane ; (2) two reduplicated limbs are mirror
images of one another about a plane which bisects the angle between the long
axes of the members, and which is at right angles to the plane of these axes.
The detailed explanation of reduplication and mirror-imaging has given rise
to considerable controversy. See Harrison, 1921 a; Przibram, 1924; Mangold,
1929 A.
Stultz, 193 1.
THE MOSAIC STAGE OF DIFFERENTIATION 225
By this stage, too, the growth-coefficient of the limb relative to
the body has also been determined, as is shown by heteroplastic
experiments in grafting limbs between slow-growing and fast-
growing species of Amhlystoma. Limbs of the fast-growing species
on the body of the slow-growing one become disproportionately
large, and vice versa} (See fig. 203, p. 421.)
It is only at later stages that the unitary limb region, which forms
one of the major pieces in the mosaic of the whole organism, itself
becomes converted into a mosaic of invisibly determined sub-
regions. The precise time of onset of this stage varies in different
forms. In Ambly stoma punctatum it appears to be reached when the
visible limb-bud has attained a markedly conical form. The organ-
ism is then a larva with well-developed external gills and tail. In
Triton taeniatus, on the other hand, it appears to set in relatively
earlier, in the tail-bud stage. ^ In Triton, the limb also develops
relatively earliei than in Ambly stoma punctatum, but there is no
correlation between time of determination and time of develop-
ment, for in Amhlystoma tigrinum^ determination sets in earlier
but development does not occur until later than in Amhlystoma
punctatum.
When the stage of regional determination of subregions within
the field has been reached, division of the rudiment will no longer
result in the formation of two limbs by regulation, but each portion
will give rise to a partial structure. Progressive chemo-diflFerentia-
tion has taken place, and within the main limb-field a secondary
mosaic has been formed, each region of which, however, is still
indefinitely determined and therefore capable of regulation.
The analysis of these late stages has been undertaken in the limbs
of the embryo chick. By grafting portions of the limb-bud of a
4-day chick on to the chorio-allantois of another Qgg, it is found that
if the limb-bud is divided into pieces by cuts at right angles to its
future long axis, the proximal piece differentiates into a perfect
femur, the next piece into a perfect tibia and fibula, and the distal
piece into a perfect foot. It is important to note that even the
structure of the joints appears to be predetermined in almost all its
details.* (See figs. 109, 11 1.)
Harrison, 1924 a; Huxley. 1932. ^ Brandt, 1924.
Ruud, 1926. * Murray, 1926.
15
226
THE MOSAIC STAGE OF DIFFERENTIATION
It is impossible to imagine that the cuts which were made passed
exactly in each case between the limits of the zones allocated to
thigh, shank, and foot, and it is necessary to conclude that these
I •
11
V
I
Fig. 109
Mosaic determination and partial regulation within the limb-rudiment of the
chick. Differentiation of a small basal fragment of a very early (4-day) left hind-
limb bud, grafted on the chorio-allantois of another egg after 5 days. Right,
microphotograph of entire graft, in longitudinal section. The connexion with the
chorio-allantois is seen on the right: the graft has differentiated into a femur,
7*5 mm. long, mesenchyme, and some muscle-fibres (right bottom). Left, re-
construction of skeletal elements. The curve of the bone is in the same direction
as in a normal left femur, i , head ; 2 , shaft ; 3 , ectopic fragment of pelvis ; 4, sheath
of perichondral bone; 5, attachment of muscles (on far side) ; 6, patella; 7, tro-
chanter. Being the basal region, the graft has formed only basal structures ; there
has however been some intra-regional regulation, leading to the formation of a
femur complete at either end. (From Murray and Huxley, Joz/rw, Anat. lix, i 925 .)
segments or constituent parts of the limb are roughly determined
at varying levels down its length, but that they are determined
THE MOSAIC STAGE OF DIFFERENTIATION
227
IS no
only roughly, and their frontiers appear to overlap. There 11 .
doubt that a cell which in one experiment forms part of the thigh,
would, in another experiment with the cut in a slightly different
Fig. no
Mosaic determination within the hind-Hmb rudiment of the chick. A, Com-
plete hind-limb developed from a chorio-allantoic graft of a whole limb-bud in
the stage shown at B. A graft of the distal half of the limb-bud shown at C
resulted in a distal half-limb (D). A still smaller distal region (E) produced only
a foot (F). In D and F, the sub-regions (shank, foot) are complete. (After
Murray, from Wells, Huxley and Wells, The Science of Life, London, 1929.)
place, form part of the shank. The cuts must have roughly separ-
ated the sub-zones from one another, and each sub-zone, though
irreversibly determined to give rise only to its own segment, is
still capable of regulation to give a z^/zo/^ sub-zone. In a similar way,
15-2
228
THE MOSAIC STAGE OF DIFFERENTIATION
some of the original presumptive limb-area does not give rise in
normal development to limb, but merely to flank, skin, and muscle.
Similarly, if the 4-day leg-bud be divided longitudinally, so as
to separate preaxial and postaxial halves, the fragment usually
forms only those digits which it would have produced if left in situ
and either a tibia or a fibula. The femur rudiment regulates to a
miniature whole in each portion. The limb-bud thus appears to be
Fig. Ill
Self-differentiation of the femur-head joint, without function. The femur shown
in the section belonged to a limb from which all nerve supply was excluded. In
spite of the fact that the limb never functioned at all, the cartilage cells and fibres
of the femur-head show the normal configuration. (From Hamburger, Arch.
Entwmech. cxiv, 1928.)
a thorough-going mosaic of predetermined but slightly overlapping
regions (fig. no).
The recognition of the existence of organ-fields, i.e. regions
possessing a general determination for the production of certain
structures, and undergoing progressive regional specification of
detail, constitutes one of the major advances made in the analytical
study of development. The organ-fields resemble the whole organ-
THE MOSAIC STAGE OF DIFFERENTIATION
229
Neural Tube Field
Fig. 112
Diagram of an amphibian neurula to
show the approximate localisation of
the main regional fields as yet dis-
covered by experimental analysis. The
ism in the pre-mosaic stage, in combining a general determination
with an epigenetic mode of development.
The arm-field and the leg-field are each of them self-differenti-
ating in a general way, and will produce only an arm or a leg, as the
case may be.^ But certain details of the development of the field
are not independent of the im-
mediate environment, i.e. the re-
mainder of the organism, and
in particular, its gradient-field.
We have seen that the arm-field
of Amblystoma is polarised from
its inception. The leg-field also
has some polarity from the start,
as is revealed when it is grafted
heterotopically in an abnormal
orientation: the original antero-
posterior axis of polarity of the
disc persists and becomes that of arrows indicate that the fields are
the limb. After rotation of the known to be polarised from their first
. . appearance. (Origmal.)
leg-disc at the origmal site (ortho-
topic), however, it acquires a new polarity in relation to that of the
body as a whole, i.e. the leg-disc has a new antero-posterior axis
impressed upon it.^ It may be hazarded that this result is due to the
size of the disc rotated not being large enough to cover the whole
limb-field. Later, however, the antero-posterior axis of the leg-disc
is entirely fixed, but the dorso-ventral axis is not. This is also the
condition of the arm-disc at the earliest stage studied (middle
gastrula). Rotation experiments with discs at this stage show that
the determination of the dorso-ventral axis is dependent on the body
of the host. Later still, the dorso-ventral axis of the discs is also
determined, and the limb-disc has by then proceeded far on the
way to becoming a mosaic of determined subregions.
The fact that a field, although qualitatively determined, is capable
of being quantitatively influenced in its development, is well shown
by those experiments in which an ^arm-disc of an Amblystoma
neurula is grafted in place of an extirpated leg-disc. The graft
^ Harrison, 1918; Ruud, 1929, 193 1 1 Stultz, 193 1.
2 Stultz, 193 1.
230
THE MOSAIC STAGE OF DIFFERENTIATION
develops into an unmistakable arm, but it may have five digits. This
is the normal number of digits of the amphibian leg, while the
normal arm only has four ^ (fig. 113). In this case, we may suppose
that the larger nerve-trunk supplying the leg exerts a trophic effect
on the growing rudiment, leading to a condition in which the
distal region tends to be meristically divided into five instead of
four digits. This is the converse of the results obtained after
Fig. 113
Modification of the arm-bud when grafted into the leg-region. The amphibian
arm ends in four fingers (i , 2, 3, 4) ; the leg in five toes (i , 2, 3 , 4, 5). An arm-bud
of a white axolotl, grafted into the leg-region (after extirpation of host's leg) of
a black axolotl, develops into a typical arm {Tr), except that it possesses five
fingers. (From Ruud, Arch. Entzumech. cxviii, 1929-)
inducing subnormal development in the centres of the mid-brain
by extirpating the rudiments of fore-limb or eye : - in these cases
the hind-limb was usually malformed and under-developed, and
frequently possessed only four or even three toes. Curiously enough,
totally denervated hind-limb rudiments, though their growth is
reduced, are not malformed, and develop the normal complement
of five toes.^ It may be supposed that abnormal conditions in
1 Ruud, 1929. ^ Diirken, 1925, 1930. ^ Hamburger, 1928.
THE MOSAIC STAGE OF DIFFERENTIATION 231
the mid-brain exert a specific ''negative" trophic eflfect in the
hind-Hmbs, but the question cannot be regarded as settled. (See
P- 430-)
Before leaving the limb-field, there is a further point which re-
quires consideration. As early as the neurula stage, the mesoderm
of the limb-field is found to be self-diflFerentiating. At the same
time, it is clear, if only from the fact that the limbs are sym-
metrically placed with regard to the plane of bilateral symmetry,
that the localisation of limb-forming potencies is in some way de-
pendent on something else, which we may at present call the general
gradient-field of the organism (see Chap. ix). We are ignorant as
to the causes which are normally operative in calling forth these
limb-forming potencies in ordinary development, but we do know
that these potencies may be experimentally called forth by a variety
of agents. Grafted ear-vesicles,^ celloidin beads, or the free termina-
tion of various nerves deflected so as to underlie the tissues of the
field,- all result in the formation of limbs. The quality of the struc-
ture produced is therefore a specific property of the field, the
activities of which may be "released" by a variety of non-specific
agents. As we shall see, the same is true of other fields, and
probably of all.
§6
A curious contrast to the regulative capacity of the limb-field is
the mosaic nature of the rudiment of the shoulder-girdle. This
rudiment consists of three centres of chondrification, representing
the coracoid, precoracoid, and scapular elements, but they are not
all contained within the 3 J-somite limb-disc, for when the latter is
grafted it will give rise to a shoulder-girdle of about one-third
normal size.^ Conversely, after extirpation of a limb-disc, portions
of the shoulder-girdle rudiment are left in sitii.^ Removal of, or
grafting of, portions of the limb-girdle rudiment at the early tail-
bud stage in Amblystoma results in the development of partial
structures, and regulation to form a complete girdle does not take
place, while regulation does take place to form a perfect limb.^
In some experiments on Amblystoma in which limb-discs
^ Balinsky, 1925, 1926, 1927; Filatow, 1927. - Detwiler, 1918.
^ Locatelli, 1925; Guyenot and Schotte, 1926; Bovet, 1930.
^ Harrison, 1918.
232 THE MOSAIC STAGE OF DIFFERENTIATION
(3 J somites in diameter) were grafted after rotation, it was found
that the limb occasionally underwent a rotation at the shoulder-
girdle, so as to become correctly oriented. ^ When the disc was
rotated through an angle up to 235°, the limb might right itself by
a rotation in the reverse direction : on the other hand, if the disc
was rotated through 270°, the limb might complete the circle by
rotating the remaining 90' in the same direction.^
It appears that the rotation of the limb is in some ways dependent
on the shoulder-girdle. If the rotated disc is only i| somites in
diameter, it contains none of the girdle-rudiment, and no regu-
latory rotation takes place. In the case of rotated discs of the normal
diameter of 3 1 somites, parts of the shoulder-girdle are formed, and
regulatory rotation may take place. If the rotated disc is 5 somites
in diameter, a complete girdle is formed and the limb conforms to it,
without regulatory rotation.^ Lastly, if in a graft 5 somites in
diameter a 3J-somite disc is separated from a peripheral ring, and
then both central disc and peripheral ring are rotated independently,
the limb undergoes postural regulation with reference to the ring.*
Apparently, therefore, the portions of the girdle whose rudiments
lie outside the 3j-somite disc but within the 5 -somite ring, act as
determining factors on those portions of the girdle whose rudi-
ments are included within the 3J-somite disc. The girdle then
brings about the rotation of the limb, but in a manner which is still
obscure.
§7
Turning now to other examples of fields, we may take that of the
amphibian ear. This occupies a region of ectoderm on each side
of the head, behind the eye, and must in some way be dependent on
the organiser, since dorsal lip grafts are capable of inducing the
formation of ears. Here again, it is found that the field is more ex-
tensive than the normal presumptive rudiment, for if a portion of
the presumptive ear-area be removed at the early neurula stage in
Rana nigromaculatay a normal (though smaller) vesicle is formed,
and this can be shown to arise from the neighbouring cells which
have closed over the wound, though these would normally have
^ Harrison, 1921 a. - Nicholas; 1924 b,
2 Nicholas, 1926. * Nicholas, 1925.
THE MOSAIC STAGE OF DIFFERENTIATION
233
given rise to epidermis.^ This shows that more cells are capable
of ear-formation than normally exert this capacity.
This is confirmed for Urodeles by experiments on Amhlystoma.
At later stages, when the rudiment has invaginated to form the ear-
vesicle, the power of ear-formation is lost by the neighbouring
epidermis,^ for if the vesicle is extirpated it is not regenerated.
Just as in the case of the limb-field, the ear-field very soon shows
a polarisation. If a piece of the ear-area of Rana nigromaculata (at
the stage when the rudiment is just thickened) is rotated through
180°, the auditory vesicle which sub-
sequently develops is reversed. Further,
a piece of the ear-area of one side grafted
on to the opposite side of the body de-
velops with the asymmetry of its side
of origin.^ This shows that the rudi- WHP- * - -^ — ^^
ment was already determined as regards
two at least of its axes.
The gill-region in Amphibia also
constitutes a field in the ectoderm of
the embryo. At the early neurula stage
in embryos of Amblystoma,^ Rana fiisca
and esciilenta, and Bombinator,^ rotation ^ig. 114
of a piece of the gill-area through 180° Ventral view of larva of Boju-
is followed by development of the gills ^^'"^^^r in which the ectoderm
. , 1 • A \ • J of the gill-region on the left
(and operculum m Anura) m reversed gj^e had been rotated through
orientation. This shows that the field is 180°. * limit between rotated
polarised along an antero-posterior axis. ^^fi.i^roTfLetlS.bT^; ^\
At the same time, the fact that at this op, operculum. (From Braus,
early stage it is still only a field and Zeitschr.f.MorpJuu.Anthrop.
. ,, , • tt 1 XVIII, 1914.)
not a spatially and regionally deter-
mined rudiment is shown by the capacity of two rudiments grafted
together to regulate and give rise to a single normal set of gills
(provided that the antero-posterior axes coincide)^ (fig. 114).
Turning now to the heart, it is found in embryos of Bomhinator
at the neurula stage that if the presumptive heart-area (which occu-
pies a region of the mesoderm) is extirpated, a heart is formed from
^ Tokura, 1925.
^ Harrison, 192 1 b.
- Kaan, 1926.
* Ekman, 1913, 192;
234
THE MOSAIC STAGE OF DIFFERENTIATION
neighbouring regions : ^ it may therefore safely be concluded that
the heart-field is more extensive than the actual presumptive heart
Fig. 115
Experiments on the development of the heart in Bombinator pachypiis . A, Single
normal heart formed from an enlarged rudiment, that of another embryo having
been added in the mid-ventral line : the parts arising from the graft shown dark.
B, C, Partially doubled hearts formed after grafting foreign tissue (pharyngeal
wall) into the mid-ventral line ; ventricle and bulbus show duplication ; the graft
has been used up in the formation of the heart in B, but in C a portion remains
as undifferentiated material. D, Heart duplication almost complete, as a result of
grafting a large piece of foreign tissue in the mid-ventral line; most of the graft
has remained undifferentiated, and only a small portion (shown dark) has con-
tributed to the heart in front and behind. E, Complete duplication as a result of
grafting a piece of foreign tissue which has remained undifferentiated ; the right-
hand member shows situs inversus. (Redrawn after Ekman, Arch. Entiomech.
cvi, 1925.)
region. Heart-forming potencies decrease with increasing distance
from the normal presumptive heart region^ and at the neurula
^ Ekman, 1921. " Ekman, 1925.
THE MOSAIC STAGE OF DIFFERENTIATION 235
Stage, a piece of the heart-field can be rotated through i8o° and
still give a normal heart, but this is no longer possible at later stages
such as the tail-bud. ^ As in the case of the limb, ear, and gill-fields,
therefore, the heart-field is polarised along an antero-posterior axis
from an early stage.
A normal heart can be formed from a longitudinal half-rudi-
ment ; a single rudiment, split lengthwise, can be made to give rise
to two or even three hearts ; and two rudiments grafted together at
^f
^--^ .#:..
A B
Fig. 116
Power of regulation of the heart-field, in Bombinator. At the neurula stage, the
left half of the heart-field was extirpated; the right half, remaining in situ, has
regulated to form a complete heart, with sinus venosus {S) ; atrium {A), ventricle
(F), and bulbus {B). The histological differentiation of the parts is normal.
Certain details of morphological differentiation are, however, abnorrnal; the
ventricle projects to the left, and its long axis may be longitudinal (as in A) or
transverse (as in B). Figures taken 14 days after operation. (From Stohr, Arch.
Entwmech. cxii, 1927.)
the neurula stage can regulate to form one normal heart, provided
that both the antero-posterior axes are similarly oriented.'^
All these results have been confirmed for Urodela by experiments
on Amblystoma.^
The epidermis itself may be regarded as a large field, the deter-
mination of which is characterised not so much by any positive
differentiation (for this is comparatively slight), but by the pro-
gressive incapacity to differentiate mto other structures, e.g. lens.
Nevertheless, the epidermis possesses a polarity, and this is ex-
1 Stohr, 1925. - Ekman, 1924.
^ Copenhaver, 1926.
236 THE MOSAIC STAGE OF DIFFERENTIATION
pressed in amphibian embryos by the direction of beat (mostly
antero-posterior) of the ciHa which it bears. The ciUa arise at the
early neurula stage, and if a piece of epidermis of Amblystoma is
rotated through 180° and replanted at this stage, the cilia beat in the
normal direction. If, however, the epidermis is rotated at the late
neurula stage, its polarity is then fixed, and the cilia beat in the
reversed direction.^
The so-called balancer, present in some Urodela larvae {Triton^
Diemyctyliis, Amblystoma punctatum, but absent or extremely rudi-
mentary in Amblystoma tigrinum'^), is an organ of attachment in the
form of a cylindrical projection of ectoderm with a mesenchymal
core. Balancer-forming potencies occupy fields in the ventral ecto-
derm of the head, beneath the eyes ; they are at a maximum at a
central point, and decrease with increasing distance from it.^ If a
part of the rudiment is extirpated at an early stage, a balancer will
be formed from the neighbouring regions.* A balancer rudiment
grafted into other positions is self-diiTerentiating, and induces the
formation of the mesenchymal core from host-tissue. A rudiment
from Amblystoma punctatum can be grafted on to Amblystoma
tigrinum, or even on to the anuran Rana sylvatica, and develop into
a balancer with induced core, although these hosts normally possess
no balancer. As mentioned in Chap, vi (p. 177), a single balancer-
field can give rise to as many as four balancers.^ (See also p. 327.)
The fully formed balancer has radial but not bilateral symmetry,
and it does not appear that the balancer-field is polarised. The facts
already recorded in Chap, vi (p. 177), viz. that balancer-forming
potencies can be evoked by neural crest cells, neural fold cells, and
fore-gut-wall cells, even belonging to Urodeles or Anura which
normally possess no balancer, serve as a further illustration of the
principle enunciated above (p. 23 1), that the quality of the structure
produced depends on intrinsic properties of the field, and not on
specific stimuli of the releasing mechanism. At the same time, the
fact that tissue from species which possess no balancer is capable of
evoking balancer-forming potencies, shows that the absence of a
balancer in these species is due to the absence of such a field in
^ Twitty, 1928. ^ Nicholas, 1924 a.
^ Harrison, 1925 b. ^ Bell, 1907.
^ Raven, 193 1 a.
THE MOSAIC STAGE OF DIFFERENTIATION 237
their tissues. This has been verified by experiments in which
epidermis from Amblystoma tigrinum is grafted into the appropriate
position on embryos of Amblystoma punctatum, and is found to be
incapable of forming a balancer. ^
Evidence regarding the existence of a nose-field is provided by
experiments on Rana temporaria in which the nose-rudiment is
extirpated at a stage prior to the formation of a nasal pit ; a nose
is nevertheless formed from neighbouring tissue, which grows over
to cover the wound. More distant epidermis will, however, not do
this. 2 The nose-field is therefore more extensive than the pre-
sumptive nose-rudiment, and if a large area, representing the entire
nose-field, is extirpated, no nose is formed. In some cases, the nose-
field gives rise to a single median nasal organ in place of the normal
paired two : this monorhiny is associated with and due to the same
causes as cyclopia (see Chap, ix, p. 348).
§8
Another case in which the presumptive zones or fields of different
organs appear to overlap is seen in the capacity which the adult
newt possesses of regenerating a lens to its eye.
The material for the regenerated lens is derived from the dorsal
margin of the iris of the eye-cup itself.^ It will be remembered
(see p. 187) that in Rana esculenta the eye-cup retains for a con-
siderable time the power of inducing a lens, and that in many forms
when an eye-cup is grafted into the body of another embryo in such
a way that it is deprived of contact with epidermis, the eye-cup may
form a lens from its own margin.^ This appears to be what happens
in the regeneration of the lens in the adult newt. The eye-cup is
then of course separated from the epidermis by the cornea, and the
epidermis itself is differentiated into the conjunctiva ; the edge of
the eye-cup is represented by the margin of the iris. This power and
method of regeneration implies that the lens-inducing faculty and
the lens-producing faculty have not been lost by the eye-cup even
in the adult.
The fact that it is always the dorsal margin of the iris which
provides the material for the regenerated lens requires considera-
1 Mangold, 1931c. - Ekman, 1923.
^ Colucci, 1891; Wolff, 1895. * Spemann, 1905 ; Adelmann, 1928.
238 THE MOSAIC STAGE OF DIFFERENTIATION
tion. The fact itself is attested by numerous experiments, in some
of which the newt is made to he on its back during the period of
regeneration;^ in others, the whole eye-cup is rotated in situ
through 180°, so that the choroid fissure which is normally ventral
comes to lie dorsally f in others again, the lens-forming potencies
of all parts of the iris margin are tested by grafting definite sectors
representing one-sixth of the circumference of the iris into the
cavity of the eye of another newt from which the lens has been
extirpated.^ In all cases there is found to be a gradient of lens-
forming potencies extending dorso-ventrally through the eye-cup
and resulting, in the intact eye-cup, in the regeneration of the lens
invariably from its dorsal margin. Once lens-formation is initiated
here, it inhibits the formation of lenses at other points. In this
connexion, it may be mentioned that the presence of the normal
lens inhibits such grafted fragments of iris from regenerating a new
lens. The presence of the normal iris, however, does not act in this
way, and it may sometimes regenerate a second lens from its own
upper border.'^
If, as already noted (p. 187), there is in the early neurula stage
a labile preliminary determination of a lens-area prior to the
definitive determination of a lens, we may suggest that some of this
area overlaps the eye-area, and that a portion of it becomes in-
corporated in the eye-cup. Further, topographical considerations
make it clear that if this were so, the region of maximum lens-
forming potency in the eye-cup would be the dorsal part, since, on
the analogy of the limb, lens-forming potencies must be assumed
to decrease along a radial gradient from some central spot in the
presumptive lens-area, and this spot lies nearer the dorsal than the
ventral part of the future eye-cup. If this is so, then, in the absence
of a lens, this dorsally situated tissue in the eye-cup may well be
stimulated to exhibit its original lens-forming capacity.^
1 G.Wolff, 1901. - Wachs, 1920. ^ Sato, 1930.
4 Spemann, 1905; Wachs, 1914; Sato, 1930.
^ Recent experiments have shown that the restriction of the site of lens-re-
generation to the dorsal margin of the iris is also due to the fact that this is the
region of the eye-cup which is farthest away from the choroid fissure, which is
always formed ventrally, and appears to exert an inhibitory effect on lens-
regeneration. If at the early tail-bud stage the optic vesicle is rotated in situ
about its stalk through 180°, the choroid fissure will be formed ventrally, in
tissue which was the presumptive dorsal part of the eye-cup. If, then, in later stages
THE MOSAIC STAGE OF DIFFERENTIATION 239
§9
Turning now to the amphibian nervous system, a number of ex-
periments show that the neural fold region is first determined as a
field.
In Triton, a dorsal half of an embryo, even at comparatively late
stages such as that of the early gastrula, will develop into a normally
proportioned little embryo. ^ In such cases the neural folds of the
miniature embryos are proportional to their reduced size, although
the material which the half contains is that from which a full-sized
neural plate would normally be formed.'^ If this experiment is
repeated at the late gastrula stage, the neural folds produced from
the half are full-sized, and therefore relatively too large. Irre-
versible determination of the neural fold region has therefore
taken place towards the close of the period of gastrulation (fig. 117).
Experiments of another kind, likewise on Trito?i, lead to further
results. Gastrulae can be produced which are deficient in median
material lying in the plane of bilateral symmetry. This may be
eflFected either by cutting out a median disc of tissue and sticking
the two lateral portions together again, or by making paramedian
cuts in two embryos, on the right of the mid-line in one and on
the left in the other, and then exchanging halves and sticking
them together, so that one composite embryo so formed will be
deficient and the other overprovided as regards median material.
In the case of the deficient embryos, the region involved includes
part of the presumptive neural fold tissue. Normally, of course,
this region is broad anteriorly in the region of the brain, and
narrow posteriorly in the region of the spinal cord. The operation
to which the deficient embryos have been subjected results there-
the lens is extirpated, a lens will be regenerated from the actual dorsal margin,
which was the presumptive ventral part of the eye-cup. But the original dorso-
ventral gradient of lens-forming potency persists, though masked by the in-
hibitory effect of the choroid fissure. For if sectors representing one-sixth of the
circumference of the iris of such eye-cups developed from rotated optic vesicles
be tested for lens-forming potencies, it is found {a) that the potencies of the
actual dorsal (presumptive ventral) sector are low, and (6) that there is little
difference between the potencies of this and other sectors. In other words, the
original dorso-ventral gradient of lens-forming potency and the ventro-dorsal
gradient of lens inhibition have here cancelled out. (Sato, 1933 ; see also Beck-
with, 1927.)
1 Spemann, 1903. - Ruud and Spemann, 1923.
240
THE MOSAIC STAGE OF DIFFERENTIATION
fore in partial or complete removal of the presumptive neural
fold tissue from the trunk region, and in removal of only a
portion of it from the region of the head. Such embryos develop
complete brains with properly proportioned eyes, but the spinal
Fig. 117
Loss of power of regulation in the development of dorsal halves of newt embryos.
A, The constriction and isolation of a dorsal half at the early gastrula stage leads
to' the development of (B) a diminutive embryo with neural folds of proportion-
ately reduced size. C, Normal neurula for comparison with B. D,The constric-
tion and isolation of a dorsal half at the late gastrula stage leads to the development
of (E) a diminutive embryo with disproportionately large neural folds, unable to
close. (From Ruud and Spemann, Arch. Entwmech. lii, 1923.)
cord may be absent. The important point to notice is that the
neural tube was already determined at the time of the experiment,
for the complete removal of its rudiment in the trunk region will
resuh in absence of spinal cord altogether. But the neural tube
was only determined as a whole and not in its detailed constituents,
Oc
241
Lai
^'^ Som
.Som
^^'Oc
D
Fig. 118
Regulation in newt embryos deficient in or overprovided with median material.
At the early gastrula stage, two embryos of Triton taeniatus are cut parasagittally
into slightly unequal lateral halves, the cuts passing on the right of the middle line
in one, on the left in the other. The larger right half is then stuck on to the larger
left half; the smaller right half on to the smaller left half. In spite of the excess
and deficiency of material lying in the plane of symmetry, both embryos
develop with normal proportions, thus showing regulation. A, " Large " embryo.
B, "Small" embryo: in this case, the neural fold material was partially and
not completely removed in the trunk-region. C, Transverse section through A,
D, Transverse section through B. Oc, eye; Lab, ear- vesicle; Som, somites.
(From Spemann and Bautzmann, E., Arch. Entiumech. ex, 1927.)
HEE 16
242
THE MOSAIC STAGE OF DIFFERENTIATION
for although the presumptive neural fold region has been much
reduced in the head-region, some of it is still present, and this
Fig. 119
Power of regulation in differentiation and growth of the eyes in Triton, a, Opera-
tion performed at the early neurula stage : a piece of the presumptive eye-region is
removed without interfering with the underlying tissues. Z>, Resulting larva
17 days after operation; left eye normal, right eye 1/3 normal size, c, The same
animal just metamorphosed, 103 days after operation; both eyes approximately
normal and equal-sized. (From Mangold, Ergebn. der Biol, vii, 1931.)
remnant has undergone regulation and has differentiated into a
complete and normally formed brain with its attendant structures.^
On the other hand, the overprovided embryos also regulate to
^ Spemann and Bautzmann, 1927.
THE MOSAIC STAGE OF DIFFERENTIATION 243
produce large but properly proportioned nervous systems; there
may, however, in some cases, be a certain amount of duplication of
the extreme anterior end (fig. 118).
In Urodeles, removal at the early neurula stage of a portion
of the region of the presumptive eye-rudiment does not prevent
the embryo from developing normally as regards its brain and
paired eyes^ (fig. 119).
All these experiments show that the neural fold region is deter-
mined as a whole at these stages : within the neural fold region there
also appears to be a determination along the antero-posterior axis
of the levels of the various constituent subregions, much as in the
limb. There is, however, no evidence that the neural fold region as a
whole ever passes through a totipotent phase in which any part of
sufficient size can regulate to produce a whole, as does the limb-
field.
It is to be noticed in the experiments mentioned above, in which
the presumptive neural fold region was removed altogether in the
trunk region and only reduced in amount in the head, that regula-
tion takes place within levels on the antero-posterior axis of the
embryo, but not along that axis. The neural fold tissue of the anterior
region regulates to form a brain, level for level, but it does not
regulate longitudinally to form brain and spinal cord, i.e. structures
characteristic of other levels.
As with the limb, ear, gill, and heart-fields, the nervous system
is polarised. The existence of this polarisation or gradient is shown
by the following experiment. If at the early neurula stage in an
embryo of Amhlystoma one presumptive eye-region is cut out,
rotated through 90° and grafted back again, the resulting embryo
possesses on the operated side an eye-cup which is subnormal and
deficient in its development and differentiation. Had this region
been extirpated completely, the remainder of the neural fold field
would have regulated to give rise to perfect eyes (see above). It
follows therefore that the region in question possesses a polarity
which at this stage presents obstacles to regulation if it is interfered
with. 2 The diflFerence from such a field as that of the limb is that
here the chemo-differentiation of subregions at diiferent levels
occurs at the first formation of the field, instead of later.
^ Adelmann, 1929, 1930; Mangold, 193 1 A. ^ Woerdeman, 1932.
16-2
244 THE MOSAIC STAGE OF DIFFERENTIATION
Progressive chemo-differentiation within the neural fold field
may best be illustrated with reference to the eye-region. As we
have seen (p. 243), experiments involving extirpation of the pre-
sumptive eye-region or of part of it at the early neurula stage in
Triton and Amblystoma have shown that the eye-field is more ex-
tensive than the region which in normal development actually gives
rise to the eyes.^ Eye-forming potencies, as tested by grafting,
are higher in the mid-line than more laterally.^ The reason that
a single median eye is not normallyformed depends on the presence
of the underlying gut-roof (see below). Further, regulation takes
place most readily across the transverse axis of the neural fold
region.^ At the same time, experiments of grafting portions of the
eye-region into other parts of the body show that the eye-rudiment
is already invisibly chemo-differentiated.* If the entire eye-field is
extirpated, no eyes are developed.^
At the same time, although the eye-region is determined as a
subregion of the neural fold field at the neurula stage, it is still
capable of regulation within itself, as is shown by those experiments
on Amblystoma at the neurula stage in which a portion of pre-
sumptive eye-tissue is grafted into the belly wall and differentiates
there into a more or less well-formed eye consisting of tapetum and
retina. Curiously enough, these eyes lack a stalk, although the graft
included the region which would normally have given rise to the
optic stalk in the intact embryo. It follows, therefore, that the
various regions of the optic complex — retina, tapetum, and stalk-
are not rigidly determined within the eye-area at the stage in
question in Amblystoma.^
In Pelobates (Anuran) at the tail-bud stage, grafted portions
composed of tapetum only can regulate to form little optic cups
with retina and tapetum in correct proportions.^ The optic stalk,
however, at this stage is already predetermined. In other Amphibia
it has been found that two eye-rudiments, grafted together, regulate
to form one.''
A complication in the analysis of the progressive determination
^ Woerdeman, 1929; Manchot, 1929. - Adelmann, 1930.
^ Mangold, 1928, 1931A.
^ Spemann, 1919; Spirito, 1928; Adelmann, 1929, 1930-
■^ Adelmann, 1930; Stella, 1932.
*' Dragomirow, 1932, 1933. " Pasquini, 1927.
THE MOSAIC STAGE OF DIFFERENTIATION 245
of the eye is introduced by the fact that the organiser, in the form
of the primitive gut-roof, underhesthe neural plate, and it now plays
a part in the further determination of the eye-regions. The action of
the primitive gut-roof in this respect has been tested by grafting
portions of the eye-area without and with the primitive gut-roof.^
It is found that the primitive gut-roof reinforces the eye-forming
potencies of the lateral portions of the eye-area, and, further, it
leads to the formation from median pieces of the eye-area of two
eyes, with optic stalks, and separated by a region of the floor of the
brain representing the optic chiasma ; whereas similar pieces with-
out gut-roof produce a single eye. The explanation of this "twin-
ning" effect of the gut-roof on the eye-rudiment is still to seek.
The gut-roof or organiser further accelerates the processes
leading to progressive subdivision of the eye-region into chemo-
differentiated subregions. This is illustrated by experiments on
Rana escidenta. If at the neurula stage in embryos of this species
a rectangular piece is cut out from the neural fold region including
part of the eye-area (together with the underlying primitive
gut-roof), rotated through i8o°, and grafted back again so that
the original anterior edge of the piece is posterior, the rotated piece
undergoes self-differentiation. The result in the later embryo is a
reversal of the normal order of the structures of the brain : the di-
encephalon with the epiphysis lies behind the mesencephalon with
its optic lobes. Anteriorly and posteriorly, these structures which
have developed from the reversed piece are continuous with the
parts of the brain which have developed undisturbed. It may be
noticed, therefore, that the diencephalon, epiphysis, and optic lobes
were already determined as subregions in the neural fold field at the
time of operation,^ and there is evidence that the infundibulum is
also determined.^
But the most interesting feature of this experiment is concerned
with the eyes. It is frequently found that there is a small pair of
eye-cups in the normal position, and another pair farther back,
situated either in front of or behind the ear-vesicles. The explana-
tion of this result is that the anterior cut by which the rectangular
piece was separated from its surroundings passed through the
^ Adelmann, 1930. 2 Spemann, 1912A.
Fig. 1 20
THE MOSAIC STAGE OF DIFFERENTIATION 247
presumptive eye-rudiments: a portion of these rudiments was
therefore left in situ, and another portion was included in the
rectangular piece. The portion left in situ developed into the eye-
cups in the normal position; the other portion after rotation de-
veloped into the eye-cups farther back. If the rectangular piece
was short, its hinder edge was in front of the ear-vesicles, and the
posterior pair of eyes then developed there. If on the other hand
the rotated piece was longer, so that its hinder edge was situated
behind the ear-vesicles, the posterior pair of eyes was behind the
ear-vesicles also^ (fig. 120).
These results show in the first place that the eye-rudiments were
determined, since they could go on diflPerentiating normally in
abnormal surroundings. In some cases the four little eye-cups are
not equal in size, but the sum of the sizes of the left front and right
hind eye-cups is equal to that of the right front and left hind cups :
the two eye-cups formed from one original rudiment divided by
the cut are, together, equal to one normal eye. This means, there-
fore, that the eye-rudiment is not only qualitatively but also quanti-
tatively determined, and that its topographical limits are now fixed.
The actual time of onset of this determination cannot be stated,
since the rudiment continues to be in contact with the underlying
gut-roof, and progressive chemo-differentiation probably proceeds
after the experiment. It is to be noted that in the experiment on
^ Spemann, 1912A.
Fig. 120
Mosaic development and self-differentiation of the brain and eyes in Rmia
esculenta. A, Dorsal view of the brain of a normal larva. B, Similar view of a
larva in which at the neurula stage a square piece of neural plate with underlying
gut-roof was rotated through 180° ; the asterisks mark the line of junction between
the rotated and non-rotated regions. C, Similar larva in which the piece
rotated was longer. Note that the rotated regions of the brain have continued to
develop by self-differentiation : the mid-brain lies in front of the epiphysis or fore-
brain. There are four eyes, owing to the fact that the cuts went through the eye-
region and, after rotation, parts of the eye-region found themselves behind the
normal position (B) or, if the rotated piece was long, behind the ears. The sum
of the sizes of the four eyes is equal to that of two normal eyes. In spite of their
reduced size, the subdivided portions of the-eye- rudiments have become rounded
into cups approximating to the normal morphological differentiation. D, Neu-
rula, showing the operation. Av, auditory vesicle; Bb, between-brain ; Cp,
choroid plexus; Ep, epiphysis; Fb, fore-brain; Hb, hind-brain; Hg, habenular
ganglion ; Mb, mid-brain ; Mo, medulla oblongata ; O/, optic lobe ; Ov. eye-cup ;
Pc, posterior commissure. (From Spemann, Zool.Jahrb. Siippl. xv (3), 1912.)
248
THE MOSAIC STAGE OF DIFFERENTIATION
Fig. 121
Self-differentiation of the eye and of its constituent tissues. The optic vesicle was
removed and grafted into the region of the ear in Rana palustris. The various
tissues undergo self-differentiation regardless of their morphological differentia-
tion, and of the proportions in which they are present. A, 19-day old graft,
showing absence of pigment cells; rods and cones project into the cavity of a
vesicle. B, 7-day old graft, showing excess of pigment cells. C, 5-day old graft,
showing absence of part of pigment layer; rods and cones projecting into coelom
of host. (From Lewis, Amer.Journ. Anat. vii, 1908, figs. 4, 5, 7.)
THE MOSAIC STAGE OF DIFFERENTIATION 249
Amblystoma above-mentioned, the eye-region when not underlain
by gut-roof is found to be still capable of regulation within itself
at the neurula stage. The results of the experiments on Pelobates
are apparently to be explained only on the supposition that chemo-
differentiation of the eye-region in this form occurs more slowly.
When these eye-cups are very small, it is frequently found that
they are abnormally proportioned in that they may have too many
or too few tapetum cells ; too much or too little optic stalk or retina,
compared with the proportions in which these constituents are
found in the normal eye-cup. The explanation appears to be that
the various constituents of the eye are now separately and in-
dividually determined. The cut, going haphazard through the eye-
rudiment, will often separate parts which possess the prospective
constituents of a normal eye in abnormal proportions. Similar
results are obtained from experiments on Rana palustris in which
incomplete eye-rudiments are grafted into Various positions. Here,
even the various retinal layers appear to be determined^ (fig. 121).
§ 10
Yet another conclusion of importance emerges from the experiment
mentioned above (p. 245). It w^ill have been noticed that although
some of the miniature eyes are abnormal in the proportions of their
constituent parts, they nevertheless round themselves off into little
spheres resembling normal optic vesicles, and some at any rate of
these little vesicles become invaginated to form cups. In some
cases, cups formed entirely of tapetum without any retinal tissue
are produced. In other words, the processes of morphological
diiTerentiation, or production of form, are not dependent on the
histological differentiation of the tissues which they are moulding.
A similar conclusion can be drawn from the results of other ex-
periments, in which it has been shown that portions of the neural
fold region, or of the heart, or gut region, roll themselves up into
tubes in spite of the fact that their histological differentiation may
be abnormal.^
At the same time, other work, and on the most diverse groups,
has shown that the histological differentiation of the tissues may
^ Lewis, IQ08.
" Roux, i885;Ekman, 1924; Stohr, 1925; Boerema, 1929; Holtfreter, 193 ib
250 THE MOSAIC STAGE OF DIFFERENTIATION
take place independently and in the absence of normal morpho-
logical differentiation. This may be seen in those experiments in
which a portion of the embryonic area of the blastoderm of the
chick is made to undergo development on the chorio-allantois of
another egg ;^ or in embryos of Cephalopods, the normal develop-
ment of which has been impeded by toxic agents. ^
The same conclusion emerges from the results obtained by
culturing in vitro various rudiments of the chick embryo, such as
those of the eyes, fore-limbs, and ears. In these cases, histological
differentiation may reach a high degree of perfection, while there
may be little or no approach to the morphological differentiation
of normal anatomy.^ (See also Chap, xi, p. 375.)
A pretty example of abnormal morphogenesis is seen in the
differentiation of reconstitution-masses of dissociated sponge cells
which contain an excess of collar-cell tissue. In this case, partial
spheres consisting of a single layer of collar-cells are produced with
the collars directed outwards instead of inwards, as in the normal
gastral lining.* (See p. 281.)
Purely morphological differentiation, then, seems to be in large
part conditioned by physical and mechanical factors of available
space, material, and pressure. Histological differentiation is in
large part independent of these factors. While these two kinds of
differentiation are sufficiently distinct during the later stages (i.e.
after their initial determination) for the one to take place without
the other, the question next arises as to what relation these two
kinds of processes bear to one another at the start.
The first visible important steps in differentiation are concerned
with the form-changes which result in gastrulation and neurulation.
These may be held to constitute a phase of morphological differ-
entiation, which, in development, is thus seen to precede histo-
logical differentiation. The question therefore arises as to whether
the latter is dependent on the former in the initial stages. If it were,
we should have another instance of the supposed effects of "dy-
namic determination ", referred to on p. 163. The problem therefore
presents itself as to whether it is possible to prevent a certain region
^ Murray and Huxley, 1925; Hoadley, 1924, 1925, 1926 A.
- Ranzi, 1928. ^ Strangeways and Fell, 1926; Fell, 1928.
* Huxley, 191 1; de Beer, 1922.
THE MOSAIC STAGE OF DIFFERENTIATION 251
of the embryo from passing through this early phase of morpho-
logical differentiation (mass movements at gastrulation and neuru-
lation), and then to see whether it is capable of undergoing
histological differentiation.^
It will be remembered that during gastrulation in Amphibia, the
presumptive neural fold material undergoes a translocation in a
particular direction for each piece of tissue, so that the material is
brought into position for the formation of the neural folds in the
neurula (see p. 25). A piece of presumptive neural fold tissue may
be grafted into the dorso-lateral region of another embryo in the
gastrula stage and orientated in such a way that the movements
of the host tissues in which it becomes involved are either directed
parallel or perpendicular to the direction in which the tissue would
have moved had it been left intact in situ. It is found that the
tissue differentiates morphologically into neural folds regardless
of its orientation and of the direction of the movements which
it has undergone. 2
It can be concluded from these experiments that specific form-
changes are not necessary for subsequent histological differentiation.
Other recent investigations of the histological differentiation of the
cell-regions in early embryonic stages of Triton have shown that
certain histological distinctions between presumptive epidermis and
presumptive neural fold are independent of form-changes. These
distinctions are already present at the earliest neurula stage. The
cells of the neural fold region are elongated, arranged in a single
layer, and have ellipsoidal nuclei ; their pigment is concentrated at
the outer end of the cells. The cells of the epidermal region are
cubical, arranged in two layers; the nuclei are spherical, and the
pigment is distributed irregularly.
If at the neurula or late gastrula stages (i.e. after the organiser has
been invaginated and underlies the presumptive neural folds)
pieces of presumptive neural fold tissue or presumptive epidermis
are grafted into atypical positions, they develop, as we have already
seen, by self-differentiation, and undergo the histological differ-
^ Goerttler, 1927. It may be noticed, however, that this experiment results
not only in forcing the piece of tissue to undergo abnormal movements, but it
also interferes with its polarity, which, as we have already seen (p. 243), plays
an important part in differentiation.
2 Holtfreter, 1933 a.
252 THE MOSAIC STAGE OF DIFFERENTIATION
entiation characteristic of their normal fate, but the morphological
differentiation is not always achieved. If, on the other hand, such
presumptive pieces are taken from an early gastrula (i.e. before the
organiser has been invaginated) and grafted, they will undergo the
morphological differentiation of their new surroundings: pre-
sumptive neural fold tissue in an epidermal area will remain flat,
while presumptive epidermis in the neural fold area will become
folded up into a neural tube. But, in spite of the morphological
differentiation which these pieces are forced to undergo, they retain
some of the histological characteristics of their normal prospective
fates. ^
In these experiments we have on the one hand the fact that histo-
logical differentiation can take place without morphological, and
on the other, the fact that morphological differentiation when forced
upon a piece of tissue does not entirely obliterate its presumptive
histological characteristics. It is necessary to conclude, therefore,
that in these cases, morphological and histological differentiation
are independent of one another.
There are other facts which point to the same conclusion. In the
larva of the sea-urchin, for instance, some histological differentia-
tions (apical organ) take place without any antecedent form-
changes of the tissue in question. In amphibian material, the
results of experiments involving the culture of pieces of tissue in
vitro likewise point to the independence of histological and morpho-
logical differentiation. We need only point to the instances men-
tioned above in Chapter iii (p. 50), in which pieces of tissue taken
from the blastula or early gastrula show far-reaching powers of
histological differentiation without having undergone any specific
form-changes, or any morphological differentiation.
Perhaps the most striking demonstration of the independence of
morphological and histological differentiation is provided by those
cases in which an insect tgg (Platycnemis) gives rise to two embryos
as a result of a transverse discontinuity in the blastoderm. The two
embryos develop, each from its ventral surface in the normal way,
and they are situated back to back. One is larger than the other,
and when it folds up its sides to form its dorsal surface, it actually
encloses its smaller brother within itself, and compels it to perform
^ Lehmann, 1928 b, 1929.
Mx
0
Sch.C/7
'Au
Au
^At
-M.Ch
SchwK
Twin embryos in the insect Platycnemis. The embryonic rudiment was split
into two unequal portions at an early stage ; the larger portion has produced an
apparently normal embryo, but within it (stippled) is the dwarf embryo produced
from the smaller portion, which is inside-out (see text). It has become enclosed
within the larger embryo as a result of the upgrowth of the sides of the latter ; the
direction of upgrowth is shown by the dotted line arrows. This process was
too strong for the sides of the smaller embryo, which were forced to follow suit
and to fuse ventrally instead of dorsally, thus enclosing the limbs within a closed
cavity lined by the outer surface of the epi4ermis. The ventral nerve-cord of the
second abdominal segment of the larger embryo is in contact with that of the
first abdominal segment of the smaller embryo, which thus appears larger.
For explanation of lettering, see fig. 60: capital letters refer to the larger
embryo, small letters to the smaller embryo. In addition: Abd^, ist abdominal
segment; Bm, nerve cord; F, fibre-tracts; R, dorsal wall. (From Seidel, Biol.
Zentralbl. xlix, 1929.)
254 THE MOSAIC STAGE OF DIFFERENTIATION
similar movements. But, for this smaller embryo, these movements
result in the folding and eventual fusion of its sides ventrally instead
of dorsally, since it is back-to-back with the larger embryo. Thus
the smaller embryo is inside-out : its limbs are contained in a closed
cavity lined by its body- wall which is completely inverted; its
organs and viscera lie outside its body-wall, and in contact with
those of its larger brother, inside which it is.^ In spite of these
form-changes being the reverse of normal, histological differentia-
tion continues as if nothing had happened (fig. 122).
While "dynamic determination", or the determinative eifects of
form-changes, may possibly be operative in the case of organisers
(see Chap, vi, p. 163), it does not seem that histological differentia-
tion in the mosaic stage of development is dependent on it.
§11
A special section may be devoted to the problem of the gonads
and sex-differentiation, which present many interesting features.
A full discussion of all aspects of the question has been given in
recent books such as The Development of Sex in Vertebrates'^ and
Sex and Internal Secretions ;^ accordingly here much controversial
detail will be omitted. Here, only such points as bear upon morpho-
genesis and the problem of differentiation will be dealt with, and
they in broad outline.
In general, the vertebrate gonad arises as what is doubtless a
special gonad-field on the dorsal side of the coelom. It first consists
of thickened coelomic epithelium (germinal epithelium) with some
underlying mesenchyme, together with primordial germ-cells. In
many vertebrates, these are undoubtedly differentiated precociously,
in most cases in the endoderm, and then migrate into the gonad-
rudiment. In other cases, especially among the higher forms, it
seems equally clear that germ-cells arise directly from the germinal
epithelium. It is possible that both these sources contribute to the
formation of germ-cells in many vertebrates.
Later, the gonad-rudiment becomes differentiated into an ex-
ternal cortex and a central medulla, but the details vary considerably
in different groups.
1 Seidel, 1929. - Brambell, 1930.
2 E. Allen, 1932.
THE MOSAIC STAGE OF DIFFERENTIATION 255
We may begin with the conditions in the Anura, which have been
very thoroughly investigated.^ The primordial germ-cells arise in
the dorsal region of the gut-wall, and then become separated from
the rest of the endoderm as a continuous ridge dorsal to the mesen-
tery. This ridge later divides into two. In these two genital ridges,
the germ cells are mixed with mesenchyme, and overlain by coelo-
mic epithelium which becomes slightly thickened. Later the core of
the ridges is invaded by the rete tissue, consisting of cords of cells
which appear definitely to grow out from the rudiment of the
mesonephros. The sexually undifferentiated gonad-rudiment is
now completely constituted, and consists of two portions, a peri-
pheral cortex composed of coelomic epithelium and primordial
germ-cells with associated mesenchyme, and a central medulla
derived primarily from the mesonephros. The cortex is broadly
homologous with the germinal epithelium of Amniota.
Sexual differentiation now occurs. In the female the cortex en-
larges, its contained germ-cells develop into oogonia and oocytes ;
meanwhile the medulla ceases growth and develops into epithelial
ovarial sacs. In the male, the rete cords of the medulla continue to
proliferate, and are invaded by the germ-cells, which leave the
cortex and migrate inwards, then proceeding to differentiate into
spermatogonia. Later the rete cords produce, among other struc-
tures, the non-germinal portions of the testis tubules. Meanwhile
the cortex, after losing its germ-cells, becomes reduced to a thin
peritoneal covering.
The evidence appears conclusive, first, that the type of sexual
differentiation of the indifferent gonad-rudiment is normally
dependent on its genetic sex-constitution, although, as we shall see
later, this can be overridden by other agencies. The case is like that
of any other mosaic differentiation, except that the gonad-field has
one of two potentialities open to it, according to the sex-chromo-
somes which it contains. Secondly, the primordial germ-cells
appear to be completely bi-potential as regards sex. What they shall
become is determined by local influences emanating from the region
in which they come to lie. In the cortex they become female, in the
medulla, male. In other words, their differentiation is dependent.
Temperature exercises a differential effect upon the cortex and
^ Full references in chapters by Willier and by Witschi in E. Allen, 1932.
256 THE MOSAIC STAGE OF DIFFERENTIATION
medulla. Low temperature causes a differential inhibition of
medullary development with consequent delay in males of the
degeneration of the cortex and of the immigration of the primordial
germ-cells from it to the medulla. As a result, the primordial germ-
cells, exposed to cortical influences, become oogonia, and 40 mm.
tadpoles are all somatically females. Later, however, in the genetic
males among them, the delayed medulla succeeds in reaching the
stage of development requisite to inhibit the cortex, upon which they
become transformed into somatic males.
High temperature, on the other hand, has a deleterious effect
upon the cortex, but not upon the medulla. As a result the sexes
are early differentiated in the normal i : i ratio, but later the females
show inhibition of the cortex. No further oocytes are differentiated,
those already embarked on differentiation degenerate after a short
period of further growth, the ovarian sacs derived from the medulla
begin to proliferate and form cords, and any undifferentiated pri-
mordial germ-cells migrate inwards and join the medullary cords,
where they differentiate into spermatogonia; thus the genetic
females become transformed into somatic males. Essentially
similar results have been obtained in Urodeles.
These experiments clearly demonstrate the existence of local
sex-inductive agencies in cortex and medulla respectively. The
nature and action of these factors is more fully revealed by a series
of beautiful experiments on parabiotic twins. ^
Amphibian embryos are united parabiotically either side by side
(parallel pairs) or in series with the anterior end of one joined to the
posterior end of the other (chains). This has been effected both
homoplastically, between partners of the same species ; or hetero-
plastically, between different species. Here we shall confine our-
selves to homoplastic parabiosis, and to two-sexed pairs, in which
the partners are genetically of opposite sex.
The most interesting experiments concern frogs {Rana). In
these, no effect on sex-differentiation is exerted in chain pairs : sex-
differentiation is normal both in the male and the female partner.
In parallel pairs, however, the sex-differentiation of one member of
the pair is modified. The affected partner is normally the female,
and the modification consists in a certain degree of inhibition of
^ Witschi, 1932.
257
Fig. 123
Antagonistic sex-differentiation in Amphibia. Diagram of parabiotic twin pairs
of unlike sex in toads (left), frogs (centre) and Urodeles (right). Above, parallel
pairs; below, chain pairs. In the gonad: male differentiation, black; female
differentiation, white. In toads, there is no effect of one twin upon the other
(note Bidder's organ at the anterior end of the gonad in both sexes). In frogs,
there is no effect in chain pairs ; but in parallel pairs, the male gonad affects the
gonads of the female partner, the effect diminishing with distance (degree of
shading of circles). In Urodeles (Trm/m^), the male gonad completely inhibits the
female gonad in both types of twin pairs. (Based upon Witschi, Chap, v in
E. Allen, Sex and Internal Secretions, London and Baltimore, 1932; modified.)
HEE 17
258 THE MOSAIC STAGE OF DIFFERENTIATION
female-differentiation combined with a certain degree of encourage-
ment of male-differentiation, resulting in progressive transforma-
tion away from the somatic female towards the somatic male-type
of gonad.
The most unexpected result is that the effect always manifests
itself first in the " inside " ovary, i.e. that nearer to the male partner,
and always on the inside margin of this ovary ; from here it gradually
spreads, but with diminishing intensity, to the more distant
regions. It appears clear that the medulla of the male produces a
substance which not only promotes masculine sex-differentiation of
germ-cells, but is also antagonistic to cortical development: further,
that this substance is not strictly localised but can diffuse outwards
in diminishing concentration and with diminishing effectiveness.
In chain-pairs, the gonads of the female partner lie beyond the
limit of effectiveness; in parallel pairs, they lie across a rapidly
decreasing concentration-gradient (fig. 123).
Comparative studies on other forms provide further striking
results. In toads, no effect is ever observed on the gonads of either
partner, whether in parallel or chain-pairs. In the Urodele Triturus,
however, the effect is equally marked in both kinds of combination.
(The details here are slightly different: there is a long period of
mutual inhibition, in which both male and female gonads are
delayed and rendered nearly sterile. Later the male recovers, and
reduces the female gonads still further to small rudiments almost
free from germ-cells. There is no male transformation of the
genetic females. In other Urodeles the effect is similar in affecting
the female partner equally in chains and in parallel pairs, but
neither the mutual antagonism nor the final inhibition of female-
differentiation are so extreme.)
It would thus appear that the morphogenetic (inductive) sub-
stances produced by cortex and medulla are in toads strictly
localised within the regions where they are produced, and wholly or
almost incapable of diffusion. This is borne out by the existence in
toads of Bidder's organ, an anterior section of the gonad of ovarian
character, which develops from a portion of the gonad-rudiment
consisting wholly of cortex.^ This could not well develop, as it does,
in males if the medullary substance could diffuse even a short dis-
1 Witschi, 1933 B.
THE MOSAIC STAGE OF DIFFERENTIATION 259
tance during ontogeny. In frogs, on the other hand, the inductive
substances must be capable of moderate diffusion. The effect here
recalls the graded distribution of limb-forming potencies in the
limb-field of Urodeles (p. 223), and the probable diffusion of in-
ductive substances from the presumptive dorsal lip region during
cleavage (pp. 134, 311).
In Urodeles, on the other hand, diffusion is so complete that there
is no evidence of any concentration-gradient. It is possible, though
not demonstrated, that here the substance diffuses into the blood-
stream and is transported by it. We have thus within the boun-
daries of one class of vertebrates either a complete or a nearly com-
plete gradation between sharply-localised morphogenetic substances
and freely-circulating hormones. It has indeed been suggested that
the sex-hormones of the adult gonads are identical or homologous
with these morphogenetic substances produced by the embryonic
cortex and medulla, merely differing in being secreted into the
blood-stream instead of soaking through the tissues.
In support of it we find indications in cases of hermaphroditism
or asymmetrical development of gonads that the accessory sex
characters (male and female ducts), whose differentiation is known
to be dependent upon sex-hormones, are locally better developed
in regions of greater development of the gonad of corresponding
sex.i While no certainty can yet be reached on this point, it is a
valuable suggestion to guide further research. In any event, it is
clear that the physico-chemical conditions regulating diffusibility
of morphogenetic substances are of great importance in ontogeny.
In Urodele parabiosis the failure of the medulla of genetic female
gonads to differentiate in the male direction after regression of the
cortex under the influence of the male partner is in marked contrast
with the results in frogs. It appears to be general in the subclass. ^
No adequate explanation is yet forthcoming : in general, it appears
to link up with the subject of metaplasia (p. 211). The medulla of
all female Amphibia differentiates in a specifically female direction ;
that of the Anura retains its bisexual potency, and is capable of
metaplasia and male histo-differentiation ; that of the Urodela loses
the original bisexual potency and is capable only of continued
development or of regression within the limits of female-type
1 Witschi, 1933 B. 2 Witschi, I933 A.
17-2
26o THE MOSAIC STAGE OF DIFFERENTIATION
potency. The progressive restriction of potencies during ontogeny,
and its variation between types of tissue and types of organism,
remains a central problem for developmental physiology.
Two further points may be mentioned. Occasionally in Anuran
parabiotic twins, the female partner may obtain an unusually early
start. In this case, the male-differentiation of the male partner is
inhibited and a female phase may be passed through. This however
is only to be seen in the parts of gonads nearest to the female
partner, and is transitory, normal male-differentiation eventually
gaining the upper hand again.
Secondly, in heteroplastic parabiosis between two species of
frogs, a curious effect is visible in pairs in which both partners are
genetic females, and in which one member
belongs to the species Rana sylvatica. The
sylvatica ovaries hypertrophy, those of the co-
twin become reduced and degenerate. This
can be explained on the assumption that some
substance necessary for ovarian growth is
present in limited quantity in the embryo, ^
and that the faster-growing sylvatica ovaries
obtain a disproportionate share of it.^ In rare C-
cases, the reduced ovary of the co-twin may
even show some changes in the direction of
male transformation. This may be explained Fig. 124
on the hypothesis of antagonism between Diagram of the two
cortex and medulla; when the cortex is in- Sramprbkn^gonad-
hibited by being starved of the substance rudiment. M, medulla,
necessary for its growth, the medulla is re- ^^2^°"''^^^- ^^^ J^^^^"
•' .,.,., differentiation ; C, cor-
leased from mhlbltion. tex, responsible for
Thus we come to the general conclusion female -differentiation.
that sex-differentiation in Amphibia is under \^ in^E.^ Allen, S^x
the control of substances provided by the a7id Internal Secretions,
gonad cortex and medulla respectively: that Baltimore and London,
^ • -r: 1 1932- Modified.)
these substances are not speci6s-specinc : that
they are mutually antagonistic : that they are capable of various
^ The same hypothesis will account for the fact that removal of the gonad
proper in either sex in toads leads to the hypertrophy of Bidder's organ to form
a functional ovary. The substance in question is very possibly a hormone pro-
duced by the pituitary. See Witschi, 1933 b.
THE MOSAIC STAGE OF DIFFERENTIATION 261
degrees of diffusion from the regions where they are produced : that
in their action at a distance on another developing gonad the
inhibitory effect of each is the primary or at least the stronger,
and that the male-differentiating substance normally develops
earlier and is more potent than the female-differentiating sub-
stance. It is also possible that they are or become converted into
the sex-hormones of the adult.
Further highly interesting results have been obtained as the
result of fertilising over-ripe eggs in frogs. ^ Below a certain degree
of over-ripeness (about 3 days), no effects of any kind are to be
noted in the resultant embryos. Beyond this critical point, the
following main effects appear, all of them increasing with the degree
of over-ripeness. First, a conversion of a certain proportion of
genetical females into somatic males. The proportion is at first
small, but finally, with eggs rather over 4 days over-ripe, all-male
offspring are produced. Meanwhile a certain degree of delay in
development is noticeable, and with increasing over-ripeness de-
fects of development and abnormal mortality also appear, cul-
minating in death of all embryos at an early stage when the eggs
are about 5 days over-ripe. The defects of development have
already been noted in Chap, v (p. 96); they arise mainly after
4-5 days' over-ripeness, and manifest themselves chiefly as a
failure of coordination, leading to abnormal cleavage, production of
double monsters, development of teratological outgrowths, and in
extreme cases the production of abnormal tumour-like structures
which partake of many of the characteristics of truly malignant
growths. Comparable phenomena have been observed in trout.^
The sex-transformations are of peculiar interest, since the whole
morphogenesis of the gonad-rudiment is modified. In extreme
cases^ the rod-like area of primordial germ-cells does not become
detached from the endoderm, and the gonad-rudiment at its first
appearance is a rudimentary fold containing no germ-cells (in such
cases the germ-cells appear later to migrate into it, but how this
occurs is not established). In other cases'* the germ-cells while still
in the endoderm become abnormally pigmented. The gonad-
rudiment in highly affected specimens passes through a female
^ Willier, 1932; Witschi, 1932. " Mrsic, 1923, 1930-
3 Kuschakewitsch, 19 10. * Witschi, loc. cit.
262
THE MOSAIC STAGE OF DIFFERENTIATION
phase followed by degeneration of the cortex and growth of the
medulla leading to male transformation as in the high temperature
experiments: in extreme cases, the cortex is inhibited from the
outset and sexual diiferentiation is male throughout. In general,
Fig. 125
The effect of late fertilisation on gonad-differentiation in frogs. Left, two stages
in the normal morphogenesis of the gonad. Above, genital ridge with primordial
germ-cells and mesenchyme. Below, later stage with peripheral cortex containing
primordial germ-cells and central medulla derived from the invading rete-cords
of nephrogenous origin. Right, corresponding stages in embryos from late
fertilised (over-ripe) eggs. Above, very small genital ridge with no primordial
germ-cells . Below, later stage with invading medulla (rete tissue) but rudimentary
cortex, with no primordial germ-cells. (Redrawn after Kuschakewitsch,
Festschr.f. R. Hertwig, 1910, vol. ii.)
testis- differentiation is accelerated^ (fig. 125). The thyroid of late-
fertilised frogs is also hypertrophied.^
We have thus a series of effects with progressive degrees of
over-ripeness. First, minor upsets of morphogenesis, notably in
regard to sex. Secondly, more general upsets of morphogenesis,
^ Eidmann, 1922.
- Adler, 1917.
THE MOSAIC STAGE OF DIFFERENTIATION 263
notably partial twinning. Thirdly, teratomorphic and malignant
effects. It is interesting that, as Dr Waddington has pointed out
in conversation, carcinogenic compounds are certainly related to
oestrin, and are probably to the chemical substance responsible
for organizing (p. 154). So our three effects, concerned with
sex, wath organiser abnormality (twinning), and with malignancy,
may conceivably all be related to one fundamental process af-
fecting substances of this type.
It is interesting that keeping the eggs for some time before fer-
tilisation in conditions of relative lack of w^ater (hypertonic salt
solutions ; keeping in air with a minimum of moisture) leads to a
large preponderance of females.^ Unfortunately no embryological
study of this case has been made, but it too is evidence that condi-
tions in the egg-cytoplasm during the earliest stages of development
may modify morphogenesis at later stages : since sex-differentiation
provides two alternative methods of morphogenesis in which the
result is determined by a balance of two competing factors, w^e
should expect to find in it the best indicator for such effects.
In Amniotes, the embryology of the gonad is not so simple. In
general, however, we may say that the distinction between mascu-
linising medulla and feminising cortex is maintained. The medulla
is largely formed by the primary sex-cords w^hich migrate inwards
from the germinal epithelium; in the male these form the testis
tubules, in the female they become inhibited and persist in modified
form. In the male, only one set of sex-cords is formed. In the
female, however, the cortex enlarges to produce a second set, which
gives rise to the main ovarian structures, in association with which
the female germ-cells differentiate. In the male, on the other hand,
the cortex (germinal epithelium) becomes reduced to a mere peri-
toneal epithelium soon after the formation of the primary sex-cords.
The chick is here the best-investigated type. In the chick the
germ-cells are formed in the extra-embryonic endoderm, in a
crescent-shaped area of the blastoderm antero-lateral to the embryo.
Embryos can be castrated by excision of this area, or by ultra-
violet^ or X-rays,^ proving the mosaic determination of the pri-
mordial germ-cell tissue. Normally these cells appear to be at-
tracted into the mesoderm when it invades the germ-cell field, and
^ King, 1912. - Reagan, 1916; Benoit, 1930. * Danchakoff, 1933.
264 THE MOSAIC STAGE OF DIFFERENTIATION
thence into the blood-stream. After being found in all parts of the
body for a considerable period, they become localised during a few
hours in the site of the future gonad-rudiment. Presumably the
gonad-field, once determined, attracts the germ-cells chemically.
By various lines of evidence it has been shown that a gonad-
field is determined and will differentiate into a gonad with typical
sex-cords even in the total absence of primordial germ-cells, and
that conversely in chorio-allantoic grafts germ-cells may be present
in considerable numbers without giving rise to a gonad. There is,
however, evidence that the germ-cells can induce some degree of
early gonad-differentiation upon peritoneum which would normally
never give rise to germinal epithelium. It is thus probable that the
germ-cells are necessary for, or at least normally assist in, the
process of gonad-differentiation, but that this effect can only be
exerted within a "gonad-field" region of the dorsal coelomic
epithelium, whose potencies are highest in the presumptive gonad-
region.
In passing, it should be mentioned that, in some mammals at
least, germ-cells appear to arise in situ at comparatively late stages
in the already differentiated germinal epithelium. Here early gonad-
differentiation cannot be dependent in any degree upon the presence
of germ-cells.
The transition between a state of aiTairs in which an early deter-
mination of germ-cell tissue occurs in the endoderm and that in
which a late determination occurs in the germinal epithelium is not
easy to envisage, but we have at least analogies with the determina-
tion of other organs, i.e. the lens (p. 189), or the limbs of Amphibia,
which in Urodeles are differentiated very early and are independent
of thyroid action, while in Anura they differentiate later and will not
display full growth in the absence of a certain concentration of
thyroid hormone. Further work is needed to elucidate this point.
The ovaries of birds and of monotremes are of course asym-
metrical, that on the left being large and functional, that on the
right reduced, and non-functional. It is interesting to find in the
bird that this difference is determined from the outset of differen-
tiation of gonad-rudiment not as yet showing any sign of sexual
differentiation. When grafted on to the chorio-allantois of other
embryos, indifferent gonads of either side may differentiate into
THE MOSAIC STAGE OF DIFFERENTIATION 265
testes, presumably in the case of grafts from genetic males, since
testes may develop on either male or female hosts. When they
differentiate into ovaries, only a left rudiment will form a true
ovary ; the right rudiment will only develop the rudimentary ovarian
structure typical of the right ovary. This difference must be deter-
mined in relation to the original asymmetry gradient of the embryo.
The different initial determination of right and left gonads is
further shown by another experiment. If all the germ-cells are
destroyed during the first half of the incubation period (which can
be accomplished by X-rays, owing to the high susceptibility of the
germ-cells to this agency) , the subsequent complicated differentiation
of the non-germinal portions of the gonad will continue, leading
to the formation, shortly before hatching, of testes, functional
(left) ovaries, and non-functional (right) ovaries, which are of
typical structure except for being sterile.^
Grafting and other experiments have also elicited other interesting
facts. In the first place, the grafting operation, and, still more, brief
exposures to low temperatures soon after visible sex-differentiation
has begun, favour the persistence of structures which normally
atrophy during development, such as the right Miillerian duct (ovi-
duct) of females, and both Miillerian ducts in males. The percent-
age of survival of such structures is raised by low temperature from
about 18 per cent, found in controls to over 70 per cent.^
Low temperature is known to inhibit or retard many develop-
mental processes: it would appear either that it has a specially
strong effect on processes leading to the reduction of organs, or that
since these processes, as shown by the 18 per cent, of persistence
in controls, are unusually labile, slight alterations in conditions
will cause large changes in their results (see also Huxley, 1932,
Chap. VI, 8).
Another interesting fact is that the capacity of grafted portions of
the gonad-field for differentiation increases with their age when
grafted. Before the time of visible differentiation of germinal
epithelium, few or no grafts of this area give a gonad at all. A little
later, gonad-like bodies of uncertain^ sex are produced. If the graft
is taken still later, when it includes gonads which are well-formed but
still microscopically undifferentiated as regards sex, sex-different-
1 Danchakoff, 1932. - Willier, 1932.
266 THE MOSAIC STAGE OF DIFFERENTIATION
iation either in tlie male or female direction occurs in the grafts, and
its completeness and frequency (as well as the size of the resultant
gonad) increases with the age of the grafts. It would thus appear that,
as a result of processes occurring in the embryo as a whole, gonad-
detemiining substances tend to accumulate in liigher concentration
in the presumptive gonad-held during, and presumably for a short
time before, the stage of its early differentiation. This is perhaps
somewhat parallel \^-ith the intensification of organiser potencies in
the grev crescent area during the time from fertilisation to gastrula-
tion (p. 68) and with the progressive capacity* of the lens of the
bull-frog for self- differentiation after detemiination (p. 189). It
would be of great interest to discover whether a similar state of
affairs can be detected for other rudiments in other groups. A
parallel increase of differential potency in whole embr\os and
large fragments vni\ be discussed in the next section.
The remarkable condition of the avian right ovary is correlated
with a marked regression of the cortex after its first formation. The
capacit}- for this must, as we have seen, be determined intrinsically
within the right gonad - ru diment ; but one of the results of this
primar}- differentiation as a right ovar\^ appears to be sensitivity^ to
substances emanating from the left oysltv ; for when the left ovary
is removed, the right hypertrophies. What it shall then produce is
determined bv the degree of degeneration which its cortex has
previouslv undergone. If considerable cortex is still left, this
dominates and it becomes a gonad of true ovarian type ; if less cortex
is left, both it and the medulla participate in the hypertrophy,
forming an ovo-testis. If the cortex had completely regressed, the
medulla hypertrophies and it forms a testis. The germ-cells
normally disappear from the right ovar}- in the first month after
hatching. If the hypertrophy takes place later, the resultant gonad is
sterile ; if earlier, as in ver\- early left ovariotomy, spermatogonia
may be formed.
In conclusion, the special interest of the gonads for our purpose
lies in the fact that two alternative paths of differentiation lie open
to them, the actual path taken being normally first decided by the
sex-chromosome mechanism, and implemented by a quantitative
balance bet^veen two mutually antagonistic male-differentiating and
female-differentiating substances, locally produced by the two main
THE MOSAIC STAGE OF DIFFERENTIATION 267
regions of the gonad-rudiment, formed in different quantities, at
different rates, and with different capacities for diffusion. The
result is a bipotentiahty, and therefore a labiUty of differentiation
which makes the gonad especially interesting for a study of the
effects of external agencies upon morphogenesis.
§12
Problems of possibly a special nature are presented by experiments
of grafting portions of chick blastoderms on to the chorio-allantois
of other embryos. In general, it seems that the power of histo-
logical differentiation of a piece of blastoderm is conditioned by its
size, and by its age at the time of its isolation from the rest of the
blastoderm. An entire unincubated blastoderm, when grafted, will
show a degree of histological differentiation approximating to that
found in normal chick embryos which have been developing for the
same length of time. As mentioned above (p. 250), the morpho-
logical differentiation of such a piece may be very abnormal indeed.^
Small pieces, representing about one-fifth of the area of a whole
blastoderm, are restricted in their powers of histological differ-
entiation. Pieces of unincubated blastoderms grafted on to the
chorio-allantoic membrane will differentiate only into epidermis
and gut. Pieces cut from blastoderms incubated for 2 hours will,
in addition, differentiate into nervous tissue. After 4 hours' in-
cubation of the blastoderm, pieces cut from it will produce brain,
eye, cartilage, and muscle. After 10 hours, corium and feather-
buds are formed."^ It may also be mentioned that the earlier a piece
is isolated, the smaller are the organs which it forms. ^
It is thus apparent that the older a piece of tissue is at the time
of grafting, the better it will differentiate. This is especially well
shown in the case of the eye. A 4-hour piece will produce an eye
consisting of pigment cells only ; a 6-hour piece gives an eye differ-
entiated into pigment cells and retinal cells. After 8 hours, the
various layers of the retina are differentiated, while complete self-
differentiation of the eye is obtained from pieces cut from blasto-
derms that have been incubated for 33 hours. In the case of the
mesonephros, 4-hour pieces give secretory tubules, 6-hour pieces
1 Murray and Selby, 1930. - Hoadley, 1926 a.
^ Hoadley, 1929,
268
THE MOSAIC STAGE OF DIFFERENTIATION
also produce glomeruli, lo-hour pieces differentiate a Wolffian duct ;
older pieces give complete self- differentiation of the mesonephros.
Feather-buds present the same picture.^
Other experiments have been performed in which the pieces cut
from the blastoderm were larger, representing one-third instead
of one-fifth of the whole area. While further results are desirable,
it seems from those already obtained that these larger pieces differ-
entiate more fully than the smaller ones, thus indicating that the
size of the piece is also a factor in the capacity of tissues to undergo
histological differentiation. ^
It will be remembered (see Chap, vi, p. i6o) that in the chick
blastoderm there is a gradient of developmental potencies at differ-
ent levels along the axis, and this must be taken into account in
interpreting the results of grafts of portions of blastoderms.^
A similar gradient of potency for differentiation has been noted
in the case of portions (thirds) of ii-day rat embryos grafted on to
the chorio-allantoic membrane of chicks, and can be stated in
tabular form.^
Tissue
differentiated
In grafts from
Anterior
one-third
Middle
one-third
Posterior
one-third
Nasal sac
Brain tissue
Hair follicles
Epidermis
Cartilage
Bone
Mesonephros
Gut
Present
Present
Present
Present
Present
Present
Absent
Absent
Absent
Absent
Present
Present
Present
Present
Present
Absent
Absent
Absent
Absent
Present
Present
Present
Present
Present
These cases in the rat are perhaps hardly comparable to those in
the chick, owing to the difference in age of the fragments tested.
Until further information is obtainable concerning the existence
and possible spread of a labile determination in these forms, it is
hazardous to attempt an interpretation of these cases.
Hoadley, 1924, 1925.
Willier and Rawles, 193 1 b; Hunt,
[932.
2 Murray and Selby, 1930.
^ Hiraiwa, 1927.
THE MOSAIC STAGE OF DIFFERENTIATION
269
§ 13
A word may be said as to certain problems of determination which
occur in later stages. As an example, we may take the case of the
spurs of fowls, grafted into hosts of the same or opposite sex, a few
days after hatching.^ As might be expected, grafts into hosts of the
same sex as the donor develop in the way characteristic for that sex,
Fig. 126
Differential behaviour of juvenile male and female spurs grafted into young
female fowls . The two legs of a hen 1 8 months old into which, when newly hatched,
two female spurs and one male spur had been grafted from day-old donors ;/. the
two grafted female spurs have remained the same size as the control spur (c.)
which has developed on the host; m. the grafted male spur has enlarged to the
dimensions characteristic of a spur on a normal cock. (Redrawn after photo in
Kozelka, jfoiirn. Exp. Zool. LXi, 1932.)
remaining rudimentary in the female, but attaining a large size in
the male. Whereas, however, female spurs in a male host are
capable of male-type development (although, owing to an inhibi-
tory effect of male environment on female tissues, this is not
universal), male spurs on a female host regularly develop masculine
^ Kozelka, 1932, 1933.
270 THE MOSAIC STAGE OF DIFFERENTIATION
size and other male-type characteristics. It would thus appear that
male-type development has been already determined in the spur-
rudiments of the young male chick, although these are still very
small. The spur-rudiments of the female chick, on the other hand,
are in a labile, undetermined state. Whether the determination in
the male has been eiTected by the testis hormone acting on the
rudiment is not known. Against this is the fact that female spurs
grafted to male hosts, and then after varying periods up to 24 days
replanted in the original (female) donors, do not show male-type
growth. Possibly only the embryonic rudiment can be sensitised by
male hormone. The alternative hypothesis of different reactivity of
ZZ (male) and ZO (female) tissues must also be included (fig. 126).
In contradistinction to this case, it is known that in many ver-
tebrate organs, exhibiting sexual dimorphism in size and other
characters, the sex-differences are only maintained so long as the
hormones responsible are acting upon them (see, e.g., Goldschmidt,
1923). The whole problem of the time-relations of determination
exerted by hormones, and of its reversibility, needs careful experi-
mental analysis.
Chapter VIII
FIELDS AND GRADIENTS
§1
If a simple animal such as a Planarian is cut transversely into two
pieces, normally the front piece will form a tail at its hind end, and
the hind piece will form a head at its front end. But if the trans-
verse cut had been made a short distance farther back in the body,
those cells which in the previous experiment belonged to the hind
piece and proliferated to form a head, will now belong to the front
piece and will proliferate to form a tail. Therefore the determina-
tion of the quality of the structure which is formed cannot be
based on any localisation of specifically different materials or
substances, for, if so, it would be impossible to understand how
either a head or a tail can be formed from identically the same
tissues. How, then, is the quality of the structure which will be
formed determined?^
A situation in some respects comparable with that just described
occurs in the regeneration of the limbs of newts. An amputated
limb gives rise to a regeneration-bud, from which an arm or a leg,
as the case may be, is eventually formed. These structures can be
easily distinguished by the number of digits and other criteria.
But at the outset of this process of regeneration there is no
qualitative determination of arm-forming as opposed to leg-
forming material in the regeneration-bud, for an arm regeneration-
bud can be grafted on to the stump of an amputated leg, where it
will develop into a leg, provided that the operation is performed
soon enough after the amputation of the arm and the formation of
the arm regeneration-bud. The converse experiment of grafting a
leg regeneration-bud on to the stump of an arm leads to the for-
mation of an arm under the same conditions.^
The tissue regenerated by an arm or a leg is at the outset not even
determined to produce a limb. The early regeneration-bud of a
^ See also J. Loeb, 1912. - Milojewic, 1924.
272 FIELDS AND GRADIENTS
limb grafted on to the base of a tail actually produces a little super-
numerary tail (fig. 127).
The undetermined stage of the regeneration-bud is of limited
duration. Whereas a bud of hemispherical form is still undeter-
mined, by the time a markedly conical shape has been attained, the
bud is determined, and if grafted elsewhere will now continue to
differentiate in accordance with its place of origin instead of in
accordance with its new situation. In this respect regenerated tissue
behaves just as do the various regions of the amphibian egg, which
also pass from a plastic to a determined phase.^
Fig. 127
Lack of determination in early regeneration-buds. Triton larva showing a tail
(against a square of white paper) developed from an early limb regeneration-bud
grafted into the tail-field. (From Guy^not, Rev. Suisse de Zool. xxxiv, 1927.)
The success of the converse experiment in which the early re-
generation-bud of a tail is grafted on to the stump of an amputated
hind-limb, or into the fore-limb field, close to the base of the (un-
operated) host-limb, and then produces a limb,^ has also been
reported, but this, though highly probable, cannot be regarded as
conclusively proved^ (fig. 128).
^ Guy^not, 1927; Guyenot and Ponse, 1930.
" Weiss, 1927 B. ■
^ The experiment was done with regeneration-bud and host belonging to the
same species, and it is difficult therefore to be absolutely certain that the limb
developed from the grafted cells. Further, as pointed out by Guyenot, the graft
may have come under the influence of the endings of the brachial nerve, which
are known to be able to produce the formation of a limb (see p. 362). However,
the presumption is that Weiss' interpretation is correct.
FIELDS AND GRADIENTS 273
The results of the regeneration experiments, as well as those con-
ducted on embryos undergoing embryonic development, agree in
demonstrating two important points. The first is that tissue which
is about to differentiate into a given structure is at the outset un-
determined, and therefore capable of differentiating into other
structures, of wholly different type. The second is that the actual
decision as to the fate of such undetermined tissue rests with its
position relative to some major system. In the amphibian egg, the
determining factors are the level of the tissue along the main egg-
axis, and its distance from the organiser region (see p. 139)- In the
Fig. 128
Lack of determination in early regeneration-buds. The smaller limb here shown
(right) was produced from the early regeneration-bud of a Triton tail grafted on to
the stump of a hind-limb. Left, unoperated hind-limb of other side. (From
Wells, Huxley and Wells, The Science of Life, London, 1929, after Weiss.)
case of the Planarian cut transversely, the determination of the
pieces of freshly regenerated tissue are controlled in relation to the
polarity of the whole organism : front edges of hind halves produce
heads, hind edges of front halves produce tails. In regeneration in
newts the type of differentiation is controlled by the local environ-
ment of the regeneration-bud; this exerts qualitatively different
effects in different regions of the body (e.g. region of leg as against
region of tail). The material of the ^arly regeneration-bud is in-
different. So far, it has been shown that its capacities of differentia-
tion include organs of such different type as limb and tail ; it would
be of great interest to determine whether it was so completely
HEE 18
274
FIELDS AND GRADIENTS
undetermined as to be able to produce any structures, internal or
external.
On pushing analysis further, 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
Fig. 129
Gradients of various kinds in the earthworm Pheretima. O O and outer
left-hand scale, oxidisable substance as determined by the Manoilov reaction.
X X and outer right-hand scale, solid content, per cent. V V and inner
left-hand scale, temperature at which heat-shortening occurs. • ©and inner
right-hand scale, electrical potential, millivolts. (Redrawn after Watanabe, Set.
Rep. Tohokii Imp. Univ. vi, 193 1.)
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.^
^ Historically, concepts of this type were first introduced into embryology by
Boveri (1901, 1910) ; later Child (191 5 a) generalised a large number of observa-
tions in the form of his theory of axial gradients ; the term field, however, was only
introduced in the last few years, notably by Spemann ("Organisationsfeld")
(1921), Gurwitsch (1922, 1927), Weiss (1927 c), Bertalanffy (1928), de Beer (1927).
FIELDS AND GRADIENTS
275
%
fer
li
'/•■vh
I
p
f
Fig. 130
Differential susceptibility in the primitive oligochaete Aeolosoma. Above, four
stages in the disintegration of a worm with a well-developed posterior zooid
nearly ready for detachment, exposed to Nj 100 KCN. Below, graph of the axial
gradient of a similar specimen. The abscissae represent the ordinal number of
the segments, the ordinate the time to death in the toxic solution, in minutes.
(From Hyman, Jourw. Exp. Zool. xx, 1916.)
18-2
276 FIELDS AND GRADIENTS
In general, the term field implies a region throughout which some
agency is at work in a co-ordinated way, resulting in the establish-
ment of an equilibrium within the area of the field. A quantitative
alteration in the intensity of operations of the agency in any one part
of the field will alter the equilibrium as a whole. A field is thus a
unitary system, which can be altered or deformed as a whole ; it is
not a mosaic in which single portions can be removed or substituted
by others without exerting any eflFect on the rest of the system.
The agencies operative within biological field-systems have not
yet been identified with certainty. In many cases, as in the re-
generation of hydroids and worms, it has been suggested with a
good deal of probability (on the basis largely of experiments on the
differential susceptibility of the regions of the system to toxic and
narcotic agents) that they concern a gradient in the rate of some
fundamental metabolic process (see p. 301). However, the precise
nature of the processes in question is irrelevant to the general dis-
cussion, and for the time being we shall refer to them under the
non-committal term of activity -gradients. In other cases, such as
the limb-producing capacities of the Urodele limb-field (p. 222)
which concern the morphogenesis of a single restricted region, the
simplest assumption is that there exists a graded concentration of
the specific chemical substances responsible for limb-production
and laid down by chemo-diflFerentiation.
In all examples so far studied, the agencies in question appear to
be graded quantitatively in somewhat simple patterns, frequently
(Hydroids, Planarians, many eggs) in the form of a single gradient
with high point at one end and low point at the other, the direction
of the gradient coinciding with the long axis of the organism. It
Fig. 131
Axial susceptibility- gradients of various oligochaete worms. The abscissae
represent the ordinal number of the segments of the worm, the ordinates the time
in minutes elapsing before death when exposed to weakly toxic solutions of KCN
(Njioo to NI500). Above, left, susceptibility-gradient of a mature Aeolosonia in
which secondary zooid formation has not begun. Above, right, the same for an
Aeolosoma in which the shape of the gradient indicates that the processes leading
to the formation of a posterior zooid have been initiated (compare also fig. 130,
in which a posterior zooid is visibly differentiated). Centre, susceptibility-gradient
of an individual of Dero without visible fission-planes. The posterior rise in
susceptibility, characteristic of most oligochaetes and associated with the sub-
terminal growth-zone, is well shown. Below, the same in Lumbriculus. The
posterior rise is more marked, and concerns a larger proportion of the body-
length. (From Hyman, Jfourn. Exp. Zool. xx, 1916.)
5
10
!5
20
25
30
35
45
0
-
/
20
-
J
40
60
1 .
">. ,
^
1 1
1 1
1
1 1
1 1 1 1
10
30
50 70 90
Fig. 131
110
50
278 FIELDS AND GRADIENTS
was this aspect of biological field-systems which first attracted
attention, and led Child to formulate his theory of "axial gradi-
ents".^ It is preferable to combine the two ideas in a single phrase
by speaking of field-gradient systems (figs. 130, 131).
In other cases (Annelids), a double gradient is found, with a
high point at both ends. As we shall see later (p. 309), the two
gradients are probably of qualitatively different nature. The gradi-
ent-system of the amphibian early gastrula also appears to be of
this type (pp. 310, 318). In other cases, as in the localised areas
of the embryo which after the phase of chemo-differentiation are
predetermined to give rise to particular organs, we appear to have
gradient-fields with a central or subcentral high-point, the gradient
apparently being concerned with the concentration of a particular
chemical substance. Cases where this form of gradient have been
definitely demonstrated are the limb-disc of Urodele embryos
(p. 222), the neural plate (p. 243), the rudiments of the auditory
vesicle (p. 232), the gills (p. 233), the heart (p. 233), and probably
that of the lens (p. 238).
Although the precise mechanisms underlying these systems are
still to seek, various important aspects of morphogenesis cannot
be understood or rationally interpreted without postulating their
existence. Further, from the large body of empirical data available
it is possible to deduce certain general rules which are perfectly
valid on their own biological level, in spite of having as yet received
no adequate interpretation on the physiological or physico-chemical
level. In what follows, an attempt will be made to give some account
of the general properties and behaviour of these biological field-
gradient systems, and to show how the field-gradient conception
illuminates certain processes of morphogenesis.
§3
It will be best to base our treatment upon the phenomena of re-
generation, since here the field-gradient systems are for the most
part less specialised and less restricted than in early embryonic
development. From such a study a number of rules emerge.
(i) Our first general rule is that where complete regeneration is
possible from a fragment of the body, the type of regenerate pro-
1 Child, 1915A.
FIELDS AND GRADIENTS 279
duced is normally controlled in relation to the polarity of the
fragment.
It is well known that the bodies of Coelenterates, Planarians and
Annelids are polarised. A differential of some sort exists between
different levels, so that in cut pieces the end nearest the apical
region usually regenerates a new apical region (see p. 271), while
the other cut surface usually regenerates a posterior end. (The ex-
ceptions to this statement are treated later (p. 296), and it will be
found that they can all be formulated in terms of another general
rule.)
In these cases, the cut piece contains a portion of the general
gradient-system of the entire individual, which piece by the fact
of cutting becomes isolated as a separate field-system in which the
factors determining polarity are still graded from apical to basal end.
This rule, however, needs some amplification. In Planaria, if a
transverse fragment is divided in the middle line into two halves,
both will form a head at the anterior end. But if it is divided into
three pieces, the central of which includes the main longitudinal
nerve-trunks, the two outer pieces will, if below a certain length,
form heads either obliquely at the medio-anterior corner, or at
right angles to the original main polarity, on the median cut surface :
the percentage of medianly directed heads increases with decrease
in the length of the piece.^ It would appear probable that in this
case the new heads are determined in relation to the cut ends of the
lateral nerves, which come off transversely from the main longi-
tudinal nerve-trunks, and are of the same essential structure, con-
taining cells as well as fibres. There is of course also a secondary
medio-lateral susceptibility gradient in the intact animal, and this is
presumably correlated both with the course of the lateral nerves and
with the determination of medianly directed heads.
(ii) The second general rule is that the origin of polarity is to be
sought in external factors. Either the polarity of the regenerating
fragment is taken over from that of the whole organism, which is
derived from that of the embryo, which in turn is due to factors
external to the ^gg (pp. 36, 60); or the regenerating fragment ac-
quires a new polarity under the influence of the external agencies
acting upon it after its isolation. In some cases, although the frag-
^ Beyer and Child, 1930.
28o FIELDS AND GRADIENTS
ment is originally polarised by virtue of possessing part of the
general gradient system of the organiser from which it has been
isolated, it is possible to abolish this original field and to substitute
another for it. This can be done, for example, with pieces of the
stem of the hydroid Corymorpha, by placing them in dilute solutions
of various narcotics (see p. 63), In these conditions the pieces
round themselves off, and dedifferentiate. If sea-water is now sub-
stituted, they redifferentiate, but with a new polarity, at right angles
to the substratum.^ This is probably to be explained by the greater
oxygen-concentration away from the substratum. Normally, the
differential established by this means, at right angles to the original
polarity and to the long axis of the piece of the stem, is less powerful
than the already existing differential due to the physiological
gradient between the two ends of the piece. But when this latter
has been abolished by narcotics, the other comes into play, and
establishes a new physiological gradient.
It is to be noted that although this differential is smaller than that
constituting the original polarity of the piece, the polyp eventually
formed is normal. Once the gradient has been established, it acts
as a realisation-factor for the production of an apical region. If the
conditions permit of this developing normally, then, as will be seen
later, the rest of the reconstituted organism will be normal, provided
that the piece is not too small (see p. 285). This is clearly similar to
the processes leading to the establishment of the plane of bilateral
symmetry and the grey crescent in Amphibia, described in
Chap. IV. Numerous differentials, of very varying intensity, can
lead to the establishment of bilateral symmetry : if conditions are
normal, the bilaterality of the embryo is always normal, whatever
the intensity of the trigger action which has released the processes
leading to its formation.
Reversal of polarity has also been obtained by appropriate
methods of grafting in Hydra and other forms. The reversal may
occur in the small engrafted fragment or in the major "host"
portion.^
In other cases, the regenerating portion of tissue is not isolated
in such a way as to take over a part of the original field-system of
the organism from which it is derived, and therefore possesses no
^ Child, 1925 B, 1927. ^ Goetsch, 1929.
FIELDS AND GRADIENTS 281
original polarity at all. This is seen in the reconstitution-masses
formed from pieces of sponges or hydroids after being strained
through bolting-silk.i Yet here, too, axiation, or the development
of polarity, later appears, presumably in response to external differ-
entials in such factors as oxygen supply. This appears to be com-
parable with the determination of the main axis of polarity in the
oocyte (pp. 36, 65 ; figs. 27, 132). In passing, it may be mentioned
that in the sponge reconstitution-masses, one important step in
differentiation, namely the attainment of the two-layered condition,
appears to be caused by the migration outwards of the dermal and
inwards of the collar-cells from their original scattered positions.
Here the fate of the cells is not determined by their position, as is the
case with undifferentiated cells (e.g. blastomeres of regulation-eggs,
early regeneration-buds), but the already acquired differentiation
determines the position taken up. When an excess of collar-cells is
present, these cannot be overgrown by dermal cells, and they form
spheres or vesicles with the collars directed outwards instead of
inwards (see p. 250).
(iii) Our third general rule is that in regeneration the apical region
or head is the first to be formed ; and that its formation, once initiated,
is an autonomous process, independent of the level of the cut, and
also of the formation of other regions, whether in the regenerated
material or within the old tissues of the piece.
The autonomy of a limited apical region is most clearly seen in
the regeneration of Annelid worms. In many of these, the tissue
actually regenerated at an anterior cut surface, whatever its level
in the body of the worm, never forms more than a restricted head
region, composed of a definite number of segments (the precise
number varying with the species: it may be as low as two).^ In
Planarians the autonomous apical region is the head ; its posterior
limits, however, seem not to be quite so sharply fixed as in Anne-
lids. The formation of the cephalic ganglion appears here to
1 H. V, Wilson, 1907; Huxley, 1911, 1921 a.
2 In some Annelids, anterior regeneration is complete; i.e. the regenerated
tissue produces just those segments needed to complete the front end of the
worm, and not a fixed number of segments only (see Berrill, 193 1 ; E. J. Allen,
1921). This appears to depend on the power of growth in the regenerated tissue.
However, the extreme anterior end would here too be the dominant region
(p. 285).
282
^^
^^^
B
Fig. 132
Reorganisation of cell-masses from dissociated cells in the sponge Sycon.
A, 2-layered stage attained after 5 days, probably by migration of the dermal cells
to the exterior and the gastral cells to the interior. B, 34 days, a typical Ascon
stage has been reached, with open osculum {osc) and uniradiate and triradiate
spicules. See also fig. 27, p. 66. (From Huxley, Phil. Trans. Roy. Soc. B, ecu,
1911.)
FIELDS AND GRADIENTS
283
be the most essential feature in the production of a new apical
region.
The autonomy of regenerated apical regions in these forms is in
striking contrast to the dependence of regenerated basal (posterior)
Fig. 133
The independence of the apical region. Partial regeneration in short stem-
fragments of Tiibularia, whether the result is uniaxial or biaxial, gives rise to
apical regions, together with as much of the rest of the organism as can be formed
from the material available. (From Child, Individuality in Organisms^ Chicago,
1915.)
regions. In all Annelids and Planarians, the tail region regenerated
at a posterior cut surface is formed as a direct continuation of the
fragment, and completes the missing parts of the animal. No
remodelling is needed, either in the new tissues or in the original
fragment ; whereas after the formation of a head region of limited
284
FIELDS AND GRADIENTS
extent, a complete animal can only be produced by a remodelling
of the organisation of the original fragment (see below).
In Hydroids such as Tubiilaria and Corymorpha, the new
hydranth is produced entirely by reconstitution within the old
Fig. 134
Reconstitution from pieces of stem in Corymorpha. A, Normal unipolar form
showing hydranth and base with holdfasts. B-G, Reconstitution of very short
pieces to form partial structures, either unipolar (B, C) or bipolar (D, E, F, G).
The extreme apical region is always present. In E, the original apical end has
formed more than the basal end. H, J, Total and partial twinning of hydranth.
K, L, Formation of numerous apical and basal regions in relation to a single
hydranth (K) or independently (L). (Redrawn after Child, Biol. Gen. 11, 1926.)
tissue, not by regeneration from the cut surface, so that here the
delimitation of the apical region is less clear-cut. The independence
of extreme apical regions is, however, very well shown in these
forms (fig. 134).
Extremely small fragments of their stems do not become recon-
FIELDS AND GRADIENTS 285
stituted into miniature whole polyps; they produce only apical
portions of polyps, but these are of normal size. Such small frag-
ments frequently form an apical region at both ends, for reasons
to be discussed later : in such cases, two sets of apical structures are
produced, without any basal portion. Similar phenomena occur
in the regeneration of very short pieces of Planaria. In Corymorpha,
reconstitution-masses produced from the aggregation of dissociated
cells may produce only apical portions of hydranths^ (see p. 65).
In all cases, what is determined in the first instance is, in fact,
the formation of an extreme apical region of a certain standard size,
this varying with the size of the piece and also with external con-
ditions. Once this extreme apical region is determined, the region
next more basal is determined, and so on, until all the available
material is used up. This process may be initiated either at one or
at both ends of the piece.
Abnormal external conditions influence the size of the apical
region produced. In Planarians, for instance, cold and narcotics
reduce its size, while heat up to a certain degree increases it. Beyond
a certain degree of cold or concentration of narcotics, no apical
region will be formed at all (fig. 135; see also p. 301).
In the most general terms, it appears that the relative size and
the degree of differentiation of the apical region depend in some
way upon the physiological activity of the regenerated tissue. If this
is depressed by cold or narcotics, the development of the apical
region is subnormal.
(iv) Our fourth rule is that, once an apical, region is produced, it
then exerts an influence on other organs and regions within the old
tissues of the fragment: this influence is, however, limited in
extent. Accordingly, the apical region has been called by Child
the "dominant" region. In terms of the field-concept, the apical
region establishes a field of a certain extent, which it dominates so
as to control the morphogenetic processes of the other regions of
the field. The control is exerted in such a way that the various
morphogenetic processes occur in harmonious relation with each
other : this is because it exerts its control through the establishment
of a field.
If the range of dominance is artificially reduced, as by removal of
1 Child, 1928 B.
286
FIELDS AND GRADIENTS
some of the more basally situated tissue, the gradient-field set
up by the dominant region is in relation to the reduced size of its
U
m
nM
l:,'S' ?ji^
'-/; S •■'
«''i?' Ifj|,
;;'.■'-• ■■;'/'
u m
mu
u U
\iy c>
mi m
mm
Kfe
'Ji!'.v^^
Fig. 135
Correlation between size and degree of development of a regenerated apical region
in Planaria, and the extent of its inductive capacity. Posterior fragments are
isolated as shown in a. h. Regeneration in standard conditions, c-e. Regene-
ration in increasing concentrations of narcotics, showing decreasing size and in-
creasing abnormality of the regenerated head. Correlated with this, the pharynx
induced in the old tissues becomes smaller and less remote from the apical
region. /, Regeneration at high (optimal) temperature. The head and eyes are
larger, the induced pharynx farther away and of greater size. (From Child,
Individuality in Organisms , Chicago, 19 15.)
possible range. For instance, the reconstitution of a polyp in a
portion of stem of Tuhiilaria of a certain length normally results in
FIELDS AND GRADIENTS
287
the formation of rudiments of distal and proximal tentacles of a
certain size, distance apart, and distance from the apical point. But
these values are smaller if the piece of stem is shorter^ (see also
p. 318). Similarly, when the Ascidian Clavellina undergoes de-
differentiation into a small mass of cells, and subsequently re-
differentiates into a well-proportioned Clavellina of reduced size,^
mill
a
III!
I %i
I tv
ill
lilfl
H '& 5- T' «
m
i
Fig. 136
Modification of the scale of organisation in reconstitution in stem-pieces of
Tubularia. a, Future mouth region; b, primordia of apical tentacles; c, future
hypostome; d, primordia of main (basal) tentacles, i, Under standard condi-
tions. 2, In optimal conditions: the scale of organisation is enlarged. 3, In sub-
normal conditions : the scale is decreased. (From Child, Individuality in Organisms,
Chicago, 19 1 5.)
one might say that the various fields are localised in terms of relative
quantitative positions along the main gradients : and these relations
holding for different total sizes, the control exerted by the dominant
region will be harmonic.
As already mentioned (p. 165), the extent of the field dominated
by an apical region can be experimentally modified. Narcotics re-
duce the size of the regenerated head in pieces of Planarians ; the
size of the reconstituted pharynx, as also its distance from the
Driesch, 1899; Child, 193 1
- Huxley, 1926.
ZSS FIELDS AND GRADIENTS
anterior cut surface, is then a function of the size and degree of
differentiation of the head, which in turn appears to be a function
of the activity of the regenerating tissue from which it was formed.^
A similar state of affairs is found in the reconstitution of pieces of
the stem of Tubularia. Here the apical tentacles constitute the
dominant region. The size of the rudiments of these determines
the distance between these and the rudiments of the basal tentacles,
and can be modified experimentally. In one series of experiments on
stem fragments of a definite length, the average length of the pri-
mordia of the two sets of tentacles was reduced by 12 per cent, by
immersion in M/i 50,000 KCN, and 23 per cent, by immersion in
M/50,000 solution^ (fig. 136).
In reconstruction from dissociated cells in Corymorpha, the
frequency of complete hydranths was much reduced and that of
partial forms, consisting of apical portions only, much increased by
moving the undifferentiated cell-aggregates about during a certain
critical time after their formation instead of leaving them attached
to the substratum. The interpretation advanced is that when
attached, the differential established between well-oxygenated
upper surface and poorly oxygenated lower surface will be large,
the resultant gradient steep; when moved, the gradient between
apical and basal regions will be less steep, and the structure can
therefore differentiate on a larger scale, whereas with a steep
gradient it is more compressed.^ In other words, the morphogenetic
field of the polyp in process of reconstitution can be altered
as a whole by altering the differentiation of the apical region.*
Interesting results have also been obtained in Sahella (p. 165). In
abdominal fragments of this worm, the number of segments of
abdominal type which are transformed into segments of thoracic
type by a regenerated head varies from o to 75 (the number pro-
duced in normal ontogeny is 5 to 1 1). Here, the agencies responsible
for the wide range in the extent of the region morphogenetically
affected by the new head in this case appear to reside chiefly in the
old tissues^ (fig. 137).
The conversion of abdominal into thoracic segments, obtained
experimentally in Sahella, occurs as a normal process in the develop-
1 Child, 1915A. - Child, 1931. ^ Child, 1928 b.
^ Child, 1915 A, p. 128. ^ Berrill, 1931.
_ o S • r3 <u y 2
HEE
19
290 FIELDS AND GRADIENTS
ment of Filigrana and Salmacina. The young forms of these worms
have three thoracic segments, and the budding zone at the hinder
end adds a number of segments of abdominal type behind them. In
subsequent development, the number of abdominal segments is
increased as a result of the activity of the budding zone, but such a
method is of course out of the question in the case of the thoracic
segments. These increase their number to ten by conversion of the
most anterior abdominal segments.^
(v) This fourth rule is really a special case of a more general fifth
rule, which is that, within a given field, the diflFerentiation of all
regions, other than an apical region, is dependent on influences
which proceed from more apical levels. For instance, a piece of a
Planarian can regenerate a tail posteriorly even if it fails to re-
generate a head. Similarly, whereas a piece of a Planarian from the
post-pharyngeal region will not form a new pharynx unless a head
is regenerated at its anterior end, a piece from the prepharyngeal
region is capable of producing a pharynx even in the absence of a
head.^
Corymorpha also provides a good example of this. In this
hydroid, grafts of a portion of the stem of one polyp inserted later-
ally in the stems of other polyps will in a certain proportion of cases
act as organisers and induce the outgrowth of a new hydranth. It
was found that grafts from the apical region inserted at basal levels
induced hydranths in nearly 85 per cent, of cases, while grafts from
basal levels inserted at the same level in another stem were only
effective in 45 per cent, of cases; in addition, the hydranths pro-
duced by basal grafts grew more slowly and arrived at a smaller
size.^ The capacity to organise does not reside in any specific tissue
but is a physiological condition, the efficacy of which varies
quantitatively down a gradient (fig. 138).
In addition, it should be noted that general stimulation such as
that produced by an incision will induce the formation of new
hydranths in Corymorpha. Here the influence of the substrate on
the result emerges clearly : for whereas at apical levels of the stem
a single incision will usually induce a hydranth, at basal levels this
is ineflfective, and lacerated incisions are required for induction.
1 Malaquin, 1919. ^ Child, 1915 A, p. 102.
3 Child, 1929 B.
FIELDS AND GRADIENTS
291
r\
u
\J
E
17
D
Fig. 138
Induction by grafts in stems of the hydroid Corymorpha. The grafted fragment
is shown stippled. A-C, Distal fragments of stem engrafted at proximal levels in
the host stem. A, An early stage. B, C, Two specimens after 48 hours' develop-
ment. The graft induces an outgrowth, which it organises to form a complete
large hydranth. D-E, Proximal fragments, of stem engrafted at proximal levels
in the host stem, after the same length of time as B and C. The resultant hy-
dranth is smaller (D) or subnormal and delayed (E). (From Child, Physiol. Zool.
II, 1929.)
19-2
292 FIELDS AND GRADIENTS
The relation between the effect of a dominant region and that of
general physiological stimulation is clearly brought out by this
experiment, and lends additional weight to the view that the
dominant region owes its inducing capacities in part at least to
its high physiological activity.
(vi) The sixth rule is that one at least of the influences exerted by
the more apical regions on regions at lower levels is that of inhibition.
There appears to exist both inhibition of general activity (as
evidenced chiefly by susceptibility experiments), and also of differ-
entiation. The inhibition of differentiation is well shown by the
following experiment. If a polyp of Haliclystus be cut across trans-
versely, it will regenerate new tentacles over the whole cut surface.
If, however, an oblique cut be made, reaching down as far as the
transverse cut in the previous experiment, and continuing upwards
so as to leave intact a small portion of the original distal rim, no
regeneration will occur on the less apical part of the cut surface.
The presence of the apical region inhibits lower levels from re-
generating.^
This rule is really another way of putting certain consequences
of our third and fourth rules. Within the region of the body capable
of regenerating a new apical region at all (which may include the
whole organism, or may be restricted to its more apical portion:
see p. 297), any piece of tissue, if by reason of an operation it finds
itself at the front cut surface of a fragment, can develop into an
apical region. That it does not do so in the intact animal is due to
the presence of the apical region. The control exerted by the apical
region is thus twofold : it inhibits the appearance of other apical
regions within the limit of its field, and it influences the tissues to
develop into subordinate organs in relation to the morphogenetic
gradients which it sets up within its field.
The inhibition set up is not merely morphogenetic; it is also
trophic. In portions of Hydroid colonies kept in suboptimal con-
ditions, the stolons that are formed frequently detach themselves
from the stock and move slowly across the substratum, their original
tip leading the way. This appears to be due to the tip being the
dominant region within the subsidiary gradient-field of the stolon :
it is able to grow by abstracting material from the proximal, sub-
1 Child, Sci. Rep. Tohoku hnp. Univ. 4th Ser. Biol, vin, 1933, p. 75.
FIELDS AND GRADIENTS
293
294 FIELDS AND GRADIENTS
ordinate region of the stolon. When, however, the hydranths are
healthy and vigorous, they dominate the stolon and maintain them-
selves at the expense of any attached stolons, which are gradually
resorbed. A similar state of affairs is seen in the Ascidian Pero-
phora (see fig. 139 and p. 425).
The most striking case of trophic dominance is found in the
flatworm Stenostomiim.^ Here it can be conclusively shown that
the dominance depends on the degree of development of the apical
region. Stenostomiim possesses asexual reproduction and forms
chains of attached zooids (up to eleven in number), separated by
fission-planes. These fission-planes are formed in a regular order,
and the relative age of the zooids can thus be determined, as well
as by inspection of the degree of development of the head. If a
fragment of the chain be isolated by cutting, the zooid possessing
the oldest head left in the fragment normally resorbs all younger
zooids and any headless portions of zooids which are anterior to it.
This is shown in fig. 140. If the posterior cut had been made a
little farther back, a still older head would have been included in
the fragment, and would have resorbed all regions anterior to itself.
If the fragment is made so short as not to contain a head, regenera-
tion occurs at the anterior cut surface, and there is no resorption.
A similar relation occurs between the earlier- and later-formed
holdfasts of Corymorpha?
These facts show that it is not merely the presence of a cut
surface which leads to regeneration: the cut surface must be in a
certain relation to the gradient-system of the fragment as a unit.^
(vii) This leads on to a seventh rule, which is a corollary of the
fourth. This concerns the origin of new apical regions as a result of
what Child has called physiological isolation. If a portion of tissue
comes to lie outside the field dominated by an existing apical region,
a new apical region will arise in this portion, even though it is still
in physical continuity with the rest of the organism. The common-
est way in which this state of things is brought about is by con-
tinuous growth. For instance, in Stenostomiim the first appearance
of a new head only occurs at a certain distance from the old, and
^ van Cleave, 1929; Child, 1929 a. 2 Child, 1928 a.
^ It is possible that the phenomena of the graded distribution of growth-
potency in the animal body (see p. 366) is correlated with this trophic effect
of one part of a morphogenetic system upon another.
295
^ ^
\ L
Fig. 140
A chain-forming flatworm, Stenostonnmi grandis. Physiological dominance of
zooids with older head-regions over those anterior to them which are headless
or have less advanced head-regions. A, Chain of five zooids, showing piece
isolated, between X-X and Y-Y. B, The headless anterior zooid-fragment is
partly resorbed. C, It is further resorbed but is attempting to differentiate a
head. D, It has been totally resorbed, and the next zooid is undergoing resorp-
tion. E, F, The original posterior zooid, with the oldest head in the fragment,
has resorbed all the material anterior to it, and has divided to form a younger, more
posterior zooid. (From Child, Arch. Entzomech. cxvii, 1929.)
296 FIELDS AND GRADIENTS
the detachment of the part of the chain dominated by this second
head only takes place when a certain greater distance has been
reached. The formation of new zooids in colonial organisms such
as hydroids is regulated in relation to this rule. The distance be-
tween zooids — i.e. the extent of the field controlled by a more apical
zooid — varies with temperature, nutrition and other conditions.^
Complete physiological isolation of an incipient new apical re-
gion can also be achieved by removing the old dominating apical
region. Some species of Planaria reproduce by transverse fission.
By cutting off the original head, precocious fission is induced.
Further, in these forms, the length of body attained before fission
occurs varies with the degree of differentiation of the head : if as a
result of regeneration in depressant solutions a subnormal head is
produced, it can only control a small field, and fission occurs at an
unusually small body-length. ^
In the regeneration of such forms as hydroids and worms,
various complications may be found. Sometimes biaxial regenera-
tion occurs, leading to the formation of two apical regions, one at
each end of the piece (or, more rarely, two basal regions — e.g. tails
in Planarians). Sometimes no apical organ is regenerated. The
percentage frequency with which this occurs almost always in-
creases with increasing distance of the front end of the fragment
from the original front end of the body. When an apical region is
regenerated, its final form and the rate of its regeneration also vary
with the level of the original body from which regeneration takes
place.^
It is unnecessary to go into detail here as to the reasons for these
complications. They appear to depend on the interplay of several
factors. The result depends in the first place on the portion of the
gradient-field of the original body contained in the cut fragment.
Secondly, on the release of the fragment from the inhibition ex-
erted by the old dominant region, which results in an increase of
1 Child, 1 929 A.
2 In plants, physiological isolation has been obtained by exposing to low
temperature a portion of the region (e.g. a runner) connecting dominant and
subordinate parts. Even though under these conditions the runner continues to
grow, the field is interrupted, and a new plant is precociously formed at the free
tip of the runner. Similar experiments have not yet been successfully carried
out on animals. See Child, 1915 A.
2 Sivickis, 1931 A and b.
FIELDS AND GRADIENTS 297
physiological activity throughout the piece ; the extent of this in-
crease will vary with the age and position of the fragment. Thirdly,
the operation of cutting also results in an increase of activity : this
is intense close to the cut, and then appears to grade away rapidly.
Fourthly, external conditions influence the activity both of the old
tissues and still more of any new tissue proliferated at the cut
surface. This question has been discussed at some length by Child. ^
Comparable results have been shown to occur in the regeneration
of fragments of certain plant tissues, such as seakale roots."
(viii) So far as the facts are relevant here, we may sum them up in
the form of the following rule : The frequency or absence of regenera-
tion, and the type of structure regenerated appear to depend (a) on
the level of the cut surface within the original gradient-field, and (b)
upon the form and steepness of the gradient eventually established
between the proliferating tissues at the cut surface and the rest of
the piece.
As Child has epigrammatically put it, when a new apical region
is regenerated, it arises not because of the activities of the rest of
the fragment, but in spite of them.
As a corollary of these various rules with regard to the establish-
ment of polarity, at the autonomy and subsequent dominance of
the apical region, the facts concerning the varying number of
structures produced by a given piece of tissue may be satisfactorily
explained. A given length of Tubularia stem normally possesses
but a single hydranth, whereas regeneration experiments show that .
it is capable of producing dozens. The limb-disc of a Urodele, if
cut up and the pieces grafted, can produce several fully developed
limbs : why in normal development does it only produce one ? In
the regeneration of a fragment of Corymorpha stem, sometimes one
new dominant region is produced, sometimes two, sometimes
several : why is this ?
The reason that a given field normally gives rise only to one of
the structures characteristic for it is due to the inhibiting effect of
a dominant region, once initiated, upon the development of other
dominant regions. Normally the gradients within the field are such
as to give one region a start ; this becomes the dominant region and
inhibits the potentialities of other regions. This is well seen in
^ Child, 1915 A. ^ Jones, 1925.
298
g c« O
S.^
5 > =3 'd
^ 2 c« rt
— ^ o o 5;
^ ^^
<U O .
t3 2 O 1^
o c t; *'^
(U v^ C c«
C3 OT C "Ij
(U (U ^
<U "^^ Vh ^ •
c w o -^ ^
C f«, <u ^^ CO
^ H ^ ^ ^
.S >
0 0) O)
rt "5 C t3 V
FIELDS AND GRADIENTS 299
Urodele limb-buds. If a limb-disc is removed and grafted on to
the flank of the same animal, it will develop into an independent
limb if sufficiently far from its original position. But if the site of
grafting is within three segments of its original position, it is within
the sphere of dominance of the limb developing from the portion
of limb-disc left in situ and becomes resorbed.^ Experiments on
Anura have had similar results.-
If an entire limb-disc be grafted, it often develops into two or
three limbs. In this case the operation has upset the normal gradi-
ent system, and permitted supernumerary centres of activity to
develop. This fact is of great interest in its bearing upon dichoto-
mous growth ; for it shows that a field which normally gives rise to
a single set of structures can under slightly altered conditions be
made to give rise to two.^
Similar agencies are at work in a Hydroid or a Planarian. The
various regions of the body, though each capable of producing a
new apical region, are all held in check by the existing head or
hydranth. When, however, growth has removed them to a sufficient
distance, the inhibition can no longer act on them, and they do
develop into apical regions.
The double multiple forms are of great interest. Two-headed
Planarians can be produced by splitting the anterior end and pre-
venting the two halves from reuniting; and types with doubled
apical region can be produced by similar means in sea-anemones.'*
In cases of normal dichotomy of branching organisms, the duplica-
tion of the axis and main structures is brought about by growth ;
however, what initiates the division of the growing-point in these
forms is not yet known.
In fragments of hydroid stems, biaxial hydranths are formed when
conditions at the two ends are such as to produce two positive
gradients of sufficient intensity to initiate the formation of an
apical region. Neither region has a sufficient advantage to inhibit
the development of the other. If either cut surface is handicapped
by being enclosed in paraffin or stuck in the sand, only the other
end produces a new hydranth. In Tiibularia the stem is enclosed in
1 Detwiler, 1918. " Hellmich, 1930.
^ The symmetry relations of the supernumerary limbs are of much interest:
this problem is considered on p. 224.
^ Child, 1924, p. 161. See also below, p. 327.
300
FIELDS AND GRADIENTS
Fig. 142
a perisarc which acts as a handicap to all regions within it, and
permits of regeneration only at the ends. In Corymorpha, however,
the perisarc is absent over most of the stem-length in large specimens.
As a result, multipolar forms often arise, especially from short
fragments^ (fig. 142).
A very striking example of the multiple production of dominant
regions is seen in the sea-anemone Harenactis, When portions of
the body are isolated, they roll up
to produce hollow structures like
a tyre, the original distal and
proximal cut surfaces meeting and
growing together. It is then found
that regeneration is initiated at a
number of places along the line
of suture. Regeneration is found
especially at places where, owing
to irregularities of the cut surfaces,
union has not been smooth. At
each of these spots, conditions are .tta^anelf^^SS'Te;"
favourable for new growth, and diagram of Harenactis to show {a and
therefore the establishment of new ^) sections isolated for regeneration.
, These roll in to torm hollow tubes, as
apical organs ; and the various shown (centre) in section : the distal
regions of new growth are isolated and proximal cut surfaces unite in a
r 1 .1 1 ^1 • suture, here shown centrally. Right,
from each other by other regions regeneration of apical regions (whole
in which smooth union, leaving no or partial tentacle groups) from the
free cut surface, has taken place.'^ suture. When the suture is irregular,
' . ^ with considerable prohteration, a
As Will be seen m the next chap- number of apical regions can arise
ter, these phenomena of double (above) ; when the union is smooth,
, . 1 . . . . one regenerate dominates and inhibits
or multiple organisations arising ^^^ development of others (below).
from a single portion of tissue are (After Child, Physiological Foiinda-
of great interest in the interpreta- tions of Behavior, New York, 1924;
° . ^ modined.)
tion of various facts in ontogeny.
Before passing to our next section, the views of Goetsch^ should
be mentioned. He finds that regeneration is frequently accom-
panied by a polarised migration of cells. In some cases, certain
types of cells have the tendency to migrate apicalwards, other types
Child, 1926.
Goetsch, 1929.
2 Child, 1924, p. 119.
FIELDS AND GRADIENTS 3OI
basalwards. In other cases, indifferent cells migrate in both
directions, but become progressively differentiated in different ways
according as they are moving apicalwards or basalwards. The
presence of a lateral graft, e.g. in Hydra, will induce a flow of cells
towards the graft; as is shown by heteroplastic experiments in
which the two types of tissue can be distinguished, some of these
cells grow out to form a base below the graft.
There is thus a form of "dynamic determination" (see p. 163),
and although the graft acts in a way resembling an amphibian
organiser, it does so largely by a different method, namely, by
stimulating directive growth-processes. Something of the sort occurs
as part of the induction of new hydranths by stem grafts in
Corymorpha (p. 164), and in the case of grafted amphibian limbs
(p. 364). On the basis of these and numerous other experiments
he comes to conclusions rather different from those of Child.
However, although it is probable that the further analysis of
these directive movements of migration and growth will throw
much light on regeneration and differentiation, they cannot explain
a number of the facts previously cited in this chapter, for which
some form of field-theory is indispensable.
(ix) Next we come to an extremely important rule, which is
that the action of external conditions upon gradient-fields and the
morphogenetic processes associated with them is always differential.
This appears to be a consequence of the quantitatively graded nature
of the fields. The differential action is revealed under three main,
heads, {a) differential inhibition, {b) differential stimulation or
acceleration, {c) differential acclimatisation and recovery.
{a) When depressant agents are used in concentrations not
permitting acclimatisation, the most active regions are the most
susceptible, and suffer most. This is well seen when regenerating
Planarian fragments are exposed to narcotics. The heads regener-
ated under such conditions are not only subnormal in size, but
abnormal in form, in that certain regions are missing. The process
takes place progressively as the toxicity increases : first of all the
eyes become approximated, and then fused (absence of interocular
region) ; then the median part of the pre-ocular region fails to form.
Higher concentrations affect more lateral and more basal parts of
the head, until finally only a small basal head-rudiment, eyeless but
302 FIELDS AND GRADIENTS
with rudimentary ganglia, is produced. Higher concentrations in-
hibit head-formation ahogether, and only healing occurs (fig. 143).
(b) DiflFerential acceleration occurs in response to exceptionally
favourable conditions. The effects are the exact reverse of those
obtained by diflrerential inhibition, though such extreme departures
from the normal are not seen, since there are no regions which fail
to form. As illustration we may take the fact that optimal high
temperature applied during regeneration of Planarian fragments
leads to the formation of heads which are not only relatively large,
but have widely separated eyes and an unusually large pre-ocular
region.
(c) Differential acclimatisation occurs in certain low concentra-
tions of depressants. In these it appears that the most active regions,
although the most susceptible, have the greatest power of acclimati-
sation, and after a time show differential development. For in-
stance, intact normal Planarians placed in weak alcohol or ether
first show a differential reduction in size of head, the whole pre-
ocular region disappearing. Later, new growth sets in, and this is
abnormally high in the most median and most anterior regions,
leading to "snouted" forms^ (fig. 143).
In slightly stronger concentrations, this differential action will
not take place during exposure to the solution, but occurs on re-
placement in water. In such cases the process is strictly speaking
one of differential recovery instead of differential acclimatisation,
but the results are in most respects similar.
(x) In addition to these statements, applicable to regeneration of
the complete type, within total fields permeating the whole body,
there must be mentioned another very important rule derived from
a study of partial regeneration in a local field. This is that the various
tissues of the regenerated region need not be proliferated from
corresponding tissues in the old region, but are determined in re-
lation to a gradient-system which extends out from the old region
into the proliferated material. Total regeneration appears normally
to take place in two phases — first the formation of a new apical
region, and secondly the remodelling of the old tissues under the
influence of this apical region. However, in partial regeneration,
e.g. of an amputated limb or tail, the new tissues are not known to
^ Child, 1921 A.
FIELDS AND GRADIENTS
303
exert any morphogenetic effect on the old tissues of the stump : if,
as usually occurs, a complete appendage is restored, this is effected
entirely by means of new growth.
y V
J K L • I M
Fig. 143
Differential susceptibility in Plmiaria dorotocephala. A-E, Various grades of
head differentiation after regeneration. A, Normal, B, Teratophthalmic (eyes
approximated or partially fused, head form nearly normal). C, D, Teratomorphic
(single median eye, lateral sensory projections approximated or fused anteriorly).
E, Anophthalmic (no eye, median or no sensory projection, rudimentary cephalic
ganglion). F-H, Diagrams showing, between the dotted lines, the regions
missing in hypotypic heads. F, In teratophthalmic forms (cf. B). G, In terato-
morphic forms (cf. C, D). H, In anophthalmic forms (cf. E). J-M, Differential
acclimatisation. J, Normal head. K, Reduction of apical region after 2-3 weeks
in dilute anaesthetics. L, M, Subsequent hypertrophy of the apical region after
iJ-2 weeks more in the solution. (Redrawn after Child, bidividiiality in Or-
ganisms, Chicago, 1915 (A-E), and jfoiirti. Exp. Zool. xxxiii, 1921 (F-K).)
For this to occur, it is clear that exactly those regions removed
by the operation must be restored by the new growth, a phenome-
non abundantly confirmed in limb -regeneration in Arthropods and
304
FIELDS AND GRADIENTS
Amphibia. It has always been difficuh to connect this with any
purely chemical specificity of the regenerating tissues at one level
as against another level of the limb, and recent work has made such
a view wholly untenable. For one thing we have the fact already
referred to (p. 271) that the regenerated material is at first wholly
undifferentiated, and is only later determined in relation to the
substrate on which it grows. This is not conclusive, for it merely
proves that the old tissues do not impart any chemical specificity
they might possess to the material just proliferated ; the later deter-
mination might be due to chemical influences specific to the level
Fig. 144
Diagram to show the independence of regenerated tissues, a, Triton with normal
fore-limb skeleton, b, The humerus is removed, and the fore-arm and hand re-
moved, c, The regenerated fore-arm and hand contains the normal complement
of skeletal elements. (Przibram, in Handh. norm. u. path. Physiol, xiv (i) (i), 1926.)
of the cut. However, it has now been shown that total absence of
one kind of tissue, or the substitution of one kind of tissue by another
in the regenerating base of the limb, does not interfere with normal
regeneration. If the skeleton be removed from the upper arm or
thigh region of a Urodele limb, and the limb later cut across in this
region, the distal regenerated portion possesses a normal skeleton,
whereas no regeneration of the missing parts occurs in the stump. ^
Similarly, if the skin is removed from a limb, an envelope of lung
tissue grafted on, and the limb cut across after healing has occurred,
the regenerated portion is found to possess normal epidermis, in
spite of the absence of such tissue in the stump.^
^ Weiss, 1925; Bischler, 1926. ^ Weiss, 1927 a.
FIELDS AND GRADIENTS 305
Such facts can only be interpreted in terms of a field theory.
Some general activity must be distributed in a graded way through
the limb so as to constitute a gradient-field. The fate of the re-
generated tissue is determined in relation to the level of the gradient
at which regeneration is made to occur, not to the specific tissues
present on the cut surface. Further, the determination of the re-
generated portion is a unitary process. The regenerated portion is
determined as a field, the morphogenetic agencies in which are in
equilibrium with those operative in the stump, so that the fractional
field of the regenerated portion and that of the stump together
make a whole (see also Chap, x, p. 362). Both the products of
undifferentiated cells and also certain types of already specialised
cells contribute to the regenerated material.^
Presumably the morphogenetic gradients in the stump extend as
it were by extrapolation into the new tissue, so that it comes to be
permeated by the missing portion of the total field: when this
occurs, the gradient activities of the whole field are in equilibrium.
As regards its gradients, the regenerated portion then constitutes
a fraction of a field: but since it alone contains undifferentiated
tissue, in its subsequent morphogenesis it behaves as an auto-
nomous field system with basal boundary set by the level of the cut.
The same type of behaviour is seen in the regeneration of a tail
in Planarians; the new tissue from the start is determined in re-
lation to the existing gradient-stem of the old piece. It would thus
appear that the basalmost regions of a limb are dominant, and
correspond, as regards their activities in the gradient-system, to the
anterior (apical) region of the whole body in animals capable of total
regeneration.
It should be noted that in such cases quite a small fraction of
the field (e.g. a short disc cut from a limb) will be able to exert this
morphogenetic effect on material proliferated from its cut surface,
even when grafted into another region of the body altogether (e.g.
a short section of fore-limb stump taken with an indifferent re-
generation-bud that has been proliferated from it, and grafted into
the hind-limb field; seep. 273). It is also important to find that when
a section of a limb is cut out and engrafted elsewhere in reversed
orientation, with original proximal cut surface away from the body,
^ Hellmich, 1930.
HEE 20
c
Fig. 145
Regeneration is determined by the level of amputation within the limb-field, and
not with reference to the organism as a whole. A, Dorsal, and B, Ventral, views
of a newt {Triton) in which the legs were amputated above the thigh, and shanks
were grafted in their place and subsequently amputated. C, Radiogram, showing
that only the tarsus and foot have been regenerated: i.e. structures distal to the
graft. (From Guyenot, Rev. Suisse de Zool. xxxiv, 1927.)
FIELDS AND GRADIENTS
307
this free cut surface does not regenerate the missing (i.e. proximal)
regions, but produces a structure representing the parts of the Hmb
distal to the level of the cut, although this duplicates regions of the
stump. 1 The same is true of tail-fragments.^ These results show that
Fig. 146
Diagram showing the morphogenetic effect of the limb-field in regeneration. In
Triton, an early regenerate bud from a fore-limb cut as in {a) is taken and grafted
on to a hind-limb stump. If (6) grafted with a portion of the original stump, it
produces (c) a fore-foot; if {d) grafted alone, it produces {e) a hind-foot. (Przi-
bram, in Handb. norm. 11. path. Physiol, xiv (i) (i), 1926.)
the explanation given above needs modification. The field is not
active within its differentiated regions : the morphogenetic influence
is exerted in relation to the character of the diflFerentiated tissue at
the cut surface.
§4
Little is known as regards the precise time-relations of some of the
processes, e.g. whether the new morphogenetic gradient-field is
established immediately the new head is determined, or not until
it has reached some degree of morphological development, such as
the formation of a brain, or the penetration of nerves from the new
brain into the old tissues. The general sequence, however, is clear.
^ References in Milojewid and Grbid, 1925.
^ Milojewi6 and Burian, 1926.
308 FIELDS AND GRADIENTS
But the precise method by which the dominant region exerts its
morphogenetic control over the rest of the field is still unknown.
However, an experiment may be described here which not only
illustrates the importance of quantitative potential difference, but
also throws light on the problem of determination in regeneration.
In the fresh-water Annelid worm Lwnbriculus, if the hindmost fifth
of the body is cut off, a head will be regenerated from the front edge
of this piece in 90-95 per cent, of cases. In a second series of ex-
periments, a small piece containing two or three segments is cut off
in such a way that its anterior edge is at precisely the same level on
the long axis of the worm as the anterior edge of the whole hindmost
fifth in the first series. These small pieces of the second series only
regenerate a head in 20-30 per cent, of cases. It might be supposed
that this lack of power to develop a head was due to insufficiency of
material in the small piece, but this is not so. If a hindmost fifth
of the worm is cut off as before, and then, 20 hours later, a large
piece of this be removed so as to leave a piece identical in size and
in level with that used in the second series of experiments, it is
found that a head will be regenerated in 70 per cent, of cases.^
Lack of power to develop a head in the second series of experi-
ments is therefore not due to lack of material, for the pieces of the
third series are of the same size as those of the second, but can re-
generate a head almost as well as those of the first series. The only
difference between the pieces of the third and second series is that
for 20 hours the anterior end of the pieces of the third series has
been in continuity with the whole hindmost fifth of the worm, and
this period of time is apparently long enough for the qualitative
determination of a head to be effected, as in the first series. After
this determination, reduction in size of the piece does not hinder
head-production. The fact that the act of cutting raises the activity
of the old tissue in small pieces more than in large pieces where the
cuts are farther apart and the stimulation consequent upon them
has to act on a much larger mass of material. The anterior edge of
small pieces will therefore have more difficulty in obtaining the
necessary threshold potential difference for head-determination.
Experiments in every way analogous to those just described on
Lumhriciilus have been performed on Planaria^ and with similar
^ Hyman, 1916.
FIELDS AND GRADIENTS 309
results.^ It appears that at room temperature the formation of a
head is determined in about 6 hours from the time of operation.
§5
There is another fact concerning the gradient-systems of adult
lower invertebrates which requires consideration, for it throws light
on certain processes of embryology. This is the double gradient
analysed by Child and his school in Annelid worms. In these
animals, as is well known, new segments are added from a growing
zone in the penultimate segment of the body. Experiments with
dilute toxic solutions show that there is a region of high susceptibility
at both ends of the worm, with a minimum at an intermediate point.
Child and his school have always attempted to reduce all gradient-
phenomena to variations in a single variable, which they have tried
to identify with oxidative metabolism, but which, theoretically,
might be any general activity of protoplasm. This conception, how-
ever, seems definitely to break down in face of the facts in Annelids.
Here, two distinct processes appear to be at work. One is the forma-
tion of new segments at the hind end associated with the presence
of undifferentiated, physiologically young tissue ; the other is the
controlling and morphogenetic activity of the front end, associated
with old tissue and a high grade of differentiation. It is worth re-
calling that the conditions of formation and the morphogenetic
effects of the dominant region in Annelids are similar to what is
found in Planarians (see p. 279). If regeneration occurs at all at an
anterior cut surface, the normal result is a new dominant region,
which never consists of more than a small number of segments,
constant for each species ; and this, once produced, causes morpho-
genetic changes in the old tissues, such as the production of a new
genital region at the correct distance behind the head, or the trans-
formation of a certain length of intestine into crop and oesophagus,
or the conversion of abdominal segments into thoracic segments."
Though both head and tail in Annelids are regions of high sus-
ceptibility, the processes at work in the two are entirely distinct.
There are therefore two qualitatively different gradients in the
organism, and there is every right to believe that the effects of the
^ Child, 1914. See also Abeloos, 1932.
^ Harper, 1904; Berrill, 1931.
310 FIELDS AND GRADIENTS
two will interact — e.g. that the morphogenetic effect of a head of
given activity will differ according to the tail-gradient and the
effects which this exerts on the old tissues, just as in regeneration
from a posterior cut surface, with a given tail-gradient, the morpho-
genetic results will vary according to the size and activity of the
head.
This leads on to a point which may prove to be of great theoretical
importance, although so far only limited discussion of it has taken
place. ^ It concerns the classification of gradient-fields into two
types. The first constitutes what Waddington refers to as an in-
dividuation- field, in which there exists some form of dynamic
equilibrium controlling morphogenetic processes. Removal of one
part of the system will, if growth is still possible, lead to the re-
generation of what is missing, as above pointed out (p. 276).
Further, the induction effect of a dominant region is exerted not
by contact as with the amphibian organiser, but apparently at a
distance, as with regenerating Sabella or Planarm; this is because
the essential effect of the dominant region is to establish a total
field. Another term for these would be gradient-fields of direct
effect.
In contradistinction to this we find what may be called gradient-
fields of secondary effect. A gradient-system exists, and exerts its
effects, not directly, but by giving rise to a graded concentration of
some chemical substance which is then responsible for certain
morphogenetic effects. It appears that in amphibian eggs the dorso-
ventral gradient with the organiser at its high point is of this type.
The reasons for this assertion are in the first place that induction
is exerted mainly by contact (see p. 135) ; secondly, that dead organ-
isers may continue to exert their inductive effect (p. 153); thirdly,
that there is no evidence of equilibrium or saturation being ob-
tained in the organiser region. This last point requires elucidation.
In the bird, a complete "organiser-field", i.e. a sheet of epiblast
(ectomesoderm) containing the whole primitive streak and an ex-
tensive area around it, is still capable of inducing a neural tube and
other organs in a sheet of epiblast from another embryo.^ The
organising capacities of the primitive streak have not been ' ' satur-
ated" in the formation of its own field, as would be expected if
^ Waddington and Schmidt, 1933.
FIELDS AND GRADIENTS 31I
organisation were an affair of equilibrium between the morpho-
genetic capacities of the organiser and the neighbouring tissues, as
is clearly the case with a regenerating Planarian (see p. 287).
These two types of gradient are of course not mutually exclusive.
The primary gradient-field of the amphibian egg is an individua-
tion-field ; but as a result of its existence, graded accumulations of
yolk and other substances occur, which then exert effects upon
development. In this case, the substances accumulated are mere
raw materials, but in other primary gradients, doubtless, true organ-
forming substances are formed in this way. It is, of course, also
possible to conceive of the graded formation within a gradient-field
of some substance which has no further effect on development, so
that there is no secondary action of the field as occurs with the
organiser. Such fields we may if we like distinguish as fields of in-
direct action but without secondary effect. Many cases of graded
distribution of pigment within organs are doubtless of this type.
The gradient-field of the amphibian organiser appears to be
essentially one of secondary effect ; but it very possibly acts as a
weak individuation-field in the stages before gastrulation. In birds,
as already mentioned (p. 160), portions of the organiser (primitive
streak) when isolated regularly produce more than their presump-
tive fates, thus showing a tendency to individuation.
Partial fields such as the limb-field in Amphibia appear to par-
take of both these aspects of field-action. They seem undoubtedly
to be areas in which there has resulted a graded concentration of a
specific chemical substance which is capable of producing limb-
formation : but they also have their own individuation-field, which
sees to it that what is produced is normally neither a partial nor a
multiple structure but one whole organ.
As with any new concept, considerable analysis, both experi-
mental and theoretical, will be needed before the different roles of
field-systems in ontogeny can be properly understood. Meanwhile,
however, this distinction between fields of direct and indirect action
is a first important step, helping considerably to clarify amphibian
development (see Chap, ix, p. 318).
Chapter IX
FIELDS AND GRADIENTS IN NORMAL ONTOGENY
§ I . Polarity in ontogeny
As already mentioned, the conclusions reached in the preceding
chapter are derived from experiments on regeneration and grafting
in adult animals. They are also, however, relevant in the normal
ontogeny of higher forms, though the conditions here are often more
complex and more specialised. In the present chapter it is pro-
posed to illustrate the various principles, so far as possible, from
early development.
(i) Polarity and the main axis of the resultant organism
The first rule mentioned in Chap, viii was that the inherent
polarity of a fragment normally determined the polarity of the
organism which arose from it. This obviously holds good in normal
ontogeny. The egg is a fragment of the mother, in which a well-
marked polarity has been set up before it is detached. In the great
majority of cases, the main animal -vegetative axis of the Qgg gives
rise to the definitive antero-posterior axis of the resulting organism,
with the head or apical region arising at the animal end. In various
Echinoderms the main axis of the tgg persists as that of the larva,
but later a new axis in a different direction is established in the
rudiment of the adult.
(ii) Polarity determined by external agencies
We next come to the point that the polarity of an organised
portion of living matter has in the long run been determined by
agencies external to it ; and that in certain cases the existing polarity
can be overridden and a new polarity imposed by external con-
ditions. Examples have already been given of how the polarity of
the developing oocyte or egg may be determined by factors external
to itself, either by conditions within the ovary, or external agencies
acting after fertilisation (Fucus). Cases have also been adduced in
which the axis of bilateral symmetry is determined from without
FIELDS AND GRADIENTS IN NORMAL ONTOGENY 313
(p. 60). The rule appears to be of general application for the
developing egg.
The overriding or abolition of the original polarity by external
agencies appears seldom to be obtainable with eggs ; but some re-
markable cases are known from Echinoderms. We have already re-
ferred (p. 83) to the fact that in developing fragments of Lytechinus
and Patiria eggs, which have been obtained by cutting before fer-
tilisation and subsequently inseminated, the first two cleavage
planes are always at right angles to the plane of the cut.
Subsequent development demonstrates that even more radical
changes have been effected. When the gastrulation of the fragments
occurs, it invariably takes place at the centre of the cut surface, and
at right angles to it. The polarity of the developing egg-fragment
and the axis of the resultant larva is therefore determined in re-
lation to the cut, and not in relation to the original polarity of the
whole tgg.^
It is to be supposed that the operation intensely stimulates the
cut surface, and that the resultant increase of protoplasmic activity
grades away across the fragment. The activity-gradient thus pro-
duced must be able to override the original gradient within the
fragment.
This is also stated to occur in the California species of Para-
centrotiis. However, in the European Paracentrotus lividiis, meri-
dional halves of the Qgg produced by isolation of the 1/2 blasto-
meres appear to retain the original polarity.^ It is to be noted that
in this case no raising of activity by cutting has occurred; the
separation also took place at a later stage. Thus the observations on
the two forms are not necessarily contradictory.
In P. lividus also, marked deformation as a result of centrifuging,
however, is incapable of altering the original polarity. The point at
which gastrulation is initiated is always at the original vegetative
pole, as indicated by its relation to the subequatorial pigment-
band. Thus gastrulae are produced which may be extremely
elongated, flattened, or obliquely deformed in the animal-vegetative
direction^ (see also Chap, iv, p. 69).
^ Taylor, Tennent, and Whitaker, 1925.
2 Horstadius, 1928.
3 Harvey, 1933.
314
* "V
Fig. 147
Atypical (A-C) and typical (D) differentiation of portions of Triton early gas-
trulae, grafted (interplanted) into the orbit of larvae. A, Notochordal tissue from
presumptive epidermis. B, Notochordal tissue from presumptive endoderm.
C, Cartilage from presumptive neural plate. D, Epithelial vesicle from pre-
315
/. d.
s. k.
sumptive epidermis, b. basal membrane; c, cornea; ch. notochord; cy. epithelial
vesicle ; d. covering layer ; ep. epithelial vesicle ; i. contents of vesicle ; k. cartilage ;
I. Leydig's cells; m. muscle; ml. pigment;/), pigment; rm. muscle; s. granules in
Leydig's cells. (From Kusche, Arch. Entzumech. cxx, 1929, figs. 9, 13, 20, 22.)
3l6 FIELDS AND GRADIENTS IN NORMAL ONTOGENY
§ 2. The domina7it regioji in ontogeny
(iii) Independence of the dominant region
Instances of this are difficult to obtain in ontogeny. The egg cannot
regenerate new tissue Uke a Planarian worm, and we can therefore
only compare the morphogenetic processes occurring in it to those
occurring by morphallaxis in the regeneration of, for example, a
piece of Tuhidaria stem. However, the gradient-system of the egg
Fig. 148
Differentiation of notochord {nc.) and mesoderm {ms. muscle) from animal pole
material (presumptive epidermis and/or brain) of Triton interplanted in the eye-
socket of a larva of Triton taeniatus. (From Bautzmann, Naturwiss. xvii, 1929.)
is almost always more specialised than that of a hydroid stem, being
in many cases partly or wholly determined as regards different
levels along its main axis. It is also more limited in size, and there-
fore its gradient is presumably steeper.
However, when both animal and vegetative portions of the ^gg
can reorganise themselves to form perfect wholes, as in some
Coelenterates (p. 97), equatorial portions of the egg, originally in
FIELDS AND GRADIENTS IN NORMAL ONTOGENY 317
the centre of the gradient, and constituting a subordinate region,
must in the vegetative half have turned into a dominant region and
come to control the new complete gradient-field of the fragment.
The curious and apparently anomalous production of notochord
and mesoderm by various isolated regions of the blastula of the
newt, although these regions may possess the most diverse pro-
spective fates (p. 139, footnote), may perhaps be explained on
these fines when it is remembered that isolation of a piece of tissue
removes it from the control of the dominant region to which it has
been subjected. As noted on p. 285, experiments on Planarians and
Hydroids have shown that the tendency in such cases is for a small
isolated piece to develop by self- differentiation into an isolated
dominant region. The dominant and only self-differentiating region
in the late blastula of the newt is the organiser, and the tissue into
which it differentiates is notochord and mesoderm : other regions
develop in subordination to it. On this assumption, therefore, a
piece from any other region, when isolated, should, if environmental
circumstances permit, come to be the site of a new dominant
region, and differentiate accordingly.^ However, the occasional
differentiation of such pieces into tissues which represent neither
the presumptive fate of the piece nor that of the dominant region
(organiser) presents a difficulty. We should however recall that
whereas in the Invertebrates only a simple field is involved, in
Amphibia there are two interacting gradient-fields (pp. 310, 318).
(iv) The modifying influence exerted by the dominant
region on other parts
This is obvious in the example just given of the formation of
miniature wholes from animal and vegetative portions of Coelenter-
ate eggs. Regions originally containing but half the length of the
main gradient become reorganised to contain whole gradients.
In most of the well-analysed types of ontogeny, however, con-
ditions are more complex than in the regeneration of Hydroids or
Planarians, for the main organising activity proceeds from the high
^ In connexion with the environmental circumstances, it is a curious fact that
pieces of presumptive neural tube tissue (which has been the tissue most fre-
quently used in these experiments) show a much greater tendency to differentiate
into notochord when interplanted into the coelomic cavity of an older larva than
when explanted in an inorganic medium (Holtfreter, 1931 a). See also Huxley, 1930.
3l8 FIELDS AND GRADIENTS IN NORMAL ONTOGENY
point of a secondary gradient, established after fertilisation, and
the precise morphogenetic effects are due to the interaction of this
with the original animal- vegetative field established in the oocyte.
The relation between the gradient-field set up by a dominant
region and an amount of tissue representing a reduced range, is
well shown in the experiments on newt embryos (described on
p. 239) in which the early gastrula is constricted into dorsal and
ventral halves. The dorso- ventral gradient is then of half the
normal length, and the dorsal half-gastrulae possess neural folds of
proportionately reduced size. Another example is provided by the
experiments on sea-urchin larvae to be described below (p. 323), in
which four micromeres are added to a single ring of mesomeres
(disc an i), and a properly proportioned pluteus larva is formed.
The main gradient is here represented by one quarter of its original
length, and in this case the amount of the dominant region has had
to be reduced in order to produce a harmonic result.
§ 3. The interaction of primary and secondary gradients
In early amphibian development, for instance, there appears
clearly to be two gradients of qualitatively diflFerent nature. One
is the gradient along the primary egg-axis from animal to vegetative
pole ; the other, a gradient whose high point or dominant region is
the organiser. The first appears to be established during the de-
velopment of the oocyte in the ovary. It must in the first instance
be quantitative and concerned only with some general activity of
the cytoplasm : but by the time that the egg is ripe, it has in addition
produced a structural effect, in the shape of the graded increase in
the proportion of yolk found when passing down the egg-axis to-
wards the vegetative pole. The existence of this gradient has been
shown by susceptibility experiments. (See p. 332, and figs. 154,
i55> 156.)
Per contra, although the other gradient, which is normally
established as a result of fertilisation, has a sharply qualitative
aspect in that the dorsal lip region alone is capable of exerting
organiser capacities, yet it is also quantitative in other aspects. For
instance, it is found that, as determined by cell-size in the late
blastula and early gastrula, the rate of cleavage in the future dorsal
side of the animal hemisphere is greater than in the ventral side.
FIELDS AND GRADIENTS IN NORMAL ONTOGENY 319
Susceptibility experiments also demonstrate the existence of a
dorso -ventral gradient in general activity, from the region of the
grey crescent ventralwards over the egg. In respect of its position
at the high end of a gradient, the organiser of the amphibian egg
shows a further resemblance to the dominant regions of a Coelen-
terate, Planarian, or Annelid. (See p. 68 and fig. 28.)
As already pointed out in Chap, vi, the action of an apical region
such as a Planarian head is extremely similar to that of an organiser
in ontogeny. Not only does it exert a morphogenetic effect during
regeneration, but also when grafted into an intact worm. But the
morphogenetic action of the amphibian organiser is normally
exerted in a way somewhat different from that of a regenerating
head, for its definitive influence is exerted on those parts which
it actually comes to underlie as a result of gastrulation, and appears
to be a chemical effect, demanding contact for its realisation. In
this respect the dorso-ventral gradient of the amphibian egg reveals
itself as a gradient-system of secondary effect, thus differing
importantly from the apico-basal gradient system (see Chap, viii,
p. 310). However, the labile determination effected before the onset
of gastrulation can only be the result of action at a distance, as with
the effects of a regenerated Planarian head.
In Amphibia the end-result, in the shape of the main morpho-
logical organisation of the embryo, is dependent on the interaction
of the organising capacity of the dorsal lip with the primary apico-
basal gradient-system. The most important action of the dorso-
ventral gradient, from the point of view of developmental physiology,
is the production of a specific inducing substance localised in the
dorsal lip region, which acts as a trigger or releasing stimulus for
the differentiation of its own and other tissues. On the other hand,
the most important action of the apico-basal gradient is the pro-
duction of an individuation-field, which sees to it that the develop-
ment released by the non-specific action of the organiser is in the
first place different in different parts of the field, and in the second
place is correlated into an organised whole.
Either system also appears to have minor effects of the opposite
type to its main effect. The organiser, as just mentioned, appears to
exert an action at a distance prior to gastrulation, and this may be
comparable with that of the individuation-field set up by a Planarian
320 FIELDS AND GRADIENTS IN NORMAL ONTOGENY
head ; on the other hand it may be that the individuative component
of this action is really due to the apico-basal individuation-field,
and that the organiser region here again only exerts a releasing
action, whether by the diffusion of chemical substances, or by
neuroid transmission, or by other means.
It is further probable that the dorso-ventral gradient contributes
to the total individuation-field of the embryo, e.g. by introducing
a dorso-ventral polarity (see p. 357 for the dorso-ventral polarity
of limb-areas).
Per contra^ the primary individuation-field also exerts indirect
effects owing to the graded accumulation of yolk, cytoplasm, fat
and other substances along its axis (p. 311). This has secondary
effects upon the rate of cleavage and relative cell-size in different
parts, which are of importance in the mechanics of gastrulation ;
and also upon the amount of raw materials available in different
parts of the body. The apical region of the primary field will always
attempt to form a brain : but it can only form a brain of normal type
if it contains less than a certain proportion of yolk, and less than a
certain proportion of fat. Thus the indirect effects of the primary
gradient are adjusted to co-operate with the direct effects.
The position of the grey crescent itself is a prior example of such
interaction. The point of sperm-entry decides the meridian of the
grey crescent, and therefore the meridian on which the high point
of the secondary gradient will lie. However, the precise latitudinal
position of this high point is not sharply predetermined at a fixed
level, but depends upon conditions in the primary gradient and can
be experimentally modified by modifying these. For instance, ex-
posure of the frog's egg to depressant agencies (e.g. Njio LiCl)
during early segmentation leads to the dorsal lip being formed
nearer the animal pole than usual; in some cases even above the
equator (fig. 149).^ Temperature gradients (p. 339) applied during
segmentation also influence the position of the dorsal lip.^
The same sort of interaction of two gradient-systems occurs in
the Echinoderms, only the high point of the secondary gradient is
here directly vegetative instead of dorsal. In all probability, similar
processes are at work in Annelids and Arthropods (see p. 309).
^ Bellamy, 1919.
^ Dean, Shaw, and Tazelaar, 1928.
FIELDS AND GRADIENTS IN NORMAL ONTOGENY 321
Obvious examples of the dependent differentiation of a sub-
ordinate region under the influence of the dominant (organiser)
region interacting with the primary gradient-field are seen in the
Amphibia in the formation of secondary embryos after the grafting
of an organiser; or the development of engrafted fragments of
organs other than the organiser, when the donor has not reached
mid-gastrulation, in accordance with their new position instead of
their original presumptive fate. The most remarkable of all such
cases are the modification of pieces of Anuran presumptive epi-
dermis, grafted into the future mouth-region of a Urodele egg, to
Fig. 149
Modification of the site of dorsal lip formation in the frog. Left, control egg;
right, egg exposed to w/ 10,000 KCN for 24 hours from the 2-cell stage. The
dorsal lip {b.p.) is much closer to the equator in the treated egg. (After Bellamy,
Biol. Bull. XXXVII, 1919; modified.)
form a part of the head and jaws which is perfectly organised with
the rest of the larva, but which differentiates Anuran structures
(e.g. suckers and apparently teeth) never found in Urodeles^ (p. 142).
(v) The influence of more apical (but not completely apical or
dominant) regions on less apical regions
This is excellently illustrated by the experiments on the 3 2-cell
stage in sea-urchin eggs, described in Chaps, v and vi (pp. 1 03 , 1 68),
in which it was shown that not only would the basalmost disc of
cells (micromeres) induce gastrulation, but so would the sub-basal
disc after removal of the micromeres.
It may also be recalled that the tendency of the animal disc, an. i,
is to produce a larva in the middle of which the cilia of the apical
organ occupy much too much space and in which no gastrulation
takes place, while the tendency of vegetative material is to produce
^ Spemann, 1932, 1933; Spemann and Schotte, 1932.
P CO P
II
o ^ .2
u cj .t;
•p .2 "o
c ?^ c
=1
o o
M ^ § ^->
£ ^-^ o ^
c T <u —
« M-c .2
lili-
^ O ^ T3
^ ^ c/^ -2
^ ^ O :S
<*. ^ ^ U
,^ <Li £« <i^
p +^ o „
^ S ..
t+H oj 5i S
c/o oj o :5:
o c; '"' «
FIELDS AND GRADIENTS IN NORMAL ONTOGENY 323
an exogastrula, without apical organ, cilia, or mouth. The situation
here is complicated by the fact that the gradient-field (vegetative-
animal) concerned with inducing gastrulation interacts with the
previously established animal-vegetative gradient-system of the
egg. As a result, not only are organising capacities graded with
distance from the vegetative pole, but so are the capacities for being
organised.
It has been found^ that the production of a properly proportioned
pluteus larva is dependent in the first place on the presence of some
of the vegetative pole material. This material acts as an organiser,
and is normally to be found in the micromeres : micromeres grafted
into abnormal situations will induce gastrulation and the formation
of a secondary set of main organs where they are grafted ; they will
organise the neighbouring tissues so as to make them conform to
the normal morphology of a larva, and to the new polarity set up
by the graft. But the organising capacities are not restricted to the
micromeres, for, if they are removed, it is found that the next most
vegetative region, disc veg. 2, is capable of forming a pluteus with
proportions approximating to those of the normal. It is well known
that lithium salts produce exogastrulation in Echinoderm larvae,
i.e. a reinforcement of the vegetative potencies." It is therefore in-
teresting to find that lithium salts induce gastrulation and the
awakening of organiser properties in isolated animal halves^ (see
also p. 337).
In the second place, the production of a perfect pluteus larva is
dependent on a balance between animal and vegetative material,
and it has been possible to study this balance quantitatively. In
order to obtain a well-proportioned pluteus it is necessary to add
one micromere to disc veg. i, two micromeres to disc an. 2, and four
micromeres to disc a?i. i. Excess of material from the animal pole
leads to imperfect gastrulation and abnormal enlargement of the
apical organ, excess of material from the vegetative pole leads to
exogastrulation and reduction in the extent of the ciliated area.
Similarly, it is possible to observe gradual approximation to
normal proportions when macromeres are added to a complete
animal hemisphere. Isolated, the animal hemisphere gives a
^ Horstadius, 193 1. ^ Herbst, 1895.
^ von Ubisch, 1929.
21-2
324 FIELDS AND GRADIENTS IN NORMAL ONTOGENY
blastula, three-quarters of the surface of which is covered by the
ciHa of the apical organ. Addition of half a macromere gives a
larva in which the apical organ is reduced almost to normal pro-
portions ; a ciliated band and a stomodaeum are formed, but no gut.
Addition of a whole macromere gives a little pluteus in which the
gut is, however, too small. Addition of two macromeres gives a
perfect pluteus. The addition of four micromeres produces roughly
the same effect as that of one macromere. Thus, in proportion to
total bulk, the organising capacity of the micromeres is far higher
than that of the macromeres, since their size is only about one-
thirtieth of that of the macromeres (fig. 150).
If in place of a whole animal half, an isolated disc an. i had been
used, the addition of four macromeres would have resulted in the
formation of a perfect pluteus. This again shows that the morpho-
genetic effects of the organiser material are dependent on the level
(within the main gradient) of the tissues which they are organising.
A further proof of this is given by the following fact. An isolated
veg. I disc will invaginate a little gut ; but the addition of an animal
hemisphere to veg. i prevents the latter from gastrulating at all.
If corresponding amounts are removed from both ends of the
gradient, the remaining tissue is still able to form a pluteus. Thus
discs an. 2, veg. i and veg. 2, together, are able to form a properly
proportioned larva. But the zones which have been removed, an. i
and the micromeres, are together also able to give rise to a proper
pluteus. It is therefore possible to obtain two perfect larvae after
section at right angles to the egg-axis, provided only that the
balance between animal and vegetative potencies is preserved.
The importance of these facts needs no emphasising. They show
that the morphogenetic properties of the organiser in the Echino-
derm larva are located at one end of a gradient ; that these capacities
are not localised in any given tissue, but diminish gradually with
increasing distance from the vegetative pole, along the gradient, and
that the degree of organisation produced is quantitatively depend-
ent, first upon the difference of level (along the main gradient)
between organising material and material to be organised, and
secondly upon the relative amounts of the two kinds of material.
FIELDS AND GRADIENTS IN NORMAL ONTOGENY 325
§ 4. Inhibition, physiological isolation, and multiple potentiality
of fields in ontogeny
(vi) Inhibition exerted by a dominant region
on other parts of the system
Perhaps the most striking example of this in early ontogeny is
found in sea-urchins. Here, the presence of the organiser (gastru-
lating) region inhibits the formation of long cilia on the late blastula
and gastrula, except for a small tuft at the apical pole. In the ab-
sence of the organiser, these cilia spread over all or most of the
surface of the blastula (see p. 103). The inhibition is here exerted
by the dominant region of the secondary or vegetative-animal
gradient (p. 320); but the principle is the same as in the example
given in the preceding chapter.
No cases of resorption of a subordinate by a dominant region are
known in early embryology. The resorption of parts occurring at
metamorphosis (Amphibia, Echinodermata), and the partial re-
sorption of one member of a pair of double monsters by the other
are clearly of rather a different nature. However, an alteration in
relative size of parts can often be obtained as the result of differ-
ential inhibition. This is so in the experiments on Chaetopterus
larvae and Echinoid plutei, described on p. 332: it indicates that
there is a competition for available food-material between the
different parts of the embryo, and that the degree of success in that
competition is, in part at least, regulated by the relative activity of
the dominant region and other parts of the organism.
(vii) Physiological isolation and the multiple
potentiality of gradient-field systems^
The fact that in many forms the early stages of development can
be made, by appropriate fragmentation, to produce more than one
normal larva was one of the earliest discoveries of the science of
experimental embryology. It attracted a great deal of attention, and
led Driesch to formulate his conception of ''harmonic equipo-
tential systems" (p. 353).
Numerous examples of this have been given. We need only recall
^ And see corollary, Chap, vin, p. 294.
326 FIELDS AND GRADIENTS IN NORMAL ONTOGENY
that multiple development can be obtained by cutting the un-
fertilised egg and inseminating the fragments (p. 120); by isolating
1/2 or 1/4, and in some cases even 1/8 blastomeres (p. 97); or by
cutting and breaking the blastula into fragments (pp. 81, 89).
The most significant example of the multiplication of potencies in
the early egg is perhaps the production of double monsters from
inverted frog's eggs (p. 94). In this case there is no spatial isola-
Fig. 151
Multiple potentiality in head-field in the Planarian Dendrocoelum lacteiun. The
anterior end was partially slit by a number of cuts ; the organism has produced
ten heads. (Redrawn from Korschelt, Regeneration und Transplantation, 1927,
fig. 269, p. 444; after Lus.)
tion of fragments; a physiological isolation between two active
regions is brought about by the intercalation of a mass of inert yolk.
The coalescence of two eggs to produce a single unitary embryo is
a converse result of the same principles. Further, just as two-
headed Planarians or bifurcated regenerated limbs can be produced
by operations, so can two-headed newt embryos be produced by
partial constriction in the 2-cell stage (pp. 75, 350).
FIELDS AND GRADIENTS IN NORMAL ONTOGENY 327
Partial or regional fields can also give rise to more than one
structure. The amphibian organiser region itself can be divided
and engrafted to produce several embryos (p. 151). The limb-field
can be made to produce a number of limbs. This can be done not
only by grafting portions of it into new situations (p. 223), but
simply by making deep cuts in the early Kmb-buds:^ the result
is a number of limbs growing out from the limb-area. Other
regional fields also show this multiple potentiality, e.g. heart,
balancer, etc. One of the most striking examples is provided by the
anterior end of a Planarian, which, by making deep cuts, can be led
to give rise to as many as ten heads^ (fig. 151).
One point which may here be mentioned is the existence in all
large-yolked vertebrate embryos and in all mammals of consider-
able areas of tissue produced by the fertilised egg but not organised
into the body of the embryo. Examples of such tissues are the
extra-embryonic blastoderm of selachians, reptiles and birds, and
the trophoblast of mammals. These do not appear to be organised
in relation to the organising centre of the embryo, and in some
cases (chorion or trophoblast of amniotes) are cast away at hatching
or birth, and thus never become incorporated in the field-gradient
system of the organism. In other cases (yolk-sac) they do ulti-
mately become incorporated by resorption within the body, and
are then organised to produce a portion of the gut.
Such extra-embryonic structures may perhaps be looked on as
composed of tissue which has grown so rapidly as to escape the .
organising action of the organiser, and thus to remain beyond the
boundaries of the embryo. It is of interest that exposure of fowl
eggs to low temperature will produce a large proportion of '*ani-
dian" blastoderms, in which no embryo is formed, but the blasto-
derm shows considerable powers of growth.^
With regard to points (viii) and (x) of our previous chapter, these
only apply to cases of regeneration. They are thus not relevant to
normal ontogeny.
These points lead on to a consideration of the problem of
twinning. The term twinning in the broad sense is applied to any
process by which more than one individual is produced during
early ontogeny from a single zygote. We may, however, profitably
1 Tornier, 1906. ^ Lus, 1924. ^ Needham, 1933.
328 FIELDS AND GRADIENTS IN NORMAL ONTOGENY
distinguish cases in which the separation of the future individuals
occurs during cleavage from those in which the process concerns
later stages. In the former cases, the separate individuals are
isolated by the process of cleavage itself, whereas in the latter,
processes of dichotomous growth and fission are involved.
In the former category, we first have certain cases in which
repeated and irregular division, leading to separation, occurs at an
early stage of cleavage. This phenomenon, usually called poly-
embryony, is found in certain Hymenoptera and Polyzoa. Here,
this process leads to the production of numerous separate in-
dividuals from one egg by the separation of its blastomeres or
groups of blastomeres. These cases are really natural experiments
of blastomere isolation, and it may be noted that axes of polarity
and symmetry relations play little part in the process. As to why
it is in these cases that the blastomeres separate and produce
wholes on their own instead of parts, little can be said except to
point out that in Hymenoptera and Polyzoa the fertilised egg
undergoes cleavage within a mass of living matter, consisting in
the case of the former of the tissues of a parasitised caterpillar
preyed upon, and in the case of the latter, of the nutritive cells of
the ovicell or brood pouch. ^ In these cases it is interesting to note
that a fertilised frog's egg grafted into the body cavity of a fully
developed frog undergoes modified cleavage and these products
become separated and develop as far as they are able on their own.^
Another group of cases comprises those where the early cleavage
stages are artificially interfered with in one way or another. This
phenomenon leads to the formation of double (or multiple) monsters,
each partner being derived from a blastomere or group of blasto-
meres which has been to a certain extent isolated, physiologically
or physically, from the others. Here we must place the double
monsters obtained in Amphioxus as a result of shaking and dis-
arranging the blastomeres (pp. 79, 123): in Tubifex and in Chae-
topterus as a result of inducing equal divisions of blastomere D
containing the essential ingredients for the formation of somatoblasts
(twinning in Clepsine is probably of this type (p. 113)) : in the star-
fish Patiria as a result of spontaneous parthenogenetic development
resulting in semi-independent development of both of the blasto-
1 Harmer, 1930. ^ Belogolowy, 1918.
FIELDS AND GRADIENTS IN NORMAL ONTOGENY 329
meres of the 2-cell stage. ^ Here also may be included those cases
in which for reasons at present unknown the heavily yolked egg of
fish and of birds may exceptionally possess two blastoderms, and
perhaps cases of double monsters in scorpions.^
Twinning in the restricted sense, however, is the result of a
dichotomy setting in, not during the earliest stages of cleavage,
but during later stages of development, and resulting in definite
Kflgf.i
Kfbgf.a
Fig. 152
Incipient twinning mechanically produced. Ventral view of Triton embryo
from an egg slightly constricted in pre-gastrulation stage, showing slight anterior
doubling. Kfbgf.a, outer gill-filaments ; Kfbgf.i, inner gill-filaments ; oc, inner
eyes ; olf, olfactory pits ; * pigment overlying heart-rudiments. (From Spemann,
Arch. Entzvmech. xvi, 1903.)
fission of one embryo into two or more. Here belong the cases of
twinning as found regularly in the armadillo (here resulting in the
formation of four or eight embryos),^ occasionally in other mam-
mals including man (leading to the production of so-called identical
twins), or in birds or earthworms leading to the production of
double monsters: experimentally leading to the production of
double monsters in the frog after reversal of the tgg, in Fiindulus
and trout after subjecting the tgg to cold or oxygen-deficiency
^ Newman, 1923. - Brauer, 1917. ^ Newman, 19 17, 1923.
330 FIELDS AND GRADIENTS IN NORMAL ONTOGENY
(fig. 153), in Patiria after fertilisation by sperm of another species,
or as a result of overcrowding.
In all cases in which twinning has been experimentally produced,
it is clear that the critical stage at which dichotomy occurs is that
of early gastrulation. In the reversed frog's egg the invaginated gut
becomes mechanically split into two in a manner described above
(p. 95) and since the gut-roof is the organiser, the resulting em-
bryo is accordingly more or less completely doubled. Similar cases
are operative in the production of anterior doubling as a result of a
ligature constricting the tgg in the plane of bilateral symmetry
(p. 156, and figs. 32, 152, 169, 170).
In other cases, the twinning is due not to a physical but to a
physiological dichotomy, and the region aff"ected appears always to
be the apical point of a gradient. This point is known to be dif-
ferentially susceptible to depressants (p. 332). All agencies which
make for abolition of polarity, by reducing the rate of activity of the
apical point and flattening the gradient, also tend to encourage the
production of twinning.
This is particularly well seen in Patina where as a result of a
lowering of the general rate of activity consequent upon abnormal
fertilisation or overcrowding, invagination of an enteron takes place
not from one, but from two or three points.^ The same phenome-
non occurs in teleosts (Fundulus and trout), where as a result of the
depressant effects of cold, or lack of oxygen, the originally single
axis of polarity is replaced by two.^
In the armadillo, there is, relatively to other mammals, a delay in
the formation of a placenta, and consequently in the establishment
of a source of supply of oxygen and nutriment for the embryo, and
this occurs at a stage corresponding to the early gastrula, just before
the appearance of the primitive streak. In those occasional cases in
which two embryos are formed on a single blastoderm in a bird's
egg, it is probable that the cold experienced by the egg after laying
and before incubation is responsible for an arrest of development
at a stage which corresponds to the early gastrula, shortly before the
appearance of the primitive streak.
The twinned worms occasionally to be found in the cocoons of
Oligochaetes are presumably to be accounted for by a delay in
^ Newman, 1923. ^ Stockard, 1921.
FIELDS AND GRADIENTS IN NORMAL ONTOGENY
331
development caused by lack of oxygen within the cocoon, which is
occasioned by the high mortality of the eggs and consequent
foulness.
A formal explanation of twinning and the replacement of a single
axis of polarity by two axes, more or less independent, is to be
found in the principle of axial gradients. The maximal suscep-
tibility of the apical point of the gradient, when acted on by
Fig. 153
Partial twinning in trout brought about by reduced oxygen supply during pre-
gastrulation stages. Left, unequal components, anterior duplication. Centre,
anterior duplication, unequal components: component on the left has a very
small head and is cyclopean. Right, a subnormal individual, with only one eye,
no mouth, gills or tail fin, and much reduced trunk, is attached to the surface
of the yolk-sac opposite to the larger normal individual. (Redrawn after Stockard,
Amer.Journ. Anat. xxviii, 1921.)
depressant agencies, brings about the depression of its level of
activity below that of the immediately neighbouring regions. These,
in all cases where the original embryonic area is a flat plate or
blastoderm, as in fish, birds or mammals, will be symmetrically
situated right and left of the original apical point.
Interesting confirmation of the truth of this interpretation is pro-
vided by the cases of twinning presented by the Oligochaetes. As
mentioned above, these worms are characterised by the presence of
332 FIELDS AND GRADIENTS IN NORMAL ONTOGENY
two gradients, with apical point to the front and hind ends respec-
tively. Here, twinning occurs most frequently at each end of the
worm, and very rarely in the middle region.^
The opposite to twinning is the merging together in the middle
line of organs which are typically paired, a good example of which is
provided by cyclopia and monorhiny (p. 348). Attention may here
be called to the part played in the development of the amphibian
eyes by the underlying organiser. It will be remembered (p. 245)
that a piece of presumptive eye-region of the neural plate, taken
from the middle line without underlying organiser, usually dif-
ferentiated into a single eye. The presence of neighbouring under-
lying organiser tissue, on the other hand, leads to the development
of paired eyes from such grafts. In other words, the organiser has
brought about twinning of the rudiments in the field : explanation
of this effect is, however, obscure.
The same principle which underlies twinning by dichotomy in
the whole organism can also be applied to the duplication of single
organs (see above, pp. 296, 327).
§ 5 . Differential susceptibility and the modification
of ontogeny in invertebrates
(viii) The effects of differential susceptibility
Remarkable modifications of normal development have been ob-
tained by applying the principles of differential activity to the eggs
of Annelid worms.^ In Chaetopteriis , susceptibility experiments
show that the animal pole is at first the most active region. This
condition persists until the young larva begins to show elongation
of the trunk, when the posterior region becomes the most active.
By immersing the developing eggs in inhibiting agents (e.g.
rnj 100,000 KCN) from fertilisation onwards, microcephalic forms
are produced. These forms also have their extreme posterior regions
inhibited, as the treatment is continued during the period when
these show high susceptibility. If the treatment is discontinued
after 11 hours from fertilisation, the posterior region is better
developed, while the microcephaly persists.
If, on the other hand, the treatment is not begun until the
^ Hyman, 1921. ^ Child, 1917, 1925 a.
FIELDS AND GRADIENTS IN NORM AL O N TOGEN Y
333
O ^ C cj ci u-
o o .22 <£
O " c
t; >< c3 <u ^
^ -f 9 ^ c 5
334 FIELDS AND GRADIENTS IN NORMAL ONTOGENY
24-hour Stage, the susceptibihty conditions are reversed and mega-
cephaly resuhs (fig. 154).
It is interesting to find that in the microcephahc forms the an-
terior trunk region is absolutely larger than in controls, while the
same is true for the heads of the megacephalic forms. This is to be
explained very simply. There is a definite limited quantity of food
material available in the egg; and when one region is inhibited,
regions which are less aflfected are able to obtain a greater share. It
is not known how long the modifications of proportion thus ob-
tained will persist, although, in Arenicola, forms with some degree
of posterior inhibition have been reared through metamorphosis.
Other interesting experiments have been carried out on sea-
urchin eggs.^ Exposure to inhibiting concentrations of KCN
throughout early development results in plutei in which apical
regions, notably the oral lobe, are relatively under-developed. The
posterior (basal) regions are therefore relatively over-developed,
and consequently after such treatments narrow-angled forms are
produced, in which the arm spicules may even be parallel.
In weaker solutions, where differential acclimatisation can occur,
the reverse process is found. The oral lobe is relatively enlarged,
and the skeletal arms diverge at a wide angle. In extreme cases,
types of highly abnormal proportions are produced (fig. 155).
The well-known "lithium-larvae " of Echinoderms may be men-
tioned in this connexion. It was early discovered^ that when sea-
urchin eggs are reared in a medium to which lithium salts have been
added, forms known as exogastrulae are produced, in which the
archenteron is present, but evaginated instead of invaginated
(fig- 156).
Exogastrulation is the most obvious eflFect of this treatment, but
it appears to be a secondary result. The primary effect of lithium is
to decrease the amount of ectoderm produced by the egg, and to
increase the amount of endoderm, progressively with increasing
concentration. The skeletogenous cells move towards the animal
pole. Exogastrulation is probably a mechanical effect, the de-
creased ectodermal area being unable to accommodate the enlarged
archenteron in its interior. If the lithium treatment is continued
until the middle blastula stage, " holoentoblastulae " may be
1 Child, 1 91 6 b. 2 Herbst, 1895.
FIELDS AND GRADIENTS IN NORMAL ONTOGENY 335
Fig. 155
Differential susceptibility in the early development of sea-urchins {Arbacia).
A, B, Normal pluteus. C, D, Differential inhibition by dilute KCN; C, applied
throughout development; D, applied for a short period. Inhibition of apical
region (oral lobe) with (in D) correlative increase in basal regions and consequent
parallel-armed condition. E-G, Diffefential acclimatisation in very dilute
solutions. The apical regions become relatively very large, with consequent
wide-angled condition of the arms. (From Child, Physiological Foundations of
Behavior, New York, 1924.)
336 FIELDS AND GRADIENTS IN NORMAL ONTOGENY
produced, which are entirely composed of endoderm, with the
usual exception of a tiny button at the apical end of the large endo-
dermal vesicle. When the treatment is discontinued at the 24-hour
stage, only moderate effects, resulting in exogastrulae, are found.
En.
Fig. 156
Differential inhibition of the ectoderm in sea-urchin larvae reared in water to
which lithium salts are added. Progressive stages of inhibition with increasing
concentrations of lithium, a, is an exogastrula; e, is completely endodermised.
Ek. ectoderm; En. endoderm; Mz. mesenchyme; Pz. pigment cells; Zr. clump
of cells at base of exogastrulated gut. (After Herbst, from Schleip, Determina-
tion der Pri?mtiventzvicklimg , 1929, fig. 323, p. 505.)
When the treatment is not begun until the late blastula or gastrula
stage, death soon ensues, but without any modification of the pro-
portions of the germ-layers, indicating that this is determined by
the mid-blastula stage.
It may be suggested that this result is in part due to differential
FIELDS AND GRADIENTS IN NORMAL ONTOGENY 337
inhibition causing a flattening of the primary (animal-vegetative)
gradient of the egg, the apical portions being more susceptible to
lithium. However, it cannot be due entirely to this, since differ-
ential inhibition brought about by KCN does not result in a rela-
tive increase of endoderm. There must be some more specific effect
of the lithium, though this again is not purely specific, since similar
exogastrulae can be obtained by treatment with the salts of other
alkali metals such as potassium, and such substances as carbon
monoxide. Examination of sea-urchin eggs under dark ground
illumination has revealed the presence of a yellow-coloured ring,
the extent of which appears to coincide with the presumptive endo-
derm. Treatment with lithium raises the upper border of this ring
towards the animal pole and thus provides a visible index of the
degree of " endodermisation". The effect of lithium appears to be
exerted on the colloid structure of the cytoplasm, which it coarsens;
and since in normal development the ectoderm cells present a
finer microstructure than the endoderm cells, it is probable that
this coarsening renders differentiation along ectodermal lines
impossible.^ (See also Appendix, p. 496.)
A remarkable contrast to the '' vegetativised " larvae produced by
lithium are the "animalised" larvae which result from a treatment
of the unfertilised eggs with sodium thiocyanide (NaSCN). Such
larvae show an expansion of the ectodermal region at the expense of
the endodermal : the cilia of the apical organ occupy more than the
normal area ; the gut is smaller or even absent ; and the number of
skeletogenous mesenchyme cells is reduced, even altogether to
zero.^
In such larvae which are completely "ectodermalised", a very
interesting feature is the appearance of a second apical organ at the
vegetative pole: in other words, the original animal-vegetative
gradient has been steepened, and the secondary vegetative-animal
gradient obliterated: its place has been taken by an additional
gradient of the animal-vegetative type, but with its apical point on
the site of the vegetative pole. The polarity of the vegetative half of
the egg has been reversed, and the larva is comparable to a biaxial
head-regeneration in Planaria (p. 285).
If now such an "animalised" larva is subjected to lithium treat-
^ Runnstrom, 1928. 2 Lindahl, 1933 c.
338 FIELDS AND GRADIENTS IN NORMAL ONTOGENY
ment, skeletogenous mesenchyme cells and endoderm are produced
from the equator of the blastula, and two guts are formed, one in
relation to each pole.
It is of further interest to note that the effects of lithium are seen
on the ventral side sooner than on other meridians, thus indicating
Fig. 157
A, Megacephalic, and B, microcephalic, larvae produced by exposing frogs' eggs
to 10 hours' adjuvant and antagonistic temperature-gradients respectively.
Above, external views ; below, sections of head in region of maximum brain
depth. Note difference in size of brain. C, Extremely microcephalic tadpole
produced by exposure to an antagonistic temperature-gradient for 32 hours from
fertilisation, then kept in water for 7 days. (Redrawn, A and B after Huxley,
Arch. Entwniech. cxii, 1927; C, after Tazelaar, Huxley and de Beer, Anat. Rec.
XLVii, 1930.)
the existence and polarity of the dorso- ventral axis (see Chap, iv,
p. 68).
All the evidence therefore goes to show that the main gradient-
systems are concerned with a number of separate physiological
processes which may be variously affected by different agencies.
This is an important extension of Child's views.
FIELDS AND GRADIENTS IN NORMAL ONTOGENY 339
§ 6 . The effects of temperature- gradients
In Amphibia, too, the primary gradient can be experimentally
modified in various ways. One is by superimposing a temperature-
gradient upon it during early development. This has been effected
by several different methods. ^ The gradient may be applied in
various directions, e.g. from side to side across the main axis
(lateral) or along it (polar). In the latter case, the temperature-
gradient may be adjuvant to the egg's original gradient, or else
Fig. 158
Effect of a lateral temperature-gradient, applied for s\ hours from the 2-cell stage,
on cleavage in the frog. The animal cells are larger on the right (cooled) side, small
on the left (heated) side. Note the sharp demarcation between large and small
cells. (From Dean, Shaw and Tazelaar, Brit. Journ. Exp. Biol, v, 1928.)
antagonistic. In the former case, the difference in size between
blastomeres of the animal and vegetative hemispheres is accentuated
at the close of cleavage, whereas in the latter it is reduced, often
to the extent of leaving the animal blastomeres scarcely smaller than
those at the vegetative pole. Various minor anomalies of gastru-
lation are produced, but the net result of adjuvant temperature-
gradients is the production of embryos and young larvae with
somewhat oversized heads, whereas, with antagonistic gradients,
the head region is subnormal. This shows the plasticity of the
^ Huxley, 1927; Castelnuovo, 1932.
340 FIELDS AND GRADIENTS IN NORMAL ONTOGENY
Fig. 159
Effects of temperature-gradients applied to frogs' eggs from soon after fertilisa-
tion, a, b, Preserved during gastrulation. a, After adjuvant gradient for 12 hours :
note marked overgrowth by the dorsal lip and absence of the ventral lip. b, After
antagonistic gradient for 16 hours. Note larger animal and smaller yolk-cells,
presence of ventral lip, and slight overgrowth by dorsal lip. c and d, Preserved as
early neurulae. c (below), After adjuvant gradient for 12 hours. Note small
neural folds (see text), no ventral differentiation of mesoderm, d (above). After
antagonistic gradient for 16 hours. Note large neural folds (see text), well-
formed ventral mesoderm but poor differentiation of notochord and myotomal
mesoderm. (From Dean, Shaw and Tazelaar, Brit. Journ. Exp. Biol, v, 1928.)
FIELDS AND GRADIENTS IN NORMAL ONTOGENY 341
1) u.
to W
fcJD O
^ X
rt
^ ^
a
-d
c
<n
OJ
C
C3
a
g
•"
U-i
D
(U
flHH-1'
^, ^ 06
-il :.
-5 s >
:s «^
■£■£■5
C O ^^
« - t-
<+- 5 "^
o .2 J2
> . - K
? 75 N
<U O a
■> ^h
"rt "rt "^
V- O C
-M •r-i C3
C t! >
<U (U ^
^.£^
X a^
cs; « S
si (U 1— '
U?^ C
■^ u C
OJ IS o
£ ££
.2 ^-^
C ^ =^
£-£-g
S c ?
O rt D
W -^ C =+-!
342 FIELDS AND GRADIENTS IN NORMAL ONTOGENY
primary field-system: in the first case presumptive trunk regions
actually become head, and vice versa in the second case (figs. 158,
160).
When the temperature-gradient is applied after mid-gastrula-
tion, antagonistic gradients often produce neural folds which are
much bulkier than normal, while the opposite effect is produced
with adjuvant gradients.^ This is also true of the mesoderm. As
Gilchrist suggests, this apparently paradoxical effect is presumably
mm $■■■■■ '.-v.
Fig. 161
Three stages in the development of an Amhlystoma embryo treated from the
4-cell stage for 3 days after being symmetrically marked with vital stains, the
whole right half was inhibited by being subjected to abnormally low temperature.
A, On removal from treatment, yolk-plug stage; the normal side shows an
incipient neural fold. B, Later; the left neural fold is well developed, the right
has still not appeared. C, The right neural fold has arisen and has united with the
left; it has, however, been formed out of material to the left of the original mid-
dorsal line. (After Vogt, from Gilchrist, Quart. Rev. Biol, iv, 1929.)
due to the fact that the neural plate is determined by the ingrowing
organiser region, whose high point is vegetative, so that high
temperature at the animal pole is really antagonistic to the processes
leading to neural plate formation (fig. 159).
In a series of experiments in which lateral temperature-gradients
were applied to Urodele blastulae,^ the plane of bilateral symmetry
of the Q^g and embryo was actually shifted towards the warmed
side. It appears that this is in the main due to alterations of
growth of the invaginated organiser; however, since no experi-
ments seem to have been performed in which the application of
^ Gilchrist, 1929; Dean, Shaw and Tazelaar, 1928, text-figs. 6 and 7.
^ Gilchrist, 1928; Vogt, 1928 b, 1932.
FIELDS AND GRADIENTS IN NORMAL ONTOGENY 343
the temperature-gradient was concluded prior to the onset of
gastrulation, we do not know whether the primary gradient-system
may not also be directly deformed (fig. 161). See Appendix, p . 494.
^J
Fig. 162
Effects of a lateral temperature-gradient (3 days from beginning incubation) on
the development of the chick. Above, the whole blastoderm on the heated (right)
side, the area vasculosa much larger and more differentiated, the optic vesicle
moderately larger. The somites on the heated side have been "stepped up" so
as to alternate with those on the other side: this is shown below on a larger scale.
(From Tazelaar, Quart. Journ. Micr. Sci.-hxxu, 1928.)
Somewhat similar results were obtained by applying tempera-
ture-gradients to chick embryos. In some specimens treated with
lateral gradients, the mesoblastic somites on the heated side were
344 FIELDS AND GRADIENTS IN NORMAL ONTOGENY
slightly shifted anteriorly, so as to alternate with those on the
cooled side: the precise meaning of this is not clear ^ (fig. 162).
A curious effect upon cleavage has been noted in some of these
temperature-gradient experiments. It was not infrequently found
that in two sets of eggs from the same batch, one exposed to an
adjuvant and the other to an antagonistic gradient, the yolk-cells
were no more divided in the latter than in the former case, although,
of course, they had been exposed to a much higher temperature.
The cells of the animal hemisphere, on the other hand, were very
much smaller in the adjuvant series. In other words, the develop-
ment of the adjuvant series was more advanced, although its mean
temperature had been the same. This can only be explained by
postulating some effect of the rapid division of the heated animal
cells which stimulates division in other parts of the egg.^
§7. Differential susceptibility in the ontogeny of vertebrates
Experimental modification of the primary gradient of the verte-
brate Qgg has also been achieved by the method of differential
acceleration. Certain treatments produce an acceleration of de-
velopment in all parts, but the acceleration is disproportionately
high in the more apical regions. For instance, by exposing the eggs
of the fish Macropodus to atropin sulphate for an hour and three
quarters during cleavage, the size of the head is increased relatively
and absolutely and it also has altered proportions, for the relative
width of the extreme anterior portion of the animal between the
eyes is much increased.^ Similar results have been obtained in
experiments on the frog,* notably with weak acids, and by means
of differential accHmatisation to very weak poisons (figs. 163, 164).
Equally interesting, and in some ways more instructive, results
have been obtained by the use of depressants on early stages, causing
differential inhibition. The depressant first used was magnesium
chloride,^ acting upon the fish Fundulus, and it was originally
thought that the effects were the specific result of that particular
substance ; but later work has shown that essentially similar effects
^ Tazelaar, 1928. ^ Huxley, 1927; Castelnuovo, 1932.
^ Gowanloch, in Child, 1924, pp. 85-6.
^ Bellamy, 19 19, 1922. ^ Stockard, 1910.
FIELDS AND GRADIENTS IN NORMAL ONTOGENY 345
Fig. 163
Differential acceleration in the development of the teleost fish Macropodus.
A, Control. B, Exposed to dilute atropin" sulphate for if hours during early
cleavage. Note large head, relatively shorter posterior trunk region. (From
Child, Physiological Foundations of Behavior, New York, 1924, after Gowanloch.)
346 FIELDS AND GRADIENTS IN NORMAL ONTOGENY
are produced by a wide variety of depressant substances. This fact
is characteristic of the experimental modification of gradients : any
specific effect of the agent employed is usually overridden by its
general effects which are exerted on the shape and the slope of the
gradient (but see p. 337). Similar results have been obtained with
Anura.i i^ toads, remarkable malformations of the mouth region
are to be noted (figs. 165, 168).
Fig. 164
DiflFerential susceptibility in the early development of the frog. A, Differential
acclimatisation of frog embryo exposed for 4 days from fertilisation to very dilute
KCN. Note very large head. B and D, Differential acceleration. Frog embryos
after 4 and 6 days respectively in NI5000 HCl from the 2-cell stage. Note rela-
tively large head and accelerated development, as against control at 6 days (C).
(Redrawn after Bellamy, Amer.Joiini. Anat. xxx, 1922.)
In moderate concentrations, the result of exposure of Fundulus
eggs to depressant substances is the production of a head of reduced
size, the reduction being disproportionately great in the inter-
ocular region— in other words, the exact converse ^f the experi-
ments with stimulants. But when more marked effects are pro-
duced, they consist in the complete non-formation of a greater or
1 Bellamy, 19 19; Cotronei, 1921.
FIELDS AND GRADIENTS IN NORMAL ONTOGENY 347
eye
Cyclopic frog tadpole produced by treatment with AZ/y lithium chloride for
3 hours in the early gastrula stage. The single median eye is beneath the surface.
The mouth is rudimentary. (From Child, Physiological Foundations of Behavior,
New York, 1924, after Bellamy.)
^
Cyclopia induced by depressant agencies in Fundulus. Above: left, normal
young fish; centre, partial cyclopia, and_^right, complete cyclopia, induced by
treatment with magnesium chloride in stages prior to eye-determination. Below,
side view of the completely cyclopic specimen, showing malformed and ventrally
situated mouth. The treatment leads to the non-formation of the most anterior
and median regions. (After Stockard. from Wells, Huxley and Wells, The Science
of Life, London, 1929.)
348 FIELDS AND GRADIENTS IN NORMAL ONTOGENY
less extent of the apical regions, resulting in animals with eyes in
contact, fused eyes, or a single median eye (cyclopia), and a single
median nostril (monorhiny). The mouth undergoes corresponding
modifications. In Amphibia, the effects may go so far as to give
rise to completely eyeless larvae, often with markedly malformed
mouths. Neighbouring parts are only very slightly affected, and
the trunk region seems not to be affected at all, or to a degree which
would be revealed only by precise measurements^ (figs- 166, 167).
These curious facts can be explained as the result of differential
susceptibility of the different regions of the gradient. A certain
level of activity is needed for the formation of apical (anterior)
structures, a slightly lower level for those next posterior, and so on.
Fig. 167
Effect of lithium chloride on apical structures in anuran development. Left,
control frog tadpole. Right, tadpole from an egg exposed for 3 hours to M/7 LiCl
during early gastrulation ; the anterior head region is inhibited, the external
nostrils {o.p.) are fused, and the eyes close together. (Redrawn after Bellamy,
Biol. Bull. XXXVII, 1919.)
While chemo-differentiation proceeds apicalwards under the in-
fluence of the organiser, the posterior levels of the body can all be
determined. But the extreme apical end, being the most susceptible
to depressant agents on account of its high rate of activity,- is now
in a state which will not permitof the formation of high-level organs.
The material of the apical (animal) region is, however, not destroyed,
and is used up in the construction of subapical structures.
This will explain why certain definite structures are absent from
an embryo which has been exposed to depressants in the early
stages of cleavage, i.e. long before the structures in question have
become determined, let alone differentiated. Another way of
putting this interpretation is to say that the whole gradient has been
flattened out in such a way that its apical end no longer reaches the
threshold potential value needed for the production of extreme
^ Cotronei, 1921. - See Child, 1915 a; Bellamy and Child, 1924-
FIELDS AND GRADIENTS IN NORMAL ONTOGENY 349
apical structures. This has been confirmed by first of all finding
the most susceptible region of Anuran gastrulae, then, in another
experiment, staining this region intra vitartiy and subsequently
producing cyclopia with LiCl and finding that the stained region
gives rise to the prechordal part of the brain. ^ A similar explana-
tion will apply to cyclopia in regenerating Planarian heads (p. 301).
Fig. 168
Effect of lithium chloride, applied for about 24 hours during late gastrulation, on
the mouth region of Bufo vulgaris, (i) Mouth of normal larva, showing horny
beak {b.), rows of horny teeth, and lateral papillae (p.a.). (2-4) Mouth of lithium
larvae ; (2) showing fusion of the two parts of the beak across the aperture, and great
lateral compression ; (3) and (4), mouth reduced to two or one projections, in some
cases without horny teeth ; no beak. (Redrawn after Cotronei, Riv. Biol, iii, 1921 .)
In such poikilothermal systems it is clear that the action of the
gradient-fields cannot be concerned solely wdth the absolute values
of some fundamental process such as oxidation, but with something
more complex, involving primarily the relative values at different
points — in other words, with the form of the gradient rather than
the absolute intensity of the processes constituting it.
^ Guareschi, 1932.
350 FIELDS AND GRADIENTS IN NORMAL ONTOGENY
We may here mention some other experimental resuhs which
may be interpreted on similar lines. It has been seen (p. 156)
that a constriction of the blastula of the newt in the plane of sym-
metry will lead to the formation of two miniature but complete
embryos if the constriction is complete, or of a double monster in
which there are two perfectly formed heads joined on to a single
posterior region of the body, if the constriction is incomplete.
Sometimes, however, the plane of the constriction is not exactly
coincident with the plane of bilateral symmetry, and one half
comes to contain more of the region of the animal pole (i.e. the top
of the gradient) than the other. In such cases, while one of the
heads of such a monster is normal, the other is cyclopic ^ (fig. 169).
The explanation is based on the same considerations as those
already used above. Since by the constriction, one half has been
deprived of the region of the extreme animal pole, that half has a
gradient of which the top is not relatively high enough to form a
perfect head, complete with extreme apical structures ; the other half,
with the complete gradient, is capable of doing this (see fig. 170).
Certain lines of evidence indicate that it is the high point of
the organiser gradient which is affected by lithium, not the high
point of the eggs' primary gradient. ^
§8
From what has already been said in regard to power of regulation,
either in isolated blastomeres (p. 102) or in particular organ-fields,
it should now be clear that regulation in early ontogeny can only
occur while the system in question is in the form of a gradient-
field : it cannot occur when the system is split up into a mosaic of
independent chemo- differentiated regions. A system, be it egg,
blastomere, or field, can only make good the loss of material in so
far as that which was lost only formed part of a field, and was not a
definitely localised determination forming part of an established
mosaic. In regeneration, the new dominant region may override and
remodel what remains of the original organisation.
From this point of view, power of regulation ceases to be a
mysterious force striving for a return to the normal : systems that
can regulate are merely in the same case as the egg, viz. gradient-
^ Spemann, 1904. ^ F. E. Lehmann, 1933, Rev. Suisse Zool. XL, 251.
351
o c o r
<u UJ
o
O
c
.2^
^3
-5 ^
C3
4)
352 FIELDS AND GRADIENTS IN NORMAL ONTOGENY
fields, which, if the expression may be permitted, have not yet cut
their coats, but will do so according to their cloth. It will thus be
apparent that when organisms regulate, they do so for reasons
which are the same as those responsible for normal development,
Fig. 170
Cyclopia in one member of a pair of anterior-doubled monsters in Triton. The
result of oblique constriction of the egg, and exclusion of the animal-pole region
from the half that will give rise to the cyclopic member. (From Spemann, Zool.
Jahrh. Siippl. vii, 1904-)
and the facts call for no transcendent regulative principle such as
is invoked by Driesch in his theory of entelechies. The problem of
regulation is identical with that of certain important phases of
normal development.
FIELDS AND GRADIENTS IN NORMAL ONTOGENY 353
From the experiments in which isolated sea-urchin blastomeres
develop into perfect larvae, Driesch was led to formulate his
principle of "harmonic equipotential systems": equipotential
since parts can give rise to wholes and must therefore possess equal
and complete potencies : harmonic since the product is of normal
proportions and affords evidence of a definite relation-equilibrium
within the system.
Driesch asserted that such systems afforded proof of vitalism.
We may however point out that the requirements of harmonic
equipotential systems are met by the theory of gradient-fields:
relative quantitative differences in activity-rate leading to qualitative
diflFerentiation : localisation being due to relative position along the
total length of a gradient. But it may be doubted whether true
harmonic equipotential systems have any existence in fact. The
1/4 blastomere of the sea-urchin regulates because it possesses the
whole extent of the gradient: divide it transversely, or, an even
more demonstrative case, divide the egg transversely (equatorially),
and no perfect larva will be formed (p. loi). The parts are not all
equipotential, although it may be possible, as in the case of blasto-
meres at the 4-cell stage, to eflFect subdivisions of a system without
segregating regions of different potencies.
The limb-disc has been claimed to be a "harmonic equipotential
partial system ", but it does not appear that this connotation serves
any more useful purpose, or even carries the analysis as far as the
simpler concept of gradient-field, since, as already mentioned
(p. 223), limb-forming potency is unequally distributed round a
sub-central high point.
The results of this chapter may be briefly summed up by saying
that in ontogeny the developing egg, prior to the stage of primary
chemo-differentiation, possesses an organisation in the shape of a
field-gradient system. The unitary and plastic nature of such a
system may be partly obscured by the unequal deposition of raw
materials, or by some degree of determination (though not an irre-
versible chemo-differentiation) having taken place before cleavage
begins. Further, matters are often complicated, notably in verte-
brates, by the existence of a second gradient- system connected with
the organiser.
23
Chapter X
GRADIENT-FIELDS IN POST-EMBRYONIC LIFE
Chap. VIII was concerned with phenomena which could only be
explained by postulating the existence in adult Hydroids, Plan-
arians and Annelids of gradient-fields concerned with morpho-
genesis and reproduction. In higher animals, such as Arthropods
and Vertebrates, in which asexual reproduction does not occur and
in which total regeneration is no longer possible, the existence of
gradient-fields in adult life is not easy to detect. In such forms, the
presence of total axial gradient-fields is especially noticeable during
the earliest stage of development when they constitute the only or
at least the major organisation of the developing embryo. Similarly
the presence of partial (regional) fields is especially noticeable during
the immediately succeeding phase, when the organism consists
essentially of a patchwork of chemo-diflFerentiated regions, each
with its own field but as yet not differentiated into organs.
It might be reasonably supposed that these gradient-systems
were only operative during the stage when they are most noticeable,
and that the organisation of one stage does not persist, but is wholly
supplanted by that of the next stage. This, however, does not in
point of fact appear to be the case, and there is considerable
evidence for the persistence of the gradient-fields of the embryo
throughout life, even in the highest animals.
There is the natural presumption that the gradient-field in
Hydroids and worms is directly derived from the primary gradient-
field of the egg which has persisted into the adult phase. But even
if this be so, in less plastic and more complex types the gradient-
fields might be imagined to fade out at a certain stage of develop-
ment. In what follows, various lines of evidence to the contrary
will be presented.
Examples of the persistence of the main axial gradient of the
organism, as evidenced by its influence upon the polarity of the
later developed regional fields, are to be found in the differentiation
GRADIENT-FIELDS IN POST-EMBRYONIC LIFE 355
of the lateral-line, the limbs, ear, gills, and heart, in Amphibia. The
lateral-line arises from an epidermal rudiment or placode situated
close behind the ear at the early tail-bud stage. It extends down the
side of the body by free growth. This can be observed in experi-
ments where an anterior half of the body of an embryo of the dark-
coloured Rana sylvatica is grafted on to the posterior half of the body
of an embryo of the light-coloured Rana palustiis. The lateral-line
then grows back as a dark structure on a light background. ^
The determination of the path along which the lateral-line will
grow is of special interest in connexion with the concept of
field-gradient systems. If the tail of a frog embryo is cut off and
replaced in an inverted position so that its ventral side is a continua-
tion of the dorsal side of the trunk, one of two things may happen.
If the tail heals on to the trunk perfectly, the lateral-line will grow
back on to the tail and remain at the same level at which it was on
the trunk. This means that it grows along a line on the side of the
tail, along which it would not have grown if the tail had not been
inverted. But if the healing of the tail on to the trunk is imperfect,
and the continuity between them is obstructed by scar-tissue, the
lateral-line, as it growls back on to the tail, changes its level for
the one proper to the tail-region before its inversion (fig. 171).
It is clear that the track along which the lateral-line grows is not
rigidly predetermined, for it can follow a line along w^hich it would
normally not have grown. At the same time, the growth of the
lateral-line is controlled in relation to the field-system of the body,
so that it grows along the antero-posterior axis of the organism at
a certain definite level on the dorso-ventral axis. If the inverted
tail heals on perfectly, it appears that it comes under the control of
the main gradient-system of the embryo, so as to form part of a
single unitary field. This will allow the lateral-line to grow back in
a straight line without changing its level. If, on the other hand,
scar-tissue intervenes between the trunk and the inverted tail, the
latter remains in some important way isolated from the main
gradient-system of the former, and preserves its old field-organi-
sation to which the lateral-line conforms when it comes into its
sphere of influence.
Confirmation of this view is obtained by the experiment of
^ Harrison, 1904.
23-2
356 GRADIENT-FIELDS IN POST-EMBRYONIC LIFE
dsl Dorsal
si gr. t hf gr. e
Dorsal
Ventral
gr.e
gr. I
Dorsal
'— si
Ventral
Fig. 171
Modification of gradient-fields in grafted tail-fragments of frog tadpoles. The
tip of the tail was removed in the embryo and regrafted upside down. When
smooth healing occurred (as in lower figure) the lateral-line growing down from
the trunk assumed a position normal for the intact organism on the dorsally
directed (originally morphologically ventral) side of the muscles of the grafted
piece. When, however, much scar-tissue was formed (as in upper figure) the
lateral-line grew along the morphologically dorsal side of the muscles of the graft,
i.e. in relation to the field-system of the graft, not of the organism as a whole.
ch, notochord; med, neural tube;^r.e, line of fusion of epidermis; gr.t, line of
fusion of myotomes ; /?/, fold in fin ; si, lateral line ; dsl, dorsal branch of lateral
line. (From Harrison, Arch. Mikr. Anat. lxiii, 1904.)
GRADIENT-FIELDS IN POST-EMBRYONIC LIFE
357
grafting the anterior half of an embryo of Rana sylvatica into the
back of an embryo of Rana paliistris from which the rudiment of
the lateral-Hne has been extirpated. The sylvatica embryo has its
antero-posterior axis at right angles to that of the paliistris embryo.
The lateral-line of the sylvatica head grows back normally under
the influence of its own gradient-system, until it reaches the tissues
Sylvatica
Palustris
Fig. 172
Effect of the main gradient-field on the direction of growth of the lateral-line. An
anterior half-embryo of Rana sylvatica (dark) was grafted on to the back of an
embryo of Rana paliistris (light). The sylvatica lateral-line {si), on growing back
to reach the paliistris component, bent back to assume the position normal for a
lateral-line in the posterior region of the body. (From Harrison, Arch. Mikr.
Anat. LXiii, 1904.)
of the palustris embryo. Here it bends round when it has reached
the appropriate level, and continues growing back under the
influence of the gradient-field of th^ palustris embryo^ (fig. 172).
§2
The fore-limb rudiment of Amhlystoma at the early tail-bud stage
is in the form of a disc of mesodermal tissue at the side of the body
(see Chap, vii, p. 222). To each disc there can be ascribed two in-
visible axes — the antero-posterior axis, and the dorso-ventral axis,
defined relatively to the axes of the whole embryo. If a left limb-
^ Harrison, 1904.
358
GRADIENT-FIELDS IN POST-EMBRYONIC LIFE
disc is grafted on to the right side of the body, the proper way up
and the proper way out (" heteropleural, antero-posterior, dorso-
dorsal"), only the antero-posterior axis has been interfered with
and reversed. In such case, the disc develops into a limb with a
left-hand asymmetry on the right side of the body, with elbow
Fig. 173
Diagram illustrating experiments on the symmetry of limbs. The circles
represent the limb-buds as grafted on to the right side of the body. The letters R
and L in the centre of the circles indicate the side of origin of the bud (right or
left). The letters A, P, D, V inside the circle indicate the antero-posterior and
dorso-ventral axes of the grafted bud, these letters outside the circle refer to the
same axes of the body of the organism. The limb which develops is shown with a
thick outline. The position of a reduplicated limb (should one develop; see
footnote, p. 224) is indicated by the fine outline; the dotted line refers to the
form which the limb would have taken if the dorso-ventral axis of the bud had
been fixed at the time of grafting. Only medio-medial combinations are shown.
(After Harrison, from de Beer, Biol. Rev. 11, 1927.)
pointing forwards and hand pointing backwards : it has preserved
its prospective antero-posterior polarity, in spite of the reversal of
this relative to the body as a whole. But if a left limb-disc is grafted
on to the right side of the body, the proper way out but upside
down (" heteropleural, antero-anterior, dorso-ventral"), the antero-
posterior axis has been respected and only the dorso-ventral axis
GRADIENT-FIELDS IN POST-EMBRYONIC LIFE 359
has been reversed. Such a disc, however, develops into a Hmb with
right-hand asymmetry on the right side of the body, akhough it
originally came from the left side, and it is the proper way up : the
palmar surface of the hand is turned down.^ It has failed to preserve
its prospective dorso-ventral polarity, and has acquired a new one
in conformity with its new surroundings (fig. 173).
These experiments show that the antero-posterior axis of the
limb-disc was irreversibly fixed before the time of the operation.
The polarity thus imposed on the limb-disc determines where a
preaxial border (that marked by the radius and first digit) will be.
But the dorso-ventral axis is not yet fixed, and the determination as
to which side will be the palm and which the back of the hand de-
pends on the orientation of the disc with regard to its host. In the
antero-posterior axis of the limb-disc it is easy to recognise the
primary axis of polarity of the embryo. The main axial gradient of
the egg persists, and permeates the limb-disc. The dorso-ventral
gradient of the embryo, however, appears to be less powerful or to
become active only at a much later stage.
As regards the medio-lateral axis, it is found that a limb-disc will
always develop outwards, away from the body, whether it was
grafted the proper or the wrong way out ("medio-medial", or
"medio-lateral"), and this shows that the medio-lateral polarity
is not fixed in the limb-disc stage. ^
^ Harrison, 1921 a; Ruud, 1926.
- A further point of interest in connexion with the grafts of Hmb-discs is that,
at these early stages, it is not "right-handedness" or "left-handedness" that is
determined at all. This is made quite clear from the fact that a left limb-disc can
be made to differentiate into a right-handed limb on the left side of the body by
reversing the antero-posterior axis. (Either, " homopleural, antero-posterior,
dorso-ventral, medio-medial " ; or "homopleural, antero-posterior, dorso-dorsal,
medio-lateral".) The geometrical configuration of right- or left-handedness is
the result of the determination of three axes. One of these, the antero-posterior,
is already determined at the stage operated upon. The second axis, the dorso-
ventral, is determined later, so that grafts of limb-buds of a more advanced stage
of development show a determination not only of the preaxial border, but also
of the palmar surface (Brandt, 1924). The third axis, the medio-lateral, seems
throughout life to be dependent on the orientation of the limb-rudiment relative
to the whole organism, and never to be irrevocably determined. In an adult
newt, the left leg may be cut off, and planted into the dorsal side of the animal
in such a way that the end which was originally proximal now points outwards.
Regeneration takes place from this end, and a bud is formed which proceeds to
differentiate into a right leg. The preaxial border and the palmar surface being
determined as a result of the original antero-posterior and dorso-ventral axes, a
360 GRADIENT-FIELDS IN POST-EMBRYONIC LIFE
A State of affairs which presents many similarities with the deter-
mination of the axes of the hmb is found in connexion with the
development of the ear (see p. 233). The auditory vesicle in Am-
phibia arises as a sac formed from the epidermis on each side of the
neural tube, behind the eye. When it is first determined in normal
development, the ear-rudiment is a field the constituent parts of
which are not yet fixed ^ and it is therefore capable of regulation.
But the field is polarised with reference to the main axis of the em-
bryo, so that if the ear-rudiment is grafted in such a way that the
antero-posterior axis is reversed, it develops with reversed asym-
metry.2 Comparable grafting experiments have shown that in the
rudiments of the external gills and of the heart ^ of Amphibia,^ and
of the operculum in Anura,^ the antero-posterior axis is already deter-
mined at a stage when the rudiment is still in the condition of a
field, capable of regulation after losing a portion of itself.
Another fact of morphogenesis which appears to depend upon the
persistence of the main gradient-system of the organism is the
phenomenon of serial heteromorphosis. As is well known, after
amputation of an appendage, certain Arthropods may regenerate one
of another type, e.g. an antenna in place of an eye in Palaemon, or a
leg in place of an antenna in various Orthopteran Insects. Natural
examples of this have also been found in various groups. The ab-
normally located appendage is, in almost all cases, one belonging
properly to a more posterior region of the body.^ This could be ac-
counted for if it is assumed that the original morphogenetic gradient
persists throughout life, but becomes flattened during later develop-
ment, so that anterior structures now come to correspond to a lower
level of the gradient than they formerly did during early develop-
ment. Since it is known that cold and depressant chemical agents
will flatten a physiological gradient (p. 337), it is by no means un-
reasonable to assume that increasing age will have the same effect.
reversal of the medio-lateral axis in the regeneration- bud results in a reversal of
the asymmetry of the limb regenerated (Milojewic and Grbic, 1925). See also
Harrison, 1925 a.
1 Kaan, 1926. - Tokura, 1925.
3 Stohr, 1925; Copenhaver, 1926. ^ Harrison, 1921 b.
^ Ekman, 19 13. ^ Przibram, 193 1 b.
GRADIENT-FIELDS IN POST-EMBRYONIC LIFE
361
Experimental evidence in support of the view that age is concerned
is provided by the fact that if the antennae of Dixippus are ampu-
tated in the first instar they regenerate as antennae, but if they are
amputated in later instars, they regenerate as leg-like organs. Mean-
while Sphodromantis provides evidence supporting the view that
the effect is correlated with general metabolic activity. In this form,
an amputated antenna will regenerate as an antenna if the animal is
Fig. 174
Diagram to illustrate serial heteromorphosis. In Palaemon (above) removal of the
eye without removal of its ganglion {a, distal cut) leads to regeneration of an
eye {h) ; with removal of the ganglion {a, proximal cut), to that of an antenna-like
organ (c). In Mantids (below) amputation of the antenna in the region of the
flagellum {d,I) leads to regeneration of a fresh flagellum {e) ; in a basal joint (d,III),
to that of a leg (/). (From Przibram, Handb. nor?n. u. path. Physiol, xiv (i), (i).)
kept at 25° C, but at lower temperatures a leg-like organ is formed.
These heteromorphoses are thus presumably produced when the
main gradient of the animal is flattened. The flattening would
be primarily due to age, but can be accentuated by external con-
ditions. The function of these heteromorphoses is of great
interest. (Lissmann and Wolsky, 1933.)
Another main gradient of the early vertebrate embryo is the
asymmetry-gradient (Chap, iv, p. 77), which is responsible for the
asymmetrical disposal of the heart and viscera. Further evidence
362 GRADIENT-FIELDS IN POST-EMBRYONIC LIFE
for the existence and persistence of this gradient is to be found in
the asymmetry of the reproductive system. It will be remembered
that the asymmetry-gradient gives rise to a general preponderance
of the left side. This is apparent also in the gonads. When only
one gonad becomes functional, as in female birds and monotremes,
it is the left. Further, in normal development, e.g. of frogs, the left
gonad in both sexes is usually the larger.^ The left testis is larger
than the right in many species of birds.^ With this may be asso-
ciated the fact that in intersexual mammals the left gonad tends to
be more female, the right gonad more male.^
We may also mention the interesting fact that in genetic Poly-
dactyly in birds, when, as sometimes occurs, the extra digit is formed
only on one leg, this leg is usually the left.*
§4
The persistence of regional fields to later stages has been demon-
strated in adult Vertebrates capable of regeneration, by experi-
ments in which nerves are deflected from their normal course and
left to end in various regions close under the skin. In the newt, for
instance, if a brachial nerve is diverted from its normal course and
led away so as to end freely within a certain area surrounding the
arm, the growth of a supernumerary arm is initiated : if it is led into
an area close to the dorsal fin (or crest), a supernumerary piece of
crest is induced. Similarly, a sciatic nerve deflected into the region
of the arm or of the tail causes an extra arm or an extra tail to arise.
In the lizard the area at the base of the tail can be stimulated to
form a supernumerary tail by the sciatic nerve. ^
Thus round the arm, in the newt, there exists an area which re-
tains the potency of arm-production even in adult life. This area
has been appropriately called the arm-field.^ Similar fields exist
for the tail, leg, dorsal crest, etc. Other evidence, confirming this,
is provided by the experiments recorded in Chap, viii (p. 271),
in which undetermined regeneration-buds of newts grafted into
abnormal situations produced organs characteristic of their new
situations, and not the type of organ by which they had been budded
1 Cheng, 1932. - Friedmann, 1927.
3 Baker, 1926. ^ Bond, 1920, 1926.
^ Guyenot, 1928. '^ Guyenot and Ponse, 1930.
GRADIENT-FIELDS IN POST-EMBRYONIC LIFE
363
out. Each field occupies only a certain definite zone surrounding
the structure to which it gave rise during development. If nerves
are deflected to "frontier" regions between the fields, mixed
structures or chimaeras are produced, partaking of the nature of
both fields. 1
From various lines of evidence, it appears that the action of the
deflected nerves in these experiments is in no way specific, but
merely trophic. What the nerve does is to stimulate proliferation :
the type of structure proliferated is a function of the specific field.
A B
Fig. 175
Effects of deflected nerves ending freely in fields. A, In the limb-field, leading to
the formation of a limb. B, In the dorsal-crest field, producing dorsal crest.-
C, In the tail-field, giving rise to extra tail. (From Guyenot, Rev. Suisse de Zool.
XXXIV, 1927.)
This view is confirmed by other work, carried out on non-breeding
newts, in which a fine silk ligature was tied tightly round the body,
passing over the amputated stumps of the hind-limbs. This was
done in order to produce a mechanical division of the limb
regeneration-buds. In addition to succeeding in this object it
caused an unexpected effect in promoting a local proliferation on
the mid-dorsal line, which developed into a typically crest-like
structure^. This occurred whether the ligature was superficial, or
was passed through below the surface in the dorsal region. (See
also Chap, xiii, p. 430.)
1 Locatelli, 1925; Bovet, 1930. " Milojewic, Grbic and Vlatkovid, 1926.
364 GRADIENT-FIELDS IN POST-EMBRYONIC LIFE
A remarkable effect sometimes occurs after implantation of a
foreign limb-bud in Urodele larvae. The grafted limb may de-
generate, but its presence may stimulate the host-tissue to pro-
liferate and replace the grafted tissue. This was proved by grafting
haploid limb-buds on to diploid larvae. After a time all the haploid
tissue had been replaced by diploid: the formation of a super-
numerary limb by the host had been induced.^ In three cases it
appears that a grafted fore-limb which degenerated after trans-
plantation into the hind-limb held was replaced by host-tissue
which then differentiated into a hind-lim.b. This recalls the move-
ments of cells induced by grafts in Hydra, etc. (p. 301).
Normally, however, the regenerate is formed definitely from the
remainder of the organ ; this is shown in cases where a haploid arm
has been grafted on to a diploid body in Triton, and then the graft
is cut through: the regenerate is entirely haploid. ^ Similar results
are found with the regeneration of Triton limbs grafted hetero-
plastically on to Salamanders.
It is important to note that the morphogenetic properties of the
regional field itself, once they have been determined, are not in-
fluenced by position relative to the whole organism. For instance,
in Salamander larvae, fore-limbs grafted into the hind-limb field,
and then cut through, produce fore-limb structures in regeneration,
and vice versa for hind-limbs grafted into the fore-limb field and
made to regenerate. A portion of the determined field has here been
transplanted, and continues to produce structures of its proper type
irrespective of its position. -
Only an extremely small portion of a determined field is needed
to determine the character of the structures regenerated. If a limb
regeneration-bud, in the stage in which it is still undetermined,
together with a small portion of stump, be grafted into a foreign
field, it will regenerate in accordance with the character of the
stump, not in accordance with the character of the new field as
would have happened if it had been grafted alone ^ (see p. 271).*
^ G. Hertwig, 1927. ^ Weiss, 1924 b. ^ Milojewic, 1924.
'^ The existence of sharply delimited fields diflfering in their histological and
physiological properties is also known from studies on Anuran metamorphosis
(see p. 427) and from work on bio-electric phenomena in the regions of the
mammalian brain (Kornmiiller, 1933).
GRADIENT-FIELDS IN POST-EMBRYONIC LIFE
365
§5
The loss of power of regeneration has been studied in connexion
with the tail and limbs in Amphibia. As is well known, the Urodela
will regenerate tails and limbs even in the adult, but in the Anura
the adult has lost this power, which is present only in the young
tadpole. It is further to be noticed that the Anuran tadpole loses
the power to regenerate its limbs before it loses the power of
Fig. 176
Absence of regeneration after total extirpation of the field. Triton from which
the entire tail-field has been removed. No regeneration at all. (From Guyenot,
Rev. Suisse de Zool. xxxiv, 1927.)
regenerating its tail. This may be compared with the fact that adult
lizards can regenerate a tail, but not a limb.
In analysing the problem as to why the power of regeneration in
the Anura is limited, it is possible straightway to discard the view
that the degree of histological differentiation of the tissues is the
deciding factor. The differentiation of the leg of the adult newt,
with its bony skeleton, functional muscles, and fibrous connective
tissue, is much greater than that of the leg of the tadpole which has
366 GRADIENT-FIELDS IN POST-EMBRYONIC LIFE
already lost its regeneratory power, in which the leg consists simply
of a cartilaginous skeleton, muscles in process of differentiation, and
mesenchymatous connective tissue. It is necessary therefore to look
for another explanation.
Grafting experiments have shown that tails and limbs of tadpoles,
transplanted on to adult frogs and then amputated, can regenerate
in their new position provided that the tadpole from which they
were taken had not already lost its regeneratory power. ^ The in-
ternal environment of the adult Anuran, therefore, does not provide
any factor specifically inimical to regeneration. Nor, on the other
hand, does the internal environment of the adult Urodele provide
any factor specifically helpful to regeneration, for a limb of a tad-
pole of the toad (Bufo) taken after the power of regeneration is lost,
grafted on to an adult Salamandra and amputated there, fails to
regenerate.^
The conclusion, is, therefore that loss of power to regenerate,
however it may originate, comes to operate regionally within the
fields themselves. It is not without interest to find that, in the
Urodeles, power to regenerate is eflFectively stopped if the whole
field is extirpated.^ This has been proved in respect of the snout,
the tail, and the limbs.
§6
Further evidence for the persistence of a total field is derived from
a study oi growth-gradients.^
In the first place the relative growth of parts, including the
phenomena seen in their regeneration, is regulated with reference
to a "growth-equilibrium" which concerns the organism as a
whole. The precise size of any part at any time depends on a
partition-coefficient of material as between the part and the rest
of the body (i.e. all the other parts). The value of this growth-
coefficient differs for different parts of the body, and depends
primarily on factors inherent in the tissues of the organ. If the
growth-coefficient is above unity, the part will increase in relative
size (positive heterogony) ; if below unity, it will decrease (negative
^ Naville, 1927. ^ Guyenot, 1927.
^ Schotte, 1926 a; Guyenot and Valette, 1925; Bischler, 1926.
* Huxley, 1932.
GRADIENT-FIELDS IN POST-EMBRYONIC LIFE 367
heterogony) ; if equal to unity, it will stay constant (isogony). The
external conditions, such as temperature^ and, notably, nutritive
level, will modify the partition of material between various parts
of the body ; but in every case a total equilibrium is concerned in
the process.
Such an equilibrium does not constitute a field-system. How-
ever, it is further found that the growth-potencies of various
regions are frequently graded in a quantitative way, so that the
body appears to be permeated by a field-system of interconnected
growth -gradien ts .
The most clear-cut examples of such growth-gradients are de-
rived from the study of the large chela of Crustacea. When, as in
the males of many species and both sexes of others, these show
marked positive heterogony, they always exhibit a growth-
gradient with subterminal high point. When they are not dispro-
portionate in their growth (approximately isogonic), all their joints
are growing at approximately the same rate — i.e. their growth-
gradient is almost flat. The same is true of the abdomen of female
Brachyura, which shows marked positive heterogony, and has a
well-defined growth-gradient with subterminal or terminal high
point, whereas the male abdomen is almost isogonic and has a very
slight growth-gradient, with central rise^ (fig. 177).
In limbs which show negative heterogony, the sign of the growth-
gradients is reversed. For instance, the limbs of sheep decrease in
relative size after birth; here the girdle is the high point of the
growth-gradient, the foot the low point ^ (see fig. 198, p. 414). In
other cases, growth within an organ is regulated in a more complex
way, though still in a graded pattern. A good example of this is seen
in the antennae of Copepods. (Forfurther details, see Huxley, 1932.)
These gradients may not only act within an appendage, or a
region of the body, but may permeate the body as a whole. Ex-
amples are seen in the relative growth of the appendages along the
axis of the body in hermit-crabs, or in the growth-profiles of male
and female stag-beetles. It is probable that the growth of the diflFer-
ent regions of the body in PlanariaHs also occurs in relation to a
simple gradient (with posterior high point) ; but the available data
only concern themselves with the proportions of the head and of the
^ Przibram, 1917, 1925. ^ Huxley, 1932, p. 83. ^ Huxley, 1932, p. 88.
368 GRADIENT-FIELDS IN POST-EMBRYONIC LIFE
rest of the body. It is of interest to note that during the reduction
in size that accompanies starvation in these animals, the trend in
change of proportions is reversed, the trunk becoming relatively
more reduced in size than the head, so that the proportions of an
animal of a given size are the same whether it has been growing
larger or becoming smaller.^
Special cases of great interest resuhing from graded growth are
those of the shells of Molluscs and Brachiopods. In these cases
2.5
"35 2.0
CO C
C "
.2 <u
o
12 3 4 5 5 7
Segments of abdomen: distal — >
Fig. 177
Growth-gradients in the abdomen of crabs. The abscissae represent growth-
coefficients (differential growth-ratios) of hnear dimensions of abdominal
segments relative to carapace length. The ordinates refer to the abdominal
segments; 7, telson. Solid line, breadth of segments: ® Teltnessus, (S; x the
same, ?; + Pinnotheres (pea-crab), ?. The dotted line refers to segment-length
in Pinnotheres, ?. (From Huxley, Problems of Relative Growth, London, 1932.)
growth takes place at a definite growing edge, and the new material
laid down solidifies and takes no further part in growth. A similar
type of growth is found in other hard structures such as the horns
of mammals, teeth, etc.
D'Arcy Thompson^ first pointed out that the form and size of
the horns of two-horned rhinoceroses could only be understood on
the assumption of a growth-gradient, decreasing posteriorly, in the
head region, affecting the proliferation of epidermal structures.
This is of some general interest, as it can only manifest itself where
Abeloos, 1928.
^ Growth and Form, 1917, p. 612.
GRADIENT-FIELDS IN POST-EMBRYONIC LIFE 369
the centres for horn-formation (which are doubtless specifically
chemo-differentiated regions) are present. In many other mam-
mals, presumably, similar gradients are present, but we are ignorant
of their existence, as no horn-centres exist by which they can mani-
fest themselves.
In all these cases, if growth-potency is evenly graded along the
growing zone, the resultant hard structure assumes the form known
mm.
14-
mid
vrentraJ
dorsal
Ant
Q
■3^
Fig. 178
Persistent gradient-fields affecting feather-growth in adult birds. In all, the
ordinates represent growth of regenerating feathers per day. The abscissae
represent distances within the breast-region, in A and C antero-posteriorly, in
B and D ventro-dorsally. A, In a capon (® single feathers, (^ means). B, In
a cock, in two regions of the breast. C, In a capon. D, In a cock and a hen.
(Based on data of Juhn, Faulkner and Gustavsen, from Huxley, Problems of
Relative Grozoth, London, 1932.)
mathematically as the logarithmic spiral. Slight departures from
a straight-line growth-gradient give rise to departures from strict
logarithmic-spiral form.
The most important of such departures is seen in Molluscs. The
growing edge of the mantle here makes a more or less circular
aperture. If growth is equally graded on the two sides of this
aperture between high and low point of growth-activity, the shell
produced is a plane spiral, as in Ammonites or Scaphopods. If,
HEE 24
370 GRADIENT-FIELDS IN POST-EMBRYONIC LIFE
however, the growth-gradient on the one side of the mantle
aperture is concave upwards, on the other side concave downwards,
the result is what is known as a turbinate spiral — i.e. a form such
as that of a whelk- or snail-shell, characteristic of most Gastropods.
The corkscrew horns of sheep, goats, etc. are due to similar asym-
metrical growth-fields.
Experiments on fowls have shown that here (fig. 178), too,
growth-gradients exist throughout life which affect the rate of re-
generative growth of feathers.^ These gradients diflFer in different
regions, but within a given region are simple in form. Similarly,
there is a gradient in regeneration-rate of anuran larval tail skin.^
Perhaps the most interesting evidence that growth-potencies are
regulated in relation to some form of field-gradient system is de-
rived from a study of the effects of a localised region of high growth-
rate on the growth of neighbouring parts. In general, these are
slightly enlarged, the effect gradually grading away with distance.
This is seen in the increased size of the walking legs on the side of
the large claw in male fiddler-crabs, where the enormous male-
type chela is confined to one side of the body.^ A similar effect on
the walking legs behind the large claw is seen in other Crustacea,
such as Maia and Palaemon, but here, as the large claws are sym-
metrical, it is found on both sides of the body.^ In male stag-beetles
the disproportionate increase in relative size of mandibles with
increase in total absolute size is correlated with a slight increase in
relative size of antennae, and of first as against third legs^.
As regards Crustacean limbs, this effect of a localised region of
intensive growth appears only to be exerted posteriorly, while
anteriorly the result is partly or wholly reversed. In some cases the
induced increase of growth is less in limbs immediately anterior
to the region of intense growth-rate than in those immediately
posterior; in other cases, their growth is even slightly inhibited.
Examples of this positional eflFect are seen in the second and third
maxillipeds of male spider crabs {Maia and Inachus), and in the
first pereiopod of male prawns (Palaemon) in which the second
pereiopod is enlarged as a large claw.^ Such a differential action
^ Juhn, Faulkner and Gustavsen, 193 1 ; Lillie and Juhn, 1932. In this case,
an additional point of great interest is the correlation found between regeneration-
rate and susceptibility to hormones.
3 Clausen, 1932. ^ Huxley, 1932.
GRADIENT-FIELDS IN POS T- EMBRYO N IC LIFE 371
anteriorly and posteriorly to a region of high growth-intensity can
only be explained by postulating some polarised agency connected
with growth-regulation, which extends through the body as a
whole (fig. 179).
260
0 =
240
-
220
-
200
-
180
-
/ '
)60
/ i^
140
// '
\ \
120
^,c
\
<fcr::
100
90
1
. .... .1
. »
3rd.
1st.
2nd.
3rd.
4th.
5th.
Mxpd Pereiopods
Fig. 179
Polarised effect of the presence of a region of high growth-rate upon the general
growth- gradient, in the prawn Palaemon carcinus. Abscissae, growth-rate (per-
centage increase for a hundred per cent, increase in carapace length) of linear
dimensions of the organs represented along the ordinates : third maxilliped, and
first to fifth pereiopods. The large claw is here the second pereiopod (not the first
as in crabs and lobsters), and is much more enlarged in males (solid line) than in
females (dotted line). Correlated with this, in males the appendages posterior to
the large claw show an increased growth-rate (the increase diminishingposteriorly) ,
those anterior to it a decreased growth-rate. (From Huxley, Problefns of Relative
Grozuth, 1932, after Tazelaar.)
24-2
372 GRADIENT-FIELDS IN POST-EMBRYONIC LIFE
These growth-fields continue to operate so long as growth con-
tinues. The processes underlying them are clearly of a different
nature from those concerned with the gradient-systems of the early
embryo, and in higher animals it is uncertain whether they are even
the directly-produced descendants of those gradients. However,
the growth-gradient of a Planarian, as revealed by the relation
between head-size and body-size (p. 287), co-exists with the axial
gradient, and is the reciprocal of it;^ which suggests that the two
gradient-systems may be connected.^
§7
Although it seems clear that the gradient- and field-systems of
the egg and early embryo may persist into later life, this does
not necessarily imply that they persist wholly unchanged. For
instance, the facts of serial heteromorphosis can best be explained
on the view that the primary gradient has been flattened. Then
we have facts such as those concerning the regeneration of skin
in lizards,^ which show that the type of scale regenerated varies
with the external conditions (e.g. temperature). The fact that re-
generated tails produce scales unlike those originally present is
thus presumably due not to "atavism" but to the fact that con-
ditions in the regeneration-bud are different from those in the
original tail-rudiment, a fact which in turn may be correlated
with an alteration of the gradient-systems concerned.
The chief points elicited in this chapter may be summed up as
follows. In the first place, strong evidence is provided for the
persistence throughout life of the primary axial gradient and of
focalised gradient-fields responsible for the morphogenesis of
particular organs, although the precise form and effects of these
gradients may alter with age. Secondly, attention is drawn to the
persistence throughout life of growth-gradients controlling the
relative growth of parts of the body. Here again, both total growth-
gradients and local growth-gradients appear to exist. It is possible,
though not certain, that these growth-gradients stand in some close
relation to the morphogenetic gradients previously described.
^ Abeloos, 1928.
^ For a more detailed discussion see Huxley, 1932, Chap. vi.
^ Noble and Bradley, 1933.
Chapter XI
THE FURTHER DIFFERENTIATION OF THE
AMPHIBIAN NERVOUS SYSTEM
§1
The differentiation of the amphibian nervous system presents a
number of special problems of great interest for the physiology of
development. A large number of experiments have been made on
this subject, and they illustrate so many of the principles which
operate to bring about differentiation, that a chapter may be pro-
fitably devoted to it.
At the blastula stage, as already mentioned, the presumptive
neural fold material occupies a zone in the form of a transverse
band, at right angles to the plane of bilateral symmetry, and passing
close to the animal pole of the egg. This presumptive neural fold
region appears to have received a partial and labile determination in
situ before gastrulation, and this is more marked in the region of the
brain than in that of the spinal cord. Then, during gastrulation, a
streaming movement of the cells of the animal hemisphere takes
place, which results in a shifting of the presumptive neural fold
material, so that it comes to occupy the position of a band running
down the dorsal side of the embryo. At the same time, the
organiser has become invaginated, and having become the noto-
chord, gut-roof and mesoderm, it underlies the neural fold region
and determines it irrevocably to develop by self-differentiation.
As already mentioned (p. 28) the definitive neural tube arises
from the anterior 4/5th of the neural folds, while the hindmost
fifth becomes caudal mesoderm.
The definitive determination of the neural fold field as a whole does
not prevent the possibility of a considerable degree of regulation
taking place within it. This implies, as explained above (Chap, vii,
p. 239), that its various constituent structures have not yet been
individually localised, delimited, and determined. Such further
determination soon follows, however; the region of the cerebral
374
THE FURTHER DIFFERENTIATION OF THE
hemispheres is now determined to evaginate to form vesicles,^ and
the eye-cup, with its stalk, retina, and tapetum, becomes quahta-
tively and quantitatively determined. In the remainder of the neural
tube, centres of differentiation of neurons from the neuro-epithelial
cells, and of their greater or lesser degree of proliferation, are deter-
mined at certain definite places.^ The main lines of the regional
determination of the nervous system are thus completed when the
neural folds have fused with one another to give rise to the neural
tube, and the optic cups have been formed.
a b c
Fig. 1 80
Diagram showing the effect on the differentiation of the neural tube of a, proxi-
mity of a notochord without myotomes; b, proximity of myotomes without
notochord ; c, absence of notochord and myotomes (mesenchymal environment) ;
as seen in transverse section. (From Holtfreter, Arch. Entwmech. cxxvii, 1933.)
At the same time, certain features of the differentiation of the
neural tube are not independent of the presence of other structures.
For instance, the notochord is responsible for the formation of the
ventral sulcus of the central canal, i.e. it determines the formation
of a thin floor on the side of the neural tube immediately overlying
it.^ On the other hand, the myotomes which flank the neural tube
are responsible for the formation of the thick lateral walls, and for
the radial arrangement of the cells in them^. These facts emerge
from experiments in which embryos were obtained possessing a
notochord but no myotomes, or with myotomes but no notochord.
It will be noticed that the action on the differentiation of the neural
tube of both notochord and myotomes tends to the same result. If
Nicholas, 1930.
Coghill, 1929; Detwiler, several works.
^ Mangold and Seidel, 1927; Bautzmann, 1928; Bytinski-Salz, 1929.
* Lehmann, 1926.
AMPHIBIAN NERVOUS SYSTEM 375
the notochord is absent and the myotomes join one another in the
middle hne beneath the neural tube, the latter has a very thick
floor and thin roof, and the central canal extends horizontally instead
of vertically (fig. 180; see also p. 220).
It is clear, therefore, that the normal bilateral symmetry of the
neural tube is dependent on the presence and normal relative
positions of notochord and myotomes. Not only does the notochord
induce the formation of the neural tube (p. 135), but it plays a part
in determining its subsequent differentiation.
If a portion of neural tube is made to develop in a region of
mesenchyme (i.e. deprived of the proximity of notochord and
myotomes) it differentiates with radial symmetry : the walls are of
equal moderate thickness, and the central canal is of circular cross-
section.^
It follows that an environment of myotomes is unlikely to be
conducive to the formation of the vertebrate brain and its numerous
outgrowths and vesicles, and, in point of fact, myotomes are absent
from the neighbourhood of the fore-brain, where the somites are
destined to become the extrinsic eye-muscles. Conversely, in
AmphioxuSj where myotomes flank the neural tube right up to its
anterior end, there is a minimum of cerebral differentiation.
The conditions of the histological differentiation of the neural tube
must now be considered. This consists of the formation in definite
regions of accumulations of the cell-bodies of the neurons or nerve-
cells forming the grey matter, and of the development from these
cells of axons or fibres in definite directions or tracts forming the
white matter. The subsequent morphological differentiation of the
brain is really only the result of the histological differentiation of
neurons in particular places, and the directed growth of their axons.
The prefacial, postfacial, and hemispheric centres in the brain, and
the anterior region of the spinal cord, appear to be places at which
a certain definite number of neurons are determined at the early
neural tube stage to develop by self-differentiation.^ The proof of
this for the centre in the anterior region of the spinal cord is given
^ Holtfreter, 1933 b. 2 Detwiler, 1925 b; Coghill, 1929.
376
THE FURTHER DIFFERENTIATION OF THE
by the following experiment. If a region of the spinal cord corre-
sponding to trunk- segments 1-3 is removed and grafted into the
spinal cord of another embryo in place of the region of segments
4-6, the amount of neuron proliferation shown by it remains
roughly the same as it would have exhibited in its normal position.^
Other regions, however, show dependent differentiation in the
proliferation of neurons. The spinal cord of a newt tapers from front
to back, which means that the tube in the region of the more
posterior segments of the body contains fewer neurons and axons
Fig. 181
Diagram to show the operation of exchanging a region of the spinal cord of
Amhlystoma corresponding to segments 3-5 for a region corresponding to seg-
ments 7-9. (From Detwiler, Natunviss. xv, 1927.)
than do the anterior regions. If a region of the spinal cord corre-
sponding to trunk-segments 3, 4 and 5 is cut out, rotated about its
long axis, and planted back again in the order 5, 4 and 3,^ the
result is the normal differentiation of the spinal cord, with the
proper taper and the proper number of neurons and axons in the
various regions. The same is true if a region of the cord corre-
sponding to trunk-segments 7, 8 and 9 is grafted in place of the
region of segments 3, 4 and 5^. (fig. 181). It is clear, therefore, that
the factors which govern the proliferation of neurons in these regions
of the spinal cord reside elsewhere. In other words, while the spinal
cord is qualitatively self-differentiating, it is still dependent-
^ Detwiler, 1925 b, 1928 a. - Detwiler, 1923 b.
^ Detwiler, 1923 a.
AMPHIBIAN NERVOUS SYSTEM
377
i% f 1/ I '^^
t %^
^1
2 ^
en s*
M Ui g
o /^
C C
M <u
^
<4j
o c o
O bJ3 ^
be C w
. "^ T3 CQ
fci) ^ ,
^ ^ Q
<: T3 o
■^ s
g- O
o > X
^ o "
^ s ~
(u ^^ s: ^->
^ OJ § o
2 _C >P N
3 "? c "
' C O r-T
(U CS ^ 'O
^ "- 'C o
1^ c'J 1:^ ti;
378 THE FURTHER DIFFERENTIATION OF THE
differentiating as regards certain quantitative features.^ But before
dealing with these factors, attention must be paid to the conditions
under which tracts of axons are formed and the direction of their
growth controlled, for they play an all-important part in the pro-
blem (fig. 182).
From the pioneer experiments on tissue-culture,^ and those of
grafting limb-buds ('' aneurogenic ") from embryos whose spinal
cords had previously been extirpated,^ it is known that axons grow
out as free projections from the cell-bodies of the neurons. It
is also important to notice that experiments in which neurons
have been made to produce axons in tissue-cultures through
which an electric current is passing, show that the direction of out-
growth of the axon is controlled by the direction of the current.
Further, if a conductor carrying an electric current is passed
through the culture, the axons grow out from their cell-bodies in a
direction at right angles to the axis of the conductor.^ The strengths
of current used (about 2 billionths of an amp.) correspond in range
with those found in living embryos.
Now, experiments on the differential susceptibility of the parts
of young embryos of Amhlystoma at the early tail-bud stage show
that two axial gradients are present. One of these appears in the
ectodermal tissue of the dorsal side of the body and has its high
end at the head, decreasing posteriorly. This gradient is clearly a
derivative of the original gradient of the primary egg-axis of
polarity. The other gradient is situated in the tissues of the noto-
chord and mesoderm underlying the neural tube, and has its high
end at the high end of the embryo, decreasing anteriorly. This has
been proved by susceptibility experiments.^ Since the hind end of
the embryo, where this second gradient has its high end, corre-
sponds to the point of closure of the blastopore, this gradient must
represent that of which the organiser was the top during earlier
stages of development (fig. 183).
In the embryo of Amhlystoma at the early tail-bud stage there are
therefore two gradients, working in opposite directions. Now the
ventral part of the neural tube is in intimate contact with, and is
^ Yamane, 1930. ^ Harrison, 1907 b, 1910.
^ Harrison, 1907 A. * Ingvar, 1920.
^ Coghill, 1929.
AMPHIBIAN NERVOUS SYSTEM 379
even firmly adherent to, the underlying notochord and mesoderm,
and is under the influence of the second gradient with the high end
posteriorly. The dorsal part of the neural tube is under the influence
of the first gradient with the high end anteriorly. x'Vn illustration of
the action of these two gradients can be obtained from a simple
study of the development of the vertebral column in trout larvae.
The basidorsal cartilages can be seen to develop in cranio-caudal
succession, while the basiventral cartilages appear in caudo-cranial
succession.^ The order of development of the cartilages is pre-
sumably another expression of the gradients.
Several experiments have shown that one of the manifestations
of axial gradients is a difference of potential when the high and low
ends of a gradient are connected with a galvanometer."^ Further,
Fig. 183
Graphs showing the gradients in ectoderm and mesoderm of Amblystoma
embryos, revealed by susceptibility experiments (KCN). (From Coghill,
Anatomy and the Problem of Behaviour, Cambridge, 1929.)
it is known that an electric current can induce an axis of polarity
and a consequent gradient in tissue exposed to it (p. 63).^ Since,
again, an electric current is known to be able to direct the out-
growth of axons, it seems very probable that the gradients in the
body determine the direction of growth of the tracts of axons which
constitute the white matter running up and down the neural tube.*
Careful observation of the initial stages of neuron-differentiation
in Amblystoma have shown that the axons and dendrites arise from
the neurons as processes which creep along the inner surface of the
membrane lining the neural tube, and this creeping always takes
place along the long axis of the tube, i.e. either in an anterior or a
posterior direction. It is therefore very probable that the direction
of outgrowth of these processes from the neurons is governed by
^ de Beer, unpublished. - Hyman and Bellamy, 1922.
2 Lund, 1923 A, 1924. * Kappers, 1917, 1921.
380 THE FURTHER DIFFERENTIATION OF THE
the gradients^ and made to coincide with their axes. The processes
grow up and down the gradients (fig. 184).
But there is a further point to notice. The axons in the dorsal
half of the neural tube conduct impulses forwards towards the an-
terior end and the brain, and form part of the aflrerent or sensory
system. The axons in the ventral half of the neural tube conduct
impulses backwards, away from the brain, and form part of the
Orientation or Neurones
Fig. 184
Sections showing three stages (A, B, C) in the differentiation of neuro-epithehal
cells into neurons. The axon and dendrite processes of the neuron creep along
the inner surface of the limiting membrane of the spinal cord, along its long axis.
The floor-plate cells, in the mid-ventral line, form processes which grow laterally
and then backwards. (From Coghill, Anatomy and the Problem of Behaviour,
Cambridge, 1929.)
efferent or motor system. Reversal end-for-end of a section of the
spinal cord at the tail-bud stage does not alter this plan, and it must
therefore be concluded that the polarisation (as well as the direction
of growth) of the neuron processes is determined by the axial
gradient under whose influence the neuron is situated, in such a
way that a process of a neuron which grows up the gradient from
the low to the high end becomes an axon, and later on conducts in
^ Coghill, 1929.
AMPHIBIAN NERVOUS SYSTEM 381
this direction, while the processes which grow down the gradient
become dendrites. It can then be easily understood why the dorsal
half of the spinal cord (subjected to the ectodermal gradient from
front to back) should contain afferent axons conducting forwards,
and the ventral half (subjected to the mesodermal gradient from
back to front) efferent axons conducting backwards.^
§3
A further application of the principles stated above gives a formal
explanation of the main architecture of the peripheral nervous
system, characterised by the formation of paired nerves growing
out in each segment of the body, at right angles to the spinal cord.
By this time, tracts of axons are present running along the spinal
cord, and one of the results of the passage of an impulse through
these tracts is the setting up of an electric disturbance, analogous
to the passage of an electric current. A neuron under the influence
of such a current will produce an axon which will grow out at right
angles to the direction of the current, as in the tissue-culture
through which a conductor carrying an electric current is passed.
In the chick it has been observed that this outgrowth of neurons at
right angles to the spinal cord normally occurs as the axons of the
" activating bundle " reach their level. ^ In Amblystoma, it has been
found that isolation of a portion of spinal cord from the medulla
(by grafting it into the side of the body), with consequent reduction
in the number of descending fibres, leads to quantitative reduction
in the development of the ventral nerve-roots.^
On each side of the neural tube, the mesoderm becomes seg-
mented into myotomes, or muscle-segments, and within each of
these there is evidence of a gradient : the high point being in the
centre and the activity-rate grading off forewards and backwards
from this central point. The existence of these gradients is ex-
pressed by the distribution of the pigment, since, in the develop-
ment of Amphibia generally, pigment is formed most abundantly
in regions of high activity-rate. In the developing muscle-segments,
pigment is usually accumulated near their centres. The septa
between the segments are therefore regions of low activity.
^ Coghill, 1929. 2 Bok, 1915. 2 Yamane, 1930.
382 THE FURTHER DIFFERENTIATION OF THE
These conditions have a bearing on the direction of growth of the
peripheral nerve-fibres when they have emerged from the spinal
cord. Those fibres which emerge in the ventral region of the cord,
continuing to grow along a gradient (and eventually becoming
differentiated into axons, since they are growing up the gradient),
will accordingly grow to the centre of each muscle-segment, and
innervate it. On the other hand, the regions of the septa, between
the muscle-segments, will attract the dendrites of the sensory
neurons,^ which will then grow to the ends of the muscle-segments
(thus providing their proprioceptive innervation), and continue in
the septum between the muscle-segments to the skin.
Both in the sensory and motor systems, therefore, the distribu-
tion of the peripheral nerves can be interpreted in terms of gradi-
ents : axons growing towards a region of higher rate, and dendrites
towards a region of lower rate. Within the central nervous system
itself, the same principle can be applied. Experiments of differ-
ential susceptibility on the spinal cord indicate that a strip of tissue
occupying the ventral mid-line, and forming the so-called keel,
has, during late embryonic life, the highest activity-rate at any
given level of the cord : this is also proved by the fact that the keel
is the site of the most rapid differentiation of neurons in the spinal
cord. It is most interesting to find that during this period any axon
outgrowths formed in the transverse plane are directed towards the
keel.
In the brain, other centres of differentiation of neurons are the
postfacial and prefacial centres, already mentioned, and, further
forward, the dimesencephalic, the postoptic and the hemispheric. Up
to the early swimming stage, the postfacial centre is the most active,
as evidenced by the relative rate at which neurons are differentiated
there, compared with the rate in other centres. Correlated with
this fact, it is found that the first neurons to become polarised in the
dimesencephalic centre send out axons towards and into the pre-
facial centre. In a similar way, all over the brain, neurons which
are differentiated in the neighbourhood of a centre grow axons to-
^ The sensory neurons considered here form part of the transient sensory
system of Rohon-Beard. They differ, of course, from the sensory neurons of the
definitive system in that they are situated in the neural tube instead of the dorsal-
root ganglia. Eventually, the Rohon-Beard neurons are superseded by the latter.
AMPHIBIAN NERVOUS SYSTEM 383
wards that centre, and a number of tracts and commissures are
formed, including the olfactory paths, the posterior and postoptic
commissures, and paths between the thalamus and hypothalamus.
An interesting but as yet unexplained point is that the relative
rates of activity of the various centres, measured by the rate of
neuron-differentiation, do not remain constant. At one period, the
hemispheric centre is more active than the olfactory, but later on
the olfactory is more active than the hemispheric. This state of
affairs allows of the formation of the reciprocal paths which are so
characteristic of various parts of the brain. It is clear, therefore,
that a simple application and extension of the principles of axial
gradients go a long way towards explaining the problems connected
with the laying down of the main lines of the systems of tracts in
the central nervous system, and in the peripheral nervous system,
although the determination of the time-relations still remains
obscure.
§4
It is now time to revert to the question of the factors which control
the proliferation of neurons in the spinal cord, in regions other
than those in which their proliferation at certain definite centres is
the result of a previous determination, followed by self -differentia-
tion. It has been found that the sensory load, as given by the
number of receptor-organs, is the governing factor controlling the
number of sensory neurons, but that the motor load, as given by
the number of muscle-fibres to be innervated, has no effect upon
the number of motor neurons. It is the number of axons which end
in any given place that determines the proliferation of neurons at
that place, but the endings of dendrites have no such effect. Thus,
planting an extra limb in the side of the body increases the amount
of muscular and epidermal tissue present ; it has no effect on the
number of motor neurons in the ventral region of the spinal cord,
but it results in an increase in the number of sensory neurons in the
dorsal-root ganglia^ at the level of the graft.
Removal of the skin from one side of the body (effected by graft-
ing together side by side two embryos each of which has had the
skin removed from one side) does not affect the number of motor
^ Detwiler, 1920 a; Carpenter, 1932, 1933; Carpenter and Carpenter, 1932.
384
^
-A.
t-.
::r^
"^
n
k
H-:
N
r>^
>
•^
^:^,
'^
F^
nr-Q
-^
p^
r-H
fv
7^
b--"
N.
^
N
^■.
'^
4
d]
^•^
>«ws=
^5^=6
M
r
J
^
^
?
t
7
8
S
4-5
4-0
3-5
3-0
2-5
2-0
1-5
1-0
0-5
^^J ^5 6- 7 8 3
Fig. 185
Graphs illustrating the degree of proliferation in the spinal cord and in the dorsal-
root gangliaof the trunk region of axolotl larvae. Abscissae, trunk-segments 3 to 9.
Ordinates, weight as estimated by wax model reconstructions of the various
regions. Graph a. Weight of entire spinal cord in segments indicated. Graph h.
Weight of grey matter of spinal cord in segments indicated. Graph c, Weight of
dorsal-root ganglia in segments indicated. Curve A (full line), normal larva;
curve B (broken line), larva in which the fore-limb was grafted farther back and
innervated from segments 5, 6 and 7 ; Curve C (dotted line), larva in which spinal
cord segments 7, 8 and 9 were substituted for segments 3, 4 and 5. Note that
the motor cell-area (ventral region of grey matter of spinal cord) is little or not
affected by interchange of segments or transposition of limb, but that the sensory
cell-area (dorsal-root ganglia) is markedly affected by transposition of limb.
(From Mangold, Ergebn. der Biol, in, 1928; after Detwiler.)
^
i:<
vl
^
N
t-
-^
>.
/
\
\^
-■-•..
\
B/
/
^
\
\
^■-;d
<^«.^
— .^
- /
/
•
>^
^
r
DIFFERENTIATION OF AMPHIB I AN NERVOUS SYSTEM 385
cells in the ventral region of the spinal cord, but it results in a
60 per cent, decrease in the number of sensory neurons in the
dorsal-root ganglia of that side.^ Unilateral removal of the
muscles (which of course contain proprioceptive receptor organs),
without injuring the skin overlying them, results in a 40 per cent,
decrease in the number of sensory neurons in the dorsal-root
ganglia of that side, again without affecting the number of motor
neurons in the ventral part of the spinal cord.^ Incidentally, it
may be observed that these results give interesting information as
to the proportion in which exteroceptive and proprioceptive
Fig. 186
I, Amblystoma embryo showing area of epidermis removed. 2, Two embryos
grafted together and lacking epidermis, and therefore sensory load, on their
inner sides. (From Detwiler, 3^07^;7z. Exp. Zool. xlv, 1926.)
neurons occur in the dorsal-root ganglia, since the skin contains
exteroceptors and the muscles proprioceptors.
All these experiments have an effect only on the number of
neurons in the dorsal-root ganglia, i.e. of sensory neurons, and the
effect has in all cases been due to an increase or decrease in the
sensory region from which the sensory cell-area (the ganglia) in
question receives impulses. On the other hand, the number of
motor neurons in the ventral region of the spinal cord, i.e. the multi-
plication of cells in the motor cell-area, is controlled by the number
of endings of axons of the descending tracts of the cord (tractus
bulbo-spinalis). Thus if the region of the first five segments of the
spinal cord is removed, and in its place an extra medulla oblongata
^ Detwiler, 1926 a. ^ Detwiler, 1927 a.
HEE 25
386 THE FURTHER DIFFERENTIATION OF THE
and first two spinal segments from another embryo are grafted
(fig. 187), there will be two medullas, and an increased number of
fibres in the descending tracts as compared with the normal. The
result is an increase in the number of motor neurons in the ventral
1.0 \ MEDULLA
272
1.0
1.0
Weight
Patios
Case
ERL8
SSI
D.89
245
SS2
D.70
218
SS3
153
SS4
0.39
130
SS5
0.35
108
SS6
0.25
93
0.23
1 SS7
\ 77
0.90
0.80
0.56
0.47
0.40
0.34
0.28
Cell
Ratios
0.74
0.92
0.77
0.40
0.31
Weight
Ratios
MEDULLA,
233
156
Trans-
plsuited
Medulla
180
TrSSl
277
TrSS2
243
SS6
113
SS7
90
Case
TrBrSc 1
1.0
0.65
0.77
1.15
1.04
0.48
0.38
Cell
Ratios
Fig. 187
Diagram showing the number of cells in various regions of the medulla and spinal
cord of Ambly stoma in a normal embryo (case ERL8) and in an embryo from
which spinal segments 1-5 were removed and replaced by an extra medulla and
spinal segments i and 2. SSi, SS2, etc., spinal segments i, 2, etc. TrSSi,2,
grafted spinal segments 1,2. The figures within each segment represent the
average number of cells seen in a transverse section. The weight ratios are
estimated from the weights of reconstructed wax models of the various regions.
(From Detwiler, Quart. Rev. Biol, i, 1926.)
AMPHIBIAN NERVOUS SYSTEM 387
region of the grafted first and second segments;^ this increase is
quantitatively proportional to the size of the extra region im-
planted.^ That the medulla is the region from which the tracts
responsible for this proliferative effect originate, follows from the
fact that interference with higher levels of the brain (e.g. removal
of the mid-brain) produces no effect on the normal proliferation
of neurons in the spinal cord,^ while removal of the medulla is
attended by a reduction in the number of neurons in the ventral
region of the spinal cord. At the same time, the nature of the action
exerted by the medulla is complex, for if a medulla is grafted in place
of segments 4, 5 and 6, of the spinal cord of another embryo, the
graft exerts no proliferative effect on posterior regions, and itself
undergoes no more proliferation than is typical for its position.^
The centre of independent high rate of proliferation at the
anterior end of the spinal cord (see p. 375) appears to exert an in-
fluence on the rate of multiplication of neurons in the dorsal or
sensory cell-area anterior to it ; for when the first three segments of
the cord are grafted into a more posterior position, so as to occupy
the position of the third, fourth and sixth, the intact segments
anterior to the graft show a higher rate of proliferation than
normal, and in some respects come to resemble a medulla oblongata.
It is therefore possible that normally the medulla may be dependent
for its rate of neuron-proliferation on influences emanating from
the anterior end of the spinal cord, possibly by way of the neurons
of the spino-bulbar tract. ^
The proliferation of neurons is, we see, under the control of
factors which are situated " upstream " relatively to the direction in
which nervous impulses will eventually be conducted by the axons
to the cell-area in question. But it is not the passage of ordinary
nervous impulses that is responsible for this effect, for embryos can
develop normally in a solution containing narcotics which prevent
the passage of impulses.^ The proliferative effect must therefore be
due to some other activity of the axon-endings. It is very possibly
identical with the trophic effect of adult nerves, which, also, is not
identical with the ordinary conducting function (see p. 431).
^ Detwiler, 1926 b. ^ Nicholas, 1931. ^ Nicholas, 1930.
* Detwiler, 1927 b. ^ Detwiler, 1928 A, 1929 B, 1930 A.
^ Matthews and Detwiler, 1926.
25-2
388
THE FURTHER DIFFERENTIATION OF THE
The discovery of the factors controlling neuron-proliferation is of
theoretical interest from two further points of view. In the first
place, it is clear why the brain differs in form from the spinal cord.
The anterior end of the body is occupied by the organs of special
sense, from which an enormous number of fibres enters the neural
tube. These fibres induce the multiplication of neurons where they
The effect of eye-extirpation on the development of the mid-brain. Transverse
section through the mid-brain of a larva of Rana fusca from which at an early
stage an eye was removed. The roof of the mid-brain on the operated side is
markedly under-developed (left half of figure), as may be seen by comparing it
with that of a normal control (right side of figure) ; the outer molecular layer (2)
and the stratum medullare superficiale (6) are absent on the operated side, (From
Diirken, Biol. Gen. vi, 1930,)
end, and the large numbers of neurons so formed find expression
in the bulges and prominences familiar as the optic lobes, restiform
bodies, and olfactory lobes, which differentiate the brain morpho-
logically from the spinal cord. It has been shown that the grafting
of an extra eye or an extra nasal pit on the head results in an in-
growth of fibres from the graft to the brain, and an increase in the
number of neurons in the brain at that point. ^ Conversely, the
^ May and Detwiler, 1925; May, 1927.
AMPHIBIAN NERVOUS SYSTEM 389
extirpation of sensory organs may cause a reduction in size of the
brain-centres to which their fibres normally run^ (fig. 188).
Secondly, these results help to supply a partial explanation of
some of the phenomena of neurobiotaxis. It has been observed
in comparative studies that corresponding centres of neurons in
different animals may occupy different positions in the brain, or, in
other words, that certain nerve-centres have shifted their position
during the course of evolution. The centre of origin of the motor
fibres of the facial nerve is situated near the centre of the medulla
oblongata in the selachian, but it lies on the floor of the medulla in
mammals.^ In each case, the nerve-centre lies close to the endings
of the axons from w^hich it habitually receives axons. Actually,
this displacement of the nerve-centre in phylogeny (the "march to
the sound of the firing", as it has fancifully been called) is only the
result of a phylogenetic change in the positions of the axon endings.
The cause of such change is another question, still obscure, but its
effect has been the proliferation of neurons and the formation of
nerve-centres in the changed positions, in each successive ontogeny.
The nerve-centres are localised and differentiated afresh in each
generation, and this may take place in new positions if the axons
(from which the centre habitually receives impulses in the passage
of reflex arcs) end in new positions.
§6
With regard to the peripheral nervous system, interesting results
have been obtained bearing on the question as to how the nerve-
fibres become connected up with their end-organs. Two different
kinds of factors appear to be at work. In the first place, the out-
growth of the nerve-fibre in the direction of the end-organ is
controlled by non-specific factors; while its intimate functional
connexion with the end-organ is controlled by factors specific to
the organ.
As an example of the general directive effect which is exerted by
the presence of an organ, we may take that of the limb-rudiment.
If in an embryo of Arnhlystoma the limb-rudiment is moved some
distance forwards or back from its normal position, the nerves
^ Durken, 1912. ^ Kappers, 1930.
390
THE FURTHER DIFFERENTIATION OF THE
which normally supply the limb do not grow to the place where it
ought to be, but to the place where it is, provided that this is not
more distant than two or three segments away from its normal
position.^ If this distance from the normal position is exceeded, the
limb becomes innervated by other nerves, corresponding to the
level of its position, which would not normally have supplied
a limb at all. In Amblystoma, the normal supply to the fore-limb
is composed of fibres from spinal nerves 3, 4 and 5, forming the
Fig. 189
The attraction of outgrowing nerve-fibres towards an abnormally situated limb.
a, The constitution of the normal brachial plexus of an axolotl formed from spinal
nerves 3-5. b, The brachial plexus of an embryo in which the limb-bud was
moved five segments further back; the plexus is formed by spinal nerves 5-9.
(From Mangold, Ergehn. der Biol, in, 1928. after Detwiler.)
brachial plexus. But the plexus can be formed from spinal nerves
2, 3 and 4, or 5, 6 and 7 (sometimes with the co-operation of
additional nerves) (fig. 189).
The attraction which the limb exerts on the outgrowing nerve-
fibres is shown still more clearly by experiments in which the whole
of one half (the right) of the rudiment of the spinal cord of the frog
is removed at the neurula stage. No nerves at all grow out from the
right side towards the hind-limb, but fibres from the sciatic plexus
^ Detwiler, 1920 b, 1922.
AMPHIBIAN NERVOUS SYSTEM 391
of the left side turn back across the mid-Hne, and innervate the right
hind-leg. In general, it appears that the pattern of the plexus
formed is largely independent of the amount, origin, and direction
of ingrowth of the immigrant nerve-fibres, and is determined by
factors intrinsic to the limb.^
In yet other experiments, on newt larvae in which a fore-limb is
grafted into the flank of the body close to an intact hind-limb, and
the nerve to the hind-limb is severed, the nerve regenerates and in-
nervates both autochthonous hind-limb and grafted fore-limb. The
actual details of innervation vary in each experiment. It is possible
for the grafted fore-limb to be completely innervated by branches
of the third lumbar nerve, which normally supplies only the ad-
ductors of the femur and the flexors of the knee. This shows that
nerves may be attracted towards and innervate muscles different
from those which they normally supply.^ The same conclusion
emerges from experiments on Amblystoma in which a limb-rudi-
ment is partially removed at the early tail-bud stage, and grafted
back into the same embryo at a distance of four segments posterior
to the normal position. From the remainder of the rudiment in the
normal position a limb is also formed, so that the embryo has two
fore-limbs on the same side, and the nerves of the brachial plexus
may be supplied to both.^ In these cases, an additional point of
interest is the fact that both limbs show simultaneous movement
of homologous muscles, although the actual nerve-fibres which
innervate them may be quite different, and their distribution varies
in each individual case.^
The attraction which is exerted by a limb on a growing nerve is
even less specific than would appear from the experiments just
mentioned, for it is also exerted by an eye or a nasal pit, grafted
on to the side of the body of a larva (in Amblystoma) , after removal of
the limb-rudiment.^ In these cases the nerve-fibres which would
normally have innervated the limb grow towards the eye or the
nasal pit as the case may be, and end in the tissue immediately
surrounding it.
^ Hamburger, 1927, 1929. ^ Weiss, 1924 A. ^ Detwiler, 1925 A.
* Experiments of this type have led to the so-called resonance theory of nerve
action. See Weiss, 1924 a, 1928 ; Versluys, 1927, 1928 ; Detwiler, 1926 c, 1930 B,
c; Detwiler and Carpenter, 1929; Detwiler and McKennon, 1930.
^ Detwiler, 1927 C
392 THE FURTHER DIFFERENTIATION OF THE
The same non-specific attraction has been shown in the case of
grafts of rudiments of chick embryos on to the chorio-allantois. If
the rudiments include those of the mid-brain, muscle-segments,
cartilage, and mesonephros, it is found that nerve-fibres grow out
from the mid-brain towards them. Normally, the neurons of the
mid-brain produce axons which do not emerge from the central
nervous system, but form visual association neurons. Under the
conditions of the experiment, however, they are attracted towards
the various structures which happen to be differentiating in prox-
imity to them.i It may also be noted that in these experiments the
mid-brain is not enclosed in a connective tissue capsule, so that
there is no mechanical obstacle to the outgrowth of axons. ^
It would appear that this non-specific attraction is a result of a
high degree of physiological activity on the part of the structure
exerting the attraction ; and in a general way the growth of a nerve-
fibre towards such a structure may be compared with its growth up
and down the gradients within the neural tube. It should also be
noted that the deflection of nerves to an abnormally situated graft
is greater if the graft is a limb than if it is an eye.
A structure or organ which is already innervated appears to exert
no attractive effect on a growing nerve ; it is, as it were, saturated.
This fact emerges clearly from experiments on Amblystoma in
which the limb-rudiment is removed and a tail-rudiment is grafted
on to the side of the body, some distance behind the normal limb
position. Contrary to what happens when a limb, an eye, or a nasal
pit is grafted, no nerves grow out towards the tail. This is presum-
ably because the tail contains its own little piece of neural tube, the
nerves from which provide for its own innervation.^ It must be for
this reason that in those cases where a limb is transplanted to an
abnormal position, the brachial nerve (which is attracted by the
^ Hoadley, 1925.
2 See also Detwiler, 1928 a. A similar alteration of morphological process in
the absence of a retaining capsule is seen in the lens. When lens-rudiments are
grafted into blastulae, they develop as regular spheroids if their limiting membrane
remains intact. If, however, it is locally damaged, a large irregular protrusion of
fibre-elements occurs (Kriiger, 1930). In a somewhat similar way, the normal
absence of capsule round the thyroid of teleost fish permits a pseudo-mialignant
growth of the organ if it is induced to hypertrophy, while this is impossible with
the encapsulated thyroid of higher forms (Marine and Lenhart, 191 1).
^ Detwiler, 1928 b.
AMPHIBIAN NERVOUS SYSTEM 393
muscles of the limb) is not attracted by the muscle-fibres of its
segmental myotomes, for the latter are already innervated whereas
the muscles of the limb are not.
It is to be noted that when a limb is grafted to an abnormal
position, nerve-fibres are not only attracted to it, but they form in-
timate functional contact with its muscles. An eye or a nasal pit,
on the other hand, can attract the nerve-fibres to their vicinity, but
no more ; no intimate functional contact is established. These facts
have led to the view that the establishment of functional contact
and innervation is controlled by factors of a specific kind for each
type of structure, possibly chemical in nature.^ If this hypothesis
should turn out to be justified, then, in the outgrowth of a nerve-
fibre and its functional innervation of an end-organ, both non-
specific and specific factors would be involved.
§7
We may now turn to the differentiation of the cells of the neural
crest. Many of these, of course, give rise to the neurons of the
dorsal-root ganglia, but it appears that the metamery and differ-
entiation of the ganglia is dependent on the presence of the seg-
mented myotomes. If at the tail-bud stage of Pleiirodeles the myo-
tomes are removed from one side of the trunk without damaging
the neural crest, the resulting embryo lacks spinal ganglia on the
operated side.^ Similarly, the spinal ganglia fail to develop norm-
ally in experiments in which portions of spinal cord are grafted
without myotomes into the flank of other ernbryos.^ Conversely,
the interpolation of an extra myotome as a result of grafting leads
to the formation of an extra spinal ganglion.^
In addition to giving rise to neurons, some of the cells of the
neural crest have been experimentally shown to produce the sheath
cells, which enclose the peripheral nerves. If in Amhlystoma the
neural crest is removed in the region of the trunk, no dorsal nerve-
roots or ganglia are developed: the ventral nerve-roots develop
normally, but have no sheaths. On the other hand, if the ventral
1 Cajal, i9o6;Tello, 1923.
^ Lehmann, 1927.
^ Yamane, 1930.
* Detwiler, 1932, 1933 b.
394
THE FURTHER DIFFERENTIATION OF THE
half of the spinal cord is removed, the dorsal nerve-roots are un-
affected, and their nerves possess sheaths in the normal way.^
Removal of the neural crest in the region of the head leads to re-
sults which are in many ways remarkable, and difficult to interpret.
It is found that embryos of Amhly stoma punctatum from which the
neural crest of the head has been extirpated on one side show de-
ficient chondrification of the anterior part of the trabecula cranii
and of the cartilages of the visceral arches, including the jaws and
7.^*77
Fig. 190
Left side view of the chondrocranium of a larva of Amblystoma showing (shaded
by dots) the regions which fail to develop after extirpation of the neural crest.
Au.cap. auditory capsule; B.oc. basal plate; Cbr. ceratobranchial ; Chy. cerato-
hyal; C.Tr. orbital cartilage; Ex.oc.M. oculomotor nerve foramen; Ex.op.N.
optic nerve foramen; M. Meckel's cartilage; Q. quadrate; Tr.B. trabecula;
Vert, first vertebra; i Bb., 2 Bb. first, second basibranchial ; 1-4 Ebr. first to
fourth epibranchial. (From Mangold, Ergebn. der Biol, in, 1928, after Stone.)
branchial arch skeleton. ^ These results have been confirmed on
Amblystoma mexicanum^ and Rana.^ It is known that derivatives
of the cells of the neural crest extend ventrally at early stages into
the region of the visceral arches, and it would seem from these ex-
periments that these cells became directly converted into cartilage
cells. Conclusive proof would be obtained if intra vitam stains in
the neural crest at the neurula stage could be found in cartilage cells
at subsequent stages: some authors, indeed, working with de-
scriptive methods only, have professed to see special histological
^ Harrison, 1924 b.
^ Raven, 1931 b.
2 Stone, 1926.
^ Stone, 1929.
AMPHIBIAN NERVOUS SYSTEM
395
characteristics in the cells of visceral arch cartilage, and to have
traced them back to the neural crest cells^ (figs. 190, 191).
Experiments in which the neural crest cells were stained intra
vitam have not yet demonstrated the presence of the stain actually
l.CBR.
Fig. 191
Chondrocranium of a larva of Rana palustris from which the neural crest was
removed on the right-hand side; note reduction of trabecula and visceral arches.
Letters as in fig. 190. Also: IR. infra-rostral; PO. pterygo-quadrate; SR.
supra-rostral. (From Stone, Arch. Entwmech. cxviii, 1929.)
in the cartilage cells, although the colour can be seen in the correct
position in the living state. ^ Presumably, by the time the cartilage
is differentiated, the stain has beerf dissipated. However, definite
proof of this potentiality of neural crest cells has recently been
1 For the morphological bearing of these facts, see de Beer, 1930.
2 Stone, 1932.
396 DIFFERENTIATION OF AMPHIBIAN NERVOUS SYSTEM
provided by grafting experiments^ in which portions of neural fold
from the head-region of the early neurula were grafted into the
ventral epidermis of other neurulae, and there produced cartilage
as well as nerve-cord and ganglia. Grafts of the corresponding
presumptive region of the late gastrula only produced nerve cord
and ganglia : it would appear that the capacity to produce cartilage
is determined later than that to produce neural elements. From
other experiments, it appears that neural crest tissue has the power
of determining other tissue (e.g. presumptive epidermis) to dif-
ferentiate into cartilage" (see p. 193), and this might be taken as a
case of homoiogenetic induction.
Experiments on heteroplastic grafts of axolotl tissues into
Triton hosts have shown that the neural crest cells in the trunk-
region also may have various prospective fates. While some of
them give rise to the trunk spinal ganglia, others migrate in the
form of mesenchyme to the outer side of the myotomes, and
into the dorsal and ventral fins.^
Further differentiations of the nervous system may occur under
the influence of hormones. Strictly speaking, such cases fall
beyond the scope of this book. But we may mention the well-known
fact that human cerebral development is incomplete without the
presence of a sufficiency of thyroid hormone. Another case of brain
differentiation under the influence of thyroid is seen in Amphibia.
Here a marked change in the proportions and shape of the parts of
the brain occur at metamorphosis. Thyroidectomised tadpoles
preserve in the main the larval type of brain.* Further, the
morphogenetic changes occurring in the amphibian brain at
metamorphosis are known to be accompanied by psychological
changes. Salamander larvae can be tamed and trained to take
food out of the human hand; but this habit vanishes completely
from the day of metamorphosis.^ This 'forgetting is clearly
due, not to a psychological process of suppression ' (as suggested
by W. H. Rivers in his Instinct a?td the Unconscious, 1920), but
to morphological changes in the nervous system.
^ Raven, 1933 a. ^ Holtfreter, 1933 b. ^ Raven, 193 1 b.
* B. M. Allen, 1924. ^ Flower, 1927.
Chapter XII
THE HEREDITARY FACTORS AND
DIFFERENTIATION
§1
One of the most important resuks obtained from the experimental
study of development is the fact that all the evidence points to the
equality of nuclear division as being the general rule. Also, many
of the results of regeneration would be unintelligible except on this
idea. Now, genetic research has revealed the existence of unit
hereditary factors or genes, whose only visible effect is upon some
local characteristic of the organism. For instance, in Drosophila,
there exist genes whose primary effect is to modify the colour of
the eye, while other genes are more particularly concerned with the
shape of the wing. But since the factors which control the forma-
tion of an eye are present not only in the cells of the eye but also in
the cells of the wing and everywhere else in the body, the question
immediately arises as to why the genetic effects are localised in
particular regions. It is useless to appeal to other hereditary factors
in order to account for this phenomenon, for such factors, on the
same evidence, must be present in all cells, and therefore will be .
unable by themselves to establish a differential anywhere.
The answer to this question has already been provided. It is that
primary differentiation is not an effect of the hereditary factors, but
of external factors. Their first effect is to establish a system of
gradients, as a result of which the various regions of the developing
egg come to exhibit differences of a quantitative nature, both in
respect of the activity of their processes, and of the proportion of
materials such as yolk which they contain. There are several gradi-
ent-systems in the pre-mosaic stage of development of a newt's egg
— the primary apico-basal (animal-vegetative, or future antero-
posterior) with high point at the animal pole; the dorso-ventral
gradient with high point at the grey crescent ; the exterior-interior
gradient, presumably with low point at the centre of the egg ; and,
398 THE HEREDITARY FACTORS AND DIFFERENTIATION
apparently, the asymmetry gradient with high point to the left.
These interact to form a complex compound system, no two points
in which will be in entirely identical conditions.
It is these quantitative differences between regions of the embryo
which are responsible for initiating the processes of differentiation.
Of themselves, the hereditary factors are insufficient to account for
differentiation, and their action must be considered in relation to
the external factors and to the new internal factors which are con-
stantly arising as a result of antecedent processes of development :
internal factors which as such were not present in the undiffer-
entiated oocyte.
A clear-cut example of the direct influence of the cytoplasmic
environment upon the chromosomes is furnished by the develop-
ment of Ascaris. Here, a process takes place known as the diminu-
tion of the chromatin, which occurs in all the blastomeres except
that one which will give rise to the reproductive organs. The fer-
tilised egg has normal chromosomes which divide at the first
cleavage, but in one of the resulting two blastomeres the ends of
the chromosomes are thrown off into the cytoplasm and their
middle portion breaks up into fragments. In the other blastomere
the chromosomes remain entire. In the subsequent divisions of the
blastomere with diminished chromosomes, all the chromosomes
appear in the diminished form. On the other hand, in the division
of the blastomere with entire chromosomes, one blastomere retains
the entire chromosomes, while those in the other blastomere under-
go diminution. A similar process occurs in the subsequent divi-
sions of the blastomere (always a single one) in which the chromo-
somes are entire, until it gives rise to the gonads (fig. 192).
It has been shown by experiment that the presence in any blasto-
mere of the cytoplasm of the vegetative pole of the egg (containing
the so-called "brown granules") prevents the diminution of the
chromosomes. Normally, since the first cleavage division in Ascaris
is in the equatorial plane of the tgg, the division spindle being ver-
tical in the plane of the egg-axis, only one blastomere of the 2-cell
stage contains the vegetative-pole cytoplasm, and therefore only
one blastomere preserves the entire chromosomes. If a ripe tgg is
placed in a centrifuge apparatus and rotated at 3800 revolutions per
minute for several hours, the egg, being free to revolve, orientates
THE HEREDITARY FACTORS AND DIFFERENTIATION 399
itself with its axis along a radius of the centrifuge, and the stratified
distribution of its contents is accentuated. Further, the egg be-
comes flattened, and the cleavage-spindle, adapting itself to the
longest axis of available cytoplasm, lies horizontally instead of in
the vertical position. The result is a cleavage division in the vertical
2-cell
taking
Fig. 192
Cleavage and chromatin-diminution in the normal egg of Ascaris. i,
stage. The first cleavage is latitudinal; chromatin-diminution is
place in the animal cell. Si(AB), first somatoblast rudiment of the primary
ectoderm. la, Enlarged view of the diminution process. 2, 3, 4-cell stage;
2, T-shaped phase. 3, Lozenge-shaped phase. Note extra-nuclear chromatin
resulting from diminution in A and B. S.^ {EM Si), second somatoblast (endo-
meso-stomodaeal rudiment). 4, At the next cleavage, chromatin-diminution
occurs in the second somatoblast. 5, 6, Later stages. iSj , secondary and tertiary
ectoderm rudiments. P4, germ-cell with undiminished chromatin. (After
Boveri, from Jenkinson, Experimental Embryology , Oxford, 1909.)
400 THE HEREDITARY FACTORS AND DIFFERENTIATION
plane, or the plane of the egg-axis, and both the resulting two
blastomeres contain a portion of the original vegetative-pole cyto-
plasm ; it is further found that both retain the entire chromosomes. ^
Each of these two blastomeres then behaves like the single blasto-
mere of the 2-cell stage which contains the vegetative-pole cyto-
Fig. 193
Results of centrifuging the egg of Ascaris. Above: left, an uncleaved egg after
centrifuging ; centre and right, resultant division into two similar cells (plus a
small centripetal mass containing yolk). Below: the behaviour of the chromo-
somes in centrifuged eggs ; left, no diminution of chromosomes in the 2-cell
stage ; right, diminution of the chromosomes in both of the two upper cells. (After
Hogue, from Morgan, Experimental Embryology, Columbia University Press,
1927; modified.)
plasm in normal development, and the embryos resulting from
such treatment are double monsters (fig. 193; and see p. loi).
The conditions controlling the retention of entire chromosomes
in the blastomeres of Ascaris, therefore, reside not in the nuclei but
in the cytoplasm. The cytoplasm produces a situation to which the
^ Boveri and Hogue, 1909.
THE HEREDITARY FACTORS AND DIFFERENTIATION 40I
chromosomes of the nuclei react, by undergoing or not undergoing
diminution.
A different but equally interesting method of chromosome
elimination is found in the fungus-fly Sciara coprophila. In this
species the cells of the male germ-line possess five pairs of chromo-
somes. In the somatic tissues of the male, only seven of these ten
chromosomes are found, one pair of large chromosomes and one
single member of another pair being eliminated. In the female the
somatic and probably the germ-cells contain eight chromosomes.
It is probable that here too elimination occurs, but extends to germ-
cells as well as to soma, and is confined to the pair of large chromo-
somes which is also eliminated in the male. There would then
exist not one but two types of chromosome elimination. It appears
that the decision as to which shall occur is predetermined in the
zygote by the genes in one particular chromosome of the mother.
Undoubtedly the reduced chromosome-complexes must differ from
each other and from the unreduced complex in their morphogenetic
and physiological effects, and the elimination process is thus here a
true link in the chain of differentiation. However, it seems certain
that this constitutes a highly exceptional method, but it is of interest
as showing that qualitative changes in the total gene-complex may
arise during early development in different parts of the embryo.^
§2
The effect of external environmental factors, in co-operating with
the hereditary factors (and other internal factors) in producing de-
velopment, is shown by experiments in which embryos are made
to undergo development in abnormal environments. A simple and
striking case is that of sea-urchin eggs made to develop in sea-
water which is deficient in calcium. The blastomeres resulting from
the cleavage of such eggs do not remain in contact with one another,
but become separated as isolated and independent cells, so that
normal development of the original embryo is of course out of the
question (although each of the blastomeres of the 4-cell stage if re-
placed in normal sea-water can produce a normally proportioned
but diminutive larva)'^ (^gs. 44, 194).
^ Metz, 1931. ^ Herbst, 1897, 1900.
HEE 26
402 THE HEREDITARY FACTORS AND DIFFERENTIATION
Another example is provided by the exposure of the eggs of the
frog or of certain fish to the action of weak toxic substances. In
such cases (already noted in Chap, ix, p. 348) the animals develop
with one median cyclopic eye instead of the normal pair.^ Since it
Fig. 194
Absence of cohesion in the blastomeres of sea-urchin eggs in calcium-free sea-
water, a-c, Successive stages in one egg. «, 2-cell stage. &, 4-cell stage, c, i6-cell
stage. The cell-membrane has become radially striated, and the cells fail to
remain united, d. Disintegration of a blastula into its component cells when placed
in the same medium. (After Herbst, from Jenkinson, Experimental Embryology ,
Oxford, 1909.)
it known from palaeontological evidence that fish have possessed
paired eyes since the Silurian dpoch, these experimental results are
an illuminating example of the fact that hereditary factors, however
long the time during which they have been transmitted to successive
generations, can only produce their normal effects by interacting
^ Stockard, 1910; Bellamy, 1919.
THE HEREDITARY FACTORS AND DIFFERENTIATION 403
with a specific normal environment.^ An equally good case is that
of the adult characters of the axolotl. As is well known, the adult
characters (the genes controlling which have been inherited for
countless generations) normally fail to appear, as the animal is almost
invariably neotenous and does not undergo metamorphosis. But
spontaneous metamorphosis does occasionally occur under certain
conditions of the external and internal environment ; in particular,
the administration of thyroid hormone. In the absence of these
environmental conditions, the genes are powerless to produce the
adult characters.
§3
While the genes are by themselves incapable of initiating the pro-
cesses of development and differentiation, it is obvious that they
play an active part in the control of these processes, once develop-
ment has been started, and their presence is essential. A good
illustration of this is provided by sea-urchin eggs when fertilised by
two sperms. Each sperm brings with it an aster which divides, with
the result that there are four, and a quadripolar spindle may be
formed in the egg. Such an egg contains three nuclei, and since
each is haploid, there will be three 7i chromosomes spread at
random over the four spindles. Each chromosome divides, thus
producing six ?i chromosomes in all, to be distributed between the
four blastomeres into which the egg divides at once. On the
average, therefore, there will be 6/Z/4, or 1*5?/, chromosomes to
each blastomere.
It is known from experiments on parthenogenesis that the hap-
loid number of chromosomes, or n, is sufficient to enable develop-
ment to occur, and therefore, if all the chromosomes were equiva-
lent, any blastomere which received at least n chromosomes might
be expected to develop. But such is not the case. If, on the other
hand, it is assumed that each chromosome of each genome is
functionally different, so that when a particular chromosome is
absent its place cannot be taken by any other chromosome of the
same genome, but it can be supplanted by the corresponding chromo-
some of one of the other genomes, then it is possible to calculate the
chances in favour of any one blastomere receiving at least one
^ Goodrich, 1924, p. 56.
26-2
404 THE HEREDITARY FACTORS AND DIFFERENTIATION
complete set of all the chromosomes. As has already been seen, the
blastomeres of the sea-urchin can be separated, and the hypothesis
can be tested by seeing how many of such blastomeres of dispermic
eggs are capable of development. As a matter of fact, the observa-
tions are in accordance with the calculated probabilities. Further,
in some dispermic eggs, there is formed not a quadripolar but a
tripolar spindle, and the egg cleaves into three. Here, the pro-
babilities of any blastomere receiving a complete set of chromo-
somes are different, but again, observation accords with calculation.
Thus the chromosomes of any haploid set (genome) are functionally
different, and the presence of all of them is essential.^
The problem has also been attacked from another angle by means
of experiments on frogs' eggs which have been subjected to X-rays
or mechanical injuries to the nucleus, and which are fertilised by
sperms subjected likewise to X-rays, ultra-violet rays, or trypa-
flavine. The effect of such treatment on the sperm is to incapacitate
the nucleus from playing any further part in development, without
destroying the activating power of the sperm. In no case can
normal development ensue if both the egg and the sperm nuclei
have been affected, but it has been possible to determine the stages
at which the normality of the developmental processes breaks down.
In the first place, it has been found that the presence of a certain
amount of chromatic material on the spindle is necessary if cleavage
is to take place at all.^ Next, it appears that as a result of slight
irradiation of the egg (the sperm having been treated with trypa-
flavine), a normal though retarded cleavage may take place, but
gastrulation is seriously affected. Either the blastopore closes very
slow^ly and nothing more happens, or the blastopore lip is merely
ephemeral, or it does not even appear at all. In all these cases it is
clear that the damaged nuclear apparatus is responsible for the
failure to develop.^
Further evidence is supplied by experiments with larval hybrids,
i.e. larvae resulting from the. fertilisation of eggs of one species
by sperm of another. This is well shown in some sea-urchins, where
the larval skeleton may show considerable specific differences.
The pluteus of Echinus microtuherculatus is of an elongated
^ Boveri, 1904, 1907. - Dalcq and Simon, 1932.
^ Dalcq and Simon, 1931.
THE HEREDITARY FACTORS AND DIFFERENTIATION 405
pyramidal form, the arms being supported by simple rods. The
pluteus of Sphaerechinus granulans is of a more rounded form, with
two of its four arms longer than the others : the skeleton is in the
form of a rough framework made up of several rods interconnected.
The hybrid obtained by fertilising eggs of Sphaerechinus with
sperm of Echinus is intermediate in shape between the parental
types, and its structures show some of the characteristics of both
parents.^ Analogous results have been obtained from a study of
hybrids between fish species.^ It is clear that those characters in
which a hybrid resembles its father are due to paternally inherited
genes.
In heteroplastic experiments in which a piece of tissue from an
embryo of one species is grafted into an embryo of another species,
artificial embryonic or larval chimaeras are produced. When the
two species are closely related, as are for instance Triton cristatus
and Triton taeniatus, the result is the production of fairly normal
embryos.^ Chimaeras may also be formed by mixing regeneration-
buds of the black and the white varieties of the axolotl* (fig. 195).
In all such cases, when the operation is performed before irre-
versible determination of the tissues has taken place, the general
pattern of diiferentiation is imposed by the field-system of the
organism or region, acting as a unit. But the detailed peculiarities
of the differentiated tissues are determined by the hereditary con-
stitution of the species to which the tissue originally belonged. This
^ A related yet separate problem is the question as to the relative importance
of the parts played by nucleus and cytoplasm in controlling the development of
the larval hybrid. The method used to investigate this matter has been to fertilise
enucleated eggs w^ith foreign sperm. Experiments of this kind have been per-
formed on Amphibia (Baltzer, 1920), where, however, the embryos do not live
long enough to enable definite conclusions to be drawn, and on Echinoderms,
where until recently the technical difficulties involved have introduced un-
certainties, particularly as to whether the nucleus really is eliminated from the
egg. These difficulties have now been overcome, and it appears that the cytoplasm
of an enucleated egg can exert some effect on the characters of the larva, although
the nucleus seems to be more powerful (Horstadius, 1932). The presence of heredi-
tary factors in the cytoplasm of the oocyte has been revealed in experiments on
sex-determination in moths and on the inheritance of dextrality in snails, and in
each case there is reason to believe that these factors are the persistent results of
genes situated in the chromosomes at a previous stage. The same may be true in
the case of the Echinoderm hybrids just mentioned. See also Boveri, 1903.
^ Newman, 1914. ^ Spemann, 1921.
* Schaxel, 1922 a.
406 THE HEREDITARY FACTORS AND DIFFERENTIATION
may concern not only such characteristics as pigmentation, but also
cell-size, specific growth-intensity, specific structures (see Chap, vi,
p. 142 and Chap, vii, p. 236), or the
time-relations of development.
An example of this last type is
provided by experiments in which a
portion of presumptive neural tube
material of Triton taefiiatus is grafted
into the side of an embryo of Triton
cristatus. It may there undergo diflPeren-
tiation into gills, but such gills pre-
serve a feature of their specific origin,
although the tissues from which they
have arisen would normally never have
given rise to gills. In Triton taeniatus
the gills develop relatively earlier than
in cristatus^ and in the experiment just
described the gills which are formed
from the graft oUaeniatus tissue show Sectorial limb-chrmaera
a greater precocity of diflFerentiation axolotl, produced by combining
than the host cristatus gills of the other ^^^ ^«^^^! ^^^^ 5^/ hind-limb
. ^ . . . regeneration- bud irom a black
Slde.l The taeniatus tissue, m its specimen with the ventral half of
diflFerentiation into a structure which a hind-limb regeneration-bud
1 , ,1 1 c 1 left in situ on a white speci-
It would normally never have formed, ^^^. one year after operation.
is still controlled by certain of its (Redrawn after Schaxel, Arch.
hereditary factors. Still more demon- ^«^^^^^^^- l, 1922.)
strative results have been obtained by xenoplastic grafting between
Anura and Urodela (Chap, vi, p. 142). Here, then, is additional
evidence of the fact that the hereditary equipment of all the cells
of the organism is the same (see Chap, v, p. 85).
It is possible to make up a compound embryo by grafting to-
gether an anterior half- embryo of Rana virescens and a posterior
half-embryo of Rana palustris, or vice versa. The compound
organism behaves as a unit in regard to its general physiology and
can undergo metamorphosis and develop into a full-grown frog.
But the two components retain some of their specific characters,
not only as regards pigmentation, but also as regards structural
^ Spemann, 1921.
Fig. 195
THE HEREDITARY FACTORS AND DIFFERENTIATION 407
features such as details of head-shape^ (fig. 196). In an analogous
experiment in which lateral halves of gastrulae of Triton taeniatiis
and Triton cristatiis are grafted together, it can be shown that al-
though the compound organism is, here again, a functional physio-
logical unit which can develop into a full-grown newt, the tissues
^13-
Fig. 196
Compound organisms produced by grafting together half-embryos of two species
of frogs in the early tail-bud stage. Above, anterior component Rana sylvatica
posterior component Rana palustris. Left, shortly after operation. Right, later
embryonic stage (note the differential growth of the anterior component). Below :
left, a compound frog produced in the same way, but with Rana virescens as
anterior component; 4^^ months after operation. Below: right, a normal Rana
virescens, showing pigmentation and pattern of trunk and hind legs differing
markedly from those of the palustris component of the compound organism
(After Harrison, from Wells, Huxley and Wells, The Science of Life, London
1929.)
not only retain some of their specific histological characteristics, but
certain specific postural characteristics are retained as well, for the
manner in which the limbs are held is typical of the species.-
^ Harrison, 1898.
^ Spemann, 1921.
408 THE HEREDITARY FACTORS AND DIFFERENTIATION
§4
Another line of work concerns the time at which the hereditary
factors in the chromosomes begin to exert their action in differ-
entiation, Cidaris and Lytechinus are two species of sea-urchins
which differ considerably in the times at which corresponding pro-
cesses take place during their developments. The larva of Cidaris
gastrulates about 20 hours after fertilisation, and later, mesen-
chyme is formed from the inner end of the archenteron at about 23
hours. The larva of Lytechiiius gastrulates after about 9 hours, but
mesenchyme has already been formed at 8 hours after fertilisation ;
this mesenchyme therefore cannot be produced from the formed
archenteron but is derived from the outer surface of the larva before
gastrulation has begun, at the place where the archenteron will later
begin to invaginate.
The hybrid obtained by fertilising eggs of Cidaris with sperm
of Lytechinus begins by developing as a larva of typical maternal
(i.e. Cidaris) character, up to the end of the blastula stage. This
indicates that the paternal factors have not yet exerted any effect
up to this stage. But the mesenchyme is produced just as the
archenteron begins to invaginate, not from its inner end (as in
Cidaris) but from the sides of its base, near the outer surface of the
larva, thus resembling the conditions in Lytechinus. In this respect
the hybrid is intermediate between the two parent-species, and it
is clear that the paternal factors begin to make their effects observable
just at the beginning of the gastrula stage. ^
It is probable, therefore, that it is in the immediately preceding
stage, that of the late blastula, that the action of the hereditary
factors in the nuclei commences. In this connexion it is most in-
teresting to note that the late blastula is precisely the stage at which
the ratio of cytoplasm to nucleus in the blastomeres reverts to the
value at which it stood in the oocyte, before maturation of the egg
took place (see Chap, v, p. 132). It may therefore perhaps be
suggested that the time of onset of the action of the hereditary
factors of the nuclei depends upon the reversion of the cytoplasmic-
nuclear ratio to its initial value. ^
^ Tennent, 1914, 1922.
2 Boveri, 1905.
THE HEREDITARY FACTORS AND DIFFERENTIATION 409
As to the intermediate steps in the chain of processes by means
of which the hereditary factors influence differentiation, Httle is
known. It is, however, becoming clear that many genetic differ-
ences, including certain apparently qualitative effects, depend upon
quantitative differences in the rate of action of the factors. The
hereditary control of the rates of certain developmental processes
has been studied in the insect Lymantria, and in the crustacean
Gammarus.
In Lymantria it has been shown that sexual differentiation is
conditioned by a competition between two sets of processes : those
controlling the production of structures that characterise the female,
and those which characterise the male. These in their turn are con-
trolled by hereditary factors, the female-determiners which seem
to be lodged mainly in the F-chromosome, and the male-deter-
miners lodged in the X-chromosome. In normal development, one
or another of these sets of processes wins before the time at which
differentiation takes place, but, by making appropriate crosses
between individuals of different races, and in pure strains under
extreme experimental conditions, it is possible to alter the circum-
stances in such a way that an animal will develop along the female
line up to a certain point, but is then switched over to the male type,
or vice versa. The sooner this switching over takes place, the more
complete is the sex-reversal.^
In Gammarus, it has been demonstrated that the difference be-
tween adult black, chocolate, red-brown, and red eye-colour, is an
effect of quantitative differences in the rate of deposition of me-
lanin pigment in the facets of the eye, and that these differences are
controlled by hereditary factors.^ The interaction between genetic
factors and environment to produce a given character is also here
very well illustrated. At normal temperatures, the "rapid-darken-
ing red" factor or gene produces adult chocolate eyes. But at
temperatures below a certain threshold, no melanin at all is pro-
duced, and the eyes remain pure red. At intermediate temperatures,
intermediate shades of adult eye-colour are produced. One and
the same gene leads to different rates of melanin-formation in
^ Goldschmidt, 1927. ^ Ford and Huxley, 1927.
410 THE HEREDITARY FACTORS AND DIFFERENTIATION
-
-
/^
^^
^
-
/
^
" "
1'
-
\
y
y
.1^
_rA'JJ-
_^'
/
/-n
'SS
-
J
/
•7
/
'/
Sexuil maturify
of d 6 begins
0
'Le,n Days''
\l3
-
/-
^i
allti.
^ad)i
rc
M
-
)/
\
\
23 °L
'
^^^
\M
Jo lU
_
'
/
^
20 °L
'
1 '
-
1
/
^ //
-
/
-
/
15' C.
^
-1
/
^
-^
L
jc/r^
1 1
sion
J
I
/
"■
5
Timt
6 7
? in ti
i
^eeks
-
10 i
1 i2
Fig. 197
The action of rate-genes in determining eye-pigmentation in Garmnarus.
Ordinates, grades of colour between pure red (o) and black (14). Abscissae, time.
Above : smoothed curves for rapid-darkening {rrSS) and slow-darkening (rm)
red-eyed types (average of 1000 specimens for each curve) at standard temperature
(23° C). Below: variation of gene-expression due to temperature. All curves
refer to animals of the same pure stock (inbred rrSS). At 10° C, no melanin is
deposited, and the eye remains scarlet; at 13° C, melanin deposition only begins
at 20-24 weeks. The figure shows the facts over the range from 15° C. to 28° C.
M. sexual maturity. (From Huxley, Problems of Relative Grozoth, London, 1932 ;
after Ford and Huxley.)
THE HEREDITARY FACTORS AND DIFFERENTIATION 411
different conditions; there is a lower threshold of temperature
below which no melanin is produced, and an upper threshold above
which no further increase in the rate can be produced by this gene
(although an allelomorphic gene causes a far more rapid rate, and,
as a matter of fact, has quite a different relation to temperature).^
Further, the precise shade of adult eye-colour produced may also
depend upon a relation between the factors controlling melanin-
deposition and those controlling rate of eye-growth. When me-
lanin-formation is only moderate, the greater the area of the facets,
the more dilute the colour will be (fig. 197).
The way in which genetic factors can exert their characteristic
effect only in a particular cytoplasmic environment is also well
shown in Gammanis. The so-called "albino" and "colourless"
mutants have no melanin in their eyes. This is due to the fact that
this pigment can only be deposited in the retinular portion of the
eye, and in these types this portion of the eye is absent. The mu-
tation has not altered the genes which produce pigment, as in true
albinos, but has prevented the appearance of the only regions in
which pigment-producing genes can exert their effects.'^
§6
Finally, it is important to note that the cytoplasm of the egg may
be modified by specific factors in the maternal hereditary con-
stitution. One of the best examples of this is afforded by the
asymmetry of the Gastropod Liimia^a peregra? As mentioned in
Chap. IV (p. 71) the spiral coil of the body and shell in this species
is normally right-handed (dextral), but a left-handed (sinistral)
type also exists, and it has been shown that the difference between
them is controlled by a pair of allelomorphic genes: a dextral-
determiner and a sinistral-determiner.
A necessary result of the effect being due to genes present in the
mother is the fact that the effects of these genes are delayed by a
generation, so that the familiar 3 : i ratio is obtained, not in F^ by
individuals, but in F3 by families. If a snail has had one dextral
parent, it is found (neglecting certain special complications) that,
after self- fertilisation, all its own offspring are dextral, but of these
^ Ford, 1929. ^ Ford, 1929; Huxley and Wolsky, 1932.
^ Boycott, Diver, Garstang, and Turner, 1930.
412 THE HEREDITARY FACTORS AND DIFFERENTIATION
offspring 75 per cent, will produce dextral and 25 per cent, will
produce sinistral forms. It is clear that segregation has taken place
in the snail in question, but the dextral-determiner has acted upon
the cytoplasm of the oocyte before maturation in such a way that,
regardless of whether the dextral-determiner or its sinistral allelo-
morph has been extruded with the polar body, the embryos into
which those oocytes will develop when matured and fertilised
will be dextral. Owing to segregation, 25 per cent, of these embryos
will possess the sinistral-determiner only; their oocytes will be
subjected to the action of this sinistral-determiner, and all their
offspring will be sinistral.
A similar case is found in silkworms. Here, the pigmentation of
the serosa membrane of the embryo is determined by the mother's
genetic constitution, and not by that of the embryo. MendeHan
segregation for this character occurs, but a generation later than for
ordinary characters.^
§7
In other cases, precursor substances may be formed in the cyto-
plasm of the egg under the influence of the maternal gene-complex.
An example of this is found in Gammariis. A mutant type known
as white body contains no carotinoid pigments, neither red in the
eyes nor green in the body : it is recessive to the pigmented type.
If a male of the white-body type is crossed with a red-eyed green-
bodied female, the offspring are red-eyed and green-bodied from
the start. But if the reciprocal cross is made, the young begin their
career without any carotinoid pigment, and the eyes and body
darken to the normal red and green shades only after some time.
In this case it would appear that a gene controls the production of
substances needed for the making of red and green pigment. When
these substances are absent from the egg, the dominant normal gene
introduced from the father takes time to produce these pigment-
precursors. But if the mutant white-body gene is introduced from the
father and the normal allelomorph from the mother, the precursors
have been already manufactured by the mother and a store of them
is present in the egg-cytoplasm." It is probable that the white-body
mutation renders the animal incapable of utilising carotinoids.
1 Tanaka, 1924. ^ Sexton and Pantin, 1927.
THE HEREDITARY FACTORS AND DIFFERENTIATION 413
§8
The last example shows how a detailed analysis is often required to
discover the mechanism by which genes exert their effect. Indeed,
it is necessary to think in terms of development before it is possible
to discover what is the fundamental process with which a given gene
is concerned. In such an analysis, the old concept of Mendelian
characters will disappear. The visible character is not Mendelian
in any real sense : it is the resultant of the interaction of a particular
gene-complex with a particular set of environmental conditions. In
investigating the effect of a given gene, it is usual to study the
difference in development and end-result obtained by substituting
one allelomorph of the given gene for another in the gene-complex.
By doing so in different environmental conditions, it is possible to
obtain an idea of the fundamental process influenced by the gene
in question. By paying proper attention to the development in this
analysis, this fundamental process is seen to be something very
different from what would have been expected if only the end-
results in the adult had been studied. The resolution of the red-
black series of adult eye-colours in Gammarus into the effects of
genes controlling relative rates of melanin-deposition is a case in
point; and this in all probability has a bearing upon other eye-
colour series, as in Drosophila and in man.
Again, the fundamental process resulting in white ("albino")
eyes in Gafmnarus concerns the failure of the embryonic eye to
differentiate any rudiment of the retinula region : only a close study
of the developmental physiology of the eye-region will be able to
shed further light on the processes involved.
This is, in a certain sense, obvious. What has not been ade-
quately recognised, however, is that the converse holds true, and
that the study of developmental processes will of itself shed Hght
upon genetics. To illustrate this point, an example may be taken
from among growth-processes. The empirical study of relative
growth has shown that a change in relative growth in an organ or
region appears always to be brought about by a change in a growth-
gradient affecting that region. For instance, in Crustacea, the
differences between a small purely female type and a large male
type of chela, and between the small male abdomen and the large
414 THE HEREDITARY FACTORS AND DIFFERENTIATION
female abdomen, are both brought about developmentally by the
substitution of a steep growth-gradient with subterminal high point
for a flat growth-gradient with subcentral high point. Any genes
controlhng chela size and shape will act first by controlling the
general form of the gradient involved, and secondly by influencing
its steepness. In addition, there will doubtless be other genes
Phase 2
(Largest newborn
to medium-sized)
0-4
1 2
distal — >
Fig. 198
Growth-gradients in the limbs of domestic sheep, from birth to half-grown
specimens. Ordinates : growth-coefficients (differential growth-ratios) for weights
of parts of limb relative to weight of vertebral column. The horizontal broken line
represents isogony (growth-coefficient = i -o) ; values above it signify positive
heterogony, values below it negative heterogony. Abscissae: i, limb-girdles;
2, humerus or femur ; 3, radius and ulna or tibia and fibula ; 4, carpals or tarsals ;
5, metacarpals or metatarsals. Solid line, fore-limbs; dotted Hne, hind-limbs.
(From Huxley, Problems of Relative Growth, London, 1932, based on data of
Hammond.)
modifying the growth of local regions of the gradient, and influenc-
ing detailed characters such as bristles, ridges, etc. ; but the main
factors operative will concern the gradients as a whole.
The importance of this way of regarding the facts is well shown
in sheep. ^ Here, in the first place, the limbs during postnatal
development show a marked growth-gradient with terminal or
subterminal low point, and high point in the limb-girdles: the
growth not only of the bones but also of the muscles is aflFected by
1 Hammond, 1929; Huxley, 1931.
THE HEREDITARY FACTORS AND DIFFERENTIATION 415
this gradient. In the second place, one of the main differences
between wild species, unimproved domesticated breeds, and im-
proved domesticated breeds, consists in larger carcass, shoulder,
and thigh size (and therefore greater proportion of meat) in rela-
tion to limb size in the improved breeds, and this on analysis is
found to depend on an accentuation of the slope of the original
gradient. Owing to this, the relative growth-intensity of the
terminal portions of the limb is decreased, that of the central por-
tions in the region of the limb-girdles is increased. In improving
the meat qualities of the sheep, it is necessary to search for genes
affecting the growth-gradients of the limbs.
In a similar way, it will undoubtedly be found that there are
genes which affect the primary gradient-fields of the early embryo,
and therefore the relative sizes of the chemo-differentiated fields
in the next stage, and thus consequently the proportions of the
developed animal.
Thus a knowledge of the nature and effects of gradient-fields will
guide the geneticist in his search for Mendelian gene- differences
and his analysis of the way in which they exert their effects.
§9
In the analysis of the genetics of qualitative characters, a know-
ledge of developmental processes may be of very great importance
to the geneticist. In many cases, for example, the relative size of
a part does not vary in linear relation with the absolute size of the
body, but is proportional to the size of the body raised to a power.
In such a case, to take percentage size of part as a "character" to
be analysed could only lead to erroneous conclusions. To put it
mathematically, if developmental study shows that the growth-
formula of the part (y) relative to the body (x) is of the form
y = ax^, then the geneticist must search for genes modifying not
only the value of the constant a, but also that of b : and if he does
not know the formula, he is not likely to search for the right con-
stants.
Again, linear dimensions would appear to be the simplest "cha-
racters" to deal with in making a genetical analysis of quantitative
differences in the size and proportions of an organ. But develop-
mental analysis appears to show that the two main variables which
4l6 THE HEREDITARY FACTORS AND DIFFERENTIATION
are here concerned are, first, the total amount of material in the
organ (which itself is likely to be related to the total bulk of the
organism by a non-linear formula), and, secondly, the relative in-
tensity of growth in the different planes of space within this mass
of material. The fundamental processes are concerned with the
ratios of the linear dimensions, not with the linear dimensions
« (^
B n f)
w f)
Fig. 199
Shape-genes in gourds {Cucurbitd). Five stages in the development of ovary and
fruit in A, elongate ; B, spherical ; and C, disc types ; showing progressive change
of shape of the fruit-rudiment. (From Sinnott and Durham, Bot. Gaz. Lxxxvii,
1929.)
separately. This concept has been applied to the analysis of the size
and shape of gourd fruits,^ where it is found that a genetic analysis
on the basis of linear dimensions leads to confused results, whereas
an analysis on the basis of ratios between length and breadth
permits of a simple interpretation of the results in terms of a few
clearly defined ** shape-genes" (figs. 199, 200).
^ Sinnott and Hammond, 1930.
THE HEREDITARY FACTORS AND DIFFERENTIATION 417
These examples will serve to show the relations between the
sciences of genetics and of developmental physiology. Hitherto,
neo-Mendelism has been concerned mainly with the manoeuvres
of the hereditary units, and in large part with their manoeuvres
Y///////////M\
y.
50
Y///////////////7\
M-H
n
I I I
un
r
r— . 1
rn
-1
c
-1 ! 1 1
1
aaBB ii
A A (a)
BBU
Fig. 200
Shape-genes in gourds (see also fig. 199)- The abscissae give the form-indices
of the fruits, expressed as breadth/length ratios, running from very elongated
shape (small breadth/length ratio) on the left to very flattened (disc) shape on the
right. The ordinates represent frequencies. Top line, range of form-indices of
parent tvpes : 6, a long type (elongate) ; 50, a rounded type (sphere). The shape-
genes involved are A, B and /. A and B produce flattening, while / inhibits
their action. The constituent of line 6 is aaBBII, of line 50 AAhhii. The F^ is
intermediate and unimodal. The F.. is multi-modal: the extreme right-hand
group represents a new recombination comprising the ABU forms, resulting in
disc fruits. An F3 from one of these (bottom line) shows a sharp 3 : i segre-"
gation. The parent must have been AaBBii and the offspring 3 ABBii : i aaBBii.
(From E. W. Sinnott and D. Hammond, Amer. Nat. LXiv, 1930-)
during the two cell-generations in which the reduction of chromo-
somes is brought about. It is now beginning to concern itself with
the mode of action of the hereditary units during the much larger
number of cell-generations involved in building up the adult
organism from the tgg: and this task it can only accomplish
satisfactorily in close contact with developmental physiology.
27
Chapter XIII
THE PREFUNCTIONAL AS CONTRASTED WITH
THE FUNCTIONAL PERIOD OF DEVELOPMENT
§1
It has aheady been noted that some, at least, of the field-organi-
sation, both total and partial, characterising the early stages of
development, appears to persist throughout life, side by side with
the organisation characteristic of later stages. However, the de-
velopmental consequences of the new processes initiated in the
functional period are very striking and overshadow most of the
effects dependent upon field-organisation.
These new processes fall under several main heads — growth,
true functional modification, the unification of the organism by the
nerves, and endocrine influences. It is impossible within the scope
of this book to give any detailed treatment of development during
this functional period, but a few instances may be presented which
will serve to make its main characteristics clear.
The true growth-period of the embryo or larva does not begin
until the organism can either feed for itself, draw upon a store of
accumulated food material (as in meroblastic eggs), or be nourished
by its parent. Previous growth takes place only by imbibition of
water, or by slow contact absorption of yolk. Without quibbling
over precise definitions of growth, however, it may be pointed out
that the determination of organs may take place without any process
of growth being involved, and that growth may and normally does
continue long after tissue- differentiation has occurred.
It appears, however, at least in some cases, as in that of axolotl
limb-buds, that degree of differentiation is correlated with absolute
size of the rudiment. If the rudiment is experimentally enlarged,
as by grafting one limb-bud on to another, the resuking single
enlarged limb (see p. 223) shows accelerated differentiation as
compared with the normal limb of the unoperated side^ (fig. 201).
^ Filatow, 1932. See also Guyenot and Schott^, 1923.
PREFUNCTIONAL AND FUNCTIONAL PERIODS
419
During the early stages of development, when the whole organ-
ism or its major organ-systems are still in the gradient-field con-
dition, removal of a small portion of tissue will not result in the
absence of any particular structure, for regulation is possible within
the gradient-field. At this stage, no structures have been locally
determined, and loss of tissue does not imply loss of any definite
rudiment. It is only later, during the mosaic stage of development
when the various rudiments are chemo-differentiated, that regu-
lation is impossible. Later on, again, the power of regeneration
Fig. 201
Correlation of size with rate of development in fore-limb rudiments of the
axolotl. The very early limb-bud of one embryo is removed and superposed on
the mesodermal portion of the limb-bud of a host embryo of the same stage. The
two rudiments fuse to produce a single enlarged limb, in which differentiation
is more advanced than in the normal limb of the other side. Top: the host
limbs; left, unoperated normal limb; right, experimentally enlarged limb with
larger digit-rudiments and more advanced skeletal condensation. Below: the
limbs of the donor; left, unoperated normal limb; right, small limb-rudiment
regenerated from the remainder of the limb-field. (From Filatow, Zool. Jahrb.
(Abt. allg. Zool. Physiol.), li, 1932.)
appears. Regeneration, as pointed out by Przibram,^ is intimately
bound up with growth, and the onset of the capacity for regeneration
after the mosaic stage of development is connected with the onset
of the capacity for growth at this stage. Regulation and regenera-
tion must therefore be carefully distinguished, since they involve
developmental processes which are very different, and are operative
at different periods of the life-cycle.-
Regeneration also, in some cases at least, appears to be con-
nected with the development and function of the nervous system.
^ Przibram, 19 19.
27-2
420
THE PREFUNCTIONAL AS CONTRASTED WITH
If the limb of a post-larval or adult newt is amputated it will re-
generate, provided that the fibres of the autonomic (sympathetic)
nervous system are intact.^ The dorsal
nerve-roots can be severed and the
dorsal ganglia destroyed, or, the
ventral nerve-roots can be severed
close to their exit from the spinal
cord, without destroying the power of
regeneration of a limb. But if the
sympathetic ganglia are destroyed,
the power of regeneration is lost also.
If the nerves of the brachial or sciatic
plexus are simply severed, the post-
ganglionic sympathetic fibres are
thereby cut, and no regeneration
takes place until such time as these
fibres have themselves regenerated.^
The nervous system has been
found to play a similar part in the re-
generation of the earthworm, for the
nerve-cord must be present at the The morphogenetic influence of
r T ^- • ^ ^ 1 the nervous system. The anterior
cut surface if regeneration is to take ^^^ of an earthworm is ampu-
place from that surface. If the tated and then an incision made
anterior end of a worm is cut oflF, and, °^ "^^ ^'f^^^^^ ^""'^f^ ^° "^ '°
, remove the ventral nerve-cord
in addition, the nerve-cord is ex- from several segments. No head
tirpated for a short distance behind is regenerated from the anterior
1 f. . , cut surface of the trunk, but one
the cut surface, an anterior end may ^^^ ^j-m in relation to the an-
be regenerated from the place where terior end of the nerve-cord.
the nerve-cord ends, but never from g,tS. ^^-..^f ^"' ""''■
the original cut surface^ (fig. 202).
The precise role of the nervous system in many such cases of
regeneration is unknown, but the example of the newt's Hmb is a
warning that the relation may be difficult of analysis, and that only
fibres of a particular component of the nervous system may be
involved in these morphogenetic processes.
^ Schotte, 1926 B.
^ Morgan, 1902.
Fig. 202
THE FUNCTIONAL PERIOD OF DEVELOPMENT 421
§2
Growth is also directly responsible for a certain type of further
differentiation, namely, change of proportions. There are at least
five factors involved here. One concerns the specific growth in-
Fig. 203
Inherent growth-rates in Hmb-rudiments. Left, a larva of Amblystoma punctatum
and right, one of Amblystoma tigrinum between which the left fore-limb rudi-
ments were exchanged at the tail-bud stage: 50 days after operation, after
maximal feeding of the larvae. The grafted limbs {gr.) are approximately of the
same size as the corresponding unoperated limbs of the donors. (Redrawn from
photograph in Twitty and Schwind, ^owrn. Exp. Zool. lix, 193 i.)
tensities of the organs, which will determine the main features of
the growth-equilibrium between them and the body : this has been
dealt with in Chap, x (fig. 203). The second concerns growth-
gradients, which will influence the growth of parts within single
422
THE PREFUNCTIONAL AS CONTRASTED WITH
organs (such as crustacean chelae), within single regions of the
body (such as in the crustacean abdomen), or within the body as
a whole (as in stag-beetles or Planarians).^ The third is concerned
16
15
/4
13
12
II
JO
0-9
0 8
07
0 6
r-Q) T wh<jk eye on P
-<2) T optic vesicle onP
-(D T lens ectoderm on P
g)P lens ectocLerm on T
[5) P optic vesicle on T
-@P whole eye on T
T whole eye on P 0-
Tlens ectoderm onP@^
T optic vesicle on P (9)-
P optic vesicle on T(id)^
P lens ectoderm on T(Th
P whole eye onT {0
(a)
Fig. 204
The mutual influence of regions with diflferent specific growth-intensities.
Whole eyes, or their parts (optic vesicle and lens ectoderm), were grafted reci-
procally between embryos of Amblystotna tigrinum (T) with high growth-
intensities, and Ajnbly stoma pwictatum (P) with low growth-intensities. The
ordinates represent the ratios of the diameters of the parts of the eye on the side
receiving the graft to the diameters of the corresponding parts of the intact eye
of the other side, (a) for optic vesicle, (b) for lens. Fast-growing whole eye on
slow-growing host (i and 7) gives high ratios. The association of a slow-growing
host-lens with fast-growing grafted optic vesicle (2 and 9), or a slow-growing
host optic vesicle with grafted fast-growing lens (3 and 8) reduces the ratios.
Similarly, low ratios are found for slow-growing whole eyes on fast-growing
hosts (6 and 12); 5 and 10, 4 and 11 show the increase of ratio when a slow-
growing grafted component is associated with a fast-growing host component.
(From Huxley, Problems of Relative Growth, London, 1932; based on data of
Harrison.)
with the time-relations of development. In general, development
occurs in an antero-posterior direction, so that at a given time an-
terior organs are further differentiated than those at a more posterior
^ See Huxley, 1932, Chap. in.
THE FUNCTIONAL PERIOD OF DEVELOPMENT
5*i*-*^ •' »^' ♦^'-* 4 3«'-*» -V^l w^-^-w
423
m.g.
'. ^c-y^^
S..V*
f-
B
Fig. 205
The effect of mechanical conditions on morphogenesis. In larval axolotls kept
out of water, the dorsal fin disappears. This is due to its falling over and becoming
fused with the skin of the back. A, Section of early stage of fusion. The meso-
dermal fin-axis {f.a.) is bent at the tip. B, Section of a stage showing complete
fusion. The fin-axis still shows a curved tip. The limit of the fused fin is
marked by a sudden thinning of the epidermis {ep.); bl.v. blood-vessel; vi.g.
mucous glands ; muse, muscles ; c.t. connective tissue. (From Huxley, Proc. Roy.
Soc. B, xcviii 1925.)
424 THE PREFUNCTIONAL AS CONTRASTED WITH
level. Within vertebrate limbs, development takes place centri-
fugally, and, as a result of this, the later- differentiating parts will
increase in proportionate size during development. ^ Fourthly,
there are growth-processes directly concerned with the functional
demands made upon an organ : these also involve change of pro-
portion and will be dealt with later. Fifthly, the growth of one
structure may be modified by the specific growth-rate of neigh-
bouring structures.
In illustration of this last point, it is found that the structures
composing the eyes in Amhlystoma punctatmn and tigrinutn have
diflFerent specific growth-intensities (Chap, x, p. 366). By making
grafts of eye-cups and of lens-forming epidermis between these two
species, it is found that the presence of a fast-growing eye-cup is
correlated with an increase in the growth-rate of a slow-growing
lens associated with it, and vice verso? (fig. 204).
Mechanical modification of growth-processes is readily brought
about. It is only necessary to recall the artificial deformations of
skull, lips, waist, feet, etc., practised by various human societies.
In this connexion may be mentioned the fact that when axolotl
larvae are reared in dishes with only a little water so that their
backs protrude above the level of the water, the dorsal fin falls over
owing to its weight, and becomes completely united to the skin of
the back. But, internally, the structural and histological features
of this finless condition are quite distinct from those produced as
a result of normal metamorphosis,^ although externally they are
more or less similar (fig. 205).
§3
The unification of the organism by means of the nervous system
brings the various parts into more intimate relations with each
other as regards their functional activities, and brings the organism
as a whole into a more intimate and more delicately adjusted re-
lation with the environment. This is responsible for a greater
dehcacy of functional adjustment on the part of the various organs.
^ See Huxley, 1932, Chap, iv.
2 Harrison, 1929; Twitty 1930; Twitty and Schwind, 1931. This is also true
of limbs (Rotmann, 193 1, 1933), but does not happen with parts of the shoulder-
girdle. ^ Huxley, 1925.
THE FUNCTIONAL PERIOD OF DEVELOPMENT 425
The unification of the organism by means of the circulatory
system has in some ways a similar effect. It also makes possible a
competition between organs and regions for available nutriment,
and this may have marked effects upon development.^ The pro-
portions of parts of growing mammals, {a) fed maximally, {b) fed
so as to permit of only slight growth, and {c) fed so as to permit only
of maintenance of weight, are quite different.^ In extreme cases,
whole regions may disappear as a result of being drawn upon by
the rest. For instance, if a zooid together with an attached piece of
stolon of the Ascidian Perophora are isolated and starved in normal
conditions, the stolon will be completely resorbed by the zooid ; but
when placed in dilute toxic solutions the zooid is more affected, and
is then resorbed by the stolon (p. 294).^ In organisms without a
skeleton, starvation may produce reduction in total size, and then
different parts will be reduced at different rates, as for instance in
Planarians^ and in jelly-fish^ and hydroids.*^
The establishment of the circulation has a further consequence
which in vertebrates at least has far-reaching effects upon develop-
ment. It permits of the transport of hormones, some of which have
striking morphogenetic functions. Some hormones may be liber-
ated more or less continuously into the blood. This is apparently
the case with that amount of thyroid hormone needed to produce
normal development in man : when this threshold is not available,
the child is a cretin, stunted in growth and subnormal in intelligence.
In other cases, the hormones may be produced cyclically, and
this appears to apply to the hormone of the anterior pituitary con-
cerned with stimulating the cyclical growth of the ovarian follicles.
Or the hormones may be produced in markedly different amounts
as a result of nervous impulses to the gland, which in their turn are
controlled by external stimuli. In Amphibia, for instance, darkness
stimulates the post-pituitary to liberate the hormone which causes
expansion of melanophores : and while growth is taking place, this
1 Roux, 1 88 1.
- Jackson, 1925; Hammond, 1928; Huxley, 1932.
3 Huxley, 1921 b. * Abeloos, 1928.
^ de Beer and Huxley, 1924. ® Huxley and de Beer, 1923.
426 THE PREFUNCTIONAL AS CONTRASTED WITH
also causes extra multiplication of melanophores.^ A similar
result, doubtless brought about in the same way, is seen in fish
(Lebistes). Specimens reared on white backgrounds have con-
tracted melanophores, few in number; specimens reared on
dark background have expanded melanophores in large numbers.
Functional activity increases the rate of multiplication^ (fig. 206).
Similarly in salamander larvae (S. maculosa), yellow backgrounds
$
y. ...... .M
4
Fig. 206
Functional activity and rate of multiplication of pigment-cells. Dorsal view of
the trunk region of two specimens of the teleost fish Lebistes retkulatus, one (i)
reared for 6 months on a white background, the other (2) for the same length
of time on a black background. In both cases the pigment-cells (melanophores)
have been induced to assume the contracted state by adrenalin treatment. Note
the much larger number of melanophores in the black-adapted specimen, in which
during life they were expanded normally, while in the white-adapted specimen
they were contracted. (From Sumner and V^eWs, Journ. Exp. Zool. LXiv, 1933.)
favour the increase of the yellow areas, black backgrounds that of
the black areas. After metamorphosis, however, a gradual regula-
tion towards the control type sets in, indicating that what we may
call "functional multiplication" of pigment-cells is only important
in certain stages.^
A sudden change in the activity of a gland may take place at a
certain stage in development, as occurs with larval Urodela, in
which the sudden onset of metamorphosis is brought about by the
^ Smith, 1920. 2 Sumner and Wells, 1933.
3 Herbst, 1924.
THE FUNCTIONAL PERIOD OF DEVELOPMENT
427
thyroid throwing its stored secretion into the blood. ^ In this re-
spect, the Urodele may be contrasted with the Anuran, where the
thyroid becomes progressively more active during larval life, with-
out any such extreme change in its activity.
The morphogenetic effects of hormones are varied. Some of the
most marked are those concerned with amphibian metamorphosis,
in which the growth or differentiation of some organs and the
Fig. 207
Sharply delimited fields in a thyroid-treated frog tadpole. Section showing on
the left the epidermis of the fore-limb bud, on the right the lining of a branchial
cleft. The former has reacted to the thyroid hormone by growth (mitoses,
crowded nuclei); the latter by degenerative changes (vacuolation, shrunken
nuclei). The limit (/.) between the two zones is clear-cut, without transition.
(From Champy, Arch. Morph. Gen. Exp. iv, 1922.)
atrophy of others will only take place under the influence of the
thyroid hormone. All gradations are to be found, however, be-
tween such marked morphogenetic effects and effects of a transitory
physiological nature. The morphogenetic effect of hormones
may be linked with the pre-existence of qualitatively different
fields. E.g. in the frog, one region of epidermis will proliferate,
and another degenerate, under the influence of thyroid- (fig. 207).
1 See Huxley, 1923; Uhlenhuth, 1922. ^ Champy, 1922.
428
THE PREFUNCTIONAL AS CONTRASTED WITH
These regional differences in reactivity are established very early
(Schwind, J. Exp. Zool. Lxvi. 1933). The relation between hor-
t.m
'♦ 1- t
1 BV
< :
L
;-— E
— sc
SBS
EL
P
2 BV
■;ih~E
"i.:- ss
- ~~sc
- --T_w E
'•_/^-^-SS
K"^--sc
^-'^^.^.
3^
BV
SBS
EL
Fig. 208
The perforation of the operculum in the frog {Rana claniitans). Sections
showing the histolysis leading to normal perforation, i, First sign of histolysis
(at X). 2, Histolysis well under way; the stratum compactum and stratum
spongiosum have become interrupted; there has been a marked invasion of
the area by lymphocytes, and the skin in this region is decreasing in thickness.
3, The skin is reduced to the epidermis, which the fore-limb then ruptures.
BV, blood-vessel; E, external epidermis; EL, epidermal lining of branchial
chamber; P, pigment; L, lymphocytes; SC, stratum spongiosum; SBS, con-
nective tissue; SS, stratum spongiosum; X, site of histolysis. (From Helff,
jfourn. Exp. Zool. xlv, 1926.)
mones and growth-gradients is shown by studies on regeneration-
rate and hormone-susceptibility in birds' feathers. ^
Lillie and Juhn, 1932.
THE FUNCTIONAL PERIOD OF DEVELOPMENT
429
An interesting half-way stage between chemical eflFects due to
contact, as in the determination of a lens by the eye-cup, and those
due to circulatory hormones, is seen in the perforation of the right-
hand side of the operculum in Anuran tadpoles during meta-
morphosis. As is well known, the rudiments of the fore-limbs
develop beneath the operculum, and while the left fore-limb makes
its way out through the open spiracle, the right protrudes through
a special perforation. After extirpation of the right fore-limb rudi-
m
V' — 1^'-^-
'^ '':i.
W^
OG
— p
J
Fig. 209
Perforation of opercular skin of Rafia paiustris, grafted on to the back, over pieces
of atrophying tail-muscle. The histolysis of the opercular skin leading to per-
foration is the same as that normally due to the atrophying gills, though slower.
a. Larva, showing graft of opercular skin {OG), perforated (P). h, Enlarged
view of graft showing atrophying tail-muscles seen through the perforation.
(From Helff, JoMr«. Exp. Zool. xlv, 1926.)
ment, perforation of the operculum still occurs,^ thus demon-
strating that it is not due to mechanical pressure. Actually, it is a
substance liberated by the gills during their atrophy that is re-
sponsible for the perforation, as is shown by experiments in which
metamorphosing gills are grafted beneath the skin of the back and
cause perforation here too.^ Other atrophying organs, such as the
1 Braus, 1906. ^ Helff, 1926.
430 THE PREFUNCTIONAL AS CONTRASTED WITH
muscles of the tail during its resorption, will produce the same
effect, but more slowly. Thus presumably some substance pro-
duced during autolysis is the agent responsible (figs. 208, 209).
§5
The next subject to consider is the trophic effects of the nervous
system. In view of the fact that innervation (by fibres of the auto-
nomic nervous system) is a prerequisite condition for regeneration
of limbs to take place in adult newts, it is most interesting and
curious to find that the nervous system is not essential for the
embryonic development of the amphibian limb. It is difficult to
obtain embryos in which the limbs are not supplied by some, even
abnormal, nerves, for, as already explained (Chap, xi, p. 389), the
limb exerts an attraction on the
growing axon. But limb-rudiments
have been seen to develop when
free of any nerve-fibres. This con-
dition can be realised by grafting
the limb-rudiment of a frog into
a lymph-space of another larva,
or by extirpating the neural tube
opposite the limb region on one
or both sides in the neurula stage.
The limbs are normally differen-
tiated as regards all their con-
stituent tissues and parts : cartilage,
muscles, skin, blood-vessels, and
the joints between the skeletal The trophic effect of the nervous
11 , „ system on the development of the
segments, all these are normally limb. Ventral view of a larva (shortly
differentiated in the absence of before metamorphosis) of i^awa/w^c^
,• _^ ^^.: 1 „4. .1 _ i:»^u „^ ^ from which at the neurula stage the
mnervation, but the limb as a .^diment of the lumbo-sacral region
whole is too small. ^ In other words, of the spinal cord was extirpated on
the nerves have a trophic but not a ^^^, "S^t side. Note normal form of
. . rr 11 right leg but subnormal size and
morphogenetlC ettect on the de- development. (From Hamburger,
velopment of the limb (fig. 210). In ^rch. Enuumech. cxiv, 1928.)
this respect the effect of the nerve is similar to that of thyroid hor-
mone on limb-growth in larval Anura^ (see also Chap, x, p. 363).
* Lebedinsky, 1924; Hamburger, 1929. ^ Champy, 1922.
THE FUNCTIONAL PERIOD OF DEVELOPMENT 431
The Stimulation of the muhiphcation of the nerve-cells in the
spinal cord (Chap, xi, p. 383) in Amhlystoma is another example of
the effects of nerve-endings. There are also the cases in which the
presence of a nervous connexion is necessary for the maintenance
of structure in an organ. As is well known, muscles atrophy when
the motor nerves to them are cut. But the best-analysed examples
concern the lateral-line organs, and the taste-buds on the barbels of
the catfish Amiuriis. When the nerves to these organs are cut, the
organs themselves undergo marked dedifferentiation, and rediffer-
entiation when the regenerating nerve restores their nerve-supply.^
The trophic stimulus has been found to pass down the nerve from
the cell-body at a rate of 2 cm. per day, and the indications are that
it is due to percolation of a hormone-hke substance.^ It clearly
cannot be due to normal impulse-conduction (see p. 387).
Though the precise mechanism of their action is still obscure,
the interest of these examples for the present purpose is clear. They
demonstrate that once the nervous system becomes functional, new
methods of influencing development are available in the organism.
These methods concern such diverse processes as local cell-
multiplication, large-scale regeneration, and the maintenance of
differentiation in organs.
§6
Finally, there are the effects of function per se. This is perhaps the
most pervading of all the new effects which take their origin at the
onset of the functional period.
Function can influence the multiplication of cells and the size
of organs, the histological appearance of cells, and the arrangement
of cells and tissues within an organ. Often more than one of these
processes is involved at one time. The most obvious example of
purely quantitative change concerns compensatory hypertrophy.
When a portion of a functioning organ complex is removed, the
remainder increases in bulk in response to the increased demands
made upon it. The simplest instance concerns the kidneys. When
one kidney is removed, the other^ enlarges ; the enlargement is
considerable, though not to double its original bulk.^
1 Olmsted, 1920. ^ g. H. Parker, 1932 a, b.
3 Ribbert, 1894.
432 THE PREFUNCTIONAL AS CONTRASTED WITH
Conversely, when extra demands are made upon an intact organ,
it also may respond by increased growth. The excess growth of
striated muscle under the influence of heavy work is the most
familiar case. The heart, too, is an excellent example. In small
birds, the relative heart-size is greater in specimens from high lati-
tudes than in those of the same species from milder climates, owing
to the greater demands made upon the circulation in cold condi-
tions.^
In voluntary muscle, it is probable that the direction of the fibres
is also influenced by function, in the first instance by the tension
to which the muscle is exposed by the growth of the skeletal parts
to which it is attached.'^ The directive eflFect of stress has been ex-
perimentally demonstrated in connective tissues. By subjecting
thin tissue-cultures of fibroblasts to variations in surface tension
it has been possible to show that whereas in regions free from
directional stress, fibres are formed at random in all directions, in
regions subjected to directional tension the medium is condensed
along the lines of stress. The fibres orient themselves along these
condensations, and the cells multiply more rapidly in these regions^
(fig. 21 1 ). This case falls perfectly into fine with the experiments on
regenerating tendons. If the achilles tendon of an animal is cut,
the space between the cut ends is filled with debris, blood, and
phagocytes, and resembles a tissue-culture. Fibroblasts soon grow
into it, and the fibres which they produce are at first chaotic ; next
they form a meshwork with diagonal interlacings ; and finally form
parallel bundles. The muscle, exerting a pull on one of the cut
ends of the tendon, sets up lines of stress in the ground-substance,
and this orientates the growth of the fibres.
But if the muscle also is cut, so as to abolish the pulling eflFect,
no tendon is formed. If now a silk thread is drawn through
the regenerating tissue, in a direction at right angles to that of the
original tendon, constant gentle pulling on the silk thread will pro-
duce a bundle of fibres orientated according to the artificially-
produced lines of stress. A tendon has here been formed, but at
right angles to its normal direction.^
In respect of the orientation of the cells to the lines of stress, and
^ Hesse, 1921. ^ Carey, 1921 a.
^ Weiss, 1929, 1933. ^ Lewy, 1904; see also Nageotte, 1922.
THEFUNCTIONAL PERIOD OF DEVELOPMENT
433
of the more rapid multiplication of the cells subjected to the stress,
these experiments have completely confirmed the epoch-making
essay of W. Roux (1881), by whom the principles of functional
differentiation were first clearly stated. From these and other lines
of evidence, it appears highly probable that the size and fibre-
direction of all the tendons of the body have no direct hereditary
■^'/T^ .- y
i4\ ^--v.
■■■*
Fig. 211
Portion of a tissue-culture of chick fibroblasts exposed to regional tension (by
cultivation as a film in a quadrangular frame). In the region under tension (left)
the cells are arranged in fibres parallel to the directions of the tensile force, and
are more numerous than in the remainder, where they are scattered and of
irregular form. (From Weiss, Arch. Entwmech. cxvi, 1929.)
basis, but are determined epigenetically de novo in each individual
by the stresses and strains to which they are exposed during develop-
ment. The fact that fibroblasts arrange themselves along lines of
mechanical stress, and multiply faster when exposed to tension,
automatically accounts for the production of a mechanically
adaptive structure.
HEE 28
434 THE PREFUNCTIONAL AS CONTRASTED WITH
The fine architecture of bones appears also to be determined
in the same way. Here, too, structures which are mechanically
adapted in great detail to their functions are not determined heredit-
arily. On the other hand, the general form of bones is predeter-
mined in great detail by chemo-differentiation. Certain depressions
in the surface of avian bones appear to result from mechanical in-
teraction with neighbouring bones, but all the projections from the
surface, including the joint-structures, will arise in isolated bones
grown in culture media (see p. 225). It may prove that the
cartilaginous rudiment is rigidly predetermined, whereas the bony
structure, being secondary from the start, is always dependent
in its differentiation.
The coarse structure of a bone is, then, a result of chemo-differ-
entiation during the prefunctional period, but function is necessary
for the perfection of its finer structure, viz. the orientation of its
spicules. Function is also necessary for the normal growth of
bones. If one leg of a new-born animal is kept immobile and non-
functional, the long bones remain much slenderer than in the used
limb of the other side. On the other hand, if a leg is subjected to
changed function, as in the case of the hind legs of puppies born
without front legs, the hind legs, from the practice of hopping,
assume the proportions characteristic of hopping animals such as
the kangaroo."
With regard to the blood-system, little is known as to how much
of the broad lines of its architecture may be determined by chemo-
differentiation. What is certain, however, is that a very great deal
of its detailed architecture, as regards the size of vessels, the angles of
their branchings, and the courses which they follow, are determined
hydrodynamically. The pressure of the blood moulds the vessels in
such a way as to offer the least resistance to its flow.^
Lastly, instances may be given of functional changes involving
cell-form as well as the total size of an organ and the development
of its parts. The first case, like so much of the functional differ-
entiation of the blood-vessels, shows the effect of pressure of a
contained fluid on the walls of its container. The urinary bladder
of a dog of medium size normally evacuates a quarter of a litre of
fluid per day. The wall of the bladder is composed of smooth
muscle cells and is about half a millimetre thick. By means of a
'*■ Fuld, igoi. - Oppel and Roux, 1910.
THE FUNCTIONAL PERIOD OF DEVELOPMENT
435
tube connected with the bladder, large quantities of a neutral fluid
can be introduced into it, with the result that its internal pressure
^'n
':i
/
U:l
V.
»
0^
'i^!to^.
^
Fig. 212
Functional activity and morpho-
genesis in amphibian gills. Below:
two salamander larvae; left, reared
in conditions of oxygen-deficiency
(in water under an atmosphere with
II per cent. Oo) ; the gills are long
and feathery; right, reared in con-
ditions of oxygen-excess (in water
under an atmosphere of pure O2) ;
the gills are short and stumpy.
Above : sections of gill-filaments
from two similar larvae; right,
oxygen-deficiency : epithelium one
layer thick, of flattened cells ; left,
oxygen-excess : epithelium often two
layers thick, of rounded cells. (From
L. Drastich,Zeif5c/;r./. vergl. Physiol.
II, 1925.)
is raised. The quantity of fluid evacuated per day may reach
50 litres under these experimental conditions. As a result of this
increased work to which the wall of the bladder has been put, it
28-2
436 THE PREFUNCTIONAL AS CONTRASTED WITH
was found that it had become ten times as thick, that its cells
had developed striations very similar to those which characterise
heart-muscle, and that the whole bladder pulsated rhythmically.^
The other case is that of salamander larvae, brought up in water
which is deficient in oxygen. Such larvae show much enlarged
external gills, while the gills of specimens reared in water with
excess of oxygen are extremely small. In the enlarged gills, upon
which extra respiratory demands are being made, the capillaries
are larger, nearer to the surface, and the epithelium of the surface
and the endothelium of the capillaries are thinner, thus permitting
of a more rapid diffusion of gases. The converse changes are seen
in the reduced gills^ (fig. 212).
§7
It is important to note that no sharp line can be drawn between
functional responses of considerable morphogenetic extent, as in
the cases just cited, and transitory adjustments of a physiological
nature which leave no structural traces, such as a temporary local
vaso-dilation. The connexion between the degree of expansion of
melanophores and their rate of multiplication has been noted
above (p. 426). Further, it should be remembered that one and the
same kind of organ can respond by a morphogenetic change to one
degree of functional stimulus and not to another. For instance, it
appears that only severe demands on muscles will cause them to
hypertrophy; movements involving little mechanical strain, even
when rapid and prolonged, have no effect — e.g. those of knitting
or piano-playing.
It must also be remembered that functional adaptation can only
take place within certain limits prescribed by heredity. The
thyroid responds very readily to the demands made upon it by in-
creasing or decreasing its supply of hormone and its size. Yet by
selection, it has been possible to establish separate genetic strains
in pigeons, a high-thyroid strain and a low-thyroid strain, which
differ from each other in the size and activity of their thyroids even
under identical external conditions^ (see also p. 409).
Most important of all, it must be borne in mind that functional
modification may be very active in one group of animals, and
1 Carey, 1921 b, 1924. ^ Drastich, 1925. ^ Riddle, 1929.
THE FUNCTIONAL PERIOD OF DEVELOPMENT 437
negligible or absent in another. For instance, it is impossible for
holometabolous insects to produce functional modifications
during individual ontogeny in their skeletons. The hard parts of
these animals are definitively formed, with all their adaptive de-
tails, on emergence from the pupa, and no further growth is
possible. The same is true for the development of their muscles
and tendons : these must be preformed during the pupa stage so as
to permit of perfect function and locomotion of the animal as soon
as they are called upon.
There is thus a remarkable contrast between the development of
vertebrates and that of higher insects. In the former, prefunctional
differentiation lays down a rough sketch of the organism, upon
which most of the finer adaptive details are later inserted by means
of functional response to the demands made upon the parts. In the
latter group, on the other hand, although doubtless some details,
such as those of the blood-vessels, may be determined through
functional response, the greater part of the structure, including
even the finer adaptive details, must be laid down by elaborate
chemo-diflFerentiation, unaided by functional response.
There are, of course, other equally fundamental diflFerences in
developmental methods between groups. Hormones play a very
large part in the later stages of vertebrate morphogenesis ; but in
insects their role appears to be altogether subsidiary. Similarly,
the adult form of a vertebrate is determined by changes in pro-
portion of parts which are brought about by diflFerential growth in
already functioning organs, and which continue through a large
fraction of the Hfe-span; in holometabolous insects, no growth
occurs in differentiated parts, and proportions must be definitively
fixed during the short pupal period.
The subordination in Ascidians of the period in which the total
gradient-field system is the sole form of organisation, as contrasted
with its long persistence in Amphibia, is another example, in this
case concerning early stages of development, of the diflFerences
which may exist between groups as regards their developmental
mechanisms.
Chapter XIV
SUMMARY
§1
It is now possible to give a brief summary of the chief points which
have emerged from our study of development, during which
attention was focussed on differentiation and its origin as the
central problem.
In the first place, animal development is truly epigenetic, in that
it involves a real creation of complex organisation. It is also pre-
determined, but only in the sense that an egg cannot give rise to an
organism of a species different from its parent. The development
of each individual is unique. It is the result of the interaction of a
specific hereditary constitution with its environment. Alterations
in either of these will produce alterations in the end result.
Determination is progressive. In the earliest stages, the egg ac-
quires a unitary organisation of the gradient-field type in which
quantitative differentials of one or more kinds extend across the
substance of the egg in one or more directions. The constitution
of the egg predetermines it to be able to produce a gradient-field
of a particular type ; however, the localisation of the gradients is
not predetermined, but is brought about by agencies external to
the egg. The respective roles of internal predetermination and
external epigenetic determination are clearly seen in regard to the
bilateral symmetry of the egg. The amphibian egg is predeter-
mined to be able to give rise to a gradient-field system of bilateral
type through the establishment of the grey crescent at a particular
latitude of one meridian. The particular meridian is not predeter-
mined, but is normally decided by the point of sperm-entry ; the
precise latitude is determined as a result of the primary axial
gradient of the egg, impressed upon it by factors in the ovary. On
the other hand, the egg of a radially symmetrical animal like a
Hydroid is incapable of developing bilateral symmetry; the pre-
determined capacity to react to stimuli localised in one meridian is
not given in its constitution.
SUMMARY 439
The agencies which determine the position of the various axes
involved in the gradient-field system may be of very various nature ;
they may be factors in the maternal environment (ovarian con-
ditions), biological factors (point of sperm-entry), or external
physical factors (as in the determination of the polarity of the egg
of Fiicus). In any case, they are external to the ^gg. They may also
operate at Ytry different times relatively to fertilisation.
A number of chemical processes are set going by fertilisation.
These will proceed differently in the quantitatively different en-
vironments provided in different parts of the gradient-field
system, until qualitative differences are set up. In most cases,
these differences are at first not visible, and are presumably of
chemical nature ; this step in differentiation is therefore spoken of
as chemo-differentiation. These chemical differences appear at
first to be reversible (e.g. labile determination of the presumptive
neural tube region in the Urodele before gastrulation) but after a
certain point to become irreversible. From this moment onwards,
the organism consists of a mosaic of chemo-differentiated regions,
each determined to give rise only to one or a limited number of
kinds of structure. These are what we have called partial fields.
The attainment of the mosaic stage often takes place under the in-
fluence of a dominant region or organiser. This may determine the
extent and form of the whole gradient-field within which chemo-.
differentiation occurs, as in Planarian regeneration, or may interact
with a previously established gradient-field orientated in another
direction, as in amphibian organiser grafts.
The organiser may exert its effects at a distance, as does the re-
generated head on a cut piece of a Planarian, or may supplement
such distance effects by more powerful contact effects, as happens
when the amphibian organiser comes to underlie a certain portion
of the animal hemisphere, and at once determines it irrevocably as
a nervous system.
Modifications of the gradients by external agencies will entail
alterations in the structures produced. These alterations may con-
sist in changed proportions, or in the total absence of certain
regions (temperature-gradient experiments with frogs' eggs,
440 SUMMARY
cyclopia in fish, modification of regeneration in Planarians). Here
again, there is a predetermined capacity to produce a certain type
of structure in certain conditions ; but the precise locahsation of the
structures produced depends upon the form of the gradients in the
field-system.
Once the mosaic stage has set in, further diff"erentiation may be
brought about by the influence of one point on its neighbours. The
classical example of this is the induction of a lens from epidermis
by the optic cup.
During the period when the organisation of the developing
animal consists of a single field-system, far-reaching regulation is
possible; after irreversible chemo-differentiation has occurred, it
is not. The precise time at which irreversible chemo-differentiation
sets in varies markedly in different groups. In Amphibia it occurs
during gastrulation ; in Ascidians at fertilisation.
After the establishment of a mosaic of partial fields, it does not
follow that all the cells of any given partial field necessarily give rise
to the organ characteristic of the field. Thus, more cells are capable
of giving rise to the amphibian fore-limb than do in fact give rise
to it in normal development. Further, the boundaries of the partial
fields overlap : a given group of cells in the limb-rudiment of the
chick may contribute to the formation of either a thigh or a shank,
according as to whether it is allowed to remain attached to or is
isolated from one partial field or the other. Gradients may exist
in such fields : the capacity of cells within the fore-limb field to give
rise to a limb decreases with their distance from a subcentral
portion of the field: the same is true for many other organ-fields.
§3
Up to a certain time, regulation is still possible within each of the
partial fields ; but as development proceeds, each of these becomes
split up into progressively smaller fields, each with its own deter-
mined fate : for instance, the fields for leg, shank, and foot, within
the originally single hind-limb field.
Each area in the mosaic passes from the state of invisible chemo-
differentiation by the process of histo- differentiation to full visible
differentiation, and so reaches the functional stage. After the
organism as a whole has reached the functional stage, many new
SUMMARY 441
morphogenetic agencies come into play. The organism also, through
acquiring the power of regeneration, reacquires much of the regu-
lative capacity which it lost in its passage through the mosaic stage.
The type of organisation characteristic of one stage appears to
persist, in whole or in part, throughout subsequent stages. Thus,
the main gradient-system of the embryo permeates the partial fields
of the limb, neural folds, ear, gills, and heart, and determines their
axis ; and the growth of the lateral line along a particular level of the
flank can best be interpreted in terms of a persistent total gradient-
field.
Again, a total field-system certainly exists in adult Planarians and
appears to reveal its presence in late stages of other groups through
the presence of growth-gradients permeating the whole organism.
The persistence into adult life of the partial field- systems of the
mosaic stage is shown by the phenomena of regeneration, by the
existence of localised growth-gradients within single areas, and
notably by phenomena such as those found in newts, where, for
instance, indiflferent regeneration-buds produced by an amputated
limb will produce legs when grafted into a certain area round the
leg, while if grafted near the base of the tail they will produce tails.
§4
With this, of course, only a start has been made with the scientific
analysis of development. It remains for the future to discover such
fundamentals as the physiological basis of the field-systems, and
the elaborate physico-chemical processes which must be operative
at the time when the quantitative diflFerences of the early gradient-
field system are being converted into the qualitative differences of
the chemo-diiferentiated mosaic stage.
It is, however, already a good deal to have arrived at this first
outline of development on the biological level. To have established
the fact that organisations of quite different type succeed one
another during development is important. The recognition of the
gradient-field system, with its purely quantitative differentials, as
the basis of early organisation, is a great step forward, since it pro-
vides an adequate formal explanation of many phenomena of regu-
lation which have been considered by various authors, notably by
Driesch, as affording proof of vitalistic theories of development.
442 SUMMARY
Further, the epigenetic analysis of development is pointing the way
to a large extension of the field of heredity, in the shape of physio-
logical genetics. It is only through a study of development that it
will be possible to understand what the term ''genetic characters"
really stands for — in other words, what are the basic processes in-
volved in the action of a particular Mendelian gene.
Experimental embryology as a separate branch of science was
initiated by Roux; in its next phase, in which Driesch, Boveri,
Wilson, Herbst, Morgan, Brachet and Jenkinson are outstanding
names, a large body of facts was amassed, and the experimental
proof of epigenesis provided; in the third phase, Spemann and
Harrison are the outstanding figures within the sub-science, while the
theories of Child have not only linked the facts of regeneration with
those of embryonic differentiation, but have provided a scientific
basis for a field hypothesis for early development, thus filling a large
gap in the theoretical aspect of the subject. Meanwhile, experi-
mental embryology has been making fruitful contacts with physio-
logy, notably in the field of hormone action, with genetics, and with
growth studies.
The fourth stage is now beginning, in which this framework of
general principle will be filled in through intensive research, and
the whole science deepened by a search for the physico-chemical
bases of the empirical biological principles which have been dis-
covered in its earlier stages.
BIBLIOGRAPHY AND INDEX
OF AUTHORS
Note. This bibliography and index of authors has been specially designed
to facilitate reference both to the text of this book and to the original
works in a library. For this purpose, where two or more works by an
author are concerned with the same subject, they are referred to together;
and where such works are in the same periodical, their references are
placed together. This system involves a certain trifling disturbance of the
chronological order in some cases, which should, however, present no
inconvenience owing to the facility with which the date-figures in bold
type can be picked out.
444
4->
o
cr
bo
?
OO O
bJD.2
^ 03 (U
c
C
-73
c 2
.2P
o >.
C
Oh'
3
0} ■»->
1^ J,
a; -^
s |.
> >
o <u
o 3
o
c
c« --2
u •:p
tJO
(n
>
u
X
Co -
M O
00 ^^
CO <M
.2 J
^05 ^
t^
^ d d
CO ^ ^. HH- S
^ ^ ^ ^ ^
^2^ ■ "^
CO
■H Co
o
u
Is
00
r^ M
'^ d .
• " Oh
C<I ^ 05
M >{'^ d~S
O Tl
Oa CO
05 05
<ij
O- g
^t5^.C§
I
S? ^ 05 "Ki ^1?*
N CO
^
a ^ a
u
O
m pq
^
w
KK
•— '
l-H
13
<
^
s
<
a a
^
H-^
q^ m
m
fe
fo
^
C/3
o
c/T
H-i
m w
w
-^
u
u
o
o
a
Vh
C C
c
^ c
^
N
(D
<D
C3
CU OJ
1 1
<u a>
<u
-^^
Xi
rO
-73
xi xi
TS
<—t 1—1
C3
a
C3
<
<
<
<<
<
<<
<
m m
m
m
m
445
00
to
ON Th
00 1-H
?
^ 00
00 m m
1-1 N c^
00
oo"
\0
o
oo"
00
00
N
00 CO o
ro N 1-1
.>--
hfi
c
fcJO
(i)
(7)
^
C)
M-H
X,
c
u
Oh
c«
-n
u
Ui
CC
OJ
Ul
d
<
I I
> 3
c« O
o:z;
o
C
_o
u
<u
' C
^ c
J '^
C Oh
u
i^ g
C3 TJ
O
•S g
IB
o o
O
biD
ri "O O
< few
1-1 Sis ^^
aj nj cc
a3 Q ^
bcCo C
a
o
tS Cu
00
W
^ a .
X >< "
o u r;
fe fc
00 -^
i-i O
00 S5
05 S
CM
si
^ ^ ^ o^
rH _ S
CM CO
05 05
^1
d
[Coo'
05
H-1 C3
S~^ ^^
§ >< > -
^ ^-^ a
a :? '^ S
^ o
a ;i
(M Co
s &
X ^
^1
52 aco
S -05
^ o tH
^ CO .
,-ro5 «
i-i th r:3
J > S
-^ . ^
a^Ci^ I
? 00
•^
H-1
t^2
JS
^.£
t-S
N
K -
-«*
• 2
t5 *
G C
a a
N N
■M +->
■i;^^
^^'O
N i^ =^ ^ W
n •""» I
X-pS":^
<U 0) <u
n PQ n
lain"
0) o <u <u
PQ m m m
C3
c
c
o .
(U "^ <u
PQ m
446
o
<u
u
be
00 00
vO
CO
O
t^ TJ-OO vO
O O
00
o
CO
o
h
o
X
w
p
Q
<:
>^
<
o
o
O
00
^ o
H- 1
a
>
>
X
<u
C
<u
00
o
?«
a ^ :^
Oh i^
00 ts
00 "^
^ . E 5 ^ ^
2S X
is
>2
o ^
■s-S^'^
1 ^
.o
-^' -^ -S
J< 5^ o
-^
^
II
^ T-(" "^ ^ ^
^05
,^
r
^'^stS
SI
=o
Q
t1
ton
^
i^-j
^
-^
.^
t^
«
^
^
i^)
ti
^
^
^
^
i
^^
c^
<ij
H^
"^
O
^
•?«
=0
s
bJD
►ii
:3 o
o o
C/2 Q
:c3 ..^ 5;
N
CO E?
'-' S
d|
ill
::^^ A> :>
^ «i f^; a
XX g > .s
u ^
m m
c
ID
o
PQ PQ
go
S ^ c
o o o
m m m
u
C
C
o
h Hh
r!
<u
<U QJ
u
>
> >
O
o
o o
m PQ PQ PQ
PQ
g
^
hh
<u
:3
u
^
>
>
^1
o
o
MPQ
447
CO M
CO
ON
rooo
w CO
■^ lo C^ O
lO N 01 00 lO
01 CJ CO '-I O
CO
lO in
O o
0) (U a>
4->
o
a;
CO
1
C3
XI
o
o
<u
C
>
0
CJ
;-l
o
^
tS)
dJU
bf)
o
QJ
CO c
II
•43 ^
OJ
fa
C
<u .. _
fl biD fa
-S .CO CO 2
o
C3
<u
CO CO
o "
a;
3 2
a; :-
O O (J
C.2-S
2 ^ ^
'a g-c
O D o
C/^ CL, (1,
c3 C
•- bJD O
:2 ^ -^
C 03
^ - >. O
_^ C u c,
O ^ C
. C3 C CO
Hi I
Off; whS
OJ -S
C
_o
oj be
CO
C3 O
c£ JM CO
w "a 5^
— < g u
03 i5 ^
CO 3 ^
1^
o
I"
r
CO 10
Tt- OJ
CO CO
o <^
05 Oh
n-- =^ ^
> a.
X -^ -JS
^ ^ ^
J" 05
fc 03
■« o 2
^ co' 1^ -^ §
S^ C3
'^J^
•^05
X
On i-J
O
a. 2^
0;§
d . ^
Oi CCl
^ 03 ^
-is;
M to
SI
- X ^-^
OJ .
. a
CD W
X
>
X
X
U
o
,0
V)
N
1
1^ E^
o
CO
2 ^ ^ ^ •-
'I^rS =
^ ^
tt^ ^
-^^
b- o
J CO C
<■<
03 03
P5
,^ffi
o; CO
3 3
Cd 03
. O
" " " ■ " ^ ^ ^
mm m m m mmm mmmm
C
O 3
I
'^ CO
l-l 4^
3 >>
m m
448
o
u
N \0 CO
CO CO 00
'^ rt- CO
CO
O
O CO
CO Th
CO 1> N \0
Ti- M Tj- CO
00
ON
r^
o\
C4
N
0^
N
00
^
\0
Tj-
10
ON
co
vO
N
of
0
d
ON
CO
•K
on
0
l^
00
0
c^
CO
r-
C/2
O
h
P
C
X
w
Q
Q
<
Oh
<
o
o
m
(U
w ^^ .— I
a
s a
SI
.2 B ^
.2 2 2
•^ •- <i^
cd S 03 3 "^
'Hi) S 'bi) oj 3
C 2 d Id o
CC ^ C3 c/3 Jh
s
n
3
■t->
•13 <u
0 s
G
0
B
lyT 'O
u
0
C3
_o
73
Vh
<l>
3 OJ
■^
tC
+-< >
p
a;
1
C
0
^3
be
ho
•^^ tH
tH
^ X
ON *: g
N ?; ^
• '^ .
> ^ "4i
*-' S S
05 1-1 00 Lj
^ ^ ^ •
^- S >^ ^
Dh S h4 S
>? >^
^
vO
OS
X .
X i\
M O
N
05 ^
Oi
^ O
^ ^ .^ O"
I §2 1
•s I
bc^
.g 05
5" s?
^ o
C3 03
pd
(U
C
(U
Oh
03
o
w
p^
bc
'a;
C
o
O ^
Si };i <L> oj 13 P
O^r^ J-i J-^ ^" +^ S
C3
UO
UU
bJD
C
u
u
449
00
^
ON
O
00
CO
o"
00
CO
00
(^
00 lO
00 I>
N CO
1> O Th rf
00 r^ o r^
CO
00 r^
r^oo
o •■r
o
W
ii
-1-1 I r,
0) &i
> O
>>.5 t;
^1
o <u
a. P 5
^Zt^s
c c
2 G3 OJ (U
j3 j3 ^ be
m q5
<" c ^;
J:^ S oj
O
c
o
(3 UA
"I. o i
o -^^ s-i
i 2;^
o ^ ^
O O O E^OcoQ PC
C3 C
w V C
o S g
i ^£
S C 2
-3 .2^ ^
-Op
c3 a; J^
4-' ^ Dh
> c
•^ o
S 2
o <u
13 tH
0. S
TO JH
C/2
<M a
^2
CD
^ >H tH
i
^z
5 -2 ^ »
§5^
CO a
-05
5 ^ ^ o o
^^
«^
!>!>
c^
CO
> CO S
^ 05 J
•-T-( X
. " "o
« o .
^ . -^
N "^ ^^
Oi .g fcl
"^ oq
^ 05 .
O tH Oh
ON
M
d
m
00
<M
at?
oa .g
05 to
^1
cq
o\
N 00
o^ dS
CO (M M^
05 Oi
05 §
CO
d§
05 cq
>
CO '-'
ON,^
in o
2^
H <^ CO
M •- ^
^^
^ O
ft^
1.^
05 ^
00 i:
^
^^'
^
^'
^*
^
^
, CM.
en, H. J.
e, C D. vai
11, G. E.
*—>
^>
d
u
OO
U
o
o
u
U
fe'
^•G
,d^
T3
'O T5
Ti
73
Ti
'U
-a
-O ^ > rl
n
^ u
1 ^
•>H
••1
ja
-C j:3
^
-C
J3
j::
Si
o
o o
o
(J
UU
u
O
O
U
U
uuuu
u
UU
U
29
450
4-»
o
00
ON
1>
O 00
so ^
CO CO
CO -^ M
N CO t^
o
01
^
O
m
ON
-o
M
N
00 tJ- vO CO
O O 00 NO
pi;
o
h
o
X
w
Q
Q
X
Oh
<
o
o
o fa
o
.4_)
a
Si
:3
H
C/2
o
CO
'u O
bjo C
SI i2
2i.S2
(U c«
0 QJ
1 o
0) l-H
"o ^
"^ O
<u
>
<\i
,-M
03
;h
G
bJO
Tt
o
CO
o
aj
a
03
03
O
'u
O
U
^ a
- be
U <4H
3 O
o
00
o 2 -
• -^ c
\0 coC<l
. Oi
. a -
Q, CO
-co „-
CO S T^ J
f^2 - >^
►^^ T-( ^ C
S ^ OJ
o ^ r -^
s.^ ^ i3 ^
1^"
fc] . 03
ill
O
u 05
T3 tH
X
X > >-! -^
K K S K
05 >
5^^
d^
CO ^
>2 >^ >2
M 5
^ .
d-^
^J ^ CO
X ^^ c^^
s: ^ S^ fa
di
^■5
•^Kj
CO Cij ^ . •
dit.s
i-T 5*.'
X §
X '==;
o .
2. ^S ^
■* fa '^ CO*"
<M . .^ CO
05 -^ ^^ 05
1
6
d
-i
.
<
W
w
c
G
^
^
c
c
o
o
U
u
w w
00
.S .S
c c
o o
o .
G
a>
a
o
U
O u
(n
o cfa o
-ijj O Ja5
«^ ,^ 5^
c3 CO IJ3
Q Q
451
Ov
d
(^
o
M
CO
^
t^
N
CO
00
^
N
CO
CO
CO
UO
_
vO"
o
N
CO
o
r^
N
n
CO
CO
CO
CO
^
^
»^
•s
o
■^ in
lO
Cv
CO
o
liO
•^
CO
U-)
.^
r^ o^
CJ
O
00
o
r^
On
Tt-
CJ
><!
CJ CO
^
vO «
CO
CO
CO
CO
T3
1
^
-o
'^
C3
.a
1
CJ
03
bJO
CO
tUD
o
.«
-d
3
13
CO
u
t3
'to
C
-3
CO
£^
CO
1
C3
'o
a
u
O
a
CO*
!5
.a
o ^
o
-d
c
.o
'd
CO
O
+-<
>
'd
CO
C
o
O
1
c
13
11
c
1
o
•6
u
+->
S.
S
'o
a
(U
tJD
C
o
a
13
C
o
03
O
<U CO
0>
t3D
.S -
OQ
3
a c
o
.2
o
TO
u
■M
<
3
a
i
C
fS
N
1
o^
**
2 ??3
<
J>
^
<
CO
a
n
£
hT
s
CO
05
1-1
CO
C3
05
CO
X
Cn
T»-
><
5
>
tH
_co OhtH
cot
05
d'
^
is
d
05
T-»
05
d
<
d
Oi
>
1
•i '
CO hJ
>
►J
« 00 >
i-1
o
05
T-t
^
s
^
CO
05
CO
c3
?5
05
X
f5
-w
4
Co
52 > uo
'i
CO
X
X
X
CM X
05 .
T-l
CO
CO
d
^^
>
X
!
1
o
O
• 2
dl
O
1
2§
d 3
6
11°:
Oh ^ a J
1-
d
o ^ o
. 0 05
d
<
CD
(N
05
S
X
i
di
d
05
T-t
X
X C3
.^05
C>tH
CO
CO ►_"
CO
d
^d
dS
o §
2^
w
-
^
<
^"
'i
c
C3
n
C3
,
C/3
^
P^ ffi
pi
CD (ii
pi
p^
pi
P^
H-J
|o
6
do
d
c/5
(A
CO
HH
H uT
0)
^al
^
3.
3
^
■§1
pq CQ
PQ
1
'1
'1
"5
1
■S
QQ
Q
Q
Q
Q
29-2
452
W)
M OJ W
O"^ O^ 0\
CO CO ro
\0
CO
ON
O
t^ w Tt- CO
M N 00 M
CO
1> ON
00 ^
M 1>
C/2
o
X
h
<
o
X
w
Q
Q
>^
X
<
o
o
!j
m
PQ
> c
o o
111
s a
biD
c
be
c
"a
C/2
in ON CO X
5?l
o^ M t^ a
00 M frj w -^
CO d d d"^
^ a £ c<r^
^'~
dooo Sb
s
&g§ n.^
X
N hJ .J ^ ^- ^
^
2^-^ di S
■ei
•gk^Iiq^ o .
K . r vj
CO
2tur
urn.
urn.
LVIl
urn
Pro
52
^
s
^ ^? ^ >^
*^
d'o
. eg
^ o
S -2
a d
o o
vO • -
O "^
<vj CO
s
u
c
::: .> -j^ ^ be. 2
a t^ ^ >• ^ -^
C3 Qj o iri Irt -^
:^ o ^ D^ ^-C
OU^ W O
be
be
O
a^?^
gw
t5 S
o o
^; c«
CS
w
.^ ,o
h '^
^. ^
03
S c^
.s o
^'^
M \0
CO -^
• Oh On
-eo CO
lO O] .00
05 ^ ^ CO
Oi ^ . . .
NO
CO
a
00
^ M
O vo
'"' CO
d a
52 05
-fi -5i
^1
(U <U (U <u
QQQ Q
'2 W
Carpenter
etwiler, S.
McKenno
ragomirow
ragomirow
rastich, L.
rew, A. H.
rew, G. H
riesch, H.
o
'u
X
XX
QQ QQQQ Q Q QQ
453
.to
OvOO O
00 1> CO
ro 1-1 N
§ O
(73
C
B
u
<D
Qh
><
C
o
(73
fcX) ^
o c
o
vO
CO
CO
CO
IT)
CO
CO
d
CO
d
GO
^t! o
c 2
o c
6C ^.
O J ^ O
i2 C o
C3 -^ O
2 c ^
o
c^oo
00 lO
M CD
I ^-^2 ^
!=«
•^ o
05^,0 ^ -
■I -s -s "^ ^
2i 00
2t^
§5^ ^ ^
K o r-
I
(U
(/)
o
o .
CO g
q;
05 tH
^ >
3^"
^ ^
^
CO <:>
I §*
u
^ c
X .2 o
fH <U «0
1 i
X .
.
1—^
^^^
:i:
^
m ffl m
cq
CQ
«
^. ^^ c
^
<u
C G C
c
c
c
tCJ=, c
0
JO
0 D OJ
a>
(U
0
c S^ ?3
en
w
^ ^ ^
^
^
-7^
3
V^ Vh IH
:3 :3 :3
a
:3
Q
Q
QQQ
Q
Q
Q
w w
w
O
w
5 CO
454
o
cr
u
00
Tl-.-,
o r^ M
M M 01
\0 "-I OnOO Tf N
a\ i-i o o i> ^o
CO '^ -^ i-l M CO
CD
o
h
O
X
w
Q
Q
<
<
O
O
(U
c §
g ^
jH C3 c3
ui <u n,
,(U bX) ST
he 1^ o
■^ '^ -^
Q (Ih
W
^3
C3
a
o
I
a;
C bc.2
i ^ g
T3 XJ v^
^ i^ T3
"o) 'oi ^
<u a; I
> > X)
I I r*
U U G
W W J
3 ^
c3
^ -43
St
la
o -
O 0)
-"0 3
O ^ 3
a 3 (u
§ V =5
^ /iV <^
(U OJ O)
c a
03 O
c c
s s
c/3 <I> <U ^H
- > > JL
-d T5 O
^ c
O) C3
C 3
+-. .^--i t^ - 1J C
o
fl CO
Cl,.-h
1^ ^
^ c«
CO 'T3
• ^ C
-o o
cd tM
C O
'Sd >.
c3 t<
■!1
^ S
CO >0
CO ^
ON
d
(»
05
00 ^
•^ IT) Oi
M o th >:
>o CO f^ ^:
o- do'
■§ ^ i>
r« S C<l
k!2 05
r X
5:1
a
T-(
05 '^
tH 'IS
,^ a
00 .
•> d
■: lo a
OhvO ^
to . 01
<M ao5
TH CD - 2^
. th ^^ a
o
^ IC «^
(M (M ,
05 03 -^
S S §
o Co
a o
o
in N
05 «o
Caq -^ tii S
• 2 . ^
'^^ "^O Ci^x^cq -^oq cq
W
O
d tf
3
<
a
^ i-
ed
m gQQ
ddd
=5 fc
ffi
^^ ^"^ ^"^
o o o
03 «
ij
^p^^^
S S S
fc fo
Uh
fc fe fc
fc fo liH
<<<
OJ <U (U
^ ^ ^
u o o
(/i (Ti in
m
• DQ
a; ,
X
ed .-r
w -a
-a -a
^-^ ^ i=: 1^
fe [Jh fe fc Uk fe [Jh
• 03
455
CO
m O -"^ tJ- c> CO
O t£
.2 ^
o ^
HI O
o o
o t^
Tj- o" fi
-I N HH
o
o
o o .-
^ ^
o
'o
a
c
a
s= S
a
< o
(x
C3
? o ^
Sea
U, -^ <U
■2 c
^ .2
u to
eg w C
5 ^ ^ t^
-M "^ ^
^ C3
C c3
O <1J
"" o
cc O -^
s
t^.2
>^5
c« C
+2 O
U( ni ?3
a
c
is
-o o
1>05
c
^
«
s
o
g X N
wq < CD ^
a >
< ^
(M ^
Oi g
ti N
CO
^ Co
' — » • K
^ -9 ft;
N
^
ex, O
o X
5 5^-
^ d
05 ^
u X
N
CD fcl ^
c -a
(L» T
Q^ in
ooo
^
00
•^ o ^
o
on ^^
oo b£)
COT3
<* . "C
-O 05 ^
o g-^
to" . ?^
'^ S" ^ s
00
c
p
o
o
c
a
a
o
I
w
^ 00 -^
00
• ^ 2
^^
,N
CO
«:; -^ ^ 03 -^ 05 »^
"^ .CO
S^'^tt^
in
« m
w g
N~
N
^J
'c
c
'3 C3
<u
<U
feO
0
0
o
p^
C/2
o
fc
1^
^
^^
oerttler, K
oetsch, W.
oldschmidt
-o
w
o
o
B
1
'o
oodrich,
raper, L
ray, J.
d
1
:3
0
O
o
00
000
O
OOO
0
0
456
o
Oh
en N N CO
en
N CO N
1> M
02
O
<
O
X
w
g
Q
<
a;
<
O
o
m
PQ
(U -is
a ^
O OT
2 t
<u
o
a
O (U
0)
C
a>
bX)
u
-a
c
C3
C
be
<:«
3 bJD ar;i
■M X ^
W3 c« Vm CD ^
'^ ^ M (U r'
£ J o
a
>
a
> bfi fl
^ c«
bJD C
be O
(U -^
.^ g
bc a
fe Ph
bJO
I
I
O
d <
o
O
3
O
WW
TS TJ TS ^3
C ^ S S .
WK/W^^W.W^H-:,
ifi|iiiii
00 O O O O K
bJ3
s
a; o
s a
s a
457
o
CO
00 vO
CO M
I^ lOOO 00
O lo t^ t^
■^ CO CO CO
00
N
On
lO
CO
N CO
CO
ON W
M 1>
CO U^
CO OJ
\0 CO C\
CO w CJ
N CO rt-
O rj- hT o'
t^ N N 00
00 On
r^ ON
M 0}
o o
O a;
c 2 ^
^ ^ A ^
"2 O ^ O
^ C ^ G
'c B <^ Q
«S -S > QJ >
~^ g <u C3 "JJ
W Q < Q
d o .^
H Ui <i>
'a ^3
a
"o o 5 ii'
q t^ -rt ^ ^ =3 -^
(U _, bjo CO
O O
C X5
S o ^ a CO
^ Q <
o ^
■M -M "H c/;
^ ^ 2 I
QO ;z;
•^ 5
>^ S^
(U o
C O >
o o
■^ ^ -a
a u
ffi O W
3
o
a
o
Oh
a ^
<u o
oT o
s §
a c
>^ ^
a
a
>
C3
(U --^ C
O w ^
a, o
-^ a ^
3 1
C3
C -C
ON
o o
o .
^ X
> ^•
^1
o . a. •
o3 a ^ to
• o ^ g
d
oo"
tH
ON §
d ■^
5 eg C^^^Q:; ^<^^E^
.8^^
^
X >
^ -5J Oi '^
On
co
d ^
'^ kJT CO
x o2
^
Et^
ON^
n. S^
05 E
I"
a
O \0 o
- d-^g
Oi '<-' 03 Q
^ > -^ .
u^ • ^
^^;t§;^
^ ^^N
Q^ .-^C^
"^ '^ ?
00
t
CO
.
0000
0
0
0
00
0
00 .
c/5 ^
c c c c
pi
c
d
pi pi
d d
pi
d
dd"
6
d
s -
o o o o
o
o
o
o o
o
o o >;
'>
M
en
O)
7D O)
ii
"u "u >
fr 2
S a
'u 'u 'u 'u
Ul
u
U(
u u
u.
jd
n
Vh S-h
;h
;h
u u
Ih
C t. c
ca C3
C« C« C3 IS
c^
rt
c«
a a
C3
(S CS C3
!U
<U <U
<u
KK
ffi!::KE
X
K
X
XX
X
Ka:^
X
XX
X
458
O
cr
u
00
IT) (Ti rf-
00
vO
MD CO
lo O
M 01
<^ ui d
^ ^ a
o
M
o o
c -a
o be
C C ^
oj .2 c«
bc o o
O c)^
£ .2
rj Oh C
_ ^ c« <u
o
^ :3 c
(U ^ o
= ^3
o -J ^
a.'s ^
o o ^
^ 53 §
^ o o
n ^t; «
J ^
3 bJo
o ^
a:
.Si
u o
<v _,
C O
a CCS
03
(-1 k>^
5 bX)
II
O
Ov
•^
Tj-
f^
S
•o
.■^
«^
d
E^
o c
i>
^.
2^
(U
05
't^
u
00
bx) ~
Lh
>
.Si
<-M
<4J
<u
a;
1
01
1
^-^
^^^
g
d
<^
S^ ^
c§
o^
.g
5,P
o
c^
1
01 N
OJ ^
fe^
I
01
C3
S 52 «
Oi 05 ^
00 TH -
en o
oq
> x:
t§E§^
"^ ^ ^
^
CO
■^ d ti
rvO .
CO n: :
„- ^ to
2 ^
ttl
g
<4) bo
„3
o
13
<
U
U
o
d
^
C3 ^ ^
6
J_J^
^ bJ3 be
hf)
^
^
M
^
^
J 1 1
B
u
Vh V-( ^
;-!
<u
D
OJ
<u
<L)
<u <u <u
OJ
ff;
Dh K Dh
459
O
^ <^
- o
oc
00 f^
„ n CO lo o
ooo o o o
o o
lO o o
l-.H-.vO
c/3 o S
<^ s
o
C c« C
.2 g|
c c .2
'^ ^ c
^ ^ ^
Q
TS C
<u o
CO CO
^ ^ - -*
CO Oh -M •!->
< c
c«
J .2
ii
3 C
O X2
t; o
■^ -^
^ O
(U CJ
02 W
C
a
-T3
C
^a
>
V §
w J
2 =
0
CO
a ^
CO c:
II
oo
'u O
O u
^ i
«o o
CD
-05
1> ><
CD •~
00 §
o <
Kl
^ C o
X
X
^ l-H C;^
•- ^ "^
< CO aft^
. o
a
CO Si
CO ?:
03 S
a ^
^: § ^.
Bd
f^
CO :z
o
o
CM CN, O Q
05 CO C^ ^
tH (33 CT»
r: 00
> (M
X 03
J > >< ^
X ^ X. .
1-1 e^
„ «t> ^ ?? s3
d 2 S
<
a
Oi ^^
cd
O
^1
C/}
m (n
• . P3
c/2 c/]
.
1 — i
W
CO CO '^
wwS'^
u,
3 3
3 3"
. . . 1 — >
be
2 B
rstadi
rstadi
wlanc
nt, T
nt, T
rd, A
xlcy,
o
o
:0 O
:0 :0 O
3 3 3 3
:i:
X
EE
EEK
seed:
460
O
h
O
X
w
Q
Q
<
<
O
o
PQ
pq
0
CO
-!->
0
t^
13
vO
cr
3:
10
ii
^ vO
0)
r--
vO
^
00
On CO
Tp
^
N
CO
N
^
^ li^
Ti-
<u
10
CO
N
bJD
On
ON
CO N
oT
«^
M
HH
Th
N
P^
lO
t^ ^ rh
^
- 1> ThsO
T}- 10
On
M
00
M 0 On
N
N 0 (M
10
ON M M 0
N
0
HI
0
CO 10 J>
^
^ ^ Tj-
vO
CO CO -^ N
^
M
^
CO
CO CO CO
S
CO
C/D
0
0
1
.0
1
^1
•S
1
.0
0
1
2 ^ S
P S w
_o
0
'u
0
i2
^3
1^ <1
s § s
1
B V ^ &c
2 c ^-^
c -^ ^ 2
C cd 0 C
(U t^ Ui 0)
hOOO
C3
3 > 1)
0 .13
f
0
'u
0
PLh
."2
s
.0
00
'c
0
0
cr
1 s"i g
^ ^ X »H
^
^
^
^
w
>— 1
CD
-: 1
d 10 ON^
- 0 00 ""^
CO
T-C
Oi
>
ON ><;
CM
>
10
00
CO
d
CO
0
a3
>"
2
^
fe
>
1
0 -H rH .g
d
CM
O)
d
of
CO
05
rH
cd"
rH
CD
■rH
x"
CO
CO " a
pi;
1
■O HH VI ^
si
>
X!
0
t5
s s ^
X J X
§ a s e ? '^
. !< .• <ii CO -^^I ^
X
1
t^f^f^
|d1l IIS 1 d
2
w
i^
[0 [0 'o
a
^s^^c;
Q.
^^^l-^Q.
a
'^^
1
1
cqi^cg
Pih'
73
-d .
T3 ,
T3
g^
0
gQ
03
3
(A
(n (n (n
c/5
c/5 :n ai c/5
c/id
C/5 PC (72 <
X
d: K X <
<
t—i
I— ^ H-^ 1— ^
1— %
1— ^ H- ^1— ^ H- ^
^ ^r
H-^ >i
»— > >.
a
hJ J hJ ^
. ^
^-^
^
. . . s
>^
>^ >i >i
>^
>;>;>; >^
>> J2
c
C C C 03
^
(U <u 0
^
^ jj jj ^
0 -3
'x 5
S 'o
«
rt 03 rt :z)
1^
'x 'x !<
'x
'x 'x 'x 'x
'x '^
s&
D
3 3 D
:3
3333
3 '^
D <^
K^
X
ffiffiffi
•r
KP^Kffi
ffi
X
a:
d:
^iCffiffi
46i
OO lO
00
CO
00
to
1> o
ON r^
n CO
o o
\0 CO
m en
CO c>
en i>
I §
ID (-1
-^ bB
t: > =«
w 2 g
.^ ^t:
t^ o o
o X a
(u it: o
W £
o 1^
bpl3
p o
<u
o Sis
C >> c
-n ^
c«
O
^ 2
Oh Q
O
■_> W OO
^ 8
2.2
fa£) ^
I C3
2 w)
•M ^
O
'^ ]i ^
^ s s
bo
a
>
C
o
!0 cz
■*- a
O
^:5
h
~ o
CO
I
Ei^
05 .
. a !3
r^ < § m
Si O S O
eg c3
2 X
o o
00
01
c^
^^1
t I a
^ «? s
in a
a . th
CO '^05
^ 05 S .
' X §
s
^
> 03 li)
^ ^
SJ <^ Si
^ o
05 -^
II
CO ^
OP^
m
U
^ ^ ^ ^^ ^
c
C
c
j:j
c
c
o
o
O
o
o
o
C/)
w
C/3
O)
CO
c/)
C
c
c
C
c
C
^
^
^
,^
^
^
c
c
c
c
c
G
(U
CJ
<u
<u
<u
OJ
Ph !3 ■•-'
O O D K^
I — > I — >i — > hit
u
a
a
462
73
O
j:3
to 00 00
vO 10 O
N M N
CO O
M N
CO M
C3
u
+-»
O c3 (u >> -
(-1 ''•'
O lU
C P Q
•2 ^ -a
o o
r o
'+:< (u a;
O CO
•ills
^ -B <^ ^
s ^ -
S c« ?* S
Oh C3 Ci-
03 W
Z^ ss
o ^
s a
o £
Q
T3 D • r3
bo-O ^
C« 03 !«
«=« c/^-O
-S " ^ ^
vj o <U '^
Q^
CO
d
Oi
00
x^
u
10
.
tH
a
tH
05
r^
th
w
a
l-l
• ^ K
^<^ CO
CO ^ 05 .^
bit • o ^
^^^^
NO
d
H. X ^
00 . '^
go:;
^^
CO
CO
X"^
X ^
"^ft.
s g
Sx
. X
o. X
05 (5
c^
in f^ G "^
X ts<
II
> ,
-§
o ^ 00 N ^
" r~ fa ^
N . s •
05 i2 X .,
^^
^^
I ^ x^
d"^ S
w > .
<u ex
o o
w
<
^
Ul
,
OJ
<
a
f\?
a
^
c
OJ
;-!
N
0
0
^^
C/3
fe s
&
^^
„
tH 03
a;
a; ^
^
bX) 0
0
a '^
c«
^H 3
:3
ui;^
^
Vh -M
03 03
o
>>
^
w
03 03
a a
W
463
■^OO 10 Ov o *-<
•1 I> t^ Tf O «
HH M M M i-H N
^ § "^
o I ,
<u <u ^,
hJ W W
00
ON M
ro <^ N l^ O ON
CO
so
CO
1> w
CO CO
c 2 "
2 oj c ^
a; ^"^ -2
wis
.0 J
•-*
CO
0
■£
1
^
i
Rec. I
Journ
Rec. ^
to
ar
^•
5«I tJ
^
^
2
^ C
S K
^
o r-
. ^
§
^ -^
05 O
05
fO coco
1^ ^N
^ n CO
30 M Oi
o . th
o5 a .
S CO
XX--
« S S
TH kl tt^
CO . .
05 -^ ~S;
iH ^ ^
^ 5- 5<
Si s
N^S'^"^'^'^^ "^"^
. a
CO ^
05
CO „
CO 1:3
. >
05 s ">*
O S :J
00
ON
« 00
■-S O^ •S
eg -^
05 «^
^^
^ I
N
Tj-os
il
s ><^
> On
^O
X a
<< <
§co
Co
- ^
5 2^
fe X re Hi s X
O lyT £«" lyT M (tT
o "? 'S "^ "^ '?
<U 1) <U (U <U QJ
X ^
w
Q^ Ph Ph Ph
s: ?^ a ?^ .s .s .s .^
C3 C O -M
o o
W
464
w
P +s
c .5 .^
CJ be '^
111
> <U 13
Q K
c
C3
O
U
o
u
O '^
On -^ O O
cw O
b£
be 00
^ z ^
5 o
s
O c/2
bX)
CO
r^
r^
Th
■+
CO
^
N
CM
i>
CO
Tt-
10
M
r^
r^ «
N
'^
0
i^
^ 0
M
M
OJ
'-'
M ON
<U
il
c«
0
I^TJ
<u
bC
> c
v;
2 --5
o
Q
75 c
^ .2
W CQ
.2 a
^ 2
a; 4_,
biD >
.2 %
C a;
o C
o
!3 tH
oq
> .
3^
o\ to
Th O
M 10
^ < m
05 T^ tH
01 N N
03 03 O)
tH ^ 05 «-l
2 >
^ ^^'^
10
a
<» <
05 CO
tH 05
c^\0
On 00
M 10
d d
VO 10
1> o
-^ On
d d
C; CO CO (M ><
• -^ <
•^ Co 03
S -^ ^
tH too >
05 Oi CO U
^ tH 5 .
^ .Co .Co .«o 5;
. 5^ ^ ^ .
^ s s s ,^
w
c
o
w
woo
'5
0 ^ J:i
u
W
^00
0
0
0 be bc
bJD
be
C C C
C
C
05 C3 c«
03
03
o ooooOfi;
"^ "73 "^ "^ "^ >— I
O O O O O -Tl
bJ3 be bJD be bD-^
C C C C C ,J^
c3 03 c^ cs rt C/2
g sss g s sssss
465
O CO u-j C. 00 O
00 00
00 00
CO CO
ro
HH ro o CO I> O
N ON tJ- t^ N CO
o
vO
CO
o
CO
CO
CO
On
- O On
M N M
ON '^ c^
c« O
o (u
I -^^
•^ P be
h; fe PO
O
2 c
CO
be
bC«4H
(U o
.S c
cs (U ui
ZQ
li
a-) £
3 §
M ,
CO C/3 Cd
^ CO
C3 rj Oi
-d a
'o 1>
■-^ o B^<:K ^
o
u
03:^0
■=3 g
1 o -a
"5 ^-, ^j
O fe CO
-73
<u
bo
C
C3
rS >< C0.2
^ <u bjo t;
(u 2
C a;
c« biD
CO I
3 C
CO o
cd
a 2
•S 5 g.C
to (U > CO
c/2 pi; <!
IS ^
0) O
Ui •<-; CO
<u C3 -M
> '-' C
C ^ l>
s c a
_, <u C
>2 cu H)
^ "^ a
d fi X
OJ C (U
3 O 1>
•-H ^ (+H
I S fi
o-u ^^
"^ o .G
TO Q ^
^ ^ I
S c<i '^
^
■<* VJ ^
g !< ^
■^ ON
CO N
d d
'H CO
a o5
I
05
K d >
•^ >oq
<^ X 'ii
■^ n •
a
ON
a .
^ a
tH ^
CO O
CO ^ -^
s >
O K 00 K
00^ ON^
a a
^ "i> CO li)
-^ 1-^ I
.0
On 2
M CO
VO ON
ij-i in
d d
^ 2 ^
10 >
CO
o
o
■ij
O
<<
'O £ of
^ J^ £ i^'^ >r
ci IT} a <n a
g SgS
c«
p:^
bS §'5 g
CO +->/->, ><J
c3 C3 (— 1 cd
pi
-d P^
Q Q
M X ^
r! C -r:
^ .2 ^
^5P3
u
q|
p:5n^
>.ri^ >.U 1:3 >>^ ^ rn ^
c3(-JK!0(ua;rz: r;30-5r;:3
^ ^ S ^ ^^^^^ ^ ^
K K 3^
"d
c .. „ ..
cd cd cd C3
o^z
bx) b€ bX)
vh ;-! vh
000
30
466
o
u
0)
00
00
*-H
HH
M
N
M
^
^
HH
N
vO
ro
vO
vO
o
•-1
M
00
vO
1>
N
o
M
00
^ lO IT) lO CO rj- N
MM O O vO \0 CO
I> N d w N w -^
C/2
Pi
O
h
ID
O
X
w
P
iz;
Q
<
Oh
<
O
o
I— I
pq
PQ
^ o
73
<u
bfl
3
o
bJD
bJD
C
bfi
3
O
'o
>
o
^^
Vi
a2
^ O
tn 'r^ rO
o
u
o
o
^ a
^ S
u ^
o
O o
en C
-^ a
o a
<u o
-fi^i:
o
s
O c/2
CO
CO O
tJ-OO
CO -
CO CO
05 O
,5: >
1^
c^
1^
fe^
-ft
M
(M
o
CO
'^
CO
Oi
• r^
a
On
1— 1
N
o
►J
CO
X
•
05
U
U
T-(
CO
^
o
CT>
><
N
T-(
X
u
ri2
^ to"
m
ct;
^•^
. 00 M
^ ^^
03 n, .
o '^ «r
CO ,^ 'u
th ^ Ph
'^K^
X
c
O
t3 73
h ^* . h w h ^"^ J
», ^ fajO -> ^ -^ d
cc c3 "r^ cd o c3 o r72
ooPQOh-Joc/Do
^ ^^
si-
as :3
X ^ r<
c ^ c ^w ^
c3 X (S J2 ,. o
00
467
o
O
CO
CO
r^ 10 o
N O N
cn t}- CO
\0 N
CO
U-) r^ ro M
i>oo o r^
M CO 04 CO
-I 00
o
CO
r^ CO
bJD
0 ^
QJ
1^ ^
0
J- w C
0
1, bioc
hybrid
1, twin
<-M C^
0 u
ner
rval
ner
s S
0 <u
<y ca <u
JO
0 JO
o
111
^ 3 1^
C :3
03 O
PQ Oh
.2 o.
S 2
C .H-
-2 c
u o
C
o
3
G
c ^
.2 c
t^.2
S !=
il
Xi o
i2 -C
OJ
■M M ii -C
u
:3
OS g
C/2 h
OS
t5"
^ d S?
tH .0 ,K
CO 00
M O
CO a
< N (
^
"-^ "^S CO
^ Si -H
o\2
N
§. fee §^
r^ ~9 n^
ti^ -^ ^
- Si ^
o 2 ^
d1
^S'^
■iJ
rl
o
CO 1^
05 -^
^ <3
> ^-
l> . ^
05 ^ ^
cq -^
^^
CO .
05 ^
^1
^•1
CO .
• -^
CO t;
05 s:
d
05
ti^
C0 05
o ^
CM -^
Si
X K
5: ^
^ r^ ^ V
0
X te!
■H
N
0 =s
>
•ci
N^
X!
Ei^
^^
2,^
r.^ ^
a.<
•
0 .
h5 '^
XX
*—>
*— 1
:i:
X
<
F?
F?
g-
c
0
a
rt
C3
C3
-C
C
'>
a;
^
a
<U
(U
tu
QJ
^^ zz;z;
1.1^1
^"
C/^ (X!
C/2
(n m
:/2 Q ^ ^ H^ ^
Q
►— ii— .
►— >
t— J h— >
1 •
rt c3
C3
-a
0 0
0
0 0
u u
0 0
r- 3 -t: Ui c« X :?
:§p^'§m so a
g
gg
;<^
^^
^ ^ :? 0
0
:0 Ch
30-2
468
(U
PL,
T^OO o
N M M
o
N
OS M
t^ c> -^
CO I> CO Th N
O
h
<
o
X
w
Q
Q
Ph
<
O
O
m
c« a,
o
^
a;
5/)
c
c
C3
n
>
+j
g
(1)
3
u
o
(U
T5
Vh
>^
hfl
a
n
^
p
u
(X
j::3
a
be C
a
O 0)
C3 ,^
C +"
S,3"
C O
be b
ir! ^ ^
^ :3 o
1^
a; O
V ^
o c
oo
.?^
^ ^
05 a
00
> CO "^
.CO '
S rH CO
I .05
ft; ><^
^ ^ tf
^ i-T
00 «
^ o
, aw I
CM T-( "^
05 X «^
bo be
Ic3
1^
00
ar-<
Co
^1^
O § ^.
a "^
a ^ t:^ .:
Oi !-! X J
TH >< X X
X? S S
05 S
X +J
q;
^ b^ th -^ a'
^ t^ ^
N
CO ^ . .
f§§
^ ^ I §
O g « -73
^ ^ ^o
CIh Ph
<u
<u
a 3
^
^
u" cr
u
u
cr> o)
c^
C3
c3 ca
Plh
p^
^^
w c« ;3 C
C3 C« (M <u
d^ a, Ph pL,
=:^
HH p-4 HH Ph HH
^,a
w '5
. ,- s a a
^1 s <^ ^ F!
a a
C3 (S
C 0
be c« X2 JD ^
^ Xi
c r/)
i^ oj 'n 'n 'n
N N
<u
CC U< tH M M
Vh ^h
PW
a piH a pi. Oh
a a
469
o
CO
-t
o
o
00"
o
\0
CO
00
o
-d- ro u^ i>
10 O J> N
CO O
CO 1>
CO CO
(U
dj
j::
Vi
,__(
r!
c«
•c
U)
D
<h
0
73
O
a
a
a
■^
OJ
c^
u
2
C
C
§
a
C3
i „
-Q
c
O
C
C3
IS
O -M
'S ij
o
S ^ fl
n :>^
^"^ B .
2 P i=: "^
C3 C ■'-' (-1
C2 a; ^ -S.
F" u oj S
H <U C 4->
2 ^^
2 13
<^ 2
'^ o
^ a
c« O
.2 >
c« >>
00
1 a>
n
Oh O
^ c^
„ G
o o
C '^
O
1 ^
^ CQ .'^
^ rH <i \0
"4i CO 10
N
^ O g
^ T-( _.
O CO ^
§ ^ ^
. 00 ^
ft, a.
eo
^S as
. X . 1
a X! a'
dS
o^^ x:
00 o5 xj
05 ^ 05
CO -vT^
05 ^ w- - I
ON
d^
00
<^
Q.C2 rf
-CD (35
CO
05
:s °-
X *^ :s '^
- > hH
X
X
00
CO .
dt
2*
1> tH O)
S §§3
N 00
Q 00 r^ 42
eo
05
a^
-^05-42
CO ^ .
CO ^ ^
05 s.^.^ ^
^ftq
Qoq
d
d ^
-< .
^. ^
o c
pqj
^ G
^ fa
Oh'
eu
Oh'
Ph Ph
*-| iii oj 0
w
0
d
d
G G^
g'^
G
^^
^
OJ
c
0
G
0
H
X X
X
>
>
(S rt ."G
'^
C^
c«
c^J
0; CL) 0
0
0 0
0
p^
csi
p:;
C!!rt «
Pi pi; Pi p;
Pi
Pi Pi
^
470
Oh
O^ Os XT)
M O^ ON t^ ON 0\
C> M M 1-1 VO ^
O
<
O
X
W
Q
Q
<
<
o
o
?^ ll^ D
0^ &fil3
tD
<u a; +j 03
1^ ,^
to >. C "TJ
3
c
•si
O CO
O ^_,
3 5h
"■' -u
- C3
fcJD n
a; -"
2 2 2
■M 3 3
<u a a
c/2 y^
B b
c
2
a
>
a
o
^^. c«
cl3i^
c o -o
O I— ' •--,
>
^3
a; c/}
Co
£
Q
10 \0
00 o
00 N <^
CO 0} HH O
". ^ ^ d
CO 05 2 05
00 ^ ^ tH
?<
X X
^ vO CO
O 00 O
s s ^
o
X S
"^ «X CI,
1-^1
M 05
q;
■^li i>» to
CD ^
CM K
a >
>
^ ^
^
d^
CO -
X "^
>< S
CO .
CO
X c c c
P^ P^ P^ p:i
p:;p:;
c/D
3 3
P^ P^
o o
3 3
3 3
I
c/:
oo"
o
471
1000
\D en
a O
•a ^
O J
d
00 _-
HH CO
o
O 0^.-'~
I
'■>->
C
O ^
Tf ON
03 TS
i ^
>. -M -^
bf)
a<^ O ° 5h
tH K> <U
O JO
1.2
<+-i CI,
bX) bJO'
O bJO
+3 o
1)
o o
fe »^
00 HH
00 Tt-
(Z)
o
4J C5 +T
(D ^ t-i
bo o
CO
> Si
05
Ei^
:i1
ON <
CO >
X ^
. .0
til
Oi o
bJO .
-0;
Cu ^ J-
-;? ^
CO f^ CD
CO . (N
.^
>{ 5 |£
>< '^ '^T
O <ij -^
^ |.§
die)
c3
M f<5
pT d
CO -
<^ s
05 CO
X! O
CO .
On
NO
Etl
w Oi
CO 03
X O
^-^
^ i^
?5 •- g"^
. ^^ CO ^
^ NT-( -^
^ CO > I
C<1 (^
. 05 .^ -^
oq
I
u
bfl
— O
.^ o
U -M
o
o
CO
S^05 J
^o g
O -
C3 03
^ ^&
rh 'a?
CC I— I
O
00
o o
u o
c3 .
§1 ^
(n
2 c3
472
-a
o
bo
Oh
00
On
m M 00
o en r^
On
O ro On rh
moo O CO
CO M (VJ M
00 I> t^
O Th 00
ov Th 00
• " C iS
3-^
P .2 -^ t5 P
o
22 fi d
C3 »--)
3^.
>>.2
• " 1)
o o
'^ 2
- <u
be
^ .2 s
<u
c
o
II
S >
<u C
> cd
Q
2
"a
ii
l3 5-1 P
'^ O -i^
^ ^ O
.2 ^ a
,^ c i>
C« (J
a - .
Q w w O
o g
be c?
if "§
13 c
w Q
a
o
CO
CO
<5 Oi
O -^ -^ - -
^ K CO
.o ^^
dp.^
> ^ m
^ CO
> o
in \o
d o a
CO CO <
^ CO ^
> 05 >
--si s<r 5?
^ .O
S' "^ R
M CO On
«u
CO
a ON*"^ oi
i^?
m -
-^ >^
t <=> r; •
X o3 a ^
00 -To 1^
i > 05 "^
a -S^ s ^
Co X ^
IN . X ^
O --O . .
Oi h ^ g
> . -^ -^ ^
N Ni ^
^2 -
a^ N
>
On X
<: "^
CO S^
o .
Oi ^
> ^
05
T^ >
^ X
X
K- a
'^
Oi
-^ . -^ ^
aoa.
w
ffi K s^
XX
c
c
G C C
c
c c
c
hA^ C
c
c c c
c
c c
fl
' C3
cd
03 CTJ 03
(^
« ra
c^
r^ J^ S
a
sag
a
s a
g
(U OJ 0)
(U
(U (U (U
<u
<D <U
<u
a a a
a
a a a
a
a a
a
mmin
(n
mmm
C/D
m m
m
473
Tj-
^
i-i
5
"S
10
N
§:
HH
00
tn
Cv
10
0
'^
M
Th
I-I
y_f
i-i
Tp
1>
J>
CO
10
CO
00
CO
N
T|-
^
CO
^
00
00
M
M
CO
M
CO
^^
10
CO
CO
CO
Th
6
00
0
00
1>
M
^
U-)
^
•^
■^
N
r^
CO
n"
M
M
^
0}
-t-
^
rn
ro
vo"
■^
Th
10
m"
rh
't-
10
10
N
^
-h
00 N
Tf
^
H-l
M
l-H
-^
CJ
i>
'-'
^
'"'
c-I
utO
• S
'u
:3
c/T
:3
u
1
0
2
1)
■%
3
0
TS
g
Ti
>
0
c
1
1
1
0
C/3
c
0
:3
CU
1
0
u
73
1-1
13
0
Oh
c
0
1
c
C3
c
c
bo
13
C
C3
-73
73
0
u
d
a
C
'^
'd_
3
C
O
cr
W
bfi
>
2
D
U
'5
0
13
C
1
73
fcX) 0
C/3 ^
2 -a
i^ So
w 22 d
C
Is
Q
0
(73
'c
C3
0
2
0
73
;-(
C
<
^ 0
73 CJ
^"^
m"
x"
oT
CO
CO
10 pf
^
i^
0
00
biat. Rec. xxvi, 1923, p. 41
enaisch. Zeitschr. Nat. xxviii, 1894,
p. 196
-H
S3
'"'
Th
10
W
CO
0
r^
f^
ON
05
>
i
1
p. 2i6 ; Zeitschr. Wiss. Zool. cxx
1928, p. 105
tzber. Ges. Nat. Freunde Berlin,
d
O)
tH
X
d|
ccTkl
00
d
tH
>
1
05
d^
a
S si
1-S
p. 493 ; Proc. Roy. Soc. Lond.
CII, 1927, p. 177
atiirzviss. xvii, 1929, p. 287
rch. Entwniech. cxxiii, 1931, p.
CO
d
of
CO
05
T-<
X
X
X
X
1
si
10
d
co'
CO
05
TH
X
X
<A
1
10
10
d
t>
05
x"
-^
CO
d
05"
tH
05
S
X
1
1
1
d
<M
05
X
U
1
"^
i
j
1
CO
d
Oi
dc^
05 ~
tH
>
i
N
0
^
CO
"^
"^ oq
^
-^Q^
'^
^
^
'^
05
^^
13 . 13 • 13 13 -^
c w c D:: c c c
.
X
X
X
ed:
XXX
^X
c
c
^^
c c
c c c
be -ft
c
g
C3 C3
c c c
rt c;! c3
a a a
aSa^a:5a^a'S-^
II
aB
<u
OJ
OJ
0 (U
(U 0 OJ
a,
a
Oh
a a
a a a
a^ a d. a a a
m
ai
C/2
m (n
maim
m m ai m (fi u^
C/2 c/3
474
T3
O
a'
lo -rh c\
r^ Tf ON
M M w
o vO
o
CD
O 00
lo O
Os in On
O O ON
Pi
O
Eh
o
X
w
Q
g
Q
<
<
O
O
5
.2 -«
< o .2
D
Oi
<U D u.
WWW
o 5
is ^
<^ 1:2 _C T!
be
bX)
C o C
W
00
CO
2 ^t^
>< ^ »
X >< ^
■^ ^ ►::t 03
^ ^ ^
CO
o o
05 tH
'^ ^
,Et^
ON <o
"^ tH ^
S 05 . •
^ g X ^
ir)05
ON CM
. 05
DhTH
e CD "
^ <M 1=1
■§ ^ >
^ 13 .
^ x^
-OnI^
■ ^
^ 05
aw ^
CO - o
05 O X
th be
vS £ s
CO ^
10 05
CD >
o
05 O
•",^05
O
:: >
^00 o
d^ i
ON o - to
. ^ 05 g
fin 2^
tl^
0
•5
ffi
<
w . th
N w s
•^ ^--^
2iS g
a> (D oj
mmm
^
Ph fin
CQ
X^
z"
^■H
CLh
^
<U J^
Vh
> ^
^
(U 0
:0
C/5
&^==d
hJ
giti; 53
2^"^
^^■§s
c
c^ C3 ^ (U
C/} C/2
m
m
01 (n m
-d
c
m
<
pq <I
^
'3
umne
Well
uzuki
C/2
^ <n
475
HI T:t-00
00
CO
o
CO
o
m
o
en
.5 o
c3 :«
O C
C be
cc fcjo
a o
'o ■
o
^ h
a
a
>
Q
O (L>
■(->
a. «
O <:« 1^
• S 3 OJ
j:; (u w
;!. g.S
w h:j CU
M
D
^
.2
en
>>
o
<-M
T3
o
OJ
>>
2
W)
o
o
a
<1>
a
s
X
CO
t
fi^
CO
d On
05 Oh
^.^^
-M CM
^05
o "^ 2
in >
X
I S
E " S o"
« o ^ ^
-d-d
-C) N ^C) W
CO
o
CO
^
^Os
5 &
ti^
o
^ d
S 05
2 6--:
6 -
Oh
CO
O CO
ic d d o^ '^
^ o S2 oo~ c*3
^ S S c<i 05
> J
^^l:5^N
Oh (S
cq
p tH
CM ^ . .
05 .« ^ -^
Q J«» J*.
fe
a; ^
>^
03
•— I 1-^ <U
^ (S C«
c/^ h H
^ (U
13 ^ o
{66
>>
3 'a ^ ti ti
476
o
3
Ti- O vO
N CO 00
Tt MM
en CO M f^ M
ON 00 N vO
CO M
IT) ON
M en
u
be
O a-»
'T3
CO O
C3 I-'
o 1>
O E w
.2 o
5.5
^ 2
^ a
21
+-> *-i
i2 -^ 13
.S -S .-
CJ y c«
mho
.2 a
t^.2
? o
o O
O ^
^ 2
s g
bo en
0) 0)
a
O
4-> C«
.2 .2
"5 "blD
B a
^ S
T^ o
o^'~
era x
> .
N .
NN
^00
d d
d>^.
l> X >
(M O X
05 ^ O
tH O .
xN ^
- a<
00
CM
a
00 -
N >
d;;;
"^ § §
n, "^ '^
g
^ df
^'^^ ^
dt g.§
1^ Co
01
d ^ «
^ ^ CO
X ^ S
K ^ s:
O
^ ^00
d4^"
2^ 05 05
in X -
XXX
X ^ o
^1
c5
^
'0 ,
u
a J
O
"=« ^ fl c
a c
c .
j::
d^^ ?
? ^
5 w . .
3
. - w w •
<
>'§jj
hJ J
< ^ ^ ^ >—>
>; l^"^*^
^"-C
rn ^ -^ -^ >.
^
^
^
^
^
^
4^^^ o u
u o
^ c c c ^
^
^
^
4^
.i_i
4^
4->
4^
4-1
^ c/2 ;S 3
X! X2
•^ 2 a s 0
^
^
b£)
o
bx)
o
bJO
o
bi)
o
h pp
pp
pppp>
>
>
>
>
>
>
477
ON fO
00 O
Ti- CO
CO
en (u
O -T3
.2 _o
bJD c
o ;.
o C
^ o S .c5
V)
o -n ^ o
2 J3 5
:n
1) r^ -^
O -J
C 2
O OJ
C bX)
O^
CO
X
a
u
^
^
■H
M
05
ON
Is ><
W^ ^ . CO
'^ ^ CO
r^ CO ^ r^
•^"^^
■S; g «" "
.^ "-2 g
<^ ^ .
CO . s: -ft
CO V" 2 S*
rco . s: -^ o "^
>
(3
^ 00 CI.
^ 00 ^
.42 00 00 .
••^=2 1
d^ § r^
X ^ -^ ^
in 13 CO (3
CO 5; t^ K ns
< s « § .
^ p=^ rf< (<; 10
05 --Si 05 "< Oi
0
o
bX)
j2 <ij
^ b£-- b£ g bJO
o
E >-*
^ & ^^
d
c
S
.bX)
^ ^
478
bX)
w
PL,
1>
O !>•«
ON CO
lo ro lo
M ON M
CO
NO
vO On VO
O 00 O 0\ lO
\0 I> M M N
00
01 vO
Oh 3
c ^
o r73
r/1
TJ -M
-n
^S
03 O
^^
'^
bcXi
§.s
hJhJO
1 1
c .2
fcJO OJ
tJD tUO
O -71
c« c« cC
.t; .ti -^
3 3^
rid
§.2§
t5-2.2
s - i
-d '3 -o
2 § 2
S '^ S
0<0
be
C S ^
a3 i^ ^
5 o a;
a< o '+3
>3 a
00
lO CO
d d
05 Oi
Et^
m
00
CO CJ
- Oh
05
05 CO
00 \£) NO
CD Oi 1^*^
CM 00 cq
05 tH 05
^ 03
CO «
> s
s3
> 00
O 00
o
^O 05
8 ^ d
I of
^ ^^ -w 05
^«^:-
-fi ►fi "fs ><
b^ b^ ^ o
J I 05
.9 h* X
^ '^- "^ D^ "^
> CO ^
0}
d
<
CO
05
. CO
a; §
•2^
cc;
I
S
^
■♦J ON
Si
O
j::
:3
<
P^(^Q^
cufc
wOd
^ ^ ^
•—I <— 1 1— T
c« c« O)
(U QJ <V
cfl N N
(U (U (U
(U <u
<U <U (U
&^&
&&
&^^
^*
QX
ffiffiK
x^x^p^
illier, B
'illier, B.
illier, B.
'illier, B.
Rawles,
illier, B.
Rawles,
ilson, C
^^ ^^^ & & ^
479
00
ON
t^ M
M CO
o -
M O
ON
so"
O
so
d
C> CO O
lO T^ ON
(NJ CJ H-(
OnOo" rj- o" tF
ui lO -^ On vo
(NJ (NJ (NJ l-l W
CO
00 fC
^oo
(-;
GJ
us
0
f=
Xi
^
0
m
t3
1:3
a;
(/J
3
0
-o
c«
OJ
--a
S s
3
'o
u
2 ?^ c-i
^ 2
bJO Oh
fcjO M
^^
<ii o
II
~5 ID
a ^
c > o
-w P w
^ c« cd
o jh c
^-^ o
a, oj
J5
a
o
U3
On 05
eg ><
a
a
a
/^
13
CO
<
M
W
0
05
^
-*
T)
0
0
;;
a
03
CO
.^
tH
tH
~
>
^
^
^
X
^
'"'
0
-^
-o
>^
^
^
g
^
^
1^
eg a
> ^
o >
. a X
^ -^ X
CO CO ^
S ° S
t^ I!
o
CS^TH^^
^ o \o Co
>
. CO
W 5
•-( NO O '^
lip
^ J > -^
J *^ o Co
d^
StH
^ o
1^
5-^ s t^ ^ r<:
fe^
^E^C^ ^ti<
g s ^ §
f^
t> o
^ d
. 00
s ti
Oh O ^
^ ^ tH
Si ^
g-d-S
O.T-< ^
^ th -l:^
05
m pq pq pq pq >
. . c c
C
c
fe
►-M
w w w w w d:
w
w
w
C3
s
6 6
c c c c c c
JZ
r"!
^
^ ^ -^ -c
'O
•"0
fcJO
000000
0
U
0
0 0 ;: ;:
^ ^
-a
w oi c« w w w
O)
c« M (U (U
OJ
0
0
.}^
.^
.^
•ti .ti 0 0
0
0
0 0
0
&^&^&^
^
^
&
^^^^
^
^
&&
^
480
m
P^
O
X
<
O
X
w
Q
g
Q
<
<
O
o
m
m
■^
0
p
cr
<u
ro
<u
ON
^
ro
^
^"^
0)
00
b/)
CO
J^
^
Ph
00 0
00
r^ N
0
00 r^
O On
> bD
-a ^
§ «^
'rt en
^ 0^
O C
3 B
c/3 Z
"I
o ^
a ^
0 -d
1 g
00 i>
ON M
X .0
it?
^ NO
co<^
0\ >>
N ^
d a
O Oi
05 00
ti tJ
c
J3 oj
0
ss
c«
3
•-" 0
<-M
1^ ^
03
d-p
g
CJ >.
<U
zx
;z;
NO
d^
CD Ct^
Oi .
00 -^
0
•5
:s
<
*—i .
oT^
^
c ^
2 3
5
S ^
c«
03 03
03
>H>.
>^
a;
U
V)
P^P^
c -
C/J
(U 03
M
D o'
:3
Nl N
N
APPENDIX
I . Exogastrulation in Amphibia.
The work of Hohfreter (Arch. Entivmech. cxxix, 1933, p. 669;
Biol. Zentralbl. Liii, 1933, p. 404) on this subject only appeared after
this book was in page proof. It is, however, so important that we
have decided to summarize it in an appendix ; and have taken the
opportunity of adding some other points that had been overlooked.
Head muscles
MoutK endoderi
Fig. 213
Exogastrulation in Amphibia (axolotl). Diagrams showing {a, b) the mass-
movements of the organiser-region and endoderm ; c, the structure of an exo-
embryo. Note wrinkled amorphous epidermis, exogastrulated endo-mesodermal
portion inside-out. (From Holtfreter, Biol. Zentralbl. liii, 1933.)
Holtfreter discovered that by the simple procedure of removing
the early blastulae of axolotls from their membranes and placing
them in Ringer solution of about 0-35 per cent, strength they could
be made to exogastrulate — i.e. the presumptive endoderm and
mesoderm is evaginated instead of being invaginated, leaving the
presumptive ectoderm as a hollow sac. Stages in the process are
HEE 31
482 APPENDIX
shown in figs. 213, 214, It is of interest to note that the tendency to
constriction in the marginal zone (Ch. iii, p. 42) still manifests
itself, leading to a waist between the ectoderm and the endo-
mesoderm from the earliest stages of gastrulation. Later the waist
becomes still further narrowed to a stalk, which can be easily
severed, and may break of its own accord.
All the mass-movements of the different regions involved in
normal gastrulation (p. 43) are still operative in the exogastrulae,
though their mutual interactions in the altered circumstances are a
little different, as indicated by the arrows in figs. 213, 215. For
^^^^
r'
Fig. 214
Exogastrulated axolotl embryo, 8 days old. On the right, the epidermis ; on the
left, the exogastrulated endo-mesoderm : note pharynx with inverted gill-
pouches (on left). (From Holtfreter, Biol. Zentralbl. liii, 1933.)
instance, the organiser-region, forming the dorsal side of the
marginal zone (p. 41), stretches out as a tongue on the dorsal side
of the evaginated mass, and subsequently becomes sunk in a
groove and finally overgrown by the endoderm. This confirms the
view that the dynamics of gastrulation are predetermined in the
various local regions of the germ.
The final result of exogastrulation is what we may call an exo-
embryo (Holtfreter's Exokeim). This consists of two very distinct
parts. The ectoderm has flattened down to an irregular wrinkled
mass with the blastocoel largely obliterated. It shows no medullary
differentiations, notably no trace of neural tube or even of local
thickening to form neural plate tissue.
APPENDIX
483
The endo-mesoderm on the other hand bears a considerable
resemblance to an embryo, showing well-marked regions — a head
and gill-region with gill-clefts, a trunk-region, and a tail-region.
However, it is entirely abnormal in its detailed structure. It is
morphologically inside-out; its outer
layer is endodermal, and this contains
a more or less solid mass of notochord,
somites, mesenchyme and cartilage
(figs. 213, 216).
The endodermal epithelium is, as
in the normal embryo, polarised: but
its outer surface corresponds with that
which bounds the gut-lumen in normal
ontogeny. This may be compared with
the fact noted on p. 250, that spheres
composed of gastral layer only (collar-
cells), arising in sponge dissociation
experiments, have the collars directed
outwards, whereas in normal animals
they face the gastral cavity. In both
sponges and amphibia, one surface of
the epithelium orients itself towards
the most favourable environment,
whether this be an internal lumen or
the external medium.
Exogastrulation gives us a method
by which ectoderm can be totally
separated from endo-mesoderm from
the first onset of the gastrulation-
process. In addition, it provides pseudo-
embryos, containing all the derivatives
of the endo-mesoderm, in which we can be certain that no nervous
tissue is present ; and further, in the inversion of the endodermal
and mesodermal layers, it provides a natural experiment in ab-
normal spatial relations which it would be impossible to duplicate
artificially. (See also p. 252 for a comparable case in insects.)
As might be expected, conclusions of considerable importance
have been arrived at by analysis of the results. In the first place, we
31-2
Fig. 215
Diagrams of transverse sec-
tions through Urodele em-
bryos, showing the structure
and directions of movement
of parts in a, normal embryo,
h, exo-embryo. v.P. site of
original vegetative pole. (From
Holtfreter, Arch. Entwmech.
cxxix, 1933.)
484
APPENDIX
have the complete failure of the nervous system to differentiate in
the ectodermal portion. The only difference observable between
presumptive epidermis and presumptive neural plate is that the
latter shows an autonomous tendency to elongation in the direction
of the egg's major axis. We shall later return to the absence of
ectodermal differentiations.
In marked contrast with this incapacity of the ectoderm is the
capacity of the endo-mesoderm for self-differentiation. We have in
^^^
Fig. 216
Transverse section through trunk-region of exogastrulated endo-mesoderm,
axolotl. Note superficial endoderm surrounding notochord, myotomes, con-
nective tissue, and (below on the right) heart. (From Holtfreter, Arch. Entwmech.
cxxix, 1933.)
the first place typical notochord. Then the mesoderm produces the
following derivatives: in the anterior region, head-musculature,
mesenchyme and cartilage ; in the trunk-region, somitic mesoderm,
pro- and meso-nephric tubules with coelomic funnels and asso-
ciated with gonads, smooth gut-musculature and (empty) hearts
capable of rhythmical contraction, empty endothelial sacs, masses of
blood-cells, coelomic spaces, and connective tissue. The endoderm
shows equal powers of self-differentiation, and gives rise to buccal
cavity, pharynx with endodermal portion of the gill-clefts (the
visible apertures on the surface of the exo-embryo of course
corresponding to the normal internal apertures leading out of the
APPENDIX 485
pharynx), thyroid (probably), oesophagus, stomach, lungs, liver,
pancreas, intestine and rectum. The various sections of the gut are
characterised by the same histological peculiarities as in the normal
animal, e.g. ciliation of the oesophagus and typical glands in the
i %Fi
a
Fig. 217
Self-differentiation in exo-embryos, axolotl. Extrusion of yolk-cells, a, in normal
embryo, into lumen of gut ; b and c, in exo-embryos, from the surface. (From
Holtfreter, Arch. Entwmech. cxxix, 1933.)
stomach. Perhaps the most remarkable self-differentiation is that
of the small intestine. In the normal axolotl larva of a certain
definite age, a number of yolk-rich tells belonging to this region
partially degenerate and become detached into the lumen and are
subsequently digested by the remainder of the epithelium (fig. 217).
In the exo-embryos, this same process of degeneration and detach-
486 APPENDIX
ment occurs at a corresponding stage, though of course the cells
here are detached into the surrounding medium (fig. 217 ^ and c).
The process occurs only in the central region of the gut corre-
sponding to the small intestine. This determination of a timed
degeneration recalls that of the isolated chick mesonephros (p. 205).
The attainment of functional activity by many tissues in the
demonstrable total absence of all nervous tissue is of great interest
(cf. p. 430). The epidermis, gut, and pronephros tissue, and prob-
ably thyroid vesicles, embark upon active secretion, and ciliary
activity sets in where expected. Spontaneous rhythmic movements
of the outward-facing gut-endothelium occur regularly, brought
about by the underlying smooth musculature, thus providing the
first demonstration of the independence of this tissue of innerv-
ation for its functional differentiation. The heart may also reach
this stage, confirming explanation experiments (p. 203). The
striated skeletal musculature, however, was never observed to
contract, either spontaneously, or in reaction to mechanical,
chemical or electrical stimuli: later, the degenerative changes
typical of denervated striped muscle set in. Thus the histological
difl^erentiation of skeletal muscle can be reached (though not
maintained) in the total absence of nervous connexions, but not its
functional activity. This confirms and extends other work (p. 431).
In spite of the remarkable self-diflFerentiating powers of the endo-
mesoderm, the structure of the exo-embryo is abnormal in a
number of respects. The head and trunk-musculature, though
differentiating histologically into typical striated fibres, is not
arranged in a regular metameric plan, and the direction of the fibres
is irregular. The cartilages of the head (no cartilage appears to be
formed in the trunk-region) are massed together in a single irregular
lump. The total absence of the cartilages arising from the neural
crest (p. 394) is doubtless largely responsible for the lack of regular
arrangement of the head-musculature, as well as for its small size.
The gonad appears not as a long ridge, but as a series of small cell-
masses in close connexion with the pronephric tubules. The liver
tissue is always very small in amount, and late in appearance ; no
gall-bladder has been noted.
The tail-region is of particular interest. A caudal zone of activity
is present in notochord and trunk-musculature, and a conical tail-
APPENDIX 487
bud arises in a more or less typical way; but it never becomes
large, its growth is soon arrested, and it is finally resorbed, in spite
of the absence of any degenerative signs in its tissues. Holtfreter
ascribes this (and also the absence of regular segmental arrange-
ment of the trunk-muscles) to the absence of the neural tube and
especially to the absence of the mesenchyme derived from the neural
crest, which is known (see pp. 193, 396) to have the tendency for
directive outgrowth. As the neural tube and crest are first in-
duced, by the chorda-mesoderm, and then supply material necessary
for tail-mesoderm diflferentiation, we have here an interesting case of
mutual induction on the part of an organiser and of that which it
organises. The same interaction is apparent in the head-region,
where neural crest material, originally induced by the prechordal
portion of the organiser, appears to be necessary for the proper
anatomical differentiation of the tissues (muscle and cartilage)
derived from this region. (See also p. 181 for a comparable case of
mutual dependence in sea-urchins.)
Among other special points may be mentioned the fact that teeth,
even partial or rudimentary, are never found in exo-embryos,
showing that the presence of ectoderm is necessary for their initia-
tion. Taste-buds, however, do differentiate in the pharynx, thus
demonstrating that the view sometimes maintained of their deriva-
tion from immigrant ectoderm is incorrect (see p . 498). The presence
of gill-clefts shows that their initial determination proceeds from
the pharyngeal ectoderm and is quite independent of the presence
of ectoderm. The fact that they later disappear, however, suggests
that contact with ectoderm is needed for their maintenance.
Blood-tissue is rarely found in exo-embryos, apparently because
its primary site of origin lies far back in the ventro-caudal region,
and from here it often tends to become included within the
ectodermic vesicle.
The ciliary beat on the surface of the ectoderm is also of interest.
In the normal embryo this is directed in an orderly way, in a
predominantly antero-posterior direction (see p. 236). In wholly
isolated ectodermic vesicles, however, it is completely irregular,
indicating that a polarity or polarized gradient-field is normally
imposed upon the epidermis from the underlying endo-mesodermal
tissues. This is beautifully demonstrated by cases in which exo-
488 APPENDIX
gastrulation is not complete, but a portion of the ectoderm has been
underlain by endo-mesoderm and has been organised. In such
portions, the direction of ciliary beat is regular and normal, while
remaining irregular over the rest of the epidermis (fig. 218).
This brings us to a more general consideration of partial exo-
gastrulation. Total exogastrulation is a comparatively rare occur-
I
s.
Fig. 218
Induction of polarity in epidermis by underlying organiser. The direction of
cilia-beat (indicated by arrows) of non-underlain epidermis is irregular and
chaotic; that of epidermis underlain by organiser-tissue is regular and polarised.
(From Holtfreter, Biol. Zentralbl. Liii, 1933.)
rence : in the majority of cases, exogastrulation only proceeds to a
certain point, and then the remainder of the endo-mesoderm is
invaginated under the ectoderm. All gradations are to be found
from a minimal invagination to a normal embryo. The first step is
the presence of some blood- and yolk-cells in the ectodermic
vesicle. When they are present, they induce a smooth two-layered
epithelium, with normal tempo of differentiation, in place of the
irregularly folded and wrinkled epidermis derived from wholly
APPENDIX 489
isolated ectoderm (see figs. 13 and 219). This appears to be due
largely to the formation of mesenchymatous vesicles containing
fluid, which produce normal tension in the ectodermic vesicle.
Fig. 219
Organisation of epidermis, a, wrinkled, irregular epidermis of axolotl exo-
embryo not underlain by organiser; b, two-layered epithelium induced by pre-
sence of underlying connective tissue and blood-cells. (From Holtfreter, .4ir/z.
Entwmech. cxxix, 1933.)
Neural formations are never induced in such conditions, but are
always formed if any of the ectoderm comes to be underlain by
chorda-mesoderm. If only a narrow portion of chorda-mesoderm
490 APPENDIX
is invaginated, the actual neural plate may be much narrower than
the presumptive neural region (see p. 155). Further, it does not
differentiate into a complete but undersized nervous system, but
into tail neural tube only : the organiser, in other words, has regional
properties. If only a slight degree of invagination occurs, the in-
vaginated material is presumptive caudal tissue, and the induced
structure is a tail (fig. 220 a). This then grows out as quite a normal
Partial exogastrulation and progressive organising effects of varying degrees of
organiser-invagination. a, slight invagination (of tail-organiser only) producing
only a tail ; b, medium invagination (of tail- and trunk-organiser) producing tail
and trunk ending anteriorly with gills ; c, nearly complete invagination (tail-,
trunk-, and head-organiser) producing an embryo which is complete except
for the diminutive length of the head and size of the eyes. (From Holtfreter,
Biol. Zentralbl. liii, 1933.)
tail, confirming the conclusions reached above (p. 193) as to the
co-operation of mesoderm and neural crest mesenchyme in normal
tail-elongation. With progressive increase in the amount of tissue
invaginated, there is a progressive increase in the amount of
organisation of the ectoderm, first trunk-structures appearing, then
gills, and finally head-structures (fig. 220 b and c). The direction of
ciliary beat is normal on the organised portions of the ectoderm of
such partial embryos (e.g. with tail only), irregular on the un-
APPENDIX
491
organised portions. When the embryos are almost but not quite
complete they are cyclopic (see pp. 245, 350).
There is one further interesting point to be mentioned. When
only somitic mesoderm is invaginated, the resultant neural tube is
very small and is abnormal in cross-section; only in the presence
of notochord material will a full-sized and normally constructed
neural tube be induced (see p. 374).
Fig. 221
Regional inductive powers of exogastrulated organiser, a, diagram showing
experiment of placing pieces of undetermined epidermis on the anterior (i),
middle (2), and posterior (3) regions of the organiser (chorda-mesoderm) of
exo-embryo of axolotl. b, tail induced from epidermis in position 3. (From
Holtfreter, Biol. Zentralbl. liii, 1933.)
A further set of important experiments was made by taking un-
determined ectoderm from normal early gastrulae and laying it on
different regions of the chorda-mesoderm of the exogastrulated
endo-mesoderm in the exogastrulation experiments. This must be
done in early stages, before the presumptive chorda-mesoderm has
disappeared under the surface of the endoderm (see fig. 221).
The results are striking. From whatever presumptive region the
pieces of ectoderm were taken, they differentiate in accordance
with the regional properties of the organiser on which they are
lying. If placed on the anterior (head-organiser) part of the
492 APPENDIX
chorda-mesoderm, the ectoderm is induced to form brain with eyes,
nasal pits, and ear-vesicles. If placed on the trunk chorda-meso-
derm, it produces a normal spinal cord, which becomes displaced
below the surface. And if the piece of ectoderm is placed on the
protruded tail-region, the chorda-mesoderm grows into the
ectoderm, induces a neural tube, and then the two tissues in co-
operation grow out as a typical tail. The regional differentiation of
the chorda-mesoderm into head-organiser, trunk-organiser, and
tail-organiser, is here clearly seen (see p. 147).
An interesting point concerns the behaviour of any endoderm
which happens to be overlain by such a piece of ectoderm. Instead
of the epithelium being polarised with the free or distal ends of its
cells facing the outer medium, as over the rest of the surface of the
exogastrula, its polarity is directed internally, as in a normal
embryo ; and in a number of cases a miniature gut-lumen is pro-
duced. Thus, in normal development, the topographical arrange-
ment of the germ-layers determines in the gut the normal polarity
of its epithelium, and the formation of its lumen, although as noted
above (p. 483) the determination appears to be purely mechanical
in its nature.
But perhaps the most important of the facts revealed by these
exogastrulation experiments concern the absence of neural differen-
tiation in the ectoderm. This is all the more striking in view of the
indications which previous work has given (see pp. 50, 136, 203) of
the existence of a labile determination of neural folds, as evidenced
by experiments of removal or inactivation of the organiser-region,
and of explantation and interplantation of portions of blastulae (see
figs. 18, 62, 63).
It might be held, and is held by Holtfreter {loc. cit.), that these
new results show that an invaginated organiser is indispensable for
the determination of neural folds ; that these results dispose of the
hypothesis of a labile determination (and therefore of a *' double
assurance") of the neural folds; and that the conclusions drawn
from previous experiments are erroneous. While realising the
strength of this argument, it is as well to consider the possibilities
that the non-appearance of neural differentiations in the exo-
gastrulation experiments may be due to other causes. In this
connexion, four points may be called to mind:
APPENDIX 493
i. So far as it goes, there is in the exogastrulation experi-
ments a determination for the presumptive neural fold region
to stretch more than the neighbouring presumptive epidermis
(p. 484), and this determination must be independent of an
invaginated organiser.
ii. On comparing the experiments in which the organiser-region
is removed or inactivated with the exogastrulation experiments, it
will be noticed that there is a difference in the distance between the
presumptive neural fold region and the organiser, and in the time
during which the latter could act on the former at any given distance.
While previous experiments have led to the view that there exists
a gradient-field under the influence of the organiser before in-
vagination, in which the presumptive neural fold region undergoes
labile determination, the conditions of the exogastrulation experi-
ment are such that it might be argued that the gradient-field is
deformed, or even not formed.
iii. The labile determination in question would be part of the
general effect of an uninvaginated organiser working from a
distance as the centre of an individuation-field (see p. 310). It is
held by Holtfreter that neural fold formation can result only by
contact with an underlying organiser ("evocation", Waddington
and Needham). But it is clear from the experiments on newts (see
p. 149), in which trunk-organiser is grafted at head-level and
induces the formation of head-structures, that the tissues there are
under the influence of an action exerted from a distance by the
host-organiser acting as the centre of an individuation-field. The
same conclusions emerge from experiments on birds (see p. 162).
There is therefore evidence for the existence of individuation-fields,
in both Amphibia and birds, which is not disproved by the exo-
gastrulation experiment.
iv. The variations between the results of explantation of portions
of blastulae in inorganic media, and those of interplantation of
similar portions into living embryos (see pp. 139, 317), show that
the reactivity of the tissues (i.e. their differentiation into epidermis,
neural tube or notochord) is markedly affected by environmental
changes. It is possible that the lack of differentiation of neural
structures in the exogastrulation experiments is to be explained on
such lines as these.
494 APPENDIX
2. Lateral temperature-gradients and gradient-fields in
Amphibia.
Gilchrist {Jouni. Exp. Zool. Lxvi, 1933, p. 15) has recently re-
attacked the question (raised on p. 342) as to whether the gradient-
system in the amphibian embryo can be directly altered and
deformed by means of lateral temperature-gradients so as to give
rise to asymmetrically developed neural folds, independently of the
effects of alteration of growth of the invaginated organiser. A new
method introduced consists in applying a lateral temperature-
gradient for a certain (not too great) length of time, and then to
reverse the sign of the gradient for an equal length of time. Eggs of
the Urodele Tritiinis thus treated, and heated first on the right
during the late blastula and then on the left during the early
gastrula stages, show abnormally large neural folds on the right
side.
Gilchrist draws the conclusion that the processes of neural plate
determination (in what we should call the primary gradient-system)
take place during the late blastula, for which reason the temperature-
gradient to which they are then exposed is able to bring about a
larger development on the heated side. This view receives support
from the results of other experiments in which the temperature-
gradient is applied earlier : heated on the right in the early blastula,
and on the left in the late blastula. In such cases the neural folds
are larger on the left.
We are, however, not informed as to whether the invaginated
gut-roof is symmetrical or not, and it would still be possible to hold
that the late blastula stage is the critical time for neural plate
determination, not because of any effect on the gradient-system,
but because the processes of invagination of the organiser (gut-
roof) are also susceptible to modification by temperature at this
period. In this case, the effects of the temperature-gradient on the
neural folds would be indirect, and exerted via the organising
action of the gut-roof. Against this, however, it must be mentioned
that Gilchrist presents evidence of the early blastula stage as being
that at which the embryo is most susceptible to temperature-
gradients for the production of abnormalities in the subsequent
processes of gastrulation.
APPENDIX 495
3. Gradients and prelocalisation in Polychaetes {Nereis).
The work of Spek (Protoplasma, ix, 1930, p. 370; and Parts iv
and V of Gellhorn's Lehrbtich der allgemeinen Physiologic, 1931) on
this subject presents such a number of features of interest, that it
deserves special mention here. To the technique of intra vitam
staining of the early stages of the developing egg of Nereis, he has
added the use of indicators enabling him to detect changes of pYl
in different parts of the embryo. The unfertilised tgg of Nereis
contains a large germinal vesicle (nucleus) surrounded by a number
(about 30) of drops of fat arranged around it in the equatorial plane ;
in addition, there is, scattered through the tgg, 3. large number of
small albuminous droplets containing a lemon-yellow pigment.
This pigment is found to turn violet in acid media and therefore
acts as a natural indicator.
Fertilisation results in the formation of the polar bodies at the
animal pole, as well as in a number of changes in the cortical
regions of the cytoplasm which do not directly concern us here.
What is of great interest, however, is the fact that soon after the
polar bodies are extruded, the fat drops and albumen droplets
undergo a re-arrangement, as a result of which they leave a clear
zone or "pole-plasm" (see p. 113) at the animal pole and become
concentrated in the vegetative hemisphere of the egg (see p. 119
for a comparable case of re-arrangement and prelocalisation of egg-
contents in Ascidians). The albumen droplets in the equatorial
zone take on a peculiar colour when stained intra vitam, and the
whole egg shows a clear stratification along the main axis.
The egg is in this condition when the first two (meridional)
cleavage-divisions take place, leaving four blastomeres, each pos-
sessing the characteristic stratification. There next occurs a remark-
able change in the pH of different regions along the main axis, as
shown by the natural and by the experimentally added indicators.
The region of the animal pole shows a shift to the alkaline, while
the region of the vegetative pole shows a shift to the acid side of the
scale.
The whole phenomenon, as Spek says, presents the appearance
of a natural experiment of cataphoresis. The changes in pH at the
two ends of the main axis of the egg reflect a gradient in electrical
496 APPENDIX
potential, due to the accumulation of ions of opposite sign at the
two poles, and resulting in the segregation of the egg-contents and
in their stratified re-arrangement.
The next (third, equatorial) cleavage separates the animal cyto-
plasm (mostly clear, "alkaline") from the vegetative cytoplasm
(granular, "acid"): the former goes to the formation of the first
quartet of micromeres (i a to id), while the latter forms the
macromeres {i A to 1 D). The droplets which become included in
the micromeres are those which lay in the equatorial zone of the egg,
and, as we shall see, are destined to become included in the primary
trochoblasts (see also p. 132).
The forces which were at work prior to the equatorial cleavage in
segregating the egg-contents, continue to function after that cleav-
age-division in both micromeres and macromeres. The result is that
in the micromeres, the albumen droplets which have become
included in these cells are concentrated at their most vegetative
end; while in the macromeres, what little clear cytoplasm there is,
is situated at their most animal end, the fat and albumen drops
being still further concentrated vegetatively.
At the next (fourth) cleavage the droplets in the micromeres find
themselves in the primary trochoblasts (i <2^ to i d'^), while the
clear cytoplasm becomes included in the apical cells (1 a^ to 1 d^).
At the same time, the clear cytoplasm in the macromeres becomes
incorporated in the micromeres of the second quartet (2 a to 2 d),
while the fat drops, etc. remain in the macromeres (2 ^ to 2 D).
In general, therefore, it appears that those regions of the cyto-
plasm which are characterised by an alkaline reaction give rise to
the ectoderm, while those regions with an acid reaction become
endoderm. Facts such as these throw an interesting light on the
inhibition of ectoderm in lithium-induced exogastrulae of Echinoids
(see p. 336).
The free-swimming trochophore larva provides evidence of a
ventro-dorsal gradient, for the ventrally-situated first (ectodermal)
and second (mesodermal) somatoblasts are particularly alkaline,
while, in the endoderm, a similar gradient from ventral to dorsal
side may be observed.
We do not yet know whether any of the various egg-contents
which we have seen are distributed to various blastomeres during
APPENDIX 497
the cleavage of Nereis may be regarded as organ-forming substances
rather than raw materials (see p. 217): further experiments, in-
volving isolation of blastomeres and centrifugalisation, combined
with study of /)H indicators, will have to decide on this point. As
an example, however, of the stratification of substances (perhaps,
of potencies, see p. 102) resulting from the action of an electrical
gradient, these observations and experiments are of great value.
4. Organiser-properties in living and dead tissues.
In amplification of the statement on p. 153 that certain tissues
which possess no capacity to act as organisers when alive may
show this capacity when they are killed, we may refer to further
recent experiments by Holtfreter (N aturwissenschafteyi^ xxi, 1933,
p. 766). He has found that the property to induce the formation
of a secondary embryo or parts of it in an amphibian gastrula
are possessed by the following : all parts of uncleaved amphibian
eggs that have been boiled to a state of hardness ; all parts of an
amphibian gastrula that have been preserved for six months in
70 per cent, alcohol, treated with xylol, embedded in paraffin and
brought back to water; boiled pieces of muscle of the Annelid
Enchytraea and of the molluscs Planorbis and Limnea\ heat-
coagulated cell-free extracts of the crustacean DapJmia and of the
pupae of moths ; pieces of all organs so far tested of the stickle-
back, fresh or boiled; living pieces of larval amphibian liver,
brain and retina, and of adult liver, ovary, and heart ; living pieces
of liver, kidney, testis and other organs of lizards, birds, and mice ;
coagulated bird embryo-extract ; extract of killed calf's liver, and
boiled pieces of several mammalian organs; pieces of liver, brain,
kidney, thyroid and tongue of a fresh human corpse.
No vegetable material was found to possess organising pro-
perties, but in the animal kingdom the chemical substance which
forms the basis of the organising action is clearly widespread and
may probably be regarded as universal. The fact that it is absent
(in the living state) from all regions of the vertebrate embryo
except the organiser, while it is present in a variety of organs in
the adult, is noteworthy and is probably to be regarded as an
adaptation to a specialised mode of development in which organ-
iser action by contact (p. 310) is employed.
HEE 32
498 APPENDIX
5. Development of the amphibian mouth (see also p. 179).
The recent work of Stroer (Arch. Efitwmech. cxxx, 1933, p. 131),
on mouth-development in Amblystoma, may be referred to as an
excellent example of detailed experimental analysis. He finds, by
means of grafting experiments, that the mouth-region is composed
of both ectodermal and endodermal portions, of which the ecto-
dermal alone is capable of forming teeth. The presumptive mouth-
ectoderm is dependent on underlying endoderm for the realisation
of its potencies : without this, it develops into epidermis. Ectoderm
from the ventral region of the abdomen grafted in place of the
presumptive mouth-ectoderm does not react with the underlying
endoderm to produce a mouth (contradicting Adams; see p. 179).
On the other hand, pieces of the mouth-inducing region of the
endoderm (anterior wall of fore-gut) introduced into the blastula
may produce a mouth in interaction with the ectoderm of the heart
region. Thus we do not know the limits, either in space or time, of
the ectoderm-field capable of reacting to form mouth-epithelium.
When presumptive mouth-endoderm grafted into the blastula does
not succeed in inducing the ectodermic portions of a mouth, it
develops into endodermic portions only (portions of buccal cavity,
pharynx, oesophagus, with taste-buds, but without teeth).
An interesting point is that if only a small piece of ventral ecto-
derm is grafted into the presumptive mouth-region, it is caught up
in the invagination process carried out by the remaining presump-
tive mouth-ectoderm and apparently *' infected" with its qualities,
for it then differentiates into true mouth-epithelium.
Another point is that the determination of the presumptive
mouth-ectoderm to produce teeth takes place rather earlier than
its determination to become differentiated mouth-epithelium.
The only region capable of inducing mouth-formation is the
anterior wall of the fore-gut. This was proved by implanting
pieces of various regions of the developing egg into blastulae.
Pieces of anterior neural plate or neural fold have no mouth-
inducing capacity. Neither neural tube nor neural crest material
is necessary for tooth-formation, so that the mesodermal portion of
the teeth must be derived not from mesectoderm but mesendoderm.
Taste-buds are produced by pure endodermal grafts (see p. 487).
INDEX
Bold type denotes pages on which figures zvill be found.
abdomen, Brachyura, growth, 367, 422
acceleration, differential, Macropodus,
344, 345
Acipenser, auditory capsule, 175 f.
activating centre, Platycnemis, 170
activation of egg, 15
activity-gradient, 67 f., 81, 276 f.
adrenal, mosaic development, 199
adult, 354
Aeolosoma, differential susceptibility,
275, 277
age, fibroblasts, 210
Ambly stoma {see also axolotl), auditory
capsule, 175
balancer-field, 177, 236
differential susceptibility, 378, 379
ear-field, 233
exogastrulation, 481
eye and lens growth, 422, 424
fore-limb, 357, 358
fore-limb field, 222, 223, 224
gill-field, 233
limb growth-rates, 421
limb-innervation, 389, 390
neural crest, 393
neural fold field, 243
neural tube, 376
neuron-differentiation, 381, 384,
386
organiser, heteroplastic grafts, 142
sensory load, 365
shoulder-girdle, 281
temperature-gradient, 342
visceral cartilage, 394
Amiurus, taste-buds, 174, 431
Ammonites, shell-growth, 369
amnion, Echinoderms, 180
Amniotes, embryonic membranes. 327
sex-differentiation, 255 f.
amoebocytes, 211
Amphibia {see a/50 Anura, Urodela), 13
blastomeres isolated, 53, 89 f., 98
gradient-system, 437
larval hybrids, 405
limb-region, 304
metamorphosis, 427
organ-rudiments, 199
organiser, ii, 12, 49, 50 f., 51, 52,
80 f., 89, 134 f., 149, 15^, 3io>
3H f., 327, 490 f.
rudiments explanted, 203
sex-differentiation, 255, 257, 260
Amphioxus , asymmetry, 70, 79 80
blastomeres, 100
double monsters, 123, 328
gastrulation, 16
neural tube, 375
partial larvae, 123
polarity, 63
prelocalisation, 119
regulation, 126
Afiableps, asymmetry, 70
" aneurogenic " limb-bud, 378
"anidian" blastoderms, 367
Annelids, axial gradients, 278
differential susceptibility, 332
double monsters, 172, 330
gradient, 309, 320, 496
operculum, 71
organ-forming substances, 116
prelocalisation, 119, 495
regeneration, 281
tail-region, dependence, 283
antagonism, sex-differentiation, 260
antenna, Copepod, growth, 367
regeneration, 360, 361
ants, chemo-differentiation, 127
Anura, 13, 91, 427
sex-differentiation, 255
Anuran organiser in Urodele host, 141
anus, 31
apical organ, Dentalium, no, III, 112
apical point, 330
apical region, 288, 292
autonomy of, 283
inductive capacity, 286
Arbacia, bilateral symmetry, 69
centrifugalisation, 217, 218
cleavage, 83
differential susceptibility, 335
polarity, 66
archenteron, 16
Arenicola, differential susceptibility,
334
32-2
500
INDEX
arm-field, 229
Triton, 362, 363
armadillo, twinning, 329
Arthropods, gradients, 320
limb-regeneration, 303
regeneration, 360
Ascaris, chromatin-diminution, 398,
399, 400
cleavage, 84
giant embryo, 10 1
Ascidians, blastomeres isolated, 96,
99
cleavage, 83
differential susceptibility, 294
egg, viscosity, 108
germ-layers, 153
gradient-system, 437
organ-forming substances, 118, 216
polarity, 67
precocious chemo-differentiation,
123
prelocalisation, 118, 125
regeneration, 198
Ascidiella, partial larvae, 124
Asterina, asymmetry, 81
dorso-ventral axis, 69
asymmetry, 70 f.
Amphioxus , 70, 79, 80
gradient, 77, 81, 265, 361
limb, 358
Lifnnea, heredity of asymmetry, 411
normal, 79
ovaries, 264
reduplicated limbs, 224
atavism, 372
atropine, effect of, 344
auditory capsule, 175
autonomic nervous system, and limb-
regeneration, 420
autonomy of apical region, 283
axial gradients, 278, 331, 372, 378, 379
axial structures, 135
axiation, 60, 65
axis, egg, 13
axolotl (see also A7nblystoma) , adult
characters, 403
balancer, 176
chimaeras, 405, 406
cleavage, 132
dorsal fin, 423, 424
hind-limb field, 230
limb-differentiation, 419
regeneration-buds, 405, 406
axons, 378
Bahnung, 136
balance, sense of, 209
balancer, 31, 177, 192
mosaic development, 199
Triton, axolotl, 176
balancer-field, 236
balancer-induction, Ajiihlystoma, 236
Rana, 236
Bateson's rule, 224
Beroe, cleavage, 105
organ-forming substances, 108, 117
viscosity, 126
biaxial regeneration, 296, 299
Bidder's organ, 257, 258, 260
bilateral symmetry, 15, 36 f., 53, 56,
67 f., 79, 119, 126, 139
bilaterality, Atnphioxus,' 80
frog's egg, 14
biological integration, 57
biological order, 1 1
birds, double monsters, 329
extra-embryonic blastoderm, 327
feather-growth, 369
gonad-asymmetry, 362
organiser, 159 f., 162
plasticity, 100
polarity, 63
bladder, differentiation, 434
blastocoel, 15, 17
blastoderm, 327, 331
bird, 159 f.
experimental production in frog, 40
blastomeres, 15, 33, 39
Amphibia, 98
Amphioxus, 100
Ascidians, 96
Echinoderms, 98
Hydrozoa, 96
isolated, 96 f.
Nemertines, 98
blastopore, 16, 18, 27, 29, 41, 95
blastopore-lip, amphibian, 12
blastopore-rim, amphibian, 17
blastula, 13, 40
blood, mosaic development, 138
blood-islands, chick, 205
blood-vessels, differentiation, 174, 434
blow-fly, explantation of imaginal
discs, 174
Bombinator , blastula regulation, 94
eye, self-differentiation, 48
gill-field, 233
heart, 203
heart-field, 233, 234, 235
INDEX
501
Bombinator, lens-differentiation, 185,
186 f.
organiser, heteroplastic grafts, 142
organiser in Triton, 141
situs inversus, 74
bone, 33
differentiation, 201, 434
Bonellia, determination, 140
brachial plexus, 390
Brachiopod, shell-growth, 368
Brachyura, abdomen-growth, 367, 422
brain, 30, 373, 382, 388
determination, 137
differentiation, thyroid, 396
mammalian, fields, 364
mosaic development, Rana, 246
Bufo, differential inhibition, 349
lens-differentiation, 186 f.
regeneration, 366
calcium-free water, 401, 402
Camponotus, chemo-differentiation,
127
cancer, 54, 213
capillaries, re-differentiation, 211
carcinoma, 179
cartilage, differentiation, 33, 201, 434
visceral, 394, 395
cataphoresis, 496
centrifugalisation, 66, 83, 217 f., 313,
398
Ascaris, 400
Ascidians, 67, 123, 125
frog, 67, 219
Ilyanassa, 122
sea-urchins, 69, 218, 221
Cephalopods, cleavage, 83
differentiation, 350
dorso-ventral axis, 70
mosaic development, 208
Cerebratulus, egg-fragments, 120
polarity, 62
Chaetopterus, centrifugalisation, 219
cleavage, 132
differential susceptibility, 332, 333
differentiation without cleavage, 132
double monsters, 114, 328
egg-fragments, 120
polar lobe, 113, 171, 219
polarity, 62
chela, growth, 367, 422
chela-asyinmetry, Crustacea, 71
chemo-differentiation, 46, 53, 127,
195, 440
chemo-differentiation, of subregions,
243
precocious, 122
progressive, 221, 225
chemotaxis, 263
chick, differentiation of blastoderm-
pieces, 267
eye, self-differentiation, 201, 204
lens-differentiation, 186
limb, mosaic determination, 226,
227
limb subregions, 225, 226, 227
mesonephros, self-differentiation,
207
metanephros, self-differentiation,
200
mosaic development, 195
neuron-differentiation, 381
organ-rudiments, 199
organiser, 159 f., 160, 310
rudiments explanted, 204 f.
self-differentiation, 196
sex-differentiation, 263
situs inversus, 78
temperature-gradient, 343
chimaeras, 363, 405, 406
choanae, 180
chondrification, 177
chorda-mesoderm, 50, 489
chorio-allantoic grafts, 195, 199 f., 267
chorion, 327
choroid fissure, 189, 238
Chorophilus, blastomere isolated, regu-
lation, 93
chromatin-diminution, Ascaris, 399,
400
chromosomes, 398, 399, 400, 403
elimination, Sciara, 401
Cidaris, larval hybrids, 408
cilia-beat, polarity, 236, 487, 488
ciliated band, pluteus, 181
Ciona, mosaic development, 123
prelocalisation, 119
circulatory system, 425
Clavellina, de-differentiation, 287
winter-buds, 64
cleavage, 13,39, 67, 83 f., 126, 328, 344
asymmetry. Molluscs, 71, 72
axolotl, 132
Chaetopterus, 132
Clepsine, 72
Ctenophores, 105
Dentalium, 109, no
Echinoderms, 103, 128, 130
502
NDEX
cleavage, frog, 404
frog, meroblastic, 40
Gastropods, 71
Ilyanassa, 121
insect, 126
Patella, 114
polyspermic, frog, 131
sea-urchin, 128 f., 130
under compression, 43, 84, 85
cleavage-pattern, 83, 114
Clepsine, cleavage, 72
double monsters, 114, 328
germ-bands, 114
pole-plasm, 113 f.
prelocalisation, 119
clinostat, 36
cloaca, 32
Clytia, blastomeres, isolated, 97
coelomic cavity, Amphibia, 28
collar-cells, sponge, 250, 281, 483
colloid structure, 337
compensatory hypertrophy, 431
"competence", 100, 136
competition, nutritive, 425
complex components, 9, 191
composite frog embryos, 195, 407
compression, 83
concentration-gradient, 258
conjunctiva, 140, 178
connective tissue, 33, 179
capsule, 392
differentiation, 432, 433
induction of differentiation by, 179
Copepods, antenna-growth, 367
cornea, differentiation, 178
cortex, gonad, 254
cortical layer, egg, 152
Corymorpha, holdfast resorption, 298
inductive capacity, 290, 291
multiple hydranths, 300
organiser, 164 f., 291
polarity, 63, 65, 280
range of dominance, 288
reconstitution, 284
cranium, 175, 394, 395
Crepidula, polarity and yolk, 67
Crustacea, chelae, 71
cleavage, 84
growth-gradients, 267
Ctenophores, cleavage, 105
mosaic development, 106, 107
Ciicurbita, shape-genes, 416, 417
Cumingia, centrifugalisation, 218
cut surface, regeneration at, 307
Cyclas, polarity, 62
cyclopia, 245, 332, 348, 402
frog, 347
Fundulus, 344, 347
Planarian, 349
Triton, 350, 352
Cynthia (see also Styela), 119
cytoplasm, 46, 66, 170 f., 220, 320,
398
and cleavage, 89 f.
and heredity, 405
and sex-differentiation, 263
susceptibility, 86
cytoplasmic factors, 411
cytoplasmo-nuclear ratio, 132 f.
dedifferentiation, 63, 209, 280, 287,
431
Corymorpha, 65
delayed fertilisation, frog, 96, 262
dendrites, 379
Dendrocoelum, regenerative capacity,
298, 326
Dentalium, cleavage, 108 f., 109, iio
organ-forming substances, 108, 216
partial larvae, ill, 112
polar lobe, 109, no, 120, 216
regulation, 126, 216
dependence, basal regions, 283
dependent-differentiation, 44, 45, 47,
53 f., 173 f., 188, 191, 211, 255,
321,374
Dero, susceptibility-gradient, 277
determination, 46
labile, 49, 50 f., 94, 136, 137, 140,
187, 268, 373, 492
lack of, 273
of organiser, 163
positive and negative, 216
regeneration-buds, 271
stream of, 127, 134, 173
subregions, 225
time-relations, 307
developmental physiology, 11
dichotomy, 299, 328
Diemyctylus, balancer-field, 236
diencephalon, determination, 295
differential, 38 f., 67, 280
axial. Child's theory of, 7, 274 f.
differential acceleration, 301, 344
differential acclimatisation, 301, 335
differential inhibition, 292, 301, 335,
336,425,
differential stimulation, 301
INDEX
503
differential susceptibility, Aeolosoma,
275, 277
Ainblystoyna, 378, 379
Chaetopterus, 332 f., 333
frog's egg, 68
Fiindulns, 344
Macropodus, 345
Perophora, 293
Planarians, 301, 303
sea-urchins, 335, 336
differentiating centre, Platyaiemis, 171
differentiation, 7, 34, 58
dependent, see dependent-differen-
tiation
functional, 13, 34, 418 f., 435
histological, 33, 204, 249, 267, 375
morphological, 30, 33. 83, 249
neuron-, 377, 380
self-, see self-differentiation
and size, 418, 419
typical and atypical, 314, 315, 3^6
diminution, chromatin-, 399, 400
disintegration, 68
dispermy, frog, 37
sea-urchin, 403
dissociation, Cory?norpha, 288
sponge, 65, 250, 282
Dixippus, antenna regeneration, 36 1
dog, bladder, 434
dominance, apical region, 285 f.
limb, 305
physiological, Stetiostomutn, 294,295
dominant region, 308, 316, 325
dorsal crest field, Triton, 363
dorsal fin, axolotl, 423, 424
dorsal half-embryo, Triton, 52, 53, 88,
239, 240
dorsal lip, blastopore, amphibian, i6,
38, 40, 50, 67 f., 134 f-. 321
dorsal meridian, egg, amphibian, 14,
21,37,67
dorsal nerve-root, 33
ganglia, 31, 220, 393
dorso-ventral axis, sea-urchin, 69
"double assurance", 57, 139, 187,492
double monsters, 103, 105
Amphioxus, 80, 328
birds, 329
Chaetopterus, 114, 328
Clepsine, 114, 328
earthworms, 329
Echinoderms, 105, 168, 169, 328
fish, 329
frog, 93 f., 261, 329
double monsters, Fimdulus, 329
Patiria, 328
scorpions, 329
Triton, 90, 155, 156, 157, 158, 329,
351, 352
trout, 329, 331
Tnbifex, 113, 328
dragon-fly, activating centre, 170
chemo-differentiation, 127
cleavage, nuclear equality, 88
differentiating centre, 171
Drosophila, genes, 397
duck, organiser, 159
duplicitas anterior, 156 f., 350, 351,
352
duplicitas cruciata, 95, 113, 155, 15^,
157, 128
duplicitas posterior, 156 f.
dynamic determination, 54, 149, 163,
250, 301
ear, mosaic development, 199
ear-field, 232
polarity, 360
ear-vesicle, and capsule-induction, 175
differentiation, 31
explanted, amphibian, 203
explanted, chick, 205
induction, 147, 192
and limb-induction, 231
self-orientation, 208
earthworms, double monsters, 329
gradients, 274, 275, 277
regeneration, 420
Echinocyamus, skeleton, 181
Echinoderms, asymmetry, 73
bilateral symmetry, 68
blastomeres isolated, 98
centrifugalisation, 218, 221
cleavage, loi f., 130
dependent differentiation, 180 f.
gradients, 320
hydrocoel and amnion, 180
larval hybrids, 405
organiser, 166 f., 169, 323
polarity, 312
regional potencies, loi f.
echinopluteus, 81
Echinus, cytoplasmo-nuclear ratio, 132
larval hybrid, 404
larval skeleton, 181
ectoderm, 18
gradient in, Ambly stoma, 379
egg-axis, 18, 35, 79, 191
504
NDEX
electric current, 60
"emboitement", Bonnet's theory of, 3
embryo, 13
end-organs, 389
endoderm, amphibia, 18, 20, 483
bird, 159 f.
rat, 201
endolymphatic duct, 31
endothelium, heart, 32
entelechies, Driesch's theory of, 9, 352
Entwicklungsmechanik, 9
environment, 8, 59
ependyma cells, 33
epiboly, 16
epidermis, amphibian, differentiation,
140, 482, 489
amphibian, explanted, 203
chick, differentiation, 267
growth-tendency, 41
polarity, 235, 488
post-generation, 92
presumptive, 22, 23, 24, 26
epidermis-field, 235
epigenesis, i f., 58
epiphysis, determination, 245
equator, amphibian egg, 22
equilibrium, gradient-field, 276, 305,
310
growth-, 367
equivalence of nuclei, 85 f., 87, 397
exo-embryo, amphibian, 482 f.
exogastrula, 103, 323, 334, 336, 481 f.
experimental embryology, 11
explantation, amphibian tissues, 55,
139, 194, 202, 203
chick tissues, 196, 200, 201, 204, 206
insect tissues, 173 f.
external factors, 279, 312
extra-embryonic blastoderm, 327
eye, differentiation, Amphibia, 182
differentiation, chick, 204, 267
extirpation, Rana, 388
lens-induction, 51, 54, 183 f., 184,
185, 188, 237
mosaic development, chick, 199
mosaic development, Rana, 246
self-differentiation, Bonibiriator ,■ 46,
self-differentiation, chick, 201
self-differentiation, Rana, 248
size-regulation, 242
eye-field. Amphibia, 244
eye-graft, nerve-attraction, 391
neuron, proliferation, 388
eye-growth and lens-growth, 422, 424
eye-induction. Amphibia, 147
eye-pigmentation, Gammarus, 409,
410
fat, in frog's egg, 220, 320
feather-bud, differentiation, 268
feathers, growth-gradient, 369
fertilisation, 13
fibroblasts, differentiation, 433
growth, 370
metaplasia, chick, 214, 215
metaplasia, Pecten, 212, 213
tissue-culture, 179, 209
field, 221, 274
of direct and indirect action, 311
susceptibility of, to thyroid, 427
field-gradient system, 274 f., 405
Filigrana, head, inductive capacity,
290
fish, double monsters, 329, 331
larval hybrids, 405
organ-rudiments, 201
plasticity, 100
thyroid tumour, 392
flatfish, asymmetry, 70
fore-limb, induction, 193, 231, 363
polarity and laterality, 224, 357,
358
fore-limb field, Amhlystoma, 222, 223,
224
formative stimulus, 134, 193
form-change, 30
form-differentiation, 196
fowls, spurs of, 269
frog, centrifugalisation of egg, 40,
219, 220
cyclopia, 347
cytoplasm of egg, 91
dorsal lip of blastopore, 38, 95, 321
double monsters, 95, 261, 329
fields, susceptibility to thyroid, 190,
427
gdl-field, 233
gonad-differentiation, 257, 262
neural tube, 182
operculum perforation, 180, 428,
429
partial embryos, 92
sex-differentiation, 255 f., 257
tumour-like growth, 96
frog's egg, bilaterality, 14, 37, 38, 67
cleavage, 40, 85, 404
delayed fertilisation, 96, 261, 262
INDEX
505
frog's egg, differential susceptibility,
68, 346, 348
dispermy, 37
inverted, 35, 38, 67, 94
nuclear injuries, 404
"parasitic", 328
polarity, 14, 35, 62, 67
polyspermy, 131
temperature-gradients, 338, 339,
340, 341
Fncus, polarity, 60, 61, 70, 312
function, 34, 431 f.
in heteromorphosis, 361
functional activity, 33 f., 426, 435
functional differentiation, 13, 34, 431
functional period of development, 34,
418 f.
functional response, 436
Fimdulus, cyclopia, 344, 347
differential susceptibility, 344, 347
double monsters, 329
mosaic development, 195
pigment-cells, 174
plasticity, 100
gall-bladder, 32
Gmnmarus , precursor substances, 412
rate-genes, 409, 410
ganglia, spinal, 31, 220, 393
gastral mesoderm, 20
Gastropoda, asymmetry, 71, 72 81,411
polar lobe, 113
shell-growth, 370
gastrula, 13, 40
gastrulation, 16, 17, 19, 26, 40, 95, 330
and temperature-gradients, 342,494
gelation, 122
genes, 7, 397
genetics, 5
genome, 403
Gephyrea, determination, 140
germ-bands, 114, 117
germ-layers, 18, 46, 140
early plasticity, 47
germ-plasm, Weismann's theory of, 4
germinal epithelium, 254
gill-field, 233, 360
gills, 31,46
dependent differentiation, 45
functional differentiation, 435
heteroplastic grafts, 45, 142, 406
mosaic development, 199
and opercular perforation, 180, 429
glomerulus, mosaic development, 199
glycogen, 154, 190
glycolysis, 214
gonad-field, 254, 264
gonads, 254, 263
asymmetry, 362
differentiation, 257, 262
gourds, shape-genes, 416, 417
gradient, 288, 318
activity-, 42, 67 f., 81, 276, 313
asymmetry-, 77
double, Annelids, 309
electrical, 61, 63, 497
in cell-size, 39, 67, 339
in inductive capacity, Corymorpha,
165,291
in lens-forming potency, 238
in potency of differentiation, 160, 268
in regenerative capacity, Planarians,
298
morphogenetic, 292
oxidation-, 35, 274
physiological, 36, 41
temperature-, 39, 137, 138, 33^,
339, 340, 341, 342, 343, 494
gradient-field, 39, 139, 231, 274 f.,
305, 354, 369
primary, 311
secondary, 310
gravity, 35
grey crescent, 14, 37, 38, 67, 320
growth, 13, 26, 33, 58, 418
growth-capacities, 205
growth-coefficient, 225, 366
growth-equilibrium, 366, 421
growth-gradient, 366 f., 368, 414, 421
growth-partition coefficient, 206, 421
growth-potency, 369
growth-profile, 367, 371
growth-rates, fibroblasts, 210
limb, Urodele, 225, 421
growth-tendency, of epidermis, 41, 42
gut, amphibian, explanted, 203
asymmetry, 73, 74
differentiation, 249, 484
presumptive, 22
gut-roof, 13, 20 f., 135, 155
and situs inversus, 74
hair-follicles, mosaic development, 201
Haliclystus, inhibition, 292
haploid graft, 364
haploidy and cell-size, 133
Harenactis, multiple regeneration, 300
harmonic equipotential system, 325, 3 53
506 INDEX
head, Planarian, 165, 166, 167, 372
head-organiser, 144, 146, 147, 492
heart, amphibian, 20, 32
asymmetry, 73 f., 74, 75, 234
chick, and Hver-differentiation, 179
function, 432
morphological differentiation, 249
self-differentiation, Bombinator, 203
situs inversus, 75, 76 f., 234
heart-field, 233, 234, 235
polarity, 360
heart-rudiment, Urodele, 78
hereditary factors, 15, 397 f-
heredity, 8, 59, 397, 43^
hermaphroditism, 259
hermit-crab, growth-profile, 367
Hertwig's rule, 83
heterogenetic induction, 145
heterogony, 366
heteromorphosis, serial, 360, 361
heteroplastic grafts, 42, 161, 164 f.
hind-limb field, Amblystojna, 224, 230
histo-differentiation, see histological
differentiation
histological differentiation, 33, 204,
249, 267, 375, 440
histolysis, operculum, frog, 428, 429
holoentoblastula, 334
homoiogenetic induction, 145 f., 148,
161, 193, 396
"homunculus", 2
hormones, 173, 177, 190, 396, 425, 437
sex-, 270
susceptibility and growth-rate, 370,
428
horns, sheep, growth of, 370
host-tissues, influence of, 147, 150
house-fly, precocious chemo-differen-
tiation, 127
Hydra, organiser, 164, 165
polarity, 280
hydrocoel, 73, 81, 180
hydrodynamics, blood-vessels, 434
Hydroids, differential inhibition, 292,
425
isolated blastomeres, 96
polarity, 63, 64
reconstitution, 65, 66, 281
Hyla, lens-differentiation, 186
Hymenoptera, polyembryony, 328
hypochordal rod, 21, 183
hypophysis, 31, 33, i79
mosaic development, 198
hypostome. Hydra, 164
Ilyanassa, centrifugalisation, 122
cleavage, I2i
polar lobe, 113, 120, 121, 122
imaginal discs, explanted, 174
Inachiis, chela-growth, 370
individuation-field, 162, 310, 319
induction, heterogenetic, 145
homoiogenetic, 145, 148,161,193,396
secondary, 57
inductive capacity, Corymorpha, 290,
291
Planarians, 286
Sabella, 288, 289
" infective "organiser-capacities, 151 f. ;
mouth-region, 498
infundibulum, 33, 179, 245
inhibition, 292, 297, 301, 325
of gonad-cortex, 256
Harenactis , 300
of lens-formation, 238
of sex-differentiation, 258
innervation, 389, 390, 393, 486
insect, activating centre, 170
cleavage, 126
cleavage, nuclear equivalence, 88
differentiating centre, 171
differentiation, 252
dorso-ventral axis, 69
holometabolous, differentiation, 437
rudiments, explanted, 173
integration, 57
intermediate cell-mass, 32
interplantation, 138, 209, 314, 315, 316
intersexuality, mammals, gonad-asym-
metry, 362
intersexuality, Lymantria, 172
intestine, 32, 203, 485
intra vitam staining, 21, 394, 495
invagination, 16, 17, 19, 41, 95
inversion of frog's egg, 38, 94, 152
invertebrates, organisers, 64 f.
isogony, 367
isolated blastomeres, 96 f.
isolation, physiological, 296, 325
jelly-fish, differential inhibition, 425
joint, limb-, differentiation, 228
keratinisation, 209
kidney, amphibian, explanted, 203
chick, explanted, 200, 207
compensatory hypertrophy, 431
kidney-epithelium, redifferentiation,
179, 211
NDEX
507
kidney-tubules, 20, 32, 152, 193
killed organiser, 153, 163, 497
knitting, 436
labile determination, 49, 50 t., 94, 136,
137, 140, 187, 268, 373, 492
Lamellibranch, centrifugalisation, 218
cleavage, 108
larva, 13
larval hybrids, 404, 408
late fertilisation, 95, 261
lateral half-embrvo, Triton, 52, 53,
86 f.
lateral line, 31, 56, 355, 356, 357, 431
lateral plate mesoderm, 20, 22, 28
Lebistes, melanophore-multiplication,
426
leech, cleavage, 72, 113
double monsters, 114
prelocalisation, 119
left-right axis, 79
leg-field, Ambly stoma, 224, 230
Triton, 362
lens, 30
connective tissue capsule, 392
dependent differentiation, 55, 183 f.
formation from optic cup, 187, 188,
237
grafts, 145
self-differentiation, 184, 185
lens-fibres, 188
lens-field, 187, 189, 238
lens-growth, and eye-growth, 422, 424
lens-regeneration, 187, 237
Lepidoptera, self-differentiation, 206
light, 38, 60
limb, growth-coefficients, 206, 421
induction of, 177, 193, 231, 363
mosaic development, 198
regeneration. Amphibia, 198, 271 f.,
273, 304, 306, 307, 420
regeneration. Arthropods, 303, 361
sheep, growth, 367, 414
limb-bud, chick, explanted, 205
chick, subregions, 225, 226, 227, 228
differentiation and size, 418, 419
grafts, 364
limb-determination, Amhlystoma, 56,
224
limb-field, 222 f., 276, 299, 306, 307,
311
limb-forming potencies, 231
limb-innervation, Ambly stoyna, 389,
390
limb-joint, self-differentiation, 228
limb-rotation, 232
limb-rudiments, insect, explanted, 174
Liitmcea, asymmetry, 81, 411
limpet, cleavage, 114
lithium, effects of, 181, 323, 336, 337
liver, 32
differentiation, chick, 179
liver-extract, fibroblast-metaplasia, 213
liver-rudiment, amphibian, explanted,
203
lizard, regeneration, 365
tail-induction, 362
tail-scale regeneration, 372
localisation, 271
logarithmic spiral, 369
lower layer, blastoderm, bird, 159 f.
Lumbriculiis, regeneration, 308
susceptibility-gradient, 277
Lymafitria, intersexes, 172
rate-genes, 409
Lytechiniis, cleavage, 85
larval hybrids, 408
polarity, 313
regulation, loi
macromeres, Beroe, 105
Deyitalium, 109
sea-urchin, 102, 104, 323
macrophages, 213, 214, 215
Macropodus, differential acceleration,
344, 345
Maia, chela-growth, 370
mammals, gonad-differentiation, 264
intersexual, gonad-asymmetry, 362
organ-rudiments, chorio-allantoic
grafts, 201
plasticity, 100
trophoblast, 327
twinning, 329
mammary gland, 179
man, thyroid requirements, 425
twinning, 329
Mantis, antenna-regeneration, 361
"marginal zone", 41 f.
" mass-movements ", 43 , 25 1 , 481, 482
"mechanistic" explanations, 10
medulla, gonad, 254
medulla oblongata, 385, 386
megacephaly, 334, 338, 339, 345,
346
melanophore-expansion, 425
melanophore-multiplication, 426
membrane, egg-, 13
5o8
NDEX
Mendelian "characters", 413
Mendelian factors, 72, 397 f.
meroblastic cleavage, 39, 40
mesenchyme, pluteus, 102, 181
mesocardium, 32
mesoderm, iS, 19, 20, 50, 140, 484
Ambly stoma, gradient in, 379
and Hmb-field, 223 f.
mesomeres, sea-urchin, 102, 104
mesonephros, self-differentiation, 199,
203, 205, 207
and sex-differentiation, 255
metameric segmentation, 20, 193, 393
metamorphosis, 54, 325, 427
metanephros, self-differentiation, 200,
205
metaplasia, 211, 212, 214, 215, 259
metastases, 96
microcephaly, 332, 338, 339
micromeres, Beroe, 105
Dentaliimi, 109
Nereis, 496
sea-urchin, 102, 104, 321, 322
mid-brain, connective tissue capsule,
392
mitochondria, 124, 217
Molluscs, asymmetry, 71, 72, 81, 108,
411
double monsters, 172
organ-forming substances, 116, 216
shell-growth, 368
monocytes, 213
monorhiny, 237, 332, 348
monotreme, gonad-asymmetry, 362
ovary, 264
monsters, developmental, 7
{see also double monsters)
morphallaxis, 316
morphogenetic field, 274 f., 292
morphogenetic substances, 108, 123,
216
morphological differentiation, 30, 33,
83, 204, 249
mosaic development {see also self-
differentiation), Urodeles, 56, 194,
484, 485
chick, limb, 226, 227
Ctenophores, 106, 107
Rana, brain and eyes, 246
Styela, 99, 125
mosaic stage, 57, 124, 194 f., 419
mosaic-eggs, 70, 98, 105, 108, 215
moths, cytoplasmic factors, 405
motor load, 383, 384
moulting-hormones, insect, 174
mouth, 31, 179, 498
differential inhibition, 348, 349
multiple potentiality, 325, 326, 363
Miisca, precocious chemo-differentia-
tion, 127
muscles, atrophying, and opercular
histolysis, 429
mutual influence, eye- and lens-
growth, 422, 424
myotome, 20, 28, 32, 220, 193, 393
inductive capacity, 145, 193, 374
innervation, 381
Myzostoma, polar lobe, 113
prelocalisation, 119
narcotics, 287, 301
narcotised organiser, 153 f.
nasal capsule, 175
nasal pit, 31, 180, 203
nerve-attraction, 391
neuron-proliferation, 388
Nematodes, chromatin-diminution,
399, 400
cleavage, 83
isolated blastomeres, loi
mosaic development, 10 1
Nemertines, egg-fragments, 120
isolated blastomeres, 98
prelocalisation, 120
regeneration, 211
neo-Mendelism, 4
neoteny, 403
nephrotome, 32, 152
Nereis, cataphoresis, 496
differentiation without cleavage, 132
nerve-fibres, connexion with end-
organ, 389
differentiation, 377, 380
nerves, trophic effects of, 174, 230,
363,387,430
nervous system and regeneration, 419,
420
neural crest, 31, 33, 193, 199, 393,
394, 395, 487
neural fold field, 239, 244
neural folds, 22, 25, 27, 30, 155, 249,
373
induction of, II, 49, 134 f., 141, 160
labile determination of, 50, 56, 136,
137, 138, 373, 492
presumptive, 22, 23, 24
regulation, 240, 241
neural plate, see neural folds
INDEX
509
neural tube, differentiation, 374
explanted, 203, 392
inductive capacity of, 147, 148, 193
neurenteric canal, 27
neurobiotaxis, 389
neuroblasts, 33, 377, 380
neuron, differentiation, 377, 380, 384,
386
proliferation, 383 f.
neurula, 13, 27
newt, see Triton
non-specific attraction, 392
nose-field, 237
nostril, 31, 180
notochord, 18, 21, 28, 50, 161
differentiation, 183, 220, 314, 316,
317,484
explanted, 202, 203
inductive capacity of, 145, 374
presumptive, 22, 23, 24
nuclear division, 43, 84, 85, 87, 397
nucleic acid, 132
nucleus, 46, 119, 404 f.
Obelia, polarity, 63, 64
Oligochaetes, cleavage, 113
double monsters, 113, 330
gradients, 274, 275, 277
organiser, 166
ontogeny, 8
operculum, Anura, 180, 233, 428
Anura, polarity, 360
asymmetry, Polychaetes, 71
ophiopluteus, 81
optic cup, 30, 374
determination, 244
lens-inducingcapacity,5l,54, 183 f.,
184, 185, 188, 237
optic lobes, determination, 245, 246
optic stalk, determination, 244, 249
organ-fields, 193, 221, 228
organ-forming substances, 108, 114,
117 f., 2i6, 311
" organic points ", Bonnet's theory of, 6
Organisationsfeld, 274
organiser, amphibian, ii, 12, 49, 50 f.,
51, 52, 80 f., 89, 134 f., 149, 158,
310, 317 f., 327, 488
amphibian, extract, 154
amphibian, heteroplastic grafts, 141
amphibian, " infection" by, 151, 152
amphibian, regional potencies, 144,
146, 490
bird, 159 f., 160, 162
organiser, bird, determination, 163
bird, killed, 163
Corymorplui, 164, 290, 291
Hydra, 164, 165
Planaria, 166, 167
Sabella, 165, 289
sea-urchin, 166, 169, 321, 322, 325
organising substance, 154, 497
organism, unification, 424
orientation, of organiser, 149
of primitive streak, 161
ovarian hormones, 177
ovaries, bird, asymmetry, 264, 266
bird, mosaic development, 205
and egg-polarity, 35, 63
over-ripeness, of egg, 95, 96, 261
oxygen, 60, 280, 288
owl, skull, asymmetry, 70
Palaemon, chela-growth, 370, 371
eye-regeneration, 360, 361
palato-quadrate, 178
pancreas, explanted, 203
pangenesis, Darwin's theory of, 4
Papilio, mosaic development, 206
parabiotic twins, 256, 257, 383, 385
Paracentrotus, 313
organiser, 166 f., 169, 322
regional potencies, 10 1 f., 102, 104
pars nervosa, 179
parthenogenesis, artificial, 37, 133
partition-coefficient (growth), 366
Patella, isolated blastomeres, 115
Patiria, cleavage, 83 f.
double monsters, 328
polarity, 313
regulation, loi
Pecteii, metaplasia, 212, 213
Pelmatohydra, heteroplastic grafts, 164
Pelobates, eye-regulation, 244
Pennaria, reconstitution, 66
pericardial cavity, 32
peristomial mesoderm, 20
permeability, 108
Perophora, differential inhibition, 293,
425
Petromyzon, neural tube, 182
Phagocata, regenerative capacity, 298
Phalhisia, partial larvae, 123
pharynx, Planarian, 165, 167
Pheretima, gradients, 274
Phialidiuni, polarity, 62
phosphatase activity, 205
phylogeny, 8
510
INDEX
physiological genetics, 5, 409 f.
physiological isolation, 296, 325
physiological mosaic, 193
piano-playing, 436
pigeons, thyroid-activity, 436
pigment, gradients, 381
pigment-cells, function and multi-
plication, 426
Pinnotheres, growth-gradient, 308
pituitary, 33, 179, 198
pituitary hormone, 198, 425
placode, 28, 31, 199
Planaria, apical region, inductive
capacity, 286
biaxial regeneration, 285, 296
cyclopia, 303, 349
differential inhibition, 301
differential susceptibility, 303
growth-gradient, 367, 372, 422
heteroplastic grafts, 165
multiple heads, 299, 326
narcotics, 285, 287, 301
organiser, 165, 166, 167, 286
range of dominance, 286, 287
reduction, 368
regeneration, 271, 273, 279, 308
tail, 283, 305
Planorhis, shell-form, 81
plasticity, 44, 45, 46 f., 47, 94
regeneration-buds, 272, 273
Platycnemis, activating centre, 170, 171
chemo-differentiation, 127
inside-out, 252, 253
nuclear equality, 88
regulation, 127, 128
Pleiirodeles, labile determination, 136,
137
neural crest, 393
organiser, heteroplastic grafts, 142
pluteus, arms, 174
skeleton, 174, 181, 405
polar body, 15, 119
polar lobe, Chaetopteriis, 113, 171, 219
Dentaliio7i, 109, iio, 117, 120, 216
Ilyanassa, 113, 121, 122
Myzostoma, 113
polarity, 35 f., 56, 60 f., 70, 83, 123,
279, 312, 331
blastoderm, bird, 161 f.
epidermis, cilia-beat, 235, 488
frog's egg, 14
Fuciis, 61
Hydra, 164, 280
limb-field, 224, 358
polarity, organ-fields, 229, 233, 243?
359
of organiser, 149
sea-urchin, 60, 105, 218, 323, 337
polarity-gradient, 64
pole-plasms, Clepsine, 114, 117, 119
Tubifex, 113, 171, 216
Polychaetes, organisers, 165 f.
polar lobes, 113 f.
Polydactyly, 262
polyembryony, 328 f.
polyspermy, frog's egg, 131
Polyzoa, polyembryony, 328
position, and cell-fate, 281
positive and negative determination,
216
post-generation, 91
post-trochal region, Dejitaliiim, no,
III, 112
potassium, effects of, 331
potential difference, 308
precursor substances, 412
preformation, 2 f., 58
organ-forming substances, 117
prefunctional period, 34, 418 f.
prelocalisation, organ-forming sub-
stances, 117 f,
presumptive regions, 21 f., 23, 24, 43
primitive streak, 159 f.
primordial germ-cells, 254, 263
proctodaeum, 31
pronephric duct, 32
pronephros, induction, 50, 135, 152,192
inductive capacity, 145, 193
proportions, change of, 421
prototroch. Patella, 116
Psammechiniis, bilateral symmetry, 69
Psoitis, dorso-ventral axis, 69
psychological changes, at metamor-
phosis, 396
rabbit, tissues explanted, 201
Rana {see also frog), balancer- induction
in Triton, 177
blood, mosaic development, 198
ear-vesicle and capsule-induction,
175
Rana cateshiana, lens-differentiation,
189
Ratia escidenta, egg-axis, 14
eye-determination, 245, 246
lens-differentiation, 184, 186
lens-fibres, 190
sucker formed on Triton host, 143
INDEX
511
Rana fusca {temporaria) , eye-extirpa-
tion, 388
lens-differentiation, 183 f.
nervous system and leg-develop-
ment, 430
nose-field, 237
Rana jiigromaculata, ear-field, 232
Ra7ia palustris, neuron-differentiation,
377
and sylvatica, composite embryo,
407
and sylvatica, lateral line, 355, 356,
357
visceral cartilages, 395
range of dominance, 285, 287
rat, tissues explanted, 201, 268
rate-genes, 409, 410
raw materials, 219 f., 311, 320
realisation-factor, 280
recapitulation, Haeckel's theory of, 8
reconstitution, Corymorpha, 284
Pennaria, 66
scale of, 288
Sycon, 65, 282
Tubidaria, 287
redifferentiation, 179, 211, 431
Clavellina, 287
Corymorpha, 65
reduplication, limbs, 224, 327, 358,
363
regeneration, 58, 63, 193, 195, 197,
271 f., 304, 306, 307, 350, 360,
361, 406, 419
lack of, after field-extirpation, 365
lack of, during mosaic phase, 58, 195
lens, 187, 237
limbs, Urodele, 271, 304, 306, 307
Limibriculus, 308
Nemertines, 211
Planarians, 271, 286, 298, 308, 326
regional differentiation, organiser,
amphibian, 147, 490 f.
regional effects, 411
regional factors, 177, 191, 192
regional field, 229, 354, 362
regional potencies, sea-urchin, loi,
104
regulation, 58, 93, 98, 126 f., 195, 215,
350, 373, 419
Ascidians, 124
Ctenophores, 107
eye-size, Triton, 242
heart, 235
neural fold field, 239, 240, 241, 244
regulation, Platycnemis, 127, 128
Triton, eyes, 242
Triton, neural folds, 239, 240, 241,
244
regulation-eggs, 70, 98 f.
reptiles, extra-embryonic blastoderm,
327
resonance theory, 391
resorption, 294, 299, 325
restriction of potencies, 260
rete tissue, 255
retina, 33, 204, 244, 248, 249
Rohon-Beard cells, 382
Sabella, organiser, 165, 288, 289
Salamandra, gills, functional differen-
tiation, 435, 436
metamorphosis, the psychological
changes, 396
pigment-cells, 426
regeneration, 366
salivary gland, 179
Salmacina, organiser, 290
"saturation", nerve-attraction, 392
organiser-action, 310
scale of reconstitution, 287, 288
scales, Lymantria, 172
tail-regeneration, lizards, 372
Scaphopoda, shell-growth, 369
Sciara, chromosome-elimination, 401
scorpions, double monsters, 329
sea-urchins (see also Arbacia, Cid-
aris, Echinocyamus, Echinoderms,
Echinus, Lytechiniis, Paracejitrotus,
pluteus, Psammechinus, Sphaere^
chinas, Tripneustes) , bilateral sym-
metry, 69
calcium-free water, 401, 402
centrifugalisation, 217, 218
cleavage, 83, 84, 128 f., 130
cytoplasmo-nuclear ratio, 132
dependent differentiation, 174
differential inhibition, 335, 336
differential susceptibility, 334, 335
differentiation, 252
double monsters, 103, 105, 168
larval hybrids, 404, 408
organiser, 166 f., 169, 321, 322, 325
partial larvae, 102
polarity, 60, 313
regional potencies, 102, 104
regulation, 97, 102
seakale root, regeneration, 297
secondary induction, 57
5^2
INDEX
Selachians, cleavage, 39
endolymphatic duct, 31
extra-embryonic blastoderm, 327
self-diflferentiation {see also mosaic
development), 53 f., 70, 188, 196
eye, Bombinator, 48
eye, chick, 201, 204
eye, Rana, 248
limb-joint, 228
mesonephros, chick, 207
metanephros, chick, 200
wing-cases, silkworm, 206
sensory load, 383, 385
sex-cords, 263
sex-differentiation, 254 f., 409
sex-hormones, 259
sex-induction, 256
sex-reversal, 409
shape-genes, gourds, 416, 417
sheath, nerve-, 393
sheep, growth-gradient, 367, 370, 414
shell-growth, 368
shoulder-girdle, Amblystoma, 231
silkworm, maternal inheritance, 412
self-differentiation, 206
simple components, 9
sinews, differentiation, 174
situs inversus, 73, 74, 75, 234
size, and limb-differentiation, 418, 419
skeleton, chick, explanted, 205
pluteus, 174, 181, 405
regeneration, Triton, 304
vertebrate, adjustment, 175
vertebrate, determination, 434
skin, regeneration, 304
snails, cytoplasmic factors, 405
sodium thiocyanide, effects of, 337
somatoblasts, 113, 172, 496
somites, 20 f.
space-lattice, 71
specific attraction (nerve), 393
sperm-entry, 36, 67, 320
Sphaer echinus, centrifugalisation, 221
larval hybrids, 405
Sphodrojnantis, antenna-regeneration,
361
spinal cord, 30, 373, 382, 388
spiral cleavage, 72, 108, 117
spleen, mosaic development, 199
sponge, reconstitution, 65, 250, 281,
282
spurs, fowls, 269
stag-beetle, growth-gradient, 422
growth-profile, 367
starfish, see Asterina, Patiria
starvation, Planarian, 368
Stenostomum, differential inhibition,
294,. 295
Sternaspis, polarity, 62
stomodaeum, amphibian, 31, 179
sea-urchin, 102, 166 f.
stratification, 66, 123, 217
striated muscle, 33
function, 432
Styela, blastomeres isolated, 97, 99
centrifugalisation, 67, 123, 125
mosaic development, 99, 125
organ-forming substances, Il8, 124,
217
partial larvae, 123
polarity, 67
prelocalisation, 119
yolk, 67
Stylaria, head, organiser, 166
sucker, Anuran, heteroplastic grafts,
142, 143
susceptibility (see also differential
susceptibility), 67, 86, 88, 301,
309, 331
Sycon, reconstitution, see sponge,
reconstitution
symmetry, 15, 36, 53, 56, 67, 79, 119,
126, 139
sympathetic nervous system, limb-
regeneration, 198, 420
tail, lateral line on, 355, 356, 357
mosaic development, 197
regeneration, 272, 365, 372
tail-field, 362, 363, 365
tail-formation, Urodele, 28, 29
tail-graft, nerve-attraction, 392
tail-induction, 145, 193, 487, 490
tail-muscles, 28, 30
tail-organiser, 492
tail-region, dependence, 283
tapetum, determination, 244, 249
taste-buds, Ayniurus, 174, 431
endodermal origin, axolotl, 487, 498
teeth, horny, heteroplastic grafts, 142
true, origin of, 487, 498
teleosts, neural tube, 182
thyroid tumours, 392
Telmessus, growth-gradient, 368
temperature, sex-differentiation and,
256, 263
temperature-gradient, 39, 137, 138,
338, 339, 340, 341, 342, 343, 494
NDEX
513
tendons, differentiation, 432
tension, effects of, 432, 433
tetraploidy, 134
thyroid, brain-differentiation and, 396
in late-fertilisation, 262
and metamorphosis, 54, 427
mosaic development, 199
susceptibility of fields to, 190, 427
tumours, fish, 392
thyroid activity, heredity, 436
thyroid hormone, 425
time of onset, gene-action, 408
time-relations, chemo-differentiation,
10, 215
cleavage, 131
gradient-fields, 307
growth, 422
hormones, 270
nervous system, 383
rate-genes, 409
tissue-culture {see also explantation),
204 f., 209 f., 377, 433
toads, differential inhibition, 349
sex-differentiation, 227, 258
Trip?ieustes, centrifugalisation, 221
Triton, Anuran sucker, heteroplastic
graft, 143
balancer, 31, 176, 177
balancer-field, 177, 236
blastomeres isolated, 53, 90
blastula, regulation, 94
chimaeras,' 405, 407
cleavage, nuclear equivalence, 85, 87
compound embryos, 407
cyclopia, 350, 352
dependent differentiation, 44, 45, 46,
47
double monsters, 75, 155, 156, 157,
158,329,351,352
egg, constricted, 52, 85, 87, 89, 239,
240, 350, 351, 352
eggs, fused, 90, 91
eyes, regulation, 242
gastrula, regulation, 89, 239, 240
gills, heteroplastic grafts, 45, 142,
406
heterogenetic induction, 145
homoiogenetic induction, 145, 148,
193
host-tissues, influence of, 147
labile determination, 49, 50, 94, 140,
lens-differentiation, 55, 186, 188
lens-regeneration, 187, 237
Triton, limb-regeneration, 271, 304,
306, 307, 364
mosaic development, 195, 197
neural fold field, 239, 244
organiser, ii, 50, 51, 89, 135 f., 141,
144, 146, 151, 310
regeneration-buds, 153, 271, 272,
273, 364
regional factors, 192, 193
regional fields, 362, 363, 365
regulation, 53, 90, 94, 216, 239, 240,
241, 242
situs inversus, 75
twinning, partial, 329, 351
typical and atypical differentiation,
139, 314, 315, 316, 317
Trituriis, sex-differentiation, 258
temperature-gradient, 138, 494
trochoblasts, Choetopteriis, 132
Nereis, 132, 496
Patella, 115
trophic effect, nerves, 174, 203, 363,
387, 430_
trophic inhibition, 292, 293, 295
trophoblast, 327
trout, double monsters, 75, 329, 331
egg, over-ripeness, 261
situs inversus, 75
vertebrae, order of development,
379
trunk-organiser, 144, 146, 147, 490, 492
trypaflavine, and sperm, 404
Tubifex, blastomeres isolated, 113
double monsters, 113, 328
pole-plasms, 113, 171, 216
regulation, 113, 216
Tiihularia, apical region, autonomy,
283
morphallaxis, 284, 287, 316
scale of reconstitution, 286, 287
tumours, 54, 96, 213, 261, 392
organising effect, 153
turbinate spiral, 370
twinning, 245, 327 f.
tympanic membrane, 178
ultra-violet radiation, 88, loi, 108,
113, 263, 404
tinification of organism, 424
upper layer, bird, 159 f.
urea, 183
Urodela (see also Amblystoma, Amphi-
bia, Pleiirodeles, Triton, Triturus),
blastomeres isolated, 90
SH
INDEX
Urodela, sex-differentiation, 257, 285
thyroid and metamorphosis, 190,
427
tissues and Anuran organiser, 141
uterus, 177
vegetative hemisphere, amphibian egg,
22
veins, Lymantria, 172
ventral half-embryo, Triton, 52, 53,
88 f., 89, 240
vertebrates, asymmetry, 73
differentiation, 437
neural plate, 155
skeleton, 175
visceral skeleton, 31, 394, 395
viscosity, 105, 108, 123, 126
vitelline membrane, 43
whales, asymmetry, 70
winter-buds, Clavellina, 64
wrybill-plover, asymmetry, 70
X-rays, 404
xenoplastic grafts, 142
yolk, 13, 15 f.,
220, 320
yolk-nuclei, 40
yolk-plug, 17
yolk-sac, 327
30, 35, 40, 67, 94 f.
Live Music Archive
Librivox Free Audio
Metropolitan Museum
Cleveland Museum of Art
Internet Arcade
Console Living Room
Books to Borrow
Open Library
TV News
Understanding 9/11