(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Children's Library | Biodiversity Heritage Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
See other formats

Full text of "Plant morphogenesis"



g§| ■ " ' - r ' m ' ' , \ ' " ' '' * , '-'J 5 ." 


■ i 


. ; :||^^^^i^^^^4 ^?~* ' -^ ^ ^J^jjsj? "*-< ■«»*' Js^jw'tje^ -<W ~U^ &>«J|k5%> 


?,; /-■■,-rf^:'"^->>c"* ;/ v. ■'■< ^{-;'/^ r ': : JjS&vs^ws^Si**^ "-•: :-.'•'• '1 ■'■'-';.' : -:\ < v : ' . . . : ' , -.'-'-'"■■ ^sSi-jJJio -'' ■""""-' '•"' ' : '- - ' ■ i- ■ ■ . . ' . ,<-.•:. ,; V '.'..;:■"/ 

■' '! J ' i '.' '■'/':■ ' -i;6.' ; >; "'._.,, ! -:. : ; ';-;:. :-"-'■.:.' txg z r'.:'r^'.^:. ', " : '"' ** ' '. r .' '■■ ' : '"' , si^.T' '/ -T ' "\ '. 1 -.'.■','-'. -^..' - . ! ''.'-i '■ ■-'' '• ^ ', : .-'..,' ; '" ."".'".'" ':■'.- .''. ■ -'..'-■ '!'- ,- ' . '.' '■'-'■■ ■'■■ '. ■.".■.'■''.-.'■ 





■ '- : -'< ■'<■ '■■''(■ ' '■x' r ^^'N>'^ ; - : L - " ■"" '. - "■ -^'^ : >vv^i-^B'vi;^£' ^';'->^- : v^'c--: .■- ■' RJfe ;^': -'*- -.''"•; ■', : "-'~r'<'-<r ■ ' '• w&JKw^?^ •' ■ ] --' : - r ^S*£fi^> <S """^v^":'^ ,| ;;'';''-:;'i^!?f' 

w§i IpiSiSiPlSii Ji^SSIiiiSl liiiplffpl iplSiilSli - : 38 \ ■■■. 

SoSi<^?^^^5'^^S^^S5^^^^0^^^^K^x «§ "■ --'}: i ^l :-"*'""■■"'. "'■.;"; ■;■' ;\. ■ ■'■'■ ■■■ J .\-. '•■.'■■ '■vS'.':'- : -'':' v '~; ; ::"''"c':; ' - : ^:'-v' • '" '-v^ :■(■'■' ^y^ : : '■■.'.■"■'," -.;;'-- : V .-V ' ; - ■■'-?.."' '■ -. 

. "»' ?%s§fllllll^^ 



ftiillS ^^-iffe , 

Marine Biological Laboratory Library 

Woods Hole, Massachusetts 

Plant Morphogenesis 


Edmund W. Sinnott, Consulting Editor 

Arnold An Introduction to Paleobotany 

Curtis and Clark An Introduction to Plant Physiology 

Eames Morphology of Vascular Plants 

Eames and MacDaniels An Introduction to Plant Anatomv 

Haupt An Introduction to Botany 

Haupt Laboratory Manual of Elementary Botany 

Haupt Plant Morphology 

Hill Economic Botany 

Hill, Overholts, Popp, and Grove Botany 

Johansen Plant Microtechnique 

Kramer Plant and Soil Water Relationships 

Kramer and Kozlowski Physiology of Trees 

Lilly and Barnett Physiology of the Fungi 

Maheshwari An Introduction to the Embryology of the Angiosperms 

Miller Plant Physiology 

Pool Flowers and Flowering Plants 

Sharp Fundamentals of Cytology 

Sinnott Plant Morphogenesis 

Sinnott, Dunn, and Dobzhansky Principles of Genetics 

Sinnott and Wilson Botany: Principles and Problems 

Smith Cryptogamic Botany 

Vol. I. Algae and Fungi 

Vol. II. Bryophytes and Pteridophytes 
Smith The Fresh-water Algae of the United States 
Swingle Textbook of Systematic Botany 
Weaver and Clements Plant Ecology 

There are also the related series of McGraw-Hill Publications in the Zoological 
Sciences, of which E. J. Boell is Consulting Editor, and in the Agricultural 
Sciences, of which R. A. Brink is Consulting Editor. 

5 5^ 


Edmund W. Sinnott 



New York Toronto London 



Copyright © I960 by the McGraw-Hill Book Company, Inc. Printed in 
the United States of America. All rights reserved. This book, or parts 
thereof, may not be reproduced in any form without permission of 
the publishers. Library of Congress Catalog Card Number 60-6984 



To the Memory of 



The present volume is an outcome of the author's concern with the 
problems of plant morphogenesis for over 40 years. During that time he 
has followed the literature in this field with some care and has seen the 
subject grow from a relatively minor aspect of botanical science to a much 
wider recognition as one of the central concerns of biology. New tech- 
niques and fields of work have developed in it over the years, and its con- 
tacts with other sciences have grown much closer. Few discussions of the 
subject as a whole have appeared, however, and no extensive synthesis of 
its content has been made. The book that follows is an attempt to do 
this and to organize the field of plant morphogenesis so that it can be 
studied and taught as an integrated discipline. 

This is no simple task, and botanists will doubtless disagree as to 
how it should be accomplished. The method adopted here is first to 
discuss plant growth, emphasizing descriptive and experimental work 
on cells and meristems. This is followed by a survey of the various 
categories of morphogenetic phenomena— correlation, polarity, differen- 
tiation, and the rest— mobilizing under each the major facts that are 
known about it and the work of the chief contributors to the field. This 
portion is largely descriptive and serves to pose problems for considera- 
tion. It is followed by a third section that discusses the major factors 
operative in the control of plant development and is primarily concerned 
with physiological and genetic questions. It touches these broad sub- 
jects, however, only so far as they are concerned with morphogenetic 
problems and then only in a brief and rather summary fashion. Too 
frequent repetition of material in the three sections is avoided by numer- 
ous cross references. 

Throughout the book no attempt is made to go at all completely into 
most of the topics. From the vast literature on subjects of concern to 
plant morphogenesis, however, a considerable bibliography has been 
compiled by means of which the reader is introduced to the more im- 
portant publications where detailed information may be found. In this 
bibliography many important earlier papers of general and historical 
interest have been cited, but emphasis has been placed on relatively 
recent ones. Where available, review papers have been cited. There is 


Preface viii 

some unavoidable repetition of material but this has been reduced by fre- 
quent use of cross references. The book is designed primarily for reference, 
though by means of such a presentation more teachers may be encouraged 
to offer courses in this field. 

To many individuals the author is indebted for good counsel and as- 
sistance in many ways. His thanks are due especially to his former col- 
league Dr. Robert Bloch, who for some years collaborated with him in 
the present project. Professor Ralph H. Wetmore has read a portion 
of the manuscript and made valuable suggestions, as has the author's 
colleague Prof. Arthur W. Galston for the chapter on Growth Substances. 
Various problems have been discussed with many others, and to all 
these the author is grateful. For the positions taken on particular issues 
and the opinions expressed, however, he assumes full responsibility, as 
he does for any errors of fact, attribution, or citation which may appear 
and which are hard to eliminate completely from a book of this sort. 

For several years Mrs. Charlotte Reeder has assisted in compiling 
information, in abstracting papers, and, through her genius for orderli- 
ness, in keeping the project from bogging down under the sheer weight 
of its material. To her the author is greatly indebted. Other assistants have 
been associated with the work from time to time, and their help is 
here thankfully acknowledged. 

For financial assistance, and especially for grants from the Eugene 
Higgins Trust, the author wishes to express his gratitude to Yale Uni- 
versity and its Provost, Edgar S. Furniss. 

The illustrations have been taken from the work of many individuals. 
In every case the source from which they come has been acknowledged 
in the caption, but the author here offers his further thanks for the privi- 
lege of using them. For specific figures he is indebted to the editors of 
Endeavour, the authorities of the Boyce Thompson Institute for Plant 
Research, the Department of Plant Pathology of the University of Wis- 
consin, Triarch Botanical Products, the W. Atlee Burpee Company, Dr. 
G. R. Zundel, and Rutherford Piatt. 

This book has been a long time in the making. It proved to be such a 
time-consuming task that the author was not able to bring it to com- 
pletion until his retirement from academic duties a few years ago. He 
hopes that this long period of ripening may have improved its quality 
and increased its value to those who use it. 

Edmund W. Sinnott 


Preface vii 

Chapter 1. Introduction 1 


Chapter 2. Growth in General H 

Chapter 3. The Cellular Basis of Growth 23 

Cell division — Cell size — Experimental studies — Plane of cell division - 
Cell shape. 

Chapter 4. Meristems 55 

Apical meristems — The shoot apex — Experimental studies on the shoot 
apex — The root apex — Lateral meristems — Meristems in determinate 


Chapter 5. Correlation 95 

Physiological correlations — Genetic correlations. 

Chapter 6. Polarity 116 

Polarity as expressed in external structure — Polarity as expressed in internal 
structure — Polarity in isolated cells — Polarity in plasmodia and coenocytes 
- Physiological manifestations of polarity - Polarity and developmental 

Chapter 7. Symmetry 147 

Inorganic and organic symmetries — Radial symmetry — Bilateral symmetry — 
Dorsi ventral symmetry — Development of symmetry — Symmetry and form. 

Chapter 8. Differentiation 181 

Growth and differentiation — Differentiation as expressed in structure — Ex- 
ternal differentiation - Internal differentiation - Differentiation during 
ontogeny —Differentiation in relation to environment — Physiological dif- 
ferentiation — Differentiation without growth. 

Chapter 9. Regeneration -30 

Regeneration in the lower plants — Regeneration in the higher plants — 
Reconstitution — Restoration — Reproductive regeneration. 


x Contents 

Chapter 10. Tissue Mixtures 258 

Stock-scion interrelations — Chimeras — Somatic mutations. 

Chapter 11. Abnormal Growth 275 

Abnormal development of organs — Production of new types of organized 
structures — Amorphous structures. 


Chapter 12. Introduction to Factors 303 

Chapter 13. Light 308 

Intensity of light — Quality of light - Duration of light — Relation to other 

Chapter 14. Water 324 

Xeromorphy — Experimental work — Effect of the transpiration stream. 

Chapter 15. Temperature 337 

Thermoperiodism — Vernalization — Other temperature effects. 

Chapter 16. Various Physical Factors 345 

Tension — Compression — Bending and swaying — Ultrasonics — Gravity — 
Tissue tension — Absolute size — Bioelectrical factors. 

Chapter 17. Chemical Factors in General 363 

Elements — More complex substances. 
Chapter 18. Growth Substances 374 

Types of growth substances — Growth substances and plant growth — Growth 
substances and correlation — Growth substances and the determination of 
structure — Growth substances and internal differentiation — Other formative 
effects — Mechanism of action. 

Chapter 19. Genetic Factors 415 

Genes — Cytoplasm — Chromosomes. 
Chapter 20. Organization 449 

Bibliography 461 

Name Index 527 

Subject Index 538 



A study of the form and structure of living things has a perennial interest, 
not only for biologists but for everyone. It appeals to the aesthetic in us. 
Philosophers have been concerned with it since the time of Plato, who 
distinguished between matter and form and believed that spirit was 
inherent in the latter. Most naturalists owe their first interest in animals 
and plants to the almost infinite variety of forms which these display 
and which make possible their identification. Although morphology 
(Goethe's term), the science that deals with form, has lost the command- 
ing position it once held, following the advent of physiology and the 
disciplines that connect biology with the physical sciences, it still remains 
the foundation for any thorough knowledge of living things. We must 
all be morphologists before we can be biologists of any other sort. 

In a famous sentence Charles Darwin paid tribute to morphology by 
calling it the very soul of natural history. How curious it is, he remarked, 
that the hand of a man and of a mole, the leg of a horse, the paddle of 
a porpoise, and the wing of a bat, despite the great difference in their 
functions, should all be formed on the same basic pattern. Specific 
bodily forms and structures had long been used to distinguish the major 
groups of plants and animals and were also the basis of that "idealistic" 
morphology which so intrigued Goethe and the biologists of his day. 
Darwin, however, saw in the science of comparative morphology some- 
thing far more significant— a strong support for the doctrine of evolution, 
for only by assuming a common ancestry for each of the groups that show 
a common pattern of bodily form could these similarities be explained. 
Form was widely acknowledged as the most distinctive character on 
which the phylogenetic relationships of organisms could be based and a 
truly natural system of classification constructed. 

The study of organic form, however, poses a problem deeper still, and 
one concerned with the very character of life itself. From the facts of 
embryology it is evident that in the development of an individual there 
occurs as regular a progression of changes in form as has taken place in 
evolution. Indeed, the theory of recapitulation called attention to some 


2 Introduction 

interesting parallels between the two. In a study of form in organic de- 
velopment the biologist has the great advantage of dealing with a process 
that is going on under his eyes and is thus susceptible to experimental at- 
tack. An organism is not static but displays a continually unfolding 
series of changes during its life. It has been well described as a "slice of 
space-time." As knowledge about organic development increased, biol- 
ogists came to realize that development is not only an orderly unfolding 
but that in this process all parts of the growing individual are closely 
correlated with the rest so that an organized and integrated system, the 
organism, is produced. Differences in rate of growth and in character of 
the structures developed are evident, but these various changes do not 
occur independently. They keep in step with one another. When form 
changes, it does so in a regular and predictable fashion. Still more sig- 
nificant, the experimental embryologist is able to show not only that these 
relationships are to be seen in normal development but that they per- 
sistently tend to be restored if development is disturbed. An organism is 
an essentially fluid system through which matter is continually moving 
but which nevertheless maintains a constant form much as does a candle 
flame or a waterfall. 

In the various physiological activities of the living organism there is 
evident the same coordination so manifest in bodily development. What 
occurs in an individual is not simply a series of unrelated metabolic 
processes, but these are tied together in such a precise fashion that the 
life of the organism is maintained in a steady state. Just as the normal 
progress of development tends to be restored if it is disturbed, so the 
normal state of physiological organization tends to be maintained at a 
constant level. This regulatory process of homeostasis is recognized as 
one of the major facts of physiology. 

Organic form is thus the visible expression of an inner relatedness 
characteristic of life at every level. This can be most simply designated 
as biological organization and is the most important problem that con- 
fronts students of the life sciences. Form may be thought of not only as 
the soul of natural history in the sense that it provides a measure of evo- 
lutionary relationship but as the soul of all biology, since it provides the 
most obvious and easily accessible manifestation of the basic character- 
istic of life. 

Biological organization is to be seen most distinctly in bodily develop- 
ment. It is obvious that, to produce an individual with a specific form 
and structure, growth must be more rapid in some directions than in 
others and must form tissues and organs of different character in different 
places. Embryology has shown how precisely the activities in one part of 
the developing individual are related to those in every other part. Few 
happenings in nature are as fascinating to watch as the unfolding and 

Introduction 3 

growth of a leaf or a flower from a tiny primordium, especially when this 
is speeded up to our eyes by time-lapse photography. Every step is co- 
ordinated with all the others as though a craftsman were molding it ac- 
cording to a plan. Within the whole, the cells and other subordinate 
parts do not develop independently but all are knit together into an 
organized system. 

How all this is accomplished and a specifically formed organism pro- 
duced is not yet understood, although specific parts of the process are 
now well known. Most metabolic activities are yielding to biochemical 
analysis; students of gene action find that specific substances are produced 
by specific genes, and the nucleic acids, with their remarkable properties, 
are recognized as being at the very foundation of life itself. How all the 
various metabolic and developmental activities are related in such an 
orderly fashion, however, and proceed without interference or confusion 
so that, step by step, an organism is produced poses a problem of a very 
different kind. Relations, not chemical changes, are the facts to be ex- 
plained. The problem must be approached experimentally but one should 
recognize that this may require the development of techniques and ideas 
not yet explored. 

The biological science concerned with this dynamic and causal aspect 
of organic form is evidently different from either morphology, physi- 
ology, or embryology, though partaking of all three. It deserves a 
name of its own. The Germans usually call it Entwicklungsmechanik, a 
name proposed by Wilhelm Roux. This great zoologist is looked upon as 
the father of the science of zoological morphogenesis. He founded the 
Archiv fiir Entwicklungsmechanik der Organismen, a journal which now 
occupies 16 feet of library shelf and contains a vast amount of material, 
chiefly on the animal side. Elsewhere this science has often been termed 
experimental morphology or experimental embryology. Rritish biologists 
sometimes refer to it as causal morphology. In recent years it has generally 
been given a more appropriate name than any of these— morphogenesis. 
The derivation of this word is obvious— the origin of form. Who first used 
it is not certainly known but Ernst Haeckel employed the cognate form 
morphogeny (Morphogenie) in 1859. Some have employed this term in 
a strictly descriptive sense, essentially as synonymous with developmental 
morphology. More commonly and properly, however, it includes, in addi- 
tion to a discussion of purely descriptive facts as to the origin of form, a 
study of the results of experimentally controlled development and an 
analysis of the effects of the various factors, external and internal, that 
determine how the development of form proceeds. In other words, it 
attempts to get at the underlying formativeness in the development of 
organisms and especially to reach an understanding of the basic fact of 
which form is the most obvious manifestation— biological organization 

4 Introduction 

itself. It is in these senses that the term morphogenesis will be used in 
the present volume. 

Each of the major biological subsciences is intimately related to the 
others. One cannot study genetics apart from physiology, for example, 
or physiology from morphology, or taxonomy and evolution from all 
these. It may well be maintained, however, that morphogenesis, since it is 
concerned with the most distinctive aspect of life— organization— is the 
crossroads where all the highways of biological exploration tend to con- 
verge. Its subject matter deals with some of the most elusive and intract- 
able phenomena in science, but it is here that the greatest discoveries of 
the future are likely to be made. These will be significant not only 
for biological problems but for many others that man faces. Even 
philosophy, long concerned with problems of form, is still gaining 
from this source fresh insights into its chief task, an understanding of 

More study has been given to morphogenetic problems with animals 
than with plants. A great advantage of animal material is that in many 
groups the egg is discharged into water and the embryo develops there, 
at least through its earlier stages, and is thus easily accessible for obser- 
vation and experiment. Among higher plants, on the contrary, all the early 
development of the embryo takes place within the ovule, surrounded by 
many layers of tissue and relatively inaccessible. The result has been that, 
save for rather special material like the egg of Fucus which can be 
treated much like that of an animal, very little morphogenetic work has 
been done on plant embryos. Modern techniques, however, by which 
it is possible to grow the embryos of some higher plants to maturity in 
culture, are making the science of experimental plant embryology a more 
fruitful one. 

Workers with plants have a number of advantages, however. In plants, 
permanently embryonic regions, the meristems, are available for study. 
At the tip of shoot and root and in the cambial layers these are inde- 
terminate and produce new plant structures almost indefinitely. Such 
meristems are usually numerous or extensive on the same plant so that 
ample material for the study of development, identical in genetic con- 
stitution, is available. Growth and differentiation in the development of 
a plant are thus continuing processes and not limited to a single and 
often brief life cycle. 

Organs such as leaves, flowers, and fruits, which are determinate in 
growth, pass through a cycle closely comparable to that of individual 
animals, and morphogenetic problems can also often profitably be studied 
in them. The fact that they are usually produced in abundance on a 
single plant is a further advantage, for here the investigator need not be 
concerned about genetic diversity in his material but can study strictly 

Introduction 5 

comparable organic forms under a wide range of environmental condi- 

Another important difference between botanical and zoological ma- 
terial is concerned with the behavior of individual cells during develop- 
ment. In animal embryos many cells are relatively free to move about so 
that certain morphogenetic changes are due to movements of cells or cell 
groups rather than to differences in relative growth. In plants above the 
simplest types, on the contrary, cells are almost always attached firmly 
to their neighbors so that morphogenetic movements have no part in de- 
velopment. Changes of form are the result of differences in the location 
or orientation of cell divisions or in the size or the shape to which the 
individual cells grow. This makes the study of morphogenetic problems 
somewhat simpler in plants because development leaves a record of its 
course in the structure of the growing system itself. 

Most plant cells have rather firm walls as compared with animal cells, 
and the structures that they produce are therefore not as soft and plastic 
as in many animals. A plant part tends to hold its form rather well and 
can thus be measured more easily and accurately. Its anatomical struc- 
ture is also less fluid. Certain organs, such as woody stems and hard- 
shelled fruits, retain their form when dead and dry and can then be kept 
for study without the necessity for special preservative treatment. 

Plant material is generally more tractable than that of animals, is easily 
grown, and lends itself readily to experiment. Because of their stationary 
habit, plants are more susceptible to changes in environmental influences, 
notably water and light, than are animals, and the morphogenetic effects 
of such factors may be studied more easily in them. 

The organization of a plant, too, is much looser than that of most ani- 
mals. The individual organism is less sharply marked and specific, and 
its powers of regeneration are far greater. Its structural plan is simpler, 
for the stationary habit of most plants renders unnecessary several organ 
systems found in animals, notably a digestive tract, excretory organs, 
musculature, and a nervous system. Morphogenetic problems can there- 
fore be studied in plants uncomplicated by the physiological complexi- 
ties inseparable from animal life. The absence of a nervous system, which 
has such an important role in animal development, is of particular ad- 
vantage, for developmental processes in plants are under the control of 
relatively unspecialized protoplasm, and they may thus be studied most 
directly and at their simplest level. There is no reason to believe that 
the fundamental phenomena of the development of form are not as 
manifest in these relatively simple systems as in the more complex ones 
of animals. 

What is now called morphogenesis came first into prominence in the 
late decades of the nineteenth century after the early enthusiasm over 

6 Introduction 

the idea of evolution had given place to a more sober realization that it 
did not provide a solution for all the problems about living things. The 
list of men working primarily with animal material includes some of the 
most brilliant names in the science of zoology. In plant morphogenesis, 
most of those who made important early contributions were chiefly en- 
gaged in other fields, especially morphology, physiology, and pathology. 
Here are remembered the classical studies of Hanstein on meristems and 
their derivatives, of Winkler on chimeras, of Haberlandt on plant hor- 
mones, of Kiister on abnormal growth, of Klebs on the effects of the 
environment, of Goebel on the general area to which he gave the name 
of organography, and many others. 

It is to Herman Vochting, however, long Professor of Botany at the 
University of Bonn, that botanists owe the first thoughtful discussion of 
such problems as polarity, differentiation, and regeneration. His "Organ- 
bildung im Pflanzenreich," published in 1878, is a classic and may be 
said to have founded a new field of botanical investigation. It deserves 
to be read by all students of development even today. About the turn of 
the century the zoologist Hans Driesch stated his often-quoted aphorism, 
"The fate of a cell is a function of its position," which in a few words 
sums up a central fact of biological organization. What few botanists 
know is that Vochting, in a book written 20 years earlier, said the same 
thing in almost the same words. 1 Other botanists also made important 
contributions here. Much of the work of men such as van Tieghem, Jost, 
Sachs, Pfeffer, Schwendener, and Strasburger was on problems that we 
should now call morphogenetic. 

Most studies in plant morphogenesis have been made with vascular 
plants— pteridophytes, gymnosperms, and angiosperms. It should be 
remembered, however, that many of these problems can be approached 
more directly through work on the lower ones. Polarity, for example, is 
manifest in its simplest form in some of the filamentous algae. The very 
beginnings of differentiation are to be seen in the lower thallophytes. 
Almost every cell in some of the bryophytes may easily be induced to 
regenerate. The problem of the development of form is nowhere posed 
more directly than in the formation of the remarkable fruiting bodies of 
some of the myxomycetes and of the higher fungi. These more primitive 
plants are proving to be ideal material for the study of many problems 
in physiology and genetics, and although they have been rather neglected 
in morphogenesis, they offer abundant opportunities for fruitful work 
in this field. 

The science of plant morphogenesis has never received a comprehen- 

1 Die jeweilig zu verrichtende Function einer Zelle wird in erster Linie durch den 
morphologischen Ort bestimmt, den sie an der Lebenseinheit einnimmt. Organ- 
bildung im Pflanzenreich, 1878, p. 241.) 

Introduction < 

sive formulation. To bring together in a single volume a discussion of 
the various phenomena that distinguish it and of the factors that have 
been found to affect the development of plants, together with a bibliog- 
raphy of some of the most important publications dealing with the sub- 
ject, should help to give it recognition as a distinctive botanical disci- 
pline. To attempt this is the purpose of the present volume. 

Morphogenesis is such an immense subject, however, covering most of 
the territory of biology, that to organize its facts and its problems in a 
logical and reasonably compact fashion is a matter of much difficulty. 
The method used here is to divide the subject into three parts. First is 
presented a brief discussion of plant growth as a necessary introduction 
to morphogenesis proper, placing particular emphasis on its cellular basis 
and on the activities of the meristems. The remaining subject matter is 
then divided into two sections: first, the various phenomena of plant 
morphogenesis and the more important studies that have been made on 
them and, second, a brief account of the morphogenetic factors that have 
been found to affect the development and form of plants. In the first 
there are chapters on Correlation, Polarity, Symmetry, Differentiation, 
Regeneration, Tissue Mixtures, and Abnormal Growth. In the second are 
discussed the effects of light, water, temperature, and other physical fac- 
tors; inorganic and organic substances (especially growth substances); 
and finally the various genetic factors. There is a concluding chapter on 
the problem of Biological Organization. 

At the end of the book is a selected list of references to some of the 
more important books and papers on the subject. This obviously must be 
far from complete since the literature is enormous and scattered through 
most of the fields of botany. An attempt has been made to include, both 
in text and references, some of the important early work, not alone for 
its intrinsic but also its historical value. In the more rapidly advancing 
fields, where many of the results from older studies have now been 
superseded, only a relatively few of the earlier papers are mentioned, 
and there is considerable representation of recent work. Aside from 
bringing the subject up to date, these later papers through their 
bibliographies will give the student a means of entry into the literature 
of a given field. Opportunities for further research, particularly in areas 
now less popular than in the past, are so numerous that the author has 
felt justified in calling attention to some of them from time to time. 

The problem of deciding which pieces of work to include in a discus- 
sion of this sort and which to omit has been very difficult. Among those 
mentioned there are doubtless some that will be regarded by many read- 
ers as relatively unimportant. The omission of others will be criticized. 
It is hoped, however, that the papers chosen will provide a fair picture of 
accomplishment in plant morphogenesis. A considerable number of 

g Introduction 

review papers are cited which help summarize results in particular fields. 
No one person, and certainly not the present writer, is competent to give 
a thoroughly informed and authoritative judgment on the relative worth 
of the wide variety of investigations here discussed. It is hoped, how- 
ever, that one service of the book will be to introduce its readers to the 
subjects of these studies even though in some cases a piece of work can 
be given little more than mention. 

Since there is no sharp line between morphogenesis and its neighbor- 
ing fields of morphology, physiology, genetics, and the chemical and 
physical sciences, much of the advance in it will doubtless be made, as 
in the past, by men whose chief concern is with one of these other disci- 
plines; but as morphogenesis becomes better organized and as more op- 
portunities for training and research in it are offered by our colleges and 
universities, there will be more students whose primary interests are 
directed to it. More than other biological sciences, perhaps, morpho- 
genesis will need to maintain close contact with a wide variety of other 
fields, for few can hope to be competent in its entire area. To develop 
this comprehensive subject fruitfully will require the active cooperation 
of many sciences, and by this means the morphogenetic point of view 
can thus help to integrate all of biology. 




Growth in General 

The process of organic development, in which are posed the chief prob- 
lems for the science of morphogenesis, occurs in the great majority of 
cases as an accompaniment of the process of growth. The association be- 
tween these two activities is not an invariable one, for there are a few 
organisms in which growth is completed before development and dif- 
ferentiation are finished, but far more commonly the form and structure 
of a living thing change while it grows. Knowledge about growth is 
therefore necessary for an understanding of development, and any dis- 
cussion of morphogenetic problems in plants should be preceded by a 
discussion of plant growth in general. This is the purpose of the first 
few chapters of the present volume. 

Definition of Growth. The term "growth" has been variously defined 
by biologists. For some (Hammett, 1936) 1 it includes not only increase 
but also the accompanying phenomena of progressive differentiation. 
Most regard this definition as too inclusive and would limit it in one re- 
spect or another. Since much of the increase in volume is brought about 
merely by gain in amount of water, increase in dry weight might seem to 
be the best measure of growth, but sprouting seeds kept in the dark will 
"grow" into large, etiolated seedlings through intake of water though 
their dry weight actually decreases. In one sense these sprouts have 
grown, but in another they have not. The fundamental fact in all growth, 
of course, is the self-multiplication of living material, a process of much 
biological significance. For this reason, growth might best be defined 
as increase in amount of protoplasm. Even if we could agree, however, 
as to what constitutes protoplasm (whether vacuoles, for example, are 
parts of it), it would be impossible, as a practical matter, to measure 
this. Furthermore, in every organism, and particularly in every plant, 
there is much material ( such as cell walls and starch grains ) which is an 
integral part of the organism but which presumably is not living, and it 
seems illogical not to regard increase in such material as part of growth. 

1 For bibliographic information concerning books and papers referred to in the text, 
see Bibliography, pp. 461 ff. 


12 Growth 

In most multicellular organisms growth is accomplished chiefly by cell 
multiplication, and to some observers this process seems to be an essen- 
tial part of growth. In the cleavage of many animal eggs and in similar 
processes in plants, however, there is a great increase in cell number 
but none in the actual material which constitutes the "growing" structure, 
and whether such cellular increase should be regarded as growth is a 
question. In the case of the female gametophyte in the megaspore of 
Selaginella, and especially of the young embryo which develops there, 
growth of an organized structure by cell multiplication certainly occurs, 
but at the expense of material stored in the spore. Where an entire mass 
is cut up into cells, as in the development of a male gametophyte in a 
microspore, one may doubt as to whether this should be called growth 
at all. What definition of growth one adopts depends on the particular 
problem with which he is concerned. 

For the study of morphogenesis, the most important aspect of growth 
is the permanent increase in volume of an organ or organism, regardless 
of how it is accomplished, and this is the sense in which the term will 
here be used. The ultimate problem-the self-multiplication of living ma- 
terial-is one primarily for the student of physiology and reproduction, 
but it is the gross and geometrical result of such growth with which 
morphogenesis is chiefly concerned. 

Growth in Plants. In most plants, the process of growth is different in 
one important respect from that in animals. The typical mature plant 
cell is surrounded by a relatively stout cellulose wall which under ordi- 
nary conditions prevents any further cell division or growth save in ex- 
ceptional circumstances. The cells are rather firmly cemented together 
and thus unable to move about or migrate. Plant tissues are therefore in- 
capable of growth and renewal except through the activity of thin-walled, 
relatively undifferentiated embryonic regions, or meristems, where occur 
the divisions that produce new cells and the changes by which these at- 
tain their final size. These meristems are rather sharply localized. In plant 
axes where growth is continuous and often indeterminate, growth in 
length is controlled chiefly by the activity of meristems at the tip of each 
root or stem. The older portion of the axis, having once attained maturity, 
does not make further growth in length. A tree increases in height only 
at the tips of its twigs and not elsewhere. In the stems of some mono- 
cotyledons, however, growth of the stem in length may continue for a time 
by the activity of intercalary meristematic regions at the base of each 
internode. Perennially growing roots and stems increase in thickness 
through the activity of a lateral meristem, or cambium, situated between 
xylem and phloem, by which the growth of both these tissues is accom- 
plished. There are other sharply localized meristematic regions, such as 
the phellogen, or cork cambium. 

Growth in General 13 

In organs which (unlike the axes) have a limited or determinate 
growth, such as leaves, flowers, or fruits, the meristems are usually not 
localized but are diffuse, so that the whole organ, or most of it, is grow- 
ing throughout and not at any particular point. Such structures have a 
growth cycle of their own, much as does an animal body, and when they 
reach maturity all their tissues stop growing and there is no embryonic 
region set apart by which further growth may be accomplished. 

Graphical and Mathematical Analysis of Growth. One of the most 
obvious facts which a study of growth reveals is that it does not proceed 
at a constant rate. Many factors influence this rate, but under normal 
and favorable conditions a growing organ or organism undergoes a 
characteristic course of increase, first growing slowly, then with increas- 
ing speed, and finally slowing down again until growth stops entirely. 

It is possible to picture this graphically in various ways. In Table 2-1 
are presented the data for the increase in diameter of a gourd fruit from 
its early state as a small ovary primordium until maturity. If these diam- 
eters are plotted as ordinates against time in days, the growth curve 
shown in Fig. 2-1 results. This is an S-shaped, or sigmoid, curve and is 
typical of most growing organisms, both plant and animal, though it is 
subject to much variation. It presents the changing size of the growing 
organ throughout its course but does not give a very clear picture of the 
changing amounts of daily growth. If daily increments are plotted in the 
same way against time, the curve in Fig. 2-2 results. These increments 
are small at first, then progressively larger, and then smaller again. 

Such graphical representations of growth have long interested biol- 
ogists and mathematicians, who have endeavored to analyze them in 
mathematical terms and thus obtain clues as to the character of the 
growth process itself. In many cases such analyses have proved helpful 
in providing a simple statement of the course of growth, but there are so 
many variables involved in growing organisms that one can hardly expect 
to express their increase completely in an equation. 

Table 2-1. Growth of a Gourd Fruit from a Small Primordium to Maturity 

Diameter, Diameter, 

Date mm. Date mm. 

July 30 2.4 Aug. 9 30.0 

" 31 3.1 " 10 35.2 

Aug. 1 3.9 " 11 40.0 

" 2 5.1 " 12 43.8 

" 3 6.5 "13 46.0 

" 4 8.4 " 14 47.0 

" 5 11.0 " 15 47.5 

" 6 14.0 " 16 47.9 

" 7 18.0 " 17 48.0 

" 8 23.5 " 18 48.0 



Time in days 
Fig. 2-1. Sigmoid growth curve. Fruit diameters in Table 2-1 plotted against time in 

The first period of growth, in the most typical instances, shows the 
regular acceleration of a mass increasing, so to speak, at compound in- 
terest, the growth during any period being a constant proportion or per- 
centage of the amount already present, and the "interest" being com- 
pounded continuously. In Table 2-1 it will be observed that for the first 10 
days the increase in diameter each day is an approximately constant 
proportion of the diameter of the day before, though the daily incre- 

Time in days 
Fig. 2-2. Graph of daily increments in fruit diameter, from Table 2-1. 

Growth in General 15 

ments themselves continually increase. The equation for "compound- 
interest" growth is the familiar one 

?i = P e rt 

where Pi is the size at any time t; P the size at the beginning of growth; 
e the base of the natural logarithms, 2.18; and r the rate of growth (in- 
terest or exponential rate). This can be expressed, by using common 
logarithms, as 

logP 1 = logP + loge(rt) 

To find r for the first 10 days in Table 2-1, we substitute in this equa- 
tion as follows: 

1.4771 = 0.3802 + 0.4343 X lOr 

1.4771 - 0.3802 

r = 

= 0.25 

10 X 0.4343 

This is the rate of diameter increase per day at which, continuously com- 
pounded, this fruit is growing, expressed as a per cent of its previous 
growth. If the logarithms of the successive size of gourd fruits in Table 2-1 
are plotted against time (or the data plotted on semilogarithmic paper) 
the graph in Fig. 2-3 results. Here the growth for the first 10 days is seen 



Time in days 

Fig. 2-3. Curve of the logarithm of fruit diameter in Table 2-1 plotted against time in 

16 Growth 

to fall along an essentially straight line, showing that the fruit was growing 
at a constant exponential rate. The slope of the line is a measure of this 


The resemblance between such organic growth and compound-interest 
increase has long been noted, but it was particularly emphasized by 
V. H. Blackman (1919). He proposed the term efficiency index for the 
"interest rate" in such growth. That living things in the early stages of 
their development should grow in this way is not surprising, for if 
embryonic material is self-multiplicative, the increase per unit of time 
should be proportional to the growing mass. 

To explain the rest of the growth curve is more difficult. Evidently 
growth cannot proceed in any organism at a continually accelerating rate, 
if for no other reason than that building material would soon be used up. 
The gradual slowing down and final cessation of growth are far too 
regular a process, however, to be due to mere exhaustion of materials. A 
plant or animal provided with a superabundance of nutrients will rarely 
exceed the size characteristic for its species. Each organ or body has a 
specific growth cycle through which it passes, and the second part of 
this cycle, in which growth is falling off in rate, is much like the first part 
in reverse, so that the entire growth curve thus tends to be symmetrical. 
In such a case the periodic increments form a curve (Fig. 2-2) which 
much resembles the so-called normal curve, or curve of probability. This 
relationship has been observed and discussed by various workers, 
especially Pearl and his school (1915), but any causal relationship be- 
tween the two types of curves is not easy to see. 

The similarity between growth and the chemical phenomenon of auto- 
catalysis, in which the products of a catalytic process accelerate the 
process itself, has been noted by many observers. Robertson ( 1923 ) and 
his followers have attempted to analyze the whole growth curve as a sim- 
ple ( monomolecular ) autocatalytic reaction by which an enzyme breaks 
down a mass of substrate. During the first part of the process, growth will 
therefore accelerate, but when the amount of substrate becomes seriously 
reduced, rate of growth will also be reduced, and when the substrate is 
exhausted, growth will cease. 

Robertson derives the typical sigmoid curve from the equation for 

l ^A^x = k{t ~ tl} 

where x is the volume of the organic structure (amount grown) at any 
time t; A the final size of the structure; U the time at which it attains half 
its final size; and k is a specific growth constant, or exponential growth 

Growth in General 


The essential feature of this growth equation is that the rate of growth 
is determined by the amount of growth which is yet to occur. This im- 
plies that the final size is established at the start of growth, either by the 
amount of available building material or in some other way. If only one 
"master reaction" were concerned, the course of growth might well fol- 
low Robertson's equation, but there are evidently many substances in- 
volved and many processes going on simultaneously which probably 
make the growth process too complex to be analyzed by any one reac- 

Growth curves for certain organs and organisms fit Robertson's equa- 
tion fairly well (Reed, 1927). For others the fit is not so close. In a struc- 
ture which in its growth fits the equation perfectly, the rate of exponential 



Fig. 2-4. Growth of large and of small fruits in Cucurhita. Log of fruit volume plotted 
against time in days. Early growth in all races is at a constant rate (straight line) the 
slope of which here is the same in large and small races. Solid circle, time of flowering. 
( From Sinnott. ) 

growth will constantly decrease (since it is proportional to the amount 
yet to grow, which is decreasing) so that the curve of the logarithm of 
size against time will be convex from the beginning. When the rate is 
relatively low the difference between this and the straight line of con- 
stant exponential growth at first is not great, since the absolute amount 
of growth in these early time intervals is slight. In many growth curves, 
such as those of fruits of various races of cucurbits presented by Sinnott 
( 1945b ) , the early growth is at a constant exponential rate and shows no 
indications of convexity of line (Fig. 2-4). Furthermore, large and small 
fruits which are here compared grow at essentially the same rates but 
for very different durations. Under these conditions, Robertson's formula 
would require that the length of time between half size and maturity 

18 Growth 

should be the same for all races, but actually in the small-fruited races 
this period is much less than in the large-fruited ones. 

The mathematical analysis of growth involves many complexities and 
has been developed in much greater detail than is possible in the present 
brief discussion. For a fuller treatment of this subject the reader is re- 
ferred to the work of Pearl ( 1939 ) , D'Arcy Thompson ( 1942 ) , Erickson 
(1956), and others. 

It is evident that no single mathematical statement will express all types 
of growth nor perhaps any of them with complete exactness. Growth is 
a very complex process involving many variables, and it is not to be ex- 
pected that it can be compressed into a single equation. Even if it could 
be, this would not tell us a great deal for, as D'Arcy Thompson well says, 
a formula "which gives a mere coincidence of numbers may be of little 
use or none, unless it go some way to depict and explain the modus 
operandi of growth." That growth under some conditions proceeds as at 
compound interest and at others like an autocatalytic process is of some 
importance in providing a clue to the mechanism of growth, but so far 
mathematical analysis has added comparatively little to our understand- 
ing of the fundamental character of growth itself. For this we must look 
to a more concrete study of the growth process in terms of genetics, bio- 
chemistry, and physiology. 

Variation in Growth. Many structures do not show the simple sigmoid 
growth discussed in the preceding section. Just as the smooth course of 
growth in mammals is interrupted by birth and by puberty, it is modified 
in various ways in plants. In fruits of peach and cherry, for example, 
Tukey and Young (1939) and others have shown that after these struc- 
tures are partly grown there is then for some time no increase in volume. 
This is the period in which the endocarp, or "stone," is being formed. 
Later the fruit begins to enlarge again, so that a curve like that in Fig. 
2-5 results. Duncan and Curtis ( 1942 ) have shown a somewhat similar 
growth curve in the fruit of certain orchids where one epoch of growth is 
associated with meiosis and a later one with seed maturity. In vegetative 
structures, aside from annual periods, there are also sometimes discon- 
tinuous cycles, as in the pear shoots studied by Reed ( 1927 ) , where there 
may be three such in one season. The dandelion scape shows a some- 
what similar growth pattern (Chao, 1947). It is rapid during flower de- 
velopment, much slower after the flower opens, and then accelerates 
greatly as the fruit becomes mature. 

Borriss ( 1934a ) reports that etiolated stipes of Coprinus show a marked 
periodicity of growth with maxima 3.5 to 4.5 hours apart. This and similar 
cases may be manifestations of endogenous rhythms such as have fre- 
quently been reported in other processes ( p. 322 ) . 

Growth in General 


In many plants, particularly herbaceous ones, growth is not evenly 
distributed throughout the length of the stem. Thus in tobacco (Wolf, 
1947) and in maize (Heimsch and Stafford, 1952) the internodes are 
progressively longer from the base to about half way up the axis and 
then are progressively shorter to the apex. The distribution of growth in 
a developing leaf blade also shows local differences (Avery, 1935). Many 
alterations in form arise from local changes in growth. 

Determinate and Indeterminate Growth. In most animals, growth is 
part of a definite life cycle and produces a determinate structure. In many 
plants, on the other hand, the growth of the body is essentially indetermi- 

A |0 



Fig. 2-5. Intermittent curves of growth of a cherry fruit from flowering to maturity. 
( From L. D. Tukey. ) 

nate and, within certain limits, may go on indefinitely through the activity 
of terminal and lateral meristems. Even such theoretically unlimited 
growth (such as that of a tree in height or a vine in length), however, 
usually reaches a limit and in its growth follows a curve which is S- 
shaped. In some plants, such as the sunflower and most grasses, height 
is not indeterminate but is limited by a terminal inflorescence, and stem 
growth in such cases is typically sigmoid. Lateral organs, such as leaves 
and fruits, which do not grow by localized meristems have still more 
definite growth cycles and are quite comparable to single animal indi- 
viduals and show similar growth curves. Examples of these are fruits 
(Fig. 2-1), leaves (Wolf, 1947), and ovules and embryos (Rietsma et al, 



1955 ) . Plants provide examples of all types of growth from that in loosely 
organized, essentially indeterminate structures to highly organized and 
sharply determinate ones and therefore are particularly good material 
for a study of the mechanism by which growth, presumably free and 
continuous in primitive organisms, becomes controlled and molded into 
a definite cycle or pattern. Such cyclical, controlled growth is one mani- 
festation of the general phenomenon of biological organization. 

Growth and Size. The size that an organism attains is often an im- 
portant factor in determining the character of its development, and size 
is intimately related to growth. Differences in ultimate size may be due 




- / / 


/ / 






- / / 



Fig. 2-6. Diagram of growth of stipe and pileus of the common mushroom, Agaricus 
campestris. Homologous points are connected by lines. Growth is most active in the 
region intermediate between base and apex. (From Bonner, Kane, and Levey.) 

to differences in rate or in duration of growth or in both of these. Little 
is known in plants as to the relation between growth and size. The in- 
creased size of heterozygous corn plants is apparently associated with a 
higher growth rate (Whaley, 1950), and this may be true rather gen- 
erally for size difference in indeterminate structures. In determinate 
ones such as the fruit, however, rate may not be important. The great size 
differences between small-fruited and large-fruited cucurbits of the same 
species studied by Sinnott (1945b) are due in almost every case to differ- 
ences in duration of growth, for growth rate is essentially the same in all 
of them ( Fig. 2-4 ) . This difference in duration applies to all recognizable 

Growth in General 21 

parts of the growth cycle— from primordium to flowering, from flowering 
to the end of exponential growth, and from this point to growth cessation. 

Growth in plants has usually been studied in the higher forms because 
of their generally larger size and the greater ease with which observations 
can be made upon them. Some lower plants, however, offer good oppor- 
tunities for growth studies. Borriss (1934a) found that growth is not 
evenly distributed in the sporophore stalk of Coprinus but is progressively 
more rapid toward the apex. This has been confirmed by Bonner, Kane, 
and Levey (1956; Fig. 2-6), who find that, after the early stage, growth is 
accomplished chiefly by elongation of the cells of the hyphae. By dusting 
the tips of young sporangiophores of Phycomyces with starch grains and 
recording changes photographically, Castle ( 1958 ) has analyzed the 
distribution of growth here, both as to longitudinal and circumferential 
increase. The ratio between these two components is not constant but 
changes with location on the sporangiophore. 

Brown, Reith, and Robinson ( 1952 ) examined the mechanism of growth 
in plant cells, both in intact organs and by culture of isolated fragments. 
Lindegren and Haddad (1954) found that in yeast cells growth rate is 
constant and that it begins and ends abruptly, thus differing from growth 
in most higher organisms. 

Physiology of Growth. The essential fact in growth is the increase in 
amount of the various components of the organism. This results from the 
self-multiplication of its essential portions, the genes and their basic 
constituents, the nucleic acids. Everywhere syntheses are involved. This 
general field is closer to physiology than to morphogenesis. Also essen- 
tially physiological are problems concerning the rate and duration of 
growth. These traits may be affected by many factors, some in the genetic 
constitution of the plant and others coming from its environment, such 
as temperature, light, water, and chemical substances of many kinds. To 
consider these aspects of growth would require much space and is outside 
the purpose of the present volume. The physiology of plant growth has 
been frequently discussed, as by Thimann ( 1954 ) . 

It is not growth itself that is of morphogenetic importance but its 
relative distribution, for this is what determines form. Richards and 
Kavanagh (1945a) call attention to the fact that a study of growth by 
geometrical changes alone, as is commonly done, does not tell the whole 
story. Density (mass per unit volume) and volume may be increasing at 
different rates in different regions. The forces of stretching and compres- 
sion that result may affect the distribution of growth. Under the discus- 
sion of various factors in the latter part of this book, growth and its con- 
trol will from time to time be mentioned, but as part of a larger problem. 
This problem is the development of a specifically formed and organized 

22 Growth 

plant body. One of the leading students of morphogenesis has recently 
expressed this well: 

I think, after we have surveyed the facts, that the whole subject of growth 
will seem bigger than the chemistry of synthesis, and that it will be more likely 
that this latter will seem a small (although important) part of a larger scheme 
in which growth is used here and there, sometimes encouraged, sometimes dis- 
couraged, and in such a way that a consistent, whole, individual organism is 
created in an orderly and masterful fashion. (J. T. Bonner, 1952a, p. 61.) 


The Cellular Basis of Growth 

One of the great biological generalizations of the nineteenth century is 
the cell theory, commonly attributed to the botanist Schleiden and the 
zoologist Schwann and formally stated in 1839. This was a recognition of 
the fact that all organisms are composed of living units, the cells. The 
theory provided a common foundation for a study not only of structure 
but of growth (cell multiplication) and development (cell differentia- 
tion ) . It has served as a unifying concept for all biology, somewhat com- 
parable to the atomic theory in the physical sciences. 

The implications of the cell theory for morphogenesis are important. 
In the minds of those who promulgated it, it meant that the cell is the 
true biological individual and that an organism is the result of the ac- 
tivities of its constituent cells. That the cell is thus the primary agent of 
organization is the opinion of some biologists today. In such a view the 
organism is looked on as a sort of cellular state, built by the cooperative 
efforts of its citizens among whom, as in a human society, there is a high 
degree of division of labor. In support of this idea are cited cases such 
as that of certain of the slime molds, where some thousands of individual 
cells (myxamoebae), entirely independent in the early stages of the life 
cycle, become aggregated into a cellular mass and then by their mutual 
interactions build up a fruiting body of a specific size and form (p. 223). 

Other biologists, however, believe that the true individual is the 
organism, essentially a mass of protoplasm divided into cellular units. 
Such division has the advantage that it makes possible the differentiation 
of parts and the segregation of various physiological activities within 
particular cells. The organism may thus be said to make the cells rather 
than the cells to make the organism. In this conception the multicellular 
plant body is to be thought of as having arisen not through the aggrega- 
tion of individual cells, originally separate, but by cellular multiplica- 

That this organismal theory gives a better picture of the growth and 
activities of plants and animals is suggested by the high degree of co- 
ordination and self-regulation that exists in a living thing. The produc- 
tion of individuals essentially alike by a variety of developmental routes 


24 Growth 

in regeneration is difficult to explain as a result of the interaction of essen- 
tially independent units. The cellular society, if it is one, must have a 
strong central government which regulates the activities of its individual 
members. A certain amount of self-differentiation undoubtedly exists, in 
which a given organ or structure, once its development has begun, 
proceeds more or less independently of the rest, but the parts are usually 
interdependent. The problem of organization, the central one for biology, 
can be attacked more hopefully by a study of organized systems as 
wholes than simply of the units of which these are composed. 

There are obviously considerable differences in the degree and level of 
organization. In plants with indeterminate growth, especially in some of 
the lower groups, the "individual" is little more than a colony of cellular 
individuals, which are so nearly independent that if isolated they 
will produce new plants directly. Among higher forms it is much more 
closely organized. Even here one can hardly tell, in types such as straw- 
berries and many grasses, for example, which spread by stolons or 
rootstocks, how much should be regarded as a single individual. In many 
cases, however, growth is determinate, the number of parts is relatively 
constant, and the individual is a distinct and specific thing. In no plants 
does it reach the high level of organization that most animals display. 

In support of the concept that the organism is the developmental unit, 
one may point to the many cases in plants where, as in the alga Caulerpa, 
a very considerable degree of differentiation occurs into "roots," "stems," 
and "leaves" but where there are no cellular boundaries at all. The whole 
plant is a coenocyte, a simple mass of cytoplasm in which great numbers 
of free nuclei are embedded or move about. In other algae where the 
general character of the plant body seems to be similar to this, some 
species have uninucleate cells, others multinucleate ones, and others are 
entirely coenocytic, with no cell walls save where reproductive organs 
are formed. In most of the true molds, or Phycomycetes, the hyphae are 
multinucleate and not divided into cells, and this is true of certain of the 
higher fungi also. In some other algae and fungi the partitions across the 
filaments are incomplete and have a central perforation through which 
cytoplasm can flow, so that there is no true cellular structure. In the 
developing endosperm of the higher plants there is usually at first a 
large number of free nuclei in a mass of cytoplasm, but these gradually 
become separated from each other by the growth of walls. 

The difference between these two views of the relation between the 
cell and the organism is of much importance for morphogenetic theory. 
The individual cells are certainly significant, particularly in physiol- 
ogy, and their presence makes possible much useful analysis of de- 
velopmental processes, but just how a group of cells develops into an 
organism still remains the central problem. 

The Cellular Basis of Growth 



Growth of plants and animals, in the last analysis, is an increase in 
amount of living stuff in them, but this growth is almost always accom- 
panied by an increase in the number of their cells. This takes place by 
the process of cell division, which thus assumes much significance for 
problems of growth and differentiation. The precise method by which 
new cells are formed was not understood for some time after the cell 
theory was established. In the seventies of the last century a number 
of botanists and zoologists, Strasburger prominent among them, made 
clear the mechanism of mitosis and the leading part played by the 
nucleus in cell division. 

Division does not take place in all parts of the plant individual. In 
higher forms it is limited chiefly to apical and lateral meristems and to 



- J 





Fig. 3-1. Division of a vacuolate cell showing the development of the phragmosome, 
which precedes the cell plate. ( From Sinnott and Bloch. ) 

the growing regions of determinate organs, though there may be cell 
division under certain conditions in other parts of the plant. 

Division is usually studied in small-celled meristematic regions where 
the cells are not vacuolate or have only small vacuoles. In many cases, 
however, particularly in the rib meristems of root and shoot and below 
wounds, cells that are relatively large and in which a vacuole occupies 
the bulk of the cell may continue to divide. In such cases the nucleus 
moves from near the wall to a position in the center of the cell, where it 
is held by strands of cytoplasm. Here it undergoes mitosis. The position 
where the cell plate, and later the cell wall, will form is usually indi- 
cated early bv a plate of cytoplasmic strands, the phragmosome, which 
extends across the cell and in which the nucleus is embedded (Sinnott 
and Bloch, 1941; Figs. 3-1, 3-2). The cell plate itself is laid down later 
by the phragmoplast, a group of fibers which are a continuation of the 
fiber system between the nuclei at telophase. This spreads across the cell, 
following the course of the phragmosome where the latter is present. In 



side view, as in the dividing cambium cells figured by Bailey (1920«), 
the phragmoplast appears in section as two spindles at the edge of the 
developing cell plate. In face view it looks like a cytoplasmic "halo." The 
phragmosome and the phragmoplast, and the function of each in cell 
division, have sometimes been confused. The difference is made clear by 
Esau ( 1953£>, her Fig. 3-10 ) . In some dividing vacuolate cells, such as 
those of the cambium, the phragmosome is either absent or has not 
been observed. 

Fig. 3-2. Phragmosomes in various cells, a-c, normal tissue; c, in face view showing 
anastomosing strands; d, e, mature cells near wound face beginning to divide. (From 
Sinnott and Bloch. ) 

The significance of the cytoplasm in cell division has been emphasized 
by Muhldorf ( 1951 ) . The general problems of cell division in plants are 
treated at length in Tischler's monumental book ( 1951 ) . 

The factors that determine whether a cell will divide or not are various 
and have been much discussed. The size of the cell itself is evidently 
one of these factors. In actively meristematic regions, the dividing cells 
are usually of about the same size. This means that each daughter cell, 
after division, enlarges until it reaches the size at which its mother cell 
divided, and then itself divides. After division ceases, the cells usually 
expand considerably. 

Size in dividing cells is by no means always constant, however. Wagner 

The Cellular Basis of Growth 27 

( 1937), studying the distribution of mitoses in root tips, found that these 
occurred not only in the small cells at the apex but in the progressively 
larger ones back from this until division ceased. Incidentally, he reported 
that in many cases dividing cells are not evenly distributed through the 
meristematic region but tend to occur in several waves, moving backward 
from the tip (Fig. 3-3). 

In developing cucurbit ovaries during the period of cell division there is 
a progressive increase in the size of the dividing cells in each region 
(epidermis to placental region), and this increase is greater in successive 
tissues from the epidermis inward ( Sinnott, 1939; Fig. 3-4 ) . The daughter 
cells from a division must therefore increase to a size somewhat greater 
than that at which their mother cell divided before they themselves di- 
vide again. The largest cells to divide were many times the volume of 
the smallest ones. 

Fig. 3-3. Changes in cell length in microns (lower line) and frequency of mitoses 
(upper line) at successive distances from the root tip (at left) in periblem of onion 
root. ( From Wagner. ) 

Another factor in division, emphasized especially by zoologists, is the 
ratio of nucleus to cytoplasm (the nucleoplasmic ratio). R. Hertwig 
( 1908 ) believed that, as a cell grows, the cytoplasm increases faster than 
the nucleus so that a tension is set up which is finally relieved by the 
division of the cell. This restores the equilibrium of nucleus and cyto- 
plasm, since presumably the size of the nucleus is at once restored. Popoff 
( 1908 ) was able to remove some of the cytoplasm from certain cells by 
micropipette and found that these divided more slowly than their un- 
treated sister cells, as one would expect on Hertwig's theory. There is 
little evidence from meristematic plant cells that a changing nucleo- 
plasmic ratio is significant in cell division, though perhaps it may be. In 
larger and vacuolate plant cells it is difficult to measure the cytoplasm 
since it is distributed in a thin layer lining the wall. The cytonuclear 
ratio (volume of cell to volume of nucleus), however, can be de- 
termined. In vacuolate cells of progressively larger size the volume of 



the nucleus tends to keep pace with the surface of the cell and thus 
perhaps with the volume of the cytoplasm if the thickness of the cyto- 
plasmic layer is constant (Trombetta, 1939; Fig. 3-5). This suggests a 
relationship which in earlier stages may have a bearing on cell division. 
An important element in growth and development is the rate at which 
cell division takes place. This is essentially a physiological problem and 

60 80 100 





Fig. 3-4. Relation of cell diameter to fruit diameter in Cucurbita pepo. TA, small- 
fruited race. CF, large pumpkin. In early development, cell size increases less rapidly 
than fruit size, showing that division is occurring. Later growth is by cell enlargement. 
Solid circles, cell diameter (in microns) at last division. Final cell diameter at end of 
each line. Lowest curve, epidermis; next higher, outer wall; next, middle wall; next, 
inner wall; uppermost, placental region. (From Sinnott.) 

involves various internal and external factors, some of which will later be 
discussed. There are certain techniques by which it can be measured, 
however, which are of importance for the student of morphogenesis. 

Root tips are especially favorable material for this. Brumfield (1942) 
recorded the rate of division in the apical meristems of small roots by 
photographing the surface cells at measured intervals of time, and this 
has since been done by others ( p. 78 ) . A method for measuring rate of in- 
crease in cell number by macerating the root meristem and counting the 

The Cellular Basis of Growth 


,3.J r 









2 - 28 

3- " MATURE 

Fig. 3-5. Logarithmic graph of relation of nuclear volume to cell volume in cells of 
stem hairs of tomato. Nucleus increases about two-thirds as fast as cell. ( From Trom- 
betta. ) 

numbers of cells at intervals has been developed by Brown and his stu- 
dents (p. 41). Erickson (1956) has analyzed mathematically the rate of 
division in certain root tips. 

In growing gourd fruits Sinnott (1942) determined cell number by 
dividing tissue volume by cell volume and found that increase in cell 
number takes place at approximately the same rate in epidermis, outer 
wall, inner wall, and placental region, regardless of the marked differences 
in cell size in the four. Cell division in the epidermis takes place at the 
rate necessary to maintain a constant cell size, and this tissue may thus 
serve in a sense as a pacemaker for division in the whole ovary primor- 
dium. Jahn (1941) has made a detailed study of the localization and 
degree of cell division and cell expansion in the epidermis of growing in- 
ternodes of Vicia faba. 


The size of plant cells is obviously an important element in growth, 
differentiation, and other morphogenetic problems. Cells are relatively 
small objects, presumably because the ratio of surface to volume, and thus 
the ease of exchange of material between a cell and its environment, is 
inversely proportional to its size. Cells with high metabolic rates tend to 
be very small, and large cells are relatively inactive. Rapidly dividing 
cells, for example, are much smaller than those of storage parenchyma. 
There is a wide range in cell size among various tissues of a plant. In meri- 
stematic regions they are often as small as 1,000 cu. microns or less but in 
pulp of watermelon may be almost a million times this volume. Stras- 

30 Growth 

burger ( 1893 ) believed that the nucleus has a certain "working sphere" 
and that this limits the size to which a particular cell will grow. 

Studies of comparative cell size have often been made but chiefly on 
mature cells. This problem, however, is one that must be attacked devel- 
opmentally. Two processes are involved in it: the rate and amount of 
growth or increase in size in a given region and the rate and duration of 
cell division there. Cell size is the result of the relationship between these. 
The faster the cells divide, in proportion to the total amount of growth, 
the smaller will they be, and vice versa. 

At the end of a cell division each daughter cell is about half the volume 
of its mother cell. At this time it begins to enlarge, and if it is in a meri- 
stematic region it will soon divide again. Where division rate is relatively 
rapid, the cells may divide before they have time to enlarge to the size of 
their mother cells, and cell volume decreases, as in some early embryos. 
Sometimes the new cells do not expand at all and a process of cleavage 
takes place, much as in the first stages of many animal embryos, where 
the egg is cut up into a mass of smaller and smaller cells. In plants such 
cleavage may be seen in endosperm formation; in the development of the 
female gametophyte within a megaspore; in the renewal of meristematic 
activity in large, mature cells during regeneration; and elsewhere. Where 
cell enlargement is relatively rapid, the cells will become larger than their 
mother cells before they divide again and cell size in the meristematic 
region will increase. Thus in growing gourd ovaries, where the tissues are 
still meristematic and the cells all start from a very small size, they gradu- 
ally enlarge, though not as fast as the ovary itself. In most roots, dividing 
cells increase in size with increasing distance from the root tip. 

In their detailed analysis of the growth of the oat coleoptile, Avery and 
Burkholder ( 1936 ) found that in the outer epidermis cell division ceased 
after the organ was first initiated, so that during all later growth cells here 
elongated greatly, sometimes becoming 150 times as long as at the begin- 
ning. The inner tissues, however, grew in part by cell multiplication until 
the coleoptile was 10 to 20 mm. long. There were progressively fewer 
divisions from the subepidermis inward. Thus at maturity the longest cells 
were in the outer epidermis, the shortest in the layer next below it, and 
the cells then increased in length toward the inner layers (Fig. 3-6). 

In some meristems, especially the vascular cambium, the size of divid- 
ing initials may undergo permanent increase. Sanio ( 1873 ) and others 
since his time have found that xylem cells in trees have different lengths 
at different distances from the center of the trunk or from the ground and 
that these differences are mainly established in the fusiform initials in the 
cambium from which the mature cells develop. 

Most increase in cell size, however, comes after the final division. In 
plants, as contrasted with animals, this increase may be very great. Usually 

The Cellular Basis of Growth 


it is related to the absorption of water and increase in size of the vacuole. 
Differences among tissues in mature cell size are very considerable and 
are one of the most important aspects of differentiation. In tissues where 
division persists relatively late there is little time for cell expansion before 
maturity is reached, and the cells remain small, as is often the case in the 
epidermis. Where division ceases early, as in storage parenchyma, the cells 
grow to a much greater size. In many axial structures there is a gradient 
from without inward, the cells becoming progressively larger toward the 
center either because of more rapid increase or earlier cessation of divi- 



-1 120 

li- 100 

£ 80 

z 60 


20 - 




10 20 30 




Fig. 3-6. Number of cells, lengthwise, in various cell layers of the oat coleoptile at six 
stages in its growth. There is evidently no division in the outer epidermis after the 
coleoptile begins its development, and division ceases early in the other layers, though 
there are differences among them in the frequency of division. ( From Avery and Burk- 
holder. ) 

sion. This is not always the case, for the epidermis may have larger cells 
than the other tissues, as Avery and Burkholder found in the Avena 

There is a question as to just where in the cycle of cell division growth 
actually occurs. Abele (1936) distinguishes between Teilungswachstum, 
growth during division itself, and Streckungswacristum, growth after divi- 
sion has ceased. There is certainly a considerable visible increase in size 
during prophase but not much more until telophase. Of course the dura- 
tion of these phases must be taken into account. It is probable that non- 
aqueous material increases at a constant rate throughout growth. Sinnott 
( 1945a ) found that in gourds there was no change in rate of growth of the 



ovary, as measured by gain in dry weight, at the time when cell division 
ceased and cell expansion began (Fig. 3-7). 

Cell Size and Organ Size. There are several important implications of 
the problem of cell size for morphogenesis. One is that of the relation be- 
tween the size of an organ or body and the size of the cells that compose 
it. Is a body large because its cells are large or because they are more 
numerous? This problem was discussed by Gregor Kraus in 1869 in con- 
nection with his work on structural changes during etiolation, but Sachs 
(1893) and his student Amelung (1893) seem to have been the first to 
attack it directly. Sachs called attention to the fact that the size of a cell 

Time in Days 

Fig. 3-7. Relation of cell division and cell enlargement to growth. Logarithm of ovary 
volume plotted against time for three races of Cucurbita differing in fruit size. The 
period between the vertical bars is that during which cell division ceases. To the left 
of it, growth is by cell division; to the right, by cell enlargement. Despite this change, 
the rate of growth at this period remains constant. ( From Sinnvtt. ) 

must be closely related to its physiological activity and that cells of a par- 
ticular tissue should thus be expected to be of about the same size. If this 
is so, size in a plant would be related to the number rather than the size 
of its cells. His measurements supported this conclusion. Amelung made 
a much larger series of measurements and found the same general result, 
although he observed a good many cases where cell sizes differed consid- 
erably between comparable plants or tissues. 

The problem is not quite as simple, however, as these early workers 
thought. It is true that most of the size differences between plants, espe- 
cially those in the indeterminate axial structures, result from differences 
in cell number. In determinate organs, on the other hand, notably in bulky 

The Cellular Basis of Growth 33 

structures such as tubers and fruits, the greater size is due to an increase 
in both the number and the size of their cells. Lehmann (1926) found a 
positive correlation between the size of a potato tuber and that of its cells. 
Since the increase in cell size was by no means proportional to that in 
tuber size, it was evident that large tubers have more as well as larger 
cells. The same relations are found in tomatoes (Houghtaling, 1935). A 
more detailed study of this problem, in large-fruited and small-fruited 
races of gourds, was made by Sinnott (1939; Fig. 3-4). Here cell size in- 
creases in the young ovary but much less rapidly than organ size, showing 
that cell division is taking place. During this period there is more increase 
in cell size in the larger races. The size at which the cells divide steadily 
increases. At about the time of flowering, however, division in most of the 
young fruit ceases, first in the central region and then progressively out- 
ward, so that nearly all later fruit growth is by cell expansion. In large- 
fruited races the period of cell division and that of cell expansion are both 
longer than in small-fruited ones, so that the greater size of the former is 
due to both more and larger cells. This general developmental pattern 
was found by Riley and Morrow ( 1942) in Iris ovaries and fruits, by W. H. 
Smith (1950) in apple fruits, by Ashby and Wangermann in leaves (p. 
210), and by others. In avocado fruits, however, cell division in the fruit 
wall continues to some extent until maturity ( Schroeder, 1953b ) . In gen- 
eral, larger fruit size results from an extension, so to speak, of all parts of 
the developmental history. 

In several genera Ullrich (1953) studied the relation between epider- 
mal cell size and leaf size in the series of successive leaves up the stem. 
He found that the size of the cells decreases steadily whereas that of the 
leaves increases for several nodes and then decreases. Thus there is no 
close relationship, at least in this tissue, between cell size and organ size. 
Under unfavorable conditions, however, both tend to decrease together. 
A somewhat similar variation in the correlation between cell and organ 
size has been reported in wheat (Nilson, Johnson, and Gardner, 1957). 

It is noteworthy that tissues differ considerably in the relation of the 
size of their cells to that of the organ of which they are a part. This rela- 
tion is usually closest in storage parenchyma and least in the epidermis. In 
general, as Sachs pointed out, cells that are physiologically important, like 
most of those in the leaf blade, are relatively constant in size and show 
little relation to the size of the organ. In mosses, unlike higher plants, cell 
size and leaf size are usually rather closely proportional to each other, 
cell number being much more constant. 

Size relationships sometimes extend below the level of the cell. The 
ratio of cell size to nuclear size has already been mentioned here. Both 
Budde ( 1923 ) and Schratz ( 1927 ) found a rather close correlation be- 
tween the total surface area of the plastids and the volume of the cell. 

34 Growth 

Mobius (1920), however, observed no relation, in 215 species, between 
chloroplast size and that of cells or organs. Irmak (1956) confirmed this. 
There are other complications in the problem of cell size, some of which 
are of morphogenetic significance. One involves the dwarfing of plant 
structures. This has been discussed in a number of early papers, among 
others by Gauchery (1899), Sierp (1913), Oehm (1924), Sinoto (1925), 
and Abbe (1936). It is generally agreed that where dwarfing is the result 
of unfavorable environmental conditions cell size is reduced, though not 
equally in all structures. A scanty water supply chiefly affects the second 
phase of cell enlargement in which considerable quantities of water nor- 
mally are absorbed. The problem is complicated by the fact that dwarfing 
is often the result of genetic as well as environmental influences. 

o 30 

- 20 

_■ 10 

38 DAYS 

•Ok * "j"* 


,xx xx "x * X 

, .•*•• xxx" r " " x .-d*arf 

HjflS't"'* x - tall 

° , 20 30 40 50 60 70 80 90 100 110 120 130 140 


Fig. 3-8. Relation of cell size to plant height. Length of successive cells along the 
terminal meristematic region of a dwarf race (above) and a tall one (below) of to- 
mato. Cells of the dwarf are somewhat longer because they attain maturity, and thus 
stop dividing, at an earlier stage. The tall plants have many more cells. (From Bind- 
loss. ) 

Genetic Factors. The relation of genetic factors to cell and body size is 
complex and will be discussed more fully in a later chapter (Chap. 19). 
Genetic analyses of size differences in plants have been made repeatedly 
but the histological effects of gene and chromosome differences are widely 
various. Most genetically large plants are so because of more rather than 
larger cells. Thus in the tall races of Lycopersicon and Zinnia studied by 
Bindloss (1942) there are many more cells, lengthwise, than in dwarf 
races, though the cells are somewhat shorter (Fig. 3-8). In other cases, 
however, cell size is involved. Thus the difference between large-leaved 
sugar beets and small-leaved vegetable beets is due chiefly to the greater 
cell size of the former. What is inherited here is evidently the amount of 
postmitotic expansion, for the meristem cells are the same size in both. 
Von Maltzahn ( 1957 ) found that the difference in size of vegetative struc- 
tures between large and small races of Cucurbita was related to both cell 
size and cell number. 

The Cellular Basis of Growth 35 

Some "giant" races, however, owe their large size to larger cells. The 
first instance of this was reported by Keeble (1912) in a mutant of 
Primula sinensis. Tischler (1918) found a similar case in the reed Phrag- 
mites communis, and here large cell size was accompanied by larger than 
normal chromosome size, a fact reported by others ( Schwanitz and Pirson, 
1955). Much more commonly, giant forms with large cells result from 
polyploidy ( p. 436 ) . The first case of this to be observed was Oenothera 
gigas of de Vries, which was found to be a tetraploid. Many similar ex- 
amples are now known. Tetraploids are not always giant in character, 
however, and many polyploid series in nature show no difference in body 
size or cell size. Sinnott and Franklin ( 1943 ) found that in young tetra- 
ploid gourd fruits the gigas condition, both as to ovary and cells, is pres- 
ent until after flowering but that later growth is reduced so that at matu- 
rity there are no great size differences between diploid and tetraploid (p. 
439). A diploid giant moss race reported by von Wettstein (p. 437) 
returned to normal size of cell and organ after a few years of vegetative 

The increased cell size due to polyploidy is not uniform but is consid- 
erably greater in some tissues than in others. In the diploid moss races 
produced by von Wettstein ( 1924 ) the ratio of size increase from the In 
to the 2nwas found to be characteristic for each race (p. 437). In general, 
the increase of organ size due to polyploidy is not as great as the increase 
in cell size, since cell number tends to be somewhat reduced. 

Increased cell size may also result from increased number or bulk of 
chromosomes (p. 445), quite apart from polyploidy (Navashin, 1931; Lor- 
beer, 1930), and from extra or accessory chromosomes (Randolph, 1941; 
Miintzing and Akdik, 1948). Particular chromosomes, when present in 
trisomies, have different effects on cell size (and on other characters), 
presumably because of the specific genes which they contain (p. 447). 
Geitler ( 1940 ) observed that chromosome volume was correlated with 
nuclear volume and that in some tissues the chromosomes were more than 
four times as large as in others. In species of four genera, Mrs. Sax ( 1938 ) 
found that cell size was correlated with the chromosome number of the 
species but that in three others there was no such correlation. Somatic 
polyploidy or polysomaty ( p. 441 ) is a factor of importance both for cell 
size and for differentiation. 

Cell size has been found to be inherited in a number of lower plants, 
as in yeast (Townsend and Lindegren, 1954). 

Heterosis is usually not related to an increase in cell size (Kostoff and 
Arutiunova, 1936). 

Cell Size and Position. Many workers have found a great variation in 
size among comparable cells in the same plant. Often this is not ran- 
dom but follows a certain pattern. The problem has been studied most 



intensively in the size (chiefly length) of cells in the xylem of woody 

Sanio ( 1872, 1873), working with pine, was the first to attack this prob- 
lem intensively. He came to several general conclusions, the more impor- 
tant of which are the following: 

1. Tracheids increase in length from the center of the trunk or branch 
toward the outside through a number of annual rings but finally reach a 
constant size. 

75 - 


• 65 









S 20 


= 16 


< 12 




Age in years 







825 I 5 

800 5 

Fig. 3-9. Relation of vessel diameter and fiber length to annual increment and age in 
trunk of Acer pseudoplatanus. ( From Desch. ) 

2. This final tracheid size increases from the base of the trunk upward 
to a maximum at a specific height and then decreases somewhat. 

3. The final size of tracheids in a branch is less than in the trunk but 
depends to some extent on the position of the branch. 

Sanio's "laws" have been confirmed by most workers since his time 
(Kribs, 1928, and others). Bailey and Shepard (1915), however, found 
that, although the length of tracheids increases from the pith outward in 

The Cellular Basis of Growth 37 

a number of conifers, it does not reach a constant size but fluctuates rather 
widely, falling and rising in cycles, perhaps climatic ones. Laing (1948), 
Bissett, Dadswell, and Wardrop (1951), and Bannan (1954) report that 
tracheid length tends to be less where the growth of the tree in diameter 
is rapid, presumably because of the more frequent pseudotransverse divi- 
sions of the cambial initials. Tracheid length is largely determined by 
length of the cambial initials ( Bailey, 1920fr ) . 

Dicotyledonous woods follow the same general pattern as conifers but 
the situation is more complex because of the greater variety of cell types 
(Desch, 1932; Fig. 3-9; Kaeiser and Stewart, 1955). Fibers may increase 
considerably in length over their cambial initials (Chattaway, 1936) but 
vessel segments do not (Chalk and Chattaway, 1935). In storied woods 
neither fibers nor parenchyma cells show any tendency to increase in 
length from the pith outward (Chalk, Marstrand, and Walsh, 1955). In 
all growth rings of pine, Echols (1955) finds a close correlation between 
the fibrillar angle in the cell wall and tracheid length. The subject of fiber 
length in woody plants has been reviewed by Spurr and Hyvarinen 


The division and enlargement of cells are essentially problems in the 
physiology of growth, a subject too extensive to discuss here in any detail. 
Much experimental work has been done, however, on certain aspects of 
cell growth which are of particular morphogenetic interest and which it 
will be profitable to review briefly. 

The role of growth substances (Chap. 18) is particularly important. 
Auxin was first recognized because of its stimulation of cell enlargement, 
and in many cases it also affects cell division. Other substances are effec- 
tive here. Jablonski and Skoog ( 1954 ) observed that the cells of tobacco 
pith tissue in culture did not divide even under optimum amounts of 
auxin unless extracts from vascular tissue, coconut milk, or certain other 
things were added. This suggested that a substance specific for cell divi- 
sion but different from auxin was here operative, and such substances, 
the kinins, are now recognized. Gibberellin especially influences cell size. 
Wound hormones induce division in many mature cells. The presence of 
vitamin C seems to be related to cell expansion (Reid, 1941). Lutman 
(1934) assembled a mass of data on the effects of various inorganic sub- 
stances on cell size. The stimulating and inhibiting influences of these vari- 
ous chemical factors are key problems in the physiology of develop- 

Metabolic factors are also important. Oxygen consumption is related to 
cell division (Beatty, 1946). Interesting observations here have been re- 



ported by Transeau ( 1916 ) for the seasonal distribution of various species 
of Spirogyra. These differ markedly in the size of their cells, those of the 
largest being about 150 times the volume of the smallest. The small-celled 
forms are the first to appear in the spring, when temperatures are low, 
and the larger-celled ones come on progressively as the season grows 
warmer. This presumably is because of the higher metabolic rate of the 
smaller cells which results from their greater ratio of surface to volume. 

Progressive physiological changes (p. 210) seem also to be involved. In 
successively higher leaves on the stem of Ipomoea, Ashby and Wanger- 
mann (1950; Fig. 3-10) found that the cells became smaller and suggest 



_ »o 
. .2 

* T. 


100 — 


o- o 

__ -to 



0-1 1-0 

Leaf area (cm. 2 , logarithmic scale) 


100 cm 

Fig. 3-10. Relation between area of epidermal cells and of leaf lamina in developing 
leaves of Ipomoea caerulea, plotted logarithmically. Solid circles, second leaves; 
crosses, fifth leaves; empty circles, eighth leaves. Early growth is chiefly by cell di- 
vision, since cell size increases little. Later growth is by cell enlargement since cells 
and lamina grow at the same rate. Compare with Fig. 3-4. Cell size becomes smaller 
in successive leaves. (From Ashby and Wangermann.) 

that this is symptomatic of a process of aging in the apical meristem. This 
problem of possible senescence in plants has other implications for cell 
size. Benedict ( 1915 ) presented evidence that in a number of vegetatively 
propagated plants, notably Vitis, cell size tends progressively to decrease 
with the age of the clone, a fact interpreted by him as the result of 
senescence. He believed that the "running out" of certain varieties was 
due to this cause, but it has now been shown that in many cases such a 
change is due to virus infection. Benedict's results have had some con- 
firmation, notably by Tellefsen ( 1922) and Bergamaschi ( 1926) in studies 
of cuttings from trees of different ages. Ensign (1921), however, found 

The Cellular Basis of Growth 39 

no correlation between age of plant and vein-islet area (and thus cell 


Even growth habit may involve differences in cell size, for H. B. Smith 
( 1927 ) reports that annual sweet clover has considerably larger cells than 
does the biennial race. 

Correlation (p. 95), a sufficiently vague term to describe certain phe- 
nomena about which we understand little, also affects cell size. In various 
growth compensations, removal of an organ results in greater growth of 
another one and often in larger cells there. In "topped" tobacco plants 
(where the terminal flower cluster has been removed) Avery (1934) 
found that the leaves on the upper third of the stalk grew larger and that 
much of this extra growth was due to increase in cell size (Fig. 3-11). 
Lindemuth (1904) removed and rooted mature begonia leaves and ob- 

Fig. 3-11. Effect of topping a tobacco plant. Portion of the vein network in the 
twentieth leaf from the tip. At left, untopped; at right, topped. The leaves from topped 
plants are larger and have more space between the veins because of increased cell 
size. ( From Avery. ) 

served that they then increased considerably in size, chiefly because of 
cell enlargement. While they were attached to the plant this presumably 
was prevented by "correlative inhibition." Similar results have been re- 
ported by others. In linden leaves, however, which had been induced to 
grow much larger than their normal size by the removal of other leaves, 
Ewart ( 1906 ) found the increase to be due chiefly to a larger number 
of cells. 

Factors within the cell itself are doubtless related to the onset of its 
division. Cytoplasmic viscosity generally rises in prophase, falls in meta- 
phase, and rises again in telophase. Mole-Bajer ( 1953 ) has explored the 
effect of artificially increased viscosity of cytoplasm in slowing down the 
rate of mitotic division. Gustafsson ( 1939 ) found that the difference 
between meiotic and mitotic division was related to the degree of hydra- 

40 Growth 

tion of the nucleus. The effect of dehydration in checking mitosis has also 
been reported by Mole-Bajer (1951). 

Osmotic concentration of the cell sap was observed by Becker ( 1931 ) 
to be inversely proportional to cell size and to number of chromosome 
sets in polyploid moss protonemata. 

Various external factors are important both for the division and the 
enlargement of cells. Light frequently tends to check division, and ultra- 
violet radiation may inhibit it. The hypothetical mitogenetic rays of Gur- 
witsch (1926) and his school were thought to stimulate mitosis. The effect 
of light on cell size has been emphasized by Straub (1948). Most of the 
elongation of etiolated plants (p. 309) is due to increase in cell length. 
Giese (1947) has reviewed 300 papers dealing with the effects of various 
kinds of radiation on the induction of nuclear division. 

Temperature, so important in many protoplasmic processes, has an 
effect on mitosis. P. C. Bailey (1954) has shown that in Trillium the 
maximum rate of cell division takes place at considerably lower tempera- 
tures than does the maximum rate of increase in root length. Burstrom 
(1956) reports that under higher temperatures the final cell length in 
roots is less because of the shorter period of cell elongation. 

Wagner (1936) found evidence that gravity influences mitosis. In root 
tips placed horizontally he observed that after about an hour there was a 
marked increase of mitoses on the upper side, so that the very tip of the 
root bent down. After 4 hours mitoses were equally distributed, and after 
10 to 12 hours they were more abundant on the lower side and the tip 
straightened out again. These changes were quite independent of the geo- 
tropic bending due to auxin and cell expansion, which was evident much 
farther back from the tip. When plants were grown on a clinostat. Brain 
(1939) observed in lupine seedlings that cells of the cortex, endodermis, 
and pith in the hypocotyl were larger than those of upright plants but 
that in the radicle they were smaller. 

In some cases pressure stimulates division, as can be seen in the cortical 
tissues through which a lateral root pushes its way ( Tschermak-Woess 
and Dolezal, 1953). 

Water is of marked influence in determining cell size. The amount of it 
available often determines how much a cell can expand (Thimann, 1951). 
Zalenski and others (p. 325) have observed that at successively higher 
levels on a plant the cells of the leaves are smaller, presumably because 
of their inability to become fully turgid while they were expanding. Water 
may have other effects. Funke (1937-1939) found that if some water 
plants are put into deep water their petioles elongate rapidly by cell ex- 
pansion, sometimes lengthening tenfold in 2 days. 

The role of the wall in cell growth has been much discussed. Does the 
wall merely stretch under the pressure of an expanding vacuole or does 

The Cellular Basis of Growth 41 

it grow independently of this? Burstrom ( 1957 and earlier papers ) pre- 
sents evidence that cell elongation is not primarily a matter of water 
uptake but is due to growth of the cell wall. He believes that this occurs 
in two steps, the first a plastic stretching of the wall and the second the 
production and deposition of new wall material. Auxin promotes the first 
but probably inhibits the second. Others believe that auxin, known to 
stimulate cell enlargement, directly increases the plasticity of the wall and 
thus its extensibility (Heyn, 1940). This view has found recent support 
(p. 412). 

It obviously is necessary to know just how the wall grows and particu- 
larly whether this is by apposition of new material on its inner face or by 
intussusception throughout. Green (1958), using techniques for measur- 
ing radioactivity, treated elongating Nitelh cells with tritium (H H ) and 
found by test that the inner part of the wall became radioactive. The 
outer portion, which was not, grew thinner as the cell lengthened, thus 
suggesting that new cell wall material was being laid down only on the 
inside and not throughout. 

Rate of wall thickening sometimes has a direct effect on plant size. Thus 
in a dwarf mutant of Aquilegia (Anderson and Abbe, 1933) this trait was 
found to be due to the precocious thickening of its cell walls (p. 426). 

The important problem of the relation of deoxyribonucleic acid to 
cell division has often been investigated. Grundmann and Marquardt 
( 1953) determined the content of DNA in successive phases in the mitotic 
cycle of the nuclei of periblem cells of the root tip of Vicia. This increases 
steadily throughout the interphase. It is reduced at telophase since it is 
roughly proportional to nuclear volume. 

Brown and his students have used various modern techniques for a 
study of the problems of the multiplication and growth of cells. Brown 
and Rickless (1949), for example, cut off Cucurbita root tips of equal 
length (1.6 mm.) and grew them in culture for 3 days, taking samples 
every 12 hours. These tips were macerated, and in a haemacytometer the 
total number of cells and the number of nonvacuolate cells were counted. 
From these counts, together with a measurement of the length of each 
root examined, it was possible to determine the rate of cell division and 
the index of extension ( ratio of root length to number of vacuolate cells ) . 
This method is subject to a number of errors, particularly from the as- 
sumption that only the nonvacuolate cells were dividing. However, it 
gave consistent results, and these were in general agreement with the 
more laborious method of measuring cell size and volume from microtome 
sections. The authors found that there was no division in the absence of 
sugar in the culture medium and that the rate of division increased with 
the addition of sugar and inorganic salts and even more with the addi- 
tion of yeast extract. At 15° C the rate of division was higher than at 5 or 



25°. The greatest increase in cell size was with sugar and mineral salts. 
Yeast extract tended to decrease extension. 

Brown and Wightman ( 1952 ) grew root tips of pea 3.0, 6.0, and 10.0 
mm. long in sterile culture and found that the peak rate of division occurs 
later and that its peak value is greater the shorter the initial tip. They 
conclude that cell division in the meristem depends partly on synthesis of 
metabolites there and partly on a supply of metabolites from more mature 
regions of the root. 

I 3 5 7 9 II 


Fig. 3-12. Changes in protein (circles) and water content (crosses) of bean-root cells 
at increasing distances from the root apex. ( From Brown and Robinson. ) 

A basic problem in this field is that of protein synthesis. Brown and 
Broadbent ( 1950) sliced a series of root tips into segments 0.2 to 0.8 mm. 
thick and in each successive section determined the number of cells, the 
protein content, the dry weight, and the amount of respiration. During 
development from the meristematic to the fully extended state the average 
cell volume increased thirtyfold and there was an increase in protein con- 
tent and in respiration (Fig. 3-12). 

Genetic factors are concerned in cell activities in many ways. Beadle 
( 1931 ) , for example, found a gene for supernumerary divisions in maize; 

The Cellular Basis of Growth 43 

Moewus (1951), one for cell division in Protosiphon which is linked to 
sex manifestation; and Nickerson and Chung (1954), one in yeast that 
seems to block the sulphydryl mechanism of division. Dorries-Ruger 
(1929) grew protonemata from the spores of plants produced by various 
combinations of genome and plasmon in mosses, among races developed 
in Wettstein's laboratory. She cut off and cultured the tip cells of these 
protonemata and recorded marked differences in the rate of cell division 
in the filaments growing from them, thus comparing the effects of dif- 
ferent genotypes under the same environment. 


Cell size and cell number are important elements in growth and differ- 
entiation, but the problem of form is primarily dependent not on these 
factors but on the relative directions in which growth occurs. These, in 
turn, are closely related to the planes of cell division in the developing 
tissues. Whatever determines the position of the new cell wall between 
two daughter cells will determine the direction in which these cells ex- 
pand, since this direction will normally be at right angles to the new wall. 
At the time of cell division, therefore, the direction of growth in this par- 
ticular region of the meristem is determined. If plant cells could change 
their relative positions, as is possible in many animal tissues, the deciding 
factor in the direction of growth would often be cell movement rather 
than plane of cell division. 

The position of the cell plate at telophase, and thus of the new cell wall, 
follows the position which the equatorial plate of chromosomes finally 
assumes at metaphase. The mitotic figure may roll around somewhat be- 
fore it settles down to a permanent position, but there must be something 
that determines that position. This raises the question as to whether the 
plane of division is controlled by whatever decides the final orientation 
of the mitotic figure or whether this, in turn, is itself determined by other 
factors. That the latter may be the case is suggested by the way in which 
vacuolate cells divide. In such cells a series of cytoplasmic strands forms 
a loose diaphragm, the phragmosome, across the cell, and in the middle 
of this the nucleus is supported. In tissues where the plane of division 
can be predicted, observation shows that the position of this diaphragm 
is the one which the future cell wall will occupy. The diaphragm is laid 
down considerably before the nucleus enters metaphase, and the meta- 
phase plate of chromosomes may not at first lie parallel to the diaphragm, 
though it finally does. This seems to indicate that the plane of division is 
determined early and for the cell as a whole rather than by factors acting 
on the mitotic figure alone. There is also evidence that the mother cell, 
before division, begins to elongate at right angles to the direction in 

44 Growth 

which it will divide. All this raises the fundamental question as to whether 
morphogenetic factors operate directly on each dividing cell or whether 
relative directions of growth, and thus form, are determined by factors 
affecting the entire growing organ, the whole mass of living stuff, and that 
the degree and manner in which this is cut up into cells are a secondary 
result. This is simply another aspect of the main problem raised by the 
cell theory. 

Factors Determining the Plane of Cell Division. Many suggestions have 
been made as to the factors that determine the position of a new cell wall. 
Years ago Hofmeister ( 1863 ) stated the general rule which bears his 
name, that growth precedes division and that the new wall is at right 
angles to the long axis of the mother cell. There are many cases, espe- 
cially in parenchymatous tissue, where this rule holds, but frequent ex- 
ceptions to it occur in which the new wall is parallel to the long axis. An 
extreme example of this is the longitudinal division of very long cambial 
initials. Sachs (1878) noted that in most dividing cells the new wall meets 
the old one at an angle of 90°, even though this requires that the new 
wall be curved, and proposed this as a rule for cell division. 

About a decade later a number of biologists were impressed by the 
close resemblance between many cell configurations and masses of soap 
bubbles. The behavior of molecules in liquids and the principle of surface 
tension were then being worked out by physicists. One of the implications 
of surface tension is that, because of molecular forces acting at their sur- 
faces, liquids tend to pull themselves into forms with the smallest possible 
surface area. This is why drops of liquid, for example, or soap bubbles are 
spherical. The principle of least surfaces was applied to liquid film sys- 
tems by the physicist Plateau (1873), who showed that in a mass of 
bubbles the partition walls in every case arrange themselves so that they 
have the least possible area. He also observed that where walls intersect 
there are only three at a given point and that the angles between them 
tend to be 120°, the point at which surface forces are in equilibrium. 

The biologists Berthold (1886) and Errera (1888) applied this prin- 
ciple to young cell walls, assuming that these walls in the beginning are 
essentially weightless liquid films. The rather striking resemblance often 
observed between a mass of cells and a mass of bubbles on this assump- 
tion is easy to understand. Some interesting implications of the principle 
of least surfaces for the problem of cell division have been developed by 
D'Arcy Thompson (1942). 

Only a few examples need be cited here. If, in a cubical box, the sides 
of which are liquid films, a film partition extends across the middle, the 
partition will be flat. If it is gradually moved toward one of the sides so 
that the two "cells" become more and more unequal in size, it will sud- 
denly shift to a position across a corner of the box and, as seen in section, 

The Cellular Basis of Growth 


Fig. 3-13. Hypothetical cube of film 
with a film partition moving from left 
to right across it. When a position is 
reached 31.8 per cent of the distance 
from the right-hand side, this partition 
slips into the corner and becomes 
curved, as shown. ( From D'Arcy 
Thompson. ) 

will now be curved instead of straight as it was before (Fig. 3-13). The 
point where this shift occurs is the point where the wall, now ( in section ) 
a quarter of the circumference of a circle, has the same length that the 
flat partition wall had, for the wall will have the least possible area that 
will enclose the volume of the smaller "cell," the latter now being part 
of a cylinder. If this smaller cell is then made still smaller, the wall that 
separates it from the larger one will continue to be curved and to be less 
than any other wall area that could bound the volume of the smaller cell. 
Just where the point of shift from flat wall to curved will occur can be 
calculated by determining the point at which (before the shift) the length 
of a curved wall (a quarter of the circumference of a circle) across the 
corner will be the same as that of the flat partition wall. Both will enclose 
the same area. If it is assumed that each side of the cube, and thus the 
length of the flat partition, equals 1, then 


= 1 %rr = 4 and r = ^- = 0.637 


The area of a quarter circle with this radius is tt( 0.637 ) 2 /4 = 0.318. This 
also is the area of the smaller rectangular cell just before the shift. Thus 
the distance from the partition wall to the side wall, when it shifts from 
straight to curved, is 0.318 of the diameter of the cube. Experiments with 
films essentially confirm this theoretical expectation. In cases of unequal 
division of actual cells, such as the formation of companion cells in sieve 
tubes, the new cell is usually cut out of a corner of the old one, as this 
theory of least surfaces requires. 

Many dividing cells in plants and animals are spherical, and here the 



application of the liquid film theory is particularly interesting. In the di- 
vision of an egg into two equal cells, for example, the position of the 
wall between the two daughter cells, if they behave like soap bubbles, 
can be determined. This new wall should form an angle of 120° with the 
tangent to the circumference of each daughter cell at the point where 
these meet the partition wall, since this is the position where the sur- 
face forces will be in equilibrium and where the film system thus is stable. 
It is obvious geometrically that this wall is in such a position that the 
distance between the centers of the two new cells is equal to their radii 
(Fig. 3-14). When a single spherical cell, such as an egg or an algal cell, 
divides thus equally, the position of the two daughter cells relative to 
each other is approximately what this theory demands. 

Fig. 3-14. Stable partition and walls of minimum surface assumed by two equal 
bubbles which are in contact. Angles OPQ and OPR are 120°. The distance between 
the centers equals the radii. ( From D'Arcy Thompson. ) 

Where such a divided bubble divides again but now by a partition at 
right angles to the plane of the first one, these two walls usually do not 
meet at an angle of 90° but there is a readjustment in the film system so 
that they meet at 120°, the stable position. Arrangements like that of 
Fig. 3-15d may thus result, which resembles a group of actual cells. Any- 
one familiar with cellular structure and who draws a bit of it comes almost 
instinctively to make the cell walls intersect at angles of about 120°, 
much as they would if they were liquid films. 

Where a spherical bubble is divided unequally, the curvature of the 
partition wall can be calculated. Since the pressure is inversely propor- 
tional to the radius of curvature, a small bubble pulls itself together, so 
to speak, more strongly than a larger one. Thus P = 1/R, where P is the 
pressure and R the radius. The pressure that determines the radius of the 

The Cellular Basis of Growth 47 

partition wall between unequal bubbles is thus the difference between 
the pressure of the smaller bubble and that of the larger one. If R equals 
the radius of the partition wall, / that of the smaller bubble, and r that 
of the larger one, then 1/R = 1// — 1/r, or R = rr7(r - r'). In other 
words, the radius of the partition wall is the product of the two bubble 
radii divided bv their difference. If two bubbles have radii of 3 and 5, for 
example, that of the partition wall will be 7.5. In spherical cells of un- 
equal size which are dividing, the new wall does tend to have this theo- 
retical radius. 

Where the dividing cell has a relatively firm wall, however, as in micro- 
spores within which a small prothallial cell is cut off, the situation is 
different, since only the new dividing wall now acts as a liquid film. It 
will be curved and will occupy a position such that it intersects the old 
wall at the stable position of 90°. If the linear distance between the two 

Fig. 3-15. Plate of eight cells (or 
bubbles) assuming a position of equi- 
librium where cell surfaces are of mini- 
mum area. (From D'Arcy Thompson.) 

points of intersection (as seen in section) and the radius of the large 
cell are known, the radius of curvature of the new wall can readily be 

There are other cases of division walls, notably in "rib" meristems 
where the cells are in parallel rows and growth is strongly polar, which 
may also be interpreted on the liquid-film theory even though the re- 
semblance to a bubble system is much less close. In such rows of cells 
it can be observed that the new cross walls, even in the phragmosome 
stage, always tend to avoid a position that would put them opposite a 
cross wall in an adjacent row ( Fig. 3-16 ) and would thus bring four walls 
together, unstably, at a point. The walls are always "staggered," like 
bricks in a wall. This often prevents a new wall taking its natural posi- 
tion, which would divide the cell into two equal parts. The angles be- 
tween the walls are larger than 90° but do not reach the theoretical 120°. 
It may be that surface forces are operative in pulling the new wall away 



from the intersection point with an old one and thus tending to form an 
angle of 120°, even though the longitudinal walls, which are relatively 
firm, remain essentially straight and the theoretical angle cannot be 


The theory that the position and curvature of dividing cells are what 
they would be if the walls were liquid films lends itself to some interest- 
ing geometrical analyses, for which the reader is referred to Thompson's 
book. Giesenhagen's work (1905, 1909) also has a discussion of the 
theory and its applications. Various experiments with actual liquid films 
have been reported by van Iterson and Meeuse (1942), and Matzke 
(1946) has discussed the role of surface forces in determining cell shape. 
There is no doubt that the configurations of the cells in an actively grow- 
ing mass often do resemble a system of bubbles, for there are usually no 
more than three walls intersecting at a point and the angles between them 
tend to approximate 120°. The young walls are at least semiliquid, so that 

-f # 

Fig. 3-16. Walls in dividing cells (as shown by position of phragmosome) tend to avoid 
continuity with adjacent partition walls. ( From Sinnott and Block. ) 

surface forces are doubtless operative to some extent in determining their 
position. In any morphogenetic analysis the least-surface theory therefore 
must certainly be taken into account. It greatly oversimplifies the prob- 
lem, however, and fails to explain some facts with which the student of 
plant development is confronted. Among the chief objections to it are the 

1. The theory in its simplest form is applicable only to weightless 
liquid films, and young cell walls obviously are not such, though they 
mav approach this condition. To account for their position the theory 
would require correction. 

2. Many division walls are formed in positions different from those 
which the theory demands. Often the new walls are parallel to the longer 
axis of the cell instead of at right angles to it. The most extreme case of 
this occurs in dividing fusiform initials at the cambium which are 50 times 
or more as long as wide but which nevertheless divide lengthwise. 

The Cellular Basis of Growth 49 

3. The early wall formed by the cell plate, and certainly the phragmo- 
some which precedes it in vacuolate cells, are not at first continuous films 
and would thus not follow the law of least surfaces. 

4. In many cases, as often in the unequal division that cuts off a 
stomatal mother cell, the new wall is at first straight instead of curved 
and becomes curved only later, as the turgor of the cell increases. 

5. Frequently, as in growing cork layers, the new division wall is laid 
down exactly opposite a partition wall in an adjacent cell so that four 
walls do come together at a point (p. 195). This also happens in tissue 
which is to form aerenchyma and in which the cells are in regular rows 
with cross walls opposite. Here, however, at the point where the four 
walls meet, a small air space (which later may enlarge) is commonly 
formed by the pulling apart of the walls so that the wall angles do tend 
to reach the theoretical 120°. 

6. In a system of film bubbles increasing in number by the formation 
of new walls, the equilibrium least-surface configuration is reached by a 
shifting of the wall positions within the film system. This involves some 
gliding or sliding of the bubbles in relation to each other. Such a change 
could happen in animal tissues where the cells are free to move about, at 
least to some degree, but would be impossible in most plant tissues, where 
they are cemented to one another. 

For these reasons it is clear that the theory of surface forces alone is by 
no means sufficient to explain all the facts as to the position of new cell 
walls and the planes of cell division. Other physical factors are doubtless 
involved in determining these events. Among them pressure is important. 
Kny ( 1902 ) found that pressure applied to a dividing cell forced the 
mitotic figure into a position in which its long axis was oriented at right 
angles to the direction of the pressure, and the new wall consequently 
was parallel to this direction. This fact, incidentally, makes an important 
contribution to our knowledge of the character of the cytoplasm, at least 
at this time in the history of the cell. If the cytoplasm were essentially 
fluid, pressure from without should not change the orientation of struc- 
tures in it but would do so if the cytoplasm had a structural framework. 
Other evidence for the conclusion that walls are formed parallel to pres- 
sure on the cell can be found in the cortex of the young stems of many 
woody plants. Here the cells tend to be elongated tangentially, presum- 
ably because of the pressure exerted by the expansion of the vascular 
cylinder below through cambial activity. If these cells divide again, 
radial walls, parallel to the direction of cambium pressure, are often to be 
seen. As Kny points out, however, in cambial cells, which are presumably 
under radial pressure, division is chiefly periclinal (at right angles to the 
pressure) instead of anticlinal. This he attributes to "inner factors." In the 



case of radial divisions in the phellogen, Bouygues (1930) concludes that 
pressure is not a factor. 

The plane of division is evidently related to the polarity of the cell and 
is further discussed under this topic (p. 131 ). It has been studied particu- 
larly in the egg of Fucus. Here centrifugal force, light, electricity, and 
gradients in concentration of various substances have been found to affect 
this plane. 

In certain colonial blue-green algae and flagellates, in pollen mother 
cells, and in some other cases where division in all the cells is simul- 
taneous and in the same plane, the planes of each successive division tend 








Fig. 3-17. Distribution of angles between mitotic spindles and longitudinal axis of the 
ovary in an elongate type of cucurbit fruit ( above ) and an isodiametric one ( below ) . 
There is evidently a higher proportion of divisions nearly at right angles to the axis 
(spindles with low angles) in the former. In the latter, divisions are approximately 
equal at all angles. (From Sinnott.) 

to be at right angles to each other so that a regular pattern of cells in 
twos, fours, eights, sixteens, and so on, all in one plane, is produced. Be- 
tween divisions the cells tend to grow but not enough to make them 
isodiametric, so that the next division is at right angles to the longer axis 
of the cell, as it would be in a least-surface configuration (Geitler, 
1951). Division in three planes sometimes occurs, producing cubical 

In many instances there is no obvious explanation for the particular 
plane in which a cell divides, and we are forced to attribute this to genetic 

The Cellular Basis of Growth 51 

factors. Steward (1958) finds that in cells freely suspended in culture the 
planes of division are highly irregular and unpredictable, since such cells 
are not subject to the organizing restraints that are operative in the nor- 
mal plant body. 

The forms of most plant structures are presumably related to the planes 
in which their constituent cells divide. In a few cases this relationship has 
been demonstrated. Thus in ovary primordia of elongate gourds such as 
Trichosanthes and the "club" variety of Lageruiria, Sinnott ( 1944 ) meas- 
ured in the growing ovary the angles between the mitotic spindles and 
the longitudinal axis of the primordium and found that there were many 






Fig. 3-18. Angles between mitotic spindle and ovary axis in metaphases, anaphases, 
and telophases in Trichosanthes, the snake gourd, where the great preponderance of 
divisions are transverse and thus predictable as to orientation. The mitotic figure evi- 
dently becomes more stabilized in direction as mitosis proceeds. ( From Sinnott. ) 

52 Growth 

more of them parallel to the primordium axis, in the direction of growth 
in length, than at right angles to it, in the direction of growth in width. 
In isodiametric ovaries, however, spindle angles were almost equally 
distributed between and 90° to the axis (Fig. 3-17). 

The question arises as to just what determines the plane of cell division 
in cases like this. It cannot be simply the orientation of the mitotic spin- 
dle, for this can be shown to change during mitosis. Thus in the fruit of 
Trichosanthes, which is very long and narrow, practically all the divisions 
are transverse to the long axis and their orientation can thus be predicted. 
The spindles in metaphase, however, are by no means all parallel to the 
axis but vary considerably. In anaphase the variation is much less, and 
in telophase the cell plates are almost all transverse ( Fig. 3-18 ) . Evidently 
the spindle rolls about somewhat during mitosis (as it has been seen to 
do in living material of other forms ) but finally settles into position. What 
this position will be seems to be determined by the cytoplasmic body 
since, in vacuolate cells, the phragmosome is formed at prophase in the 
position of the final cell wall (p. 25). 

Something certainly controls not only the plane of cell division but the 
distribution of divisions and the amount and character of cell expansion. 
Whatever this may prove to be, it is concerned with the origin of organic 
form. If one looks at a section through a young and growing plant struc- 
ture, such as an ovary primordium, he sees a mass of cells of various 
shapes and dividing in many planes. Here chaos seems to reign. When he 
observes how such a structure develops, however, and finds that it is 
growing in a very precise fashion, each dimension in step with all the 
others, he comes to realize that this is not the seat of chaos but of an 
organizing control so orderly that a specific organic form is produced. 
This realization is one of the most revealing experiences a biologist can 
have and poses for him the major problem that his science has to face. 


One of the simplest manifestations of organic form is in the shape of 
individual cells. This obviously involves plane of cell division, cell size, 
polarity, microstructure of the wall, genetic constitution, and other fac- 

Since the cell, at least at first, is a fluid system, its natural shape, other 
things being equal, is that of a sphere, for this has the least surface in 
proportion to its volume. Most cells, however, are parts of tissues and 
thus are closely packed against neighboring cells on all sides. This re- 
sults in a modification of the basic spherical shape to that of a polyhedron 
with flattened sides, each representing a plane of contact with an adja- 
cent cell. How many faces should such a cell have, and what sorts of 

The Cellular Basis of Growth 53 

polygons should these faces be? At first it was believed that cells were 
12-sided figures, since when spheres are stacked together like cannon 
balls, each touches 12 others. Lord Kelvin (1887, 1894), approaching the 
problem mathematically, showed that when space is divided into similar 
units, each with a minimum area of partition and with stable angles, 
each unit will be a 14-faced figure or a tetrakaidecahedron and that eight 
of its faces will be hexagons and six will be squares. This, he thought, 
was what the shape of a cell in pith or similar tissues theoretically 
should be. F. T. Lewis ( 1923 and others ) found that the average number 
of faces in such cells was, indeed, close to 14 but that only very rarely 
did a cell with this number of faces show eight hexagons and six squares. 

This problem has been studied with particular care by Matzke and 
his students and reported in a series of papers. The results have been 
briefly reviewed by Matzke (1950), who cites the more important papers 
from his laboratory. The general conclusion is that parenchyma cells do 
tend to have 14 sides but that "ideal" ones, conforming to Lord Kelvin's 
rule, occur very infrequently. Matzke points out that many factors other 
than mathematically ideal space-filling are involved in determining cell 
shape, among them pressure, surface forces, differences in cell size, direc- 
tion of cell division, unequal growth, and genetic constitution. The prob- 
lem is being attacked developmentally by an analysis of cell shapes at the 
meristem (Matzke and Duffy, 1956). In dividing cells, the number of 
faces here rises to about 17 and in daughter cells drops at first to be- 
tween 12 and 13. The total cell population has an average number of 
about 14 faces. 

In more specialized tissues there is a wide variety of cell shapes. 
Palisade cells are elongated at right angles to the leaf surface. Most cells 
of the vascular and conducting tissues are elongated parallel to the axis. 
Hairs and glandular cells have many forms. Some cells expand equally 
on all sides. Others, like root hairs, grow only at one point. Still others, 
such as the more fantastic sclereids, have many growing regions ( Foster, 
1955, and others ) . Galston, Baker, and King ( 1953 ) found that benzimida- 
zole promotes the transverse as opposed to the longitudinal extension of 
cortex cells in the pea epicotyl. Doubtless the polarity, or polarities, of the 
cell and the plasticity, elasticity, and microstructure of its walls are in- 
volved in shape differences. 

Tenopyr ( 1918 ) found that in leaves of different shapes the shapes and 
sizes of cells were constant. Riidiger ( 1952 ) , however, observed that in 
tetraploid plants the subepidermal cells of leaves, hypocotyls, and other 
organs were not only absolutely but relatively wider than in diploids, a 
fact which he relates to the greater relative width found in most tetra- 
ploid organs as compared with diploid ones. 

Even in microorganisms where the cells are free from contact with 

54 Growth 

others, they often display shapes by no means spherical. Many unicellular 
green algae and the simpler fungi are examples of this. Von Hofsten and 
von Hofsten (1958) have explored the effect of various factors on cell 
shape and thus on vegetative characters in the ascomycete Ophiostoma. 
In the development of cell shape genetic factors are doubtless impor- 
tant, and these appear to control cytoplasmic patterns, wall differences, 
and other factors. Cell shape is one aspect of the more general problem 
of differentiation. 

Much of morphogenetic significance can be learned from a study of 
individual cells. A knowledge of their relations to each other, and par- 
ticularly of the way in which they form cell aggregates, is still more im- 
portant. This involves the general problem of meristematic activity, the 
subject of the next chapter. 



In many of the morphogenetic problems which they present, plants and 
animals are very similar. The fundamental physiological differences that 
distinguish these two groups of living things, however, produce a number 
of developmental differences between them. Among these, that in method 
of growth is conspicuous. Because of their ability to synthesize food from 
inorganic substances, plants have developed, in all forms but the smallest 
and simplest, a body which is nonmotile and anchored to the soil or other 
substratum. This doubtless resulted, during the course of evolution, from 
the fact that motility in a plant is not necessary for obtaining food, as it is 
in animals. 

The motility of animals requires that their skeletons be jointed and 
the rest of their bodies relatively soft and plastic. Plants, however, gain 
the necessary rigidity not by a specially differentiated skeleton but by a 
thickening of the walls of most of the cells. This is especially conspicuous 
in the fibrovascular system of higher plants but it occurs in other tissues. 
The plant cell wall, because cellulose is characteristically deposited in it, 
is a much firmer structure than the rather tenuous membrane which sur- 
rounds typical animal cells. As a result, plant tissues themselves are also 
firmer, save in exceptional cases such as certain short-lived floral parts. 
As a consequence of this distinctive character, most plant cells, as soon 
as their final size is reached, become locked up, so to speak, in a firm box 
of cellulose. Such a cell ordinarily does not divide further, or if it does 
its daughter cells cannot expand, so that mature plant tissue usually grows 
no more. In almost every part of the soft-celled animal body, on the 
contrary, growth occurs not only during development but in the restora- 
tion and repair of tissues throughout the life of the individual. It should 
be remembered, however, that under certain conditions a plant cell or a 
group of cells may become embryonic again and begin to divide (p. 232), 
setting up a new growing region. There is no doubt that most cells— per- 
haps all— are potentially able to do this. What prevents it is not simply 
mechanical confinement by the wall but so-called correlative factors that 
limit each cell to the development appropriate for its particular place in 
the organism. 


56 Growth 

If a plant is to grow, this must be accomplished by allowing some of 
its cells to escape the general fate and remain capable of division, pro- 
gressively forming new tissues but preserving a remnant that persists in 
a perpetually embryonic condition. The plant body grows in size by the 
activity of such localized growing points or regions, the meristems, which 
are centers of cell division and cell expansion. 

The axis of the plant grows in length by a meristem at the apex of stem 
and root, and in width by a sheath of lateral meristem, or cambium. De- 
terminate organs such as leaves, however, rarely have sharply localized 
meristems but enlarge throughout much of their extent, as an animal body 
does, until growth ceases. Meristems are obviously of much interest to 
the student of plant morphogenesis. They provide, in a sense, a con- 
tinuous embryology for the plant and offer an important point of attack 
on the problems of plant development. 


In the simplest plants, the lower algae and fungi, growth is hardly 
localized at all. Cells capable of division are either present throughout 
the plant body or in considerable portions of it, and nothing which might 
be called a meristem is to be found. In Spirogyra, for example, growth 
in length of the filament is produced by cell division almost anywhere in 
it. In such a membranous type as Ulva, growth results from divisions at 
right angles to the surface throughout most of its area. In coenocytic 
forms, the whole thallus enlarges, and growth is not related to cell di- 
vision at all. 

In some of the simpler filamentous brown algae, however, growth in 
length is limited to the tip of the filament, which is occupied by a single 
large cell. This divides transversely, and a series of daughter cells is thus 
produced from its basal face. They and their daughter cells divide a few 
more times, but division finally ceases. The only permanently embryonic 
cell is the apical one, which thus dominates the development of the plant 
body. Branches originate by the lengthwise division of this apical cell 
(Fig. 4-1). 

In types like Fucus, with larger and more complex plant bodies, growth 
still originates by the activity of an apical cell, which occupies the base 
of a terminal notch in the thallus. This cell cuts off daughter cells from 
its two lateral faces. From these and their descendants are formed the i 
various tissues of the thallus. A fern prothallium grows in much the same 
way, developing under the control of the meristematic region in the 
notch. This control may be relaxed, however, and almost any cell in the 
structure may begin to divide. Many prothallia never form the typical 
heart-shaped structure. 



Fig. 4-1. Terminal portion of the alga Sphacelaria, 
showing how thallus is produced by activity of the 
apical cell and its descendants and how a branch 
originates. ( From Haberlandt. ) 

Throughout the bryophytes, the ferns, the horsetails, and many of the 
lycopods, growth of the plant body is governed by the activity of apical 
cells, one at the apex of the shoot and the other at the apex of the root. 
These cells are usually pyramidal with the base of the pyramid outward, 
and division takes place parallel to the three inner sides of the pyramid. 
Most growth of tissues results from the later division of these daughter 
cells and their descendants, but growth seems to be initiated and domi- 
nated by the apical cell (Figs. 4-2, 4-3). It is not clear, however, just 
what the function of the apical cell is. Wetmore ( communication to the 
author ) states that he has very rarely seen an apical cell dividing and he 
suggests that these cells may function as do the groups of large cells just 
below the apex of root and shoot in angiosperms, which are thought to 
be centers of metabolic activity. Most of the actual cell division in the 
meristems of these lower vascular plants takes place in the cells just be- 

Fig. 4-2. Longitudinal section through apex of a fern root, showing origin of tissues 
from the apical cell. ( From Sacfis. ) 



side or below the apical cell. In ferns, lycopods, and horsetails this con- 
siderable body of embryonic cells at the apex of the axis somewhat re- 
sembles the terminal meristems of higher plants. 

Among bryophytes, the origin and arrangement of the leaves and the 
structure of the various tissues can usually be traced back to precise 
divisions of the apical cell and its daughter cells so that there is a very 
definite pattern of cell lineage in the plant body. This is especially 
diagrammatic in such a form as Sphagnum, where the two markedly 
different types of cells in the leaves can be seen to originate in differen- 
tial cell divisions. Such a precise cell lineage is less conspicuous in higher 

Fig. 4-3. Longitudinal section of shoot apex Fig. 4-4. Selaginella wildenovii. Me- 
of Equisetum, showing apical cell and its dian longitudinal section of young 
derivatives. ( From Golub and Wetmore. ) shoot, showing apical cell and its de- 

rivatives. ( From Barclay. ) 

forms but often can still be traced even there. In one species of 
Selaginella (Barclay, 1931), for example, the derivation of the epidermis, 
cortex, pericycle, endodermis, and vascular cylinder can be traced back 
to direct descendants of the apical cell (Fig. 4-4). 

Although the distribution of the leaves and the general organization 
of the shoot are determined in many of the lower forms by the activity 
of the apical cell and the arrangement of its derivatives, Golub and 
Wetmore (1948) found that in Equisetum there is no relation be- 
tween the cellular pattern of the apex and that of the mature axis derived 
from it. 

Meristems 59 

In a few of the lower vascular plants, notably Lycopodium, no single 
apical cell can be distinguished, and the same is true of most gymno- 
sperms and angiosperms. Instead, the meristem at the tip of both root 
and shoot consists of a considerable group of embryonic cells. Many of 
these divide actively during the growth of the plant, and they produce 
all the tissues of the axis (save those formed by later cambial growth) 
as well as the leaves and branches. 

Much attention has been paid in recent years to the structure, organiza- 
tion, and activity of apical meristems, particularly in the ferns and seed 
plants. These regions of persistent embryonic character have often been 
compared to animal embryos. Botanists have tried to find a correspond- 
ence between their structure and that of the parts that grow from them 
so that the developing plant might be analyzed in terms of embryonic 
regions, as zoologists have been able to do by using the germ layers 
established in the animal embryo. A wide variety of plant meristems 
have been studied and compared, but differentiation into layers as pre- 
cise in their fate as ectoderm, mesoderm, and entoderm seems rarely to 
occur. Some botanists, however, do regard meristematic layers as true 
germ layers (Satina, Blakeslee, and Avery, 1940). 

Although these apical meristems do not provide a precise classification 
of plant tissues, much information of importance for morphogenesis may 
be derived from them. Observation of the way in which meristems have 
produced the various tissues and organs of plants has been of service in 
the solution of problems in growth, differentiation, and phyllotaxy. Plant 
meristems offer the great advantage that a single plant may produce 
many of these embryonic regions, which are thus genetically identical. 
Though small, meristems are open to direct experimental investigation, 
and this has already provided results of much morphogenetic significance. 

The apical meristems of shoot and root, though alike in many respects, 
show certain characteristic differences, and further consideration of their 
structure and activity will be more profitable if each is considered by 


The length of the growing region in the shoot is considerably greater 
than in the root and may often extend over a region of several centi- 
meters. Cell division persists longer in some tissues than in others and 
usually stops first in the pith. No very sharp line is to be found between 
the developing region and the mature portion behind it. The strictly 
meristematic zone, however, where cell division chiefly occurs, is usually 
limited to a few millimeters or less, and most growth of the stem in length 
results from cell elongation back of this. 



The tip of the meristem in seed plants is usually a rounded, dome- 
shaped mass of cells around the base of which the leaf primordia appear 
in succession (Fig. 4-5). Some earlier investigators reported the presence 
of apical cells here but later work did not confirm this. Newman (1956), 
however, finds dividing cells in the very center of the apical dome in 
Tropaeolum and Coleus and believes that they are to be regarded as 
true apical cells. A similar situation has been reported in certain roots. 


b 4"'" •'■ ?w W>. "• • 

; ■ ' 

-„ <%~ - -\V< r- - - r t ■"..:' ••■.. v; 


"••V f 

Fig. 4-5. Longitudinal section through shoot apex of Coleus, showing meristem, leaf 
primordia, and two bud meristems. (Courtesy Triarch Botanical Products.) 

Much attention has been paid to the structure of the dome itself. 
Hanstein (1868) was the first to give careful study to the shoot meri- 
stem. He noted the presence in it of well-marked layers of cells and 
distinguished three regions, or histogens, each of which, he believed, 
gave rise to a particular tissue or tissues of the stem. The outermost, 
or dermatogen, is a single layer and produces the epidermis. Under this, 
several layers thick, is the periblem, giving rise to the cortex. The inner- 
most core, or plerome, without well-marked layers, forms the vascular 
cylinder and pith ( Fig. 4-6 ) . 

This hypothesis would have important implications for morphogenesis 



Fig. 4-6. Diagram of shoot apex according to Hanstein's interpretation. D, dermatogen; 
Pe, periblem; PI, plerome. (After Buvat.) 

if it could be supported. There is now much evidence, however, that no 
constant relation exists, valid for all plants, between these "histogens" 
and the structures formed by them. Some of this evidence comes from 
direct observation, as in Schoute's (1902) studies on the origin of the 
vascular cylinder. Some is derived from the structure of periclinal chi- 
meras (p. 268) in which the layer or layers derived from one graft com- 
ponent can be distinguished by the size of their cells from those coming 
from the other, a distinction that persists in the mature structures and is a 

Fig. 4-7. Vinca minor. Longitudinal section through shoot apex, showing three-layered 
tunica and unlayered corpus beneath it. ( From Schmidt. ) 

62 Growth 

useful means of determining their particular meristematic origin. Evidence 
from these sources shows that a particular tissue may come from one 
meristematic layer in one plant and from another in another plant. 

Hanstein's histogen theory has largely been superseded by another, 
first proposed by Schmidt (1924; Fig. 4-7). This recognizes an outer zone 
of layered cells, usually from one to four cells thick, the tunica, covering 
a core of unlayered cells, the corpus. The tunica-corpus theory does not 
maintain that either of these regions produces specific organs or tissues 
but describes the common type of organization of the shoot apex (see 
Reeve, 1948). 

The significance of layering in the meristem has often been overem- 
phasized. Whether or not a layer is formed depends on the plane of di- 
vision of the meristematic cells. When the apical initials always divide 

Fig. 4-8. Longitudinal section of shoot apex of Torreya californica, showing almost 
complete absence of layering. ( From Johnson. ) 

anticlinally they obviously will produce a layer, and its growth will be 
entirely growth in surface. Where divisions occur in other planes or ir- 
regularly, layers are not produced. Specific factors such as mechanical 
pressure (p. 49) which influence plane of division may thus determine the 
presence and number of layers. If the central region of the meristem 
is growing faster than the surface, the latter will be subjected to pres- 
sure, its cells will tend to divide parallel to the direction of that pressure, 
and a layer will be formed. Perhaps this is the only real significance of 
the layered structure. It is noteworthy that the shoot meristems of many 
gymnosperms (Korody, 1938, and others; Fig. 4-8) show little or no 
layering but that they produce structures in a perfectly normal fashion. 
Layering as such, in the sense of marking out particular regions of the 
meristem that are significant morphogenetically, seems to be of much 
less importance than many workers have regarded it. 



Fig. 4-9. At left, shoot apex of Abies pectinata, semidiagrammatic. At right, diagram 
of confocal parabolas as postulated by Sachs from such an apex as that of Abies. 
( From Sachs. ) 

Even though layering may not be of primary significance, the general 
pattern formed by the planes of cell division in the meristem is of in- 
terest. Reinke (1880) and Sachs (1878) many years ago called attention 
to the fact that the divisions approximately at right angles to the surface 
of the meristem and axis and those parallel to it tend, if extended, to 
form two sets of essentially parabolic curves with a common focus 
just below the apex of the meristem ( Fig. 4-9 ) . This somewhat diagram- 
matic interpretation of the situation has largely been neglected by 
recent writers. Such a pattern can be found both in shoots and roots, 
however, though it is often inconspicuous in small meristems. Foster 
( 1943 ) called attention to the observations of these early workers in his 
study of the broad apices of certain cycads; and Schuepp, both in his 
volume on meristems ( 1926 ) and in a later paper ( 1952 ) , has emphasized 
it. The pattern made by these two series of curves is modified as they 

V ^2 

Fig. 4-10. Diagram of longitudinal section of broad apical shoot meristem of Micro- 
cycas. 1, initiation zone; 2, central mother-cell zone; 3, peripheral zone; 4, zone of rib 
meristem. Arrows represent lines of convergence of cells. (From Foster.) 



suffer displacement, transversely and longitudinally, if growth is more 
rapid in some regions and directions than in others. In such broad 
apices as those of Microcycas, the normal pattern has been greatly 
modified and the tip flares out in a fan-like fashion (Fig. 4-10). What 
these facts mean morphogenetically we do not know, but they show that 
the growing apex has a pattern of organization which develops in a pre- 
cise fashion. 

The shoot meristem is by no means homogeneous or structureless in 
other particulars. In recent years many students have come to recognize 
a rather uniform series of zones within it, distinguished not primarily by 
layers or planes of cell division but by differences in the character of 
their cells. A general survey of zonation in vascular plants has been 
made by Popham ( 1951 ) , who, from his own work and a long series of 
published descriptions of meristems, has grouped them into seven classes. 

Fig. 4-11. Diagram of zonation in the shoot apex of Chrysanthemum morifolium. 1, 
mantle layer; 2, central mother-cell zone; 3, zone of cambium-like cells; 4, rib meri- 
stem; 5, peripheral zone. (From Popham and Chun.) 

In the vascular cryptogams there are one or more apical cells or a sur- 
face meristem, with tissues below sometimes differentiating into a cen- 
tral and a peripheral meristem. Among seed plants, four or five zones 
can be seen (Fig. 4-11). These are a surface zone, or mantle, including 
two to several cell layers and corresponding roughly to the tunica; a 
zone of subapical mother cells, irregular in shape, often rather highly 
vacuolate and dividing less rapidly than the surrounding ones; a central 
zone giving rise to the rib meristem and pith; and a peripheral zone just 
outside this, producing cortex and procambial tissue. In some plants, 
just below the mother-cell zone there is a somewhat cup-shaped arc of 
cells stretching across the axis, the cambium-like zone. Popham and 
Chan ( 1950 ) and Popham ( 1958 ) have described a typical case of this 
last type. The particular functions of these zones are not well understood 

Meristems 65 

but they doubtless differ physiologically. The subapical mother-cell zone 
is perhaps comparable to a somewhat similar region in the root (p. 78) 
where the rate of protein synthesis is lower than in surrounding cells. 

An attempt to follow cellular changes at the surface of living shoot 
apices has been made by Newman ( 1956), using Tropaeolum and Coleus. 
By an ingenious technique he was able to follow and draw, for as long 
as 9 days, the divisions of individual surface cells. At the very tip of the 
meristem he observed that divisions were frequent and believes that in 
this region there is a small group of cells that may be regarded as apical 
cells. His results fail to confirm those of Lance (1952), who reported 
that divisions were infrequent at the very apex, as Plantefol's theory 
(p. 156) assumes. 

There is a considerable literature dealing with the structure of the shoot 
meristem in particular plants and under different conditions. Much of 
this has morphogenetic interest. Cutter (1955), for example, finds that 
the organization of shoot apices in eight saprophytic and parasitic species 
of angiosperms is essentially like that in plants with normal nutrition. 
Boke in a series of papers (1955 and earlier) described the stem apices 
and shoot histogenesis in a series of xerophytes, especially Cactaceae. 
Stant ( 1954 ) compared the shape of the shoot meristem in five species 
of monocotyledons and found a relationship between this character and 
the growth habit of the plant. In general, where the meristem is long 
and narrow, as in Elodea, the plant has well-developed internodes. Where 
it is relatively short and wide, as in Narcissus, the stem is much reduced 
and the internodes very short. The size of the apical dome is essentially 
the same in cucurbits with large fruits as in those with small fruits, and 
differences in organ size do not appear until a short distance below the 
tip of the meristem (von Maltzahn, 1957). The difference between the sin- 
gle-gene maize mutant "corn grass" and normal corn arises in the 
meristem, the mutant having a relatively larger meristem and a more 
rapid production of leaf primordia (Whaley and Leech, 1950). Bouffa 
and Gunckel (1951) examined 54 species of Bosaceae but found no sig- 
nificant relation between the number of tunica layers and the taxonomic 
position of the plant. The development of the shoot meristem from its 
early appearance in the embryo has been studied by various workers 
(Beeve, 1948; Spurr, 1949; and others). 

The implications of results from the study of the shoot apex for 
morphological problems, especially the nature of the leaf, have been 
considered by various observers. Philipson (1949) believes that the evi- 
dence from this source supports the idea that the leaf is an enation and 
not a consolidated branch system. 

Leaf primordia are formed in regular sequence below the dome of the 
shoot apex, and it is here that many of the structural characters of the 



plant seem to be determined. The period between the initiation of two 
primordia (or two pairs, if the leaves are opposite) is termed a plasto- 
chron (Askenasy, 1880; Schmidt, 1924). These periodic changes in the 
meristem can be seen best in opposite-leaved forms. As the two primordia 
begin to appear, the apical dome between them becomes relatively flat. 
When they have developed further but before another pair appears, the 
dome bulges upward again and reaches its maximal surface area. In the 
lower vascular plants the primordium arises from one or more of the 
surface cells of the meristem but in the higher ones it develops as a 
swelling on the side of the apex at the base of the dome, generally as 
the result of periclinal divisions in one or more layers below the surface 
one. The term plastochron index, for the interval between corresponding 
stages of successive leaves, has been proposed (Erickson and Michelini, 

Shoot Tip during Rutmg Prion 

Shoot Tip during Second Growth Phatt 

Fig. 4-12. Diagram of shoot tip of Abies concolor. At left, during resting phase. At 
right, during second growth phase. That portion of the shoot apex above plane abed, 
which marks the level of the youngest leaf primordium, has a very different zonal 
topography in the two stages. 1, zone of apical initials; 2, mother-cell zone; 3, 
peripheral zone; 4, zone of central tissue. ( From Parke. ) 

1957 ) as a better measure of the stage of development of a growing shoot 
than is its chronological age. In many plants the shape and structure of 
the meristem change somewhat with the season (Parke, 1959; Fig. 4-12). 

The phyllotaxy of a shoot is determined by the arrangement of the 
leaf primordia around the axis. This phyllotactic pattern has been studied 
developmentally in the meristem, both through observation and experi- 
ment, by a number of workers (Chap. 7). The regularity and precision 
with which the leaf primordia arise at the shoot apex are evidence that 
this region has a high degree of organization. 

Branches are formed from meristems arising in the axils of the leaf 
primordia. They are at first much smaller than the main apical meristem 
but do not differ essentially from it. Whether the potentially meristematic 
tissue here will grow into buds and whether these buds will produce 
branches are dependent in most cases on the stimulatory or inhibitory 



influence of auxin or other growth substances (p. 386). The development 
of axillary buds has been discussed by Sharman (1945) and by Garrison 
(1955), who finds that they originate from a region of residual meristem. 
After the organization of an apical meristem in the bud, the procambial 
strands develop acropetally into the leaf primordia, as does the phloem. 
Xylem forms at several loci and develops in both acropetal and basipetal 

directions (p. 204). 

When the shoot is producing leaves, the meristematic dome is relatively 
low and rounded but when flower buds begin to be formed it becomes 
steeper and more elongate. Flowers arise as modified branches, and the 
floral parts develop from a series of primordia (Fig. 4-13). In the forma- 


Fig. 4-13. Longitudinal section through young inflorescence, showing stages in de- 
velopment of floral primordia. i, bract; t, trace to primordium; v, procambial strand. 
Most active meristematic areas are stippled. (From Philipson.) 

tion of more complex inflorescences, however, the meristem changes 
markedly. Since growth in length usually ceases at this time, what is 
formed is essentially a determinate structure rather than an indeterminate 
one like that of the vegetative shoot. What its character will be is de- 
cided by the size and number of the flowers and the character of the 
inflorescence. Various factors, notably the carbohydrate-nitrogen ratio 
and the photoperiod, determine whether the meristem forms vegetative 
or reproductive structures (p. 184). For accounts of the development of 
the reproductive apex, see papers by Gregoire (1938), Philipson (1948), 
Gifford and Wetmore (1957), and others. 

The shoot meristem is not constant in size but changes during de- 
velopment. In the young embryo it is very small, and it enlarges as the 

68 Growth 

plant grows. At the onset of reproductive maturity or at the end of the 
life cycle it often becomes reduced again in size. In maize, a plant 
essentially determinate in its growth, Abbe and his colleagues (1951a, 
1951b, 1954) have studied the size of the apical shoot meristem (above 
the first leaf primordia) and the size and number of its cells as these 
change with time. In plastochrons 7 to 14 (the seedling stage until just 
before flowering) the apex increases by a constant amount in each 
plastochron but the duration of the plastochron decreases exponentially, 
from a length of 4.7 days to one of 0.5 day. Cell size is essentially con- 
stant throughout, so that all growth is by cell multiplication. The growth 
rate per plastochron accelerates exponentially. In the five or six plasto- 
chrons during early embryogeny, on the contrary, the duration of suc- 
cessive plastochrons increases and the growth rate of the apex decreases. 

Sunderland and Brown (1956) have determined the cell number and 
average cell volume in the meristematic dome and the first seven 
primordia and internodes, back from the apex, in the shoot of Lupinus. 
The primordia increase exponentially in successive plastochrons but there 
is little increase in cell volume in the internodes. 

Cell shape in the shoot apex of Anacharis (Elodea) has been studied 
by Matzke and Duffy (1955) with particular reference to the number 
of faces. These range from 9 to 21 and are in general agreement with 
the shape of cells in other undifferentiated tissues. 

For a statement of conditions in the apical meristem of the shoot in 
ferns the reader is referred to Wardlaw (1945). This author has also 
published an extensive series of papers on experimental and analytical 
studies of pteridophytes, many of which are cited in his books (1952«, 
1952&) and in later papers by himself and his colleagues. The shoot 
meristems of gymnosperms are described by Camefort (1956) and John- 
son ( 1951 ) . General accounts of this region in the angiosperms, with re- 
views of the literature, have been written by Foster ( 1939, 1949 ) , Sif ton 
(1944), Philipson (1949, 1954), Popham (1951), Buvat (1952), and 
Gifford (1954). 

The ontogeny of a typical shoot apex (Xanthium) has been described 
in detail by Millington and Fisk (1956). 


Early work on the shoot apex was primarily descriptive, and much of 
this still continues. It has been concerned with apical cells, planes of 
division, cell lineages, layering, zonation, and the relations of the meri- 
stem to differentiation and organ formation. This work has been of great 
value morphogenetically for it has provided a fund of information as to 
the structure and developmental activity of this determinative region 

Meristems 69 

of the plant, but it has yielded little knowledge of the meristem as a 
living and functioning center of morphogenetic activity. 

In recent years, however, an increasing number of workers have used 
experimental methods to attack the problems of the meristem with 
techniques like those which have proved so fruitful in the experimental 
embryology of animals (Wetmore and Wardlaw, 1951). This part of the 
plant is a perpetually embryonic region and thus offers many advantages 
for work of this sort. Most shoot meristems are minute, however, and so 
enfolded by protective structures that experimental work upon them has 
had to await the development of specialized techniques. Means for direct 
operative attack on the meristem have now been developed and have 
begun to yield valuable results. In this work the Snows, Wardlaw, Ball, 
Wetmore, and their colleagues have been particularly active. The methods 
of tissue culture have recently added a wealth of information. Biochemi- 
cal analysis by means of experiments with growth substances, chromato- 
graphic techniques, and other methods is yielding further knowledge 
of meristem physiology. This experimental work has powerfully supple- 
mented earlier descriptive studies. 

Direct operations on the meristem involve procedures of much delicacy 
and are performed under a lens by tiny scalpels. Pilkington ( 1929 ) seems 
to have been the first to do such work. She split the meristem of Lupinus 
down the middle and found that each half regenerated a normal meri- 
stem so that the original axis was now divided into two branches. Ball 
( 1948 ) , also using Lupinus, divided the meristem into four parts 
longitudinally and each of the four, by regenerative development, was 
able to produce a normal shoot. Later ( 1952a ) he went still further and 
split the meristem into six strips. Each of these, unless it was below a 
minimum size, regenerated a new meristem and shoot, though leaf for- 
mation was somewhat delayed and vascular tissue was poorly differen- 
tiated until leaves had developed. 

A major problem here is to find to what extent the meristematic tip is 
autonomous and thus independent from the tissues below it in develop- 
ment. In a fern, Dryopteris, Wardlaw (1947 and other papers) isolated 
the apical meristem from the adjacent leaf primordia by four longi- 
tudinal incisions (Fig. 4-14). It was thus continuous with the rest of the 
plant only by the parenchyma of the pith below it, all the vascular tissue 
having been severed. Despite this isolation, the meristem continued to 
grow and to produce leaf primordia and leaves. Whatever material en- 
tered it came through undifferentiated parenchyma. Provascular tissue 
was developed below the tip but this did not make connection with the 
vascular bundles in the stem below. 

Ball ( 1948 ) did much the same thing in Lupinus, isolating the central 
axis of the meristem by four deep incisions. In this axis, however (unlike 



all cases studied in the ferns), vascular bundles were regenerated in the 
former pith tissue and became connected basipetally with the vascular 
system of the axis below. A normal shoot was thus restored. Similar 
results were obtained by Wardlaw (1950) with Primula. Furthermore, 
Ball ( 1946 ) succeeded in growing complete plants from isolated apices 
in sterile culture, and this has now been done repeatedly in other cases, 
both with ferns and seed plants ( see Wetmore, 1954 ) . 

As a result of such experiments it is now generally agreed that the 
shoot apex is totipotent and independent of the rest of the plant. This is 
hardly surprising since many— perhaps all— cells are totipotent under 
favorable conditions. The subjacent tissue must have some influence on 
the apex, however, since from below there come into it not only water 

Fig. 4-14. Longitudinal median section through the stem apex of Dryopteris, show- 
ing how the region around the apical cell ( a ) has been isolated by deep cuts extend- 
ing through the vascular tissue (v.t. ). (From Wardlaw.) 

and nutrient materials but specific substances of morphogenetic im- 
portance. The induction of flower buds at the meristem, for example, 
results from a hormone brought thither from the leaves. There evidently 
must be a reciprocal relation between the shoot meristem and the axis 
below it for both are parts of the same integrated organic system. 

The shoot apex has a morphogenetic role which goes beyond regenera- 
tion, however, for it exerts a strong influence on the differentiation of 
tissues and organs in the region adjacent to it. By a long series of ex- 
periments reported in many papers (see Wardlaw, 1952a, 1952/7, and 
Cutter, 1958) Wardlaw and his colleagues have made important con- 
tributions to a knowledge of this differentiation. Much of the work was 
done on the apical meristems of ferns. In Dryopteris, Wardlaw (1949&) 
determined the region where the next leaf primordium would arise 

Meristems 71 

(through its place in the phyllotactic series) and then isolated it from 
the apical cell by a deep tangential cut. Under these conditions a bud 
rather than a leaf primordium developed. Evidently some influence com- 
ing from the apex determines whether a new lateral outgrowth will form 
the dorsiventral primordium of a leaf or the radially symmetrical, po- 
tentially indeterminate one of a bud, a discovery of much importance 
not only morphogenetically but morphologically. Cutter (1956) has 
found that the three youngest primordia respond in the same way but 
that older ones do not. These same young primordial areas (in Osmunda) 
if excised and grown in culture will form buds and finally mature 
plants, whereas the older areas, if cultured, will grow into typical leaves 
(Steeves and Sussex, 1957). There is evidently a point before which the 
lateral structure has the potentiality to form either a bud or a leaf, but 
after a certain early stage has been reached its fate is determined. 

Wardlaw has extended these studies further (1956a, b). When in- 
cisions between apex and primordium were so shallow that the pre- 
vascular strands were not cut, a leaf primordium still developed, sug- 
gesting that this incipient vascular tissue is a pathway for morphogenetic 
stimuli. When deep cuts were made on the radial and obaxial sides of a 
primordium site, thus without isolating it from the apex, a leaf grew from 
it, but this usually showed abnormally rapid growth. 

In flowering plants results like these were not obtained, for isolated 
primordia do not develop into buds but into dorsiventral leaves or 
radially symmetrical leaf-like structures (Sussex, 1955). However, Cutter 
(1958) found that in Nymphaea and Nuphar (favorable material be- 
cause of their large meristems), although tangential cuts separating a 
primordium from the apex did not change it to a bud, buds under these 
conditions were formed more often and very close to the primordium. 
The critical time for the determination of the fate of a primordium thus 
seems to be earlier in flowering plants than in ferns. 

The problem of phyllotaxy is closely related to conditions at the meri- 
stem, for the arrangement of the leaves is presumably determined by 
the distribution of their primordia. This subject will be discussed more 
fully in the chapter on Symmetry, but it should be mentioned here that 
the experimental work of the Snows and of Wardlaw on the factors that 
determine where a given primordium shall arise has yielded much in- 
formation. It is evident that the apex has an important influence on the 
differentiation of primordia, but whether the exact determination of the 
position of these structures results from chemical, mechanical, or geomet- 
rical factors is still not clear. Experimental manipulation of the meristem 
is a hopeful way of approaching this problem directly. 

The differentiation of vascular tissue seems to depend on a stimulus, 
probably a growth substance, passing basipetally from the apex. The 

72 Growth 

influence of buds on the production of vascular tissue below them is 
well known. When growth of an adventitious bud is induced in the epi- 
dermis, for example, a vascular strand usually differentiates in the paren- 
chymatous tissue below it which establishes connection with the vascular 
tissue of the leaf or stem (p. 245). More direct experimental evidence on 
this problem is available from other sources. When Camus (1944, 1949) 
grafted a bud of endive into root callus in culture, provascular strands 
were formed below it. Camus attributes this to the effect of auxin (and 
possibly other substances) produced by the bud. Wetmore and Sorokin 
(1955) and Wetmore (1956) have shown much the same effects from 
lilac buds grafted into lilac callus in culture. Here the amount of 
provascular tissue induced in the callus below was much increased if 
auxin was also added. 

Wardlaw ( 1952c ) has called attention to the important effects on stelar 
structure which are related to the strength of the meristematic stimulus 
in the development of vascular structures. When leaf primordia in ferns 
are removed, gaps in the vascular ring below are much reduced. In this 
way it was possible to transform the axis of the normally dictyostelic 
Dryopteris into a solenostelic form with a continuous ring. By further 
reducing the size of the meristem through isolating it on a small piece of 
tissue, even a protostelic condition was produced. The basis for such im- 
portant morphological differences thus seems to be in the degree of de- 
velopment of the meristem. 

In recent years the physiology of these tiny shoot apices has also begun 
to be investigated. Growth substances are evidently synthesized there, 
but just what these are and how they act are still not known. Ball ( 1944 ) 
found that when auxin in paste was applied to the meristem apex of 
Tropaeolum no changes were produced in it, presumably because of the 
large amount of native auxin present; but below the tip hypertrophied 
tissues, multiple leaves, and abnormal development of vascular tissue 
appeared. The role of growth substances at the meristem involves many 
important problems in plant physiology and morphogenesis (Chap. 

By use of the techniques of paper chromatography, information is 
being gained as to the biochemistry and particularly the nitrogenous 
components of apical meristems. Some of the pioneer work here is de- 
scribed by Steward and others (1954, 1955) and by Wetmore (1954). 
The meristematic region is well supplied with free amino acids. The 
basic ones, arginine and lysine, are more abundant in the tip of the 
meristem than in tissues farther back, in the primordia, leaves, and 
stem. A substantial beginning has been made toward a knowledge of 
the distribution of these substances and similar ones, as well as of DNA 
and various enzymes, throughout the meristematic region and at dif- 

Meristems 73 

ferent stages in its development. Something is thus being learned about 
the chemical as well as the histological organization of the shoot apex. 

The techniques of tissue culture have also proved to be very useful for 
a knowledge of meristem physiology. It has been found (Wetmore, 
1954), for example, that the shoot apex of vascular cryptogams, when cul- 
tured with only inorganic substances and sucrose, will produce entire 
plants but that they will grow better if supplied with auxin and some 
nitrogen source other than nitrates. Angiosperm apices (Syringa), how- 
ever, will not grow in this simple medium, but if coconut milk and casein 
hydrolysate are added, the tips root, and growth is much better. When in 
the culture medium the same amino acids and amides are provided, and 
in the same proportions, as are found in meristem tissue, growth is still 
very slow and far from normal. Evidently something more is necessary. 

Wetmore's demonstration (1954) that when apices of sporeling ferns 
are cultured with successively higher concentrations of sucrose the leaves 
that they produce correspond to those formed in progressive stages of 
normal ontogeny (p. 222) shows the important morphogenetic and mor- 
phological implications of nutritional factors. 

The rates of metabolic processes in the shoot apex have also been in- 
vestigated. Ball and Boell (1944), using the Cartesian-diver technique 
by which it is possible to measure the rate of respiration in tiny bits of 
living tissue, compared this rate in the apical dome of cells, the region 
just below this where the first primordia are appearing, and a third region 
below this (Fig. 4-15). In Lupinus, respiration was most active at the tip 
and progressively less so below. In Tropoeolum, however, there was less 
respiration in the extreme tip than in the region below it. The occurrence 
of a descending metabolic gradient in the plant apex thus seems not to 
be universal. 

These various experimental studies on the shoot meristem have directed 
attention even more strongly to this embryonic region. Many believe that 
it is of primary importance for development and that in it the major 
problems of morphogenesis, at least as far as the shoot system of vascular 
plants is concerned, come sharply to focus. This conclusion is supported 
by the facts that the apex is autonomous, a small portion of the tip of it 
being able to produce an entire plant; and that if it is removed, the 
development of the tissues below it finally stops. It is recognized, how- 
ever, that certain structures, such as leaf primordia beyond a certain 
stage, are partially removed from its control since they will develop in- 
dependently in culture and are thus self-differentiating. 

Ball has compared the shoot apex to an organizer such as has been 
postulated in animal embryology. Such, in a sense, it is, but the com- 
parison is not very exact since the organizer is a part of the embryo 
which controls the development of the rest, wheeras the shoot apex 



corresponds to the whole embryo itself and thus to the organism in minia- 
ture. Recent work shows more and more clearly that the apex is an 
organized system ( Wardlaw, 1953b, 1957a ) with a structure that is not 
only histologically but biochemically integrated. Before this structure 
can control development it must itself develop. Growth of the meristem 







Fig. 4-15. Lupinus alhus, left, and Tropaeolum majus, right. Below, location of A, B, 
and C pieces in the two shoot apices. Above, oxygen consumption of these three 
pieces in milliliters of 2 , at successive hours in the apparatus. In Lupinus there is a 
gradient from A to C but in Tropaeolum the oxygen consumption is greater in B than 
in A. ( From Ball and Boell. ) 

precedes that from the meristem. The initiation of a meristem may take 
place in various ways— from the tip of a young embryo, from adventitious 
buds by regeneration, or from groups of cells or even single cells— but in 
every case from simple and undifferentiated tissue. 

A study of the origin of meristematic centers within masses of tissue 

Meristems 75 

in culture promises to be enlightening. Here Steward and his collabora- 
tors (1958) have done some significant work. They were able to grow 
in culture dissociated phloem cells from the root of carrot. Some of these 
produced multicellular masses. When a mass reached a certain size there 
formed in it a sheath of cambium-like cells enclosing a nest of lignified 
elements such as often is found in tissue cultures. In this spherical 
nodule there developed a root meristem and then a shoot initial opposite 
it. Thus an embryo-like structure was formed which was able to grow 
into an entire normal carrot plant. Steward emphasizes the fact that the 
change from random cell multiplication to organized development and 
the formation of meristems comes only after the group of inner cells has 
become enclosed by a wall of outer ones which cuts it off from direct 
access to the coconut-milk medium outside and subjects these inner cells 
to physical and presumably physiological restraints. Before this happens 
they multiply irregularly and form simply an unorganized callus-like 
mass. Such studies open up an important line of attack on the problem 
of the origin of organized meristems. It will not be possible to understand 
the role of meristems in development until we learn through experiments 
like these how such an organized apical system comes into being. 


The apical meristem of the root differs in several respects from that of 
the shoot. It is relatively short, the elongating region of the root rarely 
exceeding a millimeter in length. No lateral organs have their origin at 
the apex, and thus there are no rhythmic changes here as in the shoot. 
The lateral roots arise farther back, in the pericycle, and push out 
through the cortex. The apical meristem produces not only the structures 
of the root itself but, from its outer surface, the root cap, or calyptra. 
Root meristems received much attention in the early work of Eriksson 
(1878), Flahault (1878), Holle (1876), Janczewski (1874), and van 
Tieghem and Douliot (1888). 

In those lower vascular plants where shoot growth is centered in an 
apical cell, the root grows in the same way (Fig. 4-2). In the higher 
plants, however, although there is a meristem which is in many respects 
like that of the shoot, there is less uniformity in its organization. In 
most roots there are seen well-marked layers and to these some workers 
have applied Hanstein's terminology. The direct origin of particular tis- 
sues from particular layers is not uniform, and the same objections to 
regarding the layers as histogenetic ones may be made as for the shoot. 
The presence of the root cap prevents smooth and continuous layering 
over the root apex, and this is probably the reason why periclinal chi- 
meras are not found in roots. 



In many forms growth seems to be centered in a small group of cells. 
Brumfield (1943) produced specific chromosomal changes in root-tip 
cells of Crepis and Vicia by X radiation which could be recognized in 
the descendants of these cells farther back along the root. They were 
found to form wedge-shaped sectors of the entire root, including parts 
of the root cap, epidermis, cortex, and vascular cylinder (p. 268). Such a 
sector usually occupied about a third of the area of the root cross section, 
and Brumfield concluded that there were about three cells at the tip from 

Fig. 4-16. Longitudinal diagram of root development as observed by Williams. The 
tissues all arise from a small group of cells at the very tip. 1, epidermis; 2, hypo- 
dermis; 3, endodermis; 4, pericycle; 5, mitotic figure; 6, young cortical cells; 7, stelar 
initials; 8, dermatogen; 9, metaxylem initial. B, C, D, transverse diagrams showing 
origin of cortical cells from a single cell of the endodermis. ( From B. C. Williams. ) 

which all the tissues of the root were derived, though these cells could 
not be recognized in sections of the root apex. Von Guttenberg ( 1947 ) 
later presented evidence from a considerable variety of dicotyledonous 
plants that there is a single apical cell that gives rise to the whole root 
and which thus is comparable to the apical cell of lower vascular plants, 
but this has received little confirmation. 

Popham ( 1955a ) found in Pisum sativum a transverse row of meri- 
stematic initials across the root apex that gives rise to all the tissues of 
the root and the cap, and Clowes (1954) observed a somewhat similar 



condition in Zea. Both these workers believe that Brumfield's results can 
better be explained on the assumption of such a small meristematic cen- 
ter than of a group of apical cells, thus far unobserved. Clowes ( 1950 ) 
found meristematic layers in the root tip of Fagus that seemed discrete 
enough to be called histogens. 

Williams (1947) observed in many roots of vascular plants a rather 
simple pattern of development. The epidermis, hypodermis, and endo- 
dermis could all be traced back to a small group of cells at the very tip of 
the plerome. The endodermal row, coming from this, gives rise by re- 

Fig. 4-17. Phleum root tip. Graph showing rate of root elongation (A), average 
length of epidermal cells (B), and new transverse cell walls (C), at various distances 
from the root tip. ( From Goodwin and Stepka. ) 

peated tangential divisions to the cortex. This accounts for the fact that 
the cortex, particularly in its inner layers, is often made up of radial rows 
of cells (Fig. 4-16). A second small group of cells, just below the plerome 
tip, produces all the stelar tissues. The progress of cell division, par- 
ticularly in the surface layer and the cortex, was followed by Wagner 
( 1937 ) by means of tracing cell groups, or "complexes," each of which 
had descended from a single meristematic cell. Sinnott and Bloch (1939) 
studied cell division in living root tips of small-seeded grasses by camera- 
lucida drawings. Brumfield (1942) continued this work by the use of 

78 Growth 

photography, and the situation was analyzed more fully by Goodwin and 
Stepka (1945; Fig. 4-17) and Goodwin and Avers (1956). Erickson and 
Goddard (1951) used still more refined photographic methods. The lo- 
cation and rate of cell division in the root offer problems of considerable 

There are a number of developmental patterns in the root apices of 
seed plants. They have been classified into various "types" by Janczewski 
(1874), Kroll (1912), and Schuepp (1926). These are well described 
in Esau ( 1953&, p. 116). In the structure of the root apex there is ob- 
viously less uniformity than in that of the shoot. 

The development of the apical meristems in both root and shoot from 
their first appearance in the early embryo has been studied by a number 
of workers. A typical example is described in Pseudotsuga by G. S. Allen 

Root tips, with the regions just behind them, were among the first 
materials to be used for plant-tissue culture (p. 296) and much has been 
learned through this technique as to the physiology of the root. In many 
plants, for example, the root cannot synthesize thiamin but depends for 
this vitamin on a supply produced in the shoot. 

Mention has already been made (p. 41 ) of the work on the physiology 
of the root meristem by Brown and his colleagues, who determined the 
changes that take place in the activity of the apical cells in various re- 
gions, particularly as to growth rate, respiration rate, and protein syn- 
thesis. In a general discussion of this work, Brown, Reith, and Robinson 
(1952) show that there is a considerable difference in the composition 
of the proteins at different distances from the apex. Jensen (1955) has 
also made a biochemical analysis of the cells near the root tip in Vicia 

Clowes (1956, 1958) discovered between the active meristematic re- 
gion and the root cap a cup-shaped group of cells, the quiescent center 
(Fig 4-18), which from their appearance divide rarely. He reports that 
these cells synthesize DNA more slowly than do the surrounding ones. 
They presumably have some specific metabolic function. Jensen and 
Kavaljian ( 1958 ) have made a census or cell divisions in the root tip of 
Allium cepa. They found a definite apical initial region where there are 
few divisions and agree with Clowes that these have a low DNA content. 
Cell division is slower to start in the axial than in the peripheral region of 
the tip. They report a very definite daily periodicity in division, with a 
maximum about noon. 

The growth-substance relations of growing roots have received much 
attention (p. 391). Auxin and various synthetic substances stimulate the 
initiation of root primordia but usually check the later growth of the 
root. Auxin tends to be basipetal in its flow, a fact that helps to account 



Fig. 4-18. Median section of root tip of Zea, showing the quiescent center (stippled). 
Its cells are physiologically different from the surrounding ones and seem rarely to 
divide or grow. ( From Clowes. ) 

for the normal preponderance of roots at the base of the plant axis. A 
high carbon-nitrogen ratio (p. 366) also favors root growth. Torrey ( 1950) 
presents evidence that a growth substance, not auxin, is produced in the 
root and moves toward the apex, stimulating the formation of lateral 

Intercalary Meristems. Growth of an axis in length may sometimes take 
place at other points than its tip, by the activity of an intercalary meristem. 
Thus in many monocotyledons cell division persists in the base of an 
internode when it has ceased elsewhere, and the stem elongates all along 
its course, somewhat like an extending telescope (Prat, 1935). The gyno- 
phore of the peanut, which carries the young fruit down and into the 
ground, grows in a somewhat similar way, as has been described by 
Jacobs (1947). Such material is excellent for a study of the relations of 
cell division and cell elongation to growth. 


The Vascular Cambium. Apical meristems govern growth in length and 
produce those tissues commonly called primary. Their cells tend to be 
arranged in longitudinal rows, each row being the descendants of a single 

80 Growth 

meristematic cell. In woody plants, the root and stem axes continue to 
grow not only in length but in diameter. This is accomplished chiefly by 
the activity of another meristem, quite different in character, the vascu- 
lar cambium. The tissues that such meristems produce are termed 


The vascular cambium is a sheath of embryonic cells extending from 
the beginning of secondary growth in the shoot tip to a corresponding 
position in the root. It arises between the xylem and the phloem of the 
primary vascular bundles and forms xylem on its inner face and phloem 
on its outer one. Each cambium cell produces a radial row of daughter 
cells on either side. The tissues thus formed can usually be recognized 
by this cell pattern, though it may sometimes be altered by the marked 
increase in diameter of certain cells, notably the vessels and sieve tubes. 
Cells of the apical meristem are relatively uniform, varied though their 
products may be. Cambium cells, on the other hand, from the beginning 
are differentiated into two quite dissimilar types of cells, corresponding 
to the longitudinal and transverse cellular systems in vascular tissue. 
Those cambium cells that produce tracheids, fibers, sieve cells, and other 
elements of the longitudinal system are termed fusiform initials and are 
usually much elongated in the dimension parallel to the axis. The ray 
initials are much smaller and essentially isodiametric and produce the 
rays in wood and phloem. 

The fusiform initials, especially those destined to form tracheids and 
fibers, may be from 50 to 100 times as long as wide. Often they do not 
differ greatly in length from the mature cells that they produce, though 
in some cases the fibers of the summer wood may become much 
longer than their initials (Bosshatd, 1951). Most of the divisions of these 
initials must be longitudinal and tangential since only in this way can 
additions be made to the width of the axis. The division of such an 
elongate cell violates Hofmeister's law. It is a remarkable process and 
was first clearly described by I. W. Bailey (1920a and b; Fig. 4-19). 
The nucleus, usually in the center of the cell, divides mitotically. Be- 
tween the two daughter nuclei the cell plate is then laid down by the 
phragmoplast, an extension of the system of fibers at telophase. This ap- 
pears in longitudinal section as two sets of fibers (sometimes called 
kinoplasmosomes) connected by the cell plate, one moving upward and 
one downward until the basis of the new division wall has been com- 
pleted. An account of division in cambial cells was also given by 
Kleinmann ( 1923 ) in a paper written during the war and without a 
knowledge of Bailey's work. 

Since the circumference of the axis continually increases, it is necessary, 
if the cambium cells are not to enlarge in tangential diameter, that they 
increase in number by occasional radial divisions. In storied cambia, 



where the initials are in tiers, this is done by radial divisions much like 
the tangential ones. More commonly, however, the initial cell divides by 
an obliquely radial (pseudotransverse) wall, and the daughter cells move 
past one another until the two initials reach normal length and lie side 
by side tangentially. The number of such divisions is usually more than 
enough to make the number of initials conform to the enlarging circum- 
ference of the cambium (Bannan, 1953). In many cases the daughter 
cells fail to maintain themselves, and the rows of cells coming from 
them are pinched out and gradually disappear (Bannan and Bayly, 
1956). This process is so regulated, however, that the normal tangential 
diameter of the mature cells is essentially maintained. The length of the 
vascular elements originating from the cambium is also regulated to an 
approximately constant size. 

Fig. 4-19. Early stages in the division of a cam- 
bial cell, near the middle of a long initial. The 
nucleus has divided, and the cell plate is being 
formed by the fibers of the kinoplasmosome 
(fc). (From I. W. Bailey.) 

The change in relative position of these enlarging cambial daughter 
cells involves a problem in cellular readjustments that is of importance 
morphogenetically. In 1886 Krabbe published a monograph on what he 
termed "sliding growth," presenting evidence that during the differentia- 
tion of tissues there was a good deal of change in intercellular position 
brought about by the slipping or gliding of one cell past another. This 
is common in animal embryology, where cells are more plastic and 
often migrate for some distance, and is responsible for many of the 
changes that take place in the development of these organisms. Its oc- 
currence in plants might therefore be expected and was generally ac- 
cepted as true for some time. Krabbe was supported in his position by 
some other botanists, notably Neeff ( 1914 ) and Grossenbacher ( 1914 ) . 

Priestley ( 1930 ) criticized Krabbe's conclusion and believed that all 
intercellular changes were brought about by what he termed "sym- 



plastic" growth, a rather vague concept that there is a readjustment 
of all the cell walls operating in one "common framework" and with 
no slipping between adjacent ones. Sinnott and Bloch ( 1939, 1941 ) studied 
living and growing tissues of young roots where the gliding of one 
cell past another would be recognized, if it occurred, by alterations in 
relative wall positions and found no evidence for it. They suggested 







Fig. 4-20. Origin of ray initials in gym- 
nosperms. Radial sections from Thuja, 
showing progressive subdivisions of fusi- 
form initials and the consequent origin 
of several ray initials. (From Barman.) 

that changes in intercellular relationships come about by intrusive growth 
limited to a particular region (such as the tips of cambial cells), so 
that the cell may grow in between its neighbors without requiring that 
it slide past them. Bannan and Whalley (1950) have shown how this 
is accomplished in elongating fusiform initials. Schoch-Bodner and Huber 
( 1951 ) present evidence that the phloem fibers of flax, which become very 



long, grow not only by cell stretching as the internode elongates but 
also by localized growth at both tips, as a result of which the fibers push 
in between adjacent ones. In the readjustments thus made necessary 
it is essential that where pits are present the pit fields in adjacent cells 
develop opposite each other, since there must be a corresponding opening 
in each wall. This evidently takes place after the relative position of the 
walls has become fixed. The problem of "sliding growth" has been dis- 
cussed at length bv Meeuse (1942). 

The increasing circumference of the axis also requires that the number 
of rays be continually increased if the proportion between rays and verti- 


Fig. 4-21. Radial section of wood of Chamaecyparis, showing transitional cell types 
associated with the origin of a ray from a fusiform cell. ( From Bannan. ) 

cal elements is to be maintained. The origin of new rays has been de- 
scribed for gymnosperms by Bannan ( 1934 and Figs. 4-20 and 4-21 ) and 
by Barghoorn (1940a and Fig. 4-22) and for certain angiosperms by 
Barghoorn (1940b and Fig. 4-23). A new ray arises from a short cell cut 
out of the radial face of a fusiform initial, the nucleus first migrating to 
the particular place where the new daughter cell is to be produced. The 
height of the ray is then increased by transverse divisions of this cell and 
its products, and its width by radial divisions. High rays may break up 
into shorter ones. The rays as seen in tangential section maintain an al- 
most constant distance from one another. How this is accomplished is 
described by Bannan (1951). 



The cambial region is obviously a much more active and plastic one 
than early workers regarded it. New initials are being produced and 
others are disappearing. Rays are being formed, fusing and dividing. 
Changes and rearrangements are continually taking place among the 
initials. The two sides of the cambium are forming quite different types 
of cells, and further differentiation within the xylem and the phloem is 
beginning. All these changes, however, are so well coordinated and regu- 
lated that a specific pattern of structure, constant enough for taxonomic 
purposes, is produced and maintained. There are few places in the plant 
where histological differentiation can be so well studied as in the prod- 
ucts of the vascular cambium. 

ii i 


Fig. 4-22. Radial sections of wood of Ginkgo (A) and Amentotaxus (B and C),- show- 
ing relation of young ray to ends of wood cells. ( From Barghoorn. ) 

The cambium proper consists of a single row of cells, though on the 
xylem side there are usually several rows of mother cells, developed from 
it, which by their division produce the xylem. Relative activity in xylem 
and phloem production differs considerably in different forms. In Thuja, 
Bannan (1955) found that phloem began to develop later than xylem but 
then continued at a steady rate. In larch, however, Knudson ( 1913 ) had 
reported that phloem development preceded that of xylem, though the 
most rapid growth of each took place at the same time. In Acer, Cocker- 
ham (1930) and Elliott (1935) observed that the first cambial divisions 
formed the spring type of sieve tubes. This was followed by xylem 
growth, during which no new phloem was formed. As xylem production 
ceased, a second phase of phloem development occurred in which the 



smaller summer sieve tubes were formed. The various steps in the develop- 
ment of xylem from cambium to heartwood, with particular reference to 
changes in the cell wall, have been described by Bailey (1952). Lade- 
f oged ( 1952 ) has published a detailed study of cambial activity and wood 
formation in six conifers and 13 hardwoods in Denmark, based on obser- 
vations from March to November. 

The character of cambial products, particularly on the xylem side, as 
to the number and size of the cells and the thickness of their walls, is 
influenced by various factors though the precise effect of these has had 

Fig. 4-23. Origin of ray initial in an angiosperm. Serial tangential sections of wood of 
Trochodendron, showing origin of ray initial (stippled) from the end of a fusiform 
initial. (From Barghoorn.) 

little experimental study. In temperate climates, during the cold weather 
of much of the year cambial activity ceases. The contrast between the last 
formed wood of one season and the first of the next makes it possible to 
identify boundaries of the annual rings formed in each season. These are 
absent in regions where growth is continuous. The relations between 
climatic factors, particularly annual rainfall, and the width of these rings 
has been studied ( Glock, 1955, and others ) . Lines of denser wood within 
an annual ring may be related to rainfall differences in a single season 
(Dobbs, 1953). Injuries from frost, fire, and insect attack can also be 
recognized by their effects on the growth ring. These changes in the 

86 Growth 

products of the cambium, through the permanent record they leave in the 
tree, are of importance in the study of past climatic changes and in 
the dating of ancient timbers, and have been actively studied especially 
by Schulman ( 1956) and others in the Laboratory of Tree Ring Research. 
The course of cambial activity differs among various plants. In herba- 
ceous ones it is associated with vegetative growth and generally ceases at 
flowering (Wilton and Roberts, 1936). In most woody plants, growth of 
the new shoots in length is complete or nearly so before there is much 
cambial activity. The age of the tree may make a difference in woody 
plants, for Messeri (1948) observed that in old trees cambial divisions 
began a month earlier in the twigs than in the main stem whereas in 
young ones they started simultaneously throughout. In most conifers and 
in ring-porous angiosperms, cambial growth begins at about the same time 
throughout the stem (Wareing, 1951). In diffuse-porous angiosperms, 
however, division starts just below the buds at about the time they open 
and proceeds downward into the branches and then the trunk (Cocker- 
ham, 1930; Priestley, Scott, and Malins, 1933). This is apparently related 
to the production of auxin in the buds, for there is a close relation be- 
tween the appearance of auxin there and the onset of cambial activity 
(Avery and others, 1937b). 

The relation of auxin to cambial growth has also been studied by Brown 
and Cormack (1937), Soding (1940), Kiinning (1950), and others. 
Chowdhury and Tandan (1950), working with both evergreen and de- 
ciduous trees in India, report that buds burst in February or March and 
that growth in length continues until May. Not until the new leaves are 
fully expanded and length growth has ceased does cambial activity begin. 
It starts at the tip of the last year's shoot and proceeds down the tree and 
up into the new shoot. Growth in length begins again in the summer 
and is accompanied by cambial activity until both cease in the fall. The 
authors suggest that there are two types of substances operating here, 
one concerned with apical growth and one with cambial. Aspects of cam- 
bium physiology were discussed by Priestley in a series of papers ( 1930 
and others). 

A continuous cambium, laying down a solid ring of vascular tissue, is 
found in all woody gymnosperms and angiosperms and in many herba- 
ceous forms. In other herbaceous stems the cambium ring is discontinu- 
ous and produces distinct vascular bundles. These may be separated by 
undifferentiated fundamental tissue. Across these gaps and connecting the 
cambium of one bundle with that of the next there is often an inter- 
fascicular cambium, probably a vestigial structure persisting from the 
time when the cylinder was continuous and the stem was woody. In many 
cases it consists of a row of cells with only a few tangential divisions. 
Under suitable conditions, as in the base of a stout herbaceous stem, the 

Meristems 87 

interfascicular cambium may become active and produce typical xylem 
and phloem. Its development offers some interesting morphogenetic 


The distribution of cambial activity over the axis in woody plants also 
deserves further investigation. The amount of wood produced by the two 
branches at a given fork of a stem, for example, bears a rather close ratio 
to the amount in the main axis below them, but this ratio will depend on 
the relative size of the two branches, the angle between them, and the 
orientation of the main axis itself ( p. 108 ) . 

Another relatively unexplored field but one which may become of much 
interest for morphogenesis is that of anomalous secondary growth. In 
certain families the normal situation of a continuous cambial sheath is 
altered (Mullenders, 1947). An additional cambium may arise outside the 
phloem and start another vascular cylinder or series of bundles. There 
may be more than one of these. Such anomalous bundles may also appear 
in the pith. In other cases the surface of the cambium, instead of being 
circular in cross section, may be irregular so that radial lobes of secondary 
tissue are formed. In more extreme cases the cambium may become quite 
atypical and patches of secondary xylem and phloem may be intermingled 
in the vascular cylinder. Anomalous growth is often found in stems such 
as those of lianas or rhizomes which have other functions than support 
or conduction. It is frequently present in fleshy roots. One sometimes has 
difficulty in drawing a line between anomalous growth of this sort, 
which is really normal for a particular plant, and truly abnormal, or tera- 
tological, structures (Chap. 11). 

The vascular cambium has been little explored from a morphogenetic 
viewpoint. Although it is much more difficult to study directly than are 
the apical meristems, its products, especially wood, are so firm and meas- 
urable that they offer attractive material for a quantitative study of many 
problems in growth relationships. 

The Cork Cambium. The vascular cambium and the root and shoot 
apices are not the only localized embryonic regions in the plant. Increas- 
ing diameter of the axis necessarily results in the rupture of its outer 
layers, notably the cortex and the outer phloem. Infection and water loss 
would take place through these breaks in the tissue were it not for the 
formation of layers of suberized cells, the cork, or phellem. This is sec- 
ondary tissue formed by a cork cambium, or phellogen. It has its origin in 
a row of cells, tangentially adjacent to each other, which divide peri- 
clinally and link up into a meristematic layer that produces a series of 
daughter cells on its outer side. There may be from one or two to many 
of these and their walls become suberized and impervious to water. On 
the inside are formed one or a few layers of daughter cells, the phello- 
derm, presumably vestigial in character. 

88 Growth 

The first phellogen usually arises just under the epidermis, but as axis 
diameter increases and outer tissues are ruptured, new phellogens appear 
in the deeper layers. Sometimes these form a single continuous layer of 
cork which may be peeled off. In other cases the phellogen arises as a 
localized, often slightly concave sheet which isolates a scale-like patch of 
tissue. In older stems these phellogens appear in the earlier and more or 
less crushed and functionless phloem. The phellogen cells obviously are 
alive and must originate in still living cells of this outer tissue. In many 
cases, corky cells arise beneath wounds in various regions. In the abscis- 
sion of leaves and fruits a layer of cork is formed at the region of separa- 
tion. In both these cases, growth substances (wound hormones or auxin) 
have been shown to be related to the origin of the cork cambium. Corky 
layers often show an unusual histological trait in having the new division 
walls in tangentially neighboring cells laid down directly opposite each 
other instead of being staggered, thus forming a characteristic stratified 
structure unlike that of most plant tissues (p. 195). 

From a morphogenetic point of view the most interesting thing about 
cork-forming cambia is the way in which a continuous layer of such cells 
may suddenly arise in a mass of old and partially collapsed tissue. A host 
of dormant cells become embryonic again, link themselves up with neigh- 
boring cells, and form a cambial layer. In the typical rhytidome form of 
bark, this may be somewhat irregular in outline and often is not closely 
parallel to the surface of the organ or to the vascular cambium below. 
Where cork forms under a wound or just below the epidermis, its position 
may be explained by its location at a particular point in a physiological 
gradient, but in these more complex cases such an explanation is less 
satisfactory. Their origin resembles the way in which a pattern of wall 
thickenings or a net of fibers (p. 197), which transcends cellular bound- 
aries, may become differentiated in a mass of tissue. The origin of such 
phellogen layers is a problem in differentiation which deserves more 


Potentially, the plant axis can grow indefinitely in length through the 
activity of its apical meristems and in width through the activity of the 
vascular cambium. Actually, of course, growth finally ceases for various 
reasons, but these axial meristems are essentially indeterminate in their 

The organs of the plant other than stem and root, however— the leaves, 
floral parts, and fruits— are structures of limited, or determinate, growth. 
They finally reach maturity and cease to enlarge. In this respect, one of 
them is much like an animal individual with a definite life cycle of its 

Meristems 89 

own. Such organs provide an opportunity for a study of plant develop- 
ment which has been somewhat neglected in favor of the more sharply 
limited meristems in apex and cambium. Because of the more diffuse 
character of their growth, a study of these determinate organs will prob- 
ably throw more light on the development of form than can be gained 
from those of indeterminate growth. 

How, we may ask, does a determinate organ grow? Is it through the 
activity of localized groups of dividing cells, as in the axes, or by un- 
realized, interstitial growth, as in most animals? The fact is that both 
methods are usually employed. 

The determinate organ which has been most extensively studied is the 
leaf. The first step in its development is the appearance of a small swell- 
ing just below the dome of the shoot meristem. This grows into a leaf 
primordium and finally, through a series of developmental steps, into a 
mature leaf (p. 187). As to just how much of the meristematic tissue 
actually takes part in forming a leaf primordium, there seems to be con- 
siderable variability among different groups of plants. Rosier (1928) 
reports that in wheat only the outermost layer (dermatogen) is con- 
cerned. This pushes out and then pulls together from all sides to meet in 
the center, like a collapsing glove finger, so that the whole leaf grows 
from this one layer. Schwarz (1927), on the contrary, found that in 
Plectranthus and Ligustrum the first two layers produced the entire leaf, 
and this part of the meristem he termed the phyllogen. Most other 
workers ( see Foster, 1936 ) have found that tissue below the second layer 
also contributes often to the formation of the young leaf, particularly the 
veins. Whether this is simply tunica or both tunica and corpus depends 
on the extent of layering and seems not to be important. 

Critical evidence in this problem is provided by a study of leaf pri- 
mordia formed by periclinal chimeras (p. 268). Here one or two of the 
outer layers come from one of the graft partners and the rest from the 
other. In chimeras between nightshade and tomato the tissues from each 
can be distinguished by the fact that in nightshade the cells are much 
larger. Here Lange ( 1927 ) was able to show that although a leaf primor- 
dium in this chimera was formed chiefly from the two outer layers the 
third layer also contributed to it. In periclinal chimeras between forms of 
Datura stramonium differing in number of chromosome sets (and thus in 
cell size) Satina, Blakeslee, and Avery (1940) observed the same thing, 
as did Dermen ( 1947« ) in cranberry. In all these cases the third layer 
gave rise to the vascular tissue of the leaf. 

The way in which the primordium develops into the leaf also differs 
considerably in different forms. In fern leaves, growth of the lamina is 
largely determined by an apical cell resembling that in the shoot and 
root (p. 58). In many higher plants the early growth of the primordium 



in length is chiefly apical and seems to be governed by a group of cells 
at the tip, just under the epidermis-essentially a meristem. This produces 
the central tissues of the young primordial axis, or midrib, and growth of 
the epidermis keeps step with it. Such apical growth soon ceases, how- 
ever, and later growth is diffuse. 

The development of the tobacco leaf studied by Avery ( 1933 ) may be 
taken as a typical example of the growth of a determinate organ (Fig. 
4-24 ) . After the axis is about 1 mm. long and while still it is very narrow, 
growth of the lamina begins on both sides of this axis, pushing out like a 
wave of developing tissue. It increases faster in the middle than at either 




Fig. 4-24. Early developmental stages of a tobacco leaf, from a young primordium 
(upper left) to later ones where lamina and veins are being formed. (From G. S. 
Avery. ) 

end, and this produces the characteristic leaf shape. The rate of growth, 
as a result both of the division of the cells and of their increase in size, 
is greater in certain dimensions than in others. Avery contrasts the growth 
differences resulting from such polarized growth, primarily due to differ- 
ences in plane of cell division, with localized differences in rate of divi- 
sion. Differences in cell shape due to differential cell expansion have little 
share in over-all shape changes in the organ as a whole. In most leaves, 
growth in the various dimensions of the blade, whatever its cellular basis, 
is unequal, so that blade form changes somewhat during development. 
These changes are under morphogenetic control, however, and show close 

Meristems 91 

allometric relationships ( p. 105 ) . The development of a few similar organs 
of determinate growth has been studied, such as the thorn shoots of 
Gleditsia (Blaser, 1956). 

In leaves which do not show the usual dorsiventral character but vari- 
ous, more complex shapes, such as pitcher-like or peltate blades, the 
origin of these structures is by a system of localized meristems (Roth, 
1957). These have not been studied extensively and present some impor- 
tant developmental problems. 

The growth of other organs, such as perianth parts, ovaries, and fruits, 
resembles that of leaves in showing certain localized differences in rate 
and direction of growth, but growth is mainly diffuse and nothing com- 
parable to a true localized meristem is operative save in exceptional cases. 
Not only are the different dimensions of such an organ clearly correlated 
in a progressively changing pattern but different parts of the organ, such 
as blade and petiole, and fruit and pedicel although often growing at 
different rates, also keep in step with each other. Growth of cer- 
tain structures, notably the fruit stalk, involves some cambial activity. 

Whatever type of growth a plant organ may show, whether by apical 
meristems, cambium, or diffusely distributed embryonic activity, it is 
under strict developmental control. The problem of this control is some- 
what more complex in a plant, where both diffuse and localized growth 
occur, than in an animal, where the latter is generally absent. The pres- 
ence in a plant of these two somewhat different methods of growth offers 
certain advantages because of the possibility of studying in the same in- 
dividual two different types of morphogenetic control. However growth 
occurs, its activities are correlated and not isolated events. 


The Phenomena of Morphogenesis 



The most significant fact about organic growth, as described in the pre- 
ceding chapters, is that it is a process under definite control and thus leads 
to the development of bodies of definite form and size. This control is 
shown in the character of the growth cycle itself which, as we have seen, 
marches forward in an orderly fashion to the attainment of a specific size. 
It is shown even better in the distribution of this growth during develop- 
ment. If growth were equal in all parts and directions, organisms would 
be spherical. The remarkable variety of forms that living things display 
and that constitutes one of their important differences from most lifeless 
objects is due to the fact that the amount of growth in one region is dif- 
ferent from that in another and that its rate in the various dimensions of 
a structure is unequal. These differences are not random ones, for if they 
were, a jumble of fantastic forms would result; they arise in an orderly 
sequence and progress in a regular fashion until a specific organic struc- 
ture is developed. Something evidently guides the growth and differenti- 
ation of the organic mechanism. Occasionally this control is seriously 
disturbed and in such cases abnormal growths and monstrosities of various 
kinds appear (p. 275), but in the great majority of cases orderly develop- 
ment and the production of specifically formed structures take place. It is 
clear that in some fashion the parts of an organism are so related to each 
other that a change in one affects the rest and that the whole is thus inte- 
grated into an organized system. 

All the phenomena of development which are to be discussed— polarity, 
symmetry, differentiation, regeneration, and the rest— are simply different 
aspects of this developmental relatedness, and the various factors con- 
cerned are those which have been found to affect it in one way or another. 
The fundamental causes of this integrated development are yet unknown. 
They are often attributed to correlation, a term which, because it is in 
most cases merely a name for our ignorance, has with many students of 
morphogenesis fallen into disrepute. Nevertheless correlation is a fact, 
explain it how we will, and no one can approach a study of the phe- 
nomena of morphogenesis without recognizing this. Therefore at the be- 


96 The Phenomena of Morphogenesis 

ginning of a discussion of these problems it will be useful to consider 
some typical examples of the ways in which growth of one part or dimen- 
sion is related to growth elsewhere or to the plant's various activities. 
These will illustrate how plant forms arise and an integrated organism is 
produced and will serve as an introduction to the fundamental problem 
of morphogenesis, approached from many directions throughout this book 
—the problem of biological organization. 

There have been many discussions of correlation in the literature of 
plant development, and for some of them the reader is referred to the 
works of East (1908), Harris (1909-1918), Love and Leighty (1914), 
Murneek (1926), Goebel (1928), Thimann (1954b), and others. 

Correlations have been classified in many ways, as environmental, 
physical, morphological, physiological, genetic, compensatory, or meristic, 
depending on the characters and factors involved. For purposes of con- 
venience in the present treatment, there will be grouped together, as 
physiological, those correlations for which a physiological mechanism- 
metabolic, hormonal, or other— seems to be operative and as genetic, those 
which seem to depend primarily upon the genetic constitution of the 
individual and its formative relationships and are thus produced by 
mechanisms more deeply seated and obscure than the ordinary physio- 
logical ones. 


Physiological relationships are of particular morphogenetic interest 
since through an analysis of them the mechanisms for other types of cor- 
relation may be discovered. The various factors concerned will be treated 
in later chapters. The particularly important role of growth substances in 
plant correlation has been discussed by Thimann ( 1954fo ) . 

It will be useful here to mention a few typical examples of physiological 
correlation and to formulate some of the problems that they present. 

Nutritional Correlations. The simplest type of correlation is one which 
depends on nutrition. A region that does not produce or contain food must 
depend for its growth on one that does. Correlation of this sort between 
root and shoot must obviously occur. The root-shoot ratio is a favorable 
one in which to study correlation and the factors that modify it, and con- 
siderable attention has been given to the problem. Kny (1894) cut off 
part of the roots from growing seedlings and part of the shoot from others. 
When a considerable amount of reserve food was still available in the 
seed, loss of one part did not greatly affect the growth of the other. 
Pearsall (1923), Keeble, Nelson, and Snow (1930), and others found 
that removal of the seedling shoot sometimes actually stimulated growth 
of the root, presumably because of reduced competition for food stored 

Correlation 97 

in the seed. More commonly, however, in older plants and in cuttings, a 
rather close balance becomes established between root and shoot and is 
restored if altered experimentally. 

This ratio is subject to change during development, for in most plants 
the shoot grows consistently faster than the root. Other factors also affect 
it. In poorly nourished plants the root is relatively large and in etiolated 
ones, relatively small. Crist and Stout (1929) found that in some plants 
it was affected by soil acidity, soil fertility, and day-length. Roberts and 
Struckmeyer (1946) observed that temperature and photoperiod modified 
the ratio but not in the same way in all plants. The top-root ratio was 
studied by Shank (1945) in maize inbreds with low and with high ratios, 
and in their hybrids, under different amounts of phosphorus, nitrogen, and 
water in the soil. Increase in each of these substances tended to increase 
this ratio. Richardson ( 1953 ) measured root growth microscopically in 
small maple seedlings growing in glass tubes under controlled conditions. 
Any change in the environment of the shoot which modified photosyn- 
thetic activity had a commensurate effect on rate of root growth. Correla- 
tions depending on nutrition are evidently rather susceptible to change 
by environmental factors. 

The influence of shoot on root is not always nutritional but may result 
from the action of auxin, vitamins, or other growth-regulating substances. 
The nutritional influence of root on shoot is well shown by the horticul- 
tural practice of producing dwarf trees by grafting scions from normal- 
sized varieties on roots of genetically dwarf types in which the root system 
is too small to supply the growth requirements of a large tree. 

Among other correlations which have their basis in nutritional factors 
are those between the size of a fruit and the amount of leaf area avail- 
able for the support of its growth (Haller and Magness, 1925). There is 
also a close relation between the amount of foliage on a tree (the size 
of its crown ) and the amount of stem growth. Young and Kramer ( 1952 ) 
and Labyak and Schumacher (1954) have studied this problem by re- 
ducing experimentally the size of the crown in pine through pruning and 
observing the effect on trunk growth. In apples, Murneek (1954) found a 
relationship between fruit size and leaf area (presumably nutritive) and 
also between fruit size and seed number per fruit (presumably stimu- 
latory ) . 

Because of its practical importance, many studies have been made of 
the relation between the size of seed planted and the size of the plant 
growing from it. If a positive correlation existed between these characters 
it would pay to use only large seeds in many agricultural operations. 
Agronomists have sought all such characters in fruit and seed that might 
be correlated with high yield but have had little success. Where such a 
relation has been found, in most cases it is simply between seed size and 

98 The Phenomena of Morphogenesis 

early plant size. Passmore (1934), working with reciprocal hybrids be- 
tween large-seeded and small-seeded cucurbits, and Oexemann (1942), 
with several vegetables, observed that plants from large seeds have an 
initial advantage in size because of the larger amount of food stored in 
the seed but that this usually disappears after a time. 

Similarly, in vegetative propagation the size of a "seed" piece in pota- 
toes, though it may influence early sprout growth, has no effect on yield 
(Wakanker, 1944), though if the bigger pieces have more buds on them, 
sprout number will be larger and yield somewhat increased. 

A positive correlation between the size of a fruit and of the seeds in it 
has often been found, as by Schander (1952) in apple and pear and by 
Simak ( 1953 ) for seed size and cone size in pine. In fruits and cones of 
the same size, however, seed size was inversely proportional to seed num- 
ber. Both nutritional and compensatory correlations are probably in- 
volved here. 

Ashby (1930) suggested that the larger plant size resulting from 
heterosis was due to greater size of the embryos that produce the heterotic 
plants, thus giving them an initial advantage which was maintained 
throughout growth. Present evidence, however, does not support this idea. 

Compensatory Correlations. The nutritional factor in the relation be- 
tween two parts of a plant may be evident in other ways than by transfer 
of food from one to another. Each growing part or organ constitutes what 
Goebel called an "attraction center" which under normal conditions draws 
to itself a specific amount of building material. This may be small or large, 
depending on its genetic constitution. In one of the higher plants, 
which has many similar growing parts such as leaves, flowers, and fruits, 
the number of these parts may be reduced by accident or experiment. In 
such cases there is often a compensatory increase in the growth of the 
remaining structures, so that a negative correlation results between the 
size and the number of parts (Lilleland and Brown, 1939). Thinning of 
fruits by mechanical or chemical means is sometimes practiced so that 
the remaining fruits will grow larger. In the same way, the removal of all 
buds but one in a certain type of chrysanthemum results in the develop- 
ment of this single flower head, through compensatory growth, to a size 
very much larger than normal. 

The reverse of this relationship also may occur, for if many fruit are 
set, they will be small. In such cases, some may drop off. Thus in apples 
there usually occurs a "June drop" in which many of the young fruit, 
unable to attract to themselves a sufficient supply of food or auxin, stop 
growing and are cut off by abscission layers. In a somewhat similar way, 
the more seed developing in a tomato fruit and the more fruits in a 
cluster, the smaller will be the weight of each seed (Luckwill, 1939; 
Schander, 1952; and Simak, 1953). 

Correlation 99 

Where flowering and fruiting are continuous, as in squashes, if a certain 
number of fruit are set, related to the food-producing capacity of the 
plant, the development of more flowers ceases and will not be resumed 
unless the growing fruits are removed. There is thus a continuous compen- 
satory balance between the development of multiple plant organs and 
the amount of material or hormone available for their growth. 

A balance also occurs between the vegetative and the reproductive 
phases of a plant ( Murneek, 1926 ) . Tomatoes in which fruits are allowed 
to form abundantly will soon cease vegetative growth, but if flowers and 
young fruits are continually removed, the plants will grow to a much 
greater size. A potato plant in which tuber formation is prevented will 
often bear a large crop of fruits, structures which normally fail to develop 
presumably because of the diversion of food to the tubers. Mirskaja 
( 1926 ) removed all flower buds from plants of a number of species and 
found that this stimulated formation of lateral shoots and increased the 
size of leaf blades, tubers, and pith cells and the amount of lignified 

Removing the axillary buds from Coleus plants was found by Jacobs 
and Bullwinkel (1953) to induce longer stems, larger leaves, and more 
rapid growth of the main shoot ( Fig. 5-1 ) . The ancient art of topiary is 
simply a manipulation of these compensatory correlations. The removal 
of certain buds stimulates the growth of others which would have re- 
mained dormant, and by this means the form of the plant can be 

Such correlations may perhaps be called competitive rather than com- 
pensatory. In certain hybrid cherries, for example, the embryos start their 
development but when partly grown they shrivel and die. Tukey (1933) 
and others were able to bring such embryos to normal maturity by re- 
moving them from the seed and growing them in culture. In normal plants 
the embryo may be thought of as competing successfully with maternal 
tissues for food during development, but in these unusual cases most of 
the food is drawn instead to maternal tissues, and the embryo dies. Re- 
lease from such maternal competition allows it to grow. 

Compensatory correlations are also to be observed in the development 
of individual organs. MacDougal (1903k), who has reviewed the early 
literature, described many examples of this, as did Goebel and others. In 
some plants, for example, if the blade is removed from the young and 
growing leaf, the stipules will become much enlarged. The building ma- 
terial available to the leaf is employed in its growth but the distribution 
of this material is not the usual one. 

A good instance of compensation is reported by Johnston ( 1937 ) be- 
tween the coleoptile and the first internode of Avena. Light stimulates 
the growth of the former but depresses the latter. Regardless of light, the 

100 The Phenomena of Morphogenesis 

total growth is much the same, reduction in one structure being compen- 
sated by increase in the other. 

Stimulatory Correlations. Many correlations, however, do not depend 
upon the distribution of building materials but upon the operation of 
other factors which affect development, particularly the stimulatory and 
inhibitory action of auxin and other growth substances. 

The stimulatory effect is well shown in the control of root growth. 
Van der Lek (1925) and others have found that in many cases cuttings 

160 - 

140 - 






A I 2 

Leaf Position 

Fig. 5-1. Compensatory correlation in Coleus. Increase in leaf length in 27 days after 
removal of axillary buds and branches, as compared with controls. Leaves at left of 
vertical line had not unfolded from apical bud. ( From Jacobs and Bullwinkel. ) 

on which buds are present will root much better than those without buds. 
This evidently is due to a root-stimulating substance produced by buds 
which passes down to the base of the cutting. The character of the buds 
may also be important, for O'Rourke (1942) has shown that blueberry 
cuttings root better if the buds on them are leaf buds than if they are 
flower buds. 

The relation between a leaf and the development of a bud in its axil 
is a complex one. Felber (1948) observed that in apple the size of a 

Correlation 101 

vegetative bud at maturity is proportional to the size of its subtending 
leaf, suggesting a nutritional relation. Champagnat (1955 and other 
papers) presents evidence that there are several distinct stimulating or 
inhibiting influences that the leaf exerts on its bud. Snow and Snow 
(1942), on the basis of experiments at the meristem, believe that an axil- 
lary bud is determined by the primordium of the leaf that subtends it, 
particularly the basal part. If the primordium is partially isolated from 
the stem apex, its bud grows larger than it otherwise would. 

Related structures often affect each other. The cotton boll and its seeds 
will not reach normal size if the involucre of the flower is removed 
(Kearney, 1929). Knapp (1930) reports that the perianth of a liverwort 
grows only if the archegonium that it covers is fertilized. The ovary in 
most plants will not grow into a fruit unless at least a number of ovules 
are forming seeds. These developing parts produce substances, appar- 
ently, that stimulate the ovary wall to grow. This stimulation can be 
imitated by the use of certain synthetic growth substances to produce 
parthenocarpic fruits (p. 378). In case of metaxenia (p. 407), where the 
male parent has a direct effect on the character of the fruits, this pre- 
sumably results from something introduced through the pollen tube. 

Inhibitory Correlations. There are many developmental relationships 
which are just the reverse of stimulatory and in which one part inhibits 
the growth of another by some other means than competition for food. 
These relations, like those of stimulation, commonly involve the action 
of auxin and related substances. 

The best known case of such inhibition is the dominance by a terminal 
bud which prevents the growth of lateral buds below it (p. 386). Simi- 
larly, the epicotyl and its bud, in seedlings like those of beans, inhibit the 
growth of buds in the axils of the cotyledons. Often a leaf can be shown 
to inhibit the growth of the bud that it subtends, for if the inhibiting 
organ is removed, the bud will then grow. Sometimes physiological iso- 
lation has the same effect as removal. Child ( 1919, 1921 ) chilled a portion 
of a bean epicotyl and found that the cotyledonary buds then began to 
grow. Shading a leaf sometimes results in removing its inhibiting in- 

Preventing the growth of the apical bud by encasing it in plaster some- 
times has the same effect as removing it. Many of the early studies in 
growth correlation involved this plaster technique (see Hering, 1896). 
For example, if the portion of pea epicotyl between the terminal bud and 
the cotyledons is so encased that it cannot grow in width, growth in 
length is much reduced as compared with the control. 

Nutritional factors may have something to do with the inhibition of 
cotyledonary buds, for Moreland ( 1934 ) observed that in bean seedlings 
the growing foliage leaves have a greater inhibiting effect on these buds 

102 The Phenomena of Morphogenesis 

than does the epicotyledonary bud itself and believes that this is owing 
to the removal by these leaves of some food material necessary for bud 


Other structures may be inhibited. If root nodules and root tips are re- 
moved from roots of red clover inoculated with an effective strain of 
nodule bacteria, the number of nodules subsequently formed will be in- 
creased (Nutman, 1952). This is thought to be owing to the removal of 
inhibitory activity centered in the meristems of nodules and root. 

Inhibition by terminal buds has various practical implications. Reed 
( 1921 ) found that heavily pruned young pear trees have a greater growth 
of new shoots than do unpruned or lightly pruned ones and suggests that 
this results from the removal from them of much growth-inhibiting sub- 
stance present in the buds near the tips of the branches. 

Correlations of Position. Many parts of the plant can be shown to have 
the capacity for much more extensive growth than they normally display. 
If a leaf is removed and treated as a cutting, it will frequently grow to a 
greater size and live much longer than if it had remained a part of the 
plant (Mer, 1886; Riehm, 1905; Winkler, 1907; and others). Single cells, 
under suitable conditions of isolation and stimulation, will sometimes de- 
velop into whole plants. All parts of the plant tend thus to be totipotent. 
Why these potentialities are not realized when the part is a member of 
an organic whole is a problem. Not only is each part of this whole limited 
in its growth, but the particular way in which it develops depends on 
where it is. Driesch's famous dictum emphasizes the fact that an or- 
ganism is an organic pattern in which every part develops in a specific 
relation to the rest. The correlations that these parts display with one an- 
other are simply manifestations of the control that this pattern exercises 
in development. 

Experimental change of the position of a part in this pattern often 
effects marked alteration. Ward and Wetmore (1954) partially released 
young fern embryos from their contact with the prothallus and found 
their growth to be slower and somewhat abnormal. Wetmore asks the sig- 
nificant question as to why a spore and an egg should grow so differently. 
Each is a haploid cell and they presumably are identical genetically, but 
the surroundings under which they develop are very different. He sug- 
gests that perhaps the difference between sporophyte and gametophyte 
in ferns may be the result of this positional correlation. 

Mason (1922) reports that the terminal bud from a cotton shoot that 
has stopped growing will grow vigorously if it is budded on a young plant. 
A flower bud inserted on a vegetative shoot where it would not normally 
occur often changes in its development and may produce a flower cluster 
which is gigantic or otherwise abnormal. 

The operation of such a constantly regulated balance among activities 

Correlation 103 

is well shown in the formation of reaction wood (p. 356). This wood (in 
conifers) elongates faster than normal wood and thus tends to bend a 
branch away from the side on which it occurs. The branches have a spe- 
cific angle of orientation to the main axis, or to gravity, which is main- 
tained by the development of reaction wood on one side of the branch. 
If this normal orientation is experimentally altered, reaction wood will 
appear at the precise place and in the precise amount elsewhere which 
will tend to restore the normal branch pattern. The origin and character 
of this pattern are the essential problem. What happens to any com- 
ponent of it depends on the place that this occupies. 

There are many other examples of the operation of such developmental 
patterns in the plant body. Among these are the studies of Dormer ( 1950) 
on the development of xylem in different internodes of the young plant of 
Vicia; of Friesner and Jones ( 1952 ) on the relation of primary and sec- 
ondary branches in length growth; and of various workers on the struc- 
ture of leaves borne at different levels on the stalk. Ashby and his 
colleagues (1948) have emphasized the structural and physiological dif- 
ferences among successive leaves along the axis and have related this to 
the problem of aging. Instances of positional differences shown in topo- 
physis (p. 212) are particularly clear and may become irreversible. 

The control that the organized whole exercises over its parts is some- 
times termed "correlative inhibition." This term explains nothing but it 
emphasizes the fact that inhibitory action is certainly involved. In the 
physiology of development both inhibition and stimulation are important. 
A number of workers, among them Libbert (1954, 1955), have discussed 
the various interactions between substances which promote and those 
which check the growth of buds. Thimann (1956) has called attention to 
the fact that in most physiological processes there is a balance between 
reactions tending to promote the process and others tending to inhibit it. 
No single factor is solely responsible, but physiological activities, includ- 
ing those of development, are often under multiple control. Furthermore, 
certain factors such as auxin may stimulate under certain conditions and 
inhibit under others. 

Some students of development are therefore inclined to look on the 
growing organism as the seat of constant competition between different 
and distinct processes, a state of equilibrium between opposing forces. 
This resembles the concept of the organism as a balance between distinct 
cellular individuals each with specific tendencies of its own. It also calls 
to mind the older idea of the "battle between the parts" as the basic fact 
in development. Analysis of the structures and the activities that go to 
make up an organism gives some support to this interpretation of devel- 
opment. The close correlations that are everywhere present in develop- 
ment, however, and particularly the persistent tendency toward regula- 

104 The Phenomena of Morphogenesis 

tory action by which a specific norm or pattern of form and function is 
restored if disturbed, are difficult to explain on the basis of independent 
action by many variables. The organism more closely resembles an or- 
ganized army under disciplined control than it does a mob where each 
individual acts competitively for himself. 

The balance between stimulation and inhibition, however, is worth 
careful study by students of morphogenesis. In a few cases it has been 
investigated in the lower plants. In the coenocytic alga Caulerpa, for ex- 
ample, the "assimilators" (leaves) produce strong growth inhibition but 
the rhizoids have the opposite tendency (Dostal, 1945). The balance 
between the influences of these two sets of organs has an important effect 
on the character of the plant as a whole. 

More favorable material for a study of this aspect of correlation is 
found in the over-all form of the plant body, especially in such higher 
plants as trees. A tree is a rather loose aggregation of axes which usually 
does not show as precise a form as does an individual organ such as a 
leaf or a flower but which, nevertheless, is characteristic and recognizable. 
This has been found to result from an interaction of factors in the ter- 
minal buds and in the growing tips of the branches. Some of these factors 
tend to push the branches down, in relation to the main axis, and others 
tend to lift them up. The relative length of branches and main axis is also 
similarly controlled, evidently by domination of the terminal bud over 
those of the lateral branches. Munch (1938) has discussed the diverse 
tendencies that govern such tree form in conifers and interprets these in 
terms of hormonal action, but he emphasizes the harmony and balance 
that exist among them. Others (Snow, 1945) have considered the problem. 
It is a basic one for morphogenesis since the form of the plant body as a 
whole, although relatively variable, is nevertheless a true organic one. 
Presumably the factors that govern it resemble those that bring about the 
much more constant and specific forms of the separate organs. The body 
is an aggregation of these parts, less tightly organized than are its indi- 
vidual organs, but clearly showing organization. The beginnings of organ- 
ization and of the emergence of those correlations that determine form 
may profitably be studied in these plant bodies, which in a sense are in- 
termediate between colonies of semi-independent parts and true organic 


Organized bodily patterns, with their localized differences and specific 
characters, are examples of physiological correlation though the mecha- 
nisms involved are obscure. They doubtless have some genetic basis. Many 
other growth correlations, including those concerned in the form of plant 

Correlation 105 

parts, are more precise but are even further from a satisfactory biological 
explanation. They are inherited, but the genetic mechanisms involved 
have hardly begun to be explored. At present we can simply describe and 
classify these correlations. 

The various structures in a growing organic system tend to increase 
together and thus to be correlated in size. In a given organ its dimensions 
are likewise correlated. Since growth usually is not uniform, as develop- 
ment proceeds, the relations between the parts of the system or between 
the dimensions of the organ may change progressively and thus produce 
differences in form. Growth is usually exponential in character, and there- 
fore the relationship between the sizes of two structures growing at dif- 
ferent rates may best be found by plotting the logarithms of their sizes 
against each other. If the rates are different but the relation between the 
two is constant, these values will fall along a straight line the slope of 
which measures the growth of one structure relative to that of the other. 
It is noteworthy that in most cases where two parts of the same growing 
system, or two dimensions of a growing organ, are compared, their 
relative rates are found to be constant, however different their absolute 
rates may be. 

This relationship can be described simply by an equation. If y is the 
size of one variable, .t that of the other, b the value of y when x is of some 
arbitrary size, and k the ratio of the growth rate of y to that of x, then 

y = bx k 

log y — log b + k log x 

This phenomenon of constant relative growth (heterauxesis) has been 
observed by many biologists but was first widely emphasized by Julian 
Huxley (1932). He termed this type of growth heterogony, a term now 
replaced in much of the literature by allometry. The constant b measures 
differences in level, or at the beginning of growth, between two variables. 
The constant k provides a measure of relative growth rate and may some- 
times offer a clue to the mechanisms involved. It may be used to express 
differences when these are based on genetic, environmental, embryologi- 
cal, biochemical, or even evolutionary factors. This method of analysis 
has proved useful in the study of many kinds of growth correlations. 

Correlations of Part and Whole. Among the familiar growth correla- 
tions are those between an organ and the rest of the body or between 
members of a series of multiple parts and the structure that they con- 
stitute. In animals, large individuals typically have their organs cor- 
respondingly larger than those of small ones. In plants, however, with 
their lower level of organization, their often indeterminate growth, and 
their multiple organs, this relationship is not so simple. In beans, for 

106 The Phenomena of Morphogenesis 

example, Sinnott (1921) has shown that there is a positive correlation 
between size of leaf and size of entire plant up to a certain plant size. 
Beyond this the size of additional leaves is no greater even if leaf 
number and plant size may increase considerably ( Fig. 5-2 ) . Size of pod 
and of seed show a similar relationship to plant size. These facts suggest 
that organ size may depend on the size of the embryonic mass or the 
shoot meristem and that this may increase up to a certain point only, 
beyond which increase in total plant size involves only the addition of 
more units (internodes, leaves, and others). 

In cucurbits and many other types, although organs on the same plant 
tend to be correlated in size (forms with large fruits also having large 
leaves, thick stems, and long internodes), there is a certain amount of 
flexibility in these relationships, depending on genetic constitution. Thus 



30 50 

Dry weight of shoot (g ) 



Fig. 5-2. Relation of size of leaf to size of shoot in progressively larger bean plants. 
For a while, leaf and shoot increase together, but after a certain point, shoot size 
increases without further increase of individual leaves. {From Sinnott.) 

if a pumpkin type, which has all its parts large, is crossed with an egg 
gourd, where they are all small, the Fo generation contains plants that 
show some differences in the relative size of their parts, but there are 
none that have the large fruit size of the pumpkin and the small vine 
type of the egg gourd. The general physiological correlation of parts 
within the same plant makes it impossible for sizes of individual organs 
to segregate independently in inheritance. 

The size of the meristematic region bears some relation to that of 
plant and organ size. Crane and Finch ( 1930 ) have shown that the size 
of buds has an effect in determining the size of shoots that grow from 
them. In a comparative study of large-fruited and small-fruited races in 
Cucurbita pepo, von Maltzahn (1957) found that, although the dome- 
like undifferentiated meristem is essentially the same size in all types, 

Correlation 107 

the region just back of this and the primordia of flowers and leaves that 
originate there are considerably larger in the large-fruited type. 

In this general category of correlations are many of those described 
by J. Arthur Harris. In Nothoscordum and Allium (1909), for example, 
he found that flower clusters with relatively large numbers of flowers are 
borne on relatively long peduncles. Size changes are not always propor- 
tional, however, for in Ficaria (1918) he found that flowers with a large 
number of sporophylls have relatively more pistils than stamens. 

The relation between cell size and body size belongs to the part-to- 
whole category. This has been discussed in a previous chapter (p. 32) 
and is the basis of a very considerable literature. In a single organ 
there are often marked differences among the various tissues in the 
strength of the correlation between cell size and tissue size (Sinnott, 
1930). In general, it is clear that body size usually does not depend on 
cell size but on cell number. In many cases, however, it has been shown 
that in organs of limited growth, such as fruits, large cell size is associated 
with large organ size, though the range of the former is much less than 
that of the latter. 

Correlations between Different Parts. There are many genetic correla- 
tions which do not involve part-to-whole relationship but one between 
different parts and are thus less obvious as to origin. Sometimes these 
parts grow at the same rate but more frequently they do not. 

One of the most conspicuous of such growth relationships in higher 
plants is that between the two main organ systems of the body, the root 
and the shoot. Its nutritional aspect has already been discussed. The rela- 
tion is often so precise, however, as to suggest that it has a basis in the 
genetic constitution of the plant. Its value differs in different plants, at 
different stages of development, and under different environments. In 
most cases the root is relatively large in the seedling but grows less 
rapidly than the shoot. One increases at a rate which maintains a con- 
stant proportion to that of the other. Pearsall (1927) plotted the dry 
weight of the root against that of the shoot, both logarithmically, in a 
series of growing plants in various species. In most cases the allometric 
constant k was greater than 1, though its value differed in different 
species. In other words, the shoot grew more rapidly than the root. In 
etiolated plants, however, it was much greater than 1, and in those with 
storage roots, it was much less. 

Tammes ( 1903 ) made a study of the growth relationships between an 
internode and the leaf above it. She found that removal of a leaf would 
shorten the length to which the internode below would grow but would 
not reduce the number of its cells. This relationship does not hold in 
climbing plants, where internode length is usually attained before the 
leaf above becomes very large. In Ipomoea the excision of leaves on the 

108 The Phenomena of Morphogenesis 

main shoot has a variety of effects on the shape of leaves produced later 
by the terminal bud and on the size and number of their cells (Njoku, 


The various parts of the shoot system also show growth correlations, 
and these are responsible for the form of the shoot. They are readily 
observable in coniferous trees where the growth of the terminal shoot, 
which will form the trunk, is usually greater than that of the branches. 
This leads to the spire-like form of many of these trees. There are also 
definite relationships between the members of a branch system. The new 



Log. Lamina diam. 

Fig. 5-3. Allometric relation between lamina and petiole in Tropaeolum. I, a series of 
growing leaves. II, mature leaves in shade. ( From Pearsall. ) 

material added each year is distributed unequally but in regular fashion 
throughout the tree. 

A somewhat different type of correlation is that between the volume 
of the shoot system or any part of it and the cross-sectional area of the 
stem that supports it. Murray ( 1927 ) analyzed this relationship in a 
number of trees and finds that it is constant and predictable and that as 
the tree grows larger the cross-sectional area of its trunk becomes rela- 
tively smaller. It has been shown that where a trunk branches the cross- 
sectional area of the two branches is larger than that of the united trunk 



below them but that the degree of this difference depends on the rela- 
tive size of the two branches and on the angle between them. It is an 
expression of the polar tendency of the trunk. 

There are also correlations between the parts of an organ. In leaves, 
for example, although the length of the petiole is much more variable 













i i I j i 

l J r 1 1 1 1 1 1 1 1 










_i i i i_ 

Time in Days 

Fig. 5-4. Diameter of stalk plotted against diameter of fruit in a large-fruited race of 
Cucurbita (above) and a small one (below). Rate of growth is less for stalk than 
for fruit but at flowering the logarithmic distance between the two is approximately 
the same in both races. ( From Sinnott. ) 

than the dimensions of the lamina, there is a relation between them. In 
Tropaeolum, Pearsall ( 1927 ) found that this was allometric, with lamina 
width growing faster than petiole length (Fig. 5-3). In Acer the volume 
of the leaf blade is much more closely correlated with the cross-sectional 
area of the petiole than with its length. In the runner bean the area of 


The Phenomena of Morphogenesis 

the lamina and the cross-sectional area of the petiolar xylem are related 
allometrically, but the xylem grows only about 0.6 as fast as the lamina 
(D. J. B. White, 1954). Alexandrov, Alexandrova, and Timofeev (1927) 
observed that in Bryonia the number of vessels in any given part of the 
stem is correlated with the dimensions of the leaves in that region. These 
various facts suggest that physiological factors are here involved and 
that the amount of water transpired from the blade is important in de- 
termining the conducting capacity of the petiole. This hypothesis will 
be discussed later (p. 332). In the light of other evidence, it is doubtful 
whether such a "functional stimulus" is actually operative. 

Fig. 5-5. Relation of pith diameter to diameter of shoot in young stems of Pinus 
strobus of different sizes, showing greater relative size of pith in larger stems. ( From 
Sinnott. ) 

Somewhat similarly, the diameter of a growing cucurbit fruit and of 
the stalk that bears it are closely correlated in early growth, the fruit 
increasing more rapidly. Stalk growth ceases earlier than fruit growth, 
however (Sinnott, 1955; Fig. 5-4). 

Some correlations between parts are due to the similar effect of a 
gene or group of genes on a series of morphologically related organs. An- 
derson and de Winton ( 1935 ) studied the effect of a number of mutant 
genes, in Primula sinensis, on the morphology - of the leaf, bract, sepal, 
and petal. In several cases they had a very similar influence on develop- 
ment (producing lobing) in all four categories of organs. Such correla- 
tions are examples of what is sometimes called homeosis.. 

Many examples of growth correlation are found in internal structures. 
Thus Buchholz (1938) in Sequoia has shown that in stems of different 
sizes the vascular cylinder occupies a relatively larger portion, as meas- 
ured in cross section, in large stems than it does in small ones. In pine 



stems of different size the pith is relatively larger in the large stems and 
the cortex relatively smaller (Sinnott, 1936rt; Fig. 5-5). 

Sometimes these size relationships are found to extend below the level 
of the organ. The relation of cell size to nuclear size has already been 
discussed (p. 27). Klieneberger (1918) measured this relationship in 
a large number of plants, and the subject has been reviewed by Trom- 
betta (1942). Both Budde (1923) and Schratz (1927) found a rather 
close correlation between the total surface area of the plastids and the 
volume of the cell. 

These relationships between structures have important evolutionary 
implications which cannot be discussed here. The increasing size of the 
leaf during the development of the pteropsid stock seems to have been 
correlated with the change from a protostelic to a siphonostelic stem 
structure (Wetmore, 1943). The association of the trilacunar leaf trace 

Fig. 5-6. Diagram of a trilacunar node, 
showing relation between stipules and lat- 
eral leaf traces. ( From Sinnott and Bailey. ) 

with the presence of stipules (Sinnott and Bailey, 1914; Fig. 5-6) is 
another instance. This has been emphasized by the observation of 
Sensarma (1957) that when only one lateral trace branches only the 
stipule on that side develops. Another case is the relation of absolute 
size of the axis to its vascular development (p. 359). Among animals there 
are many examples of evolutionary allometry where increasing size of 
the organism results in a proportionally greater increase of certain organs. 
Correlations between Dimensions. Correlations between part and 
whole or part and part evidently involve coordinating mechanisms that 
bind these parts into an integrated organism. The same sort of control 
is shown, though in a somewhat different manner, in the correlation be- 
tween the various dimensions of an organ or other determinate structure. 
Here one is concerned with the very essence of form itself, with the way 
in which growth is distributed in one direction relative to that in an- 
other. This relative growth, like that between parts, is under definite con- 

112 The Phenomena of Morphogenesis 

trol and proceeds in a regular and orderly fashion (Fig. 5-7). Since 
most plant organs are determinate structures, their forms are more con- 
stant and precise than are those of the whole plant body. 

The origin of specific form in a plant or its organs may be studied in 
embryological development but more readily in the growth of organs that 
originate at the meristem, especially leaves, flowers, and fruits. In some 
cases the mature form or a close approximation to it is established very 
early, and from an examination of a tiny primordium, when its size may 
be only a fraction of a cubic millimeter, the final shape of the organ can 
be seen. The critical period in form determination here is evidently near 
the beginning of development. More frequently, however, the early 
primordium is simple, often nearly isodiametric, and the final form de- 
velops by differential growth. 





-f a * 
22 . 

z> * 
J N 

x * 





I I I I 

I I I Mill 


x 2 x 10V 

Fig. 5-7. Relative growth of length to width in developing fern prothallium. (From 
Albaum. ) 

Plant embryology in its widest sense is the record of such differential 
growth by which the complexity of organic form is attained. Most of 
our knowledge of the process is from verbal or pictorial descriptions, 
but in some cases it has been analyzed more precisely. The techniques 
of measuring allometric growth are as applicable in such cases as they 
are in the more frequently studied ones of part-to-part analysis. Richards 
and Kavanagh ( 1943 ) have extended the method further and show how 
it may be applied to three-dimensional growth. If this proves generally 
feasible, analysis of form development will become much more precise. 
Schuepp ( 1945, 1946 ) has used the methods of allometry to supplement 
others in a rather complex analysis of the development of leaf shape and 
of the origin of the leaf primordium at the meristem. 

Sinnott (1936fr) applied these methods to the study of form develop- 
ment in fruits of various races of cucurbits where form difference is due 



to differential growth rates in various dimensions ( Fig. 5-8 ) . In the long, 
narrow types, such as the "club" gourd, length increases faster than 
width but at a constant relative rate, the value of k being approximately 
1.2. In other races, such as the "bottle" gourds, width increases faster than 
length, k being about 0.8. In the latter race, which has an upper sterile 
lobe and a lower fertile one with a constricted isthmus between, the 
ratios of the diameters of these to each other and to the polar diameter 
of the ovary are specific, so that, as the fruit grows, not only the ratio of 
length to width changes but the form of the organ as a whole undergoes 


I to 

Width (mm) 

Fig. 5-8. Relative growth of length to width (plotted logarithmically) of developing 
fruits of several types of cucurbits. ( From Sinnott. ) 

precise development (Fig. 5-9). An organic pattern results not from one 
or a few correlations between dimensions but from a complex of such 
correlations. In crosses between the two gourd types mentioned, the value 
of k has been found to segregate after crossing and at least in one 
case in a simple fashion, suggesting that this is what is under direct 
gene control (p. 423). Evidently the form of the mature fruit in such 
cases depends not only upon the relative rate of its dimensional growth 
but upon the total growth attained, so that the problem of the inheri- 
tance of form involves not only the genetic basis of relative growth but 
also that of size. 

Dimensional relationships are not constant throughout the plant. Dif- 

114 The Phenomena of Morphogenesis 

ferences in shape are often found between early and later fruits, or be- 
tween leaves on different parts of the axis. Meijknecht (1955) has 
analyzed some of these differences and concludes that this variation is 
least when the structure occupies the position on the plant in which it 
shows its "ideal" development, the expression of its typical specific charac- 
ter. This calls to mind a concept of the early "idealistic" morphology. 




-, /0 



- 8 

5 ' 






6 T 8 9 /o 
Width (cm) 

40 SO 

Fig. 5-9. Relative growth of length to width in bottle gourds. Width increases faster 
than length, but the relative rate is the same in "miniature" (dots) and "giant" 
(crosses), so that although the shape of the two races at maturity is different, their 
genotype for shape is the same. The inherited difference between them is in size. 
(From Sinnott. ) 

The mechanism by which the control of relative growth is exercised 
and growth correlation established is not known, but evidently cell 
polarities are involved. Where growth is more rapid in one dimension 
than in another, it has been shown (p. 51) that cell divisions are more 
frequent in that direction. Whether the axis of the spindle or the cyto- 
plasmic polarity by which this seems in certain cases to be preceded is 
what is immediately involved in relative growth, or whether this is second- 

Correlation 115 

ary to some more general form-determining mechanism, is a basic prob- 

In this chapter there have been presented only a few of the great 
number of developmental, physiological, and genetic correlations that 
may be found throughout botanical literature, but these are representa- 
tive of the rest and emphasize an important fact in plant development. A 
plant is typically a rather loosely organized system but every part of it is 
nevertheless affected to some degree by its relations with other parts. 
These correlations are not random ones but are simple expressions of 
that general organized interrelatedness that is the distinguishing charac- 
ter of an organism. What happens to the whole affects the parts and what 
happens to a part affects the whole. An organ removed from this cor- 
relative inhibition may have a very different fate from its normal one. A 
single cell, on isolation, will often regenerate an entire plant. That it 
did not do so in its original position is owing to this inhibition. The term 
"correlation" is simply a description of the facts and explains nothing. 
It is of value, however, in emphasizing that the results of any experi- 
ment with a portion of the plant body must be interpreted not as an 
isolated event but as taking place against the background of the whole 
organism. How each portion of this organism behaves under given con- 
ditions and what its developmental fate will be depend upon its position 
in the organized system of which it forms a part. The nature of this or- 
ganized system is the fundamental problem that continually faces the 
student of morphogenesis in whatever part of the science he may be at 



In plant development, growth does not proceed at random to the produc- 
tion of a formless mass of living stuff but is an orderly process that gives 
rise to specific three-dimensional forms of organ or body. The various cor- 
relations described in the preceding chapter are manifestations of this 
formative control, which knits the developing organism together so that 
growth in one region or dimension is related to growth in the others and 
the plant thus becomes an integrated individual. A notable feature of 
these bodily forms of plants ( and animals ) is the presence in them of an 
axis which establishes a longitudinal dimension for organ or organism. 
Along this axis, and symmetrically with reference to it, the lateral struc- 
tures develop. The two ends or poles of the axis are usually different 
both as to structure and physiological activity. Thus a typical vascular 
plant has a major axis with the root at one end and the shoot at the 
other and with lateral appendages— leaves, branches, or lateral roots- 
disposed symmetrically around it. Growth is usually more rapid parallel 
to the axis than at right angles to it, so that an elongate form results, 
though this is by no means always the case. Single organs such as leaves, 
flowers, and fruits also show axiate patterns, as do the bodies of lower 
plants. These patterns appear very early in development as the result 
of differences in growth or in planes of cell division. This characteristic 
orientation of organisms, which is typically bipolar and axiate, is termed 

Polarity may manifest itself in many ways. The structures at the two 
ends of an axis are unlike, as in the case of root and shoot, "stem end" 
and "blossom end" of fruits, and petiole and blade of leaves. In re- 
generation, the organs formed at one end are usually different from those 
formed at the other. Cells and tissues may show polar behavior in graft- 
ing experiments. The transportation of certain substances may take place 
in one direction along the axis but not in the other, thus manifesting 
polarity in physiological activity. Both in structure and in function there 
are gradients of all sorts. Individual cells show polar behavior in plane 
of division and in the different character of their two daughter cells. 


Polarity 117 

It is important at the beginning to understand exactly what is meant 
by the term polarity. Sometimes this is regarded as an innate quality of 
an organism which makes its parts line up in a given direction, like iron 
filings in a magnetic field or opposite electrical charges at the two poles 
of an electrophoretic system. How far such polarizing factors operate in 
organisms we do not know. The term polarity as used most commonly, 
and certainly in the present discussion, implies much less than this and 
involves no assumption as to its causes. Polarity is simply the specific 
orientation of activity in space. It refers to the fact that a given biological 
event, such as the transfer of material through an organ or the plane in 
which a cell divides, is not a random process but tends to be oriented in 
a given direction. If this were not so, an organism would grow into a 
spherical mass of cells, like tissue in a shaken culture. This differential 
directiveness is responsible for organic form. What is the cause of it we 
do not know, but one often invokes it, although as an expression of igno- 
rance, in attempting to account for a morphogenetic fact. Polar behavior 
is no more and no less mysterious than organic formativeness but is merely 
the simplest manifestation of this, the tendency to develop a major axis 
with lateral ones subordinate to it. 

It is essential to realize, however, that polarity is not a trait that is 
originally and invariably present. There is good evidence that entirely 
undifferentiated cells, such as eggs in their early stages and other very 
simple ones, manifest no polarity at all. Within them, doubtless, there are 
polar molecules but these are arranged at random, like iron filings that 
are not in a magnetic field. Sooner or later a gradient is established in the 
cell which lines up these molecules in a specific orientation. This orien- 
tation originates in asymmetric factors in the outer environment, such 
as gravity, light, or the influence of adjacent cells, or perhaps within the 
cell from gene action. As a result, the various phenomena of polarity make 
their appearance, but not until a gradient has first been set up. Once a 
cell or a group of cells have thus become polarized, they will usually 
proceed to develop into an axiate system which then produces an or- 
ganic form without necessity for further environmental induction. 

The tendency toward polar orientation, which may be strong or weak or 
reversible and is differentially susceptible to outer influences, is the funda- 
mental fact of polarity. It must be distinguished from the various factors 
of induction that call forth and make manifest this polar tendency. To say 
that light induces polarity in the egg of Fucus is to describe a morpho- 
genetic fact, but a different problem is to explain the character of the 
cell that makes it capable of a specific polarization. An explanation of 
polarity in physical and chemical terms is difficult but a beginning at this 
task has already been made. In most biological discussions today, how- 
ever, the term polarity is primarily a descriptive one. 

118 The Phenomena of Morphogenesis 

Polarity is involved in many morphogenetic phenomena, and it will 
necessarily be referred to repeatedly in other chapters. Thus symmetry 
is the orderly distribution of structures in relation to a polar axis. Polar 
differences are the simplest aspect of differentiation. Regeneration is in 
most cases a polar process. Form results from a pattern of polarities set 
up in the developing plant. Polarity may be regarded as the framework, 
so to speak, on which organic form is built. 

The polar behavior of plants has long attracted the attention of 
morphologists and physiologists, from whose work a great body of knowl- 
edge has accumulated. Theophrastus and other ancient writers described 
the abnormal behavior of plants grown in an inverted position. There are 
a number of other early observations, especially those on regeneration 
following the girdling of trees, by Agricola, Hales, and Duhamel du 
Monceau. The term polarity was used by Allman in 1864 in connection 
with phenomena of animal regeneration and is now generally employed 
by students of morphogenesis in both botany and zoology. Vochting made 
extensive studies of polarity in plants in its relation to problems of re- 
generation, growth, and differentiation (1878, 1908, 1918). Important 
botanical work was also done by Goebel (1908), Janse (1906), Loeb 
(1924), and others. The theoretical aspects of polarity have been ex- 
tensively discussed not only by Vochting but by Sachs (1882), Pfeffer 
(1900-1906), Klebs (1903, 1904, 1913), Winkler (1900, 1933), Went and 
Thimann (1937), Lund (1947), Bunning (1952b), and others. Polarity in 
animals has been studied by many workers, notably Driesch, Roux, 
Morgan, and Harrison. Reviews of the field of plant polarity or parts 
of it have been written by Bloch (1943a), Gautheret (1944), and 
Bunning ( 1958 ) . Polarity is important not only for theoretical problems 
of plant development but for many practices of horticulture and vege- 
tative propagation (Priestley and Swingle, 1929). 

The establishment of a morphological axis in which the two ends are 
different and along which there is a gradation from one pole to the 
other may be looked upon as the first step in the process of differentia- 
tion, an important aspect of morphogenesis. The expression of polarity 
differs considerably in different plants and under different environmental 
conditions and is thus open to a wide range of experimental investigation. 
This most conspicuous aspect of organic form will probably not be fully 
understood until the mechanism of orderly and correlated growth control 
is discovered. As a relatively simple manifestation of form, however, it 
provides a useful point of attack on morphogenetic problems. 

Polar behavior in plants presents many problems. How far is it an in- 
herited character, potentially present from the beginning of development, 
and how far induced by the action of various environmental factors or by 

Polarity 119 

intercellular correlation? Once established, can it be reversed? Are the 
physiological manifestations of polarity the cause or the results of mor- 
phological polarity? Is polarity an aspect of the whole organism or do 
individual cells possess it? 

To present the various phenomena of polar behavior and the problems 
that they pose, it will be helpful to discuss the subject from several points 
of view and to describe its manifestations in external structure, internal 
structure, isolated cells and coenocytes, physiological activity, and the 
development of organic pattern. 


The most conspicuous expression of polarity is in external morphology. 
In higher plants the differences between root end and shoot end are de- 
termined very early, perhaps at the first division of the fertilized egg. This 
differentiation is not irreversible, however, for roots often appear on stems 
under favorable conditions and, less commonly, buds and shoots appear 
on roots. Polar behavior occurs in thallophytes and bryophytes, even in 
some very simple forms like those of many filamentous algae, though 
in such cases it is less sharply marked and more easily reversed than in 
vascular plants. Organisms without morphological polarity are rare. A 
few amoeboid forms have no axes in the vegetative stage but form 
polarized fruiting bodies. Algae like Pleurococcus are spherical and 
apparentlv apolar but may be induced to produce filaments, an expres- 
sion of axiation. Forms like Spirogyra, desmids, and diatoms have an axis 
of symmetry but its two poles seem to be alike. In most filamentous types, 
however, a rhizoidal pole and a thallus pole can be distinguished. 

Experimentally, polarity can best be demonstrated through its ex- 
pression in regeneration, and it is here that most of our information about 
it has been gathered. Polar regeneration has long been known and 
manipulated in the horticultural practices of vegetative reproduction. 

Vochting ( 1878 ) cut twigs of willow and kept them under moist con- 
ditions. Some he left in their normal, upright orientation and others were 
inverted. Regardless of orientation, however, roots tended to be re- 
generated more vigorously from the morphologically basal end and shoots 
from buds at the original apical end. This is the classical example of 
polarity (Fig. 6-1). If such a shoot were cut into two or more parts 
transversely, each part regenerated roots and shoots in the same polar 
fashion. Even very short pieces of stem showed this polar character. 
Vochting removed a ring of bark in the middle of a shoot and confirmed 
earlier observations that roots were formed above the ring and shoots 
below, just as if the stem had been cut in two. From these and similar 

120 The Phenomena of Morphogenesis 

experiments he concluded that polarity was a fixed and irreversible char- 
acteristic of the plant axis and that probably the individual cells of which 
the axis was formed themselves possessed a polar character. 

Experiments like these have been carried out on many plants. A wide 
variety of results, often conflicting, have been reported and several 

Fig. 6-1. Polarity in willow shoots. Left, portion of a stem suspended in moist air in 
its normal position and producing roots and shoots. Right, a stem similarly grown 
except in an inverted position. ( After Pfeffer. ) 

theories to explain polarity proposed. Klebs (1903), for example, found 
that roots would grow at the apical end of an inverted shoot, that water 
stimulated root formation at any point on the twig, and that removal of 
the bark could reverse polarity. He believed that environmental conditions 
rather than innate polarity determined the place where buds and roots 
develop on a stem. Vochting (1906) replied to these criticisms of the 

Polarity 121 

theory of polarity. The problem, however, is evidently not quite as simple 
as Vochting at first thought. 

Polar regeneration is also evident in the lower groups of the plant 
kingdom. If fern prothallia are sliced transversely, regeneration from 
their cut surfaces is polar (Albaum, 1938fr; Fig. 6-2) and is related to 
physiological gradients, especially of osmotic concentration (Gratzy- 
Wardengg, 1929 ) . In isolated primary leaves of ferns, polarity is evident, 
but both the character and the polar distribution of regenerated structures 
are somewhat diverse (Beyerle, 1932). 

Fig. 6-2. Polar character of regeneration in fern prothallia from which pieces have 
been removed by transverse cuts. An apical portion restores a single heart-shaped 
structure, but from a basal one a group of small prothallia is formed. (From Albaum.) 

In the regeneration of hepatics and mosses, polar behavior varies. The 
gemmae of Marchantia and Lunularia form rhizoids from either surface 
while they are young but only from one when they grow older ( Haber- 
landt, 1914), indicating that embryonic tissue, as it proves to be in many 
other cases, is relatively unpolarized. Polarity here can be reversed by 
gravity, light, and other environmental factors ( Fitting, 1938 ) . Vochting 
found relatively little polarity in the regeneration of the thallus of 
Marchantia. In the mosses, cuttings formed rhizoids at the lower end and 
protonemata at the upper one. This behavior could be reversed by in- 
version of the cuttings ( Westerdijk, 1907), but regenerating structures 
were always more vigorous at the morphologically basal pole. 

Polarity is evident in the sporophores of the higher fungi but here, 
also, it is not firmly fixed, for a segment of the pileus may be successfully 
grafted back to the same pileus in an inverted position ( Lohwag, 1939 ) . 

In most algae there is a sharp distinction between the rhizoidal, or 
hold-fast, pole and the thallus, or shoot, pole. Especially in the simple 
forms and in early stages of the more complex ones, this polarity may be 
reversed by changed relations to gravity, light, or other factors (Wulff, 
1910; Zimmermann, 1923). Studies on the egg of Fucus and on coenocytic 
algae are illuminating here (p. 135). 

Manifestations of polar behavior in higher plants are much more uni- 

122 The Phenomena of Morphogenesis 

form and fixed, presumably because of the higher level of organization 
and differentiation among them. It must have its origin verv early in 
embryonic development. Vochting's conclusion, however, that every 

Fig. 6-3. Transverse polarity in Dioscorea. 
Half slice of a tuber with regenerating 
shoots next the core and roots on the 
periphery. ( From Goebel. ) 

cell is polarized has been challenged by those who point out that many 
cells theoretically may become completely embryonic again and ulti- 
mately produce an entire plant and thus can have no fixed polar charac- 

Polarity 123 

ter. Pfeffer and Klebs have emphasized the probability that the cells of 
the terminal growing points have no original polarity of their own, any 
more than does an egg cell. In older parts a more stable polarization re- 
sults from the influence of conditions in the environment. Vochting's idea 
of the irreversibility of polarity in these higher plants has also been dis- 
puted. Many investigations concerned with these problems cannot be 
judged critically because of insufficient evidence, particularly as to 

anatomical facts. 

Polarity may be manifest in the transverse axis as well as in the 
longitudinal one. This is evident structurally in the transversely polar 
gradient often associated with regeneration. Thus Goebel (1908) found 
that in half slices of the root of Dioscorea sinuata shoots grew out from 
the central part of the axis and roots from the margin directly opposite 
to this (Fig. 6-3). Transverse polarity is also manifest in the flow of 
auxin in various tropisms (p. 384). The subject has been discussed in 
detail by Borgstrom ( 1939 ) . 

Stem Cuttings. In stem cuttings, polar regeneration of shoots and roots 
is clearly obvious in most higher plants, but there are considerable dif- 
ferences between species. Polar behavior may be obscured in various 
ways, as by the tendency of monocotyledons to form roots at nodes and 
by the influence in many cases of the age of the cutting upon the forma- 
tion of root primordia. The specific polar reactivity of tissues from which 
buds and roots originate must be taken into account, as well as the fact 
that a different complex of conditions may control each of the successive 
processes in the development of these structures, such as the formation 
of primordia, their growth, their final differentiation into roots and shoots, 
or the formation of callus which may give rise to either roots or shoots. 

Various modifications of polar behavior in regeneration from stem 
cuttings have been reported. Roots, for example, tend more character- 
istically to be limited to one pole in their growth than do shoots. 
Doposcheg-Uhlar (1911) observed this in Begonia, and Massart (1917) 
studied 30 species of plants, some of which showed strongly polar re- 
generation of both roots and shoots, some weakly polar regeneration, some 
only root polarity, and some only shoot polarity. Root polarity was re- 
lated to the growth habit of the plant, for species with pendant branches 
rooted readily at their apical ends. 

Polar tendency is also expressed in the manner of callus formation 
in cuttings, since in most cases callus tends to develop more vigorously 
at the basal pole than at the apical. From the basal callus, roots are 
usually formed, and shoots from the apical one. Simon (1908) noted 
certain anatomical differences between apical and basal calluses and 
made the observation that calluses from opposite poles may be made to 
fuse but not calluses from the same pole. 

124 The Phenomena of Morphogenesis 

Various investigators (Klebs, 1903; Kiister, 1904; Freund, 1910; and 
Ursprung, 1912) found that local differences in water or oxygen may 
affect root production and thus obscure the inherent polar tendency. 
Only Plett (1921), who studied phenomena of internode polarity in 
410 species, has attempted to explain the variability in distribution of 
roots and shoots on the basis of the anatomy of the plant from which the 
cutting was taken. He found that shoots from axillary buds regenerate in 
a polar fashion, as do adventitious roots that arise endogenously. Ad- 
ventitious buds growing from callus or superficial regions of the cortex, 
however, are generally distributed rather irregularly, a fact which sug- 
gests that the inner layers of the stem have stronger polar tendencies 
than do cortex and callus tissues. 

Root Cuttings. Cuttings of roots behave in polar fashion. Dandelion, 
chicory, and sea kale have been studied most frequently in this regard. 
Shoots are commonly regenerated at the basal or proximal pole (the end 
next the shoot) and roots from the apical (distal) pole. This polarity 
is maintained even when the root cutting is grown in an inverted posi- 
tion (Fig. 6-4). Wiesner (1892c) made the observation, often confirmed 
since, that, in relatively short pieces of root, shoots regenerate at both 
ends. This was also seen by Neilson-Jones (1925) and, in stem cuttings, 
by Fischnich (1939). If the growing roots were continually trimmed off 
from the apical end, shoots finally appeared there. Czaja ( 1935) produced 
roots at both ends by trimming off tissue from the basal end. Centrifuga- 
tion toward the shoot pole results in bud formation at the root pole, as 
does enclosing the base in sealing wax (Goebel, 1908). These results 
are now interpreted as due to the effect of auxin (p. 392), which tends to 
move toward the root apex. A high concentration of it tends to produce 
roots and a low one, shoots. This has been shown clearly by Warmke 
and Warmke (1950). Callus develops more vigorously at the proximal 
pole. As early as 1847 Trecul reported that, in root cuttings of Madura, 
buds and roots showed polar distribution and were formed endogenously 
but that in Ailanthus, where the buds arose in the cortex, polarity was 
much less evident. This agrees with Plett's findings in stem cuttings and 
emphasizes the more intense polar behavior of the inner tissues. 

Leaf Cuttings. Leaves when treated as cuttings behave quite dif- 
ferently from stems and roots and show a somewhat different type of polar 
behavior, evidently related to the fact that they are organs of determinate 
growth. In most cases, regeneration of both roots and shoots occurs at the 
leaf base near the cut end of the petiole. Hagemann ( 1931 ) performed 
inversion experiments on various species. In certain cases he found that 
wound stimulus or water affects regeneration. In Achimenes, shoots were 
thus obtained from the apical cut surface and roots from the base under 
certain conditions, but Hagemann concluded that, in general, polarity as 

Polarity 125 

expressed in the location of regenerating structures in leaves is determined 
by anatomical structure. Behre (1929) reports that regeneration in the 
leaves of Drosera is apolar. 

There has been much discussion as to whether, in the higher plants, 
polarity once established can be reversed. It has been the common ex- 
perience of botanists and horticulturalists that cuttings in which the apical 
end is put into the soil will not do as well as those with normal orienta- 
tion. Some inverted cuttings are found to take root, however, and may 

^ Proximo I 

Ditto I 



Fig. 6-4. Polarity of regeneration in root of Taraxacum. A root segment produces shoots 
at the proximal end ( next the base of the plant ) and roots at the distal end, whether 
the segment is normally oriented, horizontal, or inverted. Compare with Fig. 18-17. 
( From Warmke and Warmke. ) 

live thus for some time. Kny ( 1889 ) successfully grew cuttings of Hedera 
and Ampelopsis inverted for several years, and Graham, Hawkins, and 
Stewart ( 1934 ) did so with willow cuttings, which were still nourishing 
after 11 years. The tips of weeping willows will often root at the apex 
if they are dipping into water (Pont, 1934). Such inverted structures, how- 
ever, often show external malformation and anatomical distortion. Growth 
may become normal again if the cutting is restored to its upright orienta- 
tion and can form roots at the morphologically basal end. Lundegardh 


The Phenomena of Morphogenesis 

Fig. 6-5. "Inversion of polarity." Etiolated pea seed- 
ling with epicotyl decapitated, inverted, and placed in 
water. Roots now grow out from the epicotyl and a 
shoot from a cotyledonary bud. (After Castan.) 

( 1915 ) found that apparent reversal of polar behavior in Coleus was only 

Reversal seems to be easier to accomplish in seedlings than in older 
plants. Castan ( 1940 ) cut off the epicotyl and the primary root from 
etiolated pea seedlings and inverted them. Roots then grew out from the 
originally apical end and shoots from the basal one (Fig. 6-5). Rath- 
felder ( 1955 ) confirmed this observation and believes that it is a real 

Reversal of polarity is much easier to accomplish in the lower plants 
(p. 138). 


Polar phenomena are manifest not only in external form but in in- 
ternal structure. This is evident in many ways. 

Embryonic Development. In vascular plants the first manifestation of 
polar behavior is in the division of the fertilized egg. This in most cases 
seems to be related to the polar character of the gametophyte. Wetter 
( 1952), confirming earlier work of others, finds that in ferns the planes of 
division in the young embryo are related to the axis of the prothallium 
and that the segment that will form the first leaf is always directed 
toward the growing point (notch), a fact also evident later in the orien- 
tation of the young leaf itself. This relationship persists regardless of 
the direction of the incident light. In Isoetes, the first division of the 
fertilized egg is at right angles to the axis of the archegonium, and 
early embryo development is not affected by external factors (La Motte, 
1937; Fig. 6-6). 

Polarity 127 

In seed plants the embryo has a definite orientation in the ovule, the 
tip of the young radicle always being directed toward the micropyle 
and the plumular end toward the chalaza. This has its origin in the polar 
relation between embryo sac and ovule, since the archegonium, or egg 
apparatus, lies at the micropylar end of the sac. Even the group of four 
megaspores is polarized, and it is the one at the micropylar end that 
germinates into the female gametophyte. The planes of division of the 
proembryo are related to the axis of the ovule. In the young embryo as 
it develops at the end of the suspensors, the distinction between root 
and shoot begins very early, with the first transverse divisions. The direc- 
tion of the polar axis is evidently impressed upon the embryo, as upon 
the egg, by the axial organization of the embryo sac and ovule, and 
once established this polar behavior persists and is apparently irreversible. 

Fig. 6-6. Young embryos in the female gametophyte of Isoetes which have developed 
in the positions indicated. Early orientation is with reference to the polar axis of the 
archegonium, but when the leaf begins to push out it becomes negatively geotropic. 
Z, direction of zenith; R, root; F, foot; L, leaf. (From La Motte.) 

Embryonic polarity, however, may arise in other ways than through 
this simple relationship to the ovulary axis. In cleavage polyembryony 
several embryos may arise from a single egg (p. 206), each showing typical 
polar character. Adventitious embryos are sometimes formed by growth 
of nucellar cells and not from fertilized eggs, and these grow into normal 
plants. Structures essentially like embryos sometimes occur elsewhere in 
the plant ("foliar embryos" of Kalanchoe, p. 254) and these show typical 
polar behavior. The first manifestation of differentiation in any embryo, 
whatever its origin, is the appearance of a polar axis. 

A number of cases have been reported (Swamy, 1946) in which the 
polar character of the angiosperm embryo sac is reversed, an egg ap- 
paratus appearing at both ends, or even the antipodal cells at the 
micropylar end and the egg at the opposite one. 

Tissue Reorganization. Various histological changes occur in cuttings 
grown in an inverted position, as described by Vochting ( 1918 ) and 
others. Such plants are evidently abnormal in a number of respects. 
There is often a tendency in them to form swellings and tumors, par- 
ticularly near the insertion of branches, which now tend to grow upward. 

128 The Phenomena of Morphogenesis 

The cause of these swellings may lie in the fact that the original tissues 
cannot function properly under the changed orientation and that con- 
siderable cellular rearrangement must be brought about in the new tissue 
formed after inversion. These tumors resemble anatomically the "whorls" 
commonly found in wound wood and consist of parenchymatous, 
sclerenchymatous, and tracheidal elements. Vochting here has de- 
scribed the structure of such tumors in Salix fragilis and other species 
and believes them to be due to the innate polar tendency of individual 
cells. On such an interpretation, the tissues are thought to twist about 
(Fig. 6-7) until finally those of the root and of the new shoot are con- 
nected by cells of the same polar orientation ( Kiister, 1925 ) . 

There has been considerable controversy as to this hypothesis. Maule 
(1896) uses it to explain the behavior of cambium cells in wound wood. 

Fig. 6-7. Vessel polarity after budding. I, longitudinal anastomoses between vessels 
in normally oriented bud and stock. II, twisting of vessels when bud has been inserted 
upside down. At right, single vessel from the latter. (From Vochting.) 

Neeff (1922) made an extensive series of studies of the changing orienta- 
tion of cambial cells in decapitated stems, finding that these tend to turn 
until they become parallel to the newly regenerating axis instead of to 
the old one (Figs. 6-8, 6-9), and he explains this in terms of the inherent 
polar behavior of the cells, which tends to conform to that of the func- 
tional axis. Both Jaccard (1910) and Kiister, on the other hand, disagree 
with Vochting's explanation and attribute the changing orientation of 
the cells mainly to mechanical factors. Twisting whorls may also appear 
in normal callus where mechanical factors can hardly be operative. More 
intensive studies are needed of the conditions that cause change in direc- 
tion of cell growth. Altered direction of sap flow, for example, might 
affect the direction of cambial cell growth in Neeff's experiments. Similar 
changes in cellular orientation have been reported by MacDaniels and 
Curtis ( 1930 ) in spiral ringing wounds in apple, by Janse ( 1914 ) in 
bark strips left across a ringing wound in Acalypha, and by Tupper- 

Polarity 129 

Carey (1930) in tissue bridges in Acer and Laburnum. Pressure, nutrient 
movements, and basipetal cambial activity have been suggested as causes. 
The results of Went (p. 384) with inverted cuttings of Tagetes indicate 
that the direction of auxin flow in them is ultimately reversed. It is clear 
that in some way histological changes are related to the new conditions 
under which an inverted cutting has to grow. 

Fig. 6-8. Left, diagram showing direction of cambium cells (and their derivatives) 
in a normal shoot of Tilia, with a lateral root and lateral shoot growing from it. The 
cells are parallel to the particular axis of which they form a part. Right, change of 
direction of these cells when the main axis has been decapitated at both ends and the 
lateral axes are becoming the main ones. The direction of the cells in the original 
main axis has now turned to become parallel with the new ones. ( From Neeff. ) 

The results of grafting provide a direct way of testing polar differences 
in tissues. Vochting ( 1918 ) used the swollen stem of kohlrabi for a series 
of such experiments. If the top of a stem is sliced off transversely and a 
V-shaped cut made in its upper surface and if the lower portion of 
another stem is sharpened to fit this cut and inserted firmly into it, the 
tissues of the two stems will knit together. If, however, a piece is sliced 
off from the lower part of a kohlrabi stem and it is then inverted, and the 
surface now uppermost cut as before, and if a sharpened upper piece 
is inserted into this cut, the tissues will not knit. Furthermore, rootlets 
will begin to grow out from the upper piece into the lower one, as if 

130 The Phenomena of Morphogenesis 

growing in a foreign substratum. Thus a root pole will fuse with a shoot 
pole but two similar poles, when brought together, will not fuse. Vochting 
also found that a square bit of tissue cut out from a beet root and put 
directly back will knit in its former place but will not do so if turned 
through 180° before being replaced there. These facts can be explained 
by assuming that the tissues of the plant, even such relatively undif- 
ferentiated parenchymatous ones, have definitely polar behavior. Bloch 
(1952), however, observed that tissues of the fruits of Lagenaria do not 
behave in this way but that plugs, cut out and replaced, will knit in any 

Fig. 6-9. Tangential sections through tissue of an axis like those in Fig. 6-8. Left, nor- 
mal wood. Right, after decapitation, direction of cells changing to conform to the axis 
of the lateral root, now the main one. ( From NeejJ. ) 

orientation. Microscopic examination after a few days showed normal 
cellular fusion. 

In horticultural practice it has long been recognized that buds must 
be placed in normal orientation on the stock if they are to knit well. 
Colquhoun ( 1929 ) removed buds and pieces of bark in Casuarina and re- 
applied them in an inverted position. Observation of the anatomical struc- 
ture showed that the cells of the cambium joined freely and continued to 
grow regardless of orientation. Wood fibers and vessels, however, show 
the characteristic turns and twists reported by Vochting. This suggests 
that the cambial cells are unpolarized or in a condition of unstable polar- 

Polarity 131 

ity and that, as wood elements differentiate, polarity is gradually im- 
pressed upon them. Cells inversely oriented are now unable to unite, 
and the translocation of materials in them, tending in each to follow the 
original direction of flow, is seriously disturbed. This gradual assumption 
of polarity is perhaps related to changes at the cell surface as the wall is 
formed or in the structure of the wall itself. 

Another manifestation of polar activity in histological characters, per- 
haps related to the basipetal tendency in the renewal of cambial activity 
or to the polar flow of auxin (p. 384), may be observed in the reconstitu- 
tion of severed vascular strands across the ground parenchyma of pith 
or cortex in herbaceous dicotyledons (Simon, 1908; Sinnott and Bloch, 
1945; Jacobs, 1954). This always begins at the basal end of a severed 
strand and proceeds downward toward the apical end of the cut bundle 
or to uninjured ones. 

Cell Polarity. To test Vochting's contention that polar behavior of a 
tissue is the result of the polarity of its individual cells is not easy. The 
fact that very small tissue pieces retain their original polarity and that 
inversely grafted tissues do. not fuse supports Vochting. Many other facts 
can also be cited. The two daughter cells following a division are often 
unlike (p. 133). In these cases, each of the two types is found invariably 
on the same side, toward or away from the tip of the axis. Thus in many 
young roots the last division of the surface cells is unequal, the smaller 
daughter cell becoming a trichoblast and producing a root hair (p. 190). 
This cell is always on the side toward the tip of the root. Before division, 
the apical end of the mother cell is also more densely protoplasmic. In 
some cases (Phleum) the division is markedly unequal and polar. In 
others (Sporobolus) the two cells are more nearly equal and a root hair 
is not always formed (Fig. 6-10). Here the polar behavior is much less 
marked. In the leaf epidermis of monocotyledons some cells divide 
unequally, and the one toward the leaf tip becomes a stomatal mother 
cell. These facts suggest that the cells themselves have a polar orienta- 

The tendency of cells to divide in specific directions is at the bottom of 
all form determination, since it is concerned with the plane of division 
and thus the direction of growth. In the growth of elongate gourd fruits, 
for example, divisions are predominantly at right angles to the axis of 
the fruit, but in isodiametric ones they are at all angles (p. 51). Whether 
polarity is a quality of the whole developing organ or simply of its 
component cells is still uncertain and is a problem involving the deeper 
one of the relation between cell and organism. Various examples of 
polarity in unequal cell divisions have been discussed and figured by 
Bunning (1957; Fig. 6-11). 

Even when the cell does not divide, the difference between its two 

132 The Phenomena of Morphogenesis 

ends is often evident. That the cytoplasm is the seat of this polar differ- 
ence is shown by the fact that, when vacuolate cells divide, the first 
indication of the plane of division, and thus of the polar axis, is the ap- 
pearance of a cytoplasmic diaphragm in the position where the future 
partition wall will be formed (p. 25). In such cells the direction of the 
axis may be related to gradients in hormone concentration, oxygen, or 
other factors. This polar difference may be visible in the contents of the 
cell, for in Enteromorpha ( Muller-Stoll, 1952), in Isoetes (Stewart, 1948), 
and other plants the chromatophore is almost always on the side of the 
cell away from the base of the thallus, or plant body. The distribution of 
chloroplasts in higher plants is also sometimes polar. 









Fig. 6-10. Polarity in root-hair development. A, B, and C, successive stages, with root 
apex toward left. In Phleum, the last division is unequal, and the cell toward the apex 
forms a root hair. In Sporobolus, the division is essentially equal, and the cell toward 
the apex does not always form a root hair. ( From Sinnott and Bloch. ) 

The wall itself may show polar behavior, a fact which is of particular 
importance in producing differences in cell shape. Most cells are nearly 
isodiametric at the beginning, and if one at maturity is much longer than 
wide, this is the result of more rapid growth in length. Such differential 
growth, in turn, presumably comes from differences in the fine struc- 
ture of the wall, which itself is ultimately dependent on factors in the 
cytoplasm. Wilson (1955) has shown that in the wall of the large cells 
of the alga Valonia there are two systems of orientation of cellulose 
fibrils which converge to two poles at the ends of the cell. The complex 
and remarkable shapes of many cells, both in simple organisms and 
within the tissues of larger ones, are probably due to a complex pattern 

Polarity 133 

of wall polarities that determine growth in a number of directions. How 
this is brought about is a morphogenetic problem at a different level 
from most of those here discussed, and its solution may provide sugges- 
tions for an approach to other problems of form. 

It is sometimes possible to demonstrate the polarity of single cells ex- 
perimentally even though their contents are homogeneous and both ends 

Fig. 6-11. Various types of unequal and 
polar cell divisions: I, in pollen grain; 
II, in differentiation of root hairs in cer- 
tain monocotyledons; III, in differentia- 
tion of stomata in monocotyledons; 
TV, in leaf cells of SpJiagnum; V, in for- 
mation of sclereids in Monstera. (From 
Biinning. ) 












1 ® 



1 & 


\ B 



appear to be alike. This can be done by isolating cells and observing the 
structures that regenerate from them. Miehe (1905) accomplished this 
in the filamentous alga Cladophora. Here polar organization is present 
but not conspicuous. At the basal end is a rhizoid which attaches to the 
substratum, and the rest of the filament or the thallus, a single row of 
cells, is undifferentiated. Miehe plasmolyzed the cells of a filament just 

134 The Phenomena of Morphogenesis 

enough to pull them away from the walls and break whatever connec- 
tions there may have been with other cells, but without killing them. 
The plant was then deplasmolyzed. Each cell, now as effectively isolated 
as though it had actually been removed, began to enlarge, broke out of 
its wall, and proceeded to regenerate a new filament. The significant 
fact is that from the basal end of each cell a new rhizoid was formed 
and from the apical end, a new thallus. The polar character of the cells, 
otherwise impossible to demonstrate, could thus be established. These 
experiments were repeated and extended by Czaja (1930; Fig. 6-12). 
Borowikow (1914) succeeded in reversing the polarity of Cladophora 

Fig. 6-12. Polarity in a single cell. 
A cell isolated from a filament of 
Cladophora, regenerating a thallus 
from its apical end and a, rhizoid 
from its basal one. (From Czaja.) 

cells by centrifugation, showing again the close relation between the 
distribution of material in the cytoplasm and the polarity of the cell. 

In some filamentous algae, the plant's organization may disintegrate 
under certain circumstances and the individual cells thus become freed 
from their correlative inhibition. In Griffithsia, for example, Tobler 
( 1904 ) observed such cells in culture and found that, when they began 
to regenerate, rhizoids grew from the basal end (distinguishable by its 
shape) and shoots from the apical one. Schechter (1935) centrifuged 
similar ones and found that their polaritv could be altered by this means 
and that shoots always appeared at the centrifugal pole. 

The rather loosely organized tissues of these simple algae provide ex- 
cellent material for studies in cellular polarity even though their cells 

Polarity 135 

are not isolated. In bits of tissue cut from Enteromorpha, for example, 
Miiller-Stoll (1952) found that the cells near the apical portion of the 
piece regenerate papilla-like structures but that the cells at the base 
form rhizoids. 

To prove the existence of polar behavior in the cells of one of the 
higher plants is more difficult. Here it is sometimes possible to re- 
generate a new plant from a single cell or small group of cells, especially 
in the epidermis, but it is not easy to relate the polarity of the newly 
produced structure to that of the cell from which it grows. 


In many cases, polarity may be studied in cells that are isolated in 
nature and not through experiment. The most notable example of this 
is the egg of the rockweed, Fucus, which is discharged into the water and 
there is fertilized and grows into a new plant. The Fucus egg is com- 
parable to the eggs of certain animals that develop in water and that 
have proved such a rich source of knowledge of early embryology. More 
work has been done on this egg than on any other naturally isolated plant 

The unfertilized egg is naked, and its nucleus is at the center of the 
cell. It shows at the beginning no polarity whatever nor is there any 
visible differentiation in its cytoplasm. After fertilization, the egg falls 
to the bottom and in about 12 to 24 hours, under normal conditions, a 
protuberance appears on its lower surface. This develops into a rhizoid by 
which the young plant becomes anchored to the bottom. Soon the egg 
divides in a plane at right angles to the axis of the protuberance. The 
two cells that result are very different in shape and in their future de- 
velopment. The upper, rounded cell gives rise to the main portion of the 
thallus. The lower one forms little besides the rhizoid. The growth of 
the rhizoid and the first division of the egg establish a permanent polar 
axis in a system which at first is quite without one. Here is evidently one 
of the simplest expressions of polarity among plants. 

Among earlier investigators of the Fucus egg were Kniep ( 1907 ) , 
Nienburg (1922a and b), Lund (1923), and others. The more recent 
experiments of Whitaker and his colleagues have provided a large body 
of detailed information. Only the main facts will be presented here. This 
work has been reviewed by Whitaker ( 1940 ) and Bloch ( 1943 ) . 

Gravity seems not to be an important factor in the induction of the 
polar axis for, if the eggs are kept in the dark, the rhizoid develops in 
any direction. There is evidence that in eggs of the related Cystosira 
barbato, if reared in darkness, the rhizoid is formed at the point of en- 
trance of the sperm ( Knapp, 1931 ) . Light is clearly a very important 

136 The Phenomena of Morphogenesis 

factor in Fucus (Hurd, 1920). In eggs lighted from one side by white 
light of a certain intensity (or light of particular wave lengths), the 
rhizoid always forms on the side opposite the source of light, and the 
first wall is laid down at right angles to this direction. Nienburg ( 1922a, 
b) showed more specifically that it is not the direction of light but the 
intensity gradient that is the determining factor. 

Lund ( 1923 ) was able to prove that the first division wall in the Fucus 
egg was at right angles to the flow of an electric current and that the 
rhizoid grew toward the positive pole. Here the polar axis can evidently 
be determined electrically. 

A peculiar phenomenon first noted by Rosenwinge ( 1889 ) but studied 
particularly by Whitaker is the so-called "group effect." If an egg of 
Fucus lies near a group of other eggs, its rhizoid will develop toward this 
group ( Fig. 6-13 ) . Whitaker observed that this occurs even when the eggs 
belong to different species. These results have been attributed to the 

&-^-:>. : .*':''i7 Fig. 6-13. The "group effect" in Fucus eggs. Where 
there is a cluster of these, the rhizoidal pole is typically 
on the side toward the other eggs. ( After Whitaker. ) 

establishment, in the medium near the egg, of a concentration gradient 
of metabolic products from the other eggs, but Jaffe (1955) finds evi- 
dence that such a gradient, if it exists, does not involve H ion, COo, or 
2 . Jaffe also found (1956) by exposing eggs to polarized light that the 
rhizoids tended to develop in the plane of polarization. 

Whitaker ( 1937 ) subjected Fucus eggs to centrifugal force and showed 
that in such cases the rhizoid grows from the centrifugal pole. Polarity 
here seems to be dependent on the rearrangement of materials in the egg. 
Other factors, such as pH, temperature, auxin, and even the shape of the 
egg have also been shown by Whitaker to modify egg polarity. 

Not many other cases of the induction of a polar axis in isolated algal 
cells are known, although in certain green algae the polarity of the young 
plants developing from swarm spores is determined by the way in which 
these spores become attached to the substratum ( Kostrum, 1944 ) . 

The remarkable umbrella-shaped alga Acetabularia is really a single 
cell though it may be several centimeters tall. During its development it 

Polarity 137 

has only one large nucleus, situated usually at the base of the stalk. 
Here the rhizoids develop. At the summit of the stalk is the disk. If a 
nucleus is introduced into a plant or plant segment which lacks one, a 
new rhizoid system will arise wherever the new nucleus is placed and 
polar behavior of the plant may thus be modified or reversed ( Hammer- 
ling, 1955). 

In the germination of a moss spore, the young protonema pushes out on 
the side of the spore toward the light, and the rhizoid forms at the op- 
posite end, indicating that here, as in the Fucus egg, its polarity is de- 
termined by light. In several moss species, Fitting (1949) was able to re- 
verse this polarity by reversing the direction of the light. In this way the 
young protonema becomes converted into a rhizoid. 

Heitz ( 1940 ) prevented polar germination of Funaria spores by appli- 
cation of auxin. Cell division was also inhibited by this means and 
"giant" cells thus produced. D. von Wettstein (1953) confirmed this and 
found that vitamin Bi and chloral hydrate destroyed polarity without 
preventing cell division. Such apolar growth continued for 50 cell genera- 
tions, producing an undifferentiated, tumor-like body. Such a result 
emphasizes the importance of polar behavior for orderly development 
and the production of form. 

How the polar axis is determined in the spores of vascular plants has 
been demonstrated in a few cases. In Equisetwn, the spore of which 
shows no external or internal polarity, germination is followed by di- 
vision into two cells. The division wall, as in Fucus, is laid down at right 
angles to the gradient of light absorption (Stahl, 1885). The more strongly 
illuminated daughter cell becomes the primary prothallial cell and the 
one on the darker side, the rhizoidal cell. Nienburg (1924) showed that 
this alignment of the mitotic figure parallel to the direction of the inci- 
dence of light does not occur until a redistribution of cytoplasmic ma- 
terial has taken place, especially an aggregation of chloroplasts on the 
illuminated side (Fig. 6-14). 

In germinating fern spores, light modifies the polar behavior but this 
effect is different in different wave lengths (Mohr, 1956). Naf (1953), 
also working with ferns, found evidence that the spore of Onoclea has an 
inherent polarity but that this can be modified by light. He carried the 
study of polarity reversal much further by growing the young prothallia 
in a liquid culture which was constantly shaken. The prothallia thus 
developed in an environment where there were no environmental 
gradients and where the plant was exposed on all sides to equal stimu- 
lation by gravity, light, and other factors. The result was a spherical, 
tumor-like mass of tissue. Grown on agar and without movement, this 
tissue again formed structures much like the normal prothallia. The 
genetic basis for a normal prothallium, specific in character, is in the 

138 The Phenomena of Morphogenesis 

spore but such a prothallium will develop only where there is an en- 
vironmental gradient by which its polar axis is established. Such results 
emphasize the fact that neither genetic constitution nor environment 
alone controls the development of organic form, but an interaction be- 
tween them. 

Polarity is also to be found in the microspores and pollen grains of 
higher plants, though here it is not easily open to experimental analysis. 
The spore axis, as indicated by the orientation of the division of the 
spore nucleus, has a constant relation to the planes of division of the 
pollen mother cell. 

Fig. 6-14. Origin of polarity in a single cell. I, unpolarized spore of Equisetum. II, 
beginning of polarization as shown by changed positions of plastids (chl) and nucleus 
(k). Ill, first nuclear division. (After Nienburg.) 


In larger and multinucleate protoplasmic units, notably coenocytic 
forms, polarity finds a somewhat different expression than in uninucleate 
protoplasts. Thus in the plasmodium of Plasmodiophora brassicae, as re- 
ported by Terby (1933), the axes of the many nuclear division figures lie 
parallel to one another, indicating that the whole mass of protoplasm has 
a uniform anisotropic orientation, though here without a polar axis. This 
parallelity is also found in the first sporogenic division but disappears at 
the second one. In most plasmodia, however, this simple sort of polar 
behavior seems not to be present. 

True coenocytes show some remarkable examples of organized systems 
where there is pronounced differentiation of parts but no cellular parti- 
tions in the cytoplasmic body. Conspicuous among these are the algae 
Bryopsis and Caulerpo. 

Bryopsis has an axis from which "leaves" come off in a pinnate arrange- 
ment above and rhizoids below. This polar organization can be com- 
pletely reversed if the plant is held in an inverted position. The leaves 
then produce rhizoids and the rhizoids, leaves ( Noll, 1888; Winkler, 1900; 
and others ) . It now seems probable that a different relation to light rather 

Polarity 139 

than to gravity is responsible for this reversal. Steinecke (1925) has 
shown that in these inverted plants there is a movement of the cyto- 
plasm that was originally in the upper portion into the base, and vice 
versa. The easy reversal in such plants seems to be related to the fact 
that the cytoplasm can move readily throughout the whole body. The 
cellular organization of higher plants may contribute to the more fixed 
polarity that they display as well as to their higher degree of differen- 

Caulerpa has a more complex structure than Bryopsis, for it possesses 
a horizontal "rhizome" from which "leaves" grow out above and rhizoids 
below. The leaves are negatively geotropic, the rhizoids positively so, 
and the rhizome is diageotropic. Regeneration in this plant has been 
studied by many workers (Wakker, 1886; Janse, 1906 and 1910; Dostal, 
1926 and 1929; Zimmermann, 1929; and others ) . Its polar phenomena are 
rather complex. Zimmermann found that gravity determines the dorsi- 
ventrality of the rhizome and that this can be reversed. He also observed 
that in each portion of a cut leaf new rhizoids are formed below new 
leaves. An inverted leaf with its tip buried will produce leaves at its 
original base and rhizoids at its original apex, as in Bryopsis. Janse has 
shown, however, that rhizoids normally appear chiefly at the apical por- 
tions of cut leaves, and Dostal finds that, although regeneration is polar 
in young leaves, the new organs may be distributed over the entire sur- 
face of older ones. Polarity in Caulerpa is less stable than in higher plants 
and this, again, is probably because of the ease with which cytoplasmic 
movement may take place. 


Differences in the external or internal structure of the plant body are 
almost invariably accompanied by physiological differences, though the 
latter are usually more difficult to demonstrate. Among these are the uni- 
directional flow commonly shown by auxin and often by other substances; 
the differences in bioelectric potential which can be demonstrated be- 
tween different parts of the plant; and the many examples of physiological 
gradients in the plant body— in pH, rate of respiration, osmotic concentra- 
tion, auxin concentration, and others. These are doubtless related to 
visible morphological polarities but the character of the relationship is 
obscure. Whether such physiological polarities control the morphological 
ones or whether both are determined by more deeply seated morpho- 
genetic factors in the living material, which are physiological only in the 
broadest sense, is not known. 

Electrical polarities are found in many organs and in the plant body as 
a whole. Unfortunately, any discussion of the significance of electrical 

140 The Phenomena of Morphogenesis 

potentials in relation to morphological polarity and polar regeneration 
must remain for the present rather hypothetical because of the uncertain- 
ties which still exist as to the nature and origin of the potential differences 
themselves ( p. 361 ) . The phenomena of organic polarities have so many 
similarities to electrical ones that it is tempting to explain the former 
entirely in terms of the latter, but there is insufficient evidence as yet for 
such a simple solution of the problem. 

Physiological gradients of various kinds, particularly metabolic ones, 
and their significance have been extensively discussed by Child ( 1941 ) . 
Such gradients are along the major axes of the organism, and indeed 
their existence is thought by some to establish these axes and to be a 
major factor in the origin of polarity. Child believed that they arise early 
in development as the result of some unilateral difference in the environ- 
ment and that, once established, they persist. He points out that they 
often can be obliterated or redirected by external differentials and infers 
that they are of great importance in determining patterns of development. 
The inherent properties of protoplasm, unable alone to control develop- 
ment, produce their morphogenetic effects through specific reactions to 
such axial gradients. Prat ( 1948, 1951 ) has reviewed the relations be- 
tween physiological and histological gradients. 

Gradients in respiratory activity such as have so often been described 
in animal axes are found in plants ( Wanner, 1944 ) . Ball and Boell ( 1944 ) , 
however, have shown that in some plants the rate of respiration at the 
meristematic tip is less rapid than in the zone immediatelv behind this 
(p. 73). Hurd-Karrer (1926) found that in corn stalks the minimal con- 
centration of solutes is in the basal internodes and increases upward, a 
gradient reported by others for leaves at different levels in a tree. In plant 
exudates there is a concentration gradient with the highest values near 
the apex (Tingley, 1944). The proportion of ash to dry weight in herba- 
ceous plants was shown by Edgecombe ( 1939 ) to increase toward the tip 
of the plant. Many other examples might be cited. 

These gradients are often related to translocation of solutes and food 
and thus to localized and differential growth. Hicks (1928a, b) found 
that nitrogen tends to move toward the morphological tip of a stem and 
carbohydrates toward the base, even in inverted shoots, so that a gradient 
in C/N ratio results in the stem. She believes that this may be respon- 
sible for the phenomena of polarity, but this may be a parallelism rather 
than a causal relation. 

The unidirectional flow of nutrients to particular "centers of attraction" 
in shoots, roots, leaves, and other structures has been emphasized by 
Goebel as of particular importance in regeneration and other phenomena 
of development. What causes the establishment of such centers and thus 
directs the location of growth is a question closely related to that of 

Polarity 141 

polarity. Simon (1920) has suggested that the polar character of regen- 
eration in leaves is related to the basipetal movement of carbohydrates. 

How much the direction of flow in the phloem is due to polar behavior 
in the strict sense and how much to other factors is not clear, but Schu- 
macher (1933, 1936) has shown a polar flow of fluorescein there, basipetal 
in the petiole and in various directions in the stem. It seems clear that in 
most vegetative stems the flow of nutrients in the phloem is predomi- 
nantly basal. Sax (1956 and earlier papers) removed a ring of bark in a 
young tree and then replaced it in an inverted position. Under these con- 
ditions phloem transport is markedly checked and the tree is much re- 
duced in growth. This effect is not permanent, however, because the new 
bark regenerated at the seam permits phloem transport upward. 

The clearest case of physiological polarity and the one most thoroughly 
studied is that of the flow of auxin (p. 384). In the Avena coleoptile it has 
been shown that auxin normally is produced at the tip and moves toward 
its base. If the coleoptile is cut off, decapitated, and auxin applied at 
the morphological apex it will move toward the other end, whether the 
coleoptile is normally oriented or inverted. If auxin is applied to the 
morphologically basal end, however, it will not move toward the tip even 
if the coleoptile is inverted and the auxin is placed at the end now upper- 
most. Auxin flow here is therefore strictly polar. The cause of this 
polarity is not clear, for there is no histological difference with which 
it is correlated. It seems to be characteristic of auxin transport gen- 
erally, for this substance, commonly produced in buds, moves down- 
ward from them but not upward. Jacobs and others (p. 384), however, 
report that auxin may sometimes move acropetally, especially in weak 

Went ( 1941 ) has shown that the auxin flow continues to be morpho- 
logically basipetal in inverted cuttings of marigold but that after a time, 
presumably following the production of new and reoriented vascular 
bundles, the flow is reversed and auxin now moves downward toward the 
new root system. He suggested that auxin polarity is electrical in char- 
acter, but this idea has encountered some difficulties ( p. 360 ) . 

The significance of auxin polarity for many problems of plant develop- 
ment is great since this substance is so intimately related to both stimula- 
tion and inhibition of growth and to so many specific growth reactions, 
such as the initiation of shoots and roots (p. 390). 

It is tempting to explain all structural polarity in the plant as due to this 
polar flow of auxin, but here again it may be that both are the result of 
some more deeply seated factor. No satisfactory solution of the problem 
has yet been found. It is surely a remarkable fact that a simple, relatively 
undifferentiated parenchyma cell of the oat coleoptile will allow auxin to 
pass through it in only one direction. An understanding of the mechanisms 

142 The Phenomena of Morphogenesis 

involved in this polar flow would doubtless contribute to the solution of 
many problems in physiology and development. 


The chief significance of polarity for students of morphogenesis lies in 
the fact that it is the simplest expression of the general phenomenon of 
organic pattern. These patterns, which are exhibited in such profusion in 
the bodies of animals and plants, are each built around a polar axis 
which provides, so to speak, the theme or foundation upon which the 
whole develops. This, of course, is by no means the only expression of 
polarity in the plant. In a tree, for example, not only the main trunk but 
the many branches growing out from it may each have a polar axis of its 
own. The frequency and size of these branches and the angles which 
they make with the trunk produce the characteristic pattern of the tree's 
crown, one which is almost as specific as the pattern of its leaf. The char- 
acter of this crown is due to the dominance of certain buds or branches 
over others and thus to a controlled localization of growth and a balance 
between the various axes of the tree. This, in turn, seems to be governed 
by a specific polar pattern of auxin distribution. It seems probable that 
the form and development of a leaf, which involve a pattern of major and 
minor vein polarities, have a similar basis. In such cases as these, organic 
form appears to be the expression of a series of interrelated polar axes. 

Such a condition probably occurs also in forms in which the organic 
pattern is related to the polarity of individual cells. This is well shown 
in the development of shoots that grow by a large apical cell. Here the 
growth and differentiation of the axis are clearly associated with a precise 
series of divisions in various planes, both of the apical cell and of those 
cut off from it (p. 58). This pattern of diverse cell polarities is less easy to 
trace in other meristematic regions but is evidently operating there as 
well. In such a structure as the growing primordium of a young ovary, 
cell divisions are very abundant but occur in every direction, as though 
the planes of division were at random. That such divisions are all part of 
a definite organic pattern, however, is shown by the fact that the structure 
in which they occur shows a regular and progressive development toward 
its specific form. Each plane of division, presumably determined by the 
orientation of the cytoplasm, is related to the complex pattern of diverse 
polarities of which it forms a part (Sinnott, 1944). 

The polar phenomena of coenocytes and other evidence support the 
contention that the basic fact in polarity is the orientation and polar be- 
havior of the cytoplasm. Where this is confined within cell walls, more 
complex and stable patterns may be produced, but the fundamental prob- 
lem everywhere seems to be the development of polar patterns in the 

Polarity 143 

whole cytoplasmic body of the organism, whether this is cut up into cells 
or not. 

Polar patterns are not confined to organisms that have developed in the 
ordinary way by growth from a reproductive unit such as a spore or egg 
but are found in what are essentially organic communities. Thus in cer- 
tain slime molds such as Dictyostelium (p. 223) the vegetative individual 
is a tiny myxamoeba. At the end of vegetative growth some thousands of 

Fig. 6-15. Polarity in Dictyostelium. If an apical piece of a pseudoplasmodium is placed 
in close contact with the apical end of another, coalescence takes place. If it is placed 
next the basal end, either of the same plasmodium or of another, there is no coa- 
lescence, and the terminal piece moves off independently. (From K. B. Raper.) 

these become aggregated into a pseudoplasmodium where each retains its 
individuality. This colonial structure shows a polar organization, for the 
terminal portion of it can be grafted to the decapitated apex of another 
pseudoplasmodium, though not to the base (Fig. 6-15). The tip is evi- 
dently the dominant region, for if grafted to the side of a pseudoplas- 
modium it will withdraw from it a group of individuals and start out as a 
new unit. The sorocarp that ultimately develops has a vertical polar axis 
and, in some species, lateral axes as well. Polarity in organisms like these 

144 The Phenomena of Morphogenesis 

appears to be a property not of the individual cells but of the aggregate 
that they form. 

Three Aspects of Polarity. Polarity is evidently a complex phenomenon 
which is intimately related to the whole process of development. It may 
be broken down, for purposes of more detailed examination, into several 
different aspects or elements which may possibly involve different physio- 
logical or developmental processes. 

First, one may recognize the oriented behavior of living substance, as 
distinct from axiation or bipolarity. This is evident in the differential 
growth of cells and tissues, where one dimension increases more rapidly 
than the others; in the controlled plane of cell division, in which the 
cytoplasm, as evident especially in vacuolate cells, sets up a pattern 
oriented in a definite direction; and in coenocytes and plasmodia where 
growth, movement, or direction of nuclear spindles is similarly oriented. 
The fundamental basis of this behavior is not known. There may be in- 
volved the orientation of micelles or other submicroscopic units, the 
paracrystalline properties of cytoplasm, the orientation of molecules at 
cell surfaces or interfaces, or the nature of the fine structure of the cell 
wall. It is reasonable to suggest that some sort of cytoplasmic anisotropy 
is concerned in this oriented behavior. Here is evidently a major problem 
for the student of the ultimate structure of protoplasm, a problem inti- 
mately related to the whole question of directed growth and thus of or- 
ganic form. Whatever the basis of oriented behavior may be, in some 
cases it can evidently be changed in direction readily by environmental 
factors, but in others, when once established, it becomes firmly fixed. 

A second aspect or element of polarity is axiation. The oriented be- 
havior of living material most commonly, though not invariably, is ex- 
pressed in cellular systems which develop symmetrically in relation to an 
axis or plane of symmetry parallel to the direction of orientation. Most 
cells and most multicellular structures possess an axis. Such structures as 
the cells and filaments of unattached filamentous algae may show no evi- 
dent difference between the two ends of the axis, either in cell or filament, 
but they are clearly axiate. The problem of the symmetrical growth of a 
living system about this axis, so characteristic of almost all organic de- 
velopment, is an essential part of the general problem of pattern. Experi- 
mental attack upon this phenomenon of symmetry is promising, for its 
character can often be changed by modifying the environment. 

The third aspect of polarity is polar difference, the appearance of dis- 
similarity between the two ends of the axis. This is regarded by many as 
the essential characteristic of all polarity and is present in the great ma- 
jority of organic axes. In not a few cases, as we have seen, cytoplasm may 
show oriented behavior, or an axis of symmetry may develop, without any 
demonstrable evidence of difference between the two ends of the system. 

Polarity 145 

Only where polar differences occur, however, with the resulting morpho- 
logical and physiological gradients from one end of the axis to the other, 
can there develop the complex patterns characteristic of most living 

The relation between these three aspects of polarity involves the prob- 
lem of the origin of polarity itself. If they can be shown to form a pro- 
gressive series, in phylogeny or ontogeny, this would indicate that polarity 
may increase in complexity. In free-floating algal filaments (as in Spiro- 
gyra) there is no evidence that the two poles are unlike. In filamentous 
forms like Cladophora, however, where one end is attached to the sub- 
stratum, each individual cell displays a polar character in its regeneration. 
Here, and in many other cases where environmental factors are different 
at the two ends of the axis (as in the Fucus egg), it appears that this 
difference sets up an axial gradient in a system originally unpolarized, 
which results in the polar difference. Child and his school regard all 
polarity as having its origin in such environmentally induced gradients, 
which determine both the direction of the axis and the difference be- 
tween its poles. On the other hand, since instances of similar poles are 
rare, it may be held that the two ends of every axis are fundamentally 
unlike and that in cases where they seem alike the difference is merely 
masked and difficult to demonstrate. If this view is correct, polarity may 
be due to something quite different from a gradient and may be compa- 
rable to, and perhaps result from, an inherent polar tendency, presumably 
electrical in character. 

Whatever its origin, the direction of this bipolar axis is often continu- 
ally changing but under definite control, and upon this fact depends the 
orderly development of organic patterns. Thus in a three-faced apical cell 
the polar axis must shift 120° between successive divisions. In more com- 
plex meristems the planes of cell division are equally orderly, though less 
evidently so. How such a system of changing polarities is controlled so 
that growth in one direction is precisely related to that in another is a 
part of the same problem of orderly development which the student of 
morphogenesis so often meets. 

The ease with which polarity may be reversed in the simplest plants 
suggests that even in more complex ones it is not irrevocably fixed by 
genetic factors. Like any trait with a genetic basis, polarity is not a specific 
characteristic but a specific reaction to a specific environment. The en- 
vironmental factor may be external, such as the direction of light, or 
internal, like the correlation between the axis of the young embryo and 
that of the archegonium, but unless there is an environment to which the 
organism can orient itself, the phenomena of polarity will rarely appear. 
Sometimes this environmental reaction is determined early and is later 
irreversible, as in cases where polarity becomes firmly fixed in the ferti- 

146 The Phenomena of Morphogenesis 

lized egg. In other instances polar behavior is subject to induction through 
environmental factors at all stages. Many cases of polarity are like this. 
Sometimes this polar plasticity persists indefinitely, as in Caulerpa. In 
other cases, like the egg of Fucus, it lasts only till the polar axis is deter- 
mined and then remains unchanged, regardless of the environment. What 
is present in all living stuff seems to be a persistent tendency for the 
establishment of polar behavior. This, indeed, is an essential preliminary 
for the development of a formed and organized system. 

The study of polarity has thus far raised more problems than it has 
solved. Most of these, however, are of a sort more amenable to analysis 
and experimental attack than are many - others in organic development. 
Especially through subjecting each of the various aspects of polarity 
separately to experiment, aspects which may perhaps involve distinct 
developmental processes, is there hope of progress. 



The presence of an axis, so generally characteristic of the form of body or 
organ in animals and plants, is manifest not so much as an actual material 
structure but as an axis of symmetry, a geometrical core or plane around 
which or on the two sides of which the structures are symmetrically dis- 
posed. One of the most obvious manifestations of organic pattern in 
living things is this symmetrical arrangement of their parts. 

Symmetry is evident in both external form and internal structure. Lat- 
eral roots arise from a primary root in two, three, four, or more equally 
spaced rows. Leaves are symmetrically disposed around the stem in a 
phyllotactic spiral. Floral diagrams, both transverse and longitudinal, also 
provide good examples of axial symmetry, though here the axis is usually 
much shortened. 

Symmetry is equally conspicuous in internal structure. The cross sec- 
tion of almost any vertical plant axis shows symmetrical arrangement of 
its tissues, both primary and secondary. Even single cells, especially when 
they possess a considerable internal diversity like those of Spirogyra, are 

In horizontal organs the simple radial type of symmetry characteristic 
of vertical axes is replaced by a dorsiventral one where the two halves on 
either side of a vertical plane of symmetry are alike. Many prostrate stems 
and most leaves are examples of such dorsiventrality. Sometimes one type 
of symmetry may be changed to the other by modifying the orientation 
of the structure to light or gravity. In other cases the pattern of symmetry 
is inherited and cannot be influenced by environmental factors. 

Many structures are in themselves asymmetric. In leaves of Begonia 
and elm, for example, the portions on either side of the midrib are usually 
quite unlike; and there are marked internal asymmetries, as when two 
daughter cells are dissimilar. In most cases of this sort, however, the 
asymmetry proves to be part of a larger and more complex pattern which 
is symmetrical. 

Symmetry is often more conspicuous in embryonic structures or meri- 
stematic regions than at maturity, and some of its most remarkable ex- 


148 The Phenomena of Morphogenesis 

pressions are in soft and watery structures which seem to be a direct 
expression of protoplasmic configuration. Even though protoplasm is seem- 
ingly an amorphous and semiliquid material, these structures that it builds 
are far from formless, and the beautiful symmetries that they display 
seem clearly to be manifestations of the fundamentally symmetrical char- 
acter of living stuff itself. Organic symmetry presents a basic problem for 
students of morphogenesis. 


There are many examples of symmetry among inorganic objects. These 
often resemble the symmetries of living things, but there are certain 
fundamental differences between them. What the relation between these 
two types may be and whether organic symmetries have their origin in 
those of the inorganic world are problems that have long been discussed 
but are still far from solution. 

The arrangement of iron filings around the two poles of a magnet is a 
familiar example of symmetry, as are the lines of force in an electrostatic 
field. The least-surface configurations shown by liquids and especially by 
liquid film systems provide beautiful examples of symmetry. The resem- 
blance of such systems to multicellular structures in plants and animals, 
particularly the more minute ones, has been observed by many biologists 
and is discussed at length by D'Arcy Thompson ( 1942), who has analyzed 
the various forms possible in a film system. The molecular forces that 
operate here, however, are probably not important in determining the 
symmetry of large organic bodies. 

Much is now known, from X-ray studies and other sources, of the actual 
structure of molecules, and these are found to display symmetries, often 
very complex and specific ones. Whether such molecular forms have any 
relation to the bodily forms of plants and animals is a problem which has 
aroused much speculation but on which little evidence is available. Harri- 
son ( 1945 ) has discussed some of the possibilities here. 

Crystals provide the most familiar and remarkable examples of sym- 
metry in the inorganic world. Their very specific forms are the reflection 
of the forms and relationships of the molecules that compose them. The 
study of crystal symmetry is a complex science in itself and has intimate 
relationship to geometry, chemistry, and mineralogy. Crystals possess 
axes and planes of symmetry, as do organic structures, but crystalline sym- 
metry is a much more formalized and rigid phenomenon than organic. 
Many biologists have endeavored to find a relation between crystals and 
organisms in their form and symmetry, but this search, in general, has 
been a rather fruitless one. Organic symmetries can be described in the 
same geometrical language that we use for crystals, but whether there is 

Symmetry 149 

any fundamental similarity between the two is uncertain. For a discussion 
of this problem the reader is referred to the work of Haeckel (1866) and 

Although many symmetries in cells and minute multicellular structures 
resemble those in inorganic systems under the control of surface forces, 
organic symmetries are conspicuous in much larger bodies where these 
forces are not operative. Organic bodies are semiliquid systems which are 
subject to continual loss and replacement of material, as is shown by 
tagged isotopes and in other ways, and in this respect are unlike crystal- 
line structures, which are usually fixed and static. This semiliquid char- 
acter is also reflected in the almost universal presence of curved lines and 
surfaces in organic bodies as compared with the systems of straight lines 
and planes which distinguish molecular and crystalline forms. This is 
what makes possible the infinite number of similar planes of symmetry 
around an organic axis instead of the limited number of two, three, four, 
and six found in crystals. 

Aside from these differences from the inorganic, the symmetries shown 
by living plant structures also possess two distinctive features of their 
own which provide the key to an understanding of their nature. 

First, they are often expressed in multiple parts. A typical plant body 
consists of an indeterminate series of repeated, essentially similar parts, 
laterally dispersed along a continuous axis. These are leaves, branches, 
and lateral roots in higher plants and analogous repetitive structures in 
lower ones. The most conspicuous examples of organic symmetry are 
found in the relations of these repeated structures to the axis from which 
they arise. This is a type of symmetrv unlike that found in most inorganic 

Second, many plant axes, particularly those of the aerial portions of the 
plant, have either a spiral twist or a spiral arrangement of their parts. 
This complicates the expression of symmetry and, in the case of phyllo- 
taxy, has given rise to a great deal of speculation. Spirality seems to be a 
characteristic feature of protoplasmic behavior in many cases. The course 
of streaming is often spiral in a cell and thus may be reflected in the 
structure of the cell itself, as in the familiar cases of Chara and Nitella. 
Cell growth may be spiral, as has been shown by Castle ( 1936 ) in the 
hyphae of Phycomyces, and there are many other examples. 

These two traits— multiple parts and spirality— make the symmetry of 
plant parts radically different, at least in external expression, from the 
symmetries of the inorganic world. 

A few single-celled forms and some colonies like those of Volvox may 
be spherical and completely symmetrical around a point. This seems to 
be primarily an expression of surface forces, however, rather than of in- 
herent symmetry. Most single-celled plants, however, like the desmids, 

150 The Phenomena of Morphogenesis 

and most protozoa, show axial symmetry, often complex in character; and 
all typical examples of symmetry in higher plants are those manifest 
around an axis or a longitudinal plane. Three general types of such sym- 
metries are recognized: radial, bilateral, and dorsiventral. All the com- 
mon patterns for the structure of plants and their organs are based on 
these symmetries. 


In this type there is an axis of rotation around which symmetry is uni- 
form. There may be one or two evenly spaced longitudinal planes of 
symmetry, as in stems with distichous and with opposite-leaf arrange- 
ments, or these may be almost infinite in number in stems with spiral 
symmetry. Radial symmetry is present in vertically elongated axes such 
as those of the main stem and primary root and in many flowers and fruits. 
It is therefore much commoner in plants than in animals, since most of 
the latter show dorsiventrality, and it is regarded by many as the most 
primitive type of symmetry, at least in vascular plants, since their first 
axes were presumably vertical. 

In Lower Plants. Individual cells often show radial symmetry regard- 
less of their orientation, as in Spirogyra and Chara. Many plant bodies in 
the thallophytes have this type of symmetry, a familiar example of which 
is the "mushroom" form of sporophore in the fleshy fungi. Many red algae 
have radial thalli, as do some brown algae. Most true mosses also are 

In Roots. Almost all roots are radially symmetrical. This symmetry is 
shown in the straight and evenly spaced rows of lateral roots and in the 
characteristically radial primary vascular structures, in which arms of 
xylem typically alternate with bundles of phloem in a star-shaped pattern, 
with lateral roots arising opposite the xylem arms. 

In two respects the expression of symmetry in the root is different from 
that in the stem. Roots, even horizontally growing ones, are usually 
strictly radial and (save for a few cases such as air roots of orchids) 
show no dorsiventrality, regardless of their orientation, whereas horizon- 
tal stems commonly do show this. Roots also have very little twisting or 
spirality in their internal or external structures, such as most stems dis- 
play. These two differences emphasize again the fundamental diversity in 
developmental behavior of root and stem which is evident in many other 
respects. Whether these differences are inherent or are due to the radical 
differences in the environment in which roots and stems usually develop 
is an interesting morphogenetic problem. 

In Shoots. The symmetry of shoots, and particularly that shown by the 
arrangement of leaves (phyllotaxy), has attracted more attention than 

Symmetry 151 

any other aspect of the problem of organic symmetry. The external sym- 
metry of an upright stem is typically radial and often very regularly so. In 
the simplest cases, as in certain mosses such as Fontinalis, this is related to 
the activity of a three-sided apical cell, the segments cut off from its three 
sides giving rise to three rows of leaves. In higher plants, however, leaf 
arrangement is not related to meristematic structure. 

In stems with opposite leaves, at successive nodes the leaf pairs rotate 
through 90°. This decussate phyllotaxy thus shows four rows of leaves 
along the stem. More frequently, phyllotaxy shows a spiral character. 
Sometimes this is manifest, even in opposite-leaved types, by a twisting of 
the whole axis so that members of successive pairs are a little more than 
90° apart. Spirality more commonly expresses itself, however, in the ar- 
rangement of so-called "alternate" leaves. These are rarely exactly alter- 
nate but are so dispersed that a line connecting the points of attachment 
of successive ones follows a regular spiral course around the stem. The 
fact of this spiral and the various types in which it is manifest have for 
many years attracted the interest of botanists and mathematicians. Many 
of the discussions and speculations that have centered about the phyllo- 
tactic spiral are of no great significance for morphogenesis. Some are 
highly theoretical or even almost mystical. The developmental origin of 
the various types of phyllotaxy, however, is an important morphogenetic 
problem, and a knowledge of the factors involved may contribute to an 
understanding of the origin of organic form. 

Goethe was greatly attracted by the spirality of leaf arrangement and 
made it the basis of one of his theories. Charles Bonnet ( 1754 ) in the 
middle of the eighteenth century discussed the spiral structure of the pine 
cone. It was the work of Schimper (1836) and Braun (1831), however, 
that established the study of phyllotaxy on its modern basis. Various ex- 
planations of the origin and significance of spiral leaf arrangement have 
been proposed, and the literature of the subject is extensive. No compre- 
hensive review of it is available, though the earlier literature has been 
surveyed by C. de Candolle ( 1881 ) . For the more important ideas the 
reader is also referred to the works of the brothers Bravais (1837), Hof- 
meister (1868), Wright (1873), Schwendener (1878), Schoute (1913, 
1914), Church (1920), Hirmer (1922), Crow (1928), Goebel (1928), 
Snow and Snow (1934), D'Arcy Thompson (1942), Plantefol (1948), 
Wardlaw (1949a), and Richards (1950). 

Spiral phyllotaxy is not an example of symmetry in the strict sense since 
planes of symmetry, in the crystallographic meaning of the term, are 
absent. The leaves do have regular positions along the axis, however, with 
reference to each other, and these, under proper analysis, can be ex- 
pressed in terms of geometrical symmetry. The spiral formed by the 
points of attachment of successive leaves— the genetic or developmental 

152 The Phenomena of Morphogenesis 

spiral— represents the order in which the leaf primordia are formed in the 
bud. Their positions in the spiral are not indefinite but commonly fall into 
a few precise categories, the relations of which have long excited the 
interest of morphologists. Simplest of all is the distichous, or truly alter- 
nate, arrangement, with successive leaves 180° apart around the stem. To 
pass from a leaf to one directly above it involves one circuit of the axis 
and two leaves, a condition which may be expressed by the fraction Vo. In 
other types this spiral passes once around the axis but every third leaf is 
over one below it, a condition that may be represented by the fraction %. 
Commoner than either of these is a spiral where in passing from a leaf to 
one above it two circuits of the axis are made and the fifth leaf is reached, 

Fig. 7-1. Diagram showing % phyllotaxy. 

the % type. Frequently observed in stems is a % phyllotaxy (Fig. 7-1) 
and less commonly that of % 3 . In cones and other compact axes more 
complex phyllotaxies of % t , x % 4 , and 2 % 5 may be found. The series is 
thus V 2 , Vs, %, %, % 8 , % lt 13 / 34 , 2 %5, 3 %9, and so on. Each obviously 
represents the fraction of the circumference of the axis, or the angle, 
traversed by the spiral in passing from one leaf to the next. The number 
in both numerator and denominator of each fraction is the sum of those 
in the two preceding fractions. This particular series is known as the 
Fibonacci series. The higher fractions become more and more uniform 
and approach as a limit the decimal fraction 0.38197, or the angle 
137°30'28", the so-called "ideal" angle. It has been shown by Wright 
( 1873 ) that if successive leaves were formed at just this angular distance 

Symmetry 153 

around the stem from each other no leaf would ever be directly over any 
below it. The advantage sometimes suggested for this arrangement, that it 
would distribute the leaves most evenly to the light and thus be most 
efficient in preventing shading, is open to many objections. 

The fraction 0.38197 is of interest in another connection, for it desig- 
nates the "golden mean," or sectio aurea, the distance from the end of a 
line at which, if the line is cut there, the smaller fraction of the line is to 
the larger as the larger is to the whole. Thus 0.38197 : 0.61803 = 0.61803 : 
1.0. The golden mean has long been known and has received much atten- 
tion from artists and mathematicians, and its significance in the geometry 
of symmetry may be considerable, but its biological importance seems 
negligible. One should also remember that there are other series of frac- 
tions which converge to the same limit. 

This analysis of the genetic spiral assumes that, as it twists around the 
stem, a given leaf position on it is directly over one below, after passing 
3, 5, 8, 13, etc., leaves on the spiral. Thus there should be vertical rows 
of leaves, relatively few in the simpler phyllotaxies but more numerous 
in the complex ones. These have been called orthostichies and mark the 
end points of each successive fraction into which the genetic spiral is 
divided. Their presence is essential if the mathematical analysis of this 
spiral, going back to the work of Schimper and Braun and elaborated by 
so many botanists since then, is to mean very much. 

The existence of these orthostichies, however, has been challenged by 
more recent students of phyllotaxy, who have approached its problems 
not by an analysis of mature structures but by a more truly morphogenetic 
investigation of the way in which the leaves originate. The best place to 
study leaf arrangement, they maintain, is in the bud or at the apical 
meristem. Church (1920), one of the pioneers in this method of attack, 
discovered that in the arrangement of primordia as seen in a cross section 
of the bud there are no orthostichies at all, for no leaf primordium arises 
directly over one below. Thus doubt was cast on all the early conclusions- 
based on the assumption that the genetic spiral could be divided into re- 
peated portions. 

But other relationships are more important than this. A study of leaf 
primordia packed into the bud, or of other cases such as pine cones and 
sunflower heads where there are a great many structures spirally arranged 
but crowded together, shows the existence of another series of spirals, 
resembling the genetic one in certain respects but reached in a different 
fashion. If one looks at the cross section of a bud, or the face view of a 
sunflower head in fruit, or the base of a pine cone, he will notice that the 
units are not packed uniformly together like the pores of a honeycomb. 
Instead, the various structures— leaves, primordia, fruits, or scales— form 
two sets of spiral curves, starting in the center and moving to the circum- 

154 The Phenomena of Morphogenesis 

ference, one going to the right (clockwise) and the other to the left 
(counterclockwise). The effect is something like that of a spinning pin- 
wheel, or rather of two spinning in opposite directions. The inner mem- 
bers of each spiral are progressively smaller since they were formed later. 
We are looking down, in effect, on the top of a growing system, even 
though growth may have stopped. These spirals are logarithmic ones, 
since the radial distance to each successive unit on them increases geo- 
metrically and not arithmetically. The spirals are termed parastichies, 
or sometimes contact parastichies since each unit is usually somewhat 
flattened against its inner and outer neighbors in the spiral, a fact which 

Fig. 7-2. A pine cone seen from below. The scales are in two sets of parastichies, 8 
counterclockwise and 13 clockwise. 

makes the spiral easy to trace ( Fig. 7-2 ) . The spirals intersect each other 
at an angle which is near to 90°. 

In a bud or meristematic tip that will give rise to a shoot with a rela- 
tively low phyllotactic fraction ( % or % ) , the units are fewer than in 
large structures like a cone. In a cross section of such a bud (Fig. 7-3) it 
is possible to distinguish by the relative sizes of the leaf primordia the 
order in which they were produced. The genetic spiral can thus be traced, 
compact and almost two-dimensional here although later it will be pulled 
out like a telescope when the shoot elongates. In a bud like this one can 
confirm the observation of Church, that orthostichies do not exist. Were 
they here, they would appear as radial rows made by every fifth, or 

Symmetry 155 

eighth, or thirteenth primordium. These are not to be seen. Furthermore, 
if one carefully studies the angular divergence between successive pri- 
mordia he finds (in the great majority of cases) that it is close to the 
"ideal" Fibonacci angle of 137.5° which the series of phyllotactic fractions 
approaches as a limit. 

The number of clockwise and of counterclockwise parastichies in a 
given axis is not the same. In different types, however, their relative num- 
bers are fixed and specific. These also fall into a characteristic series. Thus 
in the bud section shown in Fig. 7-3 one can count five parastichies turn- 

Fig. 7-3. Cross section of apical bud of Pinus pinea showing absence of orthostichies. 
The primordia, numbered in succession, are separated by the Fibonacci angle. Five 
counterclockwise parastichies and eight clockwise ones are evident. (From R. Snow, 
courtesy of Endeavour. ) 

ing to the left and eight to the right. In simpler forms there may be three 
in one direction and five in the other. In more complex cases, such as 
many pine cones, there are 8 of one and 13 of the other, or 13 of one and 
21 of the other. Some systems have 21 and 34. Most sunflower heads show 
34 spirals in one direction and 55 in the other. Arranging these pairs of 
numbers in the form of fractions, as was done with the genetic spiral, one 
obtains the series %, %, %, 8 / 13 , 13 / 2 i, 21 / 34 , 3 %5, 5 %9, and so on, though 
the higher fractions are rare. The numbers in numerators and denomi- 
nators form a series, as in the genetic spiral, but here the denominator of 
one fraction forms the numerator of the next one instead of the next but 

156 The Phenomena of Morphogenesis 

one. The fraction which this series approaches as a limit is 0.61803, the 
larger one of the two which are separated by the golden mean. This frac- 
tion is thus the difference between 1.0 and 0.38197, the limit approached 
by the other series. The two spiral systems are evidently related but just 
how they are is a nice mathematical problem. It is no wonder, as D'Arcy 
Thompson says, that these various relationships have long appealed to 
mystics and to those who seek to square the circle or penetrate the secrets 
of the Great Pyramid! 

Parastichies are present in shoots around which leaves are borne in a 
phyllotactic spiral, but because they are pulled out so far lengthwise 
they are much less conspicuous than when many structures are packed 
together. In elongate shoots orthostichies, though absent in buds, can 
usually be demonstrated. The tensions resulting from elongation ap- 
parently operate to straighten out the spirals and in many cases to bring 
the insertion of a leaf almost directly over one that is three or five or 
eight leaves below. Not much of a twist is needed to accomplish this and 
to produce an orthostichy. One should recognize, however, that such are 
secondary rather than primary phenomena of symmetry. 

The problems of phyllotaxy were already involved enough when a 
French botanist, Lucien Plantefol (1948), added a further complexity. 
His theory has been extensively developed by others, particularly in his 
own country. It is based on a study of the insertion of the leaf traces on 
the stem. Plantefol does not regard the genetic spiral as significant. He 
traces two (sometimes more) foliar helices connecting the leaf bases in 
parallel spirals that wind up the stem, and he usually represents these 
helices as projected on a plane where their relationships can more easily 
be seen (Fig. 7-4). They originate in the traces of the two cotyledons, 
and the series remain distinct as they pass up into the bud. Here they 
terminate in a generative center of embryonic tissue just below the tip 
of the meristem ( Fig. 7-5 ) . In this the new primordia are differentiated. 
The position of each is determined, he believes, by stimulation from the 
foliar helices below, the relations being harmonized by an "organizer." 
Lance (1952) found in a number of cases a zone of abundant mitoses 
somewhat below the apex of the meristematic dome but few at the very 
tip. Crockett (1957) finds some evidence of the same thing in Nicotiaha. 
Loiseau ( 1954 ) cut off the tip of the meristem in Impatiens and observed 
that in many instances this resulted in changing the number of helices. 
This he believes was due to a disturbance of the generative center. 
Popham (1958), on the other hand, in a census of mitoses at the apex 
of Chrysanthemum, found no evidence of a generative center nor of its 
necessary corollary, a region of few mitoses at the very tip. Newman 
(p. 60) made the same observation in living material. The problem has 
been discussed by Wardlaw (1957b), who concludes that, although there 

Symmetry 157 

are various complexities in the shoot meristem, there is little good evi- 
dence from this source in support of Plantefol's theory. 

Although this theory has received strong support from a number of 
French botanists, objections have been raised against it in other quarters. 
What are chosen as foliar helices are evidently one of the parastichies or 
spiral rows of leaf traces to be seen along the axis, but which of these is 
the true helix in any instance seems difficult to determine. The leaves 
on a helix must have some vascular connection with each other, accord- 
ing to the theory, but in most stems at least two different parastichies 
could be chosen which would fulfill this requirement. A figure in one 

Fig. 7-4. Two foliar helices, the members of one connected 
by dots and of the other by dashes, seen as though 
the surface of the stem were removed and spread out. 
(From Plantefol.) 


Fig. 7-5. Diagram of shoot apex according to Plantefol's 
hypothesis, showing the generative center and the ab- 
sence of divisions at the very tip. Dots indicate mitoses. 
( From Plantefol. ) 

of Dr. Esau's papers (1943, Fig. 1), though it was not drawn to clarify 
the problems of phyllotaxy, makes these relationships evident (Fig. 
7-6). This is a diagram of the primary vascular system of Linum. The 
genetic spiral is shown, with the leaves numbered along it. In the series 
25-33-41, the bundle is continuous with a branch that passes laterally to 
the next in the series, and there are eight of these helices around the 
stem. In the series 28-33-38, the right-hand lateral of one is continu- 
ous with the left-hand lateral of the next, and there are five helices. 
Members of the series 30-33-36 have no direct vascular connection with 
one another but are in a definite row. Which of these spirals should be 
chosen as the foliar helix? One might determine the true one, perhaps, 



• 37 

Fig. 7-6. Diagram of the primary vascular strands in 
the stem of Linum perenne. The numbers from 21 
(above) to 49 (below) mark the positions of leaves in 
the generative spiral, indicated by a thin line. The 
bundles are shown by heavy lines for those on the 
nearer surface and by dotted ones for those behind. 
Various helices, in Plantefol's terminology, may be dis- 
tinguished, such as 30-33-36, 28-33-38, and 25-33-41. 
There seems no certain way to determine which are 
the "true" ones. ( From Esau. ) 







Symmetry 159 

by tracing the system back to the cotyledonary node. Camefort (1956) 
has presented a full account of Plantefol's theory and has endeavored to 
reconcile it with the classical concepts of phyllotaxy and modern experi- 
mental studies. 

The solution of these problems of leaf arrangement is evidently to be 
sought near the apical growing point where the leaf primordia actually 
originate, for their relations here will determine those between mature 
leaves on the elongated stem. This emphasis on the study of primordia 
is a return to the point of view of Schwendener ( 1878 ) , who believed 
that mechanical contact and pressure exerted by the primordia on one 
another accounted for their distribution, and especially of Hofmeister 
(1868), who proposed the general rule that a new primordium arises 
in the largest space available to it. This conclusion is generally accepted, 
but the developmental basis for it is not clear. The essential morpho- 
genetic problem beneath all this is what determines the origin of a par- 
ticular primordium at a particular place and time. 

An early idea was that a leaf-forming stimulus passes along the genetic 
spiral, but the significance of the spiral itself now seems rather slight. 
Church (1920), concerning himself chiefly with parastichies, believed 
that the point of intersection of the two major ones determined the point 
of origin of a primordium. This leaves undetermined the reason for the 
course of the parastichies themselves. Some workers are inclined to think 
that stimuli from previously formed leaves or primordia determine the 
position of new ones. Plantefol assumes that a foliar helix extends up- 
ward into the meristem to the generative center where the primordia are 
formed. Sterling ( 1945 ) finds that in Sequoia the procambial strands are 
always continuous with the older ones below and differentiate acropetally, 
pushing up into the apical meristem before the emergence of the pri- 
mordia into which they will pass, and suggests that these procambial 
strands may influence the position of the primordia. Gunckel and Wet- 
more ( 1946 ) reach the same conclusion for Ginkgo. Opposed to this 
idea is some experimental evidence, chiefly derived from isolating part 
of the meristem from regions below it by incisions without disrupting 
normal phyllotactic arrangement, a result which suggests that the stimu- 
lus for the development of a primordium does not come from below. 
Perhaps in such cases as this it is incorrect to assume that a given devel- 
opment is the cause of another which succeeds it in time. A series of re- 
lated structures and processes are part of the same organized whole and 
should be thought of as developing together rather than each step as 
inducing the one that follows it. 

The problem of what determines the phyllotactic series is open to 
experimental attack, and much work has been done on it by various 
people, among them Wardlaw, Ball, and especially the Snows. They have 

160 The Phenomena of Morphogenesis 

been able to modify phyllotaxy operatively in a number of ways. Thus 
in Epilobium hirsutum, a species in which the leaf arrangement is 
decussate (opposite), the Snows (1935) split the apex diagonally and 
found that the two regenerating shoots had spiral phyllotaxy. They were 
also able ( 1937 ) to change the phyllotaxy in the same way by applying 
auxin to the shoot apex. There may be a rather delicate balance between 
decussate and spiral phyllotaxy in this plant, for in the group to which 
it belongs (and even in a single plant of this species) both types may 
occur. In other plants the direction of the phyllotactic spiral may be 
reversed in regenerating shoots after splitting the apex. 

What determines the location of a given primordium is the basic prob- 
lem here, and as to this there are two major hypotheses. One, first pro- 
posed by Schoute (1913), assumes that the presence of a primordium 
tends to inhibit the development of another one near it, presumably by 
the sort of inhibition by which one bud checks the growth of another 
through the agency of auxin. This is the same problem studied more 
recently by Biinning (p. 199), who has evidence that each stoma produces 
a substance that prevents the development of another stoma close to it, 
thus accounting for the regular spacing of these structures. Such a hy- 
pothesis is in harmony with physiological theory, but some experimental 
results seem to be opposed to it. For example, the Snows (1952) re- 
moved the youngest actual primordium in an apex of Lupinus and after 
14 days determined the positions of the next three successive primordia 
that had appeared since this was done. In every case these later ones 
occupied the places in which they normally would have appeared, in- 
dicating that their positions in the series had not been affected by removal 
of a primordium and any inhibitory influence from it. 

That the primordia develop independently of either stimulatory or in- 
hibitory influences from neighboring ones is also shown by an experi- 
ment of Wardlaw's in which he isolated by radial cuts the areas pre- 
sumably to be occupied by the next primordia in the series, thus effec- 
tively isolating them from physiological contact with primordia already 
formed. He found that these areas developed primordia normally. 

Other factors than chemical ones may here be involved. Wardlaw 
( 1948 ) finds that each primordium tends to produce a region of tangential 
tensile stress around it but that this is absent in the area where new 
primordia are to arise. He suggests that a primordium will develop where 
tensile stress is at a minimum. 

The second hypothesis assumes that a primordium will not develop 
unless there is sufficient available free space for it. This is related to the 
ideas of Schwendener and Hofmeister and really comes down to the 
problem of the most efficient filling of the space on the surface of the 
meristem. It has been supported, in essence, by van Iterson (1907), and 

Symmetry 161 

in recent years the Snows have brought forward evidence in its favor. 
Among other experiments ( 1952 ) they isolated by two radial cuts the 
larger part, but not the whole, of the area presumptively to be occupied 
by the next-but-one leaf primordium. In such a case none develops be- 
tween the cuts, although this region grows and continues otherwise to 
be normal. They explain this result as due to the fact that the area now 
available was too small for a primordium to be formed. 

These two hypotheses, though stressing different factors, are not dia- 
metrically opposed to each other. What is to be explained is the even 
distribution of primordia, equidistant from each other (in origin) and 
regularly arranged. This is the same problem posed by the distribution 
of multiple structures. Something regulates the differentiation of each of 
these structures in such a way that each occupies an area of its own and 
that these individual areas are of about the same size. In the case of 
leaf primordia the situation is complicated by the fact that these arise 
on a curved surface and in a progressive series in time. Although me- 
chanical and chemical factors are doubtless involved in the distribution 
of primordia, as in all morphogenetic processes, it is perhaps too 
simple an explanation to regard the determination of each as due to 
crowding by its neighbors, to the presence of the "first available space," 
or to inhibition by other primordia. It seems more logical to regard the 
problem of the distribution of primordia at the growing point as another 
instance of a self-regulating biological pattern which may have its roots 
in genetic factors, the fine structure of protoplasm, or whatever else may 
be responsible for organic form. 

On either hypothesis mentioned above, if primordia are to arise in a 
spiral around the axis each should be as far as possible from its immediate 
neighbors, those coming just before and just after it in origin. In op- 
posite leaves each is placed as far away as possible, 180°. In spiral 
phyllotaxy this cannot be done. If primordium B, let us say, originates 
at an angle from A of 137.5° (the golden-mean fraction of the circum- 
ference), and if the next one, C, is placed at the same distance farther 
on (thus incidentally dividing the remainder of the circumference by 
the same ideal proportion), B is equidistant from A and C, and this is 
the maximum possible distance at which successive members can be 
placed from each other. If the distance A-B and B-C is less or greater 
than this ideal angle, C will not arise in the middle of the largest space 
available, as Hofmeister's postulate requires. What this means is that 
only if successive primordia are separated by this ideal angle will they fill 
the available space evenly and with the greatest economy. This is the 
property of golden-mean spacing that makes it significant in problems 
of this sort. 

Richards (1948, 1950) has worked out some of the implications of 

162 The Phenomena of Morphogenesis 

this fact and has returned to methods of mathematical analysis in ap- 
proaching the problem of the development of the primordia at the meri- 
stem. He emphasizes the importance of the plastochron ratio, the ratio of 
the radial distances from the center of the meristem to two consecutive 
primordia. Where this distance increases considerably in each plasto- 

Fig. 7-7. Diagram showing a spiral succession of points, each separated from the next 
by the Fibonacci angle, or about 137.5°. Parastichies can be recognized by intersec- 
tions of approximately 90° between them. At the center there are five counterclock- 
wise ones and eight clockwise. The counterclockwise series soon shift from 5 to 13, 
and later the clockwise ones from 8 to 21. (From F. J. Richards.) 

chron, both the genetic spiral and the spirals of the parastichies will open 
out rapidly. The meristem itself under these conditions will tend to be 
relatively steep, the primordia few and the parastichy numbers low. On 
the contrary, when the radial distance increases but little from one 
primordium to the next, the primordia are packed closely, the meristem 

Symmetry 163 

is likely to be flatter (as in a cone or flat head), the primordia will be 
more numerous, and the parastichy fractions will have higher numbers. 

Richards calls attention to the fact that the parastichies at a growing 
point are not limited to the two conspicuous "contact" ones emphasized 
by Church but that there may be a series of others though these are 
not obvious since they do not intersect each other at right angles. An 
advantage of this concept is that it makes clear how the parastichies shift 
from one pair of numbers to another, a problem that has troubled stu- 
dents of phyllotaxy. Richards (1948; Fig. 7-7) has constructed a diagram 
of a rather large meristem, something like a sunflower head, showing a 
series of primordial positions numbered along the genetic spiral in 
which each diverges from the last by the Fibonacci angle of 137.5°. In 
such a system one can readily trace parastichies. Near the center there 
are five counterclockwise ones crossing eight clockwise, the % arrange- 
ment, which intersect at approximately right angles. To trace this series 
very far out becomes difficult since the angles of intersection diverge in- 
creasingly from 90°. As one moves out, therefore, the system seems to 
change and the five counterclockwise spirals shift to thirteen, giving the 
% 3 arrangements of spirals that now have more nearly right-angled 
intersections. Still farther out the eight clockwise spirals are less easy 
to trace, and 21 others become more conspicuous, now making the 1 %i 
arrangement and restoring the steeper intersections. Thus in the more 
complex systems with large, flat meristems and little difference in radial 
distance between successive primordia, the parastichies, at least those 
that are conspicuous and easy to trace, may be seen to shift to pro- 
gressively higher numbers. This does not happen in ordinary shoots 
where the meristem is steeper and the primordia are fewer and larger 
and increase rapidly in size at each plastochron but it may sometimes be 
seen even in such cases (Fig. 7-8). These changes involve no biological 
mystery, as Church was inclined to believe they do, but are simply the 
result of the unique properties of the Fibonacci angle. 

Barthelmess ( 1954 ) has pointed out that the scheme proposed by 
Richards is essentially a two-dimensional one, whereas the meristematic 
region has three dimensions, a fact that must be taken into account. 
There are various other complications presented by an analysis of 
phyllotactic patterns. Bilhuber (1933) and others, for example, find 
that the situation in many of the cacti is often different from that in 
most families. These plants are essentially leafless and have angled 
stems so that in the apical regions one actually finds what look like 
orthostichies, which are related to the development of the angled pat- 
tern. Bijugate spirals (Hirmer, 1931; Snow, 1950) occur in some groups, 
where a % pair of parastichies, for example, becomes split into a % . 
Here primordia occur in opposite pairs but the plane of each pair is not 

164 The Phenomena of Morphogenesis 

at right angles to the previous one. This produces two parallel spirals. 
For a discussion of other recondite aspects of spiral phyllotaxy the reader 
is referred to the work of Church, Hirmer, Richards, the Snows, and 
others who have gone deeply into these problems. 

Of some morphogenetic interest is the direction of the genetic spiral 
itself. Observers generally agree that leaf positions around the stem are 
as likely to be in a clockwise as in a counterclockwise spiral. Beal ( 1873 ) , 
studying cones of Norway spruce, found 224 cases of the former and 
243 of the latter. Allard ( 1946 ) examined 23,507 tobacco plants and found 
that the two types were almost exactly equal. Direction of spirality was 
not inherited. How the direction is determined for a given plant is not 
known, but it is probably in some critical early cell division. This neutral- 
ity of the phyllotactic spiral is unlike the behavior of climbing plants, 

Fig. 7-8. Diagram of distribution of pri- 
mordia at the shoot apex, each diverging 
from its predecessor by the Fibonacci 
angle. As the primordia increase in size, 
the recognizable contact parastichies shift 
from 5 + 8 near the center to 8 + 13 
farther out. ( From Barthelmess. ) 

in almost all of which a given species climbs in either a clockwise or a 
counterclockwise manner exclusively. 

The difference in direction of the phyllotactic spiral, however, some- 
times alters from one part of the plant to another in conformity to a 
general pattern of symmetry, much as in the case of floral structures 
(p. 167). In shoot growth of Citrus, for example, Schroeder (1953a) re- 
ports that there are successive "flushes" with a dormant period between 
them and that a regular alternation occurs between right and left 
spirality in successive shoots. Secondary shoots have spirality opposite 
from their parent one. Thorns develop to the left of the petiole in left- 
handed shoots and to the right in right-handed ones. The direction of 
spirality in axillary shoots of View also takes place in a precise and 
alternating order which depends on their position (Dormer, 1954). 

Spirality. Spiral phyllotaxy involves a number of problems as to the 
spacing and relative position of leaves which are not present in decus- 

Symmetry 165 

sate (opposite-leaved) phyllotaxy. In the latter type the two primordia 
at a node are as far apart as they can be, and the position of each suc- 
cessive pair is at right angles to the pairs above and below. This more 
nearly fulfills the requirements for efficient spacing and maximum 
divergence than does the spiral arrangement. One therefore wonders 
why the latter is so much more common, particularly since the cotyledons 
and sometimes the first foliage leaves are opposite. The transformation 
of an opposite to a spiral phyllotaxy involves a radical rearrangement of 
the meristematic region. This seems to be an expression of an inherent 
tendency toward spirality which is evident in so many places in the 
structure and activity of plants and their parts. This inherent spirality, 
imposed on systems of different sizes and forms may, from the mere 
geometrical necessities of the case, result in the various systems of 
spiral phyllotaxy that we have been discussing. Physiological factors 
doubtless have an important role here— auxin, mechanical pressure, 
genetic determination of growth, and others— but the underlying spirality 
seems to be a phenomenon fundamental to all organisms. This may 
appear to be an oversimplification of a problem that has involved more 
diverse hypotheses than almost any other in plant morphology. If it proves 
possible, however, thus to get at the heart of this mass of facts and 
pick out one that underlies them all, we shall have come closer to an 
understanding of one aspect, at least, of the phenomenon of organic 

Spirality seems to be deeply seated in living stuff. It is evident in the 
spiral movements (nutations) seen in the growth of roots and shoots, 
particularly when this is speeded up by time-lapse photography. Tendrils 
coil spirally. Protoplasm streams in a spiral course. Molecules of DNA 
are spiral. Spiral threads (cytonemata) occur in cytoplasm (Strugger, 
1957). Spiral grain has been found almost invariably in tree trunks 
(Northcott, 1957). In protoxylem the wall markings are in spirals, save 
in the earliest cells, and there are spiral markings in many other xylem 
cells. Whether these are all due to the same basic cause may perhaps be 
doubted, but one can find spirality almost everywhere in the plant 

The simplest place to study it is in the cell itself and especially in 
the ceil wall. Much now is known about the submicroscopic structure of 
this wall and of the system of microfibrils that compose it (Preston, 
1952; Frey-Wyssling, 1953). The sporangiophore of the fungus Phycomy- 
ces is favorable material for this sort of work since, as it elongates, it 
twists spirally, as can be shown by following the course of marks placed 
on the cell surface (Castle, 1942). Spirality here seems to have its basis 
in the minute structure of the wall. Heyn (1939) believes that it is due 
to the fact that the chitin molecules which form the framework of the 

166 The Phenomena of Morphogenesis 

cell take up positions at angles of 13.5 or 27° from the long axis of the 
organ, these angles resulting from the character of the chitin molecule. 
Denham (1922) observed that in the cotton hair, at least, the spiral 
markings and striations on the wall coincide with the spiral path of the 
streaming nucleus and cytoplasm. Preston ( 1948 ) compares the growth of 
this cell to the pulling out of a flat, spiral spring which rotates as it ex- 
tends, and he has suggested an explanation for this in mechanical terms; 
but Castle ( 1936 ) thinks that this does not determine the structure of the 
growing wall, though it may produce a spiral layering where the wall is 
not elongating. He points out that not only the structure of the wall but 
its elastic properties must be taken into account and believes (1953) 
that, although the growth of the wall is helical, its course is not absolutely 
fixed by its structure, since the angle of spiral growth can be reversed by 
a change of temperature. Frey-Wyssling ( 1954 ) calls attention to the fact 
that in certain polypeptid chains the divergence angles between the 
amino acid residues show the same regularities as are found in the 
Schimper-Braun phyllotactic spiral and suggests that the same geometric 
cause— the necessity for most efficient packing— may underlie both. 

Green ( 1954 ) studied the growth of the long cells of Nitella which had 
been marked and found that these marks, as well as the two natural 
striations in the cell, showed a uniform dextral twist. Its regularity is 
maintained by growth processes evenly distributed through the whole 
cell and presumably resulting from changes in the fine structure of the 
cell wall. 

The spiral grain found in the wood of many trees is another manifesta- 
tion of spirality. This may be very conspicuous in some cases and seems 
to be most common in trees growing in exposed situations or under 
unfavorable conditions. The spiral may be right-handed or left-handed. 
This subject has been reviewed by Champion ( 1925 ) . Preston ( 1949 ) , 
using the data of Misra (1939), attempted to relate spiral grain to the 
spiral growth of single cambium initials and assumes that these twist or 
roll spirally. This would involve some slipping of cells past each other. 
From what is known of intercellular relationships, it is rather unlikely 
that such a change occurs. The essential fact in most cases of spiral grain 
is that vertical files of cells become tilted slightly to the right or to the 
left and that this results in a spiral course for the cells of the wood. Some 
slipping of the cells may be involved, but this might be accomplished 
by localized intrusive growth (p. 82) such as has been shown to take place 
at the tips of the cambial initials. The tilt seems to be related to a change 
in cell polarity. Neeff (p. 128) found that when a new polar axis was 
established the cambial cells gradually changed their direction until this 
became parallel to the new axis. That there may be a spiral polarity in 
the trunk itself is suggested by the work of Misra (1943), who reports 

Symmetry 167 

that where there is eccentricity in the woody axis the position of maxi- 
mum thickness in any eccentric ring follows a spiral course along the 
length of the axis. There is also a relation between this eccentricity and 
spiral grain, for the degree of both decreases upward in the trunk, 
and the direction of the spiral eccentricity (left or right) is the same as 
that of the spiral grain in any given axis. 

Priestley (1945) distinguishes between true spiral grain, characteristic 
of hardwoods and resulting from a twist in the primary cambium cylin- 
der, and tilted grain, characteristic of softwoods where the grain is al- 
ways straight in the wood of the first year. 

In Flowers and Inflorescences. Angiosperm flowers are apparently to 
be regarded, in an evolutionary sense, as shortened axes; and their parts, 
particularly the calyx and corolla, often show evidence of the same sort 
of spiral symmetry that exists between leaves. This can rarely be shown 
by the actual insertion of the parts, since they are at essentially the same 
level and might be regarded as a whorl, but is evident in the relation of 
their expanded portions to one another, particularly as visible in the bud. 
In flowers of dicotyledons there are usually in each circle five parts or a 
multiple of five. A very common relationship here (in the calyx, for ex- 
ample) is that two of the sepals have both edges outside the others, two 
have both edges inside, and one has one edge outside and one inside. 
This quincuncial arrangement can be interpreted through developmental 
evidence as a % spiral, since the parts appear in the same order as 
leaves in % phyllotaxy. Various modifications of this are found, but the 
typical dicotyledonous flower may be regarded in its symmetry as rep- 
resenting a % spiral. The flower of monocotyledons, on the other hand, 
has its parts typically in threes and may be regarded as a % spiral in 
symmetry. The problem of flower symmetry, particularly as expressed 
in transverse diagrams, has been the object of long study by floral 
morphologists and forms the basis of an extensive early literature (Eich- 
ler, 1875). 

For students of morphogenesis the symmetry displayed by inflo- 
rescences provides a notable example of the orderly control of growth 
relationships. Matzke (1929) has described a particularly fine example 
of such symmetry in Stelloria aquatica (Fig. 7-9). Here the inflorescence 
is a cyme, and the first flower terminates the main axis. Just below this 
flower arise two buds in the axils of opposite bracts, and from these buds 
shoots arise, each of which is likewise terminated by a flower. Below 
each of these flowers, in turn, two shoots again arise, and so on. The 
flower in this species shows quincuncial arrangement of the sepals. These 
sepals may show a clockwise spiral or a counterclockwise one. As an 
observer looks down on a diagram of such an inflorescence, it is evident 
that, of the two flowers below the first, one is clockwise and the other 


bO £ 'O 


Q a 


Symmetry 169 

counterclockwise and that this holds for each succeeding pair. This rela- 
tionship is not a random one, for the two types show a regular order as 
one progresses to successively later pairs so that the symmetry of each 
flower is predictable. Furthermore, the relative position in each flower 
of the "odd" sepal also changes with complete regularity. In one member 
of a given pair of flowers it has rotated 72° in a clockwise direction from 
the single flower next them, and in the other, 72° counterclockwise. The 
particular edge of the sepal which is inside also has a definite and pre- 
dictable position. The whole inflorescence is thus a complex pattern of 
symmetries, each successive floral meristem fitting precisely into its place 
in this pattern. The factors that determine the symmetry of each flower 
are therefore not purely local ones but operate as members of a much 
larger system. Such a system, with its parts so widely separated and so 
easy of observation, offers a particularly good opportunity for the ex- 
perimental study of symmetry. 


This is a relatively rare type in which there are two planes of sym- 
metry, so that front and back, and right and left sides, are similar. A 
bilaterally symmetrical organ resembles a radial one that has been com- 
pressed equally on two opposite sides. 

This type occurs chiefly in vertically oriented structures in which, from 
one cause or another, one of the dimensions is smaller. Thus the stems 
of certain cacti such as Opuntia are bilaterally symmetrical, as are the 
still further flattened phylloclads of Muehlenbeckia and Phyllocladus. 
These have doubtless arisen from radial types. The leaves of Iris and 
similar plants are essentially bilateral but have probably come from 
dorsiventral structures. All plants, such as the grasses and some other 
monocotyledons, which are truly distichous (the leaves arising only on 
two opposite sides of the stem) may be regarded as bilaterally sym- 
metrical. So may the flowers of the mustard family, Cruciferae, since two 
of the six stamens, directly opposite each other, are short and the other 
four long. A few of the simpler bryophytes have distichous leaves or 
leaf-like structures, as in Schizostegia, and are thus bilateral, as is the 
pinnate plant body of the coenocytic alga Bryopsis. The thallus of some 
of the larger algae, notably forms like Fucus and Laminaria, is flattened 
and shows this type of symmetry. 

In a few cases, as in some of the algae, a transition from radial to 
bilateral symmetry may be seen, and in Schizostegia the apex is radial. 
Doubtless in many instances one type could be induced from the other 
experimentally. Certain abnormal structures, such as many fasciated 
stems, are bilaterally symmetrical. 

170 The Phenomena of Morphogenesis 


In this type there is only one plane of symmetry, which extends 
vertically through one dimension of the structure. The two sides are 
alike but the front and back (or top and bottom) are not, thus distin- 
guishing it from bilateral symmetry. It is characteristic of structures grow- 
ing under an environment which is asymmetrical, as in the case of hori- 
zontal ones, of those exposed to light on one side only, and of those 
growing attached to some substratum. Among plants, creeping stems, 
rhizomes, most leaves, many thalli, a wide variety of flowers, and, in 
general, those structures which are not vertically oriented often show 
dorsiventral symmetry. 

Dorsiventrality in plants is manifest in external form, in internal struc- 
ture, and in physiological behavior. Single cells and coenocytes may 
show such symmetry. Dorsiventrality may be genetically determined and 
thus appear under various environments, or it may be directly induced 
by environmental factors. Thus a dorsiventral structure may sometimes 
become radial, and vice versa. In some cases the plant body may actu- 
ally alternate between radial and dorsiventral symmetry, as in Mnium 
undulatum and Cladonia verticillaris. 

Cases of dorsiventrality which are most obvious and easy to study are 
those in structures that are typically horizontal, either because they are 
weak and rest on the ground or because they are plagiotropic and tend 
to grow in a horizontal position. 

In Thalli. Among lower plants many thalli are dorsiventral. The 
coenocytic plant body of the alga Caulerpa is typically horizontal and 
on its lower surface bears "rootlets" and on. its upper surface, "leaves." 
The familiar heart-shaped prothallus of a fern is similarly dorsiventral, 
bearing sex organs and rhizoids on its lower surface only. This type of 
symmetry is characteristic of the plant body of many liverworts, both of 
the thalloid and the leafy types ( Halbsguth, 1953 ) . The factors which in- 
duce it in such plants have been studied by various workers (Fitting, 
1935, 1950; Pfeffer, 1871; Bussmann, 1939; and others, p. 355). Fitting 
studied especially the gemmae of liverworts. These are roundish, notched 
plates of cells the dorsiventral orientation of which is determined by the 
balance between light, gravity, and stimuli from the substrate, acting on 
preformed meristems in the notch. Fern prothallia exposed to an en- 
vironment without gradients ( shaken or on a turntable, p. 137 ) lose their 
symmetry as well as their polarity. 

In Roots. Dorsiventrality is much less evident in roots than in stems. 
Indeed, horizontally growing subterranean roots show little or no 
change from radial symmetry either externally or internally. A few forms, 

Symmetry 171 

such as Isoetes, have roots that are not radial. In a number of orchids, 
the air roots are dorsiventral in symmetry ( Janczewski, 1885; Goebel, 
1915). This is especially conspicuous where the root is in contact with 
a substrate (Bloch, 1935a). 

In Shoots. Horizontally growing stems and branches are often con- 
spicuously dorsiventral. Notable examples of this are found among the 
conifers where the lateral shoots tend to branch in a single plane and 
thus form flat sprays (amphitrophy). This form may become so firmly 
fixed that it persists even in cuttings (p. 189). In some species of Ly co- 
podium and especially Selaginella, these flattened branch systems look 
almost like much dissected compound leaves. Indeed, there is evidence 
that the large pinnately compound leaves of ferns may have evolved 
from such branch systems. In many horizontal shoots the leaves are 
usually horizontal in orientation and confined to the two sides. This may 
result from a torsion of the petioles which are actually inserted on the 
stem in a spiral or decussate fashion or, more rarely, from an actual 
modification of the phyllotaxy. 

Aside from this tendency to form flattened systems of leaves and 
branches, the dorsiventral character of shoots is also conspicuous in the 
dissimilarity of the leaves borne on the two sides. Such differences, to 
which Wiesner gave the term anisophylly (1895), are common in 
many plants and have been much discussed (Figdor, 1909; Goebel, 

Anisophylly is often induced by external factors, notably gravity and 
light. It is particularly conspicuous in woody plants with opposite, decus- 
sate leaves. In horizontal shoots of maple, for example, the upper mem- 
ber of a vertically oriented pair is much smaller than the lower; and in 
a horizontally oriented pair the upper half is smaller than the lower 
(Fig. 7-10). Experiment shows that in many cases if shoots which would 
normally be vertical are held in a horizontal position as they grow 
from winter buds they become anisophyllous. In horizontal branches 
twisted through 180° before their buds open, the new shoots show 
reversed anisophylly, the lower leaves (originally on the upper side) 
now becoming the larger. 

Anisophylly of this sort is present in certain species of Lycopodium 
(such as the common ground pine, L. complanatum), where the creep- 
ing rootstock is radially symmetrical but the ultimate branches flattened 
and dorsiventral (though they are radial if grown in darkness, Fig. 7-11). 
These branches have four rows of leaves, one on the upper side, one 
much smaller on the lower, and two lateral ones, the lateral leaves being 
the largest. Transitions from radial to dorsiventral symmetry are com- 
mon, and the differences between the two are clearly due to environ- 
mental factors. In conifers such as Thuja the ultimate branches are 

172 The Phenomena of Morphogenesis 

dorsiventral and anisophyllous and much resemble those of Lycopodium 
(all tending to grow horizontally). 

There are other plants in which the occurrence of anisophylly seems 
much less directly dependent upon environmental factors and occurs 
throughout the plant. This "habitual" anisophylly, as Goebel calls it, is 
probably to be interpreted as a genetic tendency to develop in this way 


Fig. 7-10. Diagram of a horizontally grown branch of maple, showing anisophylly. 
The vertically oriented pair of leaves (VU, VL) differ greatly in size but are sym- 
metrical. In the horizontally oriented pair (HL, HR), the lower half of each leaf is 
larger than the upper. ( From Sinnott. ) 

under such a wide range of environments that it has become essentially 
an inherited trait. In many foliose liverworts, for example, the axis has 
three rows of leaves, two of them lateral and the third, the much 
reduced amphigastria, borne on the under side. Most species of Sela- 
ginella have four rows of leaves: two lateral and relatively large and the 
other two on the upper surface between these and somewhat smaller. 

Symmetry 173 

Among some families of angiosperms this same genetic or habitual 
anisophylly occurs. Thus in Pellionia ( Urticaceae ) , in Centradenia 
( Melastomaceae, Fig. 7-12), and in Columnea ( Gesneriaceae ) one mem- 
ber of each pair is a large typical foliage leaf but the other, directly 
opposite it, is a small bract-like structure. These differences are ap- 
parently unrelated to environmental conditions. It is noteworthy, how- 

Fig. 7-11. Dorsiventral (flat) shoot of 
LycQpodium. At right is a branch grown 
in the dark, which is radially symmetri- 
cal. ( From Goebel. ) 

Fig. 7-12. Anisophylly in CeftfrL'denia. 
Leaves are opposite but one member of 
each pair is much larger than the other. 
Only the larger ones have axillary shoots. 
( From Goebel. ) 

ever, that this anisophylly is most extreme in horizontal shoots of such 
plants and is much reduced in those which grow more nearly vertically. 
On flattened plagiotropic shoot systems there are often changes in the 
pattern of symmetry that are more complex than anisophylly. In such 
shoots, for example, many leaves are asymmetric, but in a regular and 
predictable fashion (Fig. 7-13). Thus in horizontal branches of elm and 
linden, the inner half of the leaf, directed toward the apex of the shoot, 

174 The Phenomena of Morphogenesis 

is larger than the outer, and its blade often reaches farther down the 
midrib. In the beech, on the other hand, it is the outer part of the leaf 
which is the larger. Many cases of leaf asymmetry, notably the conspicu- 
ous examples in species of Begonia, are related to the position of the 
leaf on the stem, although here the stem is often short and inconspicuous. 
Somewhat similar expressions of apparent asymmetry are evident in 
the branch pattern of plagiotropic shoots. In some cases, the branches 
which arise on lateral shoots are larger on the inside, toward the apex 
of the shoot, as in flat stems of Thuja. More commonly those on the out- 
side, away from the axis, are larger, a phenomenon which Wiesner 
(1892a, 1895) has called exotrophy and which he explains as due to nu- 
tritional causes. Leaves on the outside of lateral shoots are often larger 
than those on the inside, a special type of anisophylly. 

Fig. 7-13. Anisophylly in Goldfussia. Diagram, of shoot from above. The leaves are 
opposite but the pairs are somewhat displaced. One member of each pair is larger 
than the other, and one side 'of each leaf is larger than the other side. In the axillary 
shoot, position with reference to the symmetry of the whole determines leaf size. 
( From Goebel. ) 

All such structures, which in a strict sense are asymmetric, are really 
complex patterns of symmetry induced when a fundamentally radial 
system becomes dorsiventral. What the factors are— whether nutritional, 
hormonal, or other— which determine these differences is not known. This 
is evidently the point where the relatively simple phenomenon of sym- 
metry merges into the more complex one of organic pattern in general. 
In flat, dorsiventral shoots, which are essentially structures in two di- 
mensions only, there is an excellent opportunity to analyze the problem 
of pattern in one of its simplest expressions. 

The external dorsiventrality of stems is often accompanied by dorsi- 
ventrality of internal structure. Where the stem is flattened, the vascular 
cylinder is likely to be so as well. Sometimes the symmetry changes do 

Symmetry 175 

not involve the whole cylinder. In the horizontal rhizome of Pteridium, 
for example, the outer ring of bundles is essentially circular in section, 
but the group of medullary bundles tend to be flattened dorsiventrally. 
In Selaginella the few bundles which form the vascular system also tend 
to be flattened in the same way. This flattening may even persist in those 
orthotropous shoots which have become radially symmetrical externally. 

Examples of internal asymmetry in the stems of seed plants are found 
in the horizontally growing branches of woody plants. Here the branch 
itself is not flattened but its internal structure is excentric, the pith oc- 
cupying a position some distance above the geometrical center of the 
branch in gymnosperms and below it in angiosperms. The nearer the 
branch approaches a vertical orientation, the less this excentricity is. There 
has been much discussion of the factors responsible for this internal 
dorsiventrality (p. 356). The problem is far from a simple one and seems 
to be involved with the specific pattern of branching characteristic of 
the plant. 

In Leaves. All leaves are typically dorsiventral structures, but those 
of pteridophytes and seed plants are most characteristically so. A leaf, 
to perform its usual functions satisfactorily, must be relatively broad and 
thin and oriented with its major surface at right angles to incident light. 

Dorsiventrality of leaves is especially evident in their histological 
structure. The stomata and spongy tissue tend to be confined to the lower 
part of the leaf, with the palisade layer and a continuous epidermis on 
the upper. Some vertically oriented leaves such as those of Iris are equi- 
facial and show no dorsiventrality, either external or internal. Others, such 
as those of certain rushes, may actually be tubular and essentially radial 
in their symmetry. 

The dorsiventrality of leaves in the higher vascular plants, however, 
is inherent in something more fundamental than the orientation of the 
blade. The vascular supply for each leaf is a segment, or group of seg- 
ments, of the primary vascular ring with phloem outside and xylem in- 
side. When this passes outward into the leaf as the leaf trace and finally 
becomes the vein system, the phloem therefore tends to be on the lower 
surface and the xylem on the upper, a characteristic dorsiventral orienta- 
tion from the first. Even the leaf primordia become dorsiventral very 
early. There is evidence, however, that this is the result of induction 
from the meristematic apex, for if a region where a primordium is to 
form is isolated from the apex by an incision, the structure that emerges 
may be radially symmetrical (Sussex, 1955). 

In Flowers. Floral structure provides many examples of dorsiventrality. 
The presumably primitive types of flowers are radially symmetrical, or 
actinomorphic (regular). In many families, however, such as the papi- 
lionaceous legumes, the figworts, the orchids, and others, especially those 

176 The Phenomena of Morphogenesis 

in which the flowers are borne laterally on an inflorescence, this radial 
symmetry has become dorsiventral and the flower is said to be zygomor- 
phic (irregular; Fig. 7-14). The pea flower, with its standard, two wings, 
and keel is a familiar example, and the median plane of symmetry here 
is especially well marked. Flowers of this sort provide many of the 
notable adaptations for insect pollination. In most cases zygomorphy is 

Fig. 7-14. Dorsiventrally symmetrical ( zygomorphic ) flower of Linaria vulgaris. 
( Courtesy of Rutherford Piatt. ) 

evident from the beginning of development and is unaffected by the rela- 
tion of the flower to gravity or other environmental factors. In other 
cases (such as Epilobium, Friesia, and Digitalis), if the flower develops 
in a vertical orientation or on a clinostat, it becomes radial, indicating 
that dorsiventrality here is directly affected by gravity (Fig. 7-15). In 
cases of peloria (p. 282) the flower of a species which is normally zygomor- 
phic (as in Linaria or Digitalis) may become radially symmetrical. Most 

Symmetry 177 

zygomorphic flowers are geotropic and will assume a definite position 
with relation to gravity. 

In some cases certain flowers of an inflorescence are dorsiventral and 
others radial. This is true of the ray florets of Compositae and of certain 
Umbelliferae, where that part of the corolla directed toward the outside 
of the head is much larger than that directed toward its center. In such 
cases the entire inflorescence shows a radial symmetry. Here, again, the 

Fig. 7-15. Flower of Asphodelus. 
Below, under normal conditions. 
Above, after developing on a 
clinostat. (From Vdchting.) 

whole pattern is symmetrical though certain of its elements are by them- 
selves asymmetric. The situation may be still more complex. In some 
Compositae there are as many as five types of fruits, as to size and shape, 
formed on the head but showing symmetrical distribution (Pomplitz, 

Most inflorescences (like that described for Stellaria, p. 167) are radially 
symmetrical, but some are definitely dorsiventral. A familiar example of 
this is the heliotrope and its allies, where the flower cluster is one-sided 

178 The Phenomena of Morphogenesis 

and constitutes a scorpioid cyme. The vetches and some other legumes 
are less extreme cases, and there are many others. The flowers of such 
dorsiventral inflorescences may themselves be radially symmetrical. 

Physiological Dorsiventrality. Dorsiventrality is manifest in physio- 
logical activity as well as in structure, though usually not so obviously. 
Plagiotropic roots, shoots, and other organs assume this position presum- 
ably because of specific distribution of growth substances in the growing 
tip such that the pull of gravity is counteracted and growth maintains 
either a horizontal course or one at a given angle to a vertical axis. In 
cases where the first division of a cell sets apart two different daughter 
cells, as in the first division of the egg of Fucus, there is clearly a physio- 
logical difference between the upper and lower halves. Indeed, the dif- 
ferentiation of root and shoot in the embryonic axis, with the radical 
differences in activity of these two poles, may be looked upon as an 
example of physiological (and morphological) dorsiventrality. 

In leaves of certain water plants, externally alike on both surfaces, 
Arens ( 1933 ) has presented evidence that the physiological activities at 
the two surfaces are unlike, materials from the environment entering 
through the lower surface and waste products (chiefly carbon dioxide) 
being given off from the upper. What the mechanism of such physio- 
logical dorsiventrality may be is not known, but bioelectrical differences 
(p. 361 ) are perhaps involved. 


The causes of organic symmetry are not well known, but in endeavoring 
to find them it is first necessary to determine how these relations actually 
arise in the process of development. 

The Origin of Symmetry in Coenocytic and Colonial Systems. Sym- 
metry is by no means confined to cellular structures. From the Plas- 
modium of myxomycetes, formless and unsymmetrical, there arise spe- 
cifically formed and radially symmetrical fruiting bodies of great variety. 
Here the morphogenetic process may be seen in one of its simplest ex- 
pressions, as the sporangium is molded from the plasmodial mass. 

Even more remarkable are those slime molds belonging to the 
Acrasiaceae (Dictijostelium and its allies, p. 223) where the vegetative 
body is a single myxamoeba. At the end of the vegetative period thou- 
sands of these come together into a pseudoplasmodium but do not fuse. 
This colony, after some migration, settles down and develops into a 
radially symmetrical stalked sorocarp. 

More closely resembling the bodies of the higher plants but still with- 
out cellular boundaries are the coenocytic members of the algae and 
similar groups. Here there is no formless mass of protoplasm but, from 

Symmetry 179 

the beginning, an organized system which grows at the tips of these 
branches that constitute the "rhizome," "leaves," and "roots." These sys- 
tems are symmetrical, either radially as in Bryopsis, or dorsiventrally as in 
Caulerpa. In all these cases the origin of symmetry obviously is not re- 
lated to planes of cell division or to other aspects of a multicellular sys- 
tem but is dependent upon the behavior of the entire protoplasmic sys- 

Origin of Symmetry in Cellular Systems. In cellular plants, the origin 
of symmetry can be traced more readily because it is expressed in the 
division, growth, and relationships of cells at meristematic regions. 

In simple colonial forms like Pediastrum there is a regular sequence 
of cell divisions from which a symmetrical plate of cells arises. In manv 
algae with an indeterminate thallus, growth is controlled bv a large 
apical cell. The origin of branches and the whole pattern of symmetry 
are determined here. In simple two-dimensional thalli, the apical ceil 
cuts off a daughter cell, first on the right-hand side and then on the 
left, to form the so-called pendular symmetry. In most leafy liverworts 
and mosses there is a pyramidal apical cell with three faces, and from 
each of these, in regular succession, a daughter cell is cut off. The origin 
of leaves is related to these faces, and in the simplest cases there are three 
rows of leaves produced directly by this apical cell. 

In ferns and Equisetum, however, which also grow by a three-sided 
apical cell, there is usually no relation whatever between the phyllotaxy 
of the shoot and the configuration of this cell. In the seed plants there is 
no single apical cell and no evident relation between the spiral pattern 
of symmetry and any visible structures in the meristem. It seems clear that, 
in all except the simplest plants, the origin of spiral symmetry is not re- 
lated to cellular configuration at the meristematic region but must have 
its basis in the entire embryonic mass. 

Dorsiventral symmetry in most cases is not established at the meristem 
itself but has its origin in changes which arise later. Almost all meristems 
or terminal embryonic regions are radially symmetrical. Dorsiventrality 
may arise from these in the process of normal development. This is some- 
times due to the influence of external factors such as light or gravity. It 
is sometimes the result of position in the general plant body, as when a 
branch becomes dorsiventral in symmetry. It is sometimes associated with 
particular stages in the life cycle. In plants that are dorsiventral through- 
out the mature plant body the seedlings are usually radial. In Hedera, the 
vegetative stage of the life cycle is dorsiventral but the flowering shoots 
are radially symmetrical (p. 213). Such changes are evidently due to 
alterations in the internal environment. 

Such modifications of symmetry, particularly the change from the 
radial to the dorsiventral type, involve not local regions but the entire 

180 The Phenomena of Morphogenesis 

pattern, which may be deformed much in the fashion that D'Arcy Thomp- 
son has demonstrated (p. 424). This can be seen by comparing the dorsi- 
ventral maple shoot in Fig. 7-10 with one growing vertically. 


An analysis like this emphasizes the close relationship that exists be- 
tween symmetry and organic form in general. Such form results from the 
symmetrical distribution of material around an axis in a specific pattern. 
An important part of this pattern lies in its symmetry. As we have seen, 
certain portions of the pattern (as the lateral leaves of the maple shoot 
in Fig. 7-10) appear by themselves to be asymmetrical, but they never- 
theless constitute a part of a larger pattern of symmetry which may be 
modified in various ways. A second part of the pattern is polar axiation, 
affecting the lengthenings or shortenings of the axis and the steepness of 
gradients along it. A third is the tendency toward spirality already em- 
phasized. Organic form results from the genetic and environmental 
modification of these three developmental tendencies. 



At the beginning of its development the young plant, as it grows from a 
fertilized egg or from some larger embryonic mass, is relatively simple 
and homogeneous. A characteristic feature of the developmental process, 
however, is the origin of differences in the amount, character, and loca- 
tion of growth which lead to differences between the various parts of an 
individual. Such structural or functional differentiation and its origin in 
development constitute one of the chief problems of morphogenesis. 

Differentiation is the manifestation of that "division of labor" which is 
so conspicuous a characteristic of living things. Organs are differenti- 
ated. Tissues in their development become unlike each other. Cells grow 
very diverse in character. Even the contents of a single cell are divided 
into nucleus and cytoplasm, and each of these possesses a considerable 
diversity of its own. There is evidence that even the clearest cytoplasm 
possesses submicroscopic differentiation. Strictly speaking, there is 
probably no really undifferentiated structure in a plant. Protoplasm is 
an organized system, not a homogeneous material, and this implies a 
degree of physical and chemical diversity. Furthermore, because of the 
dynamic quality of protoplasm, differentiation in living cells can never 
be entirely stable but is subject to change under changing conditions. 

Differentiation occurs wherever a true development is taking place and 
may be expressed in many ways. At a terminal meristem like that of a typi- 
cal shoot, the primordia of leaves, buds, and flowers early become dis- 
tinguishable. From cambium cells, uniform in character, there differen- 
tiate sieve tubes, fibers, tracheids, vessels, and other cell types. In the 
primordium of a fruit, where growth is diffuse and determinate, internal 
differences of many sorts begin to manifest themselves throughout the 
mass. In regenerative development a single cell or group of cells may 
dedifferentiate (p. 232) and become meristematic, and from this embryonic 
center a new series of structures then differentiates. Many differences have 
no visible expression in structure but involve physical, chemical, or 
physiological distinctions only. During ontogeny the course of dif- 


182 The Phenomena of Morphogenesis 

ferentiation often changes, not only as to the structure of the parts de- 
veloped but as to their reactivity and developmental potency. 

An important aspect of the process of differentiation is that it seems 
not to involve genetic diversity. The regeneration of an entire normal 
plant is sometimes possible from a single cell, which may come from 
almost any of the parts of the plant body, and from various tissues (p. 
253), a fact which suggests that every cell of the plant is totipotent and 
identical genetically with all the rest. This conclusion is supported by 
the common observation that the number and character of the chromo- 
somes, and thus presumably of the genes, are the same in all cells, save 
for the occurrence of somatic polyploidy. Although the process of dif- 
ferentiation is doubtless under genetic control, this cannot operate, as 
Weismann and others once suggested, by a parceling out of genetic "de- 
terminers" during development. The conclusion seems obvious that in 
these processes that part of the cell must be involved which is not 
identical everywhere in the body, namely, the cytoplasm. The origin of 
structural diversity in the midst of genetic identity is the chief problem 
that faces students of differentiation. 


Although growth and differentiation usually proceed together, they 
seem to be distinct processes, each more or less independent of the other. 
Growth may occur without differentiation by a simple multiplicative 
process, as in large parenchymatous masses such as the endosperm of a 
seed, in the tissue of an amorphous gall, or in tissue culture. In the early 
stages of many embryos, on the other hand, in the development of the 
female gametophyte in certain lower vascular plants, and in similar cases 
there is differentiation without growth. A notable example of this is fur- 
nished by the Acrasiaceae, a family of slime molds (p. 223). Here the en- 
tire vegetative growth occurs while the individuals are myxamoebae, and 
the elaborate differentiation of the colonial sorocarp does not begin 
until this vegetative phase is over. Animal embryology, particularly 
in the early cleavage stages from large eggs, provides many similar 

The independence of these two major developmental processes is 
further emphasized by the fact that conditions which favor one tend to 
be different from those which favor the other. In general, abundance of 
water and available nitrogen tend to induce growth, whereas abundance 
of accumulated carbohydrates, with less nitrogen and water, promotes 
differentiation (Loomis, 1932). Red rays of the spectrum tend to pro- 
mote growth and blue rays differentiation (p. 313). Under one photo- 
period a given species will produce nothing but vegetative growth 

Differentiation 183 

whereas another photoperiod will stimulate the differentiation of re- 
productive structures ( p. 315 ) . 

When the cycle of differentiation is complete, growth usually ceases. 
Thus, after the formation of reproductive organs in one of the higher 
plants has begun, growth of the plant as a whole is reduced and finally 
stops. When a fruit is fully differentiated, its growth in volume ceases, 
though dry weight may continue to increase for some time. The two 
processes of growth and differentiation may go on at different rates, and 
therefore their relative rates are important in determining differences in 
size. Where growth is relatively rapid, a large size will be attained before 
the completion of the cycle of differentiation stops; where it is slow, the 
cycle will be complete before much growth has occurred and the struc- 
ture will be much smaller. This is well illustrated by the analyses of 
inherited size differences in gourd fruits (p. 20). The balance between 
these two major processes in development— the addition of new material 
and its differential distribution— is of much significance. 

The process of differentiation and the problems it presents may be ex- 
amined from several different points of view. 

1. Differentiation may be studied in plant structure, for it is here that 
differences can most readily be seen. For purposes of convenience we 
may distinguish between external differentiation, which involves the out- 
ward structure and configuration of the plant, and internal differentiation, 
which involves the cells and tissue systems of which the plant body is 

2. Differentiation may be considered in its ontogenetic aspects. It is not 
a static process, evident in mature structures alone, but often changes its 
expression during development. Differentiation in a young plant is unlike 
that in a mature one, and these changes proceed in an orderly cycle 
of development both of the plant as a whole and of each of its com- 
ponents. Such differences are not in structure alone but in the reactivity 
and developmental potency of its parts. 

3. What course differentiation will take is determined not only by the 
genetic constitution of the plant but by the particular environment in 
which development takes place. External factors of many sorts affect the 
character of the structures which arise in the process of differentiation. 

4. The ultimate basis of differentiation must be in physiological 
changes in the living material itself. Most of these express themselves 
sooner or later in visible structural diversity, but there are many cases in 
which cells, structurally alike, can be shown to differ physically, chemi- 
cally, or in physiological activity. 

Examples of these four aspects of the process of differentiation, and the 
problems they involve, will be considered in the present chapter. 

184 The Phenomena of Morphogenesis 


External Differentiation 

One of the most obvious examples of differentiation is that which 
arises between the ends of a polar axis (p. 116). In all but the simplest 
axes the structures that are developed at the two ends are quite unlike. 
The most familiar instance is the differentiation of the shoot and root in 
higher plants. These two systems are set apart very early, almost at the 
beginning of embryonic development, and are fundamentally unlike in 
structure, function, and method of growth. Roots frequently develop 
from shoots but shoots less commonly from roots. Berger and Witkus 
(1954) have reported that in Xanthisma texanum the cells of the root 
always have four pairs of chromosomes but in those of the shoot, some 
plants have four pairs and some have five. The two types of plants are 
morphologically indistinguishable. How this difference in chromosome 
number arises in development is not known, but it is present in young 

Another conspicuous instance of differentiation in structure is that be- 
tween the vegetative and reproductive phases of the life cycle. In its 
early stages, the plant is becoming established. Its roots and leaves are 
formed or its vegetative thallus developed, and its career as a food-pro- 
ducing or food-acquiring organism is begun. Few plants, however, are 
permanently vegetative. When a certain stage is reached, growth no 
longer produces exclusively vegetative structures. Flower buds appear 
at the meristem, or in lower plants reproductive organs of various sorts 
begin to develop. These usually are formed as the result of internal 
metabolic changes in the plant, such as the accumulation of carbohy- 
drates or the production of specific substances. The onset of the reproduc- 
tive phase, however, is often closely related, as to time and extent, with 
certain environmental factors, notably light (p. 315). Short-day plants will 
flower only when the daily period of illumination is relatively short, and 
long-day ones only when it is longer. Some plants may never flower, and 
others may do so when they have hardly begun to develop. The balance 
between vegetation and reproduction may be tipped in various ways but 
the potency for reproduction is always present in the genetic constitution. 
This may not always be for sexual reproduction. In species which repro- 
duce chiefly by vegetative means flowers may be present but fail to 
function (as in the potato), may be much reduced (as in the banana), 
or may even be quite absent. Reproduction of some sort obviously is 
necessary, and the alternation of vegetative and reproductive phases, 
each essential in the life of the plant, is an important manifestation of 

Differentiation 185 

The difference between these two types of structure begins at the 
meristem and may often be recognized there by the presence of a large 
number of axillary buds which are destined to be flower buds. The de- 
velopment of the floral apex has been studied by many workers (p. 67). 
A single flower is a modified axis, and its parts arise from primordia 
which, although limited in number, are distributed in a precise pattern. 
The differentiation between them takes place early and produces sepals, 
petals, stamens, and carpels. The origin of these parts from particular 
layers of the meristem has been worked out, through the aid of chimeras, 
by Satina and Blakeslee ( p. 272 ) . In abnormal growth some floral organs 
may be so modified that they resemble others, as in the conversion, partial 
or complete, of petals to stamens or sepals to leaves (p. 277). 

In many trees the differentiation of flower buds begins very early, 
usually in the season before the flowers are borne. This is an important 
matter for horticulturalists and has been extensively studied (Zeller, 
1954, and others), since environmental conditions favoring flower pro- 
duction must be provided early. Whether a tree will flower (and 
fruit) well in a given season is often determined in June of the year 

Implicit in the process of reproduction is the differentiation between 
the sexes. In plants with perfect flowers this is evident only in the dif- 
ference between stamens and pistils. In monoecious plants, there are two 
kinds of flowers on the same plant and in dioecious ones these are on dif- 
ferent plants. The significance of such differences lies in the various 
mechanisms that tend to accomplish pollination, in many cases cross- 
pollination. A genetic basis has been found for some of these and is 
doubtless present in others. Environmental factors of various sorts are 
also operative here, notably nutrition and light. It is sometimes possible, 
for example, to change a staminate into a pistillate plant by altering the 
photoperiod (p. 317). The existence of sexual reproduction itself, in con- 
trast with the much less precarious method of vegetative reproduction, is 
based on the presumptive advantage of the higher variability that results 
from the recombination of genetic potencies following svngamy and 

Other traits are sometimes associated with the fundamental difference 
between the sexes, as observed in Mercurialis annua (Basarman, 1946), 
Valeriana dioica (Moewus, 1947), Urtica dioica, and Rumex acetosa 
(Umrath, 1953; Fig. 8-1). In general, the female plants are larger and 
are also different from the males in the size and shape of their leaves. 

A conspicuous example of external differentiation, since it involves an 
entire plant body, is that between the gametophyte and the sporophyte. 
In many of the simpler plants the two generations are very much alike, 
but they are markedly different in bryophytes and vascular plants. The 

186 The Phenomena of Morphogenesis 

advantage of this differentiation into sexual and nonsexual plants may 
lie in the possibility of the extensive multiplication of the products of a 
single sexual union. Among flowering plants, where mechanisms for 
effecting fertilization are more efficient than in many lower ones, the 
differentiation into two generations has almost disappeared. 

A gametophyte, coming from a spore produced by meiosis, typically 
has the haploid number of chromosomes, and the sporophyte has the 
diploid number. Many haploid plants are now known, however, which 
are undoubtedly sporophytes, and diploid gametophytes may readily 
be produced. Chromosome number is evidently not the cause of the 
difference between the two generations, but it is difficult to see why a 
haploid spore and a haploid egg (or diploids in each case) should pro- 

Fig. 8-1. Bryonia alba. Left, leaf from a shoot bearing male flowers; right, one from 
a shoot bearing female flowers. ( From Umrath. ) 

duce two structures as unlike as the prothallus and the sporophyte of 
a fern. The difference is probably attributable to the very different en- 
vironments in which these two cells develop. 

Origin of Differences. In most cases the origin of an organ or part is 
first evident as a group of meristematic cells which, by growing more 
rapidly in certain dimensions than in others, produces a definite form. 
For an analysis of such specific differentiation, however, it is necessary 
to determine how such a developing organ originates and the successive 
steps by which it becomes distinct from others. Sometimes this is 
relatively easy. In leptosporangiate ferns, for example, the sporangium 
can be shown to arise from a single cell of the epidermis. In eusporangiate 
forms, like some of the ferns and all higher plants, the sporangium 

Differentiation 187 

initial can be traced to a cell of the subepidermal layer. Analysis of 
differentiation in terms of cell lineage can often be carried further. 

The development of larger organs involves more than a single cell 
lineage. It may be studied in the differentiation of lateral organs in the 
apical regions of both root and shoot. From the root there grow only 
lateral roots, which arise in the pericycle and push out through the cortex. 
The shoot meristem, however, is more complex ( p. 89 ) . At the base of the 
terminal dome of cells arises a series of minute protuberances, the early 
leaf primordia, arranged in a precise order. 

The cause of the differentiation of these primordia from the rest of the 
apical meristem is not known. Schiiepp ( 1952 ) suggested that, since cell 
division in the outer layer of the meristem is always anticlinal but in the 
tissue below may be in various directions, this surface layer will expand 
more rapidly than the surface of the underlying tissues and will thus 
tend to buckle or pucker, starting the formation of primordia. This would 
not explain the very regular pattern in which these arise, however, and 
it can also be shown that the initial bulge results from division in a group 
of cells just beneath the surface layer. Snow and Snow (1947) have sub- 
mitted this theory to experimental test by making shallow incisions at 
the surface of the meristem. Instead of closing up, as they would do if 
the outer layers were under pressure, these gaps open, indicating that this 
region is actually under tension. 

The fate of a small lateral primordium may not always be to grow into 
a leaf. Wardlaw and his students (p. 71) have performed various experi- 
ments on the meristems of ferns in which, by deep cuts, they were able to 
isolate from the apex a young primordium or a region that was about to 
develop into a primordium. In most cases this structure, instead of form- 
ing a dorsiventral leaf, developed into a radially symmetrical bud-like or- 
gan which, in culture, was capable of growing into a whole plant. Factors 
in the surrounding meristematic tissue evidently help determine into what 
sort of structure a given primordium will differentiate. 

The growth of the leaf primordium into a mature leaf has been studied 
by many workers (p. 90, and Foster, 1936). In general, the upper and 
lower epidermis is continuous with the outer layer of the meristem, and 
what will later form the palisade and spongy layers is continuous with 
the subepidermal layer. The veins usually arise from a layer just below 
this. The differentiation of the leaf of tobacco has been described by 
Avery (1933) and of Linum by Girolami (1954; Fig. 8-2). Foster (1952) 
has reviewed the development of foliar venation. The mode of develop- 
ment and differentiation in certain leaves of unusual shape, as in 
Podophyllum and Sarracenia, is described by Roth ( 1957 ) . The growth 
of a fern frond, at least for some time, takes place by the activity of an 
apical cell (Steeves and Briggs, and Briggs and Steeves, 1958). 

188 The Phenomena of Morphogenesis 

An important morphogenetic problem here is how far the development 
of such a lateral structure depends on factors in the meristem from 
which it grew and how far it is independent. Steeves and Sussex ( 1957 ) 
removed primordia of several sizes and ages from the meristem of 
Osmunda and other ferns and grew them in sterile culture. These de- 
veloped normally into mature leaves just as they would have done if at- 
tached to the plant, except for being smaller. Evidently after a certain 
stage is reached the control of development is within the leaf itself. This 

Fig. 8-2. Early stages in development of the leaf of Linum. SI, subapical initial; x, 
procambial elements. ( From Girolami. ) 

self-differentiation has been found in other cases and shows that the 
organization of the plant is not as tight as it is in animals but that there 
is some degree of independent control of differentiation in individual 

A branch or secondary axis differentiates by the activity of a bud aris- 
ing in the axil of a leaf primordium (Garrison, 1955). In herbaceous 
plants such a bud develops directly but in woody ones it has a dormant 
period and is covered by bud scales, or cataphylls. Bud development in 
pine has been studied by Sacher (1955) and in angiosperms by Foster 

Differentiation 189 

(1931). In inflorescences the leaves may be reduced to bracts. Foliage 
leaves, bracts, and cataphylls are presumably equivalent morphologically, 
and the development of a primordium into one or the other depends on 
its function. Primordia at a meristem are thus multipotent (Foster), 
since they may form several kinds of structures. 

How a given structure differentiates is closely related to the position 
that it occupies in the developmental pattern. Not all morphologically 
equivalent organs develop alike. In Ginkgo the axis is differentiated 
into short shoots and long shoots (Gunckel and Wetmore, 1949). All 
buds form short shoots but some of these will grow into long ones. The 
ratio of the two affects the form of the tree. A somewhat similar situation 
occurs in Cercidiphyllum (Titman and Wetmore, 1955). 

The vertical axis of a tree and the lateral axes ( branches ) that it bears 
also may differ markedly, the former being radially symmetrical and 
orthotropic, the latter more or less dorsiventral and plagiotropic. Conifer- 
ous trees offer familiar examples, where the lateral branches are much 
flattened and branches of the second order occur in two lateral ranks. In 
Abies and Picea this is evidently related to gravity, for if the terminal 
bud or branch is removed, a lateral branch will bend upward and replace 
it. In Araucaria excelsa, however, this difference is so deeply seated that 
if cuttings are made from the lateral branches, the flattened character 
now persists in the new plant even if the cutting is oriented vertically. 
Not only structure but physiological behavior may be permanently 
altered, for such lateral branches will grow horizontally in whatever posi- 
tion they may now be placed. Carvalho, Krug, and Mendes (1950) 
report a similar behavior in Coffea. There are many other examples of 
such topophysis (p. 212). 

The differentiation of particular organs— root, stem, leaf, flower— 
during development is markedly influenced by growth substances of 
various sorts (Chap. 18). 

Internal Differentiation 

Visible differentiation involving external diversity in organs and their 
parts is accompanied in most plant structures by a high degree of internal 
diversity. This involves differentiation among various types of cells and 

Histological differentiation presents two chief problems : ( 1 ) How do 
cells become different from one another and (2) what is the origin 
of the various tissue patterns found in the internal structure of plant 

Origin of Cellular Differences. Cells differ from each other in many 
ways. Frequently it is possible to determine the exact cell division at 
which such a difference becomes evident. In young roots of certain grasses 

190 The Phenomena of Morphogenesis 

(p. 131), for example, the last division of many (sometimes of all) of 
the surface cells results in a small daughter cell at the apical end and a 
larger one at the basal end. This initial difference is intensified during 
the later development of these cells, for the smaller cell (a trichoblast) 
sends out from its surface an elongate sac which becomes a root hair. 
Such a structure is lacking in the larger cell (Cormack, 1949). The 
beginning of this difference may be seen even before the last division, 
for the cytoplasm at the apical end of the mother cell becomes much 
more dense than that at the basal end. Differentiation between the two 
daughter cells is thus related to the strongly polar character of the mother 

Fig. 8-3. Section through developing liverwort sporangium showing differentiation of 
alternating spores and elater cells. ( From Goebel. ) 

cell. In Phalaris the epidermal cells contain a natural red pigment which 
is deeper in color in the prospective root-hair cells, and they can be 
distinguished early for this reason (Bloch, 1943b). In many plants the 
surface cells of the root are all potentially alike, and the differentiation 
of some cells into root hairs and others into hairless cells is not de- 
termined at a differential cell division but by environmental factors. The 
difference between these two types of root-hair determination may be 
related to anatomical characters (Cormack, 1947). A close relation exists 
between the distribution of cellulose-forming enzymes and the location 
of the root hair on a surface cell ( Boysen-Jensen, 1950 ) . 

There are many somewhat similar cases of cellular differentiation. In 

Differentiation 191 

Ricinus the secretory cells are formed, in the young meristem or in later 
development, by differential cell division (Bloch, 1948), though conform- 
ing to no sharp pattern. Once formed, these cells continue to divide, pro- 
ducing rows of similar cells in a cell lineage. This may even persist in 
tissue culture. 

The differentiation between chlorophyll-bearing and colorless cells in 
the leaf of Sphagnum results from a differential division, preceded by a 
polar movement of the cytoplasm (Zepf, 1952). The origin of elaters in 

Fig. 8-4. Trochodendron. Section of leaf with a large branching sclereid. (From 
Foster. ) 

the capsule of liverworts is similar, a cell of the archesporial tissue divid- 
ing into a spore mother cell and an elater cell ( Fig. 8-3 ) . 

Certain trichosclereids develop in much the same manner. These cells, 
which in Monstera become very long and thick-walled, are set apart at 
the last division of certain cells of the meristematic cortex. The smaller 
daughter cell (this time at the basal end), possessing a relatively larger 
nucleus, develops into the sclereid and the other into a typical parenchyma 
cell. Although the sclereid begins in this case as the smaller daughter cell, 
it soon sends out one or more processes which grow longitudinally be- 

192 The Phenomena of Morphogenesis 

tween neighboring cells and often become very long (Bloch, 1946). A 
study of the origin of such idioblasts ( cells distinctly different from their 
neighbors) (Fig. 8-4) may throw light on problems of cellular dif- 
ferentiation ( Foster, 1956, and others ) . 

Stomatal initials are set apart by differential cell divisions ( Bunning and 
Biegert, 1953 ) . A smaller, more densely cytoplasmic cell and a larger one 
are formed by a late division of a surface cell. The former divides again, 
this time equally and longitudinally, to form the guard cells, the contents 
of which soon become markedly different from those of the other epidermal 
cells. In monocotyledonous plants like this, where the cells are in regular 
longitudinal rows, the stomatal initials are cut off at the ends of elongate 
cells. In most dicotyledons, where the cells of the developing epidermis 
are more nearly isodiametric, the initial is cut out of a corner of the cell 
and divides again to form the guard cells. 

A number of cases have been reported where the differentiation of one 
type of cell evidently induces changes in the character of adjacent ones. 
Thus in Sedum there are groups of cells that form tannin, and in these 
regions stomata do not differentiate (Sagromsky, 1949). In Potamogeton 
roots those cells of the exodermis that are just under the already dif- 
ferentiated trichocytes divide several times, unlike the other cells in this 
layer, and so form groups of small cells, one below each trichocyte 
( Tschermak-Woess and Hasitschka, 1953b ) . In a species of Begonia where 
there are silver spots on the leaf surface, there is a hair formed in each 
spot save the very small ones, and the larger the spot, the longer the hair 
(Neel, 1940). 

Cell Size (p. 29). Some of the most conspicuous differences between 
cells are in their size. Meristematic cells in most cases are small, and after 
the final division the daughter cells increase considerably in size. The 
extent of this increase is determined by the time in development when 
division ceases in that particular cell lineage and by the position of the 
cell in the general histological pattern. Pith cells are usually large be- 
cause they have had a long period of enlargement since their last division, 
and epidermal cells relatively small since division there lasted longest. 
Many size differences, however, such as those between the large vessels 
of ring-porous woods and the small elements around them, are due to 
local differential factors, since the cambial initials are alike. 

One type of cellular differentiation is unlike all others in that it in- 
volves fundamental alteration of the cell itself. Many instances are now 
known in which mature cells, aroused to division by various agents, are 
found to have twice as many chromosomes (or more) as they did at 
their last preceding mitosis. During the differentiation of such a cell 
from meristematic condition to maturity a doubling of the chromosome 
complement must have taken place. Such a change, though not directly 

Differentiation 193 

observable in the cell, is usually reflected in increased cell size and may 
be a factor in other changes which occur during differentiation. How 
important this factor is in cellular differentiation is not known. It may 
account for some of the diversity in cell size but probably has little to do 
with other aspects of differentiation ( p. 441 ) . 

The Cell Wall. Some of the most distinctive ways in which cells differ 
are concerned with the cell wall. The wall is of much greater variety and 
significance in plant cells than in animal cells, and the key to cellular 
differentiation in plants is often to be found in it. Walls may differ greatly 
in thickness, chemical composition, and structure, depending upon the 
function of the cells of which they are parts. In certain tissues the cells 
die early, and only the thickened walls which they formed remain. The 
size and shape of the cell and the manner of its growth seem often to 
be dependent primarily upon the character of the wall. Studies of the 
chemistry and fine structure of the wall show how complex its consti- 
tution may be and make clear that any detailed analysis of cellular dif- 
ferentiation must pay attention to changes not only in the living material 
of the cell— the true protoplast— but in the wall that is the result of its 
activity (Bailey and Kerr, 1935; Frey-Wyssling, 1955). 

In a few cases the origin of differentiation in the wall may be observed, 
especially where sculpturing occurs, as in the ringed, spiral, and retic- 
ulate cells of the xylem and in other tissues with similarly unequal 
wall thickening. Criiger (1855) and Dippel (1867) many years ago 
showed that the first indication of where such thickenings were to occur 
in developing cells was the accumulation of cytoplasm, more densely 
granular than the rest, in a definite pattern. The thickenings of the 
wall (rings, reticulations, or others) were laid down in close relation to 
this cytoplasmic pattern (Barkley, 1927). Strasburger (1882) observed 
streaming of cytoplasm along these strands. Large and vacuolate paren- 
chyma cells that are being redifferentiated as reticulate xylem cells in re- 
generation are particularly good material in which to observe the 
cytoplasmic network upon which the wall reticulum is being built ( Sin- 
nott and Bloch, 1945; Fig. 8-5). Kiister ( 1931) called attention to the simi- 
larity between such cytoplasmic configurations and Liesegang rings. 
Denham (p. 166) reports that in many cells the directions of cytoplasmic 
streaming has a definite relation to the micellar configuration of the wall. 

Differentiation and Position. In all the cases of cellular differentiation 
here described, the position of the cell in the developing system is evi- 
dently closely concerned with the type of differentiation that it under- 
goes. A notable example is the formation of reaction wood, which dif- 
ferentiates in the precise position where it will tend to bring a terminal 
or lateral axis into a specific orientation in the pattern of the whole 
(p. 356). 

194 The Phenomena of Morphogenesis 

We may distinguish position with reference to the external environ- 
ment (light, oxygen, chemical stimuli, and many others), with refer- 
ence to various factors in the internal environment (surfaces, air spaces, 
conducting strands, and cells previously differentiated), and with refer- 
ence to the autogenously unfolding and genetically controlled pattern of 
development. To distinguish these aspects of position is often diffi- 
cult or impossible. The important fact is that in the organized system 
specific parts are markedly unlike each other and that these differences, 
which in the aggregate distinguish the system, may arise from various 

Fig. 8-5. Portion of a regenerating xylem strand in Coleus, showing pattern of wall 
thickenings laid down along bands of granular cytoplasm. Earlier stage at right. 
( From Sinnott and Bloch. ) 

Intracellular Differentiation. The parts of a single cell often show a 
high degree of diversity. The distinction between nucleus and cytoplasm 
is present even in embryonic cells, but as the differentiation of the 
cell takes place, the protoplast may exhibit a wide range of structures. 
Conspicuous among these are the plastids. In the algae (as Spirogyra), 
these may be represented by large and often complex chromatophores. 
Much more minute bodies, the mitochondria, occur and multiply (Soro- 
kin, 1955; Hackett, 1955). Bodies similar to the Golgi apparatus of 
animal cells have been reported (Weier, 1932), but their general oc- 
currence is doubtful. 

Differentiation 195 

The proportions of these intracellular structures often change during 
differentiation. Thus in moss protonemata the nucleus increases in size 
from the tip of the caulonema backward but the nucleolus decreases for 
the first few cells. The nucleus also gradually changes from a spherical to 
a spindle shape, and other changes are evident (Bopp, 1955). Intra- 
cellular diversity is particularly conspicuous in the differentiation of large 
coenocvtic bodies, as in certain algae. 

Differentiation of Histological Patterns. Differentiated cells rarely occur 
separately but are grouped into tissues. Endodermis, vascular tissues, 
and many others are familiar examples. These tissue patterns begin 
to appear in the embryo (Miller and Wetmore, 1945; Fig. 8-6; Spurr, 
1949; Esau, 1954), grow more diverse in the seedling, and reach their 
maximum differentiation in the mature plant. Nowhere else are the com- 
plexities of differentiation so evident as in the development of these 


Fig. 8-6. Beginnings of differentiation in early embryo of Phlox. Successive stages, 
showing origin of central procambial core. (From Miller and Wetmore.) 

histological patterns. Specialized as these may be, each constitutes an 
element in an integrated whole. 

Wall Relationships. One of the basic elements in histological pattern 
is the relationship of cells to each other. This is determined primarily by 
the position of new walls in dividing cells relative to walls in adjacent 
cells. In most tissues a new wall is so placed that it does not come op- 
posite a neighboring cell partition, and the cells are thus "staggered" 
in position, like bricks in a building (p. 47). In a few cases, however, 
walls in adjacent cells are exactly opposite, so that they extend in con- 
tinuous lines across the tissue (Fig. 8-7). This is particularly evident in 
cork and in regenerating tissue at wounds. This arrangement lends itself 
well to the development of aerenchyma, since when such cells pull 
apart at their corners a larger volume of intercellular space results than if 
the walls were staggered. The two types of pattern may well be seen 


The Phenomena of Morphogenesis 

Fig. 8-7. Partition walls opposite adjacent ones. At left, dividing cells in tissue of 
wounded petiole of Bryophyllum, the walls being laid down directly opposite those 
in adjacent cells. At right, similar divisions in more mature tissue below wound 
surface. ( From Sinnott and Block. ) 

in the transverse section of certain roots, where the inner cortex is radially 
concentric, with opposite walls, but the outer cortex shows the alter- 
nating arrangement typical of ordinary parenchyma ( Fig. 8-8 ) . 

Endodermis. One of the simplest of these tissue patterns is shown by 
the endodermis. This is a single layer of cells differentiated in a specific 
way, as by special thickenings in the walls or the presence of a Casparian 
strip. It separates the vascular cylinder from the cortex. The position 
it occupies is usually a very definite one, and in such plants as Equisetum 
its particular pattern with reference to the bundles is specific enough to 
be valuable for taxonomic purposes. 

Fig. 8-8. Transverse sections of two roots. Left, Sporobolus, in which there are three 
layers of radially concentric cells in the cortex, the walls having been laid down 
directly opposite those in adjacent cells. Air spaces later appear at intersections. Right, 
Agrostis, in which the cell walls in the cortex always avoid adjacent ones. (From 
Sinnott and Bloch. ) 



The exact localization of the endodermis makes it of particular interest 
morphogenetically. Light is important in its development, for it is well 
differentiated in roots and etiolated stems and much more poorly de- 
veloped in the light (Bond, 1935). Venning (1953), however, finds that 
factors other than light are responsible for the formation of a typical 
endodermis. Van Fleet in a series of papers has studied differentiation 
histochemically ( 1954a and b ) , with particular reference to the position 
of the endodermis on an oxidation-reduction gradient as well as to the 
distribution of various enzymes. He has stressed the importance of histo- 
chemical determination of enzyme distribution (1952) as a means of dis- 
covering chemical differentiation before it is evident in structure (Fig. 

Fig. 8-9. Chemical differentiation of the endodermis. Its cells stain differently from 
those of adjacent tissues. ( From Van Fleet. ) 

Fiber Patterns. An example of the differentiation of a somewhat more 
complex pattern but one consisting of a single type of cells is provided 
by the development of a system of fiber strands such as that found in the 
pericarp of the cucurbit fruit, and especially well developed in the 
"dishcloth" gourd, Luff a (Sinnott and Bloch, 1943). Here the pericarp 
tissue in the early ovary primordium consists of longitudinal rows of 
squarish parenchyma cells with most of the divisions at right angles to 
the axis of the young ovary or parallel to it. Here and there begin to 
occur divisions not in these two orientations but obliquely at various 
angles (Fig. 8-10). Parallel to each such division is a series of others so 
that in a given cell or its neighbors several elongate and parallel cells 
are cut out. This group becomes connected with other groups in a con- 
tinuous series, though successive members of this series may arise at 
somewhat different angles. The result is that strands of cells are formed, 
twisting about through the original rectangular cellular system and 

198 The Phenomena of Morphogenesis 

connected in an interwoven pattern. These small elongate cells expand 
with the growth of their parenchymatous neighbors and develop into 
long sclerenchymatous cells aggregated into strands a fraction of a milli- 
meter wide which are organized into the complex fibrous "sponge." This 
sponge is not a random mass of fibers but has an organization of its own, 
for the outer members of it are arranged in rows transverse to the axis 
of the fruit and most of the inner ones extend lengthwise. They are 
united into a continuous system. This system seems to be the expression 
of a histological pattern superposed upon the fundamentally different 
system of regularly arranged parenchyma cells of the early ovary. How 
the course of its interconnected but continuous strands is established is 
a baffling problem. In somewhat the same way as these sclerenchym- 
atous strands develop, the young bundle initials of the veins arise in 

Fig. 8-10. Young ovary of Luff a. Successive early stages in the origin of a fiber strand 
differentiating in ground parenchyma. ( From Sinnott and Bloch. ) 

the mesophyll of a developing leaf blade, as described by Meeuse ( 1938; 
Fig. 8-11). 

Cambium. A familiar example of a complex pattern of differential de- 
velopment is that of the vascular cambium and its products (p. 84). The 
typical cambium consists of a continuous tangential layer of elongate 
initials in which most of the divisions are in the tangential plane. The 
cells cut off on the inside develop into tracheids, fibers, vessels, paren- 
chyma, and ray cells of the xylem, and those on the outside into sieve 
tubes, companion cells, fibers, and other phloem cells. There are pro- 
found differences between a huge vessel element in oak wood and a 
small parenchyma cell beside it but both come from similar cambium 

There are a number of morphogenetic problems presented by a study 
of this development of secondary vascular tissues. 

Differentiation 199 

1. What determines the relative frequency of divisions on the inside of 
the cambial initials to those on the outside, the relative amount of 
xylem and phloem? 

2. What determines how cambium derivatives differentiate into the 
widely diverse sorts of cells found in the mature tissues? 

3. What maintains so perfectly the anatomical pattern of the xylem 
and phloem? 

Bannan and others (p. 81 ) have shown that many radial files of cells are 
begun at the cambium and then die out and that many rays are initiated 
only to disappear, the net result being a very precise distribution of rays 
and vertical elements with reference to each other. The files of tracheids 
remain at a constant width, and the rays are evenly spaced with reference 
to each other, as can be seen in a tangential section of wood. These rela- 

Fig. 8-11. Portion of transverse section of leaf of Sanseviera. Bundle of fibers be- 
ginning to differentiate in the midst of fundamental tissue. Compare with Fig. 8-10. 
( From Meeuse. ) 

tionships are so constant and specific that they are used as taxonomic 

This same problem of a specifically patterned distribution of structures 
meets us in many other places, such as in the spacing of bundles in cross 
sections of the stems of monocotyledons, of stomata in the leaf epidermis, 
of root hairs, or of developing sclereids in the cortex. Biinning has ex- 
plored this problem (1948, pp. 173-179). He suggests that a specific 
developing structure prevents the differentiation of another like it within 
a certain distance of itself and cites some experimental evidence in sup- 
port of this idea (Biinning and Sagromsky, 1948). In the cells immediately 
around a young stoma initial, the nuclei always lie on the side of the 
cell next to the initial, as if in response to a chemotactic stimulus (Fig. 
8-12). A few cells farther out, they have normal positions. Biinning be- 
lieves that a hormonal substance passes out from the young stomatal cells 
which stimulates cell division, as shown by the production of accessory 

200 The Phenomena of Morphogenesis 

cells and others, and thus inhibits differentiation of stomata (and some- 
times of hairs or glands). Near wounds, cell division occurs but stomata 
are not differentiated. If auxin paste or juice from crushed tissue, pre- 
sumably containing wound hormones (p. 402), is applied to the young 

Fig. 8-12. Relation of nuclei to stomata. Nuclei of cells adjacent to stomatal initials, 
in young and developing leaves, are pressed closely to these initials as if chemically 
attracted to them. ( From Biinning and Sagromsky. ) 

leaf, cell divisions are plentiful but stomata do not develop (Fig. 8-13). 
This suggestion is of much interest in relation to the differentiation of 
other evenly spaced structures, but it does not explain how the inhibiting 
center itself is initiated in the first place. This is a promising point, how- 
ever, at which to attack the problem of organic pattern. 

In the histological pattern that originates back of the apical meristem, 
Biinning ( 1952# ) believes that the meristem itself inhibits differentiation 
within a certain distance. Farther back, each bundle initial, which is in the 

Fig. 8-13. Left, epidermis of developing leaf two weeks after treatment with auxin 
paste. Stomata are almost absent. Right, epidermis of untreated half of the same 
leaf. ( From Biinning and Sagromsky. ) 

Differentiation 201 

process of growth but is not part of the meristem proper and which he 
terms a meristemoid, inhibits the development of others near it. 

The differentiation of tissues produced by the vascular cambium has 
been studied by Linnemann (1953), who observed that in beech the 
proportion of rays is greater in the wood of isolated trees than in dense 
stands but that it does not vary consistently with age or as between 
trunk and branches. Rays tend to increase in width during an annual 
ring and to be wider in the narrower rings. 

The fact that vessel elements occur in longitudinal series to form a 
duct indicates the operation of a continuous stimulus longitudinally 
along the axis. Priestley, Scott, and Malins (1935) have shown that a 
single duct differentiates almost simultaneously throughout a long extent 
of trunk. A leaf trace passing down into a young stem exerts a consider- 
able correlative influence upon vessel differentiation below it. Alexandrov 
and Abessadze (1934) found that there are fewer vessels in a segment 
just below a leaf trace and that they appear earliest next the rays that de- 
limit the trace. The vessel-forming stimulus clearly moves downward, 
thus suggesting the operation of a hormonal control. 

That auxin has a role in the initiation of the ring-porous condition is 
suggested by the work of Wareing (1951), and Chowdhury (1953) has 
analyzed some of the factors responsible for the transformation of 
diffuse-porous to ring-porous structure in Gmelina. 

Continuity in differentiation of similar cells is also shown by the cork 
cambium, or phellogen (p. 88), in old cortex or phloem. Here it arises 
as a series of almost simultaneous divisions which, as seen in transverse 
section, somewhat resemble those described for the Luffa strands, since 
they are connected to one another in a series and often follow a somewhat 
irregular course through the tissue in which they arise. They form a 
continuous sheet of meristematic cells, often in localized patches, which 
cut off elements from their outer faces. Their cells suberize later, thus 
sealing off the outside tissues. The origin of these phellogens is the more 
interesting because they have their beginning in tissues where the cells 
are mature and intermixed with dead or necrotic ones. Their origin 
after wounding is related to the operation of wound hormones. 

A notable example of the differentiation of a histological pattern is 
furnished by the system of lignified thickenings (the "reseau de soutien" 
of van Tieghem, 1888) in the air roots of certain orchids, which is pre- 
sumably concerned with providing rigidity for tissue otherwise soft and 
easily collapsed. These arise as bands of lignified wall thickenings which 
surround individual cells. They may occasionally fork. It is noteworthy 
that the band in a given cell is directly contiguous to that in an adjacent 
cell, so that a continuous patterned network of thickened strands is 
established (Fig. 8-14). This reminds one of the way in which the ringed 

202 The Phenomena of Morphogenesis 

thickenings in protoxylem and regenerating xylem cells are directly op- 
posite those in adjacent ones. 

In all these examples of the differentiation of histological pattern it 
is evident that the pattern as a whole transcends cellular boundaries and 
involves an extensive and correlated series of changes. This poses in most 
direct fashion the problem of pattern in general. 

More Complex Patterns. In most cases an anatomical pattern consists 
of more than one type of tissue and thus is much more complex than 
the ones just described. In plants that grow by a large apical cell at 
the meristem, differentiation of the various tissues from particular cells 
cut off from this apical cell may be traced. Thus in Selaginella (p. 58), 


N C s 


Fig. 8-14. Continuity in the differentiation of various wall thickenings. Left, rings 
and spirals in protoxylem of Zea. Center, thickenings in reticulate vessel elements 
that have developed from parenchyma cells in bundle regeneration. Right, lignified 
bands (reseau de soutien) in cortex of air root of an orchid. In all these cases the 
thickenings form a continuous pattern across cell boundaries. {From Sinnott and 
Bloch. ) 

the apical cell, by an unequal division parallel to one of its faces, pro- 
duces on its apical side a large cell, the continuing apical initial, and 
on its basal one a cell which, seen in section, has parallel anticlinal walls 
(Barclay, 1931; Fig. 4-4). This cell divides into two by a wall at right 
angles to the first division. Each daughter cell divides again into two 
in the same way. Thus a row of four cells is produced. Proceeding down 
the shoot axis, in a longitudinal section, one can observe the fate of simi- 
lar rows of four cells which had been cut off by previous divisions of the 
apical cell. The outermost of the four becomes a cell of the epidermis. 
The second (by later divisions) produces the cortex. Descendants of the 
third and fourth form the innermost tissues. Thus the progress of dif- 
ferentiation can usually be followed in various lineages of cells. Specific 

Differentiation 203 

types of differentiation seem to be related to specific lineages, almost as 
though "determiners" were being parceled out at each division. The same 
type of differentiation has been described by Bartoo (1930) in Schizaea. 
In these cases differentiation is a true development, the unfolding of an 
internally directed pattern, with each division evidently related to the 
polarity of the cell. Cell division here seems a dominant factor in the 
determination of pattern. It should be noted, however, that there are 
often irregularities in this progression and that it is by no means always 
so precise. 

In many other plants, especially those with large apical cells, differen- 
tiation also follows a rather regular course like that just described. Chara, 
Fontinalis, and some species of Equisetum (Fig. 4-3) are examples. In 
others, however, such precise relationship between a specific type of dif- 
ferentiation and cell lineage does not occur, for a particular tissue may 
sometimes have one cell ancestry and sometimes another. The origin of 
root, stem, leaf, and foot from the quadrants of a young fern embryo, for 
example, is not rigidly determined. 

In higher vascular plants where an apical cell has been replaced by a 
mass of meristematic tissue, in most cases it becomes impossible to trace 
the origin of a group of differentiated cells from a single ancestor or to 
determine the precise divisions at which a fundamental difference be- 
tween two cells (or their descendants) originates. Such divisions may 
occur, but in these there is no great difference between daughter cells 
nor is regularity of lineage usually observable. It is probably true that in 
very many instances differentiation is the result of factors of environment 
or position and is not related at all to differential cell division. 

There is, however, a good deal of cellular differentiation to be seen in 
the apex of shoot and root, either as layers or zones (p. 62 ) and these often 
bear a close relation to the structures that develop from them. Thus in 
periclinal chimeras ( p. 272 ) it is possible to determine with much accuracy 
the derivation of particular tissues from particular layers at the apex. 
Nevertheless, in forms without meristematic layering ( as in some gymno- 
sperms) differentiation of tissues takes place equally well. The problem 
of how the histogens become distinct from each other, in forms which 
show them, is one which for its solution must go back to the young 

In the mass of relatively undifferentiated tissue below the apex arise 
the beginnings of vascular tissue. The distinction between the procambial 
or provascular cells (those which are to give rise to the primary vascular 
tissue) and the cells of the fundamental tissue begins to make its appear- 
ance early in development near the tip of the meristem. The first differ- 
ence to be observed here often is not a structural one but a difference in 
the staining reaction of the cells. The earliest structural difference to be 

204 The Phenomena of Morphogenesis 

seen, in most cases, is the elongate form of the procambial cells in longi- 
tudinal section. This form is due either to fewer transverse divisions in 
these cells as compared with their neighbors or to more frequent longi- 
tudinal ones. From groups of these elongated provascular cells arise the 
vascular bundles of the stem. There is a close relation between the differ- 
entiation of these bundles and of the leaf primordia near the apex, for the 
young leaf traces that enter the base of each primordium are continuous 
with the differentiated vascular tissue below. 

There has been some difference of opinion as to just how the pattern 
of vascular differentiation originates. It is now rather generally agreed 
(Esau, 1953Z? ) that the procambial strands develop acropetally, continu- 
ous with the mature vascular tissue below and pushing up into the bases 
of the primordia themselves. In a transverse section of the axis, the pro- 
cambium forms a ring which may be continuous or consist of a series of 
bundles. On the outside of a procambial strand the first phloem differen- 
tiates, and on the inside, the first xylem. The developmental history of 
these tissues is different, however. The phloem, like the procambium, 
develops continuously from the base toward the tip. The xylem, on the 
contrary, differentiates first in the base of the enlarging leaf primordium 
and then both upward and downward. In its downward course it meets 
the upward developing xylem in the axis below (Miller and Wetmore, 
1946). Jacobs and Morrow (1957) traced the downward differentiating 
xylem strands and found that they did not always make connection with 
the normally opposite ones below. In the root, the procambium, phloem, 
and xylem all differentiate acropetally and continuously ( Heimsch, 1951; 
Popham, 1955k). 

The physiological significance of these facts is not clear, but morpho- 
genetically they are concerned with the important question as to whether 
the course of initiation and development of structures at the apical meri- 
stem, notably the position of the leaf primordia and the pattern of internal 
differentiation, results from stimuli proceeding up from the mature struc- 
tures below or whether in its development the tip is independent of what 
has gone before. This problem is discussed elsewhere in the light of some 
experimental results (p. 238). Torrey (1955), working with root tips cut 
off and grown in culture, presents evidence that the pattern of vascular 
differentiation (triarch, diarch, or monarch) just back of the tip is not 
induced by the tissue farther back but is related to the dimensions of the 
apical meristem at the time the cylinder is differentiated. 

Much work has been done on this problem of differentiation at the 
apical meristems, and it is well covered by Esau ( 1953b ) . Among other 
recent publications are an extensive review by Esau ( 1954 ) and papers 
by Rathfelder (1954), Young (1954), Wetmore and Sorokin (1955), 
McGahan (1955), and Jacobs and Morrow (1957). 

Differentiation 205 


Differentiation, like most problems of morphogenesis, must be studied 
not only as it is found in the structure of the mature plant but as it arises 
during development. The mature plant is obviously very different from 
the embryo and the seedling, but an important question, still far from 
settlement, is whether the changes that take place here are simply the 
result of increased size and the effects of environment or are internal 
modifications arising during development and becoming manifest in the 
progressive differentiation of the individual as its life cycle unfolds. 

It is obvious that environment is of great importance in determining 
the differences that arise, and most of the experimental work in morpho- 
genesis is concerned with a manipulation of environmental factors. It is 
also clear that the specific response to an environment depends on the 
innate genetic constitution of the individual. What is not so evident, how- 
ever, is whether this response always remains the same or changes as the 
organism grows older. 

There is a good deal of evidence that changes in the plant, independent 
of environmental conditions, do indeed occur as development proceeds. 
Juvenile stages are often very different from adult ones. That these are 
real and often irreversible differences is proved by the fact that they can 
be perpetuated by cuttings. Progressive changes in the shape and char- 
acter of organs, especially leaves, at successive points along the stem have 
often been observed and by some biologists are attributed to advancing 
maturity or physiological aging. The onset of actual senescence has been 
reported in some cases. A considerable school of physiologists believe that 
the life history of a plant, particularly up to the time of flowering, consists 
of a series of successive phases, each the necessary precursor of the next 
but independent of the amount of growth attained. This concept has come 
in part from the idea of vernalization ( p. 339 ) . 

In such phasic development the major change is the onset of the repro- 
ductive period after one of purely vegetative development. This appar- 
ently begins by a physiological change, the "ripeness to flower," as Klebs 
called it. Only after this has begun do the floral primordia appear at the 
meristem. They may not be the first visible evidence of the onset of repro- 
duction. Roberts and Struckmeyer ( 1948 and other papers ) have shown 
that the induction of the flowering phase is very early indicated by a 
number of anatomical changes. Root growth is much reduced, cambial 
activity almost ceases, and the vascular tissues tend rapidly to complete 
their full differentiation. In other words, the plant structures become ma- 
ture. Reproduction is a sign of maturity, and these anatomical changes are 
evidence of a more general one that is about to take place. Many factors 

206 The Phenomena of Morphogenesis 

are involved here, either as causes or concomitants. There seems to be a 
major physiological change involved in this shift from vegetation to repro- 
duction. An important problem in both physiology and morphogenesis is 
to find what is involved in this shift. To solve it would throw light on one 
of the major formative processes in the plant. 

Embryology and Juvenile Stages. The science of embryology in the 
higher plants, in the sense in which it has been developed in animals, can 
hardly be said to exist. The early embryo is relatively inaccessible and 
is simple in structure. The divisions immediately following fertilization 
have been studied for many plants by Soueges (1939) and Johansen 
( 1950; Fig. 8-15) and show differences in certain groups, but little as pre- 
cise as the early stages in animal embryology is to be seen. Toward the 
micropylar end of the ovule the young radicle begins to differentiate and 
forms an apical meristem at its tip. At the other pole, in gymnosperms and 
dicotyledons, arise the cotyledons, with the first bud between them. The 
monocotyledons have a somewhat more complex structure here but it 
follows the same general course. 

Especially important to students of morphogenesis is Wardlaw's book 
on Embryogenesis (1955a), which discusses embryogeny throughout the 
plant kingdom, with particular emphasis on the factors that determine 
development. Maheshwari (1950) has written a general survey of angio- 
sperm embryology, including a useful discussion of experimental embryo 
culture. The ability to take embryos out of the ovule at a very early stage 
and grow them in culture has opened up a wide field of investigation 
which should be fruitful for morphogenesis. 

Several facts of significance, particularly for regeneration, have come 
from a study of plant embryos. In a number of cases the young embryo 
may spontaneously divide into several parts each of which apparently has 
the capacity to develop into a whole plant. Such cleavage polyembryony 
has been studied by Buchholz and others (p. 235). In certain plants, 
notably some members of the citrus family, embryos may arise not only 
through a sexual process but by budding from the tissues of the nucellus. 
Such nucellar embryos are important for genetics as well as for morpho- 

Of particular interest, however, are those forms in which the early struc- 
tures are markedlv different from later ones and in which characteristic 
"juvenile" stages can be seen. This type of development has been termed 
heteroblastic by Goebel in contrast to the more gradual homoblastic type. 
The difference is particularly conspicuous in the character of the leaves, 
which in the seedling are often quite unlike those of the mature plant. 
The first pair of leaves in the Eucalyptus seedling, for example, are hori- 
zontally oriented and dorsiventral in structure though all later foliage is 
characteristically pendulous and bifacial. The juvenile leaves of Acacia 

Fig. 8-15. Development of the embryo of Capsella bursa-pastoris 
from the first division of the fertilized egg to the mature embrvo. 
( From Johansen, after Soueges and Schuffner. ) 


208 The Phenomena of Morphogenesis 

are pinnately compound but the adult ones are reduced to phyllodes 
(Fig. 8-16). In pine, the seedling leaves are not in fascicles but are borne 
singly. The young plant of Phyllocladus has needle-like leaves, common 
in most other conifers, but the adult plant bears phylloclads only. Seed- 
lings of cacti have leaves but these are absent in adult plants. Many more 
such examples could be cited (Jackson, 1899). 

There are some cases in which the internal structure is markedly dif- 
ferent in young plant and adult, usually being simpler in character in the 

Fig. 8-16. Juvenile leaves of Acacia seedling (pinnately compound) contrasted with 
the flattened phyllodes that constitute the adult foliage. (After Velenovsky.) 

former. Thus in ferns which have a complex vascular system in the ma- 
ture plant, the young sporeling possesses a relatively simple protostele or 
siphonostele. Species with many-bundled leaf traces usually have only 
three in the seedling. Secondary tissues are also less complex in young 
plants. Schramm ( 1912) finds that juvenile leaves generally resemble adult 
shade leaves in structure. There are differences, particularly as to vena- 
tion, between the early, deeply pinnatifid leaves of Lacunaria and the 
simple mature type ( Foster, 1951 ) . Robbelen ( 1957 ) finds that in chloro- 
phyll-defective mutants the juvenile form of leaf is retained later than 



normally and has a relatively small meristem. Normal leaves are not pro- 
duced until the meristem reaches a diameter of 80 to 90 /*. 

Schaffalitzky de Muckadell (1954) has reviewed the literature on ju- 
venile stages. 

Juvenile traits often resemble those of plant types presumably an- 
cestral for the stock in question (Sahni, 1925, and others). This seems 
evident in many of the examples cited. Most Leguminosae other than 
Acacia have leaves and not phyllodes, and most Myrtaceae other than 
Eucalyptus have dorsiventral leaves. These facts suggest that the seedling 
repeats or recapitulates ancestral traits, much as the animal embryo has 



A-24 A-20 



Fig. 8-17. Changes in leaf shape in cotton at five successive nodes above the cotyle- 
dons, in four varieties of cotton and an Fi. ( From Dorothy Hammond. ) 

been thought to do. There is much doubt in many cases, however, as to 
what the course of evolution actually has been and so much variation in 
early ontogeny in many plants that it is impossible to establish the doc- 
trine of recapitulation as an invariably useful guide to phylogeny. 

Progressive Developmental Changes. More common than these con- 
spicuous cases of differentiated juvenile stages are those where there is 
not a sharp distinction between juvenile and later forms but a gradual 
change from the younger part of the plant to older ones. Many examples 
of this are reported in the literature. Goebel (1896) described eight suc- 
cessive leaf types in the climbing aroid Anadendrum medium which 
occurred at different levels and showed an increasing degree of com- 

210 The Phenomena of Morphogenesis 

plexity. Hammond (1941; Fig. 8-17) and Stephens (1944) have described 
similar changes in leaf shape in cotton, and Montfort and Miiller ( 1951 ) 
in mistletoe. Von Maltzahn (1957; Fig. 8-18) has compared leaf char- 
acters throughout plant development in large and small races of Cucur- 
bita and the hybrid between them. 

Similar alterations have also been found in reproductive structures. In 
Chamaecyparis there is a gradient of sexuality in the branches, the tips 
being sterile, with female cones below and male ones still farther back 
(Courtot and Baillaud, 1955). There is a flower bud in the axil of each 
leaf of Cucurbita pepo but the type of flower produced by it tends to vary 
with the position of the leaf on the plant, in the following sequence: 
underdeveloped male, normal male, normal female, inhibited male, and 
parthenocarpic female (Fig. 8-19). The order of these steps in progressive 
feminization is constant but the length of each is affected by temperature 

Fig. 8-18. Change of leaf size during 
plant growth. Lamina length of succes- 
sive leaves in cucurbit plants of small- 
fruited and large-fruited types and the Fi 
between them. ( From von Maltzahn. ) 

I 3 5 7 9 I I 13 15 I 7 19 21 

and day-length, high temperatures and long days extending the male 
phase and delaying the female one ( Nitsch, Kurtz, Liverman, and Went, 

Leaves are especially good material in which to study such changes, 
and Ashby ( 1948/?, 1950a, and Ashby and Wangermann, 1950) has made a 
thorough investigation of the changing character of the leaves in Ipomoea, 
describing the progressive differences in their size and shape and in the 
size and number of their cells from lower nodes to upper ones (Fig. 
8-20). He presents evidence that these changes are not primarily due to 
environmental factors (although such are operative) but to alteration of 
inner conditions. In his 1948 papers Ashby reviews this field and discusses 
at some length Krenke's theory (1940) that such changes are due to the 
physiological age of the plant, as contrasted to its age in time. Krenke 
regards aging as progress toward maturity, particularly reproductive ma- 
turity, which is sometimes followed by further changes, in a cycle. At 
points along this progression rejuvenescence may occur, as on shoots 
grown from lateral buds. Successive nodes are units in a developmental 



scale, and the form of the leaf is a quantitative criterion of physiological 
age. In cotton, maximum lobing of the leaf is reached at flowering, earlier 
and later leaves being less lobed. The cyclical change proceeds more 
rapidly in early-flowering than in late-flowering types, and conditions that 
hasten flowering hasten lobing. Krenke believes that rate of change in 
leaf shape is inherited and that early-maturing varieties may thus be dis- 
tinguished in the seedling stage. 





Fig. 8-19. Cucurbita. Sequence of flower 
types on a plant of the acom squash. 
( From Nitsch, Kurtz, Liverman, and 







Ashby has confirmed some of Krenke's conclusions but finds others very 
doubtful. The possibility, however, of relating successive morphological 
changes to physiological ones has important implications for the problem 
of form determination. One may question Krenke's assertion that his 
hypothesis is based on dialectical materialism, but the hypothesis itself 
should be explored as one hopeful approach to morphogenetic problems. 

The bearing of Krenke's ideas on the problem of senescence is of in- 

212 The Phenomena of Morphogenesis 

terest. Attention has already been called (p. 38) to Benedict's work on 
progressive reduction in size of structural units (cells and vein islets) in 
vegetatively propagated clones as they grow older, presumably the result 
of loss of vigor. This conclusion is still open to doubt, however. 

Topophysis. The changes so far discussed have been either juvenile ones 
or those distributed through most of the life cycle. In many plants, how- 
ever, the contrast in differentiation does not come until the onset of repro- 
ductive maturity. Diels ( 1906) observed that environmental factors which 
promote flowering also hasten the transition from juvenile to adult foliage. 






~l 1 1 1 1 1 « ■ » TTT1 

2 3 4 5 6 7 8 n-8 n-7 n-6 n. 

Fig. 8-20. Change in leaf shape at successive internodes from base to tip in Ipomoea 
caerulea. (From Ashby.) 

The reproductive stage is marked by characteristic changes at the meri- 
stem, especially in the shape of its terminal dome. In the vegetative phase 
this is typically low and rounded, but when flower buds are to be differ- 
entiated it assumes a much steeper and more elongate form. The produc- 
tion of reproductive organs marks for most plants a radical reorganization 
of their developmental processes (p. 184) and is often accompanied by 
changes so profound that they are irreversible. This is of especial morpho- 
genetic significance. 

In most cases these changes have little effect on the character of the 
vegetative organs, but such cases are sometimes found. In the conifer 

Differentiation 213 

Dacrydium ciipressinam, for example, needle-like foliage (resembling 
that of seedlings in many species with scaly leaves ) occurs not only in the 
seedling but throughout the early life of the tree. Only when it begins to 
bear reproductive organs, at the age of 20 years or more, does the foliage 
assume the scale-like character which then continues throughout the rest 
of its life. This may be interpreted as the persistence of a juvenile con- 
dition until the period of reproduction. 

A more conspicuous example and one which has been widely studied 
is that of the English ivy, Hedera helix ( Fig. 8-21 ) . The vegetative phase 
of this plant is a vine with five-lobed leaves climbing by adventitious roots 

Fig. 8-21. Hedera helix. Flowering shoot with ovate, entire leaves, and a single leaf 
of the vegetative "juvenile" region. (From Goebel.) 

and flattened against its support and often is the only form of this species 
to be seen. After some time and under conditions favorable for reproduc- 
tion, however, flower-bearing shoots arise and will grow for many years. 
Their tropistic behavior is changed, for they no longer climb but grow 
directly outward toward the light and away from their support. Their 
structure is also much altered, for the leaves are now oval rather than 
lobed and are spirally arranged. The lobed climbing form may be re- 
garded as a persistent juvenile condition. 

The reorganization of the pattern of differentiation in the transition 
from seedling to adult or from the vegetative to the reproductive phase 

214 The Phenomena of Morphogenesis 

of the ontogenetic cycle is usually reversible in the sense that cuttings 
taken from any part of the shoot system or at any stage of development 
will, by regeneration, produce a normal plant. In some cases, however, 
changes at the growing point have been so great that the newly developed 
structures seem to have undergone irreversible modification. A notable 
example of this is the English ivy described in the preceding paragraph, 
for in this plant cuttings made from the flowering shoot rarely revert to 
the climbing form but instead produce upright, radially symmetrical 
plants, the variety arborea of horticulture. These are often used as dwarf, 
tree-like ornamentals but usually die after a few years. No genetic change 
is involved here, for seeds produce the climbing, lobed form again. There 
has been a good deal of debate as to the complete irreversibility of this 
change (Bruhn, 1910; Furlani, 1914), but the usual behavior is the one 
just described. Kranz (1931), however, finds that the transition from 
juvenile to adult foliage is often not a sudden one but that the five-lobed 
type gives place to a three-lobed one before the mature, ovate leaves are 
formed. Robbins (1957) was able to change the adult form of foliage to 
the juvenile one by treatment with gibberellic acid. 

A somewhat similar case is the persistence of juvenile structures which 
can sometimes be induced by growing cuttings from the seedling stem. 
The most notable example of this is found in certain of the cypress-like 
conifers, where the seedling leaves are needle-like but are soon followed 
by the scale-like foliage characteristic of the species. If cuttings are made 
from the lateral branches arising just above the cotyledons in Thuja, for 
example, they will produce plants, often growing to small trees, which 
bear nothing but the needle-like juvenile foliage, the horticultural variety 
"Retinospora" (Beissner, 1930). Such plants do not flower and are rela- 
tively short-lived. In some way, severance of the juvenile shoot from its 
roots seems to have prevented completion of the normal ontogenetic 
cycle. Other cases have been reported in which seedlings used as cuttings 
grow very differently from those which are left on their own roots 
(M. R. Jacobs, 1939). Beissner's results were challenged by Woycicki 
(1954), who grew cuttings from seedlings of Thuja, Chamaecyparis, and 
Biota but found that the juvenile foliage did not persist. He believes that 
the Retinospora forms arose by spontaneous mutations in seedlings or 
young shoots. 

Whatever the facts in this case may be, others have been reported in 
which it is certain that cuttings taken from various parts of the plant 
produce individuals different from the normal type and like the part of 
the plant from which they came and in which these differences persist 
during the life of the cutting, or at least for a long time, but do not involve 
genetic change. This phenomenon Molisch ( 1930 ) termed topophysis. A 
familiar example occurs in Araucaria (p. 189), where the flattened, dor- 

Differentiation 215 

siventral character of the lateral branches persists in cuttings made from 
these branches. All these cases of persistent differentiation are of particu- 
lar interest in providing material for a study of the cause and character of 
differential change. 

The origin of differences arising at different times in a repeated cycle, 
rather than at a different place on the plant, Seeliger (1924) has termed 


Most of the examples of differentiation thus far cited seem to be pri- 
marily the expression of a developmental pattern controlled by the genetic 
constitution of the individual. Obviously such a constitution cannot op- 
erate except in an environment of some sort, for genes control specific 
differences in reaction to specific environmental factors. It is therefore to 
be expected that differentiation should be greatly influenced by the en- 
vironment, both internal and external. 

The basis for differentiation itself is provided by the environment, for 
the most important contribution that the physical environment makes, 
morphogenetically, is to set up a gradient in the organism. This cannot 
be done unless the environment itself displays a gradient in direction or 
intensity. Fern prothallia, for example, grown in culture on a shaking 
machine, and thus exposed equally to gravity on all sides, or on a revolv- 
ing table, and thus exposed equally to light on all sides (p. 137), are in a 
homogeneous environment which has no gradients, no single direction of 
gravity or light. As a consequence the organism produces an amorphous 
mass of tissue for it is without a polar axis, the basis for its differentiation. 
Such an axis must be induced, at least at the very start, by an asym- 
metrical environment. 

Environment and External Differentiation. The most obvious relation 
between environment and differentiation is in the effect that external 
factors have on the form and character of plant organs. Most of the final 
part of this book will be concerned with the morphogenetic effects of such 
factors. Light influences the differentiation of reproductive and other 
structures by its intensity, its wave length, and the duration of its photo- 
period. The amount of available water is important in the induction of 
xeromorphic structures. Temperature, particularly in early development, 
seems to affect the rate of certain processes that are precursors to flower- 
ing. Chemical agents, notably growth substances, have a marked influence 
on differentiation of all sorts. The discussion of these problems must wait 
until later pages. There are a few conspicuous instances, however, where 
differentiation obviously is dependent on environmental factors which can 
best be described here. 

216 The Phenomena of Morphogenesis 

One is the general phenomenon of heterophylly, where two or more 
widely different types of leaves, usually without intermediate forms, may 
occur on the same plant. This difference is most commonly, though not 
always, associated with the plant's ability to live either submersed in 
water or rooted in the ground with its shoots in the air ( Gliick, 1924, and 
p. 330 ) . This may be interpreted as a case of heteroblastic development in 
which the manifestations are reversible. It is related to the problems of 
juvenile stages and progressive development discussed in the preceding 

In many species of the pondweeds (Potamogeton) the floating leaves, 
which rest on the surface of the water, are relatively broad and have an 
internal structure not unlike ordinary herbaceous foliage, whereas the 
leaves borne under water are long, narrow, and membranous, thus being 
adapted to live as submerged organs. Somewhat similar differences may 
be seen in various "amphibious" plants (p. 332), such as the water butter- 
cup (Ranunculus aquatilis) and the mermaid weed (Proserpinaca palus- 
tris ) . These species live in environments where part of their foliage grows 
in air and part is submersed under water. Under the former condition, 
the leaves are relatively broad and well provided with stomata and inter- 
cellular air chambers. In the latter they are much dissected and thinner. 
These effects of the environment on water buttercup were observed by 
Lamarck and played an important part in the development of his theory 
of evolution. The relation of differences between the "water" and "land" 
forms in such plants to those between juvenile and adult stages has been 
discussed by various workers. Burns (1904) believes that the "water" 
form of Proserpinaca is the juvenile stage, associated with unfavorable 
conditions, and the "land" form the adult type and associated with flower- 
ing. He found that only the broad, entire leaves were formed in the 
flowering season and only the dissected ones in the winter. Whether all 
such cases of heterophylly may thus be interpreted is a question. Vischer 
( 1915 ) has called attention to the close relation between certain environ- 
mental factors (such as removal of leaves, weak light, damp air, deple- 
tion of carbohydrate reserves, and increased soil fertility ) and the produc- 
tion of juvenile foliage (see also p. 206). Factors which favor reproduction 
tend to produce adult foliage. Under certain conditions a return to the 
juvenile condition may be induced (Woltereck, 1928). There is no evi- 
dence that the ribbon-like submersed leaves of Potamogeton are juvenile 
in type, however. Arber (1919) points out that in Sagittaria (another 
"amphibious" plant) the first leaves are thin and ribbon-like even when 
the plant is growing out of the water and that they appear at maturity 
whenever the plant grows weakly. She believes that an aquatic environ- 
ment is not responsible for heterophylly but that the occurrence of hetero- 

Differentiation 217 

phylly is a necessary prerequisite for the ability of a plant to live in both 
aquatic and terrestrial habitats. 

Pearsall and Hanbv (1925) have evidence that leaf variation in Pota- 
mogeton is due, at least in part, to chemical differences in the soil, and 
Gessner ( 1940 ) and Bauer ( 1952 ) relate it to rate of metabolism in the 
buds. McCallum (1902) thought that in Proserpinaca the water type of 
leaf arose primarily because of reduction in transpiration. H. Jones ( 1955) 
has made extensive studies of the differences in development of the pri- 
mordia that produce the linear and the ovate leaves of Callitriche and 
the conditions under which these are formed. 

There are many instances where, instead of the permanent induction 
of structures at certain ontogenetic levels, there may be reversion to earlier 
stages under certain environmental conditions. This is especially frequent 
in those cases where juvenile stages are adapted to different environments 
than are the adult ones. A commonly cited example is that of Campanula 
rotundifolia, which has rounded juvenile leaves adapted to weak light, 
although the mature leaves are linear. A mature plant grown in low 
illumination will often revert to the juvenile type of foliage. Seedlings, 
even in strong light, however, bear nothing but juvenile foliage. Often 
wounding will bring about such reversion, as in shoots growing from in- 
jured regions of certain pines, which for a time bear foliage like that of 
the seedling. With many perennials there is a partial return to the juvenile 
stage at the beginning of each growing season. 

Frank and Renner (1956) found that in Hedera helix chemical treat- 
ments of various sorts did not induce reversion to the juvenile state but 
that cold shocks and X irradiation did so. De Zeeuw and Leopold ( 1956 ) 
were able to induce flowering by auxin treatment in juvenile plants that 
otherwise would not have flowered. They suggest that the completion of 
the juvenile phase may be due to the accumulation of a sufficient auxin 
level. Robbins ( 1957 ) has shown the effectiveness of gibberellin in ju- 
venile reversion. Allsopp ( 1955 ) attributes heteroblastic differentiation 
in general to changes, chiefly of size, in the shoot apex following alter- 
ation of the water balance ( p. 332) . 

Environment and Internal Differentiation. Internal differentiation, also, 
may be greatly affected by environmental factors. It is important to recog- 
nize that changes that take place in this process are part of an underlying 
pattern of relationships among the cells and between them and the en- 

This fact is made clear whenever such relationships are disturbed. If 
tissue like the cortex, for example, is exposed to the outside air by re- 
moval of the outer cell layers, structures tend to differentiate at the new 
surface which are characteristic for such a position. Thus when Vochting 

218 The Phenomena of Morphogenesis 

( 1908 ) sliced off a portion of a kohlrabi tuber, the living cells at the new 
surface differentiated into a rather typical epidermis in which even 
stomata were formed. In roots of the Araceae and air roots of orchids, 
where there is no cell division after an injury, parenchyma cells near a 
newly exposed surface redifferentiate into thick-walled ones essentially 
like those of a normal hypodermis. 

Even more complex patterns may be reconstituted under the influence 
of a different environment. In the roots of Philodendron Glaziovii there is 
a row of brachysclereids a few cell layers below the surface. After the 
experimental removal of the outer tissues, a similar row of thick-walled 
cells differentiates at about the same distance below the new surface 
( Bloch, 1926 ) . In the air root of Monstera, the cells of the cortex normally 
remain undifferentiated for a considerable distance back from the tip. 

Fig. 8-22. Air root of Philodendron. Below arrow, 
normal hypodermal tissue pattern, with layer of 
brachysclereids. Above, regeneration of similar 
layer below wound. (From Bloch.) 

At this point, however, the four or five cell rows next the outside often 
form thick, lignified walls and develop into brachysclereids. The differ- 
ence between these cells and their unlignified neighbors is not evident at 
the last cell division nor can it be traced through any cell lineage. It arises 
as these two types of cells become mature. The occurrence of lignification 
is apparently related to the position of cells with reference to the surface 
of the root and thus probably to such an environmental factor as an 
oxygen or water gradient. When a root of Monstera is wounded in such a 
manner that the parenchyma cells of the inner cortex are now exposed to 
a new, artificially produced surface, they become thick-walled brachy- 
sclereids (Bloch, 1944; Fig. 8-22). When Wardlaw isolated the central 
core of the shoot meristem by vertical incisions, he observed that the 
cylinder of vascular tissue regenerating inside the core developed at a 
constant distance from the new surface made by the cuts ( p. 238 ) . 
In a few instances where the normal ontogeny may be completed under 

Differentiation 219 

a given environment, exceptionally favorable conditions will enable the 
plant to realize developmental potencies which it never would display 
otherwise. Thus Bloch (1935/?) has shown that in Tradescantia flumi- 
nensis, which typically has bundle sheaths with only thin-walled cells, 
wounding may so stimulate differentiation that thick-walled sheath cells, 
similar to those in related species of Tradescantia, may be formed. Here 
the ontogenetic cycle has been extended beyond its normal course, either 
in reversion to a former evolutionary level or toward the realization of 
developmental potencies not yet normally expressed by this species, 
though common in related ones. 

Often the whole histological pattern may be affected. In air roots of 
orchids growing freely, adventitious roots are produced on all sides; but 
if the root is in contact with a support, these lateral roots are formed only 
next the support, presumably because of differences in moisture or other 
factors on the two sides (Bloch, 1935«). Anatomical differences are also 
evident in these two root sectors. 

In differentiation, the role of specific substances, particularly growth 
substances, is important (p. 390). Root-forming substances, shoot-forming 
substances, flower-forming substances, and others have been postulated. 
That a substance by itself has a specific organ-forming character is prob- 
ably too naive a conception, but certainly auxin and various other hor- 
mones and growth substances are effective as stimuli which call forth 
specific morphogenetic responses in the plant. Auxin influences the growth 
of cambium, the development of vessels, and other histological processes. 
It also inhibits certain activities. Beneath epidermal cells that regenerate 
new shoots (p. 245) vascular tissue often differentiates, presumably because 
of a substance coming from the young bud, which thus is able to estab- 
lish a connection with the main vascular system. Camus has shown that 
buds grafted to pieces of fleshy root in tissue culture induce the differen- 
tiation of vascular tissue in parenchyma cells beneath them, and Wetmore 
found that auxin alone does the same thing ( p. 405 ) . 

In general, environmental factors seem chiefly to affect the later stages 
in cellular differentiation and especially the character of the cell wall. The 
fundamental pattern of a structure is less affected than are its quantitative 
expression and the size and character of the cells that compose it. 

What a cell or tissue will do depends in part upon its innate genetic 
potentialities and in part upon the environment in which it happens to be. 
Cells possess different degrees of reactivity to environmental differences. 
In some, this is small, and the fate of the cell is therefore rather fixed and 
limited, regardless of its environment. In others its developmental reper- 
toire is much wider, and it may be greatly influenced by the conditions 
which surround it. It should be remembered that the degree of a cell's 
reactivity is not constant but that it may vary with the position that the 

220 The Phenomena of Morphogenesis 

cell occupies in the whole developmental pattern and in the ontogeny 
of the individual and especially that it depends upon the point that the 
cell has reached in its own cycle of maturity. The developmental expres- 
sion of the genetic constitution of a plant and of its various parts is there- 
fore not fixed and constant but is continually changing. 


All differentiation, of course, has its basis in the physiological activities 
of living substance, but it can usually be recognized most readily when 
these activities result in the production of visible differences in structure. 
It is such differences that have chiefly been considered in the preceding 
pages. Physiological differentiation itself, however, can often be demon- 
strated, and experiments in this field offer hope for the solution of many 
developmental problems. A few typical examples will be discussed briefly 
here and others in later pages. 

The diversities in structure between root and shoot are doubtless the 
expressions of fundamental physiological differences. One of the most con- 
spicuous of these is in vitamin synthesis. By culture methods it is possible 
to grow roots indefinitely from a bit of root tip. Such root cultures must 
be provided with the necessary mineral salts and also with a source of 
carbohydrates (usually sucrose). These alone prove to be not enough to 
secure indefinite growth, and they must be supplemented by small 
amounts of certain vitamins, in most cases thiamin. It is clear, therefore, 
that such roots are unable to synthesize this vitamin. Since thiamin is 
known to be present in the shoots of plants, this is evidently the region 
in which it is produced. In nature, roots must obtain their supply from 
the shoots. Just when this physiological differentiation first occurs is not 
known, but it is probably at the time when root and shoot are set apart 
in early embryology. 

It has been shown that root and shoot also differ in their ability to 
synthesize certain alkaloids. Tobacco shoots can be grafted onto tomato 
roots, and leaves and stems of such shoots are free from nicotine (Daw- 
son, 1942). If tomato shoots are grafted on tobacco roots, however, the 
tomato tissues contain large quantities of this alkaloid. It is therefore 
obvious that in such cases the capacity to synthesize nicotine is confined 
to the tobacco root and is not possessed by its leaves, as has commonly 
been assumed. Certain other alkaloids (as quinine) can be shown by such 
experiments to be synthesized in both roots and shoots. The fact that a 
substance occurs in a certain part of the plant is evidently no proof that 
it is formed there. This technique of reciprocal grafting provides a useful 
tool for the demonstration of physiological differentiation of this sort. 

Differentiation 221 

Studies of geotropic reaction of typical roots and stems show that they 
are also different in their response to auxin, the growth of roots being 
inhibited by concentrations which stimulate growth of stems, a fact which 
explains the geotropic reactions characteristic of these two organs. They 
differ physiologically in other respects, for Collander ( 1941 ) has shown 
that certain cations may be differentially distributed between root and 
shoot, sodium and manganese being more abundant in the former and 
calcium, strontium, and lithium in the latter. 

Differences between vegetative and reproductive phases of the life 
cycle are sometimes physiological as well as structural. Many early workers 
(Sachs, 1880, 1882) noticed the difference between "blind" and "flower- 
ing" stems, the former when used as cuttings producing vegetative growth 
only and the latter, flowering shoots. This difference has now been shown 
to be related to the presence of some substance or substances which 
induce flowering (p. 397). Torrey (1953) reports that three synthetic 
substances which inhibit root elongation have specific effects on the ac- 
celeration or retardation of the differentiation of xylem and of phloem. 

Physiological differentiation must evidently be important in sex deter- 
mination, and chemical differences between the sexes have been found 
by several workers. By the Manoilov reaction, for example, staminate 
and pistillate plants of poplar can be distinguished, as well as "plus" and 
"minus" strains of Mucor ( Satina and Blakeslee, 1926 ) . Stanfield ( 1944 ) 
has described chemical differences between staminate and pistillate 
plants of Lychnis dioica. Aitchison (1953) found that in several genera 
the sexes were unlike in oxidase activity, this being greater in some cases 
in males and in others in females. Hoxmeier (1953), working with Canna- 
bis and Spinacia, reports that the tissue fluids of staminate plants are more 
acid than those of pistillate ones. In Cannabis, Cheuvart (1954) observed 
differences between the sexes in chlorophyll content, especially in the rate 
at which this is reduced at the time of flowering. Reinders-Gouwentak 
and van der Veen ( 1953 ) found that in poplar the female catkins tended 
to stimulate wood formation on the stem below them whereas males did 
not, suggesting a difference between the sexes here in the production of 
a growth substance. 

Regular changes in the physiological activity of the series of successive 
leaves on a plant, related to both position and age, have been observed 
by various workers. Dormer ( 1951 ) determined the dry weight per unit 
of length in successive internodes of Vicia from the apex downward and 
found that during the unfolding of the ninth leaf there was a sudden 
change in the distribution of the dry-weight increment. The nutritional 
history of an internode thus seems to be a function of its position in the 
stem. The developing seedling has also been shown to change in its physi- 

222 The Phenomena of Morphogenesis 

ological character. Rietsma, Satina, and Blakeslee (1953a), by growing 
Datura embrvos in tissue culture, have shown that the minimal sucrose 
requirement falls steadily from the earliest stages to the mature em- 

A notable example of physiological ontogeny has been reported by Wet- 
more (1954). In the developing fern sporeling the first leaves are two- 
lobed. These are followed by three-lobed ones and finally by pinnate 
leaves in which an apical cell has appeared. Shoot apices from small fern 
sporelings, cultured in mineral nutrients and various concentrations of 
sucrose, grew into whole plants. Where the concentration was low, only 
two-lobed leaves were formed. Higher concentrations produced three- 
lobed ones and still higher, pinnate ones. The normal ontogenetic progres- 
sion here thus seems to be related to an increasing supply of sucrose. 

Metabolic gradients are marked by various physiological differences, 
especially as to the rates of reactions. Prevot ( 1940) observed that respira- 
tion in the apical region of the root of several genera was greater than 
in the more distal regions. This is not always the case, however, in shoot 
meristems (p. 73). 

Wardlaw (1952c) has found that the nutritional status of the apical 
region in ferns has an important effect on the size and character of the 
leaves and stelar structure. Apices that normally produce large and com- 
plex leaves and an elaborate vascular system, if reduced in size by poor 
nutrition, will form "juvenile" leaves and simpler stelar patterns. 

Biochemical differences of many sorts, presumably indicating physi- 
ological diversity, can be shown in cells and tissues. Differential staining 
reactions are familiar examples of this. Differences in hydrogen-ion con- 
centration between cells visibly alike can be shown by the use of indi- 
cators. Blakeslee (1921) demonstrated the presence of two chemically 
different areas of cells in the petals of certain races of Rudbeckia with 
solid petal color by dipping the petals into phenolphthalein. Sometimes 
differentiation is shown by the occurrence of natural pigments, as in the 
root tips of Phalaris arundinacea where the trichoblasts are pigmented 
but the cells that are not to produce root hairs are colorless. Van Fleet's 
work on the histochemical differentiation of the developing endodermis 
has been mentioned (p. 197). Microchemical tests of various sorts being 
out differences between many kinds of cells, even in early development, 
such as tannin cells, crystal cells, and latex ducts. Spectrometric demon- 
stration of differences in distribution of the nucleic acids are among 
notable recent examples of this sort of analysis. 

Less work has been done in demonstrating physical differences be- 
tween cells. In fern prothallia Akdik (1938) and Gratzy-Wardengg (p. 
121 ) found a definite pattern of differences in osmotic concentration over 
the surface of the prothallus, and this seems to be related to differences in 

Differentiation 223 

behavior of these regions in regeneration. In iris leaves Weber ( 1941 ) 
showed that the first indication of differentiation of stomatal mother cells 
is a difference in osmotic concentration. 

Much cellular differentiation is due to changes in the cell wall. Boysen- 
Jensen ( 1957 ) in a series of papers has demonstrated various wall changes 
in the differentiation of root hairs with particular references to the action 
of enzymes. 

These and many other observations show that morphological differen- 
tiation has its physiological concomitants. To explain how these arise dur- 
ing development is a major task of the student of differentiation. 


There are a number of instances among the fungi where development 
of the fruiting structures does not take place until the vegetative phase 
of the life cycle has ended and no more food is absorbed from the en- 
vironment. Growth, in the broader sense of the term, is therefore com- 
pleted before differentiation begins, and the latter process can be studied 
without the complications that are usually involved when growth accom- 
panies it. One of the .most notable examples of this is furnished by the 
Acrasiaceae, a family of the slime molds. 

The vegetative individual in these plants is a single amoeboid cell, or 
7ni/xamoeba. These multiply profusely by simple division and live chiefly 
on several species of bacteria. They can readily be grown in culture. After 
vegetative life has gone on for some time and when external conditions 
are favorable, a large number of these myxamoebae, in a group of from 
several thousand to about 150,000, begin to move toward a center of 
aggregation, streaming in from all sides and piling up into a mass of cells, 
the pseud oplasmodium (Fig. 8-23). This is a millimeter or two in length, 
is elongate in form, and somewhat resembles a small grub. It is sur- 
rounded by a thin sheet of slime. By the time that this aggregation begins, 
all vegetative growth has ceased, so that in the life cycle of these plants 
growth (in the sense of increase by assimilation) and differentiation are 
separate from each other in time. 

The process of aggregation seems to be controlled by the production 
of a chemotactically active substance, acrasin. The timing and mechanism 
of this process have been discussed by Shaffer ( 1957 ) . As to what deter- 
mines the origin of these centers of aggregation, however, there is some 
difference of opinion. Sussman (1952) believes that a few initiator cells 
appear in the population and attract their neighbors into a many-celled 

Wilson (1952) presents evidence that aggregation has its origin in a 
sexual process, two myxamoebae fusing early in aggregation and estab- 

224 The Phenomena of Morphogenesis 

lishing a center. Other fusions occur later and are followed by meiosis. 
The zygotes can be distinguished by their greater size. 

In the pseudoplasmodium the myxamoebae do not fuse but each cir- 
culates freely among its neighbors, and the whole mass moves over the 
surface of the substratum by means that are not yet clearly understood. 
This body of separate cells, however, is not without some degree of differ- 
entiation. It is elongated in the direction of its movement, which is toward 
the light. The apical end is slightly pointed and lifted above the rest and 
is richer in acrasin than the other regions. It is the part of the mass that is 


^0 & 

X 0' f 

°\o hy»* 


on " ^ ° 

^0 e- <g q POQtflQ 


-3 « (7 fc *> 

9 e 

o° 8 ° !) 



Aa ^ 




LJ>o<»*~» «% 





Fig. 8-23. Dictyostelium. Stages in aggregation of myxamoebae into a pseudoplas- 
modium. ( From J. T. Bonner. ) 

sensitive to the stimulation of light and seems to serve as a directive 
center for the whole. A pseudoplasmodium from which the apex has been 
removed will stop its motion and settle down at once to form a fruiting 

Two groups of cells may be distinguished in the pseudoplasmodium. 
Those near the apex and destined to form the stalk of the sorocarp are 
somewhat larger than the ones in the posterior region, which will later 
form spores. The proportion between these two types is maintained by 
a regulatory process irrespective of the size of the whole mass. Some cell 
division continues in the pseudoplasmodium but the rate is different in its 

Differentiation 225 

two regions. If the apical and the basal halves of the pseudoplasmodium 
are experimentally separated, each will form a sorocarp, but the one from 
the apical half produces the larger spores. Despite these evidences of 
the beginning of differentiation, cells in one group can be changed to 
resemble those of the other, and every cell apparently is totipotent. If a 
few cells are removed from the mass they are no longer subject to its 
organizing control, and if food is present, they will become vegetative 
cells again and proceed to multiply. 

That the pseudoplasmodial axis is polarized is shown not only by the 
difference in structure of its two ends but by their behavior. K. Raper 
(1940Z?) performed a series of grafting experiments between plasmodia 
that could be distinguished by their color, one group having fed on red 
bacteria. The apex, if cut off and placed at the rear of another plasmo- 


Fig. 8-24. Migration of the pseudoplasmodium and formation of the sorus in Dictyo- 
stelium discoideum. (From ]. T. Bonner.) 

dium, will not fuse with this one but will start off by itself. It fuses 
with the apical end of a decapitated plasmodium (Fig. 6-15). If an 
apex is cut off and placed next the side of an intact plasmodial mass it 
will attach itself there and finally draw off a considerable mass of cells 
and establish a separate plasmodium. The tip of the mass sometimes 
splits, and in this case two are formed. If two happen to come together, 
they may fuse into a single one of double size. 

Morphogenetically the most significant part of the life history is the 
formation of the stalked fruiting body, or sorocarp. After a few hours of 
migration, or when a pseudoplasmodium reaches drier surroundings, it 
stops moving and attaches itself firmly to the substratum by a disk of dif- 
ferentiated cells (Fig. 8-24). The cells of the apical region, from which 
the stalk is to be formed, become large and vacuolate and each is en- 
closed in a cellulose sheath. As this is happening, they are pushed down 
into the pseudoplasmodium by other prestalk cells climbing up around 

226 The Phenomena of Morphogenesis 

them which in turn become stalk cells (Fig. 8-25). As Bonner describes 
it, "The process is the reverse of a fountain; the cells pour up the outside 
to become trapped and solidified in the central core which is the stalk. 
In so doing the whole structure rises into the air until all the prestalk 
cells have been used up." This description applies to the species most 
commonly studied, Dictyostelium discoideum. In D. mucoroides and D. 
purpureum, however, the stalk begins to be formed during the brief 
migration of the pseudoplasmodium. 

About 10 per cent of the myxamoebae take part in the formation of 
disk and stalk. The others, still moving freely over one another, follow 
the growing tip of the stalk upward in a body and (in Dictyostelium) 

O Undrfferer*oled 
© Supportive 


Fig. 8-25. Dictyostelium discoideum. Diagram of sorocarp formation. A, B, migrating 
pseudoplasmodium. ( From J. T. Bonner. ) 

form a spherical mass of cells, the sorus, at its summit. Here each myxa- 
moeba rounds up to make a dry spore, and these spores are later carried 
away by air currents, each now capable of developing into a myxamoeba. 
Wilson presents evidence that some mitotic division occurs before spore 

In the genus Polysphondylium there is not only an apical sorus but 
several whorls of lateral stalks along the main one, each terminating in 
a smaller sorus, so that the sorocarp becomes a structure of considerable 

These plants, though so unlike higher ones, have constant generic and 
specific differences ( Fig. 8-26 ) . If myxamoebae belonging to two species 
grow intermingled in the same culture or if pseudoplasmodia of two 
species are crushed and experimentally mixed, the cells in time sort them- 



Polysphon dylium 

P. pallidum 

D. purpureum 

Fig. 8-26. Sorocarps of various members of the Acrasiaceae. (From K. B. Raper.) 

selves out and form sorocarps typical of each species. The specific char- 
acter of the sorocarp can be shown to be carried by its spores. 

Sussman ( 1955 ) has found a number of mutants of Dictyostelium dis- 
coideum in which the aggregating groups, and consequently the size of 
the sorocarps they form, is much smaller than normal. In some of these 
the fruiting body consists of as few as 12 cells but it still retains the form, 

228 The Phenomena of Morphogenesis 

proportions, and cellular structure of the larger ones, surely a remarkable 
example of the inherent formativeness of these cells. Here form deter- 
mination finds one of its simplest expressions. The behavior of the myxa- 
moebae in the Acrasiaceae reminds one of the well-known instance among 
the sponges where the entire body may be separated into its constituent 
cells and these later will come together and re-form the body of the 
sponge. The mechanism by which such morphogenetic movements occur 
in the slime molds and sponges presents some of the most baffling prob- 
lems in biology. 

The Acrasiaceae have been studied intensively in recent years. For more 
detailed accounts of experimental work on them the reader is referred to 
the publications of Bonner, Raper, Shaffer, Sussman, and Wilson, some 
of which are listed in the bibliography. The field has been reviewed by 
J. T. Bonner (1959). 

A somewhat similar example of the differentiation of a formed struc- 
ture by a mass of undifferentiated cells is shown by a specialized family 
of bacteria, the Myxobacteriaceae. This is a group in which the individual 
is a rod-like cell which divides by transverse fission, it lives on other 
bacteria. In the vegetative period these cells may be distributed through- 
out a colonial mass or may occur in radiating strands or ridges. Myxo- 
bacteria do not possess flagella, but their gliding or creeping movements 
are in some way associated with the abundant slime that they secrete. In 
the fruiting phase the individuals aggregate into masses, probably under 
the chemotactic influence of an acrasin-like substance, but no true pseudo- 
plasmodium is formed. In simple types the aggregates are merely rounded 
mounds, but in forms like Chondromyces crocatus complex stalked and 
branching systems are formed. Here, as the rods move upward, piling on 
top of one another, the mass is constricted at the base and the layer of 
slime secreted by the advancing rods hardens into a stalk. The apical 
mass of cells continues to move upward and divides to form branches 
which culminate in multicellular cysts. In the production of these com- 
plex fruiting structures by the migration of individual cells, and in the 
specific character which these structures display, the Myxobacteriaceae 
resemble the Acrasiaceae, though the groups are not closely related. The 
same morphogenetic problem as to how a specifically formed structure is 
produced by independent and undifferentiated cells is presented by both 
groups of plants. 

A general account of the Myxobacteriaceae has been written by Quin- 
lan and Raper for Volume XV of the "Encyclopedia of Plant Physiology." 

Another case resembling these but involving much larger size and a 
higher level of organization is to be found in the development of the fruit- 
ing body in the fleshy fungi, such as the common mushroom, Agaricus 
campestris. The vegetative body here is a much branched mycelium which 

Differentiation 229 

absorbs food from the organic matter in the soil. When it has a good 
supply of this, and other conditions are also favorable, the mycelial mate- 
rial is mobilized into a rounded mass just below the surface of the ground. 
This develops into a "button" and then into the familiar mushroom fructi- 
fication, with its high degree of differentiation. This is composed not of 
a mass of cells that are attached in a firm tissue, as in the higher plants, 
but of a body of tangled hyphae, free to slide past each other. As the 
stalk increases in length, these hyphae tend to be oriented parallel to 
its axis, though at the base and in the pileus ("umbrella") they remain 
much tangled. Growth takes place primarily by elongation of the cells of 
the hyphae and is entirely at the expense of food already available in the 
mycelium. How, from such an apparent chaos of snarled threads, the 
very precisely formed fruiting body of the fungus grows and differen- 
tiates poses the same difficult problem as does the development of the 
sorocarp in the Acrasiaceae, and one lying at the heart of the morpho- 
genetic process. Bonner, Kane, and Levey (1956) have reexamined the 
development of Agaricus and confirmed and extended the results of the 
classic studies of de Bary, Atkinson, and Magnus, but the problem has 
attracted relatively little attention in recent years. The development of 
these fungus fruiting bodies, however, offers to the student of the prob- 
lems of differentiation and form determination some of the most promising 
material available for his work. 



In the preceding chapters there have been considered those morpho- 
genetic phenomena which manifest themselves in normal development. 
Polarity, symmetry, differentiation, and the wide variety of correlative 
manifestations evident as the plant and its parts progress from embryo to 
maturity are all indications of the orderly control of growth processes 
which is the visible aspect of biological organization. But the progress of 
development is not always unimpeded. Accident and injuries of various 
sorts may happen to the growing plant which remove a part of its tissues 
or divide it into two or more incomplete portions. In nature this may 
result from the attacks of fungi, insects, or higher animals; from mechani- 
cal injuries of many sorts; or from unfavorable conditions which impede 
the functions of its organs. The changes that follow such injuries, losses, 
or functional disturbances often throw much light on morphogenetic 
activities, and one of the most fruitful methods of studying developmental 
processes has been to observe the consequences which follow their ex- 
perimental disturbance. Indeed, this is the only way at present by which 
many of these processes can be investigated at all. 

This field of morphogenetic research is not as active today as it was in 
earlier years, and many of the most important papers in it go back to 
some decades ago. For a review of the earlier literature the student is 
referred to McCallum (1905), Kupfer (1907), and Goebel (1908). Ban- 
ning ( 1955 ) has discussed some of the recent work. 

An important fact which such studies reveal is that the organism shows 
a tendency to restore or replace parts that have been removed and thus 
to produce again a complete individual. To this general process, which 
includes a wide range of developmental phenomena, the term regenera- 
tion is commonly applied. Regenerative activities are much more common 
in plants, with their less highly organized bodies, than in animals and can 
often be subjected to a more complete developmental analysis. 

Regeneration can be brought about not only by the removal of a part 
but by isolating it physiologically from the rest of the plant. In a young 


Regeneration 231 

bean plant, for example, if the epicotyl is decapitated the buds in the 
axils of the cotyledons will grow into shoots, thus replacing the leafy 
shoot that would have grown from the epicotyl; but these buds may also 
be induced to grow by chilling a portion of the epicotyledonary stem and 
thus preventing the interchange of stimuli or substances between it and 
the tissues below ( Child and Bellamy, 1919 ) . 

In general, the more simple and undifferentiated a plant is, the more 
completely will it restore missing parts; and the more specialized and 
differentiated it is, the less regenerative capacity it will show. Early devel- 
opmental stages are thus more likely to regenerate readily than later ones, 
and groups lower in the phylogenetic scale than those of higher position. 
Ability to regenerate is often completely lost. 

The origin of regenerative powers in plants and animals is sometimes 
explained as the result of natural selection, much as in the case of other 
traits. That a long process of competition and selection conferred the ability 
to repair almost every type of injury seems unlikely. However this may 
be, the developmental activities in regeneration are not essentially dif- 
ferent in their origin and control from those which occur in normal de- 
velopment. In both, the production of a single, complete individual is the 
final result. The factors involved in regeneration seem neither more nor 
less difficult to understand than those in normal ontogeny, nor do they 
require a fundamentally different explanation. Both are manifestations of 
general developmental control, a fundamental self-regulation in the indi- 
vidual. Both seem to be the result of the same formative process. 

Begeneration is a term that has been variously defined. The author pro- 
poses to use it here in the broadest sense, as the tendency shown by a 
developing organism to restore any part of it which has been removed or 
physiologically isolated and thus to produce a complete whole. This 
covers processes from wound-healing to the reproduction of adventive 
structures and vegetative multiplication, in which many different activ- 
ities are involved. 

In general, the method of restoration of lost parts is different in plants 
and in animals. Animal tissues are composed of thin-walled cells, usually 
able to divide and often to migrate. Most plant cells at maturity are 
relatively thick-walled and ordinarily do not divide or grow further, 
though it has been shown that many retain the power to do so. Begenera- 
tion in animals, therefore, consists largely in a reorganization of the re- 
maining portion of the organ or body. In plants, on the contrary, this type 
of regeneration is limited to meristematic or rapidly growing regions or 
to the relatively rare cases where cells become embryonic again. Much 
more commonly, at least in the higher types, replacement of lost parts 
results from the growth of dormant buds or primordia or the development 
of new ones. Such primordia arise from cells in the plant body which 

232 The Phenomena of Morphogenesis 

remain alive and are potentially meristematic. These primordia, often very 
numerous, remain dormant under ordinary conditions. 

This fact raises the question of what it is, if these primordia are 
capable of growing and forming new organs, that prevents them from 
doing so. The concept of the organism as a balance between stimulatory 
and inhibitory factors suggests itself here, but the problem remains as to 
what localizes and correlates the activitv of these factors so that a specific 
organic system is established, maintained, and restored, and what stops 
the regenerative process when this has been accomplished. 

It should be remembered that, in spite of modifications acquired during 
the process of differentiation, all the cells are probably identical geneti- 
cally, save for occasional polyploidy. The potentially meristematic cells 
thus serve, so to speak, as a "germ plasm" or genetic reserve which can 
direct the processes of regeneration and further development. Each cell, 
at least in theory, is capable of producing an entire individual. 

In regeneration, mature or nearly mature cells may sometimes become 
embryonic again and then undergo changes that restore a disturbed tissue 
pattern. How this is accomplished is of much interest for morphogenesis. 
In it the first visible step is usually a marked increase in the amount of 
cytoplasm and in the size of the nucleus and an acceleration of metabolic 
activity. The wall also tends to become thinner. This process, commonly 
termed dedifferentiation, has been described and the literature reviewed 
by Buvat ( 1944, 1945, 1950 ) . Dedifferentiated cells assume the character 
of meristematic ones and can divide and grow. The tissue thus formed 
may then differentiate again in conformity with its new function or 
position in the regenerated structure. 


Among the thallophytes and bryophytes, with their simpler bodies and 
lower level of organization, regeneration is relatively common. It will be 
discussed briefly in these groups before taking up the more complex 
aspects of it that vascular plants display. 

In most of the lower plants single cells or groups of cells have the 
ability to develop readily into a whole plant, and in many instances they 
do so regularly as a means of vegetative reproduction. Extreme instances 
of this are the conversion of vegetative cells into specialized nonsexual 
reproductive cells such as zoospores and others. 

Even when not thus differentiated for reproduction, the cells of many 
algae are readily separable from the loosely organized thallus and grow 
into new plants, as in Callithamnion (Weide, 1938) and Cladophora 
(Schoser, 1956). In the last genus the cells may be separated from one 

Regeneration 233 

another by plasmolysis, and each then grows into a new plant. Among 
simpler fungi almost any adult cell on isolation will give rise to a new 
mycelium. Kerl (1937) found that single cells from the surface of 
Pi/ronema confluens would do this. Such instances could be multiplied 
almost indefinitely. 

Individual cells, if injured, will often restore themselves, especially large 
ones like those of Vaucheria ( Weissenbock, 1939 ) , Dasijcladus ( Figdor, 
1910), and Acetabutoria (Hammerling, 1936) or the coenocytes of Bryop- 
sis and Caulerpa ( Janse, 1910; Winkler, 1900; Dostal, 1926). In such algae 
as S-phacelaria, which grow by an apical cell, this cell may be replaced, 
if injured, by the cell next below it, which first undergoes considerable 
reorganization (Zimmermann, 1923). Other cases have been described. 
Hofler (1934) observed in the filamentous alga Griffithsia that, if a cell 
dies, the one above it will send a tube either through it or around it 
which makes connection with the cell below and thus restores the con- 
tinuity of the living filament. 

In the early development of certain animal embryos, if one of the first 
two blastomeres is killed, the other develops into an entire organism. A 
somewhat similar instance in plants occurs in Fucus. Here, after the fer- 
tilized egg has formed two cells, an apical and a basal (rhizoidal) one, 
the apical cell will produce a new rhizoid at the basal pole if the first is 
destroyed (Kniep, 1907). Setchell (1905) in his studies of the kelps de- 
scribes the way in which a stipe, if the blade is cut off from its tip, will 
regenerate a new one from the cut surface. This commonly happens in 
nature where these plants are buffeted by the waves, for there has devel- 
oped an intercalary meristem near the leaf base which becomes active 
when the blade is removed. Killian ( 1911 ) describes the way in which an 
injured stem is reconstituted in Laminaria digitata. 

Some of the most remarkable cases of regeneration occur in the fruiting 
bodies of the fleshy fungi— toadstools, mushrooms, bracket fungi, and 
similar types. These are formed from masses of tangled hyphae which do 
not adhere to their neighbors as do cells in higher plants but are merely 
packed together in a weft of tangled threads. Even so, they tend, if in- 
jured, to restore the missing portions and produce a normal sporophore. 
This has been observed in Stereum by Goebel (1908), in Agaricus by 
Magnus (1906), and in other fleshy fungi. Under favorable conditions 
almost any part of one of these fruiting bodies will restore portions of its 
tissues that are removed. Such structures provide promising material for 
studies in regeneration. Brefeld and Weir maintain that every cell of 
C&prinus has the potentiality of producing an entire sporophore. 

Among bryophytes, the hepatics regenerate with particular readiness. 
Early work with these plants has been reviewed by Correns ( 1899 ) . The 

234 The Pnenomena of Morphogenesis 

process here is a common means of reproduction. In Sphoerocarpos, 
Rickett (1920) found that regeneration occurs from single cells (or some- 
times groups of adjacent cells) from almost anywhere on the thallus. At 
first the mass of cells is globular, cylindrical, or ribbon-like but it soon 
develops into a typical thallus much as does a germinating spore. 

In a study of vegetative reproduction in Metzgeria, Evans (1910) ob- 
served that certain cells on the thallus dedifferentiate and that each then 
grows into a gemma from which a new plant arises. The distribution of 
these regenerative cells is not a random one, however. A robust thallus 
produces no gemmae, and they are fewer in plants that bear sex organs. 
If the apical region is very active, there are no gemmae near it. If a piece 
of thallus is isolated, however, gemmae arise in it abundantly. Evidently 
there are factors in this plant that tend to inhibit regeneration by its cells. 

Plantlets are frequently produced from single cells in the leaves of 
Jungermanniales, and here they often, though not always, develop much 
as spores do. It is sometimes difficult to tell whether they come from ordi- 
nary vegetative cells or from ones that are predisposed to produce them. 
Fulford ( 1944, 1954 ) has described many cases of reproductive regen- 
eration in these plants. 

Many mosses also regenerate readily. The early work here has been 
reviewed by Heald (1898). Protonemata and, from these, new plants arise 
on the stem of some mosses but rarelv from the leaves unless the latter 


are detached (Gemmell, 1953). Here they grow chiefly from the surface 
cells of the midrib. 

Morphogenetically, the most significant aspect of moss regeneration is 
that under appropriate conditions protonemata develop not only from 
the gametophyte but from sporophyte tissue, both seta and capsule, and 
thus are diploid. From these diploid gametophytes, tetraploid sporophytes 
can be produced. This possibility was first discovered by the Marchal 
brothers (1907-1911) and opened up a wide field for exploration. Its 
genetic and physiological aspects have been explored by F. von Wettstein 
(1924) and his students (p. 437). In several cases (as by Springer, 1935, 
with Phascwn ) sporogonia have been observed to develop directly and 
apogamously from diploid gametophytes without a sexual process. Still 
more remarkable, Bauer (1956) observed that diploid protonemata of 
another moss, Georgia pellucida, under certain conditions form buds 
which do not develop into leafy gametophytes, as ordinarily happens in 
such cases, but produce sporogonia directly. Spores in these develop 
rarely, but when they do they germinate into normal haploid protonemata. 

Regeneration of diploid gametophytes from sporophytes of Anthoceros 
was accomplished by Rink (1935) through cutting away portions of the 
sporophyte. Here the diploid thalli are smaller and more irregular in 
shape than the haploid ones. 

Regeneration 235 


Among vascular plants there is a much higher degree of differentiation 
than in lower forms and as a consequence the processes of regeneration 
are more complex. In these plants we may recognize, for convenience, 
three rather different types of regenerative activities. Reconstitution, or 
regeneration proper, includes those cases in which, as in animal regener- 
ation, there is a reorganization of the embryonic tissue by which its orig- 
inal structure is re-formed. This is usually limited to truly embryonic 
regions, such as growing points and young embryos, and to structures 
where there is a reorganization of the tissue pattern by dedifferentiation 
and subsequent redifferentiation. Restoration describes the wide range 
of cases where missing tissues or organs are replaced through meri- 
stematic activity arising in adjacent regions. This may result from the 
activation of dormant buds or primordia already present or in the forma- 
tion of new ones such as occurs in the origin of new roots and shoots in 
the familiar processes of vegetative propagation. Reproductive regener- 
ation, or vegetative reproduction, involves the separation, by natural 
means, of a part of the vegetative body from the rest and its establishment 
as a new plant, a process which often occurs in the lower groups. Similar 
cases are those where plantlets develop on the leaves and drop off to form 
new individuals. These are all specialized instances of the ability of the 
plant, under favorable conditions, to produce a new whole from a part 
of its body, an ability that comes from the totipotency of its various 


This process, the reorganization of living material by which the normal 
structure is restored when disturbed by outer circumstances, is relatively 
uncommon in plants, since truly embryonic conditions persist in them for 
only a relatively short time before changing into a mature state where 
reorganization is difficult. Such reconstitution as does occur is of two 
sorts. In one, truly embryonic or meristematic tissues may be reshaped 
into a new whole. In the other, tissue already well along toward maturity 
may undergo a certain degree of dedifferentiation and redifferentiation 
so that its structure is reorganized and the original pattern, at least in 
part, reconstituted. 

Meristematic Reconstitution. Among the simplest cases, and one which 
not infrequently occurs in nature, is cleavage polyembryony. In many 
conifers (Buchholz, 1926), though less commonly in angiosperms, the 
early embryo rudiment, carried down into the endosperm at the tip of 
the suspensors and still consisting of only a few cells, divides and de- 

236 The Phenomena of Morphogenesis 

velops into two or more parts (Fig. 9-1). These all grow for a time but 
usually only one survives and develops into the embryo of the seed. A 
portion of the original embryo thus reconstitutes a complete whole. This 
recalls the not infrequent cases among animals where one fertilized egg 
produces two or more individuals (as in identical twins) or where a 
single blastomere, experimentally isolated, will form a whole. 

Fig. 9-1. Cleavage polyembryony 
in Torreya. This group of young 
embryos have all come from a sin- 
gle fertilized egg by cleavage. 
(From Buchholz.) 

It is with the terminal meristems of the older plant axis, however, that 
most of the experimental work on regeneration has been carried on. In 
roots it is generally agreed by observers that if only the extreme tip is 
removed, about y 2 to % mm., a new growing point will regenerate di- 
rectly at the wound surface from the underlying tissue of the plerome. If 
a little more is cut off, regeneration is only partial and chiefly by the for- 
mation of new growing centers in the outer portion of the root. If still 
more is removed, true reconstitution ceases and a callus is formed with 

Regeneration 237 

adventitious roots growing out from it. The early papers of Prantl (1874) 
and Simon ( 1904 ) on angiosperms and of Stingl ( 1905 ) on gymnosperms 
present the basic facts, which have been reviewed by Nemec ( 1905 ) . 

More recently Torrey ( 1957Z? ) has studied the regeneration of decapi- 
tated roots grown in culture media and as affected by auxin. If an abun- 
dance of auxin was present, the vascular cylinder of the new roots was 
hexarch instead of the normal triarch. Such a root reverted to triarch 
again if returned to the usual medium. Torrey interprets these changes 
as due to the direct effect of auxin on the size of the meristematic tip, 
the structure of which evidently is not determined by the mature tissue 
farther back. 

If the young root is split lengthwise, scar tissue forms on the inner 
portion of the cut surface but each tip will become reorganized into a 
new and complete meristem and will finally reconstitute a normal root 
(Lopriore, 1892). This argues against the idea that there is a single apical 
cell in the root. Ball ( 1956 ) split the hypocotyl tip of a Ginkgo embryo 
and found that the effect of this was evident for some distance upward 
in the epicotyl in the differentiation there of a divided vascular cylinder. 

In the shoot meristem the situation is complicated by the presence of 
leaf primordia. Only the terminal dome, about 80 /x back of the actual tip, 
will be regenerated if it is removed. The earlier workers believed that a 
new apex was formed here, as in the root, by direct growth from the cut 
surface, and more recently Mirskaja ( 1929 ) has reported that this occurs 
in Tradescantia. Most observers, however, have found that scar tissue 
forms over the wound and that one or more new meristems arise at the 
edges of this. 

Several plants have marked powers of meristematic reconstitution. In 
the much reduced aquatic Podostemon ceratophyllum, if the tip of the 
shoot is cut off a new one arises from a group of cells just back of the 
cut surface in or around a vascular bundle. A decapitated root is recon- 
stituted in much the same way (Hammond, 1936). In Zamio, a new shoot 
will often grow out directly from the stump of an old one, usuallv from 
the region of the central cylinder (Coulter and Chrysler, 1904). Such 
simple and direct regeneration in plants is rare. 

Karzel ( 1924 ) and others split growing shoot tips and found that each 
half regenerated more or less completely, depending on the species. 
Pilkington (1929) split simply the terminal meristem itself and observed 
the same result. From the tissue of dodder which remains within the host 
plant after the external portion of the parasite has been experimentally 
removed, Truscott (1958) observed the regeneration of a shoot meristem 
which pushed out through the surface and developed into a normal 
dodder shoot. 

Much experimental work on the regeneration of the shoot apex has 

238 The Phenomena of Morphogenesis 

been done by Ball in Lupinus and Tropaeolum and by Wardlaw in 
Primula and several ferns. Ball ( 1948, 1952« ) went still further than 
Pilkington and was able to split the meristematic apex into four, and later 
even into six, strips ( each still connected basally with the axis ) and found 
that each was able to reconstitute a whole shoot unless its tip had been 
reduced below a minimal size. Vascular tissue tended to be poorly devel- 
oped in it until leaves were formed by the new shoot. 

Both Ball ( 1952/? ) and Wardlaw ( 1950 ) isolated the central core of the 
shoot meristem from the rest of the axis by three or four longitudinal in- 
cisions (Fig. 4-14), thus leaving the meristematic dome (or the apical 
cell and its neighbors) connected with the vascular tissues below only 
by a plug of pith tissue. Both investigators found that this isolated tip 
continued to grow and in time produced procambial tissue independent 
of that in the axis below. In the flowering plants studied this differen- 
tiated basipetally and finally joined the vascular system of the main axis. 
In the ferns, however, it did not do so. At the apex of the isolated core 
new primordia were formed and (except in ferns) if this core was not 
below a minimal size, it finally developed into a normal leafv shoot. 
Wardlaw observed that the phyllotaxy of the new shoot in Primula was 
continuous with that of the original axis but in Ball's material it was inde- 
pendent and often showed reversal of the earlier spiral. Wardlaw ob- 
served that the vascular tissue developing in the central core followed the 
outline of the cut piece and differentiated at a rather constant distance 
from the cut surface, suggesting that its position was dependent on a 
gradient of some sort (Fig. 9-2). 

These experiments are of morphogenetic interest since they show that 
the apical meristem is a self-determining region which can produce a nor- 
mal shoot without any connection, other than through undifferentiated 
pith cells, with the tissues below. Furthermore, Ball (1946) and others 
have shown that small meristem tips, growing in tissue culture, will pro- 
duce entire plants. All this is not surprising, however, since many cases 
are known where a single cell (p. 253), without any vascular connection 
with other tissues at first, develops into a fully differentiated plant. It is 
to be expected that an active terminal meristem would do the same. Al- 
though the differentiated tissues below the meristem may not be necessary 
for its growth, they contribute to the character of its development, for 
through them come water, nutrients, and various morphogenetically im- 
portant substances. The terminal meristem seems neither to be completely 
self-determining nor completely under the control of the rest of the plant, 
but the two act together as an integrated system. 

Reconstitution of parts other than the meristems of the axis has often 
been reported. In a number of ferns such as Gleichenia, the leaf grows 
at the tip by a terminal meristem, thus perhaps harking back to the time 

Regeneration 239 

when the leaf was a lateral branch of indeterminate growth. In most fern 
leaves, unlike those of angiosperms, growth continues longest at the tip 
and in some species the leaves root readily there. It is thus to be expected 
that fern leaves should sometimes reconstitute new tips if the old ones 
are injured or removed. Goebel (1908) with Polypodium and Figdor 
( 1906) with Scolopendrium cut away a small bit of tissue from the grow- 
ing leaf tip and found that this meristematic region was reconstituted 
but that a double or forked lamina was produced. 

Fig. 9-2. Diagram of transverse section of apex of Primula in which a central plug of 
pith was isolated by four vertical incisions ( broken lines ) . In this plug a new vascular 
cylinder has been regenerated at a constant distance from the surface. (From Ward- 

Among higher plants, leaves of some of the Gesneriaceae are easily re- 
generated. In Saintpaulia and some species of Streptocarpus, Figdor 
( 1907 ) split a young leaf nearly to the base along the midrib and found 
that a considerable amount of new lamina was regenerated from the 
basal part of the cut surface, which is the latest to mature. If the blade is 
removed from a young primary leaf of Cyclamen, two new blades re- 
generate by reconstitution of the meristematic region at the sides of the 
petiole apex (Winkler, 1902). Figdor also reported (1926) that if the 

240 The Phenomena of Morphogenesis 

terminal leaflet of the pinnately compound leaf of Bryophyllum (Kalan- 
choe ) was cut off when the leaf was very small, it was partially regener- 
ated. Other cases of such regeneration have been reported, but the leaf, 
doubtless because of its generally determinate growth, shows meri- 
stematic reconstitution much less readily than does root or stem. 

Gametophytes sometimes display this type of regeneration. Albaum 
(1938rt), confirming earlier workers, found that if the ordinary prothal- 
lium of certain ferns is cut transversely the anterior portion, near the 
meristematic notch, will re-form the typical heart-shaped structure again. 
In the posterior region this does not happen, but new adventive pro- 
thallia are produced from the cut surfaces (p. 121). Meyer ( 1953) reports 
that the meristematic region in the notch of the prothallus, and particularly 
its apical cell, seems to inhibit the formation of other apical cells; for if 
the prothallus is divided lengthwise into three parts, the central one, 
which includes the apical cell, will regenerate its lost portions whereas 
the two lateral pieces will each first produce a new apical cell and then 
proceed to develop into typical prothallia. True reconstitution of a meri- 
stematic region thus seems to be limited to very early developmental 
stages. In later ones, even though the tissue may still be meristematic, it 
has lost some of its morphogenetic potencies and injury will result either 
in simply a wound reaction or in the production of new adventitious 
growing points rather than in a remolding of the old one. 

Reconstitution of Tissue Patterns. A somewhat different type of recon- 
stitution occurs where the structure that is removed is not a meristem but 
a part already differentiated, at least to some degree. To restore the dis- 
turbed tissue pattern involves a more complex process and is rarely as 
complete as the reconstitution that occurs in a meristem. Where mature 
or nearly mature cells are part of this pattern, some of these cells must 
evidently become embryonic again and assume a new function in the 
reorganized system. There are some remarkable examples of this which 
provide particularly interesting morphogenetic problems. 

The first step in such a reconstitution is healing of the wound itself. 
Wound reactions differ with the type of plant and the conditions. Cells 
near the wound surface generally become more active and soon, under 
the influence of wound hormones, cell division is initiated parallel to the 
surface. A phellogen here develops which forms a layer of protective cork 
over the wound. In many cases this is all that happens. Sometimes a callus 
is formed here and from it primordia of roots and shoots may develop. 
The phenomena of wound reactions have been reviewed by Bloch ( 1941, 
1952). Fourcroy (1938) has discussed the same subject and particularly 
emphasizes the accelerating influence of wounding on differentiation and 
its effects on vascular anatomy. 

In many plants, however, the tissues under the wound may be reorgan- 

Regeneration 241 

ized to some degree. This is especially true of the epidermis. In kohlrabi 
Vochting (1908) observed that when the layer of wound cork was 
sloughed off an epidermis had developed under it which was essentially 
identical with that of the normal tuber and in which typical stomata 
were present. Cells destined to be cortical in character had been radically 
altered to form a tissue appropriate for their new position in the system. 

Fig. 9-3. Vascular strands a and b regenerating in parenchymatous tissue to connect 
strand G with others; c, an earlier stage in this process, where procambial strands are 
being differentiated. (From S. Simon.) 

Such reconstitution has been reported by others. The cuticle of epidermal 
cells may also be regenerated if it has been sliced off ( Fritz, 1935 ) . 

More deeply seated tissues may be regenerated by the ^differentiation 
of others, as is the exodermis in air roots of orchids ( Kiister, 1899; Bloch, 
1926, 1935b) and hypodermal sclerenchyma (p. 218) in air roots of palms 
and Araceae (Bloch, 1937, 1944). In every case these newly developed 

242 The Phenomena of Morphogenesis 

tissues are appropriate in character for the place where they now are and 
are much like those normally present in such regions. 

Somewhat more complex is the redifferentiation of vascular bundles 
in places where these have been severed. Simon (1908; Fig. 9-3), Nemec 
(1905), Sinnott and Bloch (1945), and Jacobs (1952) have studied this 
in various herbaceous stems. If a bundle or group of bundles is severed 
by a lateral incision and the region examined in longitudinal section after 
a week or two, a new vascular strand can be seen developing behind the 
incision and connecting the severed upper and lower ends of the bundle 
( p. 193 ) . This strand is formed by the conversion of large, squarish paren- 
chyma cells of the pith into xylem cells with reticulate lignified thicken- 

Fig. 9-4. Left, regeneration of connection between severed vascular bundles in stem 
of Coleus. Right, differentiation of parenchyma cells into reticulate xylem cells in the 
development of this strand. Arrow shows the direction of its development. Note 
new walls parallel to it. ( From Sinnott and Bloch. ) 

ings (p. 194 and Fig. 9-4). Differentiation seems always to be basipetal, 
suggesting the downward passage of a morphogenetic substance, pre- 
sumably auxin, from the upper bundle toward the lower. The regener- 
ated strand is not directly at the cut surface but about the same distance 
behind it that the normal bundles are from the uninjured surface. Jost 
( 1942) caused the plant to form these vascular bridges in manv ways and 
finds that, although their general course is basipetal, they may develop 
acropetally for a while in passing around an obstacle. They do not always 
take the shortest route. The position of the strand seems to be determined 
by a gradient of some sort from the wound surface inward. It is significant 
that the conversion of parenchyma to xylem is an "all or none" reaction, 
with no cells intermediate in character. 

Regeneration 243 

A somewhat similar regeneration of vascular strands is to be seen in 
leaves where some of the veins have been cut. New strands are differen- 
tiated in the mesophyll cells which connect the severed ends. Freundlich 
( 1908 ) studied the origin of these xylem strands, and Kaan-Albest ( 1934 ) 
followed the differentiation of new sieve tubes (Fig. 9-5). The latter do 
not arise, as do those of the xylem, by conversion of whole parenchyma 
cells, but small cells are cut out of the sides of these larger elements and 
join up with one another, end to end, from one cell to the next. These 

Fig. 9-5. Impatiens. Sieve-tube connections developing between phloem bundles, one 
of which has been severed. ( From Kaan-Albest. ) 

phloem strands in their development suggest the fiber strands of Luffa 
(p. 197). 

What are the factors, one may ask, that impel the ^differentiation of 
a vascular system when intercommunication among its parts has been 
interrupted? Auxin may be diffusing from the end of a cut bundle, but 
how this operates to convert a series of parenchyma cells into a vascular 
strand is difficult to understand. Here is differentiation in very simple ex- 
pression. Doubtless the same general factors are involved as in normal 
development but the process takes place here on a greatly enlarged scale 
where it can be more readily studied than in the very small-celled tissues 

244 The Phenomena of Morphogenesis 

near the meristem. Regenerative tissues of this sort are particularly prom- 
ising material for a study of cytological differentiation. 

These are examples of the relatively simple reconstitution of a tissue 
pattern. A much more involved one is that described by Vochting ( 1908 ) 
in his classic studies of regeneration in kohlrabi. If a young and growing 
tuber of this plant was cut transversely at about half way from the tip to 
base, care being taken not to injure the leaves on the basal portion, the 
cut surface of this portion soon began to swell and by the end of the sea- 
son had developed a convex, rounded structure which in some cases 
restored the general form of the normal tuber except that no leaves de- 
veloped on its surface. Internally, however, the complex system of bundle 
connections in the reconstituted half was hardly to be distinguished from 
that of the original portion. This provides the most remarkable example 
so far described of a structure already well differentiated internally which 
proceeded to reorganize itself and reconstitute, in almost its original form, 
a large mass of tissue. There was meristematic activity here, following 
the dedifferentiation of much of the structure near the cut surface, but it 
was diffuse meristematic activity like that of a leaf or fruit rather than 
that of a localized growing point. Studies on regeneration in this very 
promising material have been largely neglected in the half century since 
Vochting's description of his work was published. 


Most regeneration in plants is not due to the reorganization of em- 
bryonic regions but to the onset of meristematic activity in regions ad- 
jacent to the place where loss has occurred. This leads to the production 
of substitute structures that restore the original whole by indirect means. 

These processes are examples of compensatory correlation (p. 98) in 
which the balance of the organism is restored after being disturbed. As 
redifferentiation after injury often throws light on the problems of differ- 
entiation and tissue pattern, a study of the restoration of lost structures 
offers a useful means of analysis of developmental potencies and the proc- 
esses of correlation. 

In many cases, as the result of wounding, a callus is produced (p. 288) 
at the cut end of a stem or root. This often originates from the cambium 
but may come from other tissues. From such calluses primordia of roots 
and shoots commonly arise. This is the most frequent type of regeneration 
in the higher plants and underlies the horticultural arts of multiplication 
by vegetative propagation, chiefly the rooting of cuttings. This field has 
been reviewed from the botanical point of view by Priestley and Swingle 
(1929) and Swingle (1940, 1952). 

Almost every plant organ has been used as a cutting— stem, root, leaf, 
hypocotyl, floral axis, and flowers— and all have been found to have some 

Regeneration 245 

ability to restore lost parts. The axial organs-stem and root-are the ones 
most commonly employed in the practices of propagation and have been 
most thoroughly studied. 

Stem Cuttings. In stem cuttings of dicotyledons buds develop most fre- 
quently at the apical end and roots from the basal one (p. 119) but this 
polarity varies considerably. The buds may be the usual axillary ones, 
many of which would not normally develop, or they may be accessory 
buds. If these are absent, dormant primordia may grow. Carlson (1950) 
has described the origin and distribution of dormant root initials on 
willow shoots. Primordia may also develop anew, from callus or from 
the normal tissues of the stem. Adventitious roots in young stems usually 
come from the pericycle but in older ones they may have a deeper origin 
in the vascular cambium (Plett, 1921). Mahlstede and Watson (1952) 
found that adventitious roots in blueberry originate in cambium or phloem 
and push out through vascular tissue, cortex, and epidermis. Priestley 
( 1926a ) stated a general rule that, of the two lateral meristematic regions 
of the axis, the phellogen is more likely to produce buds and the vascular 
cambium to produce roots. Morphogenetic problems here involved con- 
cern the causes of the differentiation of dormant or "reserve" primordia 
in particular places and especially the factors that first keep them dormant 
and then stimulate their development in regeneration. 

Bud formation is frequent on hypocotyls and has been studied particu- 
larly bv Rauh ( 1937 ) . In a few species these buds normally develop into 
shoots. In other cases they may be present but do not develop and in still 
others they may be induced only by the stimulus of regeneration, after 
the decapitation of the hypocotyl. In Linum usitatissimum, the origin of 
these buds has been traced, in decapitated hypocotyls, to single cells of 
the epidermis ( Crooks, 1933; Link and Eggers, 1946a ) in which divisions 
begin to appear. A group of cells is thus produced which develops into a 
bud initial and finally into a shoot. Several buds may begin to grow, only 
one of which becomes dominant. In undecapitated hypocotyls a few epi- 
dermal cells may divide but they rarely produce buds. Bud development 
is induced more readily in young hypocotyls than in older ones. After a 
bud begins to grow, vascular strands differentiate which connect it with 
the main vascular cylinder (Fig. 9-6). Van Tieghem (1887) described 
similar bud development in the hypocotyl of Linaria, as did Bain ( 1940 ) 
in cranberry. Such hypocotyls offer a good opportunity for the study of 
cellular totipotency and the redifferentiation of vascular tissue. 

So-called "adventitious leaves" (really reduced shoots, Rauh believes) 
are produced abundantly on the decapitated hypocotyl, or seedling tuber, 
of Cyclamen (Boodle, 1920; Rauh, 1937) and develop there from sub- 
epidermal cells. There are transitions here from simple leaves to fully 
developed buds. The great number of these buds normally produced sug- 

246 The Phenomena of Morphogenesis 

gests that this is a case of reproductive regeneration rather than of 

The stems of most monocotyledons lack a cambium, and this somewhat 
limits the possibilities of regeneration of lost parts in them. Axillary buds 
are often present, however, and the nodes and bases of leaves remain 
somewhat meristematic. New roots commonly arise in these regions. Meth- 
ods of regeneration and of vegetative reproduction are generally very 
specialized. Some members of the Juncaceae and Cyperaceae reproduce 
vegetatively from tips or nodes of culms, as in Eleocharis rostellata, where 
buds grow from the sterile culm tips ( La Rue, 1935 ) . 

Fig. 9-6. A young adventitious bud which has grown from epidermal cells on a de- 
capitated hypocotyl of flax. Note the cell divisions in the cortex which will give rise to 
a vascular connection with the stele. (From Crooks.) 

Factors in the rooting of conifer cuttings have been discussed by 
Deuber (1940). 

The physiological basis of such regenerative processes has been widely 
investigated. Gardner ( 1929 ) observed that, in both deciduous and ever- 
green trees, cuttings from 1-year-old stems rooted more readily than those 
from older ones. The influence of auxin and of various synthetic growth 
substances has been much emphasized ( p. 391 ), particularly in root forma- 
tion. Van der Lek (1925) found that cuttings rooted better if there were 
buds on them, presumably because of the production of root-stimulating 
substances by the buds. Discovery of the effectiveness of various organic 

Regeneration 247 

acids in root formation has made "root hormones" of importance in horti- 

Bud formation, also, is influenced by physiological conditions. Miller 
and Skoog ( 1953 ) report that tobacco-stem segments, in sterile culture, 
form buds much more readily if adenine is present and that indoleacetic 
acid reduces their development. Both results, presumably, are due to 
effects on nucleic acid metabolism. Ruge (1952) and a number of earlier 
workers observed that a functioning chlorophyll apparatus is necessary 
for the successful rooting of most cuttings. Shoots with variegated leaves 
or in the dark root poorly. Whether this is owing to the production by 
the leaves of food or of a growth substance is not clear, but van Overbeek, 
Gordon, and Gregory ( 1946fo ) believe that the main function of leaves in 
the rooting of cuttings is simply to supply nutrition. 

Hereditary tendencies may also influence the character of regeneration, 
as in the conversion of axillary buds into tubers in the potato (Isbell, 


Root Cuttings. Under natural conditions shoots are produced by roots 
rather infrequently and chiefly in woody plants. In many cases roots may 
be used as cuttings, however. Here the restoration of lost structures by 
the production of adventitious roots or shoots occurs much as in the stem, 
shoots tending to be restored at the basal (proximal) end and roots at 
the apical ( distal ) one. Most growth is from callus. Roots show a greater 
tendency to form adventitious buds than do stems. Naylor ( 1941 ) finds 
that both structures arise from meristematic tissue produced by paren- 
chyma cells in the younger phloem and not from the cambium. The polar 
development of regenerating structures on fleshy roots has been studied 
by various workers (p. 124). 

Way ( 1954 ) investigated regeneration on apple roots of different sizes. 
In some varieties the larger ones ( 8 to 12 mm. in diameter ) produced only 
shoots and the smaller (3 mm.) only roots. When both were formed, the 
zone of shoot production ( at the proximal end ) extended farther distally 
in the wider roots, and that of root production (at the distal end) ex- 
tended farther proximally in the narrower ones. Way interprets these dif- 
ferences as due to auxin gradients, with different concentrations in large 
and small roots. 

Buds on roots are usually endogenous in origin. In Bryophylhim they 
arise from the subepidermal layer (Ossenbeck, 1927), and Rauh finds 
them originating at the scars of the delicate branch roots. In Aristolochia 
and the Podostemaceae they grow from the cortex. Carlson ( 1938 ) reports 
that in the orchid Pogonia an adventitious shoot arises by enlargement 
and division of the surface and cortical cells at the tips of lateral roots. 
This forms a swelling in which a bud develops. 

An important morphogenetic question concerned with regeneration 

248 The Phenomena of Morphogenesis 

from roots and one which has been rather widely discussed is whether the 
early primordia of new organs are "indifferent" in nature and may pro- 
duce either buds or roots, or whether they are determined from the first 
to form one or the other. A bud is a young shoot and has the rudiments of 
leaves, which the root primordium does not have, and it should therefore 
presumably be easy to tell one from the other. This seems often to be 
difficult, however, in the very early stages, and many instances where 
root primordia have been reported to change into shoots, and vice versa, 
may be due to erroneous observation. Dore (1955) has studied the origin 
of young primordia in the regeneration of horseradish roots and finds that 
these are produced as organized meristems in close association with the 
scars of lateral roots and that they originate in the phellogen of the main 
root. He is certain that at the beginning they are capable of developing 
into either roots or shoots. That this is so is suggested by the fact that the 
ratio of buds to roots, where they can be definitely distinguished, is not 
constant but varies with conditions, as though neutral primordia were 
being tipped in one direction or the other. If this should finally prove to 
be the case and the existence of truly indifferent primordia be established, 
useful material would be provided in which to examine the very early 
stages in the differentiation of these two structures which soon become so 
dissimilar. This reminds one of the case in Selaginella reported by Wil- 
liams (1937). Here an "angle meristem" near the tip will normally pro- 
duce a rhizophore, a structure somewhat intermediate between root and 
shoot and which finally grows downward and forms typical roots. If the 
main axis of the shoot is decapitated, however, the young primordium 
which would have produced a rhizophore will now grow upward into a 
typical shoot. Mention has been made earlier (p. 71 ) of cases where a 
young leaf primordium, if isolated by deep cuts from the meristem tip, 
will grow into a bud-like structure. 

Cuttings from Other Parts. A variety of phenomena of regeneration has 
been described in typical leaves and in cotyledons, scales, and carpels, as 
well as in inflorescences, flowers, and fruits. In contrast to the axial portion 
of the plant, leaves are organs of determinate growth, and the restoration 
of lost parts by them is thus somewhat different from the process in 
axial structures. It is especially common in succulent leaves. 

The restoration of roots or shoots on leaves detached from the plant 
and with petiole placed in soil has often been observed ( Stingl, 1908; and 
Yarwood, 1946). In such cases, after adventive roots and buds have been 
formed on lamina or petiole, various anatomical changes may be observed, 
especially a marked increase in the vascular tissue of the petiole. Further- 
more, instead of being disposed in an arc, as in normal petioles, this tissue 
often enlarges to form a complete vascular ring. The petiole thus becomes 
structurally as well as functionally a stem. Winkler ( 1907o) reviewed cases 

Regeneration 249 

of such conversion and studied a particularly good example of it in 
Torenia asiatica. He believed that increased transpiration is the cause of 
the change. Simon (1929) found essentially the same thing in Begonia 
and noted that the bundles from the young roots induced new vascular 
tissue only in that part of the petiole just below them, suggesting the 
polar distribution of a hormone. Similar results were obtained by Doyle 
(1915) through grafting buds onto the petioles of rooted leaves. 

In leaf cuttings, adventitious structures are formed predominantly at 
the leaf base. Hagemann (1931), in an extensive survey of the regener- 
ative ability of leaves, investigated 1,204 species of gymnosperms, dicoty- 

Fig. 9-7. Petals of Epilobium angustifolium which produced roots when cultured on 
nutrient agar. ( From La Rue. ) 

ledons, and monocotyledons. He found that some of these showed no 
restoration, a very few formed shoots or both roots and shoots, and the 
largest proportion roots only. Schwarz ( 1933 ) examined many other spe- 
cies. The location of the regenerated structures is determined mainly by 
the anatomy of the leaf. It is noteworthy that in a number of species, pre- 
dominantly though not always succulents, restoration and vegetative 
reproduction occur in other parts of the leaf than the base, as in species 
of Drosera, Achimenes, Begonia, Torenia, and the Crassulaceae. In Utric- 
ularia Goebel (1908) found adventive shoots formed by the leaf tips. 

Restoration of organs may also occur from isolated cotyledons. Kiister 
( 1903fo ) obtained both roots and shoots from cotyledons of Cucumis and 

250 The Phenomena of Morphogenesis 

Luffa, Kowalewska ( 1927 ) shoots from Phaseolus and Visum, and Carlson 
(1953) only roots from Raphanus and Brassica. La Rue (1933) reviewed 
work on this subject and reports his own success in obtaining roots on 
excised cotyledons of 19 species and shoots on those of 22 species. 

Several investigators have studied regeneration in inflorescences. If this 
structure is cut off and treated as a cutting, root formation and subse- 
quent vegetative development of the inflorescence often follow. Bor- 
mann (1939) reviewed the literature and made extensive investigations 
himself, finding that, of 391 species studied in 65 genera and 45 families, 
the conversion of an inflorescence into a vegetative shoot by treating it 
as a cutting occurred in about 17 per cent of all the species. 

Flower cuttings of Cactaceae, where the stem is incorporated into the 
fruit, have been found to form roots and develop dormant buds ( Goebel, 
1908 ) . The ovary of Jussiaea as well as immature fruits of Lecythis react 
in the same way. Carriere (1877) describes the rooting of the capsule of 
Lilium speciosum and Kupfer (1907) that of pods of Plwseolus. 

La Rue (1942) found that, under favorable cultural conditions that 
provide both moisture and food, many flowers or their parts may be 
induced to root ( Fig. 9-7 ) , and by this means he obtained roots on flow- 
ers of three genera of monocotyledons and 22 genera of dicotyledons. He 
went further and was able to induce regeneration even in gametophytes. 
Female gametophytes of Zamia in sterile culture not only increased in 
size markedly but in a few cases developed small roots and buds. The 
latter produced leaves resembling miniature seedling ones. He later 
obtained similar results with Ctjcas (1954). 

Reproductive Regeneration 

The ability of a part of the plant to restore missing structures and 
thus regenerate a whole is essentially the ability to reproduce. Regenera- 
tion is a reproductive process, and it is understandable that during the 
course of evolution many plants should have developed means to use 
the totipotency of their individual cells and tissues as means for vege- 
tative reproduction. In many cases this has become a normal and spon- 
taneous process, as in the formation of foliar embryos on the leaves of 
many Crassulaceae or of bulbils in other forms. Isolation or injury may 
stimulate the growth of similar structures, and this process thus grades 
over into regeneration. In many cases it is difficult to distinguish be- 
tween the two. 

This type of reproduction has been described most commonly in 
leaves and leaf cuttings (Fig. 9-8). Many cases have been studied in 
both dicotyledons and monocotyledons. Goebel cites a variety of these 
from the earlier literature. In some cases plantlets occur naturally on 
leaves and drop off to form new individuals (gemmipary). In others 

Regeneration 251 

these appear only when the leaf is removed from the plant or its vigor 
reduced. Instead of actual plantlets, bulbils or bulblets may be formed, 
modified buds which drop from the plant and produce new individuals. 
Many plantlets develop from preformed meristematic cells or cell 
aggregates and thus are clearly to be regarded as reproductive struc- 
tures even though in some cases they are induced only by rather ab- 
normal conditions. Others arise from unspecialized cells, usually epi- 
dermal or subepidermal ones, much as do the shoots on the hypocotyls 
previously described. When these are frequently formed in nature they 
are usually to be regarded as reproductive rather than regenerative struc- 
tures. Only a few typical examples can be mentioned here. 

Fig. 9-8. Leaves of Achimenes, used as cuttings, regenerating roots and bulbils from 
the base and producing plantlets where veins have been cut. ( From Goebel. ) 

A familiar one is that of Tolmiea menziesii (Yarbrough, 1936a), in 
which a plantlet regularly is formed at the junction of petiole and blade 
from a preformed bud at that point. This readily separates from the 
parent plant and forms a new one. In Cardamine pratensis (Goebel, 1908) 
adventitious shoots or plantlets grow in the autumn or under special 
conditions from the axils of the leaflets by the activity of groups of 
meristematic cells. At the junctions of the larger veins occur slight 
swellings and these may also develop into plantlets (Fig. 9-9). In such 
forms there are evidently many cells that can easily be induced to become 
meristematic and form plantlets. How such cells differ physiologically 
from others it is important to discover. 

252 The Phenomena of Morphogenesis 

Some species of Drosera may also readily be induced to form shoots on 
their leaves (Behre, 1929). These develop from single epidermal cells 
on the morphologically upper side of the lamina, at the base of a tentacle, 
on the petiole adjacent to stomata or trichomes, or in young inflorescences 
adjacent to glands. In Drosera binata, a species with linear leaves, the 
young plants thus produced have roundish blades much like those of our 

Fig. 9-9. Cardamine pratensis. After a leaf is removed, plantlets develop on it from 
preformed embryonic areas (a). ( From Goebel. ) 

own common Drosera rotundifolia. It would be interesting to find 
whether this fact has any phylogenetic implications. 

In Begonia rex and some other varieties of begonia, shoots may be 
induced very readily on petiole and blade by removing the growing 
points of the shoot. Prevot (1938, 1939) found that these arose from 
epidermal cells, but only after they had reached a certain stage of ma- 
turity. Hartsema ( 1926 ) has described the changes that such cells undergo 

Regeneration 253 

at the start of shoot development ( strong protoplasmic streaming, migra- 
tion of the nucleus, and increase in cytoplasm). Here and in a good 
many other plants, shoots will appear on the blade if one or more of the 
veins are cut (Fig. 9-10), thus perhaps preventing the access of in- 
hibiting substances. Prevot was able to induce bud formation on begonia 
leaves by the application of various substances and by growing the 
plants in the absence of oxygen. He also found that strips of epidermis 
removed from the leaf would form buds. Not all begonias have high re- 
generative ability, and this seems to be an inherited character when 
various types are crossed. 

Much like these cases is the development of young plants on the leaves 
of Saintpaulia ionantha. This species is often reproduced in cultivation 
by plantlets formed on leaves that have been cut off and placed in a 
humid atmosphere. From individual cells of the upper epidermis shoots 

Fig. 9-10. Propagation of rex begonia. If a leaf is removed and placed on moist sand 
and certain of its major veins severed, plantlets will regenerate at these cuts. (From 
Avery and Johnson.) 

develop, and roots originate from parenchymatous cells near the boundary 
between xylem and phloem in the veins ( Naylor and Johnson, 1937; Fig. 

Torenia asiatica behaves in much the same way (Winkler, 1903). In 
leaf cuttings, numerous shoot primordia begin to develop over the sur- 
face of the blade, each from an epidermal cell above a vein. Only a few 
of these primordia grow into shoots, and a single leaf thus shows a 
wide range of stages in shoot development. The shoots that form come 
to flowering very quickly, sometimes when they have only one well- 
developed leaf, and should thus make excellent material for a study of 
the factors that induce flowering. 

Many members of the family Gesneriaceae, to which these plants be- 
long, regenerate readily. Leaf cuttings of Achimenes produce clusters of 
bulblets at the base of the petiole and shoots from the blade if the veins 
are severed ( Doposcheg-Uhlar, 1911). Streptocarpus (Goebel, 1908) illus- 
trates most types of regeneration. 

254 The Phenomena of Morphogenesis 

The most familiar examples of reproductive regeneration are provided 
by members of the Crassulaceae. In Kalanchoe pinnatum ( Bryophyllum 
calycinum), plantlets develop from the marginal notches of the fleshy 
leaves. This may sometimes take place while the leaf is attached to the 
plant but is more common after it has fallen to the ground. In each 
notch is a preformed foliar embryo (Fig. 9-12), long ago described by 
Berge (1877) and more recently by Yarbrough (1932) and Naylor 
( 1932 ) . This is more than a mass of meristematic cells, for it has already 
taken the first steps toward organization of a plantlet and shows the 
minute beginnings of root, stem, and leaf. In other species the degree 

Fig. 9-11. Development of an adventitious bud from cells of the leaf epidermis of 
Saintpaulia. ( From Naylor and Johnson. ) 

of differentiation of the foliar embryos varies (Stoudt, 1938). In K. 
daigremontana and K. tubiflora the plantlets attain appreciable size be- 
fore the parent leaf has reached maturity. In K. rotundifolia there is a 
residual meristem on the axial surface of the petiole which develops a bud, 
but root primordia do not become differentiated until after the leaf has 
fallen from the plant. 

The factors that induce the foliar embryos to develop into plantlets 
have been actively discussed. Loeb (1920), who made an extensive study 
of regeneration in this genus, believed that a hormonal mechanism in- 
hibited their growth as long as the leaf was attached to the plant. Reed 
(1923) attributed their behavior to the metabolic condition of the leaf 
and showed that they tend to grow if the lamina loses vigor. Ossenbeck 

Regeneration 255 

( 1927 ) regards the mechanical or physiological conduction between the 
leaf and the growing points of stem and root as important factors in 
inducing their development. Mehrlich ( 1931 ) exposed plants to a wide 
range of environmental factors in an attempt to solve the problem. In the 
activity of certain enzymes and in the relative amounts of carbohydrates, 
he noticed a difference between leaves in which the foliar embryos grew 
out and those where they did not. Varietal differences were also evi- 
dent. Gotz (1953) finds that plantlets grow out readily in long days 
but that short days tend to inhibit them. They are accelerated if the 

Fig. 9-12. Section through a notch on 
the leaf of Kalanchoe, showing a pre- 
formed "foliar embryo" with two leaf 
primordia below and a root primor- 
dium above, buried in the tissue. 
(From E. E. Naylor.) 

vessels between lamina and petiole are cut or if auxin levels are low 
(Vardar and Acarer, 1957). 

Other Crassulaceae differ in some respects from Kalanchoe. In Byrnesia 
weinbergii, Stoudt (1934) found that the foliar meristem from which a 
plantlet comes is at the base of the blade, not the margin. It is quite 
undifferentiated and will not develop until the leaf is removed from the 
plant. In Sedum (Yarbrough, 1936c) there are no preformed meristems 
at all, but these arise after the leaf is detached. 

In Crassula multicava (McVeigh, 1938) a still different condition oc- 
curs and one reminiscent of Begonia and Saintpaulia, for here, after a 
leaf has been kept for a time in a moist chamber, plantlets begin to de- 
velop. Entire plants— not the shoot alone, as in many other cases— have 

256 The Phenomena of Morphogenesis 

their origin in single epidermal cells. These cells are not only presumably 
but demonstrably totipotent. 

In a considerable number of forms the reproductive structures arising 
in regeneration are not plantlets, already differentiated and ready to 
start growth, but dormant, almost seed-like structures. In Lilium tigrinum, 
for example, in the axils of the upper leaves there are, instead of buds, 
hard black bulbils which fall off and produce new plants. Detached 
scale leaves in a number of the Liliaceae form adventive buds or bulbils 
from their bases. On leaves of Hyacinthus, removed from the plant, simi- 
lar structures may be formed, but Naylor (1940) has shown that in this 
case they do not come from preformed meristematic tissue but develop 
from epidermal and subepidermal cells. 

In many ferns vegetative cells of the prothallus may produce sporo- 
phytes by apogamy. Aposporous gametophytes are also frequently formed, 
especially on isolated juvenile leaves. Lawton ( 1932, 1936 ) was able to 
induce apospory in 13 species of ferns and obtained from them tetraploid 
sporophytes by methods similar to those used in mosses. Aposporous struc- 
tures are often strictly gametophytic, but Beyerle (1932) found adventive 
structures in 34 fern species to include sporophytic buds, undifferen- 
tiated structures, prothallia, and bodies intermediate between sporophytes 
and prothallia (reviews by Du Buy and Neurnbergk, 1938, and Steil, 
1939, 1951). 

Regeneration in the fern sporophyte often results in the production of 
new plants. Buds may be formed on leaves and roots, and foliar em- 
bryos resembling those of higher plants may be produced, as in Campto- 
sorus ( McVeigh, 1934; Yarbrough, 1936£> ) . A review of reproductive re- 
generation in ferns, covering 35 genera and 197 species, has been made 
by McVeigh (1937). 

There should also be mentioned a special kind of reproductive re- 
generation that occurs in vegetative tissues of the embryo sac and ovule, 
by which embryos develop from synergids, antipodal cells, or cells of 
the nucellus and not only from the fertilized egg (Lebeqne, 1952). 
Strasburger (1877) called this adventive polyembryony (as opposed to 
cleavage polyembryony, p. 235 ) . Examples of this are known in Funkia, 
Coelebogyne, Citrus, and others (review by Webber, 1940). The em- 
bryos here formed somewhat resemble the foliar embryos of the 
Crassulaceae and similar plants. Though sometimes started in their 
development by a regenerative stimulus, these are cases of specific repro- 
ductive processes made possible by the totipotence of cells in various 
vegetative portions of the plant and which in a way are comparable to 
asexual propagation by spores. 

In recent years experimental work on morphogenetic problems has 
been concerned primarily with the effects of environmental factors, 

Regeneration 257 

particularly light and physiologically active substances. There is still op- 
portunity for fruitful investigation through an experimental manipulation 
of the developmental process itself, especially by work on problems of 
regeneration such as have been discussed in this chapter. The significant 
results of earlier botanical workers in this field, now cultivated somewhat 
less actively than in the past, could be extended very profitably by sup- 
plementing the older methods with the advantages of modern techniques. 


Tissue Mixtures 

In animal bodies such combinations of genetically unlike tissues as grafts, 
mosaics, and chimeras are rare in nature and rather difficult to bring 
about experimentally, but in plants these are much commoner and easier 
to produce. This is presumably due, in large measure, to the presence of 
localized growing points in plants, which knit together readily. The arts 
of grafting and budding have long been known to horticulturalists and 
provide means for combining two or more varieties of related plants and 
especially for vegetative multiplication of types that cannot be propa- 
gated by seeds or in which cuttings do not easily root. 

In chimeras and in localized genetic changes, tissue mixtures may 
be much more intimate than in ordinary grafting and provide oppor- 
tunities for a study of organization and tissue relationships which are 
not available in homogeneous plant bodies. 


In the practice of grafting, a small branch or shoot, the scion, is in- 
serted into a larger rooted portion, the stock, by means of a cleft or 
other opening in such a way that meristematic regions of the two come 
into contact. The same result is achieved in a somewhat simpler fashion 
by budding. Here a bud from one type is slipped under a cut in the bark 
of another so that the two cambia are in contact, the bud later growing 
into a shoot. A third piece, in "double-worked" trees, is sometimes 
inserted as an intermediate between stock and scion. In practice, these 
methods are used chiefly in woody plants and are the means by which 
most horticultural varieties of trees and shrubs are propagated. The prob- 
lems in these plants will therefore be discussed first. 

A question of much importance both practically and theoretically 
concerns the effects produced by the stock on the scion or the scion on 
the stock. What substances can, and what cannot, pass from one to the 
other across a graft union? Do the two graft partners remain in com- 
plete physiological isolation save for the passage of water and solutes 


Tissue Mixtures 259 

from stock to scion, or in a mutual exchange between the two do the 
specific qualities of one become transferred, in any measure, to the other? 
There has been much disagreement about these matters in the past, and 
final answers are not yet available. 

There are several means by which the two graft partners may affect 
each other. Of most importance, probably, is nutrition. If a root into 
which a scion is grafted has less capacity to absorb, store, or utilize nu- 
trients than do the roots of the plant from which the scion was taken, 
reduction of size and vigor of the shoot system produced by the scion 
will result. Many cases have been found where there are such physio- 
logical differences between stock and scion clones which modify the 
effect of one on the other, particularly of root on scion. Second, there 
may be differences between the two partners in the ease with which they 
translocate water and nutrients. The wood of dwarf rootstocks usually 
has a much smaller proportion of vessels than normal roots. Differences 
in phloem transport are probably even more important, as is shown by 
the dwarfing effect of inverted rings of bark and in other ways. Finally, 
there may be differences between partners in the amount of auxin in 
each or in the rate at which it is inactivated. In several herbaceous plants 
it has been shown that dwarf types are relatively poor in auxin. This 
may be the case in woody forms, for dwarf types of fruit trees are 
much branched, an indication that auxin-induced bud inhibition is weak 
in them. Other growth substances may pass from one partner to the other. 
It is clear that water and salts from the soil pass from stock to scion 
and that carbohydrates also pass across a graft union, and in either 
direction. Certain nutritional changes are thus produced by one on the 
other. A common practice in producing dwarf fruit trees is to graft 
scions of standard varieties on stocks which are genetically small and 
thus have small roots. This reduces the amount of top growth. Most of 
this dwarfing results from the reduced water supply available from the 
roots (Colby, 1935), and dwarfing rootstocks also tends to cause earlier 
cessation of growth in the fall (Swarbrick, 1928). Dwarfing may be pro- 
duced in other ways than by reduction in root size. Tukey and Brase 
( 1933 ) found that where a dwarf variety was used either as a rootstock, 
intermediate stem piece, or top scion its effect was to dwarf the whole 
plant. Some dwarfing may also be attributed to defective graft unions 
and the consequent failure to transfer materials readily (Bradford and 
Sitton, 1929 ) . Dickson and Samuels ( 1956 ) have studied translocation 
across a graft union by means of radioactive tracers and find that there 
is a high concentration of the isotope at the junction of stock and scion, 
suggesting that the dwarfing effect may be due to a block in the flow 
of nutrients to the roots. 

But there seem also to be more subtle factors involved. A special kind 

260 The Phenomena of Morphogenesis 

of dwarfing is that produced in some plants by growth from seeds that 
have not been afterripened. A case of this sort in the peach where the 
dwarfs retained their dwarf character when grafted on normal plants 
was studied by Flemion and Waterbury ( 1945). The roots of such dwarfs 
were able to support normal shoot growth, so that the seat of the dwarfing 
seems to be in the shoot. There is no evidence of a stimulating substance 
produced in the normal plant or of an inhibiting one in the dwarf. 

Some varieties are incompatible in grafting and so do not thrive to- 
gether even though the graft union between them may be good or a 
variety compatible to both is inserted as an interstock ( Sax, 1953, 1954 ) . 
Tukey and Brase present evidence that not only the character of stock 
and scion and their compatibility are important in determining tree size 
but also the effect of environmental factors on each partner and on their 

Aside from purely quantitative effects of stock on scion in horti- 
cultural plants, other traits have been reported to be transmitted from 
one to the other. McClintock ( 1937) found that leaves of the Grimes apple 
grafted on Virginia crab stock have a greater green weight and are 
physiologically different in some respects from those of the same va- 
riety on other stocks. Blair (1938) grafted Bramley Seedling apple on 
French crab but in a number of cases inserted between stock and scion 
a 9-in. piece of one of three other varieties, Mailing II, IX, and XIII. Even 
though here stock and scion were the same, the effect of these middle 
pieces on the tree that grew out of them was markedly different in 
each case, as shown in leaf poise, general habit of branching, leaf color, 
and time of defoliation. The effect of the various middle pieces on the 
root was also evident. Rogers, Beakbane, and Field (1939), however, 
found that intermediate pieces from different sources had relatively little 
effect on the rootstocks. By grafting apple scions from various sources on 
roots, Swarbrick and Roberts (1927) found that the character of the 
root tended to be like that of the variety which contributed the piece of 
stem just above it, whether this was scion or middle piece. Amos, Hatton, 
and Hoblyn ( 1930 ) dispute this conclusion and believe that the effect of 
scion on root is simply quantitative. 

In citrus fruits Halma ( 1934 ) reports that Eureka lemon scions grafted 
on sour-orange roots greatly modify the form and color of the latter. 
In the reciprocal graft, only the color of the root was changed. These 
changes were observed in grafted but not in budded trees. 

Early work on the stock-scion relationship is reviewed by Swarbrick 
(1930), and Garner (1949) has presented the subject from a practical 
point of view. 

The mixture of tissues by grafting and the relation between the graft 
partners pose problems of wider interest than for horticulture alone. It 

Tissue Mixtures 261 

has long been known, for example, that successful grafts are usually 
limited to closely related plants, either between those of the same 
species or between species close together taxonomically. In exceptional 
cases grafts can be made between genera and very rarelv between 
families. Simon (1930), for example, has grafted Solarium and Iresine, 
belonging to the widely separated families Solanaceae and Amarantha- 
ceae. Nickell (1948) grafted white sweet clover on sunflower, and the 
plants continued to grow with normal vigor for 5 months. Silberschmidt 
(1935) studied 550 grafts between plants of the same species, of related 
species, and of unrelated species. Anatomical fusion in some cases oc- 
curred in grafts of unrelated species but here union was slower and less 
extensive, passage of nitrogen from scion to stock was reduced, and 
proteolytic activity of stock juices increased. The last fact is presumably 
concerned with the failure of unrelated grafts. 

Schroter (1955) reports that Zinnia elegans, one of the Compositae, 
can be grafted to tobacco, although not on some plants in the same 
family as tobacco. He attributes the successful Zinnia grafts to the pres- 
ence of nicotine in this plant. Mothes and Romeike ( 1955 ) grafted scions 
of tomato, petunia, belladonna, and tobacco on tobacco stocks of vary- 
ing nicotine content and found that the richer the stock was in nicotine, 
the poorer was the development of the grafted scion. 

The passage of nicotine from tobacco roots to tomato scions has been 
described (p. 220). Hieke (1942) found that in grafts between Lyco- 
persicon, Nicotiana, Atropa, and Datura the alkaloids found in the scion 
were those characteristic of the root to which this was grafted. 

Kostoff (1929) presented evidence from grafts between various plant 
types that immunity, as tested by the precipitin reaction, can be acquired 
by plants much as in animals. Chester and his colleagues, however, in a 
series of papers (1932 and others) showed that the precipitates re- 
ported are not the result of a true precipitin reaction but are simply 
calcium oxalate, a widespread substance in plants. 

Monocotyledonous plants have been found more difficult to graft than 
dicotyledonous ones, presumably because they lack a cambium. Muzik 
and La Rue (1954), however, grafted a number of species of grasses, 
including some belonging to different genera. 

The closer to the embryonic condition a tissue is, the more readily it 
can be grafted. The smallest successful graft of this sort seems to be one 
made by Gulline and Walker (1957) in which a shoot tip containing 
only about 600 cells and with a volume of less than %ooo cu - mm - was 
grafted back on the apex from which it had been cut. Later development 
was normal. 

As to the reciprocal relations of stock and scion in nonwoody plants, 
many conflicting results have been reported. Daniel described various 

262 The Phenomena of Morphogenesis 

instances of a marked qualitative effect of stock on scion or scion on 
stock. The production of tubers, for example, he believed could be trans- 
ferred by grafting from the Jerusalem artichoke to the common sun- 
flower, which normally bears no tubers. In one of his last papers (1929) 
Daniel maintains that these induced changes have sometimes become 
transmissible through seed for several generations and regards this as 
proof of the inheritance of an acquired character. 

This is essentially the position taken by Lysenko and his Russian 
colleagues. One of them, Avakian (1941), reported marked reciprocal 
effects, on fruit color and other characters, of red-fruited, yellow-fruited, 
and white-fruited tomatoes when grafted together in various ways. Simi- 
lar experiments were repeated in this country by Wilson and Withner 
(1946), who were unable to confirm these results in any respect. Bohme 
(1954) also found that no inheritable effects were produced by grafting 
between varieties of tomatoes. 

There are a number of well-authenticated cases, however, of the trans- 
mission, between graft partners, of factors that determine qualitative 
and not simply quantitative and nutritional differences. These present 
some important problems both for morphogenesis and for physiology. 
Conspicuous among them are the numerous instances where a flowering 
stimulus, from a plant which has been induced to flower by a particular 
photoperiod, can be transferred by grafting to a nonflowering plant and 
cause it to flower (p. 396). Evidently some substance is transmitted across 
the graft that stimulates flowering. 

The effect of this stimulus may be modified in various ways. Haupt 
(1954), using a late-bearing variety of peas, grafted terminal shoots of 
different ages on stocks of different ages. If scions of young plants are 
grafted on older ones, flowering takes place up to six nodes earlier than 
in controls grafted to stocks of their own age. Evidently the substance con- 
cerned with flower development is not formed in the first stages of 
the plant's growth but can be effective then if introduced from older 

The production of flowers after grafting may be due to other factors 
than a specific flower-inducing substance. The Jersey type of sweet po- 
tato rarely flowers in this country but can be made to do so by grafting 
it to another species of Ipomoea that does not form storage roots (Kehr, 
Ting, and Miller, 1953). These authors believe that flowering results 
from the accumulation of carbohydrates in the shoots after grafting. 

The tendency to form tubers in potato may be transmitted by grafting. 
If the shoot of a variety that produces tubers under a long day is grafted 
to a short-day variety and grown under long days, the short-day variety 
will now produce earlier and larger tubers than it would have done by 
itself (Howard, 1949). This effect was not transmitted through these 

Tissue Mixtures 263 

tubers when they were used as seed. Somewhat similar results are re- 
ported by Gregory (1956), who also found that the tuber-forming stimu- 
lus could be transmitted from all parts of the shoot by grafting. 

Some recent experiments on the graft transmission, or lack of trans- 
mission, of plant traits are the following: 

In English ivy (Hedera helix) Doorenbos (1954) grafted scions from 
the juvenile type (with lobed leaves) onto the adult, upright, flowering 
form and found that the latter often lost its abilitv to flower and 
showed other juvenile traits. 

Popesco (1949) reports that when Sophora japonica, a woody peren- 
nial leguminous plant, is grafted to the common bean, the bean flowers 
15 to 20 days later than it otherwise would and becomes perennial in 

Hybrids between Meliotus alba and M. dentata are deficient in 
chlorophyll because of some gene interaction and die in a few days. If 
such hybrids are grafted to sweet clover, however, they grow well, 
flower, and bear seed (W. K. Smith, 1943). 

In spruce (Picea abies), Muller-Stoll (1947a) examined grafts of shoots 
from the tops of old trees on young seedlings. After 3 years these 
flowered abundantly, far earlier than they would have done otherwise. 
Only female cones were produced, presumably because the branches 
used as scions were from the top of the tree, which bears chiefly female 
cones. This localization the author interprets as an instance of topophysis. 

Many cases have been described of the effect of stock on fruit size 
in the scion. Bitters and Batchelor ( 1951 ) report such a case in the orange, 
where Washington navel oranges were grafted on 32 different root- 
stocks and Valencia on 26. Differences were found between stocks in 
their effects on fruit size in the scion. These differences are not related 
to those in tree size or in number of fruits. 

An incompatible graft is reported in cucurbits by Wellensiek (1949). 
Grafts of muskmelon on Cucurbita ficifolia grew for a time and then 
suddenly wilted and died. Muskmelon as an interstock between cucum- 
ber and C. ficifolia has the same effect. If a few leaves are left on the lat- 
ter, however, the graft with muskmelon is successful. The reciprocal graft 
thrives, so that incompatibility is in only one direction. The author 
believes that muskmelon fails to give the stock of C. ficifolia a sub- 
stance necessary for its growth and that this is provided if a few leaves 
are left on it. 

Yampolsky (1957), in the dioecious Mercurialis annua, grafted male 
and female plants together in a variety of ways but found no alteration 
in the sexual character of either. 

Of particular interest for genetics are those cases where there is a 
known genetic difference between scion and stock. Hijoscijamus niger, 

264 The Phenomena of Morphogenesis 

the black henbane, has an annual variety and a biennial one, the annual 
forming a flower stalk and flowers in a single season but the biennial re- 
maining in the rosette stage for the first year unless experimentally ex- 
posed to low temperature for some time. This difference is due to a 
single gene, with the biennial character almost completely dominant. 
Melchers ( 1937 ) grafted a scion from the annual variety ( and also scions 
from Petunia, Nicotiana, and other related annual plants) into the 
rosettes of the biennial variety in its first year and by this means induced 
the biennial to flower in the same season. Melchers attributes this to the 
passage across the graft union of a nonspecific flower-forming substance. 
The genetic difference between the annual and the biennial varieties 
seemed to be due to the ability of the former to produce a flower-form- 
ing hormone without previous exposure to low winter temperature. 
Melchers later (1938) found that if a short-day variety of tobacco, 
grown under long-day conditions and thus unable to flower, was grafted 
into biennial Hyoscyamus in its first season, the latter soon produced 
flowers. The gene in Hijoscijamus may thus control the ability to respond 
to the flower-inducing substance rather than to form it. 

In Petunia nyctaginiflora von Wettstein and Pirschle (1938) found a 
gene d which differed from the normal D in producing plants that are 
smaller and have fewer branches, smaller and more rounded leaves, and 
a marked chlorophyll deficiency. Scions of dd grafted on DD stocks had 
slightly larger leaves and stems and more branches but were not much 
different from ungrafted dd. Scions of DD on dd showed general reduc- 
tion in size and a chlorophyll deficiency, which was greatest next the 
graft union and decreased in intensity above this. The authors believed 
that a substance was produced in the mutant that passes into normal 
scions and there either inhibits chlorophyll formation or causes chloro- 
phyll degradation. Pirschle later ( 1939 ) presented evidence that dd lacks 
a hormone, present in DD, that stimulates growth in size but does not 
affect the shape of leaves or flowers. Objection may be raised that these 
are nutritional and not hormonal effects. 

When tobacco and tomato are grafted with the dd mutant of Petunia, 
their leaves show chlorophyll deficiency when the dd plant is used as a 
stock and to a lesser degree where it is used as a scion (Pirschle, 1940). 
The gene-produced d substance is clearly not species-specific. The 
possibility cannot be disregarded that the supposed d mutant is actually 
a virus infection, although its clear genetic segregation from D would sug- 
gest that it is not. 

The single-gene mutants nana of Antirrhinum siculum and sterilis of 
Solanum lycopersicon were produced by radium irradiations and have 
been studied by Stein (1939). The former has a single unbranched 
main stalk (unlike A. siculum); is flowerless; and has larger, thicker, 

Tissue Mixtures 265 

and darker leaves. When grafted as a scion on A. siculum stock, the 
stem is somewhat . shorter and bears abnormal flowers but is otherwise 
unchanged. The tomato mutant is dwarf, has scanty chlorophyll, and 
lacks branches and flowers. When grafted as a scion on tomato stock 
its chlorophyll remained deficient but its growth became approximately 
normal as to height, branching, and flower development. The particular 
interest of these cases lies in the fact that the various effects of a gene 
here seem to be separated, some of them passing across the graft union 
and others not doing so. Such material offers a favorable opportunity 
for the study of gene action. 

Rick (1952) found a tomato mutant, wilty dwarf (wd), which differed 
from normal in having fewer and shorter internodes, smaller leaves, thin- 
ner stems, blue-gray leaf color, and a tendency to wilt in summer. This 
was grafted in various ways on homozygous normal ( + ) lines differing 
from wd in only this one gene. Control grafts of wd on wd and of + on -}- 
showed no modification. Reciprocal grafts of -f- on wd and wd on + had 
their leaf dimensions and stem lengths shifted markedly toward the 
character of the stock. In double graft combinations, wd/ -\- /wd and 
-h /wd/ +, the top scion was unaltered but the middle piece was changed 
in the direction of the stock. The author concludes that the effect of 
stock on scion here is not due to factors in the stem or leaves but in the 
root system of the stock. 

Kostoff ( 1930Z? ) observed irregular meioses in the pollen cells of tobacco 
grafted to other genera, and such scions produced much abortive 
pollen. If flowers on them were selfed, various chromosomal aberrants 
appeared in the progeny but none in the controls. 

In Petunia, Frankel ( 1956 ) grafted fertile scions to stock showing 
cytoplasmic male sterility and found sterility in the offspring of such 
scions, suggesting that cytoplasmic sterility determinants had passed 
from stock to scion. The author recognizes the possibility that nutri- 
tional changes induced by grafting may have been responsible for these 

Wagenbreth ( 1956 ) made grafts between a number of species of 
legumes and found by inoculation experiments that, although bacteria 
specific for the stock would produce nodules in such plants, bacteria 
specific for the scion would not. 

Common experience has shown that strictly qualitative characters such 
as shape are usually gene-controlled and not influenced by grafting, 
despite early claims to the contrary. A few pieces of positive evidence, 
however, have been reported in recent years. Heinicke (1935) observed 
that Mcintosh apples borne on defoliated scions grafted into Northern 
Spy, and thus grown from material produced by Northern Spy leaves, 
tended to be modified in the direction of the latter variety. Southwick 

266 The Phenomena of Morphogenesis 

(1937) also found that Mailing stock had some influence on the form 
and size of Mcintosh apples grown on it. These cases need confirma- 

In general, from the mass of literature available, one may conclude that 
stock can influence scion in producing differences in plant size, size of 
leaves and fruits, plant habit, flowering time, life span, content of inor- 
ganic and certain organic substances and growth substances, and to some 
extent in fertility and resistance to disease. Influence of scion on stock 
is much less marked. Most of these effects definitely have a chemical or nu- 
tritional basis, and few cases of strictly qualitative changes are known. 
It should be recognized, however, that a purely quantitative difference, 
as in fruit or leaf size, may influence shape by allometric correlation 
(p. 105). Truly morphogenetic effects rarely-perhaps never-pass across 
a graft union. The great preponderance of evidence also supports the con- 
clusion that no permanent genetic change is induced by one graft part- 
ner on the other. 

Stock-scion relationships have been reviewed by Rogers and Beak- 
bane (1957). Much of the literature on grafting (and many other 
things) is summarized in Krenke's (1933) monumental work. The theo- 
retical aspects of grafting have been reviewed by Roberts ( 1949 ) . 


The instances of tissue mixture just described have been by the arti- 
ficial union of two genetically different plants. These types remain sharply 
separated, each branch or other unit of the plant belonging definitely 
to one or to the other. There are tissue mixtures, however, where the fusion 
is much more intimate than this and where an organ such as a stem, leaf, 
or root is not homogeneous but is made up of two or more tissues that are 
genetically unlike. This difference may arise by somatic mutation, the 
mutated cells multiplying and forming a part of the whole, or it may be 
the result of a mixture of meristematic tissues at a graft union. The im- 
portant fact, morphogenetically, is that these diverse groups of cells 
do not each form an organism or produce developmental abnormalities 
but that they coexist as parts of the same organized system. What is pro- 
duced is a normal, whole plant. Here the organizing capacity of living 
stuff and the self-regulatory quality of the organism are particularly 

Mixtures of tissues that come from different sources are called chimeras, 
a term proposed by Winkler (1907b) from the analogy between such 
plants and the chimeras of mythology which were part lion, part goat, 
and part dragon. A number of types of chimeras are recognized, depend- 
ing on the relationship of their components. 

Tissue Mixtures 267 

The work on chimeras has been reviewed by Swingle (1927), Weiss 
( 1930), Neilsoh-Jones ( 1934, 1937), and Cramer ( 1954). 

In the so-called mixed chimeras the two kinds of tissue are mingled 
irregularly together. This mixture may persist but it is often a tem- 
porary stage and succeeded by one of the more regular types as the 
meristem becomes better organized. In mericlinal chimeras, often de- 
rived from mixed ones, one type of tissue forms a thin layer over a part 
of the surface of the other. 

The other types show a more regular relation between their two 
components. In sectorial chimeras, a definite sector of a radially sym- 
metrical structure such as a root, stem, or fruit is of one type and the 

Fig. 10-1. Sectorial chimera in apple. (From Zundel.) 

rest is of the other (Fig. 10-1). It is not uncommon to find in fruits 
such as apple or orange a sector in which the color or texture of the 
skin is different from that of the rest and which sometimes can be traced 
into the axis of the fruit. Such a sector may be distinguished in the stem, 
also, and the line between the two components sometimes runs out through 
the blade of a leaf. The term sectorial chimera may be used more 
broadly for a type in which there are large masses of diverse tissue 
adjacent to each other, regardless of whether the boundary line has 
any relation to the axis of symmetry. Thus an apple in which the terminal 
portion is of one type and the basal another, with an irregular boundary 
between, has been called a sectorial chimera. Sectorial chimeras are 
often found in shoots that arise from the vicinity of a graft union. Some 
of them may really be mericlinal ones, with one member covering a 

268 The Phenomena of Morphogenesis 

sector but only skin deep. Sectorial chimeras may sometimes be dis- 
covered by a study of their internal structure (Brumfield, 1943, p. 76, 
and Fig. 10-2). 

Of most interest to morphogenesis, however, are periclinal chimeras. 
In these remarkable plants the outer cell layers are derived from one 
graft partner and the entire inner portion of the plant from the 
other. "Graft hybrids," which have long been a puzzle to horticultural- 
ists, prove to be periclinal chimeras. They arise from grafted plants and 
partake of certain of the characters of each, but their own characters 
cannot be transmitted through seed. One of the familiar forms is 
Crataegomespihis, originating from a graft between two rosaceous 
genera: Crataegus, the hawthorn, and Mespilus, the medlar. Another is 

Fig. 10-2. Sectors of a root of Vicia after previous exposure to X radiation. Stippled 
cells are those in which observed mitoses showed that the chromosomes were un- 
changed. Black cells are those where a chromosomal change could be observed. 
These cells presumably are descended from a cell at the very tip in which a change 
had been induced and which had then given rise to a sector or wedge of similar cells. 
( From Brumfield. ) 

Cytisus-labiirnum, coming from a graft between these two leguminous 

Plants of this sort were first experimentally produced and carefully 
studied by Winkler ( 1907fc, and later papers ) . He grafted two closely 
related species, the nightshade, Solarium nigrum, and the tomato, Solanum 
hjcopersicon. After union, most of the scion was cut off, and from 
adventitious buds arising near the point of union plants developed 
which sometimes showed mixtures of the two types of tissue. Most of 
these were sectorial chimeras. Occasionally, however, Winkler found 
a plant that showed no obvious separation into two types of tissue but 
was clearly intermediate in character between nightshade and tomato. 
Several distinguishably different types of such "graft hybrids" appeared in 
these experiments, were maintained by vegetative propagation, and bore 

Tissue Mixtures 269 

flowers, fruits, and seed. Whether these were really mixtures of tissue, 
or, in some cases at least, were actual vegetative hybrids, was not clear 
at first. Baur's (1909) analysis of a Pelargonium with white-margined 
leaves showed that in this plant both the epidermis and the layer 
beneath it lacked chlorophyll. This conception of a continuous layer 
of one type of cells covering a core of another type was applied to 
Winkler's chimeras, and the latter proved to be periclinal ones. Such 
forms apparently arise at a place where there is a thin layer of one 
tissue over the other. A growing point, originating in the deeper tissue, 
pushes up and carries on its surface one or two cell layers of the other 
type. From this layered meristem a new shoot is formed. Mericlinal 
chimeras may thus be converted into periclinal ones. Jorgensen and 
Crane (1927) repeated Winkler's experiments, using five species of 
Solarium, and observed in more detail the origin of chimeras. 

In the tomato-nightshade chimeras four different forms were recog- 
nized, propagated, and even given Latin names. In one there was a 
single layer of tomato over a core of nightshade; in another, two layers; 
in a third, one layer of nightshade over tomato; and in a fourth, two 
layers of nightshade. The reason that there are rarely more than two 
layers of the outer component is presumably because a new growing 
point always arises near the surface. 

In Winkler's material it was relatively easy to distinguish the two 
components of the chimera cytologically, since tomato has 24 chromo- 
somes (2n) and nightshade 72. When chromosome counts could not be 
made, the size of the cells (much larger in nightshade) was almost as 
good a criterion. It was found that the layers could be distinguished 
at the apical meristem and that they maintained their specific character 
throughout the life of the plant. When the outermost layer at the grow- 
ing point was from one partner (tomato, for example) only the epi- 
dermis of the plant was of that type. When the second layer of the 
meristem, as well, was from tomato, the two outer layers of the plant 
were of this type. Occasionally in certain tissues of the mature plant these 
layers would become somewhat thicker by periclinal divisions and thus 
include more cells, but this was relatively uncommon. 

In periclinal chimeras (as in all seed plants) the genetic character of 
the plant is determined by the cell layer just beneath the epidermis. 
From this layer the sporogenous tissue is formed. The offspring of a 
chimera, by seed, is therefore identical with the graft partner that con- 
tributes the subepidermal layer of cells. 

Winkler maintained that two other types of chimeras that he obtained 
from grafts were true burdos, or vegetative hybrids, in which one layer 
had arisen by an actual nuclear fusion between cells of the two com- 
ponent species. The aberrant chromosome counts (cells with neither 24 

270 The Phenomena of Morphogenesis 

nor 72 chromosomes) can probably be explained in other ways. Brabec 
(1954) repeated Winkler's work and found cells with varying and 
irregular chromosome numbers but attributed this fact to the origin of 
the new shoot from highly polyploid cells. Pith cells, as has been shown 
by various workers, are often polyploid, and when chimeras come from 
such tissue the chromosome situation is often complicated. The genetics 
and .cytology of the Solarium chimeras have been studied by Gunther 
(1957). Present evidence is against the occurrence of vegetative 
hybridization and the existence of burdos. In the light of recent knowl- 
edge of periclinal chimeras, Bergann (1956) has reinvestigated the 

Control 2n, 2n, 2n 

8n, 2n, 2n 

4n, 2n, 2n 

2n, 4n, 2n 

Fig. 10-3. Apical meristems of four periclinal chimeras in Datura consisting of 2n, 4n, 
and 8n layers. Labels refer to the first two layers and the core. ( From Satina, Blakes- 
lee, and Avery. ) 

In leaf and fruit, the tomato and the nightshade are very different. 
Each of the four periclinal chimeras produced by Winkler by grafts 
between them shows distinctive combinations of these traits so that it 
is possible to determine the effects of one and of two cell layers of each 
type when it covers a core of the other. Such chimeras provide an excel- 
lent opportunity to study the morphogenetic influence of the various 
meristematic layers and the developmental origin of shape differences 
and of tissues. 

The most complete and thoroughly analyzed series of periclinal chim- 
eras are those studied by Satina and Blakeslee in Datura (1941, 1943, 
1944, 1945; Fig. 10-3). By soaking seeds of Datura stramonium in col- 
chicine solution, polyploidy was induced in certain cells of the shoot 

Tissue Mixtures 271 

meristem in 68 plants. The first cell layer (L I), the second (L II), and 
the third (L III, including everything below the first two) were often 
affected independently. L I was changed most often and L II least. These 
layers retained their specific chromosome complement throughout the 
structure of the plant that developed from this meristem. These plants 
are periclinal chimeras, not obtained from graft unions but by chemical 
induction. Unlike the tomato-nightshade forms and similar ones, here 
there may be three genetically different layers (2n, An, and 8n) instead 
of two, and these three were found to occur in almost any order from 
without inward. In the first report there were the following distributions 
of polyploidy among the layers: 2n, 2n, An; 2n, An, 2n; 2n, An, An; 2n, 8n, 
An; An, 2n, 2n; An, 2n, An; An, 8n, An; 8n, 2n, 2n; 8n, An, An. Other combi- 
nations were found later. 

There is little difficulty here in distinguishing the layers since cell 
size is approximately proportional to chromosome number and the 2n, 
An, and 8n cells are thus markedly different. Since plants belonging to 
the polyploid series differ little except in size, morphological combinations 
of characters, as in the tomato-nightshade chimeras and others where the 
partners are so unlike, cannot be seen. 

An understanding of periclinal chimeras has aided in the solution of 
a number of horticultural and morphological problems. Not only have 
the classical examples of the "graft hybrids" been given a satisfactory 
interpretation but other facts discovered in vegetatively propagated 
plants are now explained. Bateson (1921), for example, found that in a 
variety of Bouvardia with pinkish-white flowers root cuttings produced 
plants with red flowers. Here it is probable that the core of the plant 
was of a red-flowered variety and only the outer layers were genetically 
pink. Since lateral roots arise from the vascular cylinder ( here genetically 
red) and push out through the cortex and epidermis, buds from these 
lateral roots would be expected to form red flowers. Asseyeva (1927) ob- 
served in the varieties of potatoes arising by bud mutation that if the 
buds are removed from the seed tuber the new ones which now arise 
from the deeper tissues form plants like those from which the mutant 
variety had come. Zimmerman ( 1951a ) reports a similar case in roses. 
Such vegetatively propagated plants are probably periclinal chimeras 
with only the outer layer or layers of cells belonging to the mutant type. 

Other horticultural plants prove on examination to be chimeras. Einset 
and his colleagues at the Geneva (N.Y. ) Experiment Station have 
found that six large-fruited sports of several apple varieties are really 
periclinal chimeras. The core tissues of the meristem are tetraploid, and 
these are covered by one, two, or rarely three layers of diploid cells 
(Einset, Blaser, and Imhofe, 1947; Blaser and Einset, 1948). 

Dermen has worked extensively with polyploid chimeras. In apples he 

272 The Phenomena of Morphogenesis 

found some types in which (like those examined by Einset) the layers 
are 2n, 2n, An. In others, these are 2n, An, 2n, the whole core being 
diploid (Dermen, 1951). He has also studied chimeras in the cranberry 
(1947a). Einset and Lamb (1951) conclude that most of the so-called 
tetraploid grapes are actually diploid-tetraploid periclinal chimeras, as in 

Fig. 10-4. Portion of carpel wall and placenta of Datura in a periclinal chimera that 
was 2n, 8n, 2n in constitution. In material like this it is possible to trace the origin 
of tissues from specific layers at the meristem. (From Satina and Blakeslee.) 

Kerns and Collins (1947) obtained chimeras in pineapple with col- 
chicine. Some had a 2n epidermis with all the other tissue An, and these 
resembled the completely tetraploid plants. Some had a An epidermis and 
the rest In and resembled completely diploid plants. These authors be- 
lieve that there are only two "germ layers" in the pineapple. 

Periclinal chimeras have made an important contribution both to 
morphology and morphogenesis by making it possible to trace the con- 
tinuity between the regions of the meristem and the structures of the 
mature plant, since when the cells of the components of a chimera are 
distinguishable, their descendants can be traced throughout develop- 

Tissue Mixtures 


ment (Fig. 10-4). This is particularly true in the polyploid chimeras. 
The work on Datura by Satina and Blakeslee has provided much infor- 
mation here, and Dermen (1953) presented similar evidence for the 
peach. There seems to be no invariable rule as to just what mature 
structures are produced by different layers at the meristem. The particu- 
lar tissues contributed by L II and L III are not only different among 
species but between large and small individuals of the same species 
(Dermen, 1951). 

It is significant that in the root, where a root cap is present and con- 
tinuous layers of cells do not cover the tip of the meristem, periclinal 
chimeras do not occur. Sectorial chimeras, however, have been ob- 
served in roots (p. 76). 

Fig. 10-5. Somatic mutation. Sectors of cells in corolla of Pharbitis resulting from 
mutation from colorless to colored sap. The wider the sector, the earlier was the origin 
of the mutation in the development of the flower. ( From Imai and Tobuchi. ) 


Mixtures of various types of tissue may appear not as a result of graft- 
ing or experimental treatment but spontaneously. Mutations in vegetative 
cells are not uncommon in some plants. Where a mutant cell is dis- 
tinguishable, by color or in other ways, its descendants form a spot or 
stripe of tissue unlike the rest (Fig. 10-5). The earlier the mutation oc- 
curs, the larger the mass of tissue that will be produced. In annual 
Delphinium, Demerec (1931) found a gene that mutated frequently in 
petal cells, changing their color from rose to purple. An early mutation 
altered a large part of the plant but later ones formed only small spots 
on the petals. Some cases of variegation, as in maize pericarp (Anderson 
and Brink, 1952), are due to mutable genes of this sort though most 
color patterns in plants result from differentiation during development 
and not from a mixture of genetically different tissues. Some genetically 
variegated plants become chimeras, and Dermen ( 1947b ) has been able 
to determine the specific meristematic layer (LI, L II, or L III ) in which 
the mutation took place. 

274 The Phenomena of Morphogenesis 

McClintock (1929) reported a case in maize where the microsporocytes 
had 19 chromosomes but the root tips 20, suggesting that a chromosome 
had been lost during the development of the upper part of the plant. 

A number of instances are known in hybrid plants where twin stripes 
or spots occur, differing from each other and from the background 
color. These have been interpreted as the result of crossing over in so- 
matic cells. Thus in Phaseolus the Fj of a cross between plants with violet 
and with lilac flowers produced a form with light violet flowers. In one 
of these there were two stripes, side by side, one of them violet and one 
lilac (Prakken, 1938). A sectorial chimera presumably due to such 
vegetative segregation was reported for a pear fruit by Gardner, Crist, 
and Gibson ( 1933 ) . Twin spots, also apparently caused by somatic 
crossing-over or chromosome translocation, are frequent in maize peri- 
carp (Jones, 1938). 

Huskins and others (Huskins, 1948; Huskins and Cheng, 1950) re- 
ported numerous instances where, as the result of low temperature or of 
various chemical treatments, somatic mitoses occur in which the number 
of chromosomes is reduced, as it is in meiosis. Wilson and Cheng 
(1949) found that in such cases members of homologous pairs separated 
much oftener than they would have done by chance, indicating a true 
genetic segregation in the body cells of a heterozygous plant. 

In all these instances of genetic alteration in a few cells, the difference 
between these cells and the normal type is usually not very great, and 
there is less to be learned morphogenetically than in grafts and chimeras. 
Whenever genetically different tissues from any source are present to- 
gether in the same individual, however, their coexistence in a single 
whole is evidence of the organizing capacity of living stuff. 


Abnormal Growth 

The basic problem of morphogenesis, as stated frequently in the preced- 
ing chapters, is posed by the fact that every organism is an organized 
system that in its development tends to produce forms and structures of 
specific character. Within each organism there seems to be a norm toward 
which its development conforms. The expression of this norm, however, 
may vary greatly as the result of a wide range of environmental factors. 
Such variation is familiar and to be expected. There are many cases, 
however, where divergence from the norm is so great that we usually 
speak of them as "abnormal" or "atypical." Just what do we mean by 
these terms? 

Every organism, and doubtless every cell, has a far wider range of 
development potencies than it generally displays. When conditions are 
different from the usual ones, the expression of its norm is also different, 
but the norm is just as specific as before and the organism still per- 
sistently regulates its development in such a way that the characteristic 
form for that environment is produced. Norms of the individuals in a 
given species are much the same. The basis of each is the genetic con- 
stitution of the organism. It is well understood by geneticists, however, 
that what a particular gene determines is not a particular character 
but a particular reaction to a particular environment, external or in- 
ternal. In many cases it requires very special environmental factors, 
such as wounding, irradiation, application of growth substances, or 
deposition of an insect egg to bring to expression developmental poten- 
cies which would otherwise remain latent. 

Under such unusual conditions, or from genotypes which are markedly 
different from those of most members of the species, individuals some- 
times develop which are so unlike ordinary ones that they are called 
"abnormal." This does not mean that they are exceptions to the general 
biological determination that controls the growth of all living things. 
Neither do they constitute a sharply defined group set apart from all 
others, for every gradation between normal and abnormal may be found. 
One may be uncertain, for example, as to whether the occasional pro- 


276 The Phenomena of Morphogenesis 

duction of different forms of leaves on the same plant or the growth 
of plantlets on the leaf margins or many structures appearing during the 
process of regeneration are normal or not. What seems at first to be ab- 
normal may prove to be simply an intensified or exaggerated manifesta- 
tion of developmental potency. The degree of divergence from the 
average is the basis on which we term a structure or an individual ab- 
normal, and biologists often disagree as to how divergent an organism 
must be to warrant this designation. "Abnormalities" in animals and 
plants are recognized largely as a matter of convenience in order not to 
complicate still further an already difficult taxonomic and morphological 

Thus there are two concepts of what a norm is and therefore of what 
is abnormal. One is a developmental concept: the norm or standard, 
based on a specific protoplasmic pattern, to which the organism tends 
persistently to conform. The existence of this norm is the basic fact in 
biological organization. Its expression may vary greatly as the environ- 
ment changes but it always remains as the core of the morphogenetic 
process. In this sense, nothing is abnormal. The other concept of the 
norm is a purely statistical or taxonomic one. In most species the de- 
velopmental norms of its individuals do not vary widely in their expres- 
sion since these individuals are genetically very similar and have been 
exposed to a relatively narrow range of environmental influences. This 
rather constant developmental expression may for convenience be re- 
garded as a norm, and everything that differs from it substantially may 
be called abnormal. It is in this statistical sense that the term "abnormal 
growth" is generally employed. 

The student of morphogenesis, however, does not put aside these in- 
stances of abnormal growth and development as unimportant for his 
purpose. Such may well prove to be more enlightening than most "nor- 
mal" individuals. They are exceptions, extreme cases, and from exceptions 
like these often come clues to the solution of particularly difficult prob- 
lems. Furthermore, in many cases of abnormal growth certain levels of 
the very quality that we associate with life— organization— have disap- 
peared. Tissue cultures and many tumors and galls are formless, largely 
unorganized masses of cells which no longer produce the beautifully 
coordinated structures called organisms. Individual cells here must still 
retain a basic, vital organization in their living stuff, for otherwise they 
would die, but the higher levels of organization have now broken 
down. In other types of abnormal growth quite the opposite change 
has occurred and entirely new structures, specifically formed and well 
organized though on a different plan, are produced, as in many insect 

The subject of abnormal growth is therefore a promising one for stu- 

Abnormal Growth 277 

dents of plant morphogenesis since by the study of these unusual struc- 
tures development may be examined at various levels and degrees of or- 
ganization. Relatively little work has been done in this field, however, 
and most of the results are descriptive and relate to mature structures. 
In the few cases, such as crown gall, where many experimental and 
developmental studies have been made, these have proved to be very 

There is some confusion between the concepts of "pathological" and 
"abnormal" growth. Pathology is concerned with questions about the health 
and survival of the organism when it is attacked by parasites or sub- 
jected to unfavorable conditions. Abnormal growth is often produced 
by this means though here, again, it is difficult to draw the line. One would 
hardly call the aecium of wheat rust an abnormal growth, but a crown 
gall certainly is one. There are many cases of abnormal growth, on the 
other hand, which clearly are not pathological, such as inherited fascia- 
tions or the root tubercles of legumes. A student of abnormal growth is 
not concerned with the health of the plant, nor does a pathologist ex- 
amine primarily the ways in which the plants with which he deals 
diverge from the norm. Historically, however, the two fields have been 
close together, and Kuster's (1925) classic book on abnormal growth is 
entitled "Pathological Plant Anatomy." 

There is no very obvious way in which to organize the widely various 
phenomena of abnormal growth. It will be most logical, perhaps, to 
proceed from cases where divergence from the norm is relatively slight 
and move to those where it is more extreme. In the present chapter 
there will be discussed (1) the abnormal development of organs be- 
longing to the usual categories, (2) the production of new types of or- 
ganized structures, and (3) the production of amorphous structures. 
The whole field has been briefly reviewed by Bloch ( 1954 ) . 


In many cases structures still recognizable as leaves, stems, roots, flow- 
ers, or other organs have been modified in many ways, sometimes very 
radically. This is the field of teratology, the study of malformations, 
freaks, and monstrosities, which has long excited the curiosity of bot- 
anists (Moquin-Tandon, 1841; Masters, 1869; Worsdell, 1915; Penzig, 
1921; and Heslop-Harrison, 1952). Little but descriptive work has been 
done on most of them. For a long time their scientific value was chiefly 
to morphologists, who looked to malformations for evidence as to the 
morphological nature of certain organs. Thus the "metamorphosis" of 
petals and sepals into leaf -like structures (phyllody; Fig. 11-1) suggests 
that they are really leaves but have been modified in function during 

278 The Phenomena of Morphogenesis 

evolution. Heslop-Harrison lists three causes for such phenomena: ab- 
normalities in growth, of little morphological significance; abnormalities 
in development, owing to failure of hormonal systems or other form- 

Fig. 11-1. Sepal phyllody. Sepal 
of a rose flower showing abnor- 
mal development into a structure 
much like a foliage leaf. (From 
T. E. T. Bond.) 

determining factors; and minor abnormalities arising from genetic or 
environmental causes. Some teratological changes may be reversionary 
and some progressive. 

Fig. 11-2. Left, normal tendril of Cucurbita. Right, abnormal structure intermediate 
between a tendril and a foliage leaf. ( From Worsdell. ) 

There are various categories of these abnormal structures, or tera- 
tomata (Figs. 11-2, 11-3), which are only loosely related to each other. 
About certain of them a considerable literature has grown up. In many 
cases the causal agent is known, but often it is not. Little developmental 
work has been done on most of them. 

Abnormal Growth 


Organoid Galls. Under this term Kiister (1910) included many cases of 
abnormal development or distribution of organs which are clearly the 
result of parasitism, nutritional disturbances, or other known causes and 
which often show little constancy of form or structure. Such galls may 
appear at some distance from the site of the stimulating agent. With these 
are usually included similar types of abnormalities even if their causal 
factors are unknown. 

In some of these galls it is chiefly the form that is abnormal. In the 
leaves of various species of Juncus parasitized by Livia juncorum, for 
example, the sheath reaches extraordinary size while the lamina remains 
small or atrophied. In Populus tremula small stipules turn into large, leaf- 
like structures. Eriophyces also produces floral abnormalities in which 

Fig. 11-3. Abnormal flower of 
gloxinia, with extra petal-like 
structures on the outer surface 
of the corolla and an increase 
over the normal number of 
corolla lobes. ( From Worsdell. ) 

stamens or carpels become petal-like. Flower buds that have been 
grafted in a place where leaf buds would normally be sometimes produce 
unusually large and abnormal flowers, evidently because of nutritional 

Some of these changes are comparable to those occurring in regenera- 
tion. Thus in vigorously regenerating shoots of Symphoricarpos simple 
leaves become pinnately cut, and in regenerating stalks of Sambucus, 
stipules may be converted into leaves. The removal of the main shoot 
in the seedling of Vicia faba results in the formation of primary leaves 
or transitional ones instead of those of mature type. It is well known 
that decapitation, defoliation, and similar injuries lead to various 
changes. Goebel (1882) thus obtained leaves instead of bud scales in 
Prunus padus, and Blaringhem (1908) reported many morphological ab- 
normalities due to wounding. The formation of cups or aecidia on 

280 The Phenomena of Morphogenesis 

leaves seems often to be due to nutritional disturbances. Bond (1945) 
found that sepal phyllody in roses is an effect of hormones that control 
the vigor of growth and the balance between reproductive and vegeta- 
tive tendencies. This and other transformations of one sort of floral 
organ into another may be compared to changes in sex expression that 
have been found to occur after hormone treatment (Love and Love, 
1946 ) or as the result of photoperiodic change. 

In other organoid galls, abnormality consists in the formation of struc- 
tures in places where they do not ordinarily occur. Thus ovaries may 
appear in normally staminate flowers, stamens within an ovary, or 
ovules on its surface. In the well-known case of Lychnis vespertina at- 
tacked by the smut fungus Ustilago antherorum, stamens are produced 
in the female flower ( Strasburger, 1900). Flowers and cones may pro- 
liferate into vegetative shoots after attack by mites or for other reasons. 
Cecidomyia causes the formation of rootlets on the stalk nodes of Poa 
(Beyerinck, 1885) and Eriophyces fraxini small, shortened shoots on 
the leaves of Fraxinus. The attack by mites sometimes results in the 
growing out (enation) of small leaves on large ones or the formation of 
extra perianths. 

Familiar examples of abnormal development which may also be classed 
as organoid galls are "witches'-brooms" (Fig. 11-4), dense clusters of 
small, much-branched shoots, chiefly on woody plants and resulting from 
excessive production of buds which grow immediately into shoots. At- 
tacks by mites and various fungi, especially Exoascaceae and Uredineae, 
are often the causal agents, though in many instances no parasites are 
known to be involved. The physiological basis for the development of 
such structures has been thought to be the accumulation of nutrients, 
though doubtless there are other factors. 

A rather extensive literature has grown up about the character and 
causation of witches'-brooms, for which the reader is referred to Solereder 
( 1905 ) and Liernur ( 1927 ) . Liernur cites 96 instances the causes of which 
are known, occurring on 49 species of plants in 19 families, and 51 cases 
of unknown origin. They differ in morphology, anatomy, and etiology but 
resemble each other in their general character of copious, compact 
branching. As compared with normal structures, the tissues of their leaves 
and branches tend to be somewhat less highly differentiated, thus ap- 
proaching cataplasmatic galls, though mechanical elements are often well 

Fasciations. A special type of abnormalities of a rather distinct kind 
and which may be classed with organoid galls are fasciations. These are 
cases where a normally cylindrical or radially symmetrical plant part be- 
comes flattened and elliptical in cross section to form ribbon-like or some- 
times ring-like structures. The origin of fasciation has attracted a good 

Abnormal Growth 


deal of attention among morphologists. Much of the literature is reviewed 
in the papers of Schoute (1936), Bausor (1937), and O. E. White (1948). 
The term fasciation, like others in the field of abnormal growth, has 
been applied to a rather wide variety of phenomena which probably have 
different origins even though the final result in all of them is a flattened 
structure. Frank (1880) distinguished between fasciations that arise by 
a gradual expansion of the growing point in one plane and others that 
come from lateral fusion or connation between two or more separate 

Fig. 11-4. "Witches'-broom" on a fern leaf. (After Giesenhagen.) 

structures, in natural self-grafting. Schoute believes that the term fasci- 
ation should be limited to Frank's first category, and this is now the com- 
monest practice. The difference between the two concepts can be deter- 
mined only by developmental studies. Johansen (1930) has shown that 
the genesis of fasciation may be recognized even in early embryogeny. 

Fasciation is most conspicuous in the main-shoot axis but occurs also 
in roots (Schenck, 1916) and may be found in almost all parts of the 
plant. It may vary from one part to another. Many abnormalities resem- 
bling fasciation are to be seen in leaves, such as "double" leaves and 

282 The Phenomena of Morphogenesis 

Certain instances of fasciation have received special study. It has long 
been observed that if the epicotyl of a seedling of Phaseolus multiflorus 
is cut off and the buds in the axils of the cotyledons are then allowed to 
grow they will produce flattened shoots. These later revert to a cylindrical 
form and produce normal branches. This is an unstable kind of fasciation, 
and the factors responsible for it are not yet clear (Sachs, 1859; Klebs, 
1906; Georgescu, 1927; Bausor, 1937). 

Fasciations resulting from other types of mutilations and from wounds 
are discussed by Blaringhem ( 1908 ) . They frequently follow pruning in 
Salix, Populus, Robinia, Tilia, and Corylus. 

Various other factors, both external and internal, have been found to 
produce fasciation. Growth substances applied in high concentration may 
induce such vigorous local growth that ribbon-like structures result 
(Bloch, 1938). In strawberry varieties that have a tendency toward fas- 
ciation, this is emphasized by short day-lengths ( Darrow and Borthwick, 
1954). It has been suggested that fasciation is due to a high level of 
nutrition, to insufficient nutrition, to changes in correlation produced 
by growth substances, and to various factors that have been shown, 
in both normal and regenerative growth, to induce correlative disturb- 

Of especial interest are those instances where fasciation, or a tendency 
toward it, is inherited. The best known of these is in the cockscomb, 
Celosia cristata, where the inflorescence is often much flattened. Heredi- 
tary fasciation has also been found in Nicotiana (O. E. White, 1916), 
Pharbitis (Imai and Kanna, 1934), and Phlox drummondii (Kelly, 1927). 
The large-fruited varieties of tomato with more than two carpels may be 
regarded as examples of genetic fasciation. 

Fasciation is found in all groups of vascular plants. Penzig ( 1921 ) has 
reported it for Psilotum and Lycopodium; Kienholz (1932) for Pinus, and 
Schenck ( 1916 ) for other gymnosperms. It is widespread in angiosperms, 
both dicotyledonous and monocotyledonous, and many instances are cited 
by Masters. It is particularly frequent in Taraxacum, Antirrhinum, and 
Delphinium and has even been found in the giant cactus, Carnegiea 

Pelory. Pelory (or peloria) is a type of floral abnormality, quite dif- 
ferent from the others that have been mentioned, in which dorsiventrally 
symmetrical (bilabiate) flowers become radially symmetrical. It was first 
discovered by Linnaeus in Linaria vulgaris, where it may frequently be 
found in nature, and has been studied by Sirks ( 1915 ) and others ( Fig. 
11-5). In this species peloric flowers, where they occur, are found at the 
base of the inflorescence and may show a transition to normal ones at the 
top. This is perhaps another instance of phasic development. Peloric 
flowers have also been investigated in Linaria spuria, Antirrhinum majus, 

Abnormal Growth 283 

and Digitalis purpurea. Pelory is frequent in Labiatae, where it usually 
occurs at the top of the inflorescence. 

As to the cause of pelory, opinion is divided. In cases such as Digitalis 
the difference between the bilabiate and the radial condition is clearly 
induced by gravity for it can be reversed experimentally. Peyritsch found 
that in Labiatae it may be induced by strong illumination. Sachs believed 
that physiological factors are operative in causing it. There are many 
cases, however, reported by Vrolik, Darwin, de Vries, Baur, Lotsy, and 
others, where this type of abnormality, or at least a tendency to produce 
it, is inherited. 

Fig. 11-5. Peloric flower of Linaria 
vulgaris, almost radially symmetrical. 
(From Goebel.) 


In organoid galls and similar types of abnormalities that have been dis- 
cussed, structures recognizable as those found in normal plants are 
present, though their size, form, and arrangement may be altered and the 
general pattern of organization distorted. In the group now to be consid- 
ered the structures produced are entirely novel and cannot be placed in 
any such morphological category as leaves or stems. They can hardly be 
regarded as organs, but they are much more than tissue abnormalities 
for they have a constant and specific form, size, and structure and a very 
considerable amount of histological differentiation. They result from the 
attack of parasites. In most cases they undergo a definite period of devel- 
opment, or life cycle, correlated with that of the parasite and thus are 
different from the cataplasmatic abnormalities to be discussed later. In 
Kiister's terminology they are prosoplasmatic galls. There is no sharp line 
between these and the simpler gall types but their large number and 
definite forms set them apart as a rather distinct group. The majority are 
zoocecidia (galls formed by animals). The most conspicuous and best 
known owe their origin to parasitism by the gall wasps or cynipids, a 
family of the Hymenoptera. Others are produced by flies of the gall 
midges and simpler ones by mites. A few are the result of fungus parasites. 

284 The Phenomena of Morphogenesis 

Prosoplasmatic galls are so numerous and so remarkable in many re- 
spects that they have received much attention and are the basis of an 
extensive literature. They are discussed in many papers and books by 
Beyerinck (1883), Molliard (1895), Magnus (1914), Thompson (1915), 
Felt (1917), Kostoff and Kendall (1929), Kiister (1930, 1949), Ross 
(1932), Carter (1939, 1952), and others. A typical example has been 
described in detail by Hough ( 1953 ) . 

The small galls produced by some of the rusts or by Synchijtrium pilifi- 
cum on the leaves of Potentilla may perhaps be included among proso- 
plasmatic galls, as may those formed on the petioles of Populus by 
Pemphigus bursarius. Here the gall is simply a mass of expanded epider- 
mal and cortical cells which have divided anticlinally. 

Fig. 11-6. Insect galls on leaf of rose. (From Wells.) 

There is a higher degree of organization in cynipid galls, and they have 
received much more attention than any others (Fig. 11-6). In these the 
female wasp deposits an egg in the body of the plant, where the larva 
develops, and the gall results from the reaction of plant tissues to stimula- 
tion from the egg and the developing larva. Such galls possess a concen- 
tric type of organization. The histology of these structures is as varied as 
their form (Figs. 11-7, 11-8). Some are relatively simple but others con- 
sist of three, four, or even five different types of tissue. Some of these 
tissues show adaptation to specific functions such as mechanical support, 
storage, and aeration. The mechanical tissues are of particular interest be- 
cause of their relation to the position of the larva within the gall and the 
means of its escape. 

The special morphogenetic significance of these galls is that in them a 

Abnormal Growth 


specific modification of the tissue pattern of the host plant is caused by 
the presence in it of an egg and larva of the parasite. A given cynipid will 
always produce the same kind of gall on a given plant species, and the 
galls induced by different wasps on the same plant are quite dissimilar. 
On Celtis occidentalis, Carter found 17 different sorts of galls formed by 
17 species of wasps. Each type of gall is related to the character of the 
larva that develops within it. 

Doubtless the formation of these formed galls results from a specific 
stimulus coming from the wasp or the growing larva and a specific re- 
sponse by the tissues of the host plant, but how such a subtle control of 

Remains of the egg membrane 

Vein of the 2nd order of branching 
Ruptured lower epidermis 

Leaf upper epidermis 

Vein connecting leaf and gall 

Cup of platform cells 

Additional sclerenchyma ring 

Funnel-shaped sclerenchyma mass 

Nutritive tissue cells 

Plate-like sclerenchyma mass 

Gall cavity containing larva 

1 mm. 

Large air-spaces region 
Gall vascular stand 

Stellate hairs 
Gall parenchyma 
Gall-stalk sclerenchyma 
Gall epidermis 

Fig. 11-7. Diagram of a longitudinal section through a cynipid gall on the leaf of oak, 
showing its specific form and considerable structural differentiation. (From Hough.) 

the morphogenetic potencies of the host cells is exerted by the parasite is 
unknown. Various theories have been proposed and many experiments 
performed to throw light on this problem. It has been suggested that the 
gall-inducing stimulus is a mechanical one, but this seems rather unlikely, 
and most workers now believe that the stimulus is chemical in nature, 
perhaps an enzyme or a specific formative substance. 

Many investigators have tried to extract the gall-forming agent from 
the insect, inject it into a plant, and thus produce a gall artificially, but 
earlier attempts all failed. Recent ones have been more successful. Parr 
(1940) and Plumb (1953) injected extracts from the salivary glands of 
coccids and an aphid into young needles of Norway spruce and induced 
the formation of galls much like those normally produced by these insects. 

286 The Phenomena of Morphogenesis 

K. M. Smith (1920) had shown that the damage to leaves of apple by 
capsid bugs was caused by secretions of the salivary glands, and this sug- 
gested the possibility that the glands of the gall-producing larva secrete 
an enzyme which calls forth a specific growth reaction in the host tissue. 
Parr demonstrated the presence of enzymes in the gland extract but found 
that these did not stimulate gall formation when sterilized. Martin 


Fig. 11-8. Section through portion of an insect gall showing modification of normal 
leaf structure (left). The larval chamber is surrounded by mechanical tissue. (From 
Kiister. ) 

(1942), however, produced abnormal growth in sugar cane with steri- 
lized extracts. 

Substances produced by the growing larva rather than those injected 
with the egg are probably most important in gall formation. Little is 
known about these substances, however, or the place and manner in which 
they are introduced into the tissues of the plant (Kostoff and Kendall, 

A morphological problem of some interest is whether these galls and 

Abnormal Growth 287 

the tissues that compose them are really to be considered as "new" 
structures, morphologically different from the familiar categories. Cer- 
tainly they have cells and tissues unlike any normally found in their host 
plants. Such are the various hairy projections on the surface of certain galls, 
the opening mechanisms, the mechanical tissues, and others concerned 
with nutrition and aeration. As to whether these are "new" or not depends 
on our definition of that term. A given morphological category, such as 
the leaf, presumably has a continuous evolutionary history beginning 
with early vascular plants. It is part of the norm of plant structure. In 
this sense such galls are certainly new. They have arisen, however, be- 
cause of a novel factor in the plant's environment-the gall wasp. It seems 
probable that if such wasps had existed in the Paleozoic they would have 
induced galls in the vascular plants of that era. Gall-making ability on 
the part of the wasp is advantageous to it and doubtless has been devel- 
oped by selection, but the gall response by the plant to the parasite is not. 
Presumably the capacity for producing these galls has long been among 
the developmental potencies of plants. It can be regarded as new histori- 
cally but not morphologically. 

From the morphogenetic point of view the most important fact about 
these prosoplasmatic galls is that they are highly organized and specifi- 
cally formed structures induced by an outside stimulus. Here we can see 
the process of form determination manifest in a simpler fashion than in 
normal development, for here the inducing agent is not a part of the 
developmental mechanism but is introduced into the organism. If we 
understood exactly how these galls are formed, we should doubtless gain 
some important clues as to the morphogenetic process generally. Some- 
thing more is involved here than partial loss of organization, the change 
that takes place in most other galls. Here is operating a constructive, not 
a degenerative, process. 


In the two previously discussed categories of abnormal growth the 
original organization was either present in altered form or something 
entirely different from the normal was produced. There is a third group 
in which organization at its highest level, with the production of specifi- 
cally formed structures, is inoperative to a large degree, and only form- 
less, or amorphous, structures are developed. Within these structures 
there may be some histological differentiation but it is much less than in 
normal tissue. The cells remain alive, however, and perform many normal 
physiological activities. They must still possess a certain amount of organ- 
ization, evident in regulatory action. If this were not so, death would 

288 The Phenomena of Morphogenesis 

There are various sorts of amorphous structures which are not easy to 
distinguish from each other, but a number of categories may be recog- 
nized, such as intumescences, callus, tumors, and galls. The terms "gall" 
and "tumor" have no very precise meanings, but a gall is most commonly 
regarded as an anomalous growth due to an attack by a parasite and a 
tumor as one which results from other causes, though there are many 
exceptions to these definitions. 

Intumescences. The simplest sort of amorphous abnormal growth is one 
in which a group of cells at the surface of an organ expands into a wart 
or pustule. These are termed intumescences (Sorauer, 1899) and result 
from various causes (Wallace, 1928). Sometimes, as on the leaf of cab- 
bage (Von Schrenk, 1905), they are groups of watery (hyperhydric) cells, 
swollen by excessive absorption of water resulting from contact with 
spray materials or other substances. In woody plants intumescences 
usually are formed by proliferation of cork cells, sometimes from lenticels 
and sometimes elsewhere. Such intumescences involve an increase in the 
cell number (hyperplasia) as contrasted with an increase in cell size 
(hypertrophy) as in the cabbage leaf. They may result from exposure to 
ethylene gas and other substances. Intumescences often resemble natu- 
rally occurring corky spots on certain plants, especially at lenticels. 

Callus. As a result of wounding, a layer of cork cells is usually produced 
over the wound surface through the action of wound hormones (p. 402). 
This perhaps is not to be called "abnormal" tissue in the ordinary sense, 
since it is very common and indeed accompanies bark formation in trees, 
where new cork layers cover the breaks resulting from expansion of the 
axis. In many cases, however, these do not occur in the intact, uninjured 
plant and may best be included among abnormal structures. 

In cases of more serious wounds, as where a cutting is removed for root- 
ing, something more complex than a few layers of cork cells is formed 
at the surface. Here often develops callus, an amorphous mass of rather 
large-celled, loosely arranged parenchymatous tissue, produced by cell 
division in the ground tissue or more commonly from cambium (Fig. 
11-9). Its elements show relatively little differentiation, but there may be 
some meristematic growth near the surface. For its nutrition, callus de- 
pends on food from the normal tissues beneath it. 

A callus may reach considerable size but its mass has no definite form 
and there is little morphogenetic control over its growth. In its later 
stages, callus may undergo various types of development depending on 
the kind of plant, the location of the wound, and the external conditions. 
Often some differentiation appears in it, and cells are produced resem- 
bling those of normal tissue. Isolated nests of single cells or groups of 
cells may develop into tracheid-like elements, usually with reticulate pit- 
ting. Sclereids are formed in the same way. These irregular nests have a 

Abnormal Growth 289 

characteristic appearance, and their presence often indicates the origin 
of a meristem in the callus. 

The structure and arrangement of the cells in such wound tissue as 
callus are often atypical, especially near the wound itself. The cells twist 
and turn, and so-called "whorls" thus formed have been described by 
Maule (1896), Neeff (1914), and others. Krieg (1908) observed several 
concentric circles of cambium in the pith of ringed branches of Vitis 
adjacent to a wound. In the outer ring, development was inverted, phloem 
being formed toward the inside and xylem toward the outside. 

Fig. 11-9. Callus on cut stem of Cleome produced by application of growth substance. 
( Courtesy Boyce Thompson Institute. ) 

Save under particular conditions, a callus does not remain callus 
indefinitely but tends to produce normally organized structures again. 
This it does by means of new apical meristems, both of shoots and roots, 
which frequently appear in it. Such meristems arise in several ways. Cells 
abutting on nests of tracheids may produce a meristem in the form of a 
hollow sphere. Others may appear elsewhere in the callus or where it is 
in contact with the pith, cambium, or cortex. These meristems may form 
secondary wood or phloem. Meristems of roots tend to arise well below 
the surface and those of shoots either at or just below the surface. From 
such meristems typical organs may be regenerated in any region of the 
callus, thus showing that there have been no fundamental changes in 
the genetic character of the callus cells themselves. Their potencies to 
produce typically organized structures have been masked but not lost. 

290 The Phenomena of Morphogenesis 

Structures much like calluses may be induced by other factors than 
wounding, especially by various chemical substances. Among these sub- 
stances are ether, chloroform, camphor, ethylene gas, liquid paraffin, and 
especially various growth substances (p. 407). Many are nonspecific in 
their action, and it is to be assumed that their effect is primarily one of 
injury to the tissues, which produces changes in the cells and thus leads 
rather indirectly to atypical growth reactions. Others, especially the 
growth substances, produce rather specific reactions. 

One of the properties of the synthetic growth substances (such as in- 
doleacetic, indolebutyric, and naphthaleneacetic acids and paraffin) is 
their action in stimulating an increase in the number and the size of cells 
to which they are applied. Sizable overgrowths and calluses may thus be 
formed on various plant organs if sufficiently high concentrations of these 
agents are applied; and if this treatment is repeated, tumor-like masses 
are produced (Schilling, 1915; Brown and Gardner, 1936; Kisser, 1939; 
Levine, 1940; and others ) . These often resemble the ones associated with 
certain bacterial infections. 

Callus-like overgrowths are more readily obtained if the paste in which 
the inducing substances are carried is applied to the more sensitive 
regions such as those near the shoot tips, particularly after decapitation. 
Many experiments have been performed, especially on herbaceous dicoty- 
ledons, to test the effects of various growth substances on development 
(see especially the publications from the Boyce Thompson Institute and 
those of E. J. Kraus and his colleagues at the University of Chicago, 
p. 405). Differences have been found among the growth substances in the 
character and extent of abnormal growth they induce, and their effects 
are also related to the kind of plant, its age, and the region treated. His- 
tological study of these tumors shows that they resemble other calluses 
and wound tissues, especially in the absence of a constant form or size, 
the presence of some cellular differentiation as the tumor ages, and the 
development on them of root and shoot primordia. 

Amorphous Calls Produced by Parasites. Amorphous structures which 
in certain respects resemble intumescences, calluses, and chemically in- 
duced tumors are caused by various parasites such as nematodes, mites, 
insects, fungi, and especially bacteria. Even viruses are now known to be 
involved in their production (Black, 1949). There is a great variety of 
these, from small, simple structures to large and relatively complex forms. 
Some of the huge burls on trees are due to parasites but others apparently 
are caused by mechanical or other nonparasitic factors. Many amorphous 
galls have a somewhat more highly organized character than others but 
they show little constancy in size or form and their histology is less 
regular and their differentiation simpler than in normal structures. Kiister 
describes and figures many of these (1903a, 1911, 1925). He gave 

Abnormal Growth 


them the name by which they are now commonly known, cataplasmatic 


Crown Gall. Here belongs the gall that has been studied more inten- 
sively than any other, crown gall (Fig. 11-10). This is produced on a 
wide variety of plants, at least 142 genera in 61 families, by the bacterium 

Fig. 11-10. A crown gall on sunflower. (Courtesy Department of Plant Pathology, 
University of Wisconsin. ) 

Agrobacterium tumefaciens. There are reviews of work on such galls by 
Levine (1936); Riker, Spoerl, and Gutsche (1946); de Ropp (1951a); 
Klein and Link (1955); and Braun and Stonier (1958). Crown gall has 
been investigated by Erwin F. Smith and his colleagues ( 1911, 1917, and 
many others) and later by various botanists among whom Braun, 

292 The Phenomena of Morphogenesis 

Gautheret, Levine, Riker, de Ropp, and P. R. White have been especially 

The inception of a crown gall seems always to come by means of a 
wound. In the early stages of this infection the reactions of the host cells 
produced either directly or indirectly by the parasite are much like wound 
reactions, but in the young gall the new cell walls soon lose the regular 
arrangement found in wound tissues. Reparative wound calluses and those 
formed at grafts often resemble the early stages of crown gall, but as the 
gall develops, rapid cell division occurs in the outer layers and a large 
mass of callus is formed. The great difference between crown gall and 
ordinary callus, however, is that the latter is self-limiting and soon be- 
comes quiescent whereas gall tissue is capable of indefinite and amor- 
phous growth. This is a fact of particular morphogenetic significance. 
There is now good evidence that the cells of crown gall have undergone 
a permanent change in character. This seems to involve an acquirement 
of the capacity for autonomous growth, which may result, Rraun ( 1958 ) 
believes, from the permanent activation of a series of systems by which 
growth substances are synthesized. In normal cells these systems are 
precisely regulated and growth ultimately stops. Normal cells in cul- 
ture require auxin from an outside source but crown-gall cells do not. 
Crown gall thus differs from most other galls, which are self-limiting and 
do not grow indefinitely. In some cases the crown gall matures, stops 
enlarging, and undergoes some histological differentiation. Nests of vascu- 
lar cells, chiefly xylem, appear in it (Fig. 11-11), and the primordia of 
roots and shoots may develop. The cytological and histological changes 
in the development of a typical crown gall have been described by Ther- 
man (1956) and Kupila (1958). 

The process of conversion of normal cells into tumor cells is a gradual 
one. Its inception depends both on a wound stimulus and on the pres- 
ence of an auxin ( Rraun and Stonier, 1958 ) . Tumors differ in appearance, 
in the degree of their organization, and in their capacity for growth. These 
differences may be the result of their location on the plant, the virulence 
of the strain of infecting bacteria, or other factors. Crown-gall tissues 
can be grafted into normal ones, and these may be carried through an 
indefinite series of graft transfers. There is no good evidence that they 
induce adjacent normal tissue to form tumor cells, though temporary 
alterations may take place there. Gall tissues can readily be grown 
in culture and can then be grafted back to normal ones and form galls. 
In all these cases their cells remain unchanged. 

In a few plants secondary tumors may develop, often at some distance 
from the original gall or primary tumor. In certain cases this results from 
an infection near the apical meristem and a subsequent separation of 
the secondary from the primary gall by growth. Sometimes, however, 

Abnormal Growth 


secondary galls develop after growth in length is over but always in close 
association with the xylem, as though the latter were the pathway of 
induction. Secondary tumors behave much like primary ones in grafts 
and in culture, and there seems to be no very fundamental difference 
between the two. 

Remarkably enough, many crown galls, particularly secondary ones, 
seem to be free from bacteria. There is evidence, however, that bacteria 
must always be present at the very beginning of tumor growth but 
that they soon disappear. Braun and White (1943) made Vinca rosea 

Fig. 11-11. Section of a young crown gall on Pelargonium, showing a nest of vascular 
cells. ( From Noel. ) 

galls free of bacteria by heat treatment. Such tissues retained their gall- 
producing properties when grafted into healthy plants (White, 1945). 
Although bacteria are required for the inception of crown gall, once the 
change is induced they no longer seem necessary for the growth of gall 

Crown gall (and presumably other amorphous galls of this general 
type) does not result from a single cause but involves a series of factors. 
Klein and Link ( 1955 ) discuss this in their extensive account of the 
etiology of crown gall (Fig. 11-12). There is first a conditioning phase, 
perhaps induced by wounding and involving wound hormones. This 
makes the cells susceptible to conversion into tumor cells. It is possible 

294 The Phenomena of Morphogenesis 

that the activation process in ordinary wound healing and in the incep- 
tion of tumors may be the same. This is followed by an induction phase 
in which a tumor-inducing substance of some sort enters the host from 
the bacterium. A heat-labile product of virulent crown-gall bacteria has 
been found to alter conditioned cells into incipient tumor cells. How this 
is done is not clear. The substance may itself be the agent of change, 
possibly a virus or a macromolecule of DNA or even a gene or a 
hereditary agent in the cytoplasm; or it may induce the change by 
causing gene mutation or the production of permanent, self-reproduc- 
ing bodies, sometimes called plasmoids. Finally, in the promotion phase 
the gall grows to completion. Here auxin is involved, in the promotion of 
an incipient into a primary tumor cell, in the multiplication and per- 







Tumor- inducing 



* t.t 


I principle i | 

Mill' 1 

- $■ *" 






Fig. 11-12. Diagram of probable interrelations of various factors in the transformation 
of a normal cell into a primary tumor cell. ( From Klein and Link. ) 

haps the differentiation of tumor cells, and in causing various host effects 
which accompany tumor formation. The physiology of crown-gall for- 
mation has been further discussed by Klein (1958). 

A question often raised is whether crown gall and its derivatives are 
really plant cancers, as Smith vigorously maintained they were, or if 
something different from true cancer is here involved. This question 
has been discussed by Levine (1936), White and Braun (1942), and 
others. It should be remembered that such a condition as malignancy 
is difficult to define in the same terms in organisms as different in struc- 
ture and organization as plants and animals. The unrestrained, invasive 
type of growth characteristic of animal cancer, with its metastases and 
lethal quality, could hardly be expected in a plant, which has no true 
circulatory system and lacks the high degree of organization that makes 

Abnormal Growth 295 

animals so vulnerable. The animal cancer cell has lost its specificity and 
become, so to speak, an independent parasitic entity of unlimited 
growth. What this change involves and what causes it are still not 
understood. Crown-gall tumor cells are certainly in this same category 
for they grow indefinitely and do not depend on the continual presence 
of the factors that induced them. The true cell invasions and metastases, 
in which bits of cancer tissue are carried away to other parts of the body 
and there develop new centers of malignancy, are absent in plants, but 
transfer of gall tissue from place to place by grafting is readily ac- 
complished. Many students of the problem are inclined to regard crown- 
gall tumors as basically no different from animal cancers. It is obvious 
that these examples of abnormal growth provide some of the best ma- 
terial known for a study of the way in which the higher levels of or- 
ganization in the plant are broken down. For students of morpho- 
genesis they long have had a particular interest. 

Root Nodules. Another type of cataplasmatic galls rather different 
in character from the others here described and of much practical im- 
portance to man are the nodules formed on the roots of leguminous 
plants from the invasion of their cortical tissues by species of Rhizobium. 
They are an example of what has been called "controlled parasitism," 
for the relation between this bacterium and the plant may better be 
regarded as symbiosis rather than parasitism since the host plant ob- 
tains an advantage because of the atmospheric nitrogen fixed by the 
bacteria. These nodules have a higher degree of organization and pro- 
duce more specialized structures than do most cataplasmatic galls 
and perhaps should be included under prosoplasmatic ones (Allen and 
Allen, 1953). The particular character of the nodule depends upon the 
host plant and the species of bacterium that invades it. As in crown 
gall, auxin action mav here be involved. 

Abnormal Growth Due to Other Causes. Manv cases have been re- 
ported of abnormalities due to other factors than parasitism or chemical 
stimulation. X rays may produce them ( Sankewitsch, 1953), as may 
ionizing radiations (Gunckel and Sparrow, 1954). Some resulted from 
the A-bomb tests in the Pacific (Biddulph and Biddulph, 1953). 

In some plants tumors arise from no recognizable cause and are 
presumably due to somatic mutations or to a modification of organized 
development by other genetic factors. The best known case is that of 
the tumors occurring spontaneously in hybrids between Nicotiarui 
glauca and N. Langsdorfii (Kostoff, 1930a; Kehr and Smith, 1954). These 
are small amorphous structures appearing on stems and branches, and his- 
tologically resembling wound callus and crown gall. Kostoff believes that 
they are due to a disturbed growth balance, either in nucleus or cyto- 
plasm, between these two particular species. These tumors, removed 

296 The Phenomena of Morphogenesis 

from the plant, are the ones used by P. R. White in his first tissue cul- 
tures (1939). He found later that they retained their specific proper- 
ties in culture for years and continued to grow as tumors when grafted 
into young stems of Nicotiano glauca (1944). Satina, Rappaport, and 
Blakeslee (1950) studied the development of somewhat similar tumors 
appearing in fertilized ovules from incompatible crosses in Datura. 

Changes in cellular character are sometimes associated with abnormal 
growth. Prothallia of some ferns, when grown in culture, often produce 
various types of proliferations (Partanen, Sussex, and Steeves, 1955). 
Some of these remain essentially prothallial in character and show no 
fundamental deviation from normal. Their cells are still able to re- 
generate normal prothallia again. Certain tumor-like forms, however, are 
modified much further, for they have lost this ability. This loss is ac- 
companied by a modification in cellular character, visible as an increase 
in chromosome number from In to 3n or 4n. Such forms may be com- 
parable in a sense to crown gall. White and Millington (1954) have 
described a woody, nonbacterial tumor in spruce which begins in a 
single cambium cell. This has been altered, physiologically or geneti- 
cally, and forms a mass of abnormal tissue. The plant becomes what is 
essentially a sectorial chimera. 

Various aspects of the problem of plant tumors have been discussed 
by P. R. White (1951), de Ropp ( 1951a ), Klein and Link (1955), and 

Tissue Cultures. Tissue cultures can hardly be called tumors or galls, 
but in them the normal organization of the plant has disappeared to a 
greater extent than in any other case here discussed. A book on mor- 
phogenesis is not the place to consider this subject in any detail but it 
does have some important morphogenetic implications that should be 
mentioned. So-called tissue cultures of plant material on sterile media 
have been studied actively in recent years, and for an account of them 
the reader is referred to the publications of the pioneers in this field, 
especially P. R. White, Robbins, Gautheret, and Nobecourt. 

Animal tissues have been cultured for half a century but it was much 
more recently that this was done successfully with plants. The problems 
involved were first clearly stated by Haberlandt, who himself failed to 
grow isolated cells from higher plants in artificial media. A necessary 
prerequisite for the success later attained was the development of 
satisfactory media consisting of pure substances of known chemical 
character, including salts, carbohydrates, organic nitrogenous materials, 
vitamins, and growth substances. The media developed by the early 
plant workers were superior to the sera and other complex and little 
understood ones previously used by tissue culturists. 

The first plant cultures were not tissue cultures in the strict sense that 

Abnormal Growth 


they consisted of only one kind of cells, as in many animal cultures. 
Most were really organ cultures. Those grown from root tips can be 
carried through an indefinite number of subcultures and produce large 
masses of root tissue. Shoot axes may be cultured in the same way 
from apical meristems, and Nitsch (1951) and others have succeeded 
in growing fruits from small primordia. Leaves can be grown to maturity 
in the same way (Steeves and Sussex, 1957), as also can ovules 
(Maheshwari, 1958). These organ cultures have given much informa- 
tion on the nutritional requirements of various parts of the plant and have 

Fig. 11-13. Culture of stem callus of tobacco on nutrient agar, six weeks after trans- 
fer to new medium. (Courtesy Department of Plant Pathology, University of Wis- 
consin. ) 

been of importance for an understanding of their physiology. Wet- 
more (p. 222) grew fern plants from shoot apices in culture and found 
pronounced morphological effects of differences in the medium, a re- 
sult of much morphogenetic significance. 

Something closer to a true culture can be attained by growing calluses, 
tumors, parenchyma, and bits of tissue from the cambial region (Fig. 
11-13). Structures much like amorphous tumors and galls result. In these 
the strict morphogenetic control is relaxed and the explanted material 
may be grown in unlimited quantities by subculturing. Such cultures 
have been made of spontaneous tumors and of secondary crown gall. 

298 The Phenomena of Morphogenesis 

Gautheret (1945) showed that with ordinary plant material, as in carrot, 
endive, and various woody plants, a callus must first be allowed to de- 
velop from an excised piece and that this callus could then be cultured. 

There are many differences between species and between different 
parts of the same plant as to growth and structure of the culture pro- 
duced. Pieces of tissue from the cambial region, for example, which 
retain some degree of organization for a time, gradually lose it in 
successive transfers. Some cultures finally become essentially homo- 
geneous masses of parenchyma. In a few cases, however, patches and 
whorls of tracheid-like tissue appear in these, as do meristems of roots 
and shoots. The conditions under which this differentiation occurs vary, 
depending in part on the source of the material and especially on the 
character of the nutrient medium and the age of the culture. 

It is now possible to produce plant cultures where the cells are not 
united into masses but grow and divide as individual cells, much as they 
do in certain animal cultures. This may be thought of as representing 
the lowest level of plant organization, the loss of all relationships above 
the single cell. Among those who have done pioneer work in this 
field are Muir, Hildebrandt, and Riker (1954, 1958); de Ropp (1955); 
Nickell (1956); Reinert (1956); Torrey (1957a); and Tulecke (1957). 

Of particular interest for morphogenesis is the culture work of Steward 
and his collaborators (1958 and p. 75). Using a basic liquid culture 
medium supplemented by coconut milk and continually rotated, they 
grew small explants of carrot tissue taken from the phloem of the 
root. Single cells here often became separated and floated freely, dividing 
very irregularly and growing into small cell aggregates. When these 
reached a certain size, the cells at the middle of the aggregate began to 
show differences from the rest, and some developed as tracheid-like struc- 
tures, surrounded by a ring or sheath of cambium-like cells. From this 
center, one or more root meristems developed. When such an aggregate 
is grown on the surface of an agar medium it produces a shoot meristem 
and in time a carrot plant. So far, a whole plant has not been produced 
directly from a single cell but only by way of a cell aggregate. It is sig- 
nificant that although bits of carrot phloem tissue, placed directly in 
culture with coconut milk, may form large masses of callus-like tissue, 
they much less readily produce roots and shoots, perhaps because of 
inhibiting substances still present in them. 

Radical though the changes are which plant tissues display when 
cultured, there is no evidence that any permanent or genetic effect is 
produced in them or any new developmental potencies induced. A con- 
siderable degree of morphogenetic control has been relaxed, just as it 
has been in amorphous galls. 

It is evident that there is an immense amount of information available 

Abnormal Growth 299 

about abnormal growth but that students of morphogenesis have not as 
yet made very much use of it for their purposes. Three facts, however, 
which emerge from a study of this subject have already proved to be 
of much morphogenetic significance: 

1. The actual developmental potencies of most plant cells are far 
wider than ever come to expression in normal development. 

2. It is possible to break down the organization of the plant body into a 
series of successively lower levels and then to restore normal organiza- 
tion again. 

3. This can be done without modifying the genetic character of the 


A continued study of these facts, and of others in this field, will cer- 
tainly prove very fruitful. Abnormal development is only development 
under unusual conditions, and the wider spectrum of morphogenetic 
information thus made available provides the student of development 
with a powerful tool for the study of some of his most difficult problems. 


Morpho genetic Factors 


Introduction to Factors 

In earlier chapters various morphogenetic phenomena were discussed, 
but relatively little was said as to the factors that produced them. There 
now remains the task of relating these phenomena to changes in the 
outer or inner environment of the plant and attempting to account for 
their origin. This is really a part of the broader field of plant physiol- 
ogy, and no sharp line can be drawn between the two. Much of physiol- 
ogy, particularly those parts of it that deal with the various metabolic 
processes, is not of primary interest for morphogenesis. Other parts of 
it, however, such as photoperiodism, vernalization, auxin action, water 
relations, and the carbohydrate-nitrogen ratio, for example, have much 
significance for the morphogenetic phenomena of polarity, differentia- 
tion, regeneration, and others. To present these morphogenetic implica- 
tions adequately would mean going more deeply into plant physiology 
and its vast literature than can be attempted in the present volume. No 
discussion of the problems of plant morphogenesis would be com- 
plete, however, without some mention of the physiological factors which 
influence development so powerfully. The purpose of this final section 
of the book is to introduce the reader to the more important of these 
factors and to provide him with an entry into the literature of the sub- 
ject. No attempt will be made to discuss them thoroughly from the point 
of view of plant physiology. 

It is first necessary to consider the relation between the two chief 
sorts of factors— environmental and genetic. A living plant is an or- 
ganized system maintaining itself in a complex and changing environ- 
ment. Its genetic constitution (or genotype) remains unaltered save for 
occasional doubling of the chromosomes in local areas or the rare oc- 
currence of somatic mutations. Despite this, the plant does not remain 
unchanged. Its appearance (or phenotype) is often greatly modified as 
the environment is altered, and we commonly say that this change is 
the result of an environmental factor. So, in a sense, it is, but there is 
often difficulty in disentangling the effects of heredity and environment 
in morphogenetic changes. One should remember that both are always 


304 Morphogenetic Factors 

operating. A visible trait is the developmental reaction of a specific 
(and constant) genetic constitution to a specific environment. Every 
trait is therefore inherited since it will always be produced if the en- 
vironment is of a certain sort. In some traits the expression of the genetic 
constitution is essentially the same under a wide range of environments. 
The relative position of the floral parts, for example, the arrangement 
of the leaves, or the character of the pitting on the side walls of the 
vessels in the wood is usually quite unchanged under various condi- 
tions of light, moisture, temperature, or auxin concentration. Such 
traits, for this reason, are especially useful in taxonomy. Others, such 
as the height of the plant, the thickness of the cuticle on its leaves, and 
whether it flowers or not, may be very different under different condi- 
tions of nutrition, water supply, and photoperiod. Such traits 
are usually said to be determined by environmental factors. Actually, 
both types of traits are inherited and both are environmentally deter- 
mined. In the former, the repertoire of responses of the genetic constitu- 
tion to changes in the environment is relatively meagre whereas in the 
latter it may be very wide. Under a given length of day, for example, 
salvia plants will flower but lettuce plants remain entirely vegetative. 
What promotes or inhibits flowering is not simply the day-length but 
the different inherited responses of these two plants to this day-length. 

Where the developmental response of a plant varies widely under 
different environments as it often does when such factors as light or 
water or auxin concentration are changed, the obvious way to study the 
morphogenetic processes concerned is to use genetically uniform ma- 
terial but to change one or another of the environmental conditions 
under experimental control. This method has proved very successful and 
has yielded a great body of information as to the relations of environ- 
ment to plant development. This has been by far the most fruitful 
method of morphogenetic analysis since it lends itself so readily to ex- 
perimental attack. 

Traits in which environmental changes have little effect on the de- 
velopmental expression of the genetic constitution can be studied by the 
usual techniques of genetics. These consist primarily in maintaining a 
constant environment, crossing genetically pure stocks that show dif- 
ferent aspects of the trait to be studied, and analyzing the results in 
subsequent generations. There is much less opportunity here to modify 
the variables, for the genes themselves can be altered only with great 
difficulty and in an unpredictable fashion. The rise of biochemical 
genetics, however, is providing a much wider basis for experiment here. 

A question often raised in the discussion of these environmentally in- 
duced characters is whether they are adaptations and thus may serve to 
maintain the life of the plant. Many structural traits, such as the much 

Introduction to Factors 305 

reduced leaf size of microphyllous xerophytes, the nectaries in many 
flowers, or the wound cork produced on an injured surface, are present 
under almost all environments and are so deeply embedded in the 
genotype, so to speak, that the only way they may be changed is 
through genetic mutation. They have doubtless arisen by means of 
natural selection, and their adaptive character is due to this fact. 

Other traits, such as the shape and structure of the leaf blade in 
heterophyllous plants, the degree of development of vascular tissue in 
the stem, or the place of origin of roots and shoots along a regenerating 
axis, are often subject to very wide differences, depending on light, water, 
auxin concentration, mechanical stresses, and other factors. Though they 
can be greatly modified experimentally, these changes seem in most 
cases clearly advantageous under natural conditions and are thus to be 
regarded as adaptations. The adaptation here (doubtless also the result 
of natural selection) is not a specific and unchanging structure but the 
tendency to react developmentally in a favorable way as conditions 

It seems clear, however, that in many other cases, where there is a 
wide range of developmental differences induced by changes in the en- 
vironment, these are not adaptive or favorable for survival but are 
neutral in this respect. The degree of lobing in a leaf as affected by 
temperature, the relative abundance of male and female flowers as 
affected by nutritional factors, or the shape of the fruit as affected by the 
size to which it is able to grow seem none of them to have significance 
for survival. Such traits appear to be simply accidental developmental 
results of the interaction between genetic constitution and environment. 
Among these, particularly the ones induced by extreme environmental 
changes, are some of the most interesting for morphogenesis. It is there- 
fore necessary to divorce completely the problem of adaptation, which 
is an evolutionary one, from that of the environmental induction of 
characters, which is a morphogenetic one. 

In studying the various factors that are important in plant develop- 
ment, emphasis in some cases is put on changes in the environment- 
external and internal— and in others on changes in the genetic constitu- 
tion. This is purely an arbitrary classification, however, and simply for 
convenience. In the present treatment of the subject morphogenetic 
factors will be discussed in several general groups. Some, such as light, 
temperature, gravity, and some mechanical factors, originate chiefly in 
the external environment though their effects, of course, are produced 
internally. They may be grouped together as physical factors. Among 
these is included water, since its morphogenetic effects (as opposed 
to its physiological ones) are due not so much to its chemical composi- 
tion as to the physical processes of its absorption and evaporation. 

306 Morpho genetic Factors 

A second group are the chemical factors, which derive their impor- 
tance primarily from their participation in the chemical processes going 
on in the plant. Some substances, notably those in mineral nutrition, 
come into the plant from the outside, but many originate internally as 
products of the plant's metabolism. Especially important in morpho- 
genesis are the various growth substances. 

A third group of factors, the genetic ones, may also be regarded as 
part of the internal environment. Here are to be considered the genes, 
permanent and self-perpetuating; the chromosomes, which may have 
certain morphogenetic effects apart from the genes they contain; and 
the cytoplasm, the intermediary between genes and developmental 
processes. These factors, though doubtless effective because of their 
physical or chemical character, are difficult at present to reduce to such 
terms and are best considered by themselves. 

The effects of these various factors on development are complicated 
by the fact that they are operating on an organized living system which 
tends to regulate its activities in conformity to a specific norm. Three 
consequences of this should be borne in mind: 

First, a given factor does not lead directly to a given result but serves 
instead as a stimulus or evocator that sets off a reaction in the organism. 
What this will be depends on the state of the system at the time. The 
effect of light on a photographic plate is easily predictable, but its effect 
on a plant depends on the part of the plant concerned, the age of the 
plant, and its physiological condition. 

Second, the effect of one factor may be greatly modified by another. 
The photoperiodic effect of light, for example, may depend in a given 
case on the temperature of the environment, so that one factor may 
sometimes be substituted for another. Although the essence of good 
experimental work is to deal with only one variable at a time, this often 
is impossible in morphogenetic experiments (and in biology generally) 
for no one factor can be studied entirely independently of the others. 
What it will do depends on the rest of the environment and on the state 
of the organism. 

Third, the organized system is not a constant one but tends to change 
in character from one phase in its life cycle to the next and from one 
region of the body to another. The potency of a cell (the repertoire of 
developmental possibilities open to it), high at first, is reduced as the 
cell grows older. Doors continually close behind it, so to speak. The 
reactivity of a cell (the response it will make to a given environmental 
change) also is different at successive developmental stages. Both 
potency and reactivity may be unlike in different parts of the organism. 

An investigation of the effects of various factors on plant develop- 
ment, particularly environmental ones, though not as simple as might 

Introduction to Factors 307 

at first appear, may still be very fruitful. It is often possible to examine 
the effects of changes in one factor without serious complications from 
others, and the organism does remain essentially constant over short 
periods. The very considerable knowledge now available as to the 
factors that modify plant development has proved most important for 
an understanding of the problems of morphogenesis, and there are wide 
possibilities for extending it much further. The next seven chapters are 
devoted to a brief consideration of the operation of these factors. 



Light is a powerful factor in determining the course of development in 
plants and has a much more important morphogenetic effect on them 
than it does on animals. This is to be expected, since light is necessary 
for photosynthesis and thus for the production of food. Experiment has 
made it clear, however, that the morphogenetic influence of light is 
much more subtle and indirect than this and results from its control not 
only of food production but of various physiological activities in the 
plant by which this food is distributed in the processes of growth and 
differentiation. The role of light in plant development has been studied 
actively for many years and is the basis of an extensive literature. Among 
the more inclusive reviews of this field are those by MacDougal ( 1903<7 ) , 
Burkholder (1936), and Parker and Borthwick (1950). 

Many of the early results are invalid because of the impossibility in 
those days of exact control of light, as to its intensity and quality, in 
experimental work, but most of these difficulties have now been over- 
come, and light in a plant's environment can be manipulated with 
relative ease. 

It is a matter of common observation that plants reach their greatest 
size and vigor in good light and that insufficient illumination results 
in weak and spindly growth even if water, soil nutrients, and tempera- 
ture are at their optimum levels. Most of the experimental work with 
light has involved not merely differences between light and darkness 
but measured differences in the light stimulus itself. Three of these are of 
chief importance: the intensity, the quality, and the duration of the 
light. Intensity is the brightness of the illumination, the actual energy 
of the radiation. Quality concerns the wave length of the light. Dura- 
tion refers to the relative lengths of the alternating periods of light and 
darkness to which the plant is exposed. These differences are not always 
sharply separable, and one often modifies the effects of another. 


Light 309 


Since rate of photosynthesis increases with light intensity, up to a cer- 
tain point, the growth and vigor of a plant are generally proportional, 
within limits, to the brightness of the light to which it is exposed. 
Shirley (1929) grew a variety of plants in different intensities of day- 
light and found that at low ones dry weight was directly proportional 
to intensitv but that at higher ones growth was relatively less. He ob- 
served that intensity also affected certain qualitative traits, such as ratio 
of root to shoot, strength of stem, thickness of leaves, and development 
of vascular tissue. 

It has frequently been observed that plants grown in shade have rela- 
tively small root systems. In general it may be said that stem elongation 
varies inversely with light intensity but that width varies with it directly 
( Popp, 1926 ) . The effect of light may be different on different parts of 
the plant and at different stages of development. Some morphological 
effects of light may be quite specific though the mechanisms involved 
are unknown. Some herbaceous stems, for example, have zig-zag form 
in the light but are straight if grown in the dark. Plants that twine in the 
light usually lose this ability in darkness. 

Light is also important morphogenetically for some plants lacking 
chlorophyll. In certain mushrooms, for example, the fruiting body will 
not develop normally in complete darkness although the whole vegetative 
mycelium is subterranean (Borriss, 1934b). 

Etiolation. The most conspicuous effect of differences in light intensity 
on plant structure is to be seen in the phenomenon of etiolation. It is a 
familiar fact that green plants growing in darkness or relatively weak 
light tend to be tall and spindly, with small, pale leaves, weak roots, and 
poorly developed vascular tissues. Such plants soon die unless consider- 
able reserve food is available in seeds or other storage organs, in which 
case etiolated growth may continue for some time. The early work of 
Kraus ( 1869 ) showed that etiolation involves a considerable increase 
in cell length, though in most cases this is accompanied by some increase 
in cell number in the longitudinal dimension. 

Different parts of the plant and different species differ considerably 
in their manifestations of etiolation. Only shoots etiolate and not roots, 
flowers, or fruits. Avery, Burkholder, and Creighton ( 1937a ) observed 
marked differences between the first internode and the coleoptile as to 
their elongation in various light intensities. Intensity may also affect the 
proportions of parts. In Tropaeolum plants, for example, which are grow- 
ing in weak light the ratio of petiole length to lamina width becomes 
progressively greater as the leaves develop, whereas under normal 

310 Morpho genetic Factors 

illumination the two dimensions grow at about the same rate (Pearsall, 

As to the causes of etiolation there has been much discussion but no 
final agreement. Auxin is undoubtedly involved, for it is well known 
that sensitivity to it increases in darkness. Wave length of light is also 
important here, for etiolation may be very different in red light and in 
blue. The two processes of leaf growth and stem elongation may be 
affected differently. 

Priestley (1926b) called attention to the fact that in etiolated stems 
the endodermis tends to be well developed and to have thick-walled 
cells. An etiolated stem is thus somewhat like a root in structure. He 
suggests that for this reason water and nutrients, coming from the roots 

Fig. 13-1. Effect of etiolation on cell shape. Longitudinal section of cortical paren- 
chyma of the stem of Vicia faba when grown in light ( left ) and in darkness ( right ) . 
( After Kolda. ) 

into the vascular cylinder, may be confined there and prevented from 
passing outward. This would tend to accelerate growth at the tip of the 
shoot and to check the development of leaves. 

The relation of light to normal and etiolated growth evidently involves 
the problems of cell division and cell enlargement. In beans Brotherton 
and Bartlett (1918) found that in the epidermis about a third of the 
added growth in length of etiolated as compared with normal plants 
was due to more cells and about two-thirds to longer cells, the rate of 
both processes of division and enlargement being inversely proportional 
to light intensity. Cell elongation has been shown in many other cases 
to increase with diminished light intensity (Fig. 13-1). This is evident 
not only in green plants but in fungi (Castle and Honeyman, 1935). 

Light 311 

Whether light produces its effect on cell size by changing osmotic 
concentration, permeability of cell membranes, attraction of protoplasm 
for water, character of the cell wall, or other processes is not clear. It is 
significant that not merely is cell size increased in low light intensity but 
size along the polar axis of the cell. 

Meier (1934) studied the effects of the intensity of light on cell di- 
vision in the unicellular alga Stichococcus bacillaris and found that its 
multiplication in culture is proportional to the intensity of illumination 
up to a certain point but that high intensities check it. 

Thomson (1954) grew seedlings of oats and of peas with different 
amounts of light and reports that light accelerates whatever growth 
processes are going on while it acts, its effect depending on the stage 
of development of the tissues concerned. Exposure early in the course of 
either the cell-division phase or that of cell elongation hastens the 

Fig. 13-2. Light and vascular development. Portion of the vascular cylinder of the 
stem of Vicia faba grown in darkness (left) and in the light (right). (From Borg- 
strom. ) 

completion of that phase of growth, but after it is under way light 
hastens the transition to the next phase and thus reduces the final num- 
ber or length of cells. 

Shape traits may be modified by light. Smirnov and Zhelochovtsev 
(1931) found that in Tropaeolum leaves the reduction of blade expan- 
sion in weak light modified the fundamental growth pattern. Njoku 
(1956a) reports that differences in light intensity change the leaf shape 
in Ipomoea. 

Anatomical characters are also affected. Penfound ( 1931 ) observed 
that stems of Helianthus and Polygonum growing in full sunlight have a 
much greater amount of xylem and more and thicker-walled cells in the 
mechanical tissue than those in shade. The much reduced vascular tissue 
of etiolated plants is well known (Fig. 13-2). Bond and others have 
found that as light intensity is reduced the development of the endo- 
dermis increases in the stem, where it normally is weak or absent. That 

312 Morphogenetic Factors 

this tissue is much better developed in roots than in stems may be re- 
lated to the fact that roots normally grow in darkness and stems in light. 
Leaf structure is often different in different light intensities. It has 
long been known (for example, Nordhausen, 1903) that in many cases 
leaves on the south side of the tree ("sun leaves") are thicker and better 
differentiated than those on the north side or the interior of the crown 
("shade leaves"; Fig. 13-3). This has been discussed by Lundegardh 
(1931) and others. A particularly striking case is described by Cormack 
( 1955). The question has been raised (p. 327) as to whether this effect is 
actually due to light intensity or to differences in water relations, par- 

Fig. 13-3. Transverse sections through the blades of sun leaves (above) and shade 
leaves ( below ) of a, Acer; b, Quercus; c, Fagus; and d, Tilia. ( After Schramm. ) 

ticularly rate of transpiration, for sun leaves tend to be xeromorphic in 
character. Talbert and Holch (1957) studied the leaves of 37 species 
and found that sun leaves usually had smaller laminar area, shorter blade 
perimeter, deeper lobes, more pronounced veining and marginal serra- 
tions, more hairy surfaces, and shorter petioles than shade leaves. 

Anderson ( 1955 ) studied the development of sun and shade leaves in 
Cornus and Viburnum and finds that in the latter the large leaf size is 
the result of earlier and more rapid growth. The greater thickness of sun 
leaves is due to greater cell elongation. Differentiation takes place earlier 
in shade leaves. 

Light 313 


Much work has been done on the effects of different wave lengths 
(colors) of light. Many of the early results here are of doubtful value 
since it was often difficult to change the wave length without at the same 
time modifying intensity. There are many well-established facts, how- 
ever, from which conclusions can be drawn. 

It seems clear, for example, that the longer wave lengths, notably 
those in the red, promote a marked elongation of cells and thus of tissues, 
whereas the blue rays (and white light) check this effect and tend to 
prevent elongation. Teodoresco (1929) describes many examples of this, 
especially from less highly organized forms such as young plants of liver- 
worts or fern prothallia growing from spores. In these cases where white 
or blue light is used, a rather compact group of cells develops from the 
spore, but in red light a much elongated, spindly cell. The same effects 
are evident in fungi. An important fact is that the plane of cell division 
is usually controlled by the light. Mohr ( 1956 ) finds that in young fern 
prothallia in red light cell division tends to be at right angles to the 
polar axis of the structure so that filaments of elongate cells are formed, 
whereas in blue light division is in various planes, so that a plate of cells 
develops. In a normal and growing fern prothallium transferred to red 
light, many of the cells grow out to form filaments. 

The same effects are to be seen in the more complex higher plants. Thus 
Teodoresco finds that blue light checks petiole elongation but increases 
blade area, and Vince ( 1956 ) that in many plants, when grown under 
lights of equal energy levels, total stem length, internode length, and 
leaf length increase with increase in the wave length. Not all plants 
react alike, however. Whether the mechanism by which the red rays 
promote elongation is like that by which low light intensity does so is 
not clear, but presumably the same processes are affected by both factors. 
Wassink and Stolwijk ( 1952 ) used equipment by which it was possible 
to grow plants in various wave lengths of monochromatic light, and 
under these conditions there was strong elongation of the stem and curl- 
ing of leaves and petioles in green, yellow, and red light but essentially 
normal growth in blue. Fortanier (1954), however, observed that only at 
high light intensities is stem elongation greatest in red, yellow, and 
green. At low ones it is greatest in blue. Leaf number was not affected by 
wave length. 

Quality of light also affects flowering. Curry and Wassink ( 1956), work- 
ing with annual Hyoscyamus niger, found that flowers were produced in 
blue and infrared-plus-red radiation but that neither stem elongation nor 
flower-bud formation occurred in green or red light. 

314 Morphogenetic Factors 

The relation of wave length to auxin production and other problems 
of photomorphogenesis have been discussed by Stolwijk (1954). 

Other developmental traits are affected by light quality. Thus Funke 
(1931) observed that in heterophyllous water plants, where the juvenile 
immersed leaves are ribbon-like, these never develop into anything else if 
the plants are grown in red or in green light. In blue or white, however, 
normal mature foliage is produced. This change may be reversed by 
changing the wave length of the light. In root cultures of peas, red light 
inhibits the formation of lateral roots more effectively than blue or 
green, perhaps by inactivating substances necessary for root growth 
(Torrey, 1952). Many other instances of the effects of light of different 
wave length on development in higher plants have been reported. 

Less work has been done with lower plants. Meier (1936), again with 
the alga Stichococcus, found that in cell culture the individual cells in a 
given time multiplied fourfold in white light, threefold in blue, but 
only twofold in yellow and red. Green light proved to be destructive to 


In the slime mold Didynium nigripes, light is necessary for the develop- 
ment of sporangia (Straub, 1954). Green light has no effect but red 
and blue have. If plasmodia treated with these wave lengths are killed 
by freezing and fed to living plasmodia, the latter produce sporangia 
after a briefer exposure to light and much more rapidly than control 
plasmodia which had been fed untreated ones. Evidently a substance con- 
ducive to sporangium production is formed by the action of light of cer- 
tain wave lengths. Gray (1953), using the slime mold Physarum poly- 
cephalum, found that continuous irradiation with monochromatic light 
in the blue and green and a narrow band in the yellow induced fruiting 
bodies, the rate of their formation being inverse to the wave length of the 
light used. He suggests that a changed acidity resulting from the irradia- 
tion may be responsible for the production of sporangia. 

There are general discussions of the morphogenetic effects of different 
wave lengths of light by Parker and Borthwick (1950) and Wassink and 
Stolwijk (1956). 


One might expect that the longer a plant is exposed to light favorable 
for photosynthesis, the more it would grow and the more vigorous it 
would be. Keeping plants in continuous light, however, is often found to 
result in less vigor and in a disturbance of the normal reproductive 
cycle. It is evident that the production of flowers and fruits is not some- 
thing that inevitably occurs but rather that it will happen only when 
environmental factors are favorable for it. In 1920 Garner and Allard pub- 

Light 315 

lished the results of their pioneer observations which showed that flower- 
ing is not determined by the intensity or the quality of light alone but by 
the length of daily exposure to light, or the photoperiod. What essentially 
is involved is the relative length of the alternating light and dark periods, 
or, perhaps more accurately, the length of the dark period for some 
plants and of the light period for others. This phenomenon of photo- 
periodism has now been subjected to intensive study. A history of the 
work on it until 1948 was written by Murneek (1948). Other surveys 
of the subject or particular aspects of it are those by Garner (1937), 
Burkholder (1936), Hamner (1944), Leopold (1951), and Naylor 


All plants do not respond alike to photoperiodic stimulation. In the so- 
called short-day forms, flowering is induced by relatively short periods 
of daily illumination ( and thus longer dark periods ) . Such plants flower 
naturally in fall or early spring. Long-day plants require a longer period 
of light and in nature are summer-flowering forms (Fig. 13-4). Many 
plants, such as the tomato, are day-neutral and will flower under long 
or short photoperiods or continuous illumination. 

This classification is not a very exact one, for different steps in the 
reproductive process may each have their optimal photoperiod. Thus most 
strawberries flower under a relatively short day but fruit under a long 
one, and Phlox is a long-day plant for flowering but is day-neutral for 
fruiting. The optimal period for the formation of flower primordia at a 
growing point may be different from the one determining the later 
growth and opening of the flowers. These relationships have been dis- 
cussed by Roberts ( 1954 ) . 

There is a close relation between temperature and photoperiodism 
which has been studied by a number of workers, among them Purvis 
(1953) and Vlitos and Meudt (1955). High temperature will sometimes 
induce flowering even when day-length is not favorable for it. In vernali- 
zation (p. 339) it is necessary not only to expose the germinating seeds to 
low temperature but to provide the proper photoperiod for later growth. 
Schwabe (1951) concludes, from experiments with vernalized short-day 
and long-day Chrysanthemum cuttings, that the effects of vernalization 
and of day-length in this plant are operative at different stages in the 
train of reactions leading to flower initiation. Sometimes a high level of 
nutrition, especially of nitrogen, may be substituted for day-length. Short- 
day soy plants have a higher concentration of nitrogen than do long-day 
ones. In many cases an exposure for a few days to a photoperiod favorable 
for flower production will result in flowering regardless of the one to 
which the plant is later exposed. Plants vary in their sensitivity to this 
photoperiodic induction. 

Plant parts affected by photoperiodic stimulation in most cases are the 


Morpho genetic Factors 

Fig. 13-4. Above, long-day plant, Nicotiana sylvestris. Below, short-day plant, Nico- 
tiana tahacum, var. Maryland Mammoth. At left, under long days; at right, under short 
days. ( From Melchers and Lang. ) 

young but fully developed leaves. Evidence indicates that a flower-form- 
ing substance is produced in these which then diffuses through living 
cells to the meristem and there stimulates the formation and growth of 
floral primordia (p. 397). This substance can be transmitted by grafting 
from a plant in flower to one which is not, and the latter plant, even 

Light 317 

though kept under a photoperiod unfavorable for flowering, will then 
flower. The age of the plant may change its photoperiodic response. In 
Kalonchoe, for example, young plants 3 months old flower only in short 
days, but after 5 months they have become day-neutral (Harder and 
von Witsch, 1940« ) . 

Even though the initiation of floral primordia has begun under a given 
photoperiod, the later differentiation of the various structures can be 
greatly altered by changing the length of the period. Thus in the stami- 
nate inflorescence of maize, after the initiation of primordia, later growth 
under longer photoperiods will cause the flowers to be infertile and 
even to show progressive changes toward a vegetative condition. The 
glumes develop ligules and the lemmas differentiate into blade, ligule, 
and sheath until the spikelet becomes much like a vegetative shoot and 
can be propagated as such (Galinat and Naylor, 1951). When trans- 
ferred to a photoperiod unfavorable for flowering, buds which would 
have produced flowers will sometimes grow into abnormal vegetative 
shoots (phyllody), as reported by Behrens (1949) and others. Skok and 
Scully (1955) present evidence that floral development is associated 
with a dark-dependent mechanism and the elongation of the main axis 
with quite a different and light-dependent one. 

The length of the photoperiod may affect the differentiation of the 
sexes. This is well shown by the work of Schaffner ( 1931) on sex reversal 
in staminate plants of hemp, Mercurialis annua (Fig. 13-5). He planted 
seeds in the greenhouse every 2 weeks from July 15 until May 15 and 
found that in the beginning, when days were long, the flowers were all 
staminate but that the percentage of pistillate ones steadily increased 
up to the plantings of Nov. 1 and 15 (which came to flower during the 
shortest days) and that the percentage of these flowers gradually de- 
creased after this until in the long days of spring the plants were all 
staminate again. Long days obviously favor the production of staminate, 
and short days, of pistillate, flowers. Similar results were obtained by 
others, as by Jones ( 1947 ) in Ambrosia. 

In Cannabis sativa under a 16-hour day flowering takes place in from 
4 to 6 weeks, the leaves become more complex (with up to nine leaflets), 
and the plants are about half males and half females. Under 8-hour days, 
however, development is more rapid, flowering is reached in 3 or 4 
weeks, and the plants are about half hermaphrodites and half females 
(Petit, 1952). 

Day-length also affects reproduction in the lower plants. In the alga 
Vaucheria sessilis (League and Greulach, 1955) the production of sex 
organs was earlier and more abundant under 18-hour days than under 
8-hour ones. Addition of glucose and peptose to the culture medium 
hastened their formation under short days. Sex organs were not produced 

318 Morpho genetic Factors 

unless there was a high concentration of fat globules near the point of 
origin. These authors believe that this is not a case of true photoperiodism 
but that the low production under short days results from a limited 
availability of food. 

Klein (1948) reports that in the fungus Pilobolus there is a definite 
cycle of asexual reproduction caused by periods of light and darkness, 



< 60 




id 60 

u 50 


< 40 






| hji 


1 _L 

J i X 

7 JL1 

7 __i__\ 

r i V 

7 J- v 

~~t _I jL 

"7 __i V 

r T r 

_^' . 1 1 .- -^- 






Fig. 13-5. Graph showing percentage of reversal of staminate flowers to pistillate ones 
in genetically staminate plants of Mercurialis annua planted at different dates. Day- 
length markedly affects sex expression. ( From Schaffner. ) 

sporangiophores maturing at the end of a dark period and new ones re- 
maining immature at the end of the light one. Light is essential to growth, 
but a dark period seems necessary to establish a periodicity of growth 
and maturation of the fruiting bodies. Periodicities other than those 
found in nature could be established in this plant by artificial illumina- 
tion. Among these were light-dark cycles (in hours) not only of 12-12 

Light 319 

but 16-16, 15-9, and 9-15. The 16-16 period was quite consistent and evi- 
dently had been acquired by the plant. 

The physiology of photoperiodism presents many problems which are 
too compiex to be discussed here. One hypothesis proposes that in the 
light a substance is produced which persists in the subsequent dark 
period. In the latter another is formed which is destroyed by a very 
brief exposure to light. These substances interact to make a flower- 
forming substance. In the short-day species Kalanchoe Blossfeldiana, 
which has been studied intensively ( Harder, 1948 ) , a single leaf borne at 
the tip of the plant ( those above it having been removed ) if exposed to 
a long day will almost completely prevent flowering in the plant below. 
Harder, Westphal, and Behrens (1949) conclude that in it is formed a 
substance which inactivates the flower-forming hormone before this 
reaches the floral primordium. Auxin presumably is involved in some of 
these processes. This has been discussed by various workers, among them 
Konishi (1956). 

In some cases the photoperiodic reaction can be changed. Working 
with an early-blooming variety of peas which is day-neutral, Haupt 
( 1957 ) reduced flowering by removing the cotyledons and modifying 
the soil nutrients. The plants now reacted as though they were long- 
day and late-blooming types. He found that a true late-blooming 
variety which is normally a long-day plant lost its photoperiodic reac- 
tion and bloomed early if a scion from a blooming plant was grafted 
into it. 

Earliness of blooming may be due to other factors than day-length. 
Thus in an early-blooming and a late-blooming variety of Chrysanthe- 
mum, both grown under short days, Doorenbos and Kofranek ( 1953 ) 
found that the initiation of the florets took place at the same rate, but the 
time from the end of this stage until the date of blooming was 28 days 
in the early variety and 42 days in the late one. 

Photoperiodism has been studied chiefly in relation to the differentia- 
tion of reproductive structures, but it has a pronounced effect on others 
also. Among instances of this are the following: 

Pfeiffer ( 1926 ) grew buckwheat with daily illuminations of 5, 7, 12, 
17, 19, and 24 hours and found that maximum stem length and diameter 
were produced in the 17-hour day. 

The effect of the photoperiod is different on different parts of the plant 
and under different conditions. Hall (1949) grew gherkins under green- 
house conditions from seedling to maturity. At high nitrogen levels, 
plants given a 16-hour day had larger stems than those under 8 hours 
but at low nitrogen levels this was reversed and plants under the shorter 
days grew larger. Under the 8-hour photoperiod more nodes and leaves 

320 M or pho genetic Factors 

were produced but the total leaf area was smaller and there were fewer 


Deats ( 1925 ) , studying tomato and pepper, found that the amounts of 
both phloem and xylem varied directly with the length of day. 

Garner and Allard ( 1923 ) made the significant observation that as the 
photoperiod becomes less favorable for vegetative growth the structure 
of the plant becomes somewhat xeromorphic— the stem tends to become 
more branched, underground parts to enlarge, pubescence to increase, 
abscission layers to cause leaf fall, and flowers to appear. 

In a study of heterosis in beans, Malinowski (1934) crossed two races 
and grew them and their hybrids under long and short days. In long days 
the Fx plants were larger in every way than the parents. In short days 
they were about the size of the parents and flowered 6 weeks earlier than 
under long days. This acceleration of flowering seems to have cut down 
their vegetative growth and reduced heterotic vigor. 

Mac Vicar and Struckmeyer (1946) grew soybeans with a deficiency 
of boron in different photoperiods. The deficiency symptoms were much 
more severe under long-day treatment than under short. The boron 
content of all the plants was much the same, and these authors believe 
that the effect of day-length was to alter the boron requirement of the 

The relative size of the leafy shoot to roots or tubers is markedly in- 
fluenced by day -length. Radish, for example, grown under short days, 
as in a greenhouse in the winter, forms a very large root and a small 
shoot, but in the longer days of spring the root is relatively much 
smaller. Other plants with storage roots behave in the same way, as do 
potatoes in the ratio of tops to tubers ( Pohjakallio, 1953). Zimmerman 
and Hitchcock (1929) observed that in dahlias short days produce heavy, 
fleshy root systems but long days, fibrous ones. These workers also found 
( 1936 ) that growing Jerusalem artichokes under short days stimulated 
tuber production but that the same result was obtained by subjecting 
only the tip of the stem to short days by capping it part of the time with 
black cloth. This indicates that the control of tuberization by day-length 
is centered in the growing tip and its young leaves. In general, the 
growth of underground storage regions is stimulated by day-lengths 
different from those favorable for the vegetative growth of the shoot. 
Where the photoperiod is such that the shoot in its growth is unable to 
use carbohydrate beyond a certain amount, this accumulates in storage 
regions. Jenkins (1954) reports that long days and relatively high tem- 
peratures are necessary for bulb formation in shallots. In Poa, long days 
favor the growth of bulbils and short days of ears ( Schwarzenbach, 

In potatoes (Chapman, 1958), short photoperiods induce tuber for- 

Light 321 

mation. A tuber-forming stimulus is produced near the growing points 
of the plant and moves basipetally. It is able to cross a graft union and 
produce tubers in a noninduced plant. In plants with forked stems, 
one half was given short periods and the other long. Tubers were 
produced on that part of the stolon below the short-day branch. 

Reduction of growth of the axis, with the formation of leafy rosettes, 
is favored by relatively short days. Thus Oenothera forms rosettes in the 
fall and tall flowering stems in the next spring. Lettuce makes compact 
heads in the short days of early spring but shoots upward in the longer 
ones of June. 

In strawberries, long days increase leaf size and cell number (Arney, 
1956). Ashby (1950b) reports that day-length also affects leaf shape in 
Ipomoea, In 16-hour days, plants begin to flower at the fifteenth or 
sixteenth node and lobing begins at the fifth to seventh node. Under 8-hour 
days, however, flowering begins at the first node and lobing is almost 
entirely suppressed. 

Gotz (1953) has made an intensive study of the effect of day-length 
on the formation of plantlets on the leaves of three species of Bryo- 
phyllum. In short days neither plantlets nor their primordia are formed 
nor are flowers produced, and the leaves become somewhat more suc- 
culent. Under long days, however, plantlets appear in abundance. The 
effect of different day-lengths on plantlet production can be studied in 
different leaves on the same plant. Neither grafting a scion from a long- 
day plantlet-producing plant into a short-day plant, or injecting sap 
from one, will induce the formation of these structures. 

The photoperiod also affects rooting of cuttings. Some species root 
best under long days and others under short (Stoutemyer and Close, 
1946). The photoperiod under which the stock plant has been growing 
actually affects the rooting of cuttings taken from it more than does that 
under which the cuttings themselves are grown (Pridham, 1942). 

As to leaf structure, Glimmer ( 1949 ) found that in Kalanchoe a change 
in the photoperiod affects the thickness of the epidermal cell walls, the 
size of the vein islets, and the form, size, and number of mesophyll cells 
but that the number of stomata responds more slowly and the size and 
form of the epidermal cells are unaffected. In the ten species she studied, 
plants grown under short days had thicker leaves than under long ones, 
and this was almost entirely because of greater size of the mesophyll 
cells, which elongate at right angles to the surface of the lamina. They 
also increase somewhat in width. If a single leaf on a Kalanchoe plant 
growing under long days is itself exposed to short days, it grows con- 
siderably thicker and changes its form somewhat. It is significant that 
these changes are transmitted to other leaves directly above this one, 
suggesting that a morphogenetic substance is involved. Detached and 

322 Morphogenetic Factors 

rooted leaves respond to day-length by the same changes in form and 
structure as do those that remain attached to the plant ( Schwabe, 1958 ) . 

In many woody plants studied, both flowering and vegetative growth 
are markedly affected by the photoperiod (Wareing, 1956; Downs and 
Borthwick, 1956). In general, short days induce dormancy and long ones 
prolong growth. Marked ecotypes as to photoperiod have been found in 
a number of species. 

Cellular characters are also affected. Von Witsch and Fliigel (1952) 
found that in leaves of Kalanchoe Blossfeldiana ( 2n = 34 ) formed in long 
days the mesophyll cells have chromosome numbers between 128 and 
135. Under short days these cells are much larger and the degree of 
polyploidy is increased, the chromosome number going up to about 540. 
In tetraploid plants of Hyoscyamus niger produced by colchicine, the 
critical day-length for flowering and the time preceding the elongation 
of the internodes were both shortened, the number of leaves was re- 
duced, and the time of flowering delayed, as compared with diploid 
plants (Lang, 1947). 

The problems of photoperiodism are complicated by the fact, empha- 
sized by Biinning ( 1956 ) and others, that there are endogenous rhythms 
in certain of the physiological processes of the plant. It has been found, 
for example, that a light period of 12 hours alternating with 12 hours of 
darkness gives in many plants a different result from an alternation of 6 
hours of light and 6 hours of darkness and thus two cycles in a day. The 
total amount of light and darkness are the same but their effects are not. 
There is evidently a changing sensitivity in the reaction of the plant 
during the day to various environmental factors. This fact is of much 
importance for plant physiology but its significance for morphogenesis 
has as yet not been very fully considered. The existence of innate rhythms 
may account for the conflicting results obtained in experiments on the 
morphogenetic effects of various environmental factors. 


The various morphogenetic effects of light provide an excellent example 
of the complexity of interaction of factors in plant development. Light 
powerfully influences flowering, but so do temperature, growth sub- 
stances, nutrition, and genes, and in some cases water supply and gravity. 
They often have parallel effects on vegetative structures as well. These 
factors frequently can be interchanged to some extent and produce the 
same result, as when temperature is substituted for photoperiod, and 
vice versa. Auxin is closely concerned with many of the traits that light 
affects, but the exact relation between light and auxin is not clear. Some- 
times light seems to destroy it and sometimes to stimulate its produc- 

Light 323 

tion. Specific photoperiodic reactions have been found to be gene-con- 
trolled. The.morphogenetic effects of bright light and of limited water 
supply are sometimes hard to disentangle. It is difficult, as has been 
said before, to separate any one factor sharply from the others and to 
study it in isolation. All are concerned with the entire organized system 
that is the plant. 



Water is closely involved with many activities of the plant, especially 
photosynthesis and transpiration. It fills the cell vacuoles and constitutes 
the bulk of protoplasm. It maintains the turgidity of the tissues and thus 
is an important factor in growth. Botanists still are very far from explain- 
ing the complex problems of the water relations of plants. These have 
been discussed in an extensive physiological literature (see Crafts, Cur- 
rier, and Stocking, 1949; Kramer, 1945, 1955; Meyer, 1938; and Walter, 

Xeromorphy. Water is also of significance in problems of plant struc- 
ture and thus for morphogenesis. Where it is relatively scarce or the 
amount that can be absorbed is limited for other reasons (as in saline 
soils) or where evaporation is excessive, plants display characteristic 
structural features. Such xerophytes tend to have reduced leaf surfaces, 
heavy cuticle, small and thick-walled cells, high stomatal frequency, 
abundant mechanical tissue, and large root systems, and they often are 
spiny or succulent. These traits, collectively termed xeromorphy, have 
been regarded as adaptations which increase absorption or reduce tran- 
spiration and thus maintain a sufficient water supply under dry condi- 
tions. Xerophytes may show other adaptations such as hairy surfaces, 
rolled leaf blades, and stomata sunken in pits or otherwise protected 
against undue exposure to evaporation. The characteristic structures of 
xerophytic plants have long attracted the interest of ecologists and pro- 
vide much of the subject matter for the science of ecological anatomy. 

Such traits presumably have arisen through the action of natural se- 
lection and are thus not ultimately attributable to the direct effect of the 
environment. Many plants, however, if grown under conditions where 
water is scarce or transpiration high, have been observed to assume 
some degree of xeromorphism. Leaf surfaces will tend to be somewhat 
reduced, cells smaller and thicker-walled, and mechanical tissue more 
abundant. Such structural changes are clearly the result of an environ- 
mental factor— a reduction in amount of available water. What is in- 
herited here is this specific response to the environment. 

There has been considerable controversy, however, as to whether or 


Water 325 

not such changes are adaptive and are advantageous to the plant. That 
they should be so is a plausible conclusion, and for such traits as heavy 
cuticle it may be correct. Its general validity has been challenged by 
a number of people, notably Maximov, who has reviewed the problem 
comprehensively ( 1929, 1931 ) , especially as to the physiological basis of 
drought resistance. Maximov called attention to the earlier work of 
Zalenski (1904), published (chiefly in Russian) a quarter of a century 
before and largely neglected outside the country of its origin. Zalenski 
observed that the veining of the leaves in plants growing in dry, open 
spaces was much more abundant than in the leaves of those in the shade 
or in protected spots. These observations he then extended to a com- 
parative study of the structure of leaves on the same tree. Here he 
found that, as a rule, leaf structure changed with the level of insertion 
on the tree, the structure being more xeromorphic with increasing dis- 
tance from the root. The progressively higher leaves had smaller cells 
throughout, smaller stomata and more of them per unit of area, greater 
vein length per unit of area, thicker and less sinuous walls in the epi- 
dermal cells, a greater contrast between palisade and spongy layers, less 
intercellular space, and better developed mechanical tissue. These re- 
lationships were later called "Zalenski's law" and were independently 
discovered by others, among them Yapp ( 1912 ) . Zalenski's results are 
evident in herbaceous as well as in woody plants. Some data that he 
presents for Dactylis glomerata are shown in Table 14-1. Salisbury ( 1927) 

Table 14-1. Variations in Anatomical Elements of Leaves of Different 
Tiers in Dactylis glomerata * 

Tier 1 2 3 4 5 

Height of insertion ( cm. ) 10.2 25.2 37.0 51.0 

Length of leaf ( cm. ) 7.1 10.3 18.5 18.0 13.2 

Breadth of leaf (cm. ) 0.30 0,35 0.54 0.52 0.45 

Length of vascular bundles 

(mm./sq. cm. of leaf surface) 371 511 557 625 626 
Mean diameter of cells of 

upper epidermis ( mm. ).. . 0.0418 0.0294 0.0272 0.0217 0.0189 

Number of stomata in field .. . 34 42 61 80 64 

Length of stomata ( mm. ).. . 0.0434 0.0415 0.0403 0.0356 0.0384 


From Maximov (1929), after Zalenski. 

found that stomatal frequency per unit area increases with the height at 
which the leaf is borne but that the stomatal index (ratio of stomata to 
epidermal cells in the same region) changes relatively little. This is a 
necessary implication of Zalenski's observations. 

These structural characters are among those regarded as typically 
xeromorphic. Zalenski, Yapp, and others, however, have explained them 

326 Morpho genetic Factors 

without relating them to water conservation but simply as direct or 
indirect results of decreasing cell size with progressively higher leaf 
insertion. Small cell size, in turn, grows out of the greater difficulty 
with which water is obtained by the higher leaves, since they have to 
lift it farther and against the competition of the lower ones. It follows 
that at the critical period of rapid leaf growth, which results primarily 
from cell expansion through the absorption of water, the cells of the 
upper leaves cannot attain the size of the lower ones. The other struc- 
tural traits are a consequence of this basic difference. That such a con- 
clusion is correct is indicated by other evidence, such as the fact that if 
lower leaves are removed while the upper ones are still growing the 
latter will resemble in structure leaves that would normally be lower 
on the stem. 

Xeromorphy in these upper leaves therefore seems unlikely to be an 
adaptation for reducing water loss. Indeed, it was later shown that 
upper leaves may transpire more rapidly than the lower ones. These 
results have cast doubt on the adaptive character of the traits of 
xerophytes in general. Maximov calls attention to the fact that when 
water is abundant many xerophytes transpire more rapidly than meso- 
phytes and it is only under drought conditions that their water loss is 
markedly cut. He attributes this, and therefore the quality of drought 
resistance in general, not to any structural traits but to protoplasmic 
characters, notably osmotic concentration and changes in cell colloids 
that would enable the plant to conserve its water supply and thus 
endure dry conditions better than other plants. Eckardt (1953) agrees 
with this conclusion. The physiology of drought resistance has been re- 
viewed by Iljin (1957). 

In earlier years a number of Russian investigators, assuming that types 
with small cells were more resistant to drought than those with large 
cells, determined for various cereal varieties their "anatomical coefficients" 
(chiefly cell size), hoping to find a means of identifying resistant types 
by direct inspection. This possibility was not supported by the work of 
Maximov. More recently, however, Lai and Mehrotra (1949), working 
with sugar cane, have found that some cell characters, notably small 
size, seem in certain cases to be associated with drought resistance. 

Farkas and Rajhathy ( 1955 ) reexamined anatomical gradients in some 
herbaceous plants, particularly tomato, and found that cell size in 
leaves decreases from below upward and that number of stomata per 
unit area increases, thus again confirming Zalenski (Fig. 14-1). Under 
dry conditions this gradient is much steeper. They found several other 
physiological gradients some of which may be explained as dependent 
on that for water. In others the relation to the latter is not clear. Stage 
of development of the leaves also complicates the problem here. 



Shields (1950), who has reviewed the whole subject of xeromorphy, 
agrees in general with Zalenski, Maximov, and Yapp that this type of 
structure has little significance as an adaptation in drought resistance. 
She also emphasizes the importance of physiological factors in relation to 
water loss. Many of the structural characteristics of xerophytes, she sug- 
gests, may be the result of physiological differences. Thick cell walls and 
abundance of mechanical tissue may result from active photosynthesis 
in a plant where all its products cannot be used in growth because of 
the shortage of water. 

Ashby ( 1948Z? ) , however, presents evidence that the relative xeromorphy 
of the upper leaves, at least as indicated by cell size, is not due to corn- 



~ 13 

<-> 12 

^ 10 

% 9 





1 70 
- I 60 
05 50 




^^ t 







A \ 


2 3 

2 3 

2 3 

2 3 

1. Blatter 

2 Blatter 

3. Blatter 

*. Blatter 

Fig. 14-1. Gradient of cell size in tomato leaves. Graph showing stomatal number (in 
a given microscopic field ) , epidermal cell size ( in microns ) , and leaf surface ( in square 
centimeters) for the first four leaves of five young tomato plants. These are arranged 
in each case according to the ascending order of stomatal number for each leaf class. 
( From Farkas and Rajhdthy. ) 

petition for water but to the influence of immature leaves on those above 
them, an influence which may be hormonal in character. 

Leaves on the outside, and particularly the south side, of the crown 
of a tree (sun leaves, p. 312) are often considerably thicker and more 
xeromorphic in appearance than those in the center (shade leaves). 
This has been attributed to the direct action of light. This difference may 
be due in part to water relations, for Hanson ( 1917 ) showed that, on 
the outside of the crown, conditions favored much more rapid evaporation 
than in the interior. He found that leaves on the outside are smaller, 
more deeply lobed, and lower in water content and that they transpire 
faster per unit of area (Fig. 14-2). This has frequently been confirmed. 
Huber (1926) agrees that sun leaves result from a water deficit. Soding 

328 M 01 -pho genetic Factors 

( 1934 ) grew trees in pots and found that, by adding a small amount of 
salt to the soil, leaves like sun leaves were formed by the plant. Light, 
however, probably has a share in the production of xeromorphy, for 

Fig. 14-2. Effect of environment on leaf structure. Sections of leaves of Acer saccharum 
from 1, south periphery of the tree; 2, center of crown; 3 and 4, base of crown. Various 
factors are doubtless involved in these changes, but differences in rate of transpiration 
( higher in exposed leaves than in the others ) seem especially important. ( From 
Hanson. ) 

bright light has been shown to reduce leaf area and to increase blade 
thickness. The xeromorphy observable in many tropical plants may check 
the harmful effects of too intense insolation. 

Still other factors are probably involved in the development of this sort 

Water 329 

of leaf structure. Miiller-Stoll (1947b) has evidence that xeromorphy of 
plants in peat bogs is due to a deficiency of nitrogen rather than of water. 
He fertilized such plants and observed a marked increase in leaf area and 
cell size and a decrease in stomatal frequency, cell- wall thickness, and 
venation. Lack of nitrogen and lack of water thus seem to produce 
similar structural changes. These two factors are associated in other 
morphogenetic phenomena. 

Experimental Work. Many experiments have been performed to de- 
termine directly the effect on plant structure of varying amounts of 
water in the soil or in the air. Only a few of these can be mentioned 


Rippel (1919) studied white mustard growing in moist and in dry 
soil. Vein length per unit area of leaf surface was considerably greater 
in dry soil. In both, it increased progressively from the first leaf to the 
fifth, and this gradient was steeper in dry soil than in wet. 

Penfound (1931) paid particular attention to stem anatomy and found 
that, although increased soil moisture reduced xeromorphic traits, it in- 
creased the relative amount of xylem in the stem. 

Cain and Potzger (1940) brought Gaijlussacia plants into the green- 
house. They varied the amount of available water and also grew some of 
the plants in front of a fan. Though dry air and fan induced some 
xeromorphy, the mesophyll was considerably thicker in plants in the 
moister soil, contrary to most earlier observations. 

Simonis (1952) studied four genera grown in soils of high and low 
water content. In all, the leaf surface was reduced under water deficit, 
but the morphological responses of different plants were somewhat dif- 
ferent. Water content tended to be unchanged under dry conditions, 
or sometimes even was increased, and succulence was generally greater. 
Simons (1956) grew year-old apple seedlings in greenhouse pot cul- 
ture in moist and dry soil. Reduction in water supply affected leaf area 
and also thickness and size of cells in epidermis, palisade, and xylem. 

Fewer experiments have been done on the morphogenetic effects of 
differences in the humidity of the air. Eberhardt (1903) grew a wide 
variety of plants under bell jars, maintaining light and temperature the 
same in all but changing the humidity. Dry air tended to produce the 
ordinary xeromorphic traits and also an increase in hairiness ( Figs. 14-3 
and 14-4). The results of Lebedincev (1927) and Rettig (1929) were 
much the same. The effects of dry air were more pronounced when the 
soil was also dry. 

At the opposite extreme from xeromorphy, produced by water deficit, 
are those changes that result from submersion in water. As a rule, the 
roots of such plants are small or lacking, the vascular and mechanical 
tissues poorly developed, the leaves thin and often much dissected, the 

330 Morpho genetic Factors 

stomata vestigial or absent, and the cell walls thin. These traits are gen- 
erally regarded as adaptations to an aquatic habitat. A few heterophyllous 
water plants (p. 216) can live either submersed or growing in the air with 

Fig. 14-3. Diagram of cross sections of stems of Achyranthes, showing relative de- 
velopment of tissues in air which is dry ( I ) , normal ( II ) , and humid ( III ) . b, xylem; 
1, phloem; s, sclerenchyma; m, pith. Dry air tends to reduce size of pith and cortex 
and to increase development of xylem and sclerenchyma. ( From Eberhardt. ) 

their roots in soil. Among these are Polygonum amphibium and the 
water buttercup, Ranunculus aquatilis. In these plants the land form, 
essentially mesophytic in structure, is very different from the water form 





Fig. 14-4. Outer cortex and epidermis of Aster sinensis grown in air that is of normal 
humidity ( I ) and that is dry ( II ) . The latter shows greater wall thickness in collen- 
chyma and larger bundles of sclerenchyma. E, epidermis; C, cortex; S, sclerenchyma. 
( From. Eberhardt. ) 

and in some cases was not at first recognized as belonging to the same 
species. Occasionally, as in the buttercup, foliage transitional from one 
to the other may be found. In plants like the mermaid weed, Proser- 
pinaca palustris, leaves borne in the air are broad and little-lobed 



Fig. 14 5. A species of Myriophyllum, an "amphibious" plant, showing differences be- 
tween leaves grown in water and in air. ( From Fassett. ) 

whereas the submersed ones are much dissected (Fig. 14-5). Potamogeton 
and similar forms are entirely aquatic but in some species there are 
both broad floating leaves, exposed on their upper surfaces to the air, and 
delicate submersed ones. 

Various explanations have been proposed (p. 216) for the differencs 

332 Morphogenetic Factors 

between shoots grown in water and in air, especially in Proserpinaca. 
Transpiration, seasonal differences, and reversion to juvenile stages may 
be involved. Combes (1947) found that in Oenanthe low temperature 
is effective in producing deeply incised leaves. Allsopp (1955) has 
studied the water fern, Marsilea, grown under various conditions, and 
has discussed the general problem of the structure of water plants. The 
land form of leaf has four leaflets, and there are stomata on both sur- 
faces, but the water form is merely lobed and lacks stomata in the lower 
epidermis (Fig. 14-6). Raising the osmotic concentration of the culture 

5% U. E. 


5% L.E. 


Fig. 14-6. Comparable sporeling leaves of Marsilea Drummondii, an "amphibious" 
plant. Form of leaf and structure of upper and lower epidermis in leaves grown in 
media containing 5 per cent glucose (left) and 1 per cent (right). The former re- 
semble typical land forms and the latter, water forms. ( From Allsopp. ) 

medium by adding glucose produces the land type of leaf. Whether this 
is an osmotic or a nutritional effect is not certain. Allsopp concludes that 
it is the water balance of the developing tissues, determined by the 
osmotic pressure of the surrounding liquid, and the relative humidity of 
the air or, in general, the diffusion pressure deficit of the water of the 
environment, which produce the structural features distinctive of land 
or water forms. Here, again, the morphogenetic problem involves much 
more than the direct effect of a single environmental factor. 

Effect of Transpiration Stream. Another aspect of the problem of the 
morphogenetic effects of water involves the influence of water supply 



and transpiration rate on the development of vascular tissue. Does a 
strong transpiration stream stimulate the formation of conducting cells 
and thus serve as a "functional stimulus"? 

There is clearly a quantitative relation between a transpiring surface 
and the vascular tissue supplying it. D. J. B. White (1954), in a study of 
the relation of laminar area to petiolar xylem in the bean leaf, found 
that there is an allometric developmental correlation between the two, 
the cross-sectional area of xylem growing about two-thirds as fast as 
laminar area (Fig. 14-7). There have also been a good many measure- 
ments of the amount of vascular tissue at different levels in the stem, both 
absolutely and in proportion to the area of leaf lamina above. Reliable 


" Immature 



0-5 10 Log L 15 20 

Fie. 14-7. Relative growth of area of lamina (L) and cross-sectional area of petiolar 
xylem (X) in immature leaves of bean. (From D. ]. B. White.) 

data as to the amount of transpiration in relation to cross section of 
vascular tissue are difficult to obtain. Riibel (1920) measured the xylem 
area at different levels on a sunflower plant and the total leaf area above 
each level. In plants grown in a normally sunny situation there was 
about 0.21 sq. mm. of vascular tissue per square decimeter of leaf area, 
as compared with 0.10 sq. mm. in shaded plants. Since there is more 
transpiration in the sun, there is evidently a relation here between 
transpiration and the amount of conducting tissue. In the lowest stem 
levels there was from Vi to V3 sq. mm. of cross section of conducting 
tissue to every gram of dry weight of leaves above it, but in young and 
vigorous leafy plants this increased to V2 sq. mm., again showing a 
presumptive relation to transpiration. The proportion of wood to phloem 

334 Morphogenetic Factors 

decreased at upper levels. The leaves at different levels transpired at 
about the same rate. 

Alexandrov, Alexandrova, and Timof eev ( 1927 ) observed that in 
Bryonia (a running vine) the number of vessels in any part of the stem 
varies with the dimensions of the leaves in that region. The size of 
the vascular tissue in a petiole is also related to the area of its lamina. 

In fir, spruce, and beech, Huber ( 1928 ) found that the relative con- 
ducting surface (the ratio of the area of conducting tissue to the fresh 
weight of leaves above it) increases from base to apex in the stem. In 
lateral branches this ratio was smaller than in the main stem, a fact 
presumably related to the dominance of the latter. The relative conduct- 
ing surface and the amount of transpiration were determined for various 
plants by Huber and the rate of flow of water through the vascular sys- 
tem calculated. This was found to be relatively high for herbaceous 
plants, low for conifers and xerophytes, and intermediate for broad- 
leaved trees, thus seeming to be related to the amount of transpiration. 
Huber also observed ( 1924 ) that in oak branches growing in bright sun- 
light a square decimeter of leaf area transpired 75 mg./hour and was 
supplied through a cross-sectional area of 0.42 sq. mm. of vascular tissue. 
In branches growing in shade there were 46 mg. of water transpired 
through a vascular cross section of 0.20 sq. mm. In other words, the 
greater the transpiration stream passing through the vascular system, 
the larger this system was. He believes, however, that the amount of 
water carried upward depends primarily on the osmotic pull exerted 
by the leaves and the resistance to flow in the vascular tissues rather 
than solely on the size of the vascular tissue itself. 

These various facts show that there is a definite correlation between 
the amount of water passing through the vascular tissue and the amount 
of such tissue that is developed. This, consequently, suggests that the 
transpiration stream itself acts as a formative stimulus for the differentia- 
tion of vascular tissue. Doubt as to the correctness of such a conclusion, 
however, is raised by some remarkable results reported by Werner ( 1931 ) 
in maize. He was able to grow a plant of almost normal size (suspended 
in the air in a transpiration experiment) that was connected to its root 
system in the soil by only a single extremely thin root about 10 cm. 
long. The entire water supply for this plant, the stem of which was 
several centimeters wide, passed through a vascular strand only about 
0.5 mm. in diameter. Here, at least, there is little evidence that a heavy 
transpiration flow stimulates a proportionate development of vascular 
tissue. Similarly, in a frond of Osmunda, all the water transpired by a 
large blade area is drawn through a leaf trace which is very small as it 
leaves the stele. At the base of the frond this expands into a ring of 
large bundles. 



In such cases as these, one is forced to conclude that much more vascu- 
lar tissue is normally developed than is required to carry the normal 
transpiration stream. The relation so frequently observed between 
area of transpiring surface and cross-sectional area of conducting tissue 
may therefore be simply another instance of developmental correlation, 
of one tissue keeping step with another (p. 107), and may be without 
causal significance for the differentiation of vascular tissue. 

There are other ways in which water may have morphogenetic sig- 
nificance. Positive hydrostatic pressures often occur at the time of early 
and rapid leaf growth, and leaves developing then tend to be large and 
to have shallow lobes. A little later, when sap pressure is lower or absent, 
the leaves are smaller and the lobes deeper. Experiments of Pearsall 

•OOi n n n 








Fig. 14-8. Relation between the percentage of water in the soil and the percentage 
of cleistogamous florets in Stipa. ( From W. V. Brown. ) 

and Hanby (1926) tend to confirm this, since when they applied con- 
siderable hydrostatic pressure to stems while the leaves were develop- 
ing, the lobes were shallow, but they were deeper under less pressure. 
The angle at which the lateral veins go off in palmate leaves seems to 
affect their ability to deliver water. Where this angle is more than 90° 
the flow of water under pressure is reduced. This fact may be related 
to the determination of leaf shape. 

Osmotic pressure has important morphogenetic effects since it is one 
of the factors determining the amount of cell enlargement. Various cir- 
cumstances affect the osmotic concentration of the cell sap. One of these 
is chromosome number, for Becker (p. 40) observed that in a polyploid 
series of moss cells the concentration varied inversely with the degree of 
polyploidy. Schlosser ( 1935 ) found in tomato that maternally inherited 
differences in osmotic concentration in the cytoplasm, as well as in the 

336 M or pho genetic Factors 

environment, affected the expression of genes for plant height. The 
character of growth may also be influenced. If the alga Stigeoclonium 
is grown in relatively high osmotic concentrations, its cells round up 
and divide in all planes to form a so-called palmella colony. In weaker 
solutions vegetative activity is increased and the cells become cylindri- 
cal, divide in only one plane, and form filaments (Livingston, 1900). 

Water may also affect differentiation in other ways, such as the in- 
creased proportion of cleistogamous flowers formed as soil moisture 
decreases (W. V. Brown, 1952; Fig. 14-8). 

A study of the water relations of plants, and especially of the forma- 
tive effects of water on plant growth, has recently been somewhat 
neglected. These relations provide a promising opportunity, however, to 
approach some of the problems of plant morphogenesis from a direction 
different from that of most experimental work today. 



Temperature is obviously of much importance for the physiological ac- 
tivities of a plant since the rate of metabolic processes is markedly 
affected by it. Though its chief significance is in physiology, it also 
influences development in various ways. These have been summarized 
by Went (1953). 

Both light and temperature apparently produce their morphogenetic 
effects by speeding up or slowing down particular physiological processes. 
What the effect in a given case will be evidently depends on the sensi- 
tivity of the plant to these stimuli in a particular part of its body or on a 
particular phase of its development. The effect of temperature is especially 
important on rate of growth. The optimum temperature for this may be 
different in different regions of the plant, at different stages of develop- 
ment, or even at different times of day. This may result in a change in 
proportions of various parts and thus of form and structure. Biinning 
(1935), for example, observed that the later in the season the seeds of 
beans mature, the shorter is the epicotyl of seed and seedling and the 
quicker do the primary leaves reach maturity. This he found to be a 
temperature effect, for high temperature during the 5 weeks preceding 
seed maturity produces longer epicotyls and a slower development of 
primary leaves. 

In Ipomoea, Njoku ( 1957 ) found that the higher the night tempera- 
ture (with a good level of mineral nutrition) the less deeply lobed were 
the leaves (Fig. 15-1). Here what temperature seems to affect directly 
is the rate of production of leaves at the growing point, and this, in turn, 
is correlated with depth of lobing. Many other cases have been found 
where temperature thus exerts an indirect effect on form and structure 
because of the fact that different parts are differently susceptible to 
its influence. 

Thermoperiodism. There is often a daily rhythm in reaction to tem- 
perature as there is to light. At any temperature that is constant through- 
out the 24 hours many plants will grow less rapidly than if their environ- 
ment is relatively cool at night and warm during the day. The optimum 


338 Morpho genetic Factors 

temperature for growth may therefore be different under different con- 
ditions. This thermoperiodism affects growth in various ways (Went, 
1944, 1945, 1948). Went found in tomato, for example, that if the green- 
house temperature is held constant the optimum is about 26.5°C, at which 
there is a steady growth in length of 23 mm. per day. At all other tem- 
peratures, growth is less (Fig. 15-2). Plants kept warm (26.5°) during 
the day, however, but cooler at night ( 17 to 20° ) grow still better, about 
27 mm. per day. It is significant that this low temperature is effective only 
if it is applied during the dark period of daily growth. The same thermo- 
periodism is evident in fruit development, the best fruit set occurring 
when the night temperatures are 15 to 20°. Evidently two different 










-5 20 

1 5" 

J L 


a L 

Fig. 15-1. The effect of low and of 
high night temperature on shape of 
successive leaves in Ipomoea caerulea. 
Shape index measures degree of lob- 
ing. ( From Njoku. ) 

Leaves, numbered from base 

processes are involved, one going on in the light and the other in the 
dark, and with different temperature optima. Plants differ considerably 
in their response to thermoperiodism ( Knapp, 1956 ) . Sproston and Pease 
( 1957 ) have shown that thermoperiodism is related to the production of 
the sexual stage in the fungus Sclerotinia. 

There is a close relation between temperature and photoperiodism, 
for it has frequently been shown that a particular temperature can be 
substituted for day-length in determining the balance between vegetative 
growth and flowering. Thus in Rudbeckia, a long-day plant, flowers are 
produced in shorter days if the temperature is kept high (Murneek, 
1940). Flowering in beets can be controlled by manipulating the relation 

Temperature 339 

between temperature and the photoperiod (Owen, Carsner, and Stout, 
1940), a technique which has been termed photothermal induction. 

The problem of dormancy, of why plants or parts of them fail to grow 
until particular conditions are satisfied, has implications for morpho- 
genesis, particularly in relation to buds, since there are usually very many 
buds on a plant that do not develop. Whatever determines the particular 
ones that are to grow has an important influence on the form of the plant. 
Factors that inhibit or stimulate growth of these meristematic regions 
have been studied chiefly in connection with growth substances (p. 386), 
but others are involved. Among them temperature has an important place. 
It is well known that low temperature is one of the most effective means 
for breaking the dormancy of seeds, buds, and other plant parts. The 

30 -""Vday 



day temperature 16.5* 
night temperature as 
indicated on abscissa 

temperature, constant 
day and night 

3k>- 2S° 20° 15° 10° S'C 

Fig. 15-2. Thermoperiodicity in tomato. Stem growth (in millimeters per day) in 
plants kept constantly at the indicated temperature ( lower curve ) and for 8 hours dur- 
ing the day at 26.5 °C but at night at the temperature indicated (upper curve). (From 
Went. ) 

influence of temperature on bud growth has been widely studied by 
horticulturists because of its practical importance (see Chandler et al., 

Vernalization. A more significant effect of temperature for morpho- 
genetic problems is evident in the processes of vernalization, by which 
flowering is accelerated through the application of low temperature at a 
particular stage of development. Horticulturists have known for many 
years that the chilling of seeds or seedlings will in many cases force 
plants into bloom earlier than would otherwise occur. Scientific study 
of this effect, however, began in comparatively recent years and at first 
was explored chiefly by Russian plant physiologists, especially Lysenko. 
A conspicuous example of the effect of vernalization is the speeding up 

340 Morphogenetic Factors 

of flowering in some of the cereal grains so that "winter" varieties, 
superior in certain respects, can be treated like "spring" varieties. A win- 
ter race of rye, for example, is normally sown in the fall and flowers and 
fruits the next season. If planted in the spring it will produce abundant 
vegetative growth but no flowers. A spring race sown in the spring will 
fruit in that growing season. If seed of winter rye is soaked in water, 
however, and is then exposed to low temperature (0 to 10° C or there- 
abouts) for a few hours or days, it can be sown in the spring and will 
bear flowers and fruit just as rapidly as spring rye does. Grain germi- 
nated at 1°C and then planted will produce rye that flowers in 68 days, 
but if germinated at 18 °C, flowering does not take place for 150 days 
(Gregory and Purvis, 1938). The vernalized seed may be kept dormant 
for some time or even dried and it will still grow like spring rye when 
it is planted. 

These facts are explained on the assumption that development is to a 
great extent independent of growth, that in an annual seed plant there 
is a specific series of developmental stages, each a necessary precursor 
to the next, and that these stages require for their completion different 
environmental conditions, especially as to temperature and light. This 
is an aspect of the general concept of phasic development previously 
discussed (p. 205). 

In these cereals the first stage is the one in which the floral initials 
are formed, and for this process low temperatures are necessary. Winter 
rye sown in the fall will produce these initials because it is exposed to 
the low temperatures of winter but if it is sown in the spring the tem- 
perature is not low enough for this to happen. Vernalization thus acts as 
a substitute for winter temperature. 

The next stage, the development of the flowers themselves, usually 
requires higher temperatures but also relatively long days. Flowering 
will be delayed indefinitely if plants at this stage are exposed to short 
light periods. In the long, warm days of spring, both winter and spring 
varieties will thus come into flower. The difference between them is that 
spring varieties will form floral primordia in these warm days and winter 
varieties will not do so without treatment. 

In certain plants a definite number of primordial structures are formed 
at the growing point of the embryo, and it can be shown that the fate of 
some of these has not been determined in the seed but that leaves or flow- 
ers will form from them, depending on environmental conditions. Thus 
in winter rye there are usually 25 embryonic primordia. The first seven 
will always develop into leaves. The next 18 are indeterminate, and the 
lower the temperature to which they are exposed, the more of them will 
develop flowers. At high spring temperatures, none of them will do so 
unless previously chilled. 

Temperature 341 

Different plants respond very differently to vernalization, and in some 
it is without effect (Kondratenko, 1940). Vernalized plants may be de- 
vernalized, usually by high temperature following the cold treatment, 
and they may sometimes even be revernalized (Lang and Melchers, 

The induction of flowering by low temperature is by no means limited 
to the cereal grains or to seed treatment. Young plants beyond the seedling 
stage may be vernalized and thus forced into flower, and some biennial 
varieties will flower in their first season if subjected to cold. The grow- 
ing stem tip is the region sensitive to the vernalizing influence. Low- 
temperature effects on growing plants, particularly as to flowering, 
have been widely investigated. This work is reviewed by Thompson 

The exact way in which low temperature produces its effects in ver- 
nalization is not clearly known but in some cases it has been thought to 
influence the production and distribution of auxin and perhaps also of 
substances that stimulate flowering (p. 397). Hatcher (1945), however, 
finds that the auxin content of grains of winter and spring races of 
cereals is the same and that there is no detectable amount of auxin in 
the embryos either at normal or low temperatures. He concludes that it 
is not concerned in the process of vernalization. 

Although the most conspicuous effect of vernalization is the acceleration 
of flower development, vegetative characters may also be affected, such as 
leaf size. Hansel ( 1953 ) found that early leaves were longer if germina- 
tion temperatures were slightly below 0°C than if they were slightly above 
this (Fig. 15-3). Internal differentiation is also affected (Roberts and 
Struckmeyer, 1948). 

The literature of vernalization and of its relation to photoperiodism 
and phasic development is extensive. The history of research in this gen- 
eral field has been reviewed by McKinney (1940) and Whyte (1948). 
The latter is one of a series of papers on this general subject brought to- 
gether by Murneek and Whyte (1948). Among other related publica- 
tions are those of Gregory and Purvis ( 1938), and Whyte (1939). 

Other Temperature Effects. There are many other instances where 
morphogenetic effects of temperature have been observed. Some typical 
examples of recent investigation in this field are the following: 

Burstrom ( 1956 ) finds that under higher temperatures the final length 
of cells in roots is reduced because of the shorter period of cell elonga- 
tion. Cell-wall plasticity and calcium requirement are also reduced. 

Schwabe (1954), working with Chrysanthemum, limited low-tempera- 
ture treatment to the growing tip and confirmed earlier conclusions that 
this is the region where the stimulus of vernalization is perceived. The 
stimulus did not p^tss across a graft union but it was translocated to 

342 Morphogenetic Factors 

lateral buds that were distant from the one that was chilled and had 
formed some time after it was treated. 

Wittwer and Teubner ( 1957 ) observed that in tomato low-temperature 
treatment of seeds had no effect, but exposure of very young seedlings to 
low night temperatures (10 to 13°C) for 2 or 3 weeks induced earlier 
flowering and more flowers in the first cluster, in contrast to those grown 
at higher temperatures (18 to 21°C). Cold treatment of older seedlings 
increased the number of flowers in later clusters. Other factors, especially 


• 1st Lamina 
a 2d Lamina 
O 'Scores' 







- • 



3o 3 

J L 

I I I 

an- "3 -2 -60 *l *3 
treated Temperature of Vern. oq 



Fig. 15-3. Effect of vernalizing temperatures on length of lamina of first (A) and 
second ( B ) leaves in winter rye, and on stage of differentiation of spike as meas- 
ured by an arbitrary scale of "scores." Controls at left. ( From Hansel. ) 

nitrogen nutrition, also affect flower formation and complicate the prob- 
lem of studying it. 

Hall (1950) compared buckwheat plants in culture with their roots and 
shoots at different temperatures with others where the entire plant was 
grown at either high or low temperature. Development was more normal 
under the latter condition. High temperature for the shoot checked vege- 
tative growth and hastened flowering, maturity, and senescence, and low 
air temperature there prolonged ontogeny. Increase in duration of the 

Temperature 343 

vegetative phase, however, did not result in the production of more plant 


L. D. Tukey ( 1952 ) subjected bearing branches of sour cherry to sev- 
eral different night temperatures and found that higher temperatures 
accelerated development during stages I and II (early growth and stone 
formation) but checked it in stage III (fleshy pericarp growth) (Fig. 

2-5). . . 

Leopold and Guernsey (1954) treated germinating peas with various 

growth substances and followed this with low temperature. The combi- 
nation of chemical with temperature stimulation they termed chemical 
vernalization. It hastened flowering, but only if carbon dioxide was later 
present. Changes in day-length had no effect. They conclude that there 
are two stages in the growth of young pea plants which are affected, the 
first requiring auxin and low temperature and the second requiring 
carbon dioxide. The function of carbon dioxide here is not understood. 

Chaudri, Biinning, and Haupt (1956) observed that the exposure of 
young onion plants to 3 hours of low temperature during the dark por- 
tion of the photoperiod hastened the development of bulbs. This effect 
was greatest when the low temperature was applied during the latter 
part of the dark period. 

Fisher ( 1954 ) worked with a trifoliate New Zealand species of Ranun- 
culus in which the juvenile leaves are undivided. Sometimes the adult ones 
show a partial reversion to this juvenile form. He grew plants under 
controlled conditions and found that when the temperature was relatively 
high (20°C in the daytime and 15°C at night) there was a complete 
reversion to the undivided juvenile leaf shape but that at lower tem- 
peratures ( 10 and 5° ) the adult form persisted. 

Steinberg (1953) studied Mammoth Rustica tobacco, a type which 
came originally from a cross between Nicotiana rustica and N. tabacum. 
This is indeterminate in growth and very rarely flowers, but it can 
readily be made to do so if the night temperatures are dropped to 50 
or 60°F, regardless of day-length. In this respect it is unlike Maryland 
Mammoth (p. 316), which also is indeterminate in growth but flowers only 
in short days, regardless of temperature. The indeterminate character of 
growth in both is due to the fact that flowering is prevented, in one type 
by high temperature and in the other by long days. Steinberg suggests 
that there may be a separate genetic basis for the two types of reaction. 

Benson-Evans and Hughes ( 1955 ) observed that in the liverwort 
Lunularia cruciata, which is world-wide in distribution but rarely repro- 
duces sexually except in a "Mediterranean" type of climate, the induc- 
tion of archegoniophores requires subjection to low temperature, later 
followed by higher temperature and long days, thus fitting it to its par- 
ticular ecological distribution. 

344 M or pho genetic Factors 

Margalef ( 1953 ) found that in cultures of the green alga Scenedesmus 
obliquus low temperatures cause cell size to increase, although other fac- 
tors have a minor effect on it. Since cells in this species grow faster 
in length than in width, large cells have a more slender shape than 
small ones, so that temperature indirectly affects cell shape. 


Various Physical Factors 

There are a number of other morphogenetic factors which may be grouped 
together as physical ones in a general sense, notably such clearly me- 
chanical factors as external compression, tension, bending and swaying; 
gravity and inner tissue tension, together with absolute size and bio- 
electrical factors, each with its bearing on development. 

Those factors which may be called mechanical in the strict sense are 
relatively simple in character in comparison with light, electricity, and 
many chemical ones and evidently produce most of their morphogenetic 
effects indirectly through modifying physiological processes in the living 
cells. Much of the early work in this field was done by Schwendener 
( 1878, 1898 ) . Among other problems, he emphasized the importance of 
such factors in determining the arrangement of leaf primordia at the 
apical meristem. 

Mechanical effects are in many cases much like the ones produced by 
other factors, suggesting that both are acting upon the same proto- 
plasmic mechanisms. Such parallel effects are familiar to the student of 
morphogenesis and emphasize again the importance of the complex re- 
acting system rather than that of the relatively simple stimulus or evo- 
cator. There has been a good deal of disagreement as to experimental 
results in this field, much of which is probably due to the fact that the 
reactivity of the plant is very different at different stages of its develop- 
ment and under different environmental conditions. 

In studying these effects it is sometimes difficult to separate various 
plant movements and tropisms from more strictly form-producing and 
morphogenetic phenomena. Changes in the positions of parts, as in the 
leaves of Mimosa, the fly-traps of Dionaea, the stamens of various plants, 
and other structures, are due chiefly to changes in turgor brought about 
by specific substances. This is essentially a problem in plant physiology 
and offers opportunity to study the mechanisms of stimulus and response, 
with their various chemical and electrical correlates. 

Tropistic responses to gravity and light, however, are usually due to 
more rapid cell elongation on one side of the axis than on the other and 
are thus, in part, growth reactions. The various thigmotropisms, or 


346 Morpho genetic Factors 

responses to contact, involve more definitely morphogenetic changes. 
When the tip of a tendril is lightly touched on one side, as by a small 
branch or wire, the tendril will coil around it and thus tend to anchor 
the plant to a support. This coiling results from the more rapid growth of 
that side of the tendril not touching the support. There is evidence that 
the stimulus of contact tends to produce a substance that checks growth, 
although the mechanism which makes a tendril react thus, and later 
contract in a coil, pulling the plant toward a support, is not well under- 

Mechanical factors are also concerned in other growth reactions. Biin- 
ning and his colleagues (1941, 1948, 1954), working with Mimosa, Si- 
napis, and Vicia, have forlnd that mechanical stimulation (stroking with 
paper or agitating on a shaking machine) checks the lengthening of 
stems in darkness ( etiolation ) in much the same way that light does. The 
internal structure of stems grown in darkness but mechanically stimu- 
lated is quite different from that of etiolated stems and hardly to be 
distinguished from ones growing in the light (Fig. 16-1). He suggests 
that both mechanical stimulation and light partially inactivate auxin ac- 
tion. Both stimuli are more effective if repeated at intervals than if applied 
continuously, a fact which may be due to a refractive stage following 
the stimulus. 

Borriss ( 1934£> ) showed that Coprinus fruiting bodies which would not 
have matured in the darkness will do so, at least partially, if mechanically 
stimulated, and Stief el ( 1952 ) finds that the stipe of the fruiting body of 
Coprinus responds to mechanical stimulation and to light just as do the 
stems of higher plants, both stimuli tending to check elongation. 

Mechanical pressure may have an important morphogenetic effect by 
determining the plane of cell division in meristematic tissue (p. 49), the 
plane of division tending to be parallel to the direction of the pressure. 

It is with the more specifically morphogenetic effects of mechanical 
stimuli, however, that we are particularly concerned here. Chief among 
these are tension, compression, bending, swaying, and the omnipresent 
stimulus of gravity. 

Tension. Much work has been done on this problem, especially in 
earlier years, but the results are often contradictory. Hegler (1893) 
stretched seedlings of sunflower and petioles of Helleborus by attaching 
weights to them. After a 2-day application of 150 gm., it required a 
pull of 350 gm. to break these structures as compared with only 160 gm. 
in the controls. Traction seemed to have increased the tensile strength, 
and Hegler found that traction had increased the cell-wall thickness and 
the amount of collenchyma. 

Newcombe (1895) reviewed the very considerable amount of early 
literature in this field. His own work confirmed Hegler's. He also found 

Various Physical Factors 347 

that if stem bases were enclosed in plaster and thus relieved of mechani- 
cal strain they produced less mechanical tissue. 

Ball (1904), however, repeated Hegler's experiments, using pulleys 
and carefully comparing stretched plants with their controls, but found 
no difference between them in structure or tensile strength. Hibbard 
(1907) confirmed Ball. Still later Bordner (1909) studied the problem 
again and confirmed Hegler's results, using similar material. In stretched 
plants the amount of vascular tissue and the tensile strength were in- 

Fig. 16-1. Vicia faba. Outer portions of 
transverse sections through the second 
internode. a, grown in darkness without 
shaking; b, grown in light; c, grown in 
darkness and shaken. Mechanical stimu- 
lation has much the same effect on 
growth as does light. ( From Bunning. ) 

creased. He found that no effects were produced unless the plants were 
growing, a fact that may help explain the conflicting results of these 
various workers. 

Flaskamper (1910) and others subjected flower and fruit stalks to ad- 
ditional weighting but found no change in histological structure. Wieders- 
heim ( 1903 ) hung weights on branches of a weeping variety of beech 
and reported that these grew less rapidly and had shorter cells but formed 
no additional vascular tissue. 

In all this early work on mechanical factors the suggestion was natural 

348 Morpho genetic Factors 

that the plant would tend to react in such a way as to oppose the effect 
of the factor and, in the cases here discussed, to develop more mechanical 
tissue which would resist traction and prevent breakage of the plant. 
Vochting (1878), with his versatile interest in all such problems, asked 
the practical question as to whether plant axes, subject to different ten- 
sile strains in nature, showed structural differences as a result. He com- 
pared the pedicels of squash fruits hanging from trellises with similar 
ones growing on the ground and found more vascular tissue in the 
former. However, when such a fruit was supported on a platform beside 

Fig. 16-2. Effect of tension on wood structure. Left, transverse section of wood of 
Fagus sylvatica from a root grown under strong tension; right, section of root of the 
same species not under tension. Note generally thicker cell walls in the former. (From 
Jaccard. ) 

a freely hanging one, he found the same amount of vascular tissue in 
the stalks of both. Vochting suggested that the differences first observed 
were due to differences in the amount of transpiration rather than to 
tension. He also tried to induce mechanical tissue by traction in weak, 
poorly vascularized plants, but without success. However, he grafted a 
normal shoot on such a weak one and observed the development in the 
latter of a marked increase of vascular tissue. 

It might be objected that stems are not usually subject to tension but 
that roots are, and it was evidently important to study these organs as 
well. Wildt (1906) fastened the seedling stem and the adjacent part of 

Various Physical Factors 349 

its root in plaster and pulled gently. The soft central tissue disappeared 
and a solid vascular core resulted. Flaskamper, who repeated and con- 
firmed this, found that roots which had been subjected to traction had 
somewhat less tensile strength than the controls. Newcombe (1895) 
stretched roots of sunflower and squash and observed that they grew 
somewhat stouter and were definitely stronger than the controls. Jaccard 
(1914) studied some experiments of nature in this field, notably cases 
where a small root crosses a larger one and is stretched by the growth in 
diameter of the latter. In the root under tension the cell walls of the 
wood were thicker than in the control (Fig. 16-2). The amount of me- 
chanical tissue was less, however. 

Fig. 16-3. Effect of tension on a tendril. 
Device for subjecting part of a tendril 
to tension by weight over pulley (A) 
and for relieving the other part by hav- 
ing a cord bear all the tension ( B ) . 
Stimulus of contact is the same in both. 
( From Brush. ) 

Among the organs of a plant most commonly subject to tension in na- 
ture are tendrils. Many experiments have been undertaken with them 
to determine whether traction (pulling) affects their structure. Much 
difficulty was found by early workers in separating the effects of contact 
(to which stimulus tendrils are particularly susceptible) from traction. 
Brush (1912) placed a tendril in lengthwise contact with a thread to 
the free end of which, thrown over a pulley, weights were attached. In 
the control, this thread was continuous and sustained all the pull. In the ex- 
perimental one the thread was interrupted in the middle so that the 
tendril itself bore all the pull. Each tendril was thus in contact with the 
thread through all, or almost all, its length, but one was under tension 
and the other was not (Fig. 16-3). In both, there was more xylem than 

350 Morphogenetic Factors 

in tendrils having no stimulus of contact, but in the one under traction 
the walls of the fundamental tissue cells were markedly thicker than in 
the other tendrils. 

Compression. Because of technical difficulties, not as many attempts 
have been made to produce the opposite sort of mechanical stimulation, 
compression in the lengthwise direction. 

Pennington (1910) hung weights on woody and herbaceous stems of 
various sorts as they were growing in height but found no appreciable 
effect on structure or mechanical strength. Himmel (1927) used more 



Fig. 16-4. Graph showing growth of Podophyllum petioles, in inches per hour, under 
lengthwise pressure from various weights. Upper curve, control; lower, experiment. 
Growth is markedly reduced by pressure. ( From Himmel. ) 

favorable material, the growing petioles of the large, umbrella-like leaves 
of Podophyllum. On the apices of these petioles he hung weights which 
were periodically increased as growth continued. He found that growth 
rate in the weighted petioles was less than in the controls but that growth 
in the former finally equalled that in the latter (Fig. 16-4). The rigidity 
of the petioles was somewhat increased. 

Rasdorsky ( 1925 ) , working with sunflower and marigold, approached 
the problem in another way. He held up the plant by a gently stretched 
thread attached to the upper part of its stem and thus relieved it from 

Various Physical Factors 351 

supporting its own weight. Plants thus treated were definitely weaker 
than the controls. No structural changes are reported. The upward pull, 
weak as it was, may have stimulated growth in length and thus tended 
to make the plant top-heavy. 

Newcombe (1895) supported the base of the stem in young sunflower 
and squash plants by encasing it in plaster, thus relieving it of mechani- 
cal strain. In that part of the stem under the cast much less mechanical 
tissue was formed, although this developed rapidly when the cast was 
removed. Other effects of such treatment, especially on respiration, might 
account for the results obtained. 

The studies on experimental traction and compression are contra- 
dictory and indecisive, and many factors other than mechanical ones 
may well be involved. Little work has been done in this field in recent 
years. Schwarz ( 1930 ) subjected the problem of mechanical factors in 
development to critical review and concluded that they have little effect 
and that the results attributed to them may well be due to nutritional 
influences and transpiration. Rasdorsky (1931), however, took issue 
with him strongly. 

Bending and Swaying. There is much more agreement as to the effect of 
bending and swaying plant organs. Most workers find that, when grow- 
ing herbaceous stems are bent, the cells (especially collenchyma and 
bast fibers) on the convex side are smaller in cross section and thicker- 
walled than corresponding ones on the concave side. The same results 
are evident in plants grown on a clinostat, showing that gravity is not 

A good discussion of this problem is presented by Biicher ( 1906 ) . To 
this result of bending he gave the name camptotrophism. Since cells on 
the convex side of the bend are evidently under tension and those on the 
concave side under compression, the histological differences observed 
seem related to the type of mechanical strain involved. These effects 
agree with the ones from the experiments just reported where tension 
seems to increase wall thickness and reduce cell size, and compression 
produces the opposite result. Biicher obtained more direct evidence in 
support of this conclusion. He enclosed in plaster most of the lower 
portion of a growing hypocotyl of Ricinus. The upper part was enclosed 
in another casing of plaster, leaving a short portion of hypocotyl unen- 
closed between the two. The weight of the upper layer of plaster was 
supported by the hypocotyl, which was thus subjected, particularly in 
its free portion, to considerable lengthwise compression (Fig. 16-5). In 
the control plants the hypocotyl had relatively small and thick-walled 
cells. In the compressed hypocotyls, however, the cells were much 
larger and had very thin walls. This sort of experiment seems worth re- 
peating with modern techniques of analysis. What, for example, would 

352 Morphogenetic Factors 

the electron microscope show as to the character of such cell walls laid 
down under pressure? 

Biicher found that when a growing shoot, negatively geotropic, is 
forcibly kept horizontal the cells of the upper side are smaller and 
thicker-walled than those of the lower. This phenomenon he called 
geotrophism and explained it on the assumption that since the lower 
side grows more rapidly (tending to bend the stem upward) it must 
be under compression and the upper side under tension. To test the 
relative effects of gravity and bending, he inverted a plant of Ricinus 
and held its tip bent at 90° to the axis (and thus horizontal). Campto- 
trophism and geotrophism should thus be brought into opposition. Actu- 

Fig. 16-5. Effect of vertical compression on hypocotyl of Ricinus. At left, young 
seedling, protected by sheath and cotton plug, held in two blocks of plaster and there 
subjected to compression from weight of plaster. At right, bast bundle from upper part 
of such a hypocotyl ( above, under compression ) and from the normal control ( be- 
low), showing increase in cell size and reduction in wall thickness as a result of com- 
pression. ( From Biicher. ) 

ally, the plant reacted as it would to gravity, with smaller cells on the 
upper, concave side. Perhaps in such a case it is the difference in degree 
of tension that determines the result. 

Vochting ( 1908 ) placed the stem of a potted plant in a horizontal po- 
sition and put a support under it at some distance from the pot. On the 
free end he hung weights and observed a considerable increase in the 
cross section of the stem, especially of its vascular system. This was most 
marked near the point of support, where the strain was greatest, and on 
the upper and lower sides. Growth of the stem in length had ceased, 
so that these changes resulting from mechanical stimulation were evi- 
dently in the secondary tissues. Haerdtl (1927) found similar results. 

Various Physical Factors 353 

Many experiments have been conducted, beginning with Knight 
(1811), on the effect of continually swaying plants as they grow. The 
swaying was done by clockwork and pendulum, water wheel, or motor. 
Observers are generally agreed that the cross section of the axis tends to 
be elliptical under these conditions, with its wider dimension in the 
plane of sway (Fig. 16-6), and that more vascular and mechanical tissue 
is developed than in the controls (Rasdorsky, 1925; Burns, 1920). 

It may be objected that in these swaying experiments conditions are 
so abnormal that results may be due to other factors than purely mechani- 
cal ones. The experiments of M. R. Jacobs (1954) are significant here. He 
supported the trunk of a young pine tree with guy wires attached about 

Fig. 16-6. Effect of swaying on sunflower stem, cross sections. At left, control. At 
right, stem swayed for 3 weeks in the vertical plane of the figure. ( From Rasdorsky. ) 

20 feet from the ground so that lateral movement below this point was 
prevented although the trunk above the point of attachment was sub- 
ject to ordinary wind sway. This portion increased normally in thickness 
but the unswayed lower portion grew much less rapidly than the upper 
or than the unsupported controls. When the guy wires were removed, 
this part grew rapidly in thickness until its normal diameter had been 
attained. Evidently in such cases swaying stimulates cambial activity. 
This may be a factor in producing the relatively slender trunks of trees 
in a forest as compared with those grown in the open. 

A somewhat similar case has been reported by Venning (1949) for 
celery, an herbaceous plant. One series of seedlings was grown in con- 

354 Morpho genetic Factors 

stant wind and another in a windless environment. The former developed 
50 per cent more volume of collenchymatous tissue than the latter. In all 
such cases, however, it is difficult to separate mechanical effects from 
those caused by increased water loss. 

Ultrasonics. The morphogenetic effect of a quite different mechanical 
factor, intense ultrasonic vibration, has been reported by several work- 
ers. Takashima and others ( 1951 ) found that in germinating radish seed- 
lings exposed for 16 minutes the shoots were much deformed. In peas 
similarly exposed, length of root and shoot was increased through in- 
creased cell size. 

Gravity. Gravity is unlike the mechanical factors just discussed in 
that it is continuous, unchanging in intensity, and constant in direction. 
It is one of the most important formative factors, for plants must con- 
tinually regulate their growth to it. The upright position of main-shoot 
axes, the downward growth of primary roots, and the various inter- 
mediate orientations of leaves and lateral branches and roots are mani- 
festations of geotropic growth reactions. The general growth pattern of 
the plant body is a specific reaction to gravity. The problem of tropisms 
is primarily one for plant physiology but the student of morphogenesis 
should not lose sight of the fact that these tropisms, whether reactions 
to gravity, light, or other stimuli, are continually molding the pattern of 
the plant. 

One can distinguish between the tropistic effects of gravity in the strict 
sense, which involve the orientation of parts, and its truly formative ones. 
Conspicuous among the latter are the modifications of symmetry from 
radial to dorsiventral or vice versa (p. 176). Many years ago Wiesner 
(1868) and others studied the dorsiventral symmetry of plant struc- 
tures, especially leaves, when the axis on which they were borne was 
horizontal instead of vertical, and succeeded in inducing form changes in 
them experimentally. For such a difference in form between the upper 
and the lower leaves of a horizontal branch or between the upper and 
lower sides of such leaves, Wiesner proposed the term anisophyllij (p. 
171). Goebel later distinguished between anisophyllous forms which are 
constant and hereditary and those which can be reversed by changing 
the relation of the growing structures to gravity. 

Gravity has been found to modify internal structure also. Brain (1939) 
grew various plants horizontally on a clinostat and found that this modi- 
fied the cells somewhat, those on the clinostat being generally shorter 
and wider, presumably because of the greater extensibility of their cell 
walls. Larsen (1953) found that gravity has little effect on rate of cell 
elongation when acting on roots growing in the normal direction but 
checks such elongation when acting in the opposite direction or at right 
angles to it. Imamura ( 1931 ) was able to change the position of palisade 

Various Physical Factors 355 

and spongy parenchyma by reversing the orientation of Iris japonica 
leaves to gravity. Kreh (1925) examined changes in the fruiting bodies 
of the fungus Lenzites which had been turned through an angle of 180° 
with relation to gravity and found that in new growth the original 
dorsiventral structure was restored. Many other cases of the effect of 
gravity on structure have been reported. 

It is often difficult to distinguish the effects of gravity from those of 
light, since a change in the relation to one usually produces a change 
in the relation to the other unless experimental conditions are carefully 
controlled. Wiesner (1892c) in a later paper recognized that his earlier 
results in anisophylly were due to light as well as to gravity, and 
Bussmann ( 1939, 1941 ) found the same to be true of induced dorsiven- 
trality in fern prothallia. 

Gravity presumably does not modify plant structure directly but acts 
through its influence on other factors. It produces tension and compres- 
sion of tissues with consequent effects on cell division and expansion. 
The reaction of woody plants in developing stem tissues strong enough 
to resist bending and wind sway has been mentioned, and there must be 
the same regulatory control of growth to resist weight of trunk (Esser, 
1946). In this sense, gravity acts as a mechanical factor. Opatowski 
(1946) explains the oblique growth of trees under the action of prevailing 
winds as a mechanical response, based on the concept of maximum 

Equally important is the role of gravity in the distribution of growth 
substances. Just how this occurs is not clear, but differential distribution 
of substances under the stimulus of gravity must be involved (p. 380; 
Brauner, 1927). This presumably affects the form of structures when 
symmetry is changed from radial to dorsiventral, for example. It also 
has less direct effects. Van Overbeek and Cmzado ( 1948 ) and Fisher 
(1957) have shown that flower formation is geotropically stimulated in 
horizontally placed pineapple and soybean plants, presumably by 
alteration in the distribution of growth substances. Other phenomena 
of differentiation may perhaps be explained in the same way. 

Gravity serves as the frame of reference to which the whole pattern 
of plant growth is regulated. A plant develops under the constant and uni- 
form stimulus of gravity, and its tropistic and morphogenetic responses— 
hardly to be distinguished from each other here— are such that a spe- 
cific bodily form is produced. Without this regulation to a steady orient- 
ing directive, the general pattern of the plant body would doubtless be 
much less specific and might even fail to be developed. Because of their 
being anchored in one spot, plants are much more sensitive of such 
gravitational form control than are animals. 

Reaction Wood. The example of such control that has been worked 

356 Morphogenetic Factors 

out most fully is one which was first observed in the development of 
reaction wood of conifers and later in similar tissue in angiosperms. 
Students of wood structure have long noticed that horizontal branches 
of coniferous trees are excentric in cross section (p. 175), with the pith 
nearer the upper side than the lower. Below the pith is a wedge-shaped 
sector of wood, reddish in color when freshly cut, and hence often 
called "rotholz." Since the lower side of a branch is obviously under 
compression, this was long regarded as the cause for the development 
of this sort of structure, and it was called "compression wood." Its cells 
are somewhat shorter than those of normal wood, and the micellae in 
their walls are less steeply pitched. Such wood is absent, save exception- 
ally, in vertical axes and thus in the main trunk but develops wherever 
such an axis is forced to grow at an angle from the vertical, as in a tree 
bent partly over. Such wood grows somewhat more in length than does 
normal wood and thus produces considerable longitudinal compression. 
If a tree grown in an oblique position is sawn across, beginning on the 
lower side, the saw therefore tends to bind. 

This tissue has now been shown not to be the result of growth under 
compression but to have a nature and function of much morphogenetic 
significance. It is now more correctly called reaction wood. Together 
with the corresponding tissue in angiosperms, it has been studied by 
various wood anatomists, especially Hartmann (1932, 1942, 1943), whose 
work has been extended somewhat by Sinnott (1952). In young conifers, 
especially pines, before the growing season began, these workers tied 
some of the lateral branches downward and others upward. They also 
tied the tip of the main axis into a position out of the vertical, sometimes 
even in a loop. When new shoots of the current year developed from 
the terminal buds of these various axes, these new shoots tended to have 
the same direction of growth that they would have had if the shoot out 
of which they grew had not been fixed in an atypical position (Fig. 16-7). 
The main axis bent around so that it again grew upward. The lateral 
branches (in white pine) grew out at an angle of about 70° from the 
main trunk and thus from the directional pull of gravity. In other words, 
there was, in a sense, a regeneration of the normal growth pattern of 
the tree. 

In this process the reaction wood performs an essential function, for it 
always occurs in such a place that its greater lengthwise growth will tend 
to bend the new shoot (and the free portion of the old shoot) into a 
direction which would be normal for it. If a lateral branch has been 
bent upward, for example, the reaction wood will be on the upper side, 
for this will push the branch down. Such a change continues until the 
orientation of the shoot is normal, when reaction wood again develops on 
the lower side, as in ordinary branches. Its function here seems to be to 

Various Physical Factors 357 

counteract, by its upward push, the downward pull caused by the 
weight of the branch. A bent-over terminal shoot develops reaction wood 
on its lower side and thus pushes the axis up to vertical. If the terminal 
shoot is removed, an excess of reaction wood begins to develop on the 
under side of the lateral branches just below the tip and one of them 
is finally pushed up to the vertical and becomes a new "leader." Some- 
times two share this leadership, neither becoming quite vertical. 

Fig. 16-7. Development of reaction wood in pine in three branches which had been 
bent upward artificially. Reaction wood normally is formed on the lower side but in 
the new growth on these branches it develops on the upper side after bending and 
thus tends to push them back to their normal orientation. (From Sinnott.) 

Angiosperm branching is regulated in much the same way except that 
the reaction wood here is normally on the upper side of a branch and 
acts by producing tension rather than compression, thus pulling the 
branches into place instead of pushing them. Wardrop ( 1956 ) believes 
that the distribution of tension (reaction) wood in Eucalyptus is regula- 
tory and operates to maintain normal tree form much as reaction wood 
does in the conifers. 

That compression itself is not responsible for the formation of this 
wood is shown by the fact that when a terminal shoot is bent around and 

358 Morpho genetic Factors 

tied in a vertical loop, reaction wood is formed on the under side of both 
the upper and lower parts of the loop. In the former, the wood is under 
compression but in the latter under tension. That gravity is not directly 
responsible is shown by the appearance of reaction wood on the upper 
sides of branches which are being pushed down. In every case the de- 
velopment of this wood is such that it will bring back the normal pattern. 

The reaction of each part of the plant seems to be a specific orienta- 
tion to the direction of gravitational pull. This reaction is different in 
different parts of the growth pattern of the tree. In herbaceous material 
(branches of Aster) the author has found that a lateral branch tied out 
of position will tend to assume its normal angle to gravity rather than 
its normal angle to the main axis of the plant. When gravity is replaced 
experimentally by centrifugal force, reaction wood is also produced 
(Scott and Preston, 1955). 

Auxin has been shown to be responsible for the relation between a 
terminal bud and lateral ones below it (p. 386), and it is presumably 
concerned with the production of reaction wood (Wershing and Bailey, 
1942). The problem of morphogenetic significance is why there is just 
enough auxin (and thus enough reaction wood) at just the right place 
and time to produce such a specific pattern of branching that this can 
be used as a diagnostic character for the species. Here is the problem 
of organic form in one of its simplest but most puzzling manifestations. 

Spurr and Hyvarinen ( 1954a ) have reviewed the literature of reaction 
wood in the conifers. 

Tissue Tension. Another factor, mechanical in its nature, which may 
be of some morphogenetic importance is tissue tension. Not all the cells 
in a tissue are equally turgid, and cell walls differ in their elasticity and 
their plasticity. Tissues also grow at different rates. These differences 
often cause tensions between cells or groups of cells which, since plant 
cells adhere firmly to each other, cannot be reduced by cellular readjust- 
ments in position. 

In an early and thorough paper Gregor Kraus (1867) examined this 
problem. He measured the length of a piece of growing herbaceous inter- 
node and then sliced it into longitudinal strips, each consisting of only 
one tissue (pith, wood, cortex, or epidermis). When he measured the 
length of these, he found that the outer ones had shrunk in comparison 
with their original length before isolation but that the inner ones had 
expanded. Evidently there was considerable tension between them in 
the intact internode, the outer tissues being stretched and the inner ones 

The degree of tissue tension is not constant but is usually low in young 
internodes, increases farther back, and finally decreases in most cases 
to zero as growth finally ceases. The distribution of tension among the 

Various Physical Factors 359 

tissues also changes. Schiiepp (1917) found that there was tension in the 
growing point but that here it was opposite in its distribution from that 
in tissue which was extending. Schneider (1926) found no tension in the 
growing point itself but saw it first in the leaf primordia, which for this 
reason tend to bend inward. 

What bearing tissue tension may have on differentiation or on the de- 
velopment of form is not clear but it may be of some importance. In 
young ovaries, however, where presumably form and structure are being 
determined, there seems to be little tension though this increases in 
later stages of development. A remarkable instance of tension in dead, 
dry wood has been reported by M. R. Jacobs ( 1945), who found that if a 
board which includes the whole width of the log is sawed at one end part 
way down by a series of parallel longitudinal cuts the strips thus sepa- 
rated tend to spread apart fanwise, indicating the existence of a very 
considerable degree of tension between the inner and the outer parts of 
the log. 

Absolute Size. Another morphogenetic factor, which in a sense may 
be regarded as physical, is absolute size itself. It is clear, as Galileo long 
ago pointed out in his principle of similitude, that as any body increases 
in size its volume enlarges as the cube of the diameter but its surface 
only as the square. Thus the ratio of surface to volume will progressively 
decrease. In a living organism, where physiological activity is often 
limited by the amount of available surface for interchange of material 
between organism and environment, or between one tissue and another, 
the surface-volume relationship is of much importance and is obviously 
related to changes in shape and structure. 

This shows most simply in the increased elaboration of bodily form 
as size increases. Among algae, for example, the smaller types are 
relatively simple and compact but the larger ones, through branching 
or surface convolution, have a much more elaborate conformation, with 
the result that the ratio of surface to volume is not very different in the 
two. A good example of the same thing is the difference between the 
small, rounded chloroplasts of higher plants and the very much larger 
chromatophores of some of the lower ones, which are elaborately branched 
and dissected. Internal anatomy displays the same tendency, for in the 
sporeling of a fern the vascular cylinder is a solid rod but as the plant 
grows this soon opens out to form a hollow tube. It may later be broken 
up into a ring of bundles or even a series of concentric rings or tubes. 
The radial thickness of each strand thus tends to be approximately the 
same, with the result that the surface of contact between xylem and 
phloem, so important in the physiology of the plant, remains relatively 
constant. Every unit mass of phloem tissue has essentially the same 
"frontier" on the xylem as every other one, and none is limited in its water 

360 Morpho genetic Factors 

supply. Similarly, as roots increase in size, the number of radial arms 
increases, with the result that the xylem-phloem surface remains relatively 

There are many examples of this correlation between absolute size 
and complexity of conformation, a fact which Bower (1930) was the 





Fig. 16-8. The effect of absolute size on structural complexity. Steles of Lycopodium 
scariosum of seven progressively larger sizes showing the increase in complexity that 
accompanies increased size. ( From Bower. ) 

first to bring forcefully to the attention of botanists ( Fig. 16-8 ) . Although 
the advantage of such a correlation is obvious, the problem of how it is 
brought about morphogenetically is not clear. The advantage is not 
effective physiologically until the structures are mature and functioning 
but the pattern is laid down in the meristem. It may be that this is simply 
a case of inheritance of a particularly advantageous developmental pat- 

Various Physical Factors 361 

tern which has arisen through natural selection, as have other embryonic 


The relation between size and form here seems to be too immediate, 
however, to be accounted for in this long-range fashion. An inherited 
pattern ought to be evident in a small and stunted as well as in a large 
individual, but the effect of size seems to be more direct. One is tempted 
to see here another example of the regular spacing of structures to which 
attention has earlier been called (p. 199). In the same way there may be 
maintained a constant ratio between primary xylem and phloem, a ratio 
which originates within the tissues of the primary meristem. This morpho- 
genetic problem, which has an important bearing on the origin of differ- 
entiation and of structure in general, seems especially favorable for 
biochemical and perhaps even mathematical attack. 

Bioelectrical Factors. For many years the possibility that electricity, 
in one way or another, might affect the physiological and developmental 
activities of plants has interested botanists, and there is a very consider- 
able literature on the subject. Unfortunately, this is a field that is theo- 
retically and technically so complex that few workers are qualified to 
obtain dependable results in it, and much of the published work is there- 
fore of doubtful value. The present writer is certainly not competent to 
review it critically. 

Students of tropisms have discussed the possibility that differences in 
electrical potential may be involved in these activities. Went (1932) 
suggested that the polar flow of auxin is an electrophoretic process. 
Clark (1938) raised doubt as to this idea and pointed out that it is pos- 
sible by certain chemical treatments to abolish the polarity of auxin trans- 
port without changing the electrical polarity, protoplasmic streaming, or 
other characters of the system (p. 385). Schrank, however, who has dis- 
cussed the problem in some detail ( 1957 ) , has shown that in the trans- 
verse distribution of auxin in the Avena coleoptile, a transverse electrical 
polarity precedes the movement of auxin, thus tending to support Went's 
theory. The early work on the electrical control of polarity was reviewed 
by Thomas (1939). 

Morphogenetically the most significant result of the work with elec- 
tricity is the evidence that in many organisms there are continuous bio- 
electrical currents and a distribution of electrical potential differences 
so constant that a bioelectrical field is set up in the organism. By the use 
of a very delicate micropotentiometer Burr was able to demonstrate 
the existence of such a field in amphibian eggs (1941). Even in the 
young ovaries of cucurbits such relations between the form of the ovary 
and the bioelectrical pattern could be shown (1944). Burr and Northrop 
( 1939 ) have discussed the general problem of electrodynamic fields in 
living organisms. Lund ( 1931 and others ) carried out a series of studies on 

362 Monpho genetic Factors 

electrical potential differences in trees, as related to the phenomena of 
polarity, dominance, and correlation in them. 

If bioelectrical fields are actually concerned with the specific character 
of development, it would seem possible to change them by applying 
other electrical fields externally and thus perhaps to modify organic 
form. This turns out to be very difficult to accomplish, however. Never- 
theless, Lund ( 1945 ) was able to inhibit completely the growth in length 
of an onion root by passing an electric current upward through it, al- 
though the same current passing downward produced little or no inhi- 

It is tempting to explain all form differences as results of these constant 
bioelectric fields, and they may well be concerned with the control 
of organic development. Such fields are doubtless related to the fine 
structure of protoplasm and the complex pattern of surfaces which this 
presents. Whether the fields are the cause of developmental changes or 
are themselves the result of chemical or physical factors is an important 
theoretical question. The problem of bioelectric factors is too complex for 
easy solution, but it should certainly continue to be attacked vigorously 
by those who are qualified to do so. 

The volume by Lund and his collaborators ( 1947 ) presents a good ac- 
count of the problem of bioelectrical fields and their implications for 
morphogenesis. Rosene has gathered for it a bibliography of 1,406 titles. 


Chemical Factors in General 

Chemical factors are of paramount importance for metabolism and for 
physiology generally, but they also have an important part in the de- 
termination of form and structure. Physical factors-light, water, tem- 
perature, gravity-produce their effects on development chiefly through 
the external environment, but chemical ones operate morphogenetically 
both from inside and outside the organism, and in studying them it is 
necessary to recognize this fact. Nitrogen, for example, by its presence in 
the soil, markedly affects the growth and development of a plant rooted 
there, but it does so because of what happens after it has entered the 
living system of the plant. Here it may be moved from place to place and, 
as a constituent of protoplasm, it affects the course of development in 
many ways. The student of morphogenesis concerns himself, therefore, 
not only with the visible effects of changing amounts of nitrogen in the 
soil but with the history of this element as a part of the organic mecha- 
nism. This is true for other chemical substances, whether taken in from 
the outside or synthesized within the plant. Much of differentiation re- 
sults from differential distribution of substances throughout the plant 


Many effects of these substances are local. In studying them experi- 
mentally it is therefore not enough to examine their effects upon the 
plant as a whole but to discover what happens in particular parts of the 
plant as their concentration in these parts is altered. The most important 
discoveries in the field of growth substances, for example, have been 
made by studying their local effects on a developing root or leaf or ovary. 

The role of chemical substances in development also changes with time, 
for the life of the plant is a life history and this history consists of specific 
progressive changes. These are reflected not only in alterations in struc- 
ture, as between juvenile and adult foliage, but in physiological changes 
that go on within the plant. Some of these are gradual but others are 
more sharply marked. Such is the transition from the vegetative to the 
flowering state (p. 184), in which meristematic activity shifts from one 
region and one type to others that are quite different and where there is 


364 Morphogenetic Factors 

often a radical redistribution of substances in the plant body. The plant's 
life history is composed of such progressive steps. This concept is of 
importance for morphogenesis and perhaps especially for its chemical 

The subject of chemical factors in plant life, and particularly of the 
biochemistry of metabolism, is one of the chief concerns of physiology, 
but the only aspect of this field germane to the science of morphogenesis 
is the somewhat limited portion of it which deals directly with the rela- 
tion of chemical substances to development. The present discussion will 
necessarily treat this subject in nothing more than a very brief and 
general fashion. One chapter will be devoted to the role of the elements 
and the compounds that are primarily significant for their role in nutrition 
rather than in morphogenesis. A second chapter will deal with those 
substances that, even in very minute amounts, have been found to exert 
profound effects on growth and development and are commonly called 
growth substances. 


It has long been known that only a few of the chemical elements are 
essential for plant life. These, in addition to carbon, oxygen, hydrogen, 
and nitrogen, are sulfur, phosphorus, calcium, magnesium, potassium, and 
iron, together with a number of others, notably boron, zinc, copper, 
manganese, cobalt, and a few others which, in very small amounts, are 
essential for the nutrition of most plants and are known as trace elements. 

Several of the elements, or simple compounds of them, have been 
found to have some effect on development and are thus of morphogenetic 
importance, though except for nitrogen this is relatively minor. 

Nitrogen. This element is of outstanding significance in many ways. 
It is an essential constituent of all proteins and thus of protoplasm, and 
from its presence in the nucleoproteins it is concerned in the production 
of new living stuff and thus in all growth and reproduction. 

That nitrogen tends to increase the vegetative growth of plants has 
long been known, but it may also have certain specific effects on their 
structure. Burkholder and McVeigh (1940) grew maize (both inbred 
and hybrid ) with varying applications of nitrogen in sand culture. Where 
nitrogen was abundant, as compared with plants where this was limited, 
meristems were better developed, length and diameter of stem were 
greater, cell size and cell number increased as did the size and number 
of the bundles, there was greater differentiation especially in the phloem, 
and both sieve tubes and vessels increased in diameter (Fig. 17-1). 

Plants growing with little available nitrogen tend to be woodv and to 
have thick cell walls, presumably because much of the carbohydrate is 

Chemical Factors in General 


deposited in these structures, whereas if nitrogen were abundant it 
would be used in protein synthesis. Shields and Mangum (19.54) studied 
the content of total and of amino nitrogen in the leaves of '40 species of 
plants and found it highest in thin, mesophytic herbaceous leaves, next 
in small xeromorphic dicotyledons, and least in monocotyledons with 
much mechanical tissue. 

L3X 38-11 

/ 38-1 1 

0.032 0.16 08 4.0 


Fig. 17-1. Nitrogen and hybrid vigor. Two inbred lines of maize and their hybrid 
grown in sand culture with various amounts of nitrogen. The increase in size with 
added nitrogen is much greater in the hybrid than in either parent. (From'Burk- 
holder and McVeigh. ) 

Nitrogen supply may be related to the differentiation between male 
and female sex organs, for Kocher (1941), studying the dioecious species 
Melandriurh album, found markedly more nitrogen (as percentage of 
dry weight) in leaves of female plants than of male. This difference 
was slight in seedlings, rose until flowering took place, and then disap- 

There may be interaction between nitrogen and other factors. In onion, 

366 Morpho genetic Factors 

for example, Scully, Parker, and Borthwick ( 1945 ) observed that, 
with plants grown at photoperiods longer than necessary for bulb pro- 
duction, variations in nitrogen had little effect, but when the photoperiod 
was close to the critical one for bulbs, bulb development was consider- 
ably greater at low nitrogen levels than at high ones. 

Cohen ( 1953 ) reports the effect of a nitrogen compound, ammonia, on 
the development of the slime molds Dictyostelium and Polysphondylium, 
two members of the Acrasiaceae ( p. 223 ) . Treatment with a low concen- 
tration of ammonia greatly simplifies the morphogenetic expression of 
these forms and reduces the degree of their differentiation. The sorocarps 
produced resemble those of the genera Guttulina and Guttulinopsis. 

Carbohydrate-Nitrogen Ratio. Many other instances of the develop- 
mental effects of nitrogen could be cited. More widespread than these 
changes produced by nitrogen directly, however, are the ones that re- 
sult from the balance between nitrogen and carbohydrates. Horticul- 
turists know that nitrogen stimulates vegetative growth, that weakly 
vegetative plants tend to flower early, and that in vigorously vegetative 
ones flowering is either scanty or does not occur until the amount of 
available nitrogen is reduced. This problem was first studied intensivelv 
by Kraus and Kraybill (1918) in tomatoes. They found that in strongly 
vegetative plants bearing few flowers or fruit the C/N ratio was low. 
Those producing fruit abundantly had a high C/N ratio but had been 
given a good supply of nitrogen at the beginning of their growth. They 
interpret these facts to mean that when nitrogen is abundant all the 
carbohydrate produced by the plant will be used in forming new vege- 
tative tissue. If nitrogen is in short supply, however, carbohydrates will 
tend to accumulate, and when the C/N ratio becomes high enough, the 
development of reproductive structures will be stimulated. This hap- 
pens early in weakly vegetative individuals, but such plants are too small 
to produce a large crop of fruit. Maximum yield results from an abundant 
supply of nitrogen at first and an active production of carbohydrates 
later by the fully grown vegetative structures. This is what normally 
happens under favorable conditions. The important morphogenetic as- 
pect of this hypothesis is that it maintains that the differentiation of 
reproductive organs is dependent on the accumulation of carbohydrate in 
the plant. Until this happens, vegetative structures take priority over 
reproductive ones in the use of available carbohydrate. 

In general, the younger the tissue, the lower is the C/N ratio. From 
a low point in the seedling it increases as foliage develops and as the 
ratio of top to root grows larger until it reaches the point where reproduc- 
tive structures are formed. This would explain why flowering normally 
is deferred until the plant has reached a considerable vegetative develop- 

Chemical Factors in General 367 

The relation of the C/N ratio to flowering is evident in many ways. In 
biennially bearing apple trees, for example, it is high in the bearing 
years and low in the "off" ones. In fruit spurs where the buds are de- 
veloping into flower buds, starch content tends to be high and nitrogen 
low, whereas in barren spurs the opposite is true. Potter and Phillips 
(1927), however, found that flower-bud formation in fruit spurs was 
more closely related to the amount of nitrogen than to any ratio be- 
tween this element and carbohydrates. 

Loomis ( 1932 ) has emphasized the fact that the effect of water on 
development resembles that of nitrogen, both tending to stimulate vege- 
tative growth. He believes that the balance is not so much between carbo- 
hydrate and nitrogen as between the factors that tend to produce growth 
and those that tend to induce differentiation. The former include both 
water and nitrogen together with any other factors, such as temperature, 
that favor the synthesis of new protoplasm. Differentiation, on the other 
hand, requires an excess of available carbohydrate. Why this is so is by 
no means clear, though there may be a selective advantage in a mecha- 
nism which tends to defer the development of reproductive structures 
until a plant has reached the size where it is large enough to produce 
fruit abundantly. 

The carbohydrate-nitrogen ratio is related to other structures than 
reproductive ones, notably the shoot-root ratio. In general, the higher the 
C/N ratio, the larger is the relative size of the root. When nitrogen is 
abundantly available in the soil, the increased growth tends to occur 
in the shoot more than in the root. Hicks ( 1928b ) found that in willow 
cuttings the carbohydrates pass downward and the nitrogen upward so 
that the C/N ratio is higher at the base, where roots develop, than at 
the tip, where shoots are formed. Da vies (1931) also observed that in 
willow cuttings roots develop in regions where nitrogen, as a percentage 
of dry weight, is low and shoots where it is high. These facts obviously 
have a bearing on the problem of polarity in regeneration (p. 119). 

Reid (1924) made stem cuttings from tomato plants that were high in 
C/N ratio and from others that were low and found that the former made 
better roots than the latter. She also observed (1929) that the relative 
development of the root in seedlings of various species was related to 
the proportion of carbohydrate to nitrogen in the seed from which 
they grew. Where this was high the seedlings had stronger roots than 
where it was low. 

Despite the evident relation between the C/N ratio, as observable in 
the chemical composition of the plant, and the processes of flowering and 
root formation, there is now a good deal of evidence that this ratio may 
not be the cause but rather an accompaniment of these activities. In a 
number of instances, such as the soybean (Murneek, 1937) and wheat 

368 Morphogenetic Factors 

(Polster, 1938), floral initiation begins some time before the change 
from low to high C/N ratio takes place. In various plants that have been 
brought to flowering by other means, such as chemical or photoperiodic 
stimulation, the increase in C/N is much less conspicuous than under 
natural conditions. In many cases, also, there is no precise ratio that may 
be counted upon to induce flowering. Both in carbohydrates and com- 
pounds of nitrogen there are many different chemical forms, and it is a 
matter of doubt whether the ratio should regard total, soluble, or easily 
available carbohydrate, on the one hand, and total, protein, or soluble 
nitrogen on the other. 

Despite these criticisms, the general concept that in the living plant 
there is at any particular stage a balance between various chemical 
constituents and that a shift in this balance is related to a change in the 
activities of the plant is an important one. The organism tends to maintain 
a homeostatic equilibrium among its various processes but this is not a 
static condition, for the equilibrium changes progressively as the or- 
ganism develops from one phase to the next. This is so complex that 
the ratio between any two chemical substances will usually give only an 
incomplete picture of it. A study of the C/N ratio, whether this be 
cause or effect of the onset of the reproductive phase, will doubtless 
continue to provide information about this major step in the life history 
of the plant. 

Other Elements. Much of what is known about the morphogenetic 
effects of the other essential elements comes from observation of changes 
in development produced when each is deficient in amount. These changes 
are usually differences from the normal amount of growth or are ab- 
normalities of various sorts, "hunger signs" resulting from insufficient 
nutrition. The literature of physiology and pathology is full of such in- 
stances. Venning (1953) has described the developmental effects of 
deficiencies in sulfur, calcium, nitrogen, potassium, phosphorus, and iron. 
The effects of the various trace elements have been reviewed by several 
workers, among them Brenchley (1947) and Wallace (1950). 

A few of the effects of various elements are of morphogenetic interest. 

Phosphorus is of much importance in physiology and genetics because 
it is a constituent of the nucleic acids. Morphogenetically it is significant 
in relation to mitosis. Phosphate promotes cell division in roots but has 
little effect on cell elongation, whereas nitrate promotes elongation but 
not division. 

Stanfield ( 1944 ) analyzed roots and tops of staminate and pistillate 
plants of Lychnis dioica and found that the staminate had a higher per- 
centage of phosphorus than the latter in both vegetative and early flower- 
ing phases. 

Pierce ( 1937 ) observed that in violet plants grown in nutrient solu- 

Chemical Factors in General 


tions without calcium the chromosomes in root-tip cells were markedly 
smaller than when grown in complete nutrient. When phosphorus was 
in excess, however, the chromosomes were about twice normal size. 
These differences were also reflected in the size of the nucleus. 






I ■ I 









•». / 



E P 

to / 


* / 


o / 


6 / 



c y 


N P 

* / 


V / 




r / 



^ / 



I / 

o" r 


° A"**^ 




■* >*>^ 


/ Ji— - 

A__Jp / 

* t > 






/ \ 






o 6 










Fig. 17-2. Effect of zinc on tomato plants grown in nutrient solution. Curve I, plants 
to which zinc was applied from the start; Curves II and III, plants in which zinc was 
added to the medium at the times indicated by arrows. Dash lines show growth 
without further addition of zinc. ( From Skoog. ) 

There is a good deal of evidence that phosphate fertilizers increase the 
growth of roots as compared with tops, particularly in root crops. 

Calcium is closely connected with the formation of the cell wall. When 
it is deficient in amount the cytoplasm tends to break down, the walls 
to fall apart, and meristems to degenerate. Calcium itself does not enter 

370 M or pho genetic Factors 

into the composition of the wall but produces its effects through changes 
in the cytoplasm ( Sorokin and Sommer, 1929 ) . 

More conspicuous effects of calcium have sometimes been observed. 
Pearsall and Hanby (1925), for example, found that in Potamogeton 

10 r 


g 8 




^ A 

O * 




2 - 

/+>- D.G.S.B. 



ppm BORON 



§4 r 




<£ 2h 



X.J^— <l 

UTAH 10-B 





.01 .05 

ppm BORON 



Fig. 17-3: Relation of boron to cell-wall thickness in three varieties of celery. (From 
A. R. Spurr. ) 

excess calcium made the leaves broader and their cells more numer- 

In field plots abundantly fertilized with potassium, plants of flax, 
ramie, oats, and willow showed marked increase in number and size of 
bast fibers (Tobler, 1929). Hemp was but slightly affected. 

Chemical Factors in General 


Robbins and his colleagues (1929) have observed a shape difference re- 
lated to potassium. Sweet potato plants grown with little potassium have 
longer and more slender roots as compared with "chunkier" ones on 
plants grown where this element is abundant. The difference seems to be 
due to the greater cambial activity in the well-nourished plants as a 
result of increased protein synthesis resulting from an abundance of 

Zinc has indirect morphogenetic importance since it seems to be 
necessary for the maintenance of auxin in an active state (Skoog, 
1940; Fig. 17-2). 

Boron evidently is concerned with the development of the cell wall 
and affects the process of carbohydrate condensation into wall material 
(Spurr, 1957; Fig. 17-3). Reduction in amount of boron produces hyper- 


Fig. 17-4. Effect of cobalt on etiolated pea 
stem segments supplied with auxin alone 
(below) and auxin plus cobalt (above). 
( From Miller. ) 


12 3 4 


trophy and hyperplasia of tissues, especially the cambium, in tomato, 
turnip, and cotton and affects the number and the maturation of the fibers 
(Palser and Mcllrath, 1956). MacVicar and Struckmeyer (p. 320) ob- 
served that day-length altered the boron requirement of soybeans. 

Cobalt has been found by several workers to increase the size of cells, 
particularly in association with sucrose. Miller (1954) reports that in 
etiolated pea stems cobalt salts plus auxin produced only slight elonga- 
tion. Sucrose alone had the same effect, though it considerably increased 
fresh weight ( Fig. 17-4 ) . When sucrose and cobalt were applied together 
to the pea stem there was greatly increased elongation. He believes that 
water uptake and wall growth are separate processes and suggests 
that sucrose tends to increase cell volume but that cobalt promotes the 
ability of the cell to enlarge its surface. 

372 M or pho genetic Factors 

Lyon and Garcia (1944) studied the effects of anions and cations on 
stem anatomy of tomato plants in over 40 nutrient solutions varying in 
the relative proportion of nitrogen, sulfur, and phosphorus, and of calcium, 
potassium, and magnesium. Differences in stem diameter and in relative 
amounts of phloem, xylem, and pith, as well as cellular differences in 
parenchyma of pith and cortex, vessels, fibers, phloem cells, and pericycle 
fibers, were correlated with specific differences in the nutrients. 

Makarova ( 1943 ) examined the relation of various nutrient elements 
to traits described by Maximov as xeromorphic and found that, in gen- 
eral, such traits were intensified by phosphorus, boron, and manganese 
and diminished by potassium and iron. These effects chiefly concern cell 


Sucrose itself may have certain morphogenetic effects. Yates and Curtis 
( 1949 ) found that the root-shoot ratio in orchid seedlings growing in 
nutrient media is not related to the C/N ratio, as in so many cases, but to 
the concentration of sucrose alone, the optimum concentration for root 
growth being markedly higher than for shoot growth. At the best con- 
centration for root growth, shoot growth was reduced. 

In sporelings of Marsilea grown with varying concentrations of glu- 
cose, Allsopp ( 1954 ) observed that in low concentrations the plants 
showed many of the traits of the water forms of amphibious plants, and 
those at higher concentrations resembled land forms. This presumably 
is not owing, however, to the specific morphogenetic effect of glucose 
but to the fact that in the lower concentrations water is more available 
osmotically and conditions therefore to some extent simulate aquatic 
ones (p. 332). 

By using glycogen, soluble starch, or dextrin as a carbon source in- 
stead of sugar, Nickerson and Mankowski (1953) were able to convert 
the normal budding yeast type, in Candida albicans, to a filamentous 
mycelial type. 

Sossountzov (1954) tried various amino acids instead of inorganic 
nitrogen as nitrogen sources in the culture of fern prothallia and found 
that under these conditions they tended to be atypical but that in most 
cases the proportion of filamentous prothallia to cordate ones was con- 
siderably increased. 

Hammett and Walp (1943) studied 10,000 fertilized Fucus eggs, half 
of which were exposed to proline and half were not, and observed that 
proline increased differentiation here (as measured by the production 
and growth of rhizoids) much as it has been shown to do in ani- 
mals. Barghoorn (1942), however, found no definite evidence that pro- 

Chemical Factors in General 373 

line stimulates the differentiation of protoxylem in roots of cotton and 


By adding certain chemical substances to the medium, Tatum, Barratt, 
and Cutter ( 1949 ) produced various nonheritable morphological changes 
in Neurospora and Syncephalastrum which they termed paramorphs. 
Anionic surface-active agents such as Aerosol induced the formation of 
compact colonial paramorphs. Sorbose proved to be the most effective 
paramorphogen. Discovery of the mechanisms by which these substances 
cause the production of such specific forms in these very simple organisms 
might yield important information on the origin of forms at higher levels. 

There is a considerable group of chemical substances, the vitamins, 
which are of great importance in animal nutrition. Many of them are 
synthesized by plants, and some are now known to be essential for the 
growth of certain plants or plant structures. Vitamin B x (thiamin), for 
example, can be shown to be necessary for the growth of excised roots in 
culture, but it is synthesized in the shoots. Nicotinic acid also seems to be 
essential for root growth, but what other vitamins may also be necessary 
for the life of plants is not certainly known. All the vitamins are effective 
in extremely small concentrations. 

The morphogenetic importance of the vitamins for plants seems not 
to be very great. Reid (1941) has evidence that vitamin C affects cell 
size in the meristematic region, and there are a few similar instances. 
Schopfer (1950) has reviewed the problem of the vitamins in morpho- 


Growth Substances 

The chemical substances discussed in the previous chapter are primarily 
important in the nutrition of the plant, and most of them are required 
in relatively large amounts since they contribute to the composition of 
protoplasm. In relatively few cases, however, have they been shown to 
be of any very great morphogenetic significance. 

There is another group of physiologically active substances which are 
not concerned with nutrition and which, although present in most 
cases in very low concentrations, are of great importance for the growth, 
development, and differentiation of plants. They have been given various 
names-hormones, inductors, Wuchsstoffe, activators, evocators, growth 
regulators, and growth substances. They were first studied in animal 
physiology but in recent decades they have assumed major importance 
for an understanding of the development of plants as well. They are 
diverse in character and effect and are the chemical means by which 
many morphogenetic processes are controlled. Among these processes 
are growth, tropisms, many correlations, and the determination and 
differentiation of specific organs and structures in the plant body. Indeed, 
as our knowledge of these substances increases, more and more of the 
activities in development and differentiation are found to be affected by 
them, and, in interaction with the genes, they seem to be the chief 
agents in morphogenesis. It must not be forgotten, however, that they 
are agents merely and that the ultimate control of development lies in 
the factors that determine the concentration, distribution, and interaction 
of these and other chemical and physical mechanisms. Here lies the ulti- 
mate problem. 

Many of these substances, in great diversity, are produced and con- 
trolled by the plant itself and are thus of particular interest in normal 
morphogenesis. It has been shown, however, that a large number of 
synthetic compounds have effects comparable to naturally occurring ones, 
and this has greatly extended the means for experimental attack on de- 


Growth Substances 375 

velopmental problems (Zimmerman and Wilcoxon, 1935; Zimmerman, 
1951b; and an extensive series of other papers ) . 

The literature in this active field is large, and in a brief space little 
more than a very general introduction to it can be made. The student 
may be referred to a number of books, symposium volumes, and reviews, 
among which are the following: Boysen-Jensen (1936); Went and Thi- 
mann (1937); Avery and Johnson (1947); Thimann (1948); Skoog 
(1951); Soding (1952); Audus (1953); Leopold (1955); Wain and 
Wightman (1956); and the Eleventh Symposium of the Society for Ex- 
perimental Biology (1957). 

The term hormone was first used by Starling in 1904 with reference to 
secretin, a substance important in animal physiology. Such hormones are 
regarded as "chemical messengers" since they are produced in one part 
of the body and carried to some other part, where they affect develop- 
ment and various physiological processes. Their discovery marked a great 
advance in an understanding of the chemical control of growth and 

With the demonstration that there are substances in plants which are 
physiologically active in similarly small amounts, the term hormone was 
carried over into plant physiology by Fitting (1909) in relation to a 
substance in orchid pollen which produced swelling of the ovary. Various 
phytohormones are now recognized. This word is not a particularly 
happy one, however, for plants lack the efficient circulatory system of 
animals. Indeed, many of these substances exert their effects in the 
region where they are produced and thus are not "messengers" at all. 
The most important natural plant hormones, the auxins, also differ 
from typical animal ones in being relatively nonspecific and involved in 
a great variety of growth processes rather than in particular ones. Huxley 
( 1935 ) discussed the relationship of these various substances, in animals 
and plants, and suggested a classification and terminology for them. 
Evidently there are many compounds and processes involved, and for 
plants, at least, it seems preferable to use for all such morphogenetically 
active materials the relatively noncommittal term growth substances. This 
will be employed in the present discussion to refer to those substances 
of whatever sort or activity that, in low concentration, are involved in 
the control of growth, development, and differentiation. It should be 
recognized that some of these substances are effective in retarding these 
processes rather than in their stimulation. 

The history of the work on growth substances has been reviewed by a 
number of authors (Leopold, 1955, Chap. 1; Went, 1951a; and Audus, 
1953). Most of it dates from about the beginning of the second quarter 
of the present century and was largely confined, at the start, to workers 
in continental Europe. Among early students of the subject were Boysen- 

376 M or pho genetic Factors 

Jensen, Cholodny, Dolk, Fitting, Kogl, Laibach, Paal, and Went. The 
modern understanding of auxin was established by an important paper 
by the last-named botanist in 1928. 

Growth substances have been shown to be present in many species 
and throughout the plant kingdom, in algae, mosses, ferns, and a great 
variety of seed plants. They have relatively less effect on the growth of 


There are many growth substances but the best known and those ap- 
parently of the greatest general importance are the ones that occur 
naturally in plants and are usually termed the auxins. These were identi- 
fied first in the oat coleoptile but are now known to be present in almost 
all plants and to be concerned in a great variety of developmental 
processes. There has been a good deal of discussion as to how many 
auxins there are, but one of them, identified by Kogl and his colleagues 
(1933) as 3-indoleacetic acid, seems to be of primary importance, and 
all known auxins are related to it chemically. In the present discussion the 
term "auxin" will be generally limited to this and chemically similar 

Our knowledge of indoleacetic acid (IAA) is now very substantial 
but there are a considerable number of other naturally occurring growth 
substances, or substances which have been thought by many to exist in 
plants, which are unrelated to this chemically but are important in various 
ways. About these we know much less. Here belong the wound hormones 
( traumatins ) ; the various flower-forming substances (florigens); the so- 
called root-forming, stem-forming, and leaf -forming substances or calines; 
the gibberellins; the hormones concerned with the determination of sex; 
various substances that amplify or antagonize the effects of auxin; and 
others. About the chemical nature of most of these little is known, and 
the very existence of some of them is in doubt. The conception of "organ- 
forming substances," in the narrow sense, is far from being established 
and has been involved in the general discussion of determination. It is 
clear, however, that there are a great many compounds, some of them 
very specific in their influence, which in minute amounts have important 
effects on one or another of almost all phases of plant development. An 
understanding of their nature and mode of action is proving of much 
importance for the solution of morphogenetic problems. 

It should again be emphasized that most growth substances in plants 
are much less specific in their effects than are the hormones of animals. 
Thus IAA is concerned in cell expansion, cell division, cambial activity, 

Growth Substances 


abscission, parthenocarpy, tumor formation, root production, dominance 
relationships among buds, nastic responses, and tropisms generally. The 
effect in each case must evidently be a function of the specificity of the 
responding system, including other biologically active compounds, rather 
than of a particular evocating substance. 

The developmental effects of growth substances are chiefly of three 
sorts: on growth in general (defined as permanent increase in volume), 
on the correlations of growth, and on development and differentiation of 
specific structures. These may involve very different problems. 


Growth of the plant as a whole may be controlled by various growth 
substances. Van Overbeek (1935) showed that in genetically dwarf races 

12 3 4 


I 2 3 


Fig. 18-1. Yield of diffusible auxin from long shoots (left) and short shoots (right) 
at successive stages in development in Ginkgo. ( From Gunckel and Thirnann. ) 

of maize there was much less auxin than in plants of normal height, indi- 
cating that auxin was associated with growth in stem length. The 
dwarfing gene here effects a more rapid destruction of auxin rather than a 
lesser production of it. Von Abrams (1953) sprayed IAA on etiolated 
plants of dwarf and tall varieties of peas. The dwarf increased 30 per 
cent in height but the tall one was slightly reduced, an instance of the 
frequently observed fact that the effect of auxin is different with different 
material. In a number of other cases the dwarfing of plants has been 
shown to be related to a deficiency of auxin. 

Auxin particularly affects growth in length, especially of the stem, and 
in many cases has been found to be the factor that determines length 
growth of the internodes. Gunckel and Thirnann (1949) compared the 

378 Morpho genetic Factors 

amount of auxin in developing short shoots and long shoots of Ginkgo and 
found that elongation of the latter was associated with a sharp rise in 
auxin content (Fig. 18-1). 

Auxin is also involved in the growth of fruits (Fig. 18-2), though not 
all the steps in the process are yet well understood (Murneek, 1954; 
Lund, 1956). There are reviews by Nitsch (1952) and Luckwill (1957). 
An increase in auxin content often occurs at pollination, again at fertiliza- 

Fig. 18-2. Centers of origin of auxin in the development of a fruit. 1, from mother 
plant; 2, from pollen; 3, from developing ovules. (From Nitsch.) 

tion, and usually during endosperm growth and the development of the 
seeds. This increase of auxin in the ovary is often followed by its increase 
in the pedicel and adjacent regions. 

That fruit development is stimulated by growth substances is made 
clear by the occurrence of parthenocarpy, or growth of fruits without 
fertilization and seed development (Fig. 18-3). Gustafson (1936) in- 
duced parthenocarpy in cucurbits by applying synthetic growth sub- 
stances to the pistil, and this has now been done frequently with other 

Growth Substances 379 

plants. He also showed that fruits which are naturally parthenocarpic 
have a much higher concentration of auxin than do seed-bearing ones. 
In some cases parthenocarpy does not require application of a growth 
substance to the unfertilized ovary but simply the presence in the air of 
the greenhouse of vapor of a specific substance. Induction of partheno- 
carpy is by no means universally possible and has been found more 
frequently in members of the Solanaceae and Cucurbitaceae than in 
other families. The subject has been reviewed by Vazart (1955). 

Fig. 18-3. Parthenocarpy. Left, normally pollinated tomato fruit with many seeds. 
Right, parthenocarpic fruit produced by treatment with synthetic growth substance. 
( Courtesy Boyce Thompson Institute. ) 

Davidson (1950) found that several marine algae were stimulated in 
their growth by synthetic substances in the sea water (Fig. 18-4). Al- 
though auxin seems rarely to be concerned with the growth of fungi, 
Fraser (1953) reports that indoleacetic acid stimulates growth in the 
common mushroom. 

Cell Enlargement. The primary effect of auxin on plant growth seems 
to be its promotion of increase in cell size, especially in the stem and 
in its longitudinal dimension. In phototropism the side of the axis away 
from the light grows faster than the lighted one, and Went ( 1928 ) and 
others showed that auxin is more abundant on this shaded side. Avery 

380 M or pho genetic Factors 

and Burkholder (p. 30) found that the more rapid elongation here was 
the result of increase in cell length rather than in cell number. 

Most tropisms have now been shown to be due to greater growth on the 
convex side because of longer cells there (Fig. 18-5). In both photo- 
tropism and geotropism, however, the role of auxin is different in root and 
shoot. If a young plant, for example, is placed in a horizontal position, 
auxin can be shown to accumulate on the lower side, though how this 
occurs is not well understood. In the stem, this results in a bending 
upward of the axis. The young primary root, however, will bend down- 





Fig. 18-4. Effect of different concentrations of indoleacetic acid on the growth of 
Fucus, as measured by length of thallus and length and number of holdfasts, in com- 
parison with untreated controls. ( From Davidson. ) 

ward, the cells elongating faster on its upper side. Evidently the same 
concentration which stimulates cell elongation in the stem checks it in 
the root, another example of the difference in the effect of auxin under 
different conditions. Geiger-Huber and Huber (1945) found that by 
continually decapitating a root its auxin content was markedly reduced 
and that it now bent upward instead of downward. 

How much of the difference in cell size among the various tissues of 
the plant is due to auxin it is hard to say but, directly or indirectly, this 
substance is probably involved in most of the differences in degree of 

Growth Substances 


cell enlargement that occur between the meristematic condition and ma- 
turity. Gibberellin acts in a similar manner. 

Cell Division. Growth substances are also associated often with cell 
division and with meristematic activity generally. Buds, especially as they 
begin to expand, are rich in auxin. Cambial activity is closely related to 
the presence of auxin (Snow, 1935; Fig. 18-6; Soding, 1936), and the 
progressive awakening of the cambium in the spring from the stem tips 
downward is accompanied by a progressive increase in auxin concen- 
tration (Avery, Burkholder, and Creighton, 1937b; Fig. 18-7; Gouwentak 
and Maas, 1940). 

Fig. 18-5. Cell size in geotropic curva- 
ture. Lengthwise section through the con- 
vex side (upper) and the concave one 
(lower) of a root of Zea bending down 
geotropically. Bending is produced by 
the greater elongation of the cells on the 
upper side resulting from a relatively 
weak concentration of auxin there. ( After 
MacDougal. ) 

Various synthetic growth substances have been found to stimulate 
cell division even in tissues which are mature or nearly so. Such a sub- 
stance, applied to a young stem, will often produce there a callus-like 
mass of cells, so that it is easy to see why growth substances are associated 
with the formation of galls and tumors. This ability to produce cell 
division in older tissues has often been used for the determination of 
chromosome numbers in their cells, since it is now possible to observe 
mitotic figures there. 

The various wound hormones produce their effect by increasing the 
division of cells below wound surfaces. 


Morphogenetic Factors 

Cells which 
normally ' 
form fibres 


Primary , 



Fig. 18-6. Stimulation of cambial activity by auxin. Above, two bundles from control 
plant, decapitated but no auxin applied. Below, a similar region from a plant in which 
auxin had been applied to the decapitated tip. Cambial growth is markedly stimu- 
lated by the auxin. ( From Snow. ) 

Growth Substances 















Fig. 18-7. Concentration gradient of growth hormone (in terms of Avena curvature) 
at various levels in a twig of Aesculus on May 16, when the growth of the current 
year ( above arrow ) was just completed. ( From Avery, Burkholder, and Creighton. ) 


Of much significance for morphogenesis is the role of growth sub- 
stances in the correlation of growth (Thimann, 1954/?). Here is involved 
the whole problem of organized development. These substances seem to 
be the agents by which many such correlations are achieved, but one 
should hesitate to call them "growth regulators" since the actual regula- 
tion must go further back, to the mechanism that controls their distribu- 
tion and local concentration and binds the parts and processes into an 

Differential Movement. An important fact about auxin is that in living 

384 Morphogenetic Factors 

cells it does not move in all directions but has a definitely polar flow, 
going from the more apical regions of the plant downward ( p. 141 ) . This 
was first observed in the oat coleoptile, where auxin is formed at the tip 
and then passes basipetally, regardless of how the axis is oriented. If a 
portion of the coleoptile is cut off and inverted and auxin is then applied 
to the end now uppermost, it will not move downward, though auxin 
placed at the other end will move up, in the original polar direction. 
Such polar transport is also shown in the stem axis, in which auxin 
moves down through the phloem. Inverted stems, after some time and 
apparently after the development of some new vascular tissues, may gain 
the ability to transport auxin in the opposite direction from the original 
one (Went, 1941). 

Auxin transport may not invariably be polar, for Jacobs (1954) has 
reported that, if relatively weak concentrations are used instead of the 
strong ones commonly employed in experiments, there is a good deal of 
upward translocation. In a young internode of Coleus he found that 
about one-third as much auxin moved upward as downward. He also ob- 
served that in young bean hypocotyls, although auxin transport was al- 
ways basipetal, the ability to transport it at all was lacking in very young 
seedlings but increased as they grew older. It was greatest in the upper 
portion of the hypocotyl and decreased toward the base. Oserkowsky 
( 1942 ) concluded that where auxin moves only basipetally it is carried 
in living cells but that transport in both directions may result from 
diffusion through dead cells or cell walls. Leopold and Guernsey (1953fl) 
showed that in Coleus the flow was clearly basipetal in the shoot and 
in the opposite direction in the root ( Fig. 18-8 ) and that flowering stems 
transport auxin in both directions. Haupt (1956), however, reports that 
polar transport of auxin is as clear in the floral structures he studied as it 
is in vegetative shoots. Niedergang-Kamien and Skoog ( 1956 ) were able 
to reduce or inhibit polar flow by triiodobenzoic acid (Fig. 18-9) and 
suggest that the reported effect of this substance on growth correlations 
is due to this fact. 

Nevertheless, polar transport of auxin seems to be a general phenome- 
non and is obviously of much morphogenetic importance since it under- 
lies the marked structural differentiation between the two ends of the 
plant axis. It is probably involved in the polar character of regeneration 
(p. 119) and in many other developmental events. 

Auxin flow may show a transverse as well as a longitudinal polarity, 
notably in geotropic movements. The accumulation of auxin in the 
lower half of a horizontally placed axis, although doubtless a response to 
the stimulus of gravity, is not simply a downward diffusion but is made 
possible by differential and unidirectional changes in the permeability 
of the cells. Such transverse polarities have been emphasized by de 

Growth Substances 


Haan (1936), who reexamined the fact earlier noted by Noll that when 
primary roots are bent, lateral roots grow chiefly from their convex sides 
(Fig. 18-10). He observed that in such cases the root primordia, push- 
ing out through the cortex, bend toward what is at that point the convex 
side of the root (Fig. 18-11), and he interprets this as the result of a 
transverse polar gradient in a root-forming hormone. Such transverse 
reactions of plants are extensively discussed by Borgstrom (1939), who 
shows their importance in many structural and physiological characters 
of plants. 


2 0.4 06 08 1.0 

Fig. 18-8. Relative amount of basipetal Fig. 18-9. Effect of 2,3,5-triiodobenzoic 
transport of auxin in various regions of acid on polar transport of auxin. Graph 
the vegetative and of the flowering axis showing reduction in amount of auxin 

of Coleus ( From Leopold and Guernsey. ) 

transported basipetally in sunflower epi- 
cotyl cylinders which had been pre- 
treated for 2/2 hours with various concen- 
trations of TIBA. (From Niedergang- 
Kamien and Skoog. ) 

No satisfactory explanation of polar auxin transport is available (p. 361). 
The suggestion of Went (1932) that electrophoretic diffusion might ac- 
count for it has been shown by Clark ( 1938 ) to be doubtful, since the 
polar transport of auxin can be abolished by the application of sodium 
glycocholate without producing any change in the electrical polarity. 
Schrank, however, in a series of papers ( 1957 and others ) presents evi- 
dence that electrical polarity is the essential basis for the polar transport 
of auxin. The question of auxin transport is part of the more general one 
as to why many substances move about the plant in certain directions 
more than in others. Auxin is doubtless important in the development of 
growth patterns, and an understanding of the mechanism of its differen- 
tial movement would clear up many morphogenetic problems. 

386 Morphogenetic Factors 

Dominance, Inhibition, and Stimulation. Whatever the mechanism for 
the differential distribution of growth substances may be, there is no ques- 
tion that it has much to do with the effect of one part on another and 
thus with growth correlation in general (Chap. 4). 

Meristematic regions are commonly centers of auxin production, and 
from them this substance moves to others where it may either stimulate 
or inhibit growth. The most familiar example of this is the so-called 
dominance of apical buds. It is well known that in most cases, if such a 

Fig. 18-10. Curved primary root of a lu- 
pine seedling, showing lateral roots grow- 
ing from convex sides only. (After Noll.) 

Fig. 18-11. Cross section of a curved root 
like that in Fig. 18-10, showing that the 
lateral root, in its course through the 
cortex, bends toward the convex side, X. 
(After Noll.) 

bud is present, the buds below it will not grow but that they will do so 
if the apical bud is removed. Thimann and Skoog (1934) showed that if 
this apical bud is cut off and paste containing indoleacetic acid placed 
on the stump the lower buds will not develop, though in the controls, 
with paste alone, they do so. Such a result has been confirmed many 
times and has led to the widely held view that auxin inhibits growth of 
the lower buds and thus produces the dominance of the apical one. 

Snow, in a series of papers (1937, 1940, and others), calls attention to 
some facts that are not easily explained on this hypothesis. He finds that 

Growth Substances 387 

buds farthest from the apex are most inhibited, which does not seem 
logical if auxin from the terminal bud is the inhibitor. Furthermore, 
inhibition may sometimes pass upward, although auxin typically passes 
downward. Snow believes that auxin is either changed into an inhibiting 
substance (for which Libbert, 1955, has evidence) or stimulates the pro- 
duction of one and that the greater the distance from the apex, the 
greater the amount of the inhibitor. Such a substance is not polar in 
movement and can thus pass into lateral buds as auxin presumably can- 

Champagnat ( 1955 and others ) has studied especially the effects upon 
bud growth of the cotyledons or leaves that subtend them and comes to 
the conclusion that both inhibitory and stimulatory substances are in- 

Thimann (1954£>) is inclined to think that auxin itself is the major 
inhibiting influence and that the various apparent objections to this 
hypothesis may be met by assuming that auxin may sometimes move in 
an apolar direction and that its effects may be different under different 
circumstances. He does not rule out the possibility of the existence of 
other and specifically inhibiting substances. 

Meanwhile other workers have emphasized the importance of nu- 
trition in apical dominance. Van Overbeek (1938) observed that after 
removal of the terminal bud the auxin content of the stem decreases. He 
believes that in some way auxin blocks the passage of nutrients to the 
lateral buds, which have only a poor vascular connection with the main 
cylinder. When the amount of auxin is reduced a better connection is 
established, nutrients enter the buds, and thev begin to develop. 

The nutrient theory of apical dominance is strongly supported by 
Gregory and Veale (1957). They suggest that the degree of dominance 
is proportional to the supply of available carbohydrate and nitrogen and 
that competition among the various buds explains the difference in relative 
bud growth. Auxin is concerned in this competition since a high auxin 
content prevents the formation of the vascular connections between bud 
and vascular cvlinder. 

The underlying mechanism in such an apparently simple phenomenon 
as bud dominance is still by no means clear. The general fact of domi- 
nance is established, however, and helps toward an understanding of 
some of the differences among plants in their bodily patterns, for what- 
ever determines bud growth determines the shape of the plant. There are 
usually a great number of potential growing points on an axis but most 
of them do not develop into shoots, apparentlv because of inhibitory' 
action mediated in one way or another by growth substances. Plants differ 
in the degree of this inhibition. For example, in Aster novaeangliae there 
is one main stem with only a few floral branches at the top. Aster 

388 Morpho genetic Factors 

multiflorus, on the other hand, is much branched and bushy. Delisle 
( 1937 ) found that in the former species the concentration of auxin at the 
tip was high (Fig. 18-12). In the latter it was much lower, suggesting 
that the copious branching resulted from weak inhibition. After re- 
moval of the apex in the young plant, A. novaeangliae became much 
branched, and application of auxin to the tip of A. multiflorus resulted 
in the growth of a main stem with relatively small branches. Thus an 
important taxonomic character is related to the amount of auxin present 
in the plant. 

Many of the effects of growth substance in correlation are stimulatory 
rather than inhibitory. There is a close relation, for example, between the 
presence of growing seeds and the development of a fruit, the seeds pro- 
ducing the auxin necessary for fruit growth. In some cases, as has been 


5 15 




30 40 SO 




Fig. 18-12. Effect of auxin on inhibition of branching in Aster. Auxin concentrations 
at successive distances from apex in A. novaeangliae, an unbranched species. The 
amount of auxin in A. multiflorus, which is more branched., is considerably less. ( From 
Delisle. ) 

mentioned, synthetic growth substances may be substituted for the 
natural auxin and seedless fruits produced. Growth of the receptacle 
may also be related to seed production. When Nitsch (1950) removed 
the growing achenes from a strawberry, the fleshy portion stopped grow- 
ing, but if synthetic growth substance replaced the achenes, the normal 
development of the strawberry was resumed (Figs. 18-13, 18-14). If 
only some of the achenes were removed, the weight of the mature ripe 
fruit was proportional to the number of achenes remaining. Nitsch found 
that the achenes contained a large amount of auxin but that this was 
absent from the receptacle. 

Galston ( 1948 ) has described an example of competitive correlation 
in asparagus. Root primordia, stimulated by auxin, are formed during 

Growth Substances 


a period of minimum stem growth, but once they are laid down, stem 
growth again accelerates. Compensatory correlations (p. 98) of various 
sorts also have their basis in auxin action. Removal of the root tip almost 
always stimulates growth of lateral roots below it, somewhat as in 

Fig. 18-13. Relation of presence of achenes to growth of strawberry. All achenes 
have been removed early in development except three vertical rows ( left ) or three 
horizontal ones (right). Growth of the receptacle is limited to the region in contact 
with the achenes. ( From Nitsch. ) 

Mar., 1950] 








O o- — o o o 






Fig. 18-14. Auxin and growth of the straw- 
berry. Curve 1, control. Curve 2, growth of 
fruit from which all achenes were removed 
and auxin in lanolin paste substituted. Curve 
3, like 2 but without the auxin. (From 
Nitsch. ) 

Fig. 18-15. Effect of 2,4-D on 
leaf shape. Below, leaf of un- 
treated plant of Pisum sativum. 
Above, plant treated with 2,4-D 
paste. ( From Wenck. ) 

390 Morpho genetic Factors 

apical-bud dominance. Active growth of lateral roots may also inhibit 
apical growth (Street and Roberts, 1952). 

Correlations of position may be the result of auxin activity in many 
cases. Thus where an upright terminal shoot of a coniferous tree is re- 
moved, one of the lateral branches will swing up from its nearly horizontal 
position to a vertical one and replace the lost leader, evidently in re- 
sponse to the absence of auxin previously produced by the apical bud. In 
woody plants the orientation of branches with reference to the main 
axis and to gravity also seems to be due to auxin action since it is regu- 
lated by the production of reaction wood ( p. 356 ) , which seems to result 
from the presence of auxin. The precise amount of auxin (and thus of 
reaction wood formed) determines the angle that a given branch 
will assume and thus the branching pattern and form of the whole tree. 

The form of individual organs, ultimately the result of dimensional 
correlations, may be affected by growth substances, particularly in leaves 
(von Denffer, 1951; Linser and others, 1955; Wenck, 1952; Fig. 18-15). 

Such correlating activities are doubtless present in the lower plants as 
well. Moner (1954) describes the action of a substance, as yet unidenti- 
fied, which is concerned with the development of the precisely formed 
colonies of the alga Pediastrum. 

Much evidence is therefore available that the correlated and integrated 
character of the plant, whatever its final cause may be, is the immediate 
result of specific amounts of growth substances at specific places and 
times. What controls this precise production and distribution of these 
substances is a more difficult problem. 



The effect of growth substances on the specific form and structure of 
plants has attracted more attention than any other of their morpho- 
genetic activities. 

Tropisms and other auxin-mediated orientations of plant parts to cer- 
tain factors in the environment, notably gravity and light, account for 
many features of external form, though the familiar patterns of plant 
growth are produced by interaction between these tropisms and certain 
specific inner factors. Sometimes a simple tropism may produce such 
a profound change in plant form as to be significant morphogenetically. 
In "lazy" maize (van Overbeek, 1936), for example, the stalk grows 
flat on the ground, not through mechanical weakness but because of the 
characteristic distribution of auxin in it. Tropisms are primarily the 
concern of physiology, however, and there is no room here to consider 
the extensive literature in which they are discussed. 

Growth Substances 391 

The arrangement of structures in a radially symmetrical pattern doubt- 
less involves differential distribution of auxin, but little is known 
about this. When such radially symmetrical structures are placed in a 
horizontal position they often tend to become dorsiventral, as in certain 
flowers and stems. Since most such direct effects of gravity seem to be 
produced by differential distribution of auxin, as in the well-known 
cases of geotropism, auxin presumably is involved in structural dorsi- 
ventrality as well. Its role here in young liverwort plants has been 
described by Kohlenbach ( 1957 ) . 

Of wider morphogenetic interest are the effects of growth substances 
on the determination of specific structures. From his study of flower 
production Sachs (1882) suggested that organ-forming substances are 
operative in plants, especially in the determination of flowers and roots. 
The growth of knowledge of morphogenetically effective substances 
revived interest in this hypothesis, and it has stimulated a wide range of 
experiments which have thrown much light on the mechanisms of de- 
velopment. Organ-forming substances of many types have been postu- 
lated in the formation of roots, stems, leaves, flowers, galls, sexual struc- 
tures, and others. Just how such substances produce their effects is not 
known, and the actual existence of some of them is not yet proved, but 
the theoretical, and also the practical, importance of these problems is 
great. The most obvious way to account for development is to postulate 
the operation of a series of such substances. The difficulties of this con- 
ception, however, are obvious, for a very large number of them would be 
necessary. The tendency today is to assume the activity of a relatively 
small number and to explain the variety and specificity of their effects 
through their interactions and in other ways. 

Root Formation. Van der Lek (1925) observed that a piece of stem 
(used as a cutting), on which there was a bud or young leaf, formed 
roots at its base, whereas a naked stem piece did so much less readily or 
not at all. This suggested that there was a substance, formed in buds 
and leaves, that moved downward and stimulated root production. Went 
( 1929 ) demonstrated this by showing that an extract from the leaves of 
Acalypha, applied to the apex of a cutting, promoted root formation at 
its base. Several workers soon discovered that auxin and various natural 
and synthetic substances also have this effect and that cuttings could 
be made to root by the application of such substances. 

In 1935 Laibach and Fischnich described a technique by which indole- 
acetic acid in lanolin paste applied to a stem would promote root for- 
mation. In the same year Zimmerman and Wilcoxon ( 1935 ) reported 
that several synthetic substances such as indolebutyric, indolepropionic, 
phenylacetic, and naphthaleneacetic acids had this effect and could 
be used in horticultural practice to hasten the rooting of cuttings when 

392 M or pho genetic Factors 

this was otherwise slow or difficult. The root-forming activity of 
2,4-dichlorophyenoxyacetic acid, 2,4-D (Zimmerman and Hitchcock, 1942), 
was especially conspicuous. By their application in paste or other means 
under favorable conditions root initials may be produced almost any- 
where on the plant. These substances are not equally active, and some of 
them which have less value in root formation show other morphogenetic 
effects. "Root-forming hormones" are now familiar aids in plant propa- 
gation (Fig. 18-16). These substances are chiefly effective in the produc- 
tion of root primordia and in most cases they actually check the later 
growth of the roots. Because of their great theoretical and practical in- 
terest, much work has been done on the root-forming effects of growth 

Fig. 18-16. Effect of "root hormones" on cuttings of holly. At left, controls. At right, 
plants treated with mixture of indolebutyric and naphthaleneacetic acid. (Courtesy 
Boyce Thompson Institute. ) 

substances. There are a number of general reviews of this work, among 
them Pearse, 1939; Thimann and Behnke, 1947; and Avery and Johnson, 

The movement of synthetic substances, like that of natural auxin, is 
polar except when their concentration is high. If applied at the apical 
end of a cutting, they tend to pass downward and to stimulate root for- 
mation at the base. If applied basally, they form roots there. The experi- 
ments of Czaja and others have previously been described (p. 124) in 
which regeneration in pieces of root is also polar in character, shoot buds 
forming on the upper end and root primordia on the lower. This is evi- 
dently owing to the accumulation of auxin at the lower (distal) end and 
its consequent relatively low concentration at the upper one, a low con- 
centration being related to shoot growth and a higher one to root growth 

Growth Substances 


(Fig. 18-17). When slices were repeatedly trimmed off from the lower 
surface, shoot primordia began to appear there, presumably because of 
the removal of auxin and the consequent reduction in its concentration. 
When concentrations of growth substances are much higher than in 
natural conditions, they tend to have a local effect and to produce a 
downward bending of the leaves (hyponasty ) and the formation of callus. 
Upon the latter, root primordia often appear. Growth substances ap- 
plied to the soil may be absorbed, carried up in the transpiration stream, 
and affect the structure of the growing plant. 







Fig. 18-17. Auxin and regeneration. In a root segment of Taraxacum, placed hori- 
zontally (2), regeneration is normally polar, shoots developing at the proximal end 
and roots at the distal one (see Fig. 6-4). When the amount of auxin is increased at 
the proximal end, roots are produced there ( 1 ) . When it is decreased at the distal 
end, shoots develop there ( 3 ) . ( From Warmke and Warmke. ) 

It is recognized that growth substances are not the only factors con- 
cerned in root formation. A supply of sugar is necessary. Indeed, the 
stimulating effect of leaves on root formation may be due in part to their 
production of nutrients. In grafting experiments between rooting and 
nonrooting varieties of Hibiscus, van Overbeek and Gregory (1945) 
found that something formed in the leaves, in combination with auxin, 
is required for root growth, and van Overbeek, Gordon, and Gregory 
(1946) showed that this is not a hormone since it can be replaced by 
sucrose or nitrogenous substances. The importance of a high carbohy- 
drate-nitrogen ratio in root determination has already been discussed 

394 Morpho genetic Factors 

(p. 367). Torrey (1950) observed that pea root tips transferred directly 
to a culture medium provided with IAA produce lateral roots at once 
but that tips transferred after growing a week in culture do not do so for 
some time. A substance (not auxin) stimulating lateral root formation 
seems to originate in the lower part of the root and moves upward, 
producing laterals in acropetal succession. 

It has been observed that the rooting response is often altered (gen- 
erally increased ) when two different substances, such as indoleacetic acid 
and naphthaleneacetic acid, are combined. Went (1939) applied dif- 
ferent substances successively rather than in mixtures and found that 
cuttings of etiolated pea seedlings, which do not root after treatment 
with auxin alone, will do so if phenylacetic acid is first used, though 
this substance by itself is ineffective. He believes root formation results 
from the interaction of two factors and suggests that phenylacetic acid 
may act to mobilize or activate a specific root-forming factor, rhizocaline 
(Bouillenne and Went, 1933; Bouillenne, 1950; Libbert, 1956). The 
question of the existence of such a specific factor has been studied by a 
number of workers. Evidence for it is largely indirect, and rhizocaline 
has not been isolated; but auxin is evidently not the only factor operative 
in the initiating of root primordia. 

It must not be concluded that growth substances can produce roots 
anywhere on the plant. They are often formed, to be sure, in unusual 
places, as along the surface of the stem. Even in such instances, how- 
ever, the initiation of root primordia does not take place anywhere and 
indiscriminately but only in certain cells or at certain zones that are 
potentially capable of forming them. At such points roots may be formed 
under the stimulus of other factors such as ethylene, carbon monoxide, 
wounding, or abnormal nutrition. The nature and location of such re- 
gions are variable and depend on the general growth pattern of the plant 
treated. There is a difference, for example, between monocotyledons and 
dicotyledons as to rooting response. Treatment with growth substances 
is one of the methods by which knowledge may be gained as to the 
potentialities of various cells and tissues, not only for root formation but 
for other developmental activities. 

Auxin and Rhizoids. Auxin is present in the coenocytic alga Bryopsis 
and is most abundant in that part of the plant where rhizoids are com- 
monest. Jacobs (1951) finds that an application of indoleacetic acid 
stimulates rhizoid formation in the region where these are least abundant. 
He regards this as analogous to the effect of auxin on root initiation in 
higher plants. 

Leaf Formation. Attempts have been made to find substances which 
might be involved specifically in the development and growth of the 
leaf blade. In some cases leaf growth is dependent on the presence of a 

Growth Substances 


factor coming from the cotyledons. Pilet (1952) observed that Semper- 
vivum leaves parasitized by Endophyllum sempervdvi contain much 
more auxin than normal ones and, presumably for this reason, are accel- 
erated in their development. The effectiveness of adenine in leaf formation 
has been observed by various workers (D. Bonner and Haagen-Smit, 
1939). Auxin does not seem to promote the growth of the blade as a whole, 


Fig. 18-18. Section through leaf blades 
of Kalanchoe Blossfeldiana. At left, from 
a plant grown under long days. At right, 
from one grown under short days. The 
difference in thickness is entirely due to 
cell size. ( From Harder and von Witsch. ) 

however, although when embryonic leaves are treated with it, various 
changes may be produced (Laibach and Fischnich, 1936; Zimmerman, 
1951k; Applegate and Hamner, 1957). These are probably to be looked 
upon as injuries rather than formative effects. Wenck (1952) has studied 
the stimulatory and inhibitory effects of auxin and of various auxin 
antagonists on leaf growth in a number of species. 

396 Morpho genetic Factors 

Went ( 1938 ) here postulated a phyllocaline, analogous to rhizocaline. 
Later ( 1951/?) he extended his concept of the calines more fully into the 
details of leaf form and structure. He calls attention to the importance of 
adenine for mesophyll growth and shows that vein tissue, on the other 
hand, can be increased by auxin without affecting mesophyll develop- 
ment. There are thus two morphogenetic tendencies in leaf development: 
one toward the formation of veins and induced by auxin and the other 
of mesophyll, induced by adenine. Whatever factor induces the former 
( as well as the stem and petiole ) may be called a caulocaline and the lat- 
ter a phyllocaline, whatever their chemical nature may turn out to be. 
Leaf shape is affected by the balance between the two. Leaves with a 
dominance of phyllocaline will tend to be palmate for they will have an 
excess of mesophyll, whereas those with more caulocaline will tend to be 
pinnate or parallel-veined, since they will have relatively more vein 
tissue. One may question, however, whether the problem of organic form 
can be solved quite as simply as this. 

Harder ( 1948 ) observed that in certain succulents, such as Kalanchoe 
Blossfeldiarui, variations in leaf shape and structure depend on the day- 
length. Plants grown under short days have short, apetiolate, and markedly 
succulent leaves (Fig. 18-18). A single leaf subjected to short days will 
transmit this "short-day shape" to the immature leaves above it which are 
developing under a long day-length. This effect Harder and von Witsch 
(1940b) attribute to a growth substance they call metapkisin, which is 
not identical with either auxin or florigen. 

Stem Formation. It has proved difficult to demonstrate any substances 
specifically related to the growth of the stem. Went (1938) decapitated 
pea seedlings and measured the length of the secondary lateral branches. 
He gives evidence that stem growth here depended on the roots, not on 
the cotyledons, and suggests that it was due to caulocaline in conjunction 
with auxin. In later experiments neither auxin from the apex nor water 
supply from the root appeared to control stem growth, and Went again 
attributed this to caulocaline coming from the stem base and the root 

Flower Formation. The existence of flower-forming substances has a 
firmer foundation. They were postulated by Sachs, who held them re- 
sponsible for changing a plant from a vegetative to a reproductive state. 
The demonstration that such a change could be induced by altering the 
carbohydrate-nitrogen ratio (p. 366) and by photoperiodism (p. 315) cast 
doubt upon this idea. More recent work, however, has come to its support. 
Kuijper and Wiersum (1936), for example, found that if a soybean plant 
is brought to a flowering state by exposure to short day-lengths and 
another kept flowerless by long days a shoot of the former grafted into the 
latter causes the flowerless plant to form flowers abundantly. Hamner 

Growth Substances 397 

and Bonner (1938) reported that this effect could be produced in 
Xanthium through a barrier of lens paper without actual union of tissues. 
Withrow and Withrow ( 1943), however, failed to confirm this and showed 
that where transmission of the flowering stimulus had occurred there 
had been a slight fusion between cells which had grown through the lens 
paper. Nevertheless, in grafting experiments like these, a substance evi- 
dently passes from scion to stock across the graft union and induces 
flowering. To such a substance the name florigen has been applied. 

Other experiments in photoperiodism also suggest the operation of such 
a flower-forming substance. Cajlachjan (1938), by localizing the recep- 
tion of the photoperiodic stimulus, showed that this was received by the 
leaves but was effective in flower induction at growing points considerably 
distant and had therefore apparently passed, as a specific substance, down 
the petiole, along the stem, and into a lateral branch. Borthwick, Parker, 
and Heinze (1941) with soybeans found similar results. Harder (1948) 
observed that in Kalanchoe this substance passes directly down the stem 
from the site of induction but does not cross it, so that one side of the 
plant flowers but the opposite one does not. 

Unlike auxin, the movement of which is usually polar, the flower- 
inducing agent seems able to travel in any direction in the plant. Since 
local applications of cold, heat, or narcotics reduce or inhibit the trans- 
port of such substances from centers of production to those of action, 
it seems probable that living tissue is involved, a conclusion supported 
by girdling experiments of Galston ( 1949 ) and others, who showed that 
the floral stimulus cannot pass across a water gap. 

From the leaf of a unifoliate species of Streptocarpus that was ready to 
flower, Oehlkers (1955) made a series of cuttings. Those from the base 
of the leaf produced flowers at once, those from a little farther up pro- 
duced them soon, and cuttings from near the tip formed only vegetative 
shoots. Oehlkers believes this was because of the differential distribution 
of a flower-forming substance. 

Genie differences may also be involved. One variety of Hyoscyamus 
niger (henbane) is biennial and does not flower until its second year. 
Another variety is annual. If from the annual form a flowering scion, or 
leaf from one, is grafted into a plant of the biennial race during its first 
year, the latter will flower in this season (p. 264). It was also shown that 
the substance here concerned was not limited in action to this species for 
a flowering scion of tobacco or petunia (genera in the same family) has 
the same effect on biennial Hyoscyamus. 

This nonspecificity of the flowering stimulus is also evident in certain 
host-parasite relationships. Orobanche minor (one of the broomrapes), 
when parasitizing clover, flowers only when the host plant flowers ( Holds- 
worth and Nutman, 1947). Cuscuta Gronovii (dodder) flowers only 

398 Morpho genetic Factors 

in a long day if it is parasitic on the long-day plant Calendula, but in a 
short day if it is parasitic on the short-day plant Cosmos (von Denffer, 


These various lines of evidence, about which a great body of facts has 
now been gathered, suggest that there are one or more specifically flower- 
forming substances. None of these florigens has yet been isolated nor is 
there any knowledge as to their chemical nature. Some substance that 
under certain conditions stimulates flowering is certainly able to pass 
across a graft union and thus seems hormonal in character. This sub- 
stance is evidently closely involved with photoperiodism, though what it 
does depends to a great extent on the amount of auxin or other growth 
substances present. Thus J. Bonner and Thurlow (1949) completely pre- 
vented flowering in the short-day plant Xanthium canadense, grown 
under short days, by spraying it with auxin, and leaves thus treated did 
not transmit the flowering stimulus by grafting. Substances that antago- 
nize auxin action, such as triiodobenzoic acid, increased flowering in soy- 
beans (Galston, 1947). De Zeeuw and Leopold (1956) report that low 
auxin concentrations applied to two short-day species promoted floral 
initiation if applied before the induction period but are less effective 
afterward. In a Xanthium plant defoliated to a single leaf, Salisbury 
( 1955 ) found that auxin inhibited flowering if applied before the flower- 
ing stimulus (produced by photoperiodic induction) had been com- 
pletely translocated from the leaf but promoted it if applied afterward. 
Leopold and Guernsey ( 19535 ) , using the position of the first flower in 
peas as a measure of flower initiation, observed that a number of sub- 
stances, notably sucrose, malic acid, and arginine, tended to inhibit flower- 
ing but that this inhibition was removed by auxin. Although flowering 
is usually an "all or none" reaction, structures intermediate between 
flowers and vegetative shoots sometimes occur. Such phyllody has been 
produced through manipulation of the flowering stimulus by Harder and 
his colleagues (1947). Gibberellin is often effective in flower induction 
(Lang, 1957). 

The pineapple (Clark and Kerns, 1942) produces flowers abundantly 
if naphthaleneacetic acid or certain other growth substances are applied 
by spray to the center of the plant. Van Overbeek (1946«) has shown 
that plants thus treated will flower under long days, which ordinarily 
inhibit flowering in pineapple. In the sweet potato, also, Howell and 
Wittwer ( 1955 ) reported that flowering can be induced experimentally by 
a growth substance. 

The problem of the relation of growth substances to flowering is evi- 
dently a complex one. It is the basis of a considerable literature, much 
of which can be found in Melchers and Lang (1948), Lang (1952), Bon- 
ner and Liverman ( 1953), and Liverman ( 1955). 

Something analogous to the control of flowering by specific substances 

Growth Substances 


is even to be found in the algae, for von Denffer and Hustede (1955) 
were able to shift the sexual phase of Vaucheria sessilis to the vegetative 
one by treatment with indoleacetic acid (Fig. 18-19). 

Sex Determination. Both in the determination of the sex of individual 
plants and in the development of the sex organs, growth substances of 
various sorts seem to be effective. 

In the dioecious species Cannabis sativa (hemp) it is possible to dis- 
tinguish genetically male from genetically female plants before they 
flower. Heslop-Harrison (1956) grew plants under controlled photo- 
periodic conditions and during the period of differentiation of flower 
buds applied naphthaleneacetic acid to leaves at the third and fourth 
nodes. In genetically male plants, female flowers were produced, sug- 

-10 -11 . 
m m 

JES 9/an J 

Fig. 18-19. Effect of different concentra- Fig. 18-20. Young prothallia of Pteridium 
tions of IAA on the proliferation of an- aquilinum 11 days after spore germina- 
theridial primordia in Vaucheria. (From tion. A, grown on agar to which a water 
von Denffer and Hustede. ) extract of prothallia was added. B, grown 

on ordinary agar. The extract stimulates 
early development of antheridia. (From 
Dbpp. ) 

gesting that sexuality is determined by the concentration of native auxin 
during the period of primordium differentiation and that femaleness is 
associated with a relatively high auxin level. In Mercurialis ambigua 
he found ( 1957) that carbon monoxide much reduced the number of male 
flowers in genetically monoecious types, presumably by its effect on 
auxin. Laibach and Kribben (1951) painted the lower surfaces of the 
leaves of cucumber, a monoecious plant, with naphthaleneacetic acid 
in lanolin and caused an increase in the proportion of female flowers, 
sometimes altogether suppressing the differentiation of males. Indole- 
acetic acid they found to be less effective and 2,4-D more so. Extending 
his work to other plants, Laibach concludes (1953) that, in general, 
female flowers or female parts of flowers tend to differentiate under 

400 Morphogenetic Factors 

higher concentrations of growth substances than do male flowers or 

It has been suggested that hormones comparable to those of the animal 
body may influence sex in higher plants. Love and Love ( 1946) found that 
in Melandrium dioicum sex expression is influenced by various animal 
sex hormones applied in lanolin to the axils of leaves in which flower 
buds are to develop. Crystalline estrone, estradiol, and estradiol ben- 
zoate shifted the sex of the flowers toward femaleness, whereas testos- 
terone and its propionate promoted maleness. In general, hormones 
promoting maleness or femaleness in animals have the same tendency 
in Melandrium. Some doubt has been cast on these conclusions by Kuhn 
(1941), who studied dioecious species of Cannabis and Mercurialis. 
There is no evidence that substances identical with animal sex hor- 
mones are formed by plants. If sex in dioecious plants is determined by 
specific substances, these have not been isolated nor can they be passed 
from one plant to another of opposite sex by grafting (Yampolsky, 

Maleness in ferns seems to be related to specific substances. Dopp 
(1950) made a water extract of the prothallia of the bracken fern which 
stimulated the production of antheridia in sporelings 4 to 8 weeks 
earlier than in untreated prothallia (Fig. 18-20). This can be carried in 
agar media. Naf (1956) confirmed this and was further able to induce 
antheridium formation on a variety of other related ferns even though 
these did not normally produce them in culture. The extract from 
prothallia of types of ferns that form antheridia under the conditions of 
culture used was several thousand times more effective than were extracts 
from types that do not form antheridia under these conditions. Such ex- 
periments suggest that in the prothallia of all polypodiaceous ferns there 
is a substance that stimulates the formation of male sex organs. 

Sex hormones have also been found in all the thallophytes except the 
red algae and the basidiomycetes. Burgeff (1924) reported that in non- 
aquatic types such as Mucor mucedo, the hyphae of two different sexes 
("plus" and "minus" races) influence each other by means of diffusible 
substances. Kohler (1935) confirmed these results, and Plempel (1957) 
has reported the activity of four substances in sexual interactions in this 
species. Kohler found that in Phy corny ces Blakesleeanus two diffusible 
substances are produced by each sex. Krafczyk ( 1931 ) showed that in 
Pilobolus crystallinus at least three different processes are chemically 
controlled: the characteristic swelling and branching of the hyphae, the 
growth of hyphae toward each other, and the delimitation of the game- 
tangia. Machlis ( 1958 ) has found in the water mold Allomyces a hor- 
mone, sirenin, excreted during female gametogenesis that attracts the 
male gametes to the female ones. 

Growth Substances 


Similar processes have been more fully studied in the water mold 
Achlya by J. Raper (1939-1957), especially in A. bisexualis and A. ambi- 
sexualis (Fig. 18-21). Experiments involving the transfer of mycelia 
into water where plants of the opposite sex had been growing, and the 
use of cellophane membranes, gave evidence that in the very regular 
sequence of the sexual process four diffusible substances are concerned. 

A 1 Augments 
'tf Inhibits — - 

Production of 

Antheridial Hyphae 


— A Initiate 
Mutually Aug. 
— ^Initiates 

A- Complex 


Anth. HypriafF^ 

Delimitation \.\\ 
of Antheridia ^ 

Production of 
pi Initials 

Delimitation of 

ifferentiation of 


Maturation of 

Fig. 18-21. Sex hormones in Achlya. Specific activities of substances A, B, C, and D 
in the development and function of the sex organs. ( From J. Raper. ) 

402 Morpho genetic Factors 

In the male mycelium antheridial branches are first induced by hormone 
A, produced by the female mycelium. The antheridial branches then 
form hormone B, which induces the production of oogonial initials in 
the female plant. These structures now form hormone C, which causes 
the antheridial branches to grow chemotropically toward the oogo- 
nial initials. Lastly, it appears that hormone D, presumably formed 
in the antheridia, causes the oogonial initials to delimit the oogonia 
from their stalks. The chemical nature of these substances is still un- 

The most complex examples of the effects of specific substances in 
sexual reproduction and sex determination in the lower plants are those 
described by Moewus ( 1940 and many other papers ) in the unicellular 
green alga Chlamydomonas eugametos, in which the biochemical and 
the genetic basis of the various hormonal mechanisms were subjected to 
detailed analysis. A thorough reexamination of this work indicates that 
many of the facts and conclusions of Moewus are not well founded and 
that the contributions of the Chlamydomonas work to our understand- 
ing of sexuality in the lower plants are much less considerable than 
they were once thought to be. 

Work on the sexual processes and substances in thallophytes has been 
reviewed by J. Raper (1952, 1957). Hawker (1957) has reviewed the 
whole field of reproduction in the fungi. 

Wound Hormones. The substances first proved to have a determining 
effect on morphogenetic processes were the wound hormones, or necro- 
hormones. It has long been observed that in the vicinity of dying and 
necrotic cells there occur divisions in other cells which under ordinary 
conditions would not have shown such division. These have a definite 
relation, both in distribution and orientation, to the accumulation of de- 
composition products released by the injured cells. Wound meristems are 
thus developed which produce layers of cork that cut off the injured 
region and protect the healthy tissue underneath. 

Haberlandt (1921, 1922) was the first to attack this problem directly. 
He found that if the cut surface of a potato tuber is washed and the 
contents of the injured cells thus removed only a few cell divisions 
occur. It might be thought that in such cases the reduced access of 
oxygen to the flooded tissues would account for the reduction in 
metabolic activity and thus of cell division. The action of a definite 
substance, however, is strongly indicated by later experiments of 
Haberlandt and others in which the juice, debris, or extracts of tissues 
produced an effect on cell division much exceeding that from mere 
wounding (Fig. 18-22). When sap from crushed tissue was injected into 
small intercellular spaces, active cell division, presumably from wound 
hormones, was found to occur (Reiche, 1924). These substances are not 

Growth Substances 


species-specific, for those from one species will produce this effect in quite 
unrelated ones. 

Much work has been done in isolating wound hormones and deter- 
mining their chemical nature. Standard material for estimating relative 
effects of hormone activity was first sought. Wehnelt (1927) used the 
layer of parenchyma cells which lines the immature pod of the common 
snap bean. Such tissue responds sensitively to various types of stimula- 
tion by abundant cell division and the formation of intumescences, the 
size of which provides a rough measure of the intensity of the stimulus. 
This "bean test" has been used by many students of wound hormones 
(Jost, 1935; Umrath and Soltys, 1936). On such pod surfaces Bonner 
and English (1938) placed droplets of extract from crushed tissue 

Fig. 18-22. Effect of wound hormones. 
Section of internode of Kalanchoe below a 
wound, showing how cortex cells have been 
induced to divide frequently, and parallel 
to the wound surface. ( From Sinnott and 
Block. ) 

(chiefly bean pods) and found that the height of the intumescence 
which developed after 48 hours was proportional to the concentration 
of the wound hormone present in the extract. These proliferations are 
usually higher than the ones induced by other chemical or physical means. 

Considerable progress has been made toward a knowledge of the 
chemical nature of these wound hormones. Bonner and English isolated 
from bean-pod juice a substance which in low concentration was very 
active in the bean test and named it traumatin. English, Bonner, and 
Haagen-Smit later (1939) purified from the same source a crystalline 
dibasic acid similar in its effects. Traumatin appears to be active on only 
a few types of cells, such as those of the potato tuber and the bean pod. 

To understand wound reactions in any plant, organ, or tissue, account 
must be taken of many internal and external factors as yet imperfectly 
known (Bloch, 1941). Workers have often been puzzled by differences 

404 M or pho genetic Factors 

between cells in their response to wounding or wound hormones. Such 
cells evidently differ in character and reactivity. Thus the root pericycle 
and the vascular cambium respond to injuries by the production of 
wound tissue much more readily than do adjacent cells of the ground 
parenchyma. Auxin may be one factor which produces such specific 
reactivity. In Populus balsamifera, Brown (1937) found that the cam- 
bium was stimulated to active growth both by wound hormones released 
from dead cells and by a substance, presumably auxin, coming from buds 
and leaves distal to the wound. Application of auxin above the wound 
considerably increases a local wound reaction (Brown and Cormack, 
1937). Other observations confirm this (Bloch, 1941). 


Numerous instances have been reported where specific changes in in- 
ternal structure are related to the action of auxin or one of the synthetic 
growth substances. 

In the regeneration of buds on the decapitated hypocotyl of flax (p. 
245), Link and Eggers (19466) found that this was largely prevented if 
indoleacetic acid in lanolin was applied to the cut surface. Even the 
transverse division of epidermal cells, the first visible step in bud pri- 
mordium differentiation, was usually inhibited. 

Nysterakis and Quintin (1955) report that application of 2,4-D to 
growing stems of Araucaria excelsa reduced the length of the tracheids 
by more than half and changed the pitting from circular to scalariform. 

Jacobs (1956) finds that the regeneration of severed xylem strands 
and the distribution of auxin proceed together, and from this and other 
evidence he concludes that auxin is usually the limiting factor in the 
differentiation of xylem. The chief distinction of xylem cells is their 
thick secondary wall. The facts that auxin is effective only in plants- 
organisms with cellulose walls-and that the only plant group where 
auxin seems to have little effect on growth is the fungi, which have 
chitinous rather than cellulose walls, both suggest that auxin acts on the 
cell wall. 

Native auxin and synthetic growth substances have been shown to be 
effective in preventing the abscission of leaves and fruits. If a leaf 
blade is removed but the petiole left, this will soon drop off by abscission. 
If one of several growth substances is placed on the cut petiole stump, 
however, abscission will not take place. Presumably when the leaf is in- 
tact auxin is continually moving down the petiole and inhibits the 
differentiation of an abscission layer at the base. The use of sprays of 
various growth substances to prevent the fall of leaves or fruits under 
certain conditions is now a common horticultural practice. What the 

Growth Substances 405 

mechanism is by which the abscission layer is produced or inhibited is 
not known. 

Sprays of this sort are also used to stimulate rather than inhibit 
abscission, notably for the purpose of thinning young fruits when too 
many have been set. How, one may ask, does the same substance act 
in two such different ways? Evidently a normal growing and functioning 
organ will produce enough auxin to prevent its abscission. When this 
production ceases, the organ will drop off unless a fresh supply is avail- 
able through external application. Anything which checks or deranges 
normal growth, however, will tend to check auxin production and thus 
lead to abscission. Sprays of some substances and in certain concentra- 
tions will tend to do this, and hence their usefulness in the thinning of 
fruit. An answer to the problem of this double effect has been proposed 
by Jacobs ( 1955 ) , who has shown that in addition to the inhibiting 
effect of auxin on abscission there may be a speeding effect produced 
by auxin formed in young nearby leaves. This stimulates the abscission 
of a petiole whenever the flow of auxin from its leaf blade is reduced. 
Abscission is thus controlled by an "auxin-auxin balance." 

The differentiation of more specialized tissues may be stimulated by 
auxin. Camus (1949) grafted buds of Cichorium to pieces of storage 
tissue and found that vascular strands began to differentiate just below 
the bud and continued to develop until they established connection with 
the vascular tissue beneath. Buds encased in cellophane and inserted 
into tissues cultivated in vitro produced the same effect, indicating that 
a diffusible substance, possibly auxin, was involved (Fig. 18-23). 

Of significance here is the work of Wetmore ( 1956 ) on the induction 
of xylem in callus tissues. Into homogeneous callus maintained in culture 
from parenchyma cells in the cambial region of lilac, a growing lilac 
stem apex was grafted by inserting it into a V-shaped cut. After tissue 
union was effected, strands of xylem began to differentiate into the 
homogeneous callus tissue below the graft. That auxin was the factor 
responsible for this is suggested by the fact that when the cut was filled 
with agar containing auxin but without a stem tip vascular tissue ap- 
peared below it in the same way, the distribution of the strands depend- 
ing on the concentration of the auxin. It is significant that only xylem 
tissue was thus differentiated and not phloem. This was also the case 
in the regeneration of severed vascular strands in the stem of coleus 
(p. 242). It may be that the factors which stimulate xylem develop- 
ment are different from the ones involved in phloem production. 

More profound effects of growth substances on the anatomy of plants 
have been described. Much of the work here has been done by Kraus and 
his colleagues at the University of Chicago, who have tried a variety of 
substances in different concentrations and on many plants. Their gen- 

406 Morpho genetic Factors 

eral conclusions are that most of the changes produced are in the ab- 
normal distribution and proportions of tissues rather than in the produc- 
tion of new structures. Important factors in these changes are nutrition 
and the age and state of the tissues when treated. Marked differences 
in reaction to various substances were found, and morphogenetic 
processes may thus in part be manipulated. A review of this work has 
been made by Beal ( 1951 ) . 

Fig. 18-23. At left, effect of a bud grafted to the phloem region of a piece of chicory 
root in culture. The bud stimulates the development of vascular tissue below it, pre- 
sumably because of a growth substance it produces. At right, a similar experiment 
except that a sheet of cellophane, CL, has been placed between the bud and the tissue 
below. The same effect is produced, indicating that organic continuity is not neces- 
sary. B, bud; P, phloem; V, vascular parenchyma; C, cambium; A, histologically 
altered tissue; L, line of contact between bud and stock. (From Gautheret, after 
Camus. ) 


Various other substances of morphogenetic significance have been 
postulated, but little is known about them. Thus in Dictyostelium, the 
remarkable life cycle of which has been described earlier (p. 223), the 
factor controlling the aggregation of the myxamoebae into a pseudoplas- 
modium appears to be chemical in nature and diffusible (J. T. Bonner, 
1949). To this substance Bonner has given the name acrasin. 

In a few traits, such as time of fruit ripening in dates ( Swingle, 1928 ) 
and staple length in cotton (Harrison, 1931), pollen seems to have a 

Growth Substances 407 

direct effect not only on the embryo but on the tissues of the ovary 
itself, tending to make these somewhat resemble ones of the paternal 
parent. This metaxenia must evidently be due to gene-produced chemi- 
cal factors introduced through the pollen tube and modifying the de- 
velopment of such maternal tissues as the pericarp and the seed coat. 

The development of galls with specific external and internal struc- 
tures (p. 285) produced by fungi or insects seems to be dependent on 
various chemical substances. In insect galls these may be injected into 
the plant by the insect but more probably they result from secretions 
from the growing larva. 

In other galls, particularly the one most actively studied— crown 
gall— auxin is clearly involved (p. 294). Experiments with tissue cultures 
have shown that cells of normal tissue in many cases are unable to 
grow unless supplied with auxin. Cells of bacteria-free crown-gall tissue, 
however, can do so. This fact suggests that the change from normal to 
tumor tissue may result from the acquirement by tumor cells of the 
ability to synthesize auxin. It is probable, however, that the problem is 
more complex and that changes in the ability to form other growth 
substances are also involved. Thus Braun and Naf ( 1954 ) have extracted 
from crown gall a biologically active substance which is not auxin but 
which, in association with auxin, produces active proliferation of to- 
bacco-pith tissue in culture. Neither this nor auxin alone has a growth- 
stimulating effect of this sort. The question of the relation of auxin to 
crown-gall formation has been actively investigated. The subject has been 
reviewed by Braun and Stonier (1958). 

Growth substances are also involved in the production of other gall- 
like structures. Swellings and malformations somewhat resembling typical 
root nodules associated with nitrogen-fixing bacteria have been induced 
by application of synthetic growth substance on the roots of several types 
of leguminous plants (Allen, Allen, and Newman, 1953). 

A number of other growth substances deserve mention here. Adenine, 
for example, has been found to possess significant properties, especially 
for leaf growth, and the balance between it and auxin seems to determine 
the character of development in some cases (p. 414, and Skoog and 
Tsui, 1951). 

The synthetic growth substances are too numerous to be discussed 
here. Of particular note is 2,4-dichlorophenoxyacetic acid (2,4-D), im- 
portant because of its wide use as a herbicide. It produces such profound 
growth abnormalities that death usually ensues (Kaufmann, 1955; Fig. 
18-24). For some unknown reason it has relatively little effect on 
monocotyledonous plants. Work with it has been reviewed by Wood- 
ford, Holly, and McCready (1958). 

Maleic hydrazide is important in that it inhibits growth in a wide 

408 Morphogenetic Factors 

variety of plants without causing obvious morphological abnormalities. 
Plants treated with it tend to lose dominance in their apical buds and 
show certain other effects (Naylor and Davis, 1950). It seems to inhibit 

Fig. 18-24. Median longitudinal section of young adventitious root apices in rice. Left, 
of untreated plant. Right, of plant treated with 2,4-D. There is a great increase of 
periclinal divisions in the latter, which produces massive, abnormal roots. (From 
Kaufman. ) 

mitosis (Greulach and Atchison, 1953) and also checks the formation of 
flower primordia in both long-day and short-day plants (Klein and 
Leopold, 1953; Fig. 18-25). 



u. Q- 


oc o 
uj or 
m o 



o o 


.1 I 4 10 40100 MG/L. 


Fig. 18-25. The effect of maleic hydrazide on the total number of flower primordia at 
the first five nodes of soybeans. ( From Klein and Leopold. ) 

Substances such as triiodobenzoic acid and coumarin under some 
conditions increase the effect of auxin and under others markedly inhibit 
growth. A number of substances occur that inhibit or antagonize auxin 
action (J. Bonner, 1949) and have been termed antiauxins (Fig. 18-26). 

Growth Substances 


Some gases have been found to exert strong formative effects on 
plants and thus deserve to be included among the growth substances, 
although chemically they are very different from the rest. In studying 
the effects of illuminating gas and its constituents on greenhouse plants, 
workers at the Boyce Thompson Institute found that in tomato plants 
exposed under bell jars to atmospheres containing 1 per cent of carbon 
monoxide the stems became covered after a few days with an abundant 
growth of roots. Other gases produced similar effects. Carbon monoxide 
was found to induce rooting in many other plants (Zimmerman, Crocker, 
and Hitchcock, 1933 ) . These results led to the investigation of the effects 
of ethylene, acetylene, and propylene (Zimmerman and Hitchcock, 
1933). All were found to induce rooting and root-hair formation, leaf 



0.1 1.0 10.0 

Fig. 18-26. Growth of Avena coleoptile (upper curve) induced by various concentra- 
tions of indoleacetic acid. The degree of inhibition of this growth by an auxin an- 
tagonist ( 4-chlorophenoxyisobutyric acid) at concentrations of 1, 10, and 50 mg. /liter 
is shown in the three successively lower curves. ( From Foster, McRae, and Bonner. ) 

epinasty, proliferation of callus-like masses of tissue, and abscission of 
leaves, flowers, and fruits. These gases, however, do not stimulate growth 
in the absence of auxin. The relation of ethylene to auxin has been 
discussed by Michener (1938). 

In Puerto Rico Rodriguez (1932) discovered that ethylene induces 
flowering in the pineapple, and in Hawaii it was found soon afterward 
that acetylene will accomplish the same result (Lewcock, 1937). The 
effect of these gases on pineapple is much like that of the synthetic 
growth substances which induce flowering (p. 398). 

In addition to the growth substances which have here been discussed- 
auxin, various other naturally occurring substances, and the synthetic 
compounds— another is now being actively studied and is assuming an 
important place in morphogenetic problems. This is gibberellin. 

410 Morpho genetic Factors 

In a disease of rice produced by the fungus Gibberella fujikuroi, it 
was observed some years ago that many of the infected plants grew 
taller than uninfected ones (Kurosawa, 1926). Young and uninfected 
rice plants treated with culture filtrates of this fungus grew unusually 
tall. A similar increase in growth was observed when this filtrate was ap- 
plied to some other plants, both monocotyledons and dicotyledons. Sev- 
eral different but related substances were purified from Gibberella and 
are now commonly termed the gibberellins. Their nomenclature is still 
somewhat confused but they may be named gibberellin A ls A 2 , and A :i , 
the last being the best known and often called gibberellic acid. 

Gibberellin commonly produces a very marked increase in stem 
elongation ( Fig. 18-27 ) . This is particularly conspicuous in certain dwarf 




Sprayed with 20 ppm. 
Gibberel lie Acid 






weeks after spraying 
Fig. 18-27. Relation between concentration of gibberellic acid and plant height in 
bean plants. ( From Gray. ) 

plants, notably peas. Brian and Hemming (1955) induced a fivefold 
increase in height in such plants, bringing them up to the size of tall 
races, by applying a little of this substance to one of the leaves. It had 
no effect on plants of the tall races. The length but not the number of 
internodes was increased. The so-called "slender" mutants of peas, 
which are tall but spindly, showed no effects of gibberellin treatment. 
Phinney ( 1956 ) found that gibberellin caused some dwarf mutants in 
maize to grow as tall as the normal plants from which they had been 
derived but some other dwarf races showed little response. Tall plants 
were unaffected. The relation of gibberellin to dwarfing is evidently a 
complex one. Most of the elongating effect is caused by increase in cell 
length rather than in cell number. There are a few cases, however, where 
cell division as well as cell elongation has been stimulated. 

Growth Substances 411 

Leaf growth is affected by gibberellin and is often, though not always, 
increased. Radish leaves floated on a gibberellin solution in the dark 
grew larger than the controls. It is perhaps significant that kinetin has 
much the same effect and that when both substances are applied the 
increase in growth is equal to the sum of their separate effects. 

Although gibberellin and auxin are similar in some respects, notably 
in stimulating cell elongation, they differ chemically and in other im- 
portant ways. Gibberellin fails to produce typical epinasty nor does 
it induce callus formation, both of which usually result from auxin ac- 
tion. It also fails to show the polar transport within the plant so character- 
istic of auxin. It does not inhibit the growth of lateral buds but tends 
instead to stimulate their development. It does not check leaf abscission. 
It inhibits rather than promotes root initiation but does not inhibit root 

Gibberellin evidently has some relation in its effects on development 
to those produced by light, though this relation is not clear. Lang ( 1957 ) 
found that it induced biennial Hyoscyamus to flower the first season, re- 
gardless of day-length. The usual inhibition of growth produced by red 
light is removed by treatment with gibberellin. The effects of this sub- 
stance much resemble etiolation but are independent of light. This 
is unlike the effect of auxin. 

In its influence on dwarf plants, which seems to be its diagnostic 
feature, gibberellin perhaps substitutes for some essential factor that is 
normally present and which may have been lost by mutation. No effects 
of gibberellin have as yet been found in any of the lower forms. Sub- 
stances essentially like it have now been extracted from several higher 
plants (Radley, 1958), and it is probably widely spread in the plant king- 

The literature in this field has been reviewed by Stowe and Yamaki 
(1957). Further study of gibberellin should yield important information 
on the factors governing plant development. 


Relatively little is known chemically about most of the growth sub- 
stances. Some of them, like the calines, are little more than inferences. 
The existence of others, such as florigen, can be proved by experiment, 
though they have not been isolated. Others can be isolated, at least 
partially, but their chemical nature is not well known. As to auxin, 
gibberellin, and traumatic acid, fairly complete chemical information is 
now available. 

Many attempts have been made to find some common features of 
chemical structure among these substances which have formative effects 

412 Morphogenetic Factors 

and thus to gain a clue as to how these effects are produced. These sub- 
stances vary considerably. Having studied many such compounds, 
Koepfli, Thimann, and Went (1938) concluded that the minimal struc- 
tural requirement for a substance to stimulate growth, at least in the 
pea test for auxin, is to possess an unsaturated ring system with a side 
chain adjacent to a double bond in the ring, and with a carboxyl group 
in the chain separated from the ring by a carbon atom. Thimann (1957) 
points out that there are a considerable number of biologically active 
compounds which do not have this structure and at least one that does 
not even contain a ring. It seems doubtful that an understanding of 
the mechanism of action, either of auxin or similar synthetic substances, 
will be gained by a knowledge of their chemical structure without an 
equal knowledge of the reacting systems that they stimulate. The general 
question of the chemistry and mode of action of plant growth substances 
was discussed at the Wye College symposium (Wain and Wightman, 

Just how growth substances produce their morphogenetic effects is 
not well understood. The first visible result of auxin action is a speeding 
up of protoplasmic streaming, indicating that some aspect of metabolism 
is being accelerated. The marked influence of auxin on growth also sug- 
gests this, since growth requires the release of energy. Some physiologists 
believe that auxin acts as a respiratory coenzyme and thus has an im- 
portant share in the respiratory cycle. No enzyme has yet been found, 
however, that can be activated by auxin in physiological concentrations. 

Since what appears to be the primary effect of auxin is cell enlarge- 
ment, it seems plausible to conclude that water uptake is controlled by 
it, and there is some evidence for this. The suggestion has been made 
that auxin increases the osmotic concentration of the cell sap and thus 
increases cell size. Cell growth may take place, however, even with a 
decreasing osmotic concentration. Burstrom (p. 41) believes that cell 
enlargement is not primarily caused by water uptake. 

Interest at first focused on auxin-induced changes in the cell wall as 
related to growth, and Heyn (1940) has reviewed the evidence that 
auxin directly increases the plasticity of the wall and thus its irre- 
versible extensibility. Some workers believed that the effect of auxin 
was indirect and only through its influence on the cytoplasm. Recent 
studies, however, support Heyn's view. Thus Cleland and Bonner ( 1956 ) 
present evidence that auxin directly induces a loosening of the cell-wall 
structure and thus a relaxing of wall pressure, which makes possible an 
expansion of the cell. The effect of auxin is independent of cell expan- 
sion. Auxin may affect the wall by altering pectin metabolism. 

The relation of growth to protoplasmic viscosity and to the swelling 
capacity of cell colloids suggests that auxin may have something to do 

Growth Substances 413 

with these qualities. Northen (1942) observed that auxin usually de- 
creases viscosity of protoplasm. He regards protoplasm as a "reversibly 
dissociable-associable system," in which auxin (and other agents) 
cause dissociation of cellular proteins and increased swelling pressure. As 
the result of such action, it is thought that components of the fine struc- 
ture of the cell may undergo reorientation and the reactivity of the 
cell may be changed. 

Of greater morphogenetic interest is the direct relation of growth sub- 
stances to the development of organs or structures. The situation here 
is even less clear than in the control of growth. In a number of cases one 
is faced with a curious antithesis between the action of auxin in dif- 
ferent situations. For example, at a given concentration it stimulates the 
growth of stem tissue but inhibits that of the primary root, with the 
result that differential geotropic bending occurs. It stimulates the de- 
velopment of root primordia and hence is useful in the rooting of cut- 
tings, but it checks the elongation of the roots after their emergence. In 
some cases its effect is to accelerate flowering and in others to inhibit it. 
Sometimes it prevents bud growth and sometimes it stimulates this. In 
certain cases its effect is to stimulate the growth of roots rather than buds 
but in other cases it has just the opposite influence. 

In this confusing situation the hypothesis of specific organ-forming sub- 
stances appeals to many, especially those who seek direct and primarily 
chemical solutions to morphogenetic problems. To be sure, development 
often does seem to be the result of the action of such substances, as in the 
formation of roots, flowers, and abscission layers. But where, one may 
ask, does this specificity end? In the flower, are there separate substances 
for sepals, petals, stamens, and ovaries, for anther and filament, style and 
stigma? Does each tissue and each type of cell have its appropriate 
"caline"? It is easy to reduce to absurdity the more naive statements of 
this hypothesis. 

To what, then, can one attribute the highly specific results of plant 
development? One answer is that the specificity lies in the protoplasmic 
system rather than in the growth substance and that the latter serves 
primarily as a trigger or evocator that calls out a specific response. We 
should remember that auxin, the substance about which most is known, 
is markedly nonspecific. A few such biologically active substances, 
stimulating responses from a highly organized protoplasmic system, 
might account for development. A dime, it has been said, will open 
a turnstile, activate a dial telephone, or bring a tune from a juke box, but 
the dime, like a molecule of auxin, is identical in every case. The dif- 
ference lies in the complexities of the responding mechanism. The answer 
to morphogenetic problems is more difficult to come by on this conception 
than on that of specific formative substances since it involves an under- 

414 Morpho genetic Factors 

standing of the whole protoplasmic system. This is a far goal, but, as 
Thimann (1957) has remarked, "It begins to look as though the whole 
cell were necessary to auxin activity." 

But part of the living system in a plant evidently includes other 
biologically active substances. A good deal is known about several of these, 
and although they may not be "organ-forming" in the earlier sense of the 
word, their share in the control of developmental processes is more 
important than that of most other chemical compounds. A hopeful 
method of attack on morphogenetic problems is to study the relationship 
between these substances. It is now well known that there are com- 
pounds which enhance or which inhibit the effects of auxin. Still more 
promising are results such as those of Skoog and Miller ( 1957 ) on the 
relationships between auxin and adenine (or its derivative, kinetin). The 
presence of both seems to be necessary for vigorous growth, at least of 
tobacco callus in culture. If the proportions of the two are changed, how- 
ever, the character of the growth is different. Relatively high levels of 
auxin and low ones of kinetin, either in cultures of tobacco callus or in 
cuttings, will tend to produce good root growth but little bud develop- 
ment, whereas high kinetin and low auxin levels favor growth of buds 
instead of roots. To be sure, kinetin is a substance which has not yet 
been found in the living plant, and its balance with auxin has been 
demonstrated in only a few cases, but the picture this balance pre- 
sents of the possible control of differentiation through alteration in pro- 
portions within a relatively simple chemical situation opens up encourag- 
ing possibilities. 

Other factors are doubtless concerned in these cases, and the problem 
must involve more than a simple two-compound interaction, but the 
idea that there may be a relatively small number of active but non- 
specific substances, with the possibilities for complex interactions among 
them that this offers, makes understandable how an essentially infinite 
number of different structures might be produced without the necessity 
of postulating the activity of substances specific for each of them. Only 
12 different kinds of chessmen can produce, by their various relation- 
ships, an almost limitless number of combinations on the board. 

The study of plant growth substances has been of great significance 
and stimulation for morphogenesis, but it has done little more as yet than 
pose a series of deeper problems. Chief among these are three: 

1. What is it that controls the distribution of growth substances as to 
space, time, and concentration? 

2. What is it that determines the specific response which a given cell 
or tissue makes to them? 

3. How do they interact in their control of development? 

These problems are part of the deeper one of biological organization. 


Genetic Factors 

The factors discussed thus far are effective chiefly through the environ- 
ment of the plant, either its external surroundings or its inner physiologi- 
cal processes, which are open to relatively simple analysis. It is obvious, 
however, that these factors alone are not enough to explain all mor- 
phogenetic phenomena. There are also inborn differences, rooted in the 
specific constitution of each individual organism, which powerfully affect 
what it is and does. These differences are inherited, and it has been the 
great service of genetics to biology since the turn of the century to show 
that their physical basis is primarily in the genes, located in the chromo- 
somes of the nucleus. 


The various environmental factors exert their effects against this 
specific genetic background, the entire system of genes in the plant. 
Genes are sometimes thought of as though these bodies, known to be 
independent in inheritance to a certain degree, were also independent 
in development. This evidently is not true, however, for in their control 
of growth and differentiation the action of all the genes must be closely 
coordinated, in space and time, if an organism is to be produced. How 
these distinct entities are thus so precisely correlated in their action is a 
major problem for both genetics and morphogenesis. It is also clear that 
a gene does not produce its effect by determining a precise series of steps 
leading to the development of a specific trait, for the same genotype 
may have a very different effect if the environment is different. A gene 
simply determines a specific response to a specific environment. The 
genetic constitution that distinguishes a tall plant from a short one, for 
example, will not produce this difference unless the conditions of tem- 
perature, moisture, and soil fertility are such as to make vigorous growth 

The problem of gene action, of how a gene or group of genes produces 
its effects, is now one of the central concerns of genetics and is being 


416 Morphogenetic Factors 

actively investigated. The role of genes in the synthesis of enzymes and 
other substances and thus in the determination of successive steps in 
metabolic processes is yielding much information as to the relations 
between genetics and physiology. Increasing knowledge of those re- 
markable compounds, the nucleic acids, is leading to an understanding 
of the chemistry of the gene and of the manner in which it reproduces 
itself. Indeed, desoxyribonucleic acid ( DNA ) has such significant proper- 
ties that some biologists hopefully believe that it will finally produce an- 
swers for most of the basic problems of their science. All this, however, 
has as yet thrown little light on how it is that genetic factors affect the size, 
shape, and structure of plants and their parts. This is a much more diffi- 
cult problem than working out the biochemical steps in the synthesis of an 
organic compound produced in plant metabolism. How a single pair of 
genes can determine, for example, whether a tomato plant will have the 
familiar deeply lobed leaves or the unlobed "potato-leaf" type is very 
difficult to see. Here something more than a series of chemical steps seems 
to be concerned. Growth relationships are being controlled, and at present 
we must admit that very little is known about how such control is exer- 
cised. A solution of this problem must start with a knowledge of what 
actually happens in the inheritance of form and structure. To make such 
a descriptive analysis and at the same time seek hopefully for clues that 
will lead to a knowledge of the mechanisms involved is the chief task at 
present of the student of morphogenesis who is interested in the genetic 
aspects of his science. 

Genes and Growth. The underlying problem in growth is the increase 
of living substance, due ultimately to the reduplication of genes. How this 
is accomplished and how the DNA molecule divides into two new ones 
like itself are now beginning to be understood. Traits of size, either of body 
or organ, are markedly affected by environmental factors, but there is also 
a genetic basis for most of them. Since the pioneer work of East and 
Nilsson-Ehle it has been recognized that most quantitative traits depend 
not on single genes but on a series of multiple factors or polygenes, cumu- 
lative in their effect and in most cases without dominance. Such traits are 
difficult to analyze genetically since it is rare that the effect of individual 
genes can be followed, though there are statistical methods for determin- 
ing the number of genes by which two individuals differ for a given trait. 
That polygenes are operative in quantitative inheritance is indicated by 
the fact that the variability of the F 2 is markedly higher than that of the 
parents or the Fi, as would happen if segregation were taking place. 
There is now a substantial body of evidence that confirms the multiple- 
factor hypothesis. 

In a few cases the inheritance of size is not so complex, and the effect 
of individual genes can be traced. One of these, vine height in peas, was 

Genetic Factors 


found by Mendel himself and is due to a single pair of genes, tall being 
dominant over short and segregating clearly in the F 2 . De Haan (1927, 
1930) has shown that in addition to this gene there are two (perhaps 
four) others that tend to inhibit growth. In this case a group of several 
genes, all modifying the same trait, can be recognized and their indi- 
vidual effects distinguished. In a considerable number of other cases it 
can be shown that two, three, four, or more pairs of genes are concerned 
in the inheritance of a size trait. Thus Quinby and Karper (1954) have 
evidence that in cultivated sorghum varieties, ranging from 2 to 15 
ft. in height, four pairs of genes are operative. 

Genes of this sort are cumulative in their effect. Sinnott (1937), Pow- 
ers (1939), Charles and Smith (1939), and others have shown that 
this additive effect is geometric rather than arithmetic, each gene 

Fig. 19-1. Geometric action of genes determining size. A, graph of fruit weight of an 
F a population of cucurbit fruits consisting of 244 individuals plotted in arithmetically 
equal classes. B, the same population plotted in classes equal logarithmically. The 
first population is skewed, the second nearly symmetrical. ( From Sinnott. ) 

contributing not a certain definite amount of height or weight but a 
certain percentage increase of the effect of the rest. This is shown by 
the fact that the F x is closer to the geometric average of the parents than 
to their arithmetic average and is thus somewhat nearer to the smaller 
parent in size. Furthermore, if a segregating F 2 is plotted in classes that 
are arithmetically equal, it skews toward the upper side, whereas if the 
scale is a logarithmic one, the F 2 population is symmetrically distributed 
around the geometric mean (Fig. 19-1). 

A developmental study of inherited size differences shows that some 
are attained by differences in rate of growth and some in its duration. The 
size differences between plants showing hybrid vigor and their parent 
inbreds are related to a more rapid rate of growth, and some other size 
differences also have their basis in genetically controlled growth rates. In 
other cases the difference in size is due to longer duration of growth. 

418 Morpho genetic Factors 

Large pumpkins differ from small gourds, for example, simply because 
they grow for a longer time (Sinnott, 1945b, and p. 16). The actual 
growth rate of these two varieties in terms of compound-interest growth is 

the same. 

Inherited size differences are also related to cellular characters (p. 34 ) . 
Most of them are due to differences in number rather than size of 
cells, large size being the result of more cell divisions during develop- 
ment. Less frequently the period of cell expansion is more extensive in 
the larger forms and their cells are consequently larger, though usually 
not in proportion to body size. 

In the many cases where there are inherited differences in cell size, 
it is usually not the size of the meristematic cells that is different but the 
amount of increase that occurs after division ceases. Thus the fruits of 
large races of pumpkins have much larger cells than do small gourds, 
but this difference is not evident at the meristem. Some of it appears in 
the growth of the young ovary but most of it during the enlargement of 
the ovary in fruit development (Sinnott, 1939). Sugar beets have much 
larger cells (and leaves) than do table beets, but only in their mature 
structures. The meristematic cells are much the same size in both. 

Many cases have been found in which there is not a gradation be- 
tween large and small types but the small ones are so markedly dif- 
ferent as to be regarded as somewhat abnormal dwarfs. In most such 
plants there is a single gene difference from normal which seems to 
control one important growth factor. A number of these occur in maize, 
and the auxin relations of some have been worked out (p. 377). Some 
dwarfs are small-celled but a few have cells larger than those of normal 
plants. There are also a number of gigas forms which are due to gene 
differences. Large and small types are also frequently related to chromo- 
some number (p. 438). 

Differences in height may result from mutations that alter a determi- 
nate type of growth to indeterminate. These have been found, for ex- 
ample, in tobacco (Jones, 1921) and maize (Singleton, 1946). Each 
shows single-gene segregation with normal determinate plants. The dif- 
ference between bush and climbing varieties of beans, also due to a 
single gene, is really a difference between determinate and indeterminate 

Another important effect of genes on size is to be seen in cases of 
hybrid vigor or heterosis. The Fi plants, in crosses between parents that 
are homozygous or essentially so, are often much larger and more vigor- 
ous than either parent (Fig. 17-1), and this fact has wide economic ap- 
plication, especially in maize. The difference is closely associated with 
heterozygosity and disappears with inbreeding. Various suggestions have 
been made to account for it— the stimulating effect of the heterozygous 

Genetic Factors 419 

condition, the dominance of linked genes, increased embryo size, and 
others— but no satisfactory explanation has yet been reached. Gene ac- 
tion in heterosis has been discussed by Jones (1957), and there is a 
wide literature in this field (Sprague, 1953). 

Genes and Form. The chief morphogenetic significance of genetic fac- 
tors, however, is in their relation to the development of organic form. Here 
it is not the total amount but the distribution of growth that is impor- 
tant. Genes must in some way control relative growth— the amount of 
growth in one dimension as compared with that in each of the others 
so that specific shapes are produced. Many instances could be cited 
where the shape of leaf or flower or fruit is certainly inherited and where 
differences in it segregate and can be analyzed in mendelian terms, at 
least to some degree. Only a few can be mentioned here. 

Leaf shape in cotton has had particular attention. In one of the earliest 
analyses of shape inheritance, Leake (1911) found that in crosses be- 
tween broad-lobed and narrow-lobed forms the F^ was intermediate and 
the F 2 showed segregation into i/4 narrow, V> intermediate, and 1 4 broad. 
Peebles and Kearney ( 1928 ) crossed shallow-lobed and deep-lobed types 
and found F x to be intermediate and a ratio of 1:2:1 in F 2 . In some 
varieties of cotton the genetic situation is much more complex. Both 
Hutchinson (1934 and others) and Silow (1939) postulate a series of 
multiple alleles, chiefly affecting lobing. They believe that the genes are 
"compound" and vary qualitatively as well as quantitatively. Hammond 
(1941) showed the importance of developmental analyses of shape in 
individual leaves and of changes of leaf shape along the stem. This 
method was carried further by Stephens ( 1944 ) . 

In the Japanese morning glory, Pharbitis, Imai (1930) and a number 
of other Japanese geneticists have studied the complex situation presented 
by the inheritance of leaf shape in crosses among its many varieties. 

Among other traits of form the inheritance of which has been analyzed 
in mendelian terms are fruit shape in Bursa (Shull, 1914), root shape in 
radish (Uphof, 1924), and leaf lobing in Tropaeolum ( Whaley, 1939). 

Evidence for the Existence of Genes for Shape. The problem under- 
lying all these instances of the inheritance of form is to find the method 
by which genes determine what the form is to be. The fact that such 
traits show segregation suggests that genes control them directly, but 
it is difficult to see how this is done. Some geneticists have tried to sim- 
plify the problem by assuming the operation of genes for individual 
dimensions only, as in the case of vine length in peas. Thus in tobacco 
flowers Anderson (1939) studied the inheritance of tube length and 
limb width in crosses between Nicotiana Langsdorfii and N. alata, species 
which differ in corolla shape. He observed that in F 2 there was much 
sharper segregation for length than for width, suggesting that fewer 

420 Morpho genetic Factors 

genes were operative in the former character than in the latter. There 
was by no means free recombination of length and width in F 2 , however, 
as independent assortment would require. The combinations that did 
occur were only a narrow segment of those theoretically expected. When 
length was plotted against width in F 2 , these were confined to a narrow 
segment of the total, running from combinations rather like one parent 
to those like the other through others resembling the Fj. He suggests that 
factors hindering free recombination might be gametic elimination, 
zygotic elimination, pleiotropism, and linkage. He believes that all of 
these may here be operative and suggests that all quantitative characters 
of an organism may be tightly linked, surely a radical conclusion. An 
extreme instance of the hypothesis that shape is the result of genes 
determining dimensions has been proposed by Frets (1947), who postu- 
lates that in the inheritance of seed shape in beans there are a series of 
independent genes for length, breadth, and thickness, respectively. 

H. H. Smith (1950) studied a cross much like that made by Anderson 
and comes to the conclusion that there is a developmental restriction to 
free recombination but that this is due to a "correlated growth pattern." 
In simpler words, there are genes that control shape directly rather than 
through individual dimensions. 

Evidence for the existence of such genes has been presented by Sin- 
nott (1935), chiefly from studies of the inheritance of fruit shape in the 
Cucurbitaceae. This evidence is of several types, as follows: 

1. If a race with flattened, disk-shaped fruits is crossed with a spheri- 
cal one, the F! shows complete dominance of the disk shape and in the 
F 2 there is sharp segregation into % disk and y± sphere. In another case, 
two different types of spheres, when crossed, show evidence of the ac- 
tion of complementary genes, for the F x is disk-shaped and in the F 2 
there is dihybrid segregation into % 6 disk, % 6 sphere, and y 16 elongate. 
Other shape differences can be analyzed in equally simple mendelian 
terms, though more genes are usually involved. 

2. In the disk-sphere cross, F 2 segregation for shape index (ratio of 
length to width) is sharp but those for length and for breadth are much 
less so, suggesting that the segregating genes control shape directly and 
not through dimensions. 

3. In one disk-sphere cross, the fruits of the disk parent were con- 
siderably larger than those of the sphere. The size of the F x was close to 
the geometric mean between the two parental sizes, and the means of the 
segregating F 2 disks and spheres were essentially similar to each other 
and close to that of the F\. This can be explained by assuming that 
size is determined by a series of genes but that the gene for shape is 
independent of these and molds into a particular form the material 
made available by the genes for size. 

Genetic Factors 421 

4. In a considerable series of crosses between races genetically more 
complex and differing in both the size and the shape of their fruits, a 
positive correlation was observed in each case between fruit length 
and fruit width in the parents and the Fj (where presumably all size 
differences are caused by environmental factors) but a negative one in 
F 2 where segregation occurs. This again can be explained by assuming 
that shape is inherited independently of size. A certain amount of ma- 
terial is genetically available for every fruit, and if its shape genotype 
tends to produce an elongate one, this will be relatively narrow, and 
hence the negative correlation. Maximum parental length is never com- 
bined in F 2 with maximum width, or minimum length with minimum 
width, as they should be if dimensions are directly determined genetically 
and recombined independently. 

5. In the F 2 the coefficient of variation for length is twice as large as 
that for width, which is to be expected in a radially symmetrical organ 
where shape and size are genetically independent, for if the amount of 
material for growth is fixed, a unitary change in width (equatorial 
diameter) should produce a much greater change in length, since volume 
is essentially width X width X length. 

Further evidence that genes for shape are actually operative is found 
in the fact that in a number of plants, such as the tomato (Lindstrom, 
1928; Butler, 1952), genes for fruit shape are linked with others and can 
be definitely located on chromosome maps (Fig. 19-2). If it were dimen- 
sions that are directly controlled, presumably their genes would occupy 
different loci. 

If genes determining shape actually exist, the difficult problem arises as 
to how they produce their effects. The ultimate mechanisms are by no 
means clear but the visible steps in the process can be described. In 
some cases the shape of an organ, such as a fruit, is established when it 
is a very small primordium. After this, growth rates in the various dimen- 
sions are equal, and as the structure grows its shape remains constant. The 
critical step in establishing growth relationships here is taken very early. 
This is what happens in fruits of pepper, tomato, some races of Cucurbita 
pepo, and various other plants. It is usual in organs that are nearly iso- 
diametric at maturity. 

In many fruits and in most leaves, especially where the dimensions are 
markedly unlike, the primordium is a roundish mass of cells and the shape 
of the organ is produced by differential growth among the dimensions 
(Sinnott, 1936b). In various races of the gourd, Lagenaria siceraria, for 
example, mature fruit shape varies from long and narrow through round 
to flattish. All are alike in early primordia. In elongate types (like the 
"Hercules club") length increases faster than width and as the organ 
grows in size it becomes progressively more elongate. Conversely, the 


Morpho genetic Factors 


Normal (M) Mottled (m) 

Smooth (P) Hi Peach (p) 

Normal (0) Oblate (o) 

Woolly (Wo/wo) I Normal (wo) 


Normal (Ne) U Necrotic (ne) 

Simple Inflor. (S) ■ Compound Inflor, (s) 

Non-beaked (Bk) I Beaked (bk) 

Few Locules (Lc) ■Many Locules (Ic) 

Fig. 19-2. Map of chromosome 1 of tomato. Two genes for fruit shape, and genes for 
other structural characters, have places on it. ( From Butler. ) 

Genetic Factors 


flattish types grow more rapidly in width. In all such instances, form 
changes as size increases. Where a particular race is genetically small, 
its fruit shape at maturity will be different from one which is genetically 
larger but has the same shape genotype. This fact complicates a study 
of the inheritance of shape in cases where the parents differ in both shape 
and size. 

In types where shape changes during development the dimensional 
relations, if plotted logarithmically, are allometric, the points falling 
along a straight line the slope of which measures the relative growth rate 
of the two dimensions (Fig. 5-8). It is this relative rate which the genes 
control, for if two races of Lagenaria differing in the slope of this line 









— r~ 


-i 1 — 

8 10 

20 30 


Fig. 19-3. Segregation of relative growth rates. Allometric growth of length to width 
of fruits in an F 2 from a cross between a rather elongate and a rather flat variety of 
cucurbits. The two F 2 classes resemble in general the parental forms. The Fi was 
like the elongate type. What is segregating is the character of the relative-growth 
line. (From Sinnott.) 

are crossed, the trait which segregates in F 2 is the steepness of this slope 
(Fig. 19-3), the value of k in the allometric equation (p. 105; Sinnott, 

It should be remembered that what is being controlled is not simply 
the relationship between two dimensions but between all the dimensions 
that make up the organic pattern. In the "bottle" gourd, for example, the 
length of the axis during growth is related to the width of the upper, 
sterile lobe; of the isthmus; and of the lower, fertile lobe. Relative to the 
axis, the lower lobe grows fastest, the upper next, and the isthmus least. 
The form of the whole structure thus changes in a precise and predict- 
able fashion. 

This concept of shape inheritance may be illustrated most simply by 

424 Morpho genetic Factors 

inscribing the lengthwise profile of an organic form like that of a fruit 
in a series of rectangular coordinates, as D'Arcy Thompson ( 1942 ) has 
done with various structures, and then seeing how other forms may be 
derived from it by deforming these coordinates in a particular fashion 
(Fig. 19-4). A change such as might be produced by a single gene dif- 
ference is evidently not a localized one but involves, at least to some 
degree, the pattern as a whole. 

The objection may be raised that in some cases a single dimension does 
seem to be inherited, as in vine height in peas. In a strongly polarized 
organ like the stem, height may be changed with little reference to stem 
diameter. In such cases height seems to be a direct expression of size in 

— ? ^--^ — 


I I - 


V ~ / 

Fig. 19-4. Geometric modification of fruit shape. The fruit at the upper left has its 
longitudinal profile inscribed in a grid of equal coordinates. This may be changed to 
various other shapes by changing these coordinates. ( From Sinnott. ) 

which the major effect of the genes is channeled in one particular dimen- 
sion, that of the polar axis. Ear length in maize, one of the first characters 
to be analyzed in terms of multiple genes, is the major dimension of a 
polarized organ and thus strongly affected by any genes that control 
total amount of growth. Here, however, ear width is also involved to 
some extent, and there are differences in the relation of length to width 
in various races. Such a case may perhaps be regarded as intermediate 
in genetic control between one in which genes for shape express them- 
selves in a weakly polarized structure (like a fruit), thus producing a 
wide range of patterns, and one in which growth is essentially limited 
to the polar dimension. 

Genetic Factors 425 

The evidence therefore favors the view that in most cases, certainly, 
genes control form directly. How this is done poses one of the most 
difficult problems of genetics and morphogenesis and one closely in- 
volved with that of biological organization itself. Most of the work on 
gene action has indicated that genes control the production of specific 
substances; but how, one may ask, does a substance control the de- 
velopment of a specific form? This involves the control of relationships, 
like that between length and width in a developing fruit, and not only 
single relationships but a whole series of them organized into a pattern. 
The amount of growth in one dimension is related precisely to the 
amounts in all the others. For a specific substance to do this involves the 
old question of organizers and organ-forming substances which are so 
easy to postulate and so difficult to picture in physical terms. These 
gene-produced substances seem rather to act as evocators, calling out or 
modifying formative potentialities in the living stuff of the organism. We 
need to assume the immanent presence in protoplasm of something that, 
for want of a better simile, can be compared to a system of coordinates in 
three dimensions. 

A specific substance may be thought of as bending or otherwise modify- 
ing these coordinates in a particular way and thus regulating growth in 
such a fashion that a specific pattern is produced. The problem remains 
as to the nature of this underlying formativeness. It may be thought of, 
perhaps, as a molecular pattern in the cytoplasm (p. 455). To call it a 
"field" is to give it a name but not an explanation. It is evidently involved 
in the nature of the living, organized system that an organism is. To 
recognize that it exists, even though one cannot yet explain it, is a step 
in advance and may save us from a too naive conception of the nature 
of gene action in development. 

Other Structural Traits. Aside from pure forms in the geometrical sense, 
various other structural characters are under gene control, or at least 
show mendelian segregation after crossing. Thus the zigzag stem that 
appeared in descendants of a certain cross between Tom Thumb pop- 
corn and Missouri dent (Eyster, 1922) behaves as a double recessive to 
normal stem. "Corn grass," a mutation in maize with narrow leaves, 
many tillers, and few male flowers, is a single-gene dominant ( Singleton, 
1951 ) . In peas, relatively long distance between first and second flowers as 
compared with the total length of the inflorescence was found by 
Lamprecht ( 1949 ) to be the result of three genes. Abnormalities of 
various sorts have also been shown to have a genetic basis. Among them 
are double flowers in many plants, as for example Tropaeolum (Eyster 
and Burpee, 1936), where doubleness is recessive but where a dominant 
"super-double" strain appeared, female sterile and with about 135 
petals. Peloria in Linaria vulgaris and Digitalis purpurea has been shown 


M 01 pho genetic Factors 

by various workers to differ in some types by a single gene from normal 
flower. Fasciation is a single-gene recessive to normal in Nicotiana 
(O. E. White, 1916). In both peloria and fasciation individuals are found 
which are usually normal but occasionally produce these abnormalities, 
presumably because of some genetic predisposition in this direction. These 
should be of particular interest in studies of the basis of gene action. 

There are two traits in the nature of abnormalities ("tufted" and 
"polycladous"), presumably gene-determined, in the liverwort Spliaero- 
carpos (Allen, 1924, 1925) which are of particular interest in that they 
occur in the haploid gametophyte generation so that the effects of a single 
gene are directly visible, unmodified by the influence of its allele. 

Fig. 19-5. Cross section of the stem of Aquilegia. Left, normal plant. Right, mutant 
with thicker cell walls. ( From Anderson and Abbe. ) 

Anatomical characters have also been shown to be directly affected 
by genes. In Aquilegia canadense a dwarf race with bushy, compact 
growth and stiff, brittle branches has been shown by Anderson and 
Abbe (1933; Fig. 19-5) to differ from normal by a single gene. The 
direct effect of this gene is to cause precocious secondary thickening of 
the cell walls, from which all the other differences follow. Piatt, Darroch, 
and Kemp ( 1941 ) report that in wheat, solid stem differs from normal 
hollow stem by three or four pairs of recessive genes. 

In all these cases where the form or structure of a particular plant part 
has been found to be controlled by a gene or group of genes, much 
valuable information as to the method of control may be gained by a 
study of the development of this structure. In a few cases this has been 
done, as in cotton leaves and cucurbit fruits described above; in dwarf 
mutants in maize by Stephens (1948); in structure and growth rates of 

Genetic Factors 


inbred and hybrid maize by Weaver (1946), Heimsch, Rabideau, and 
Whaley (1950), and Stein (1956); in a leaf mutant in maize by Mericle 
( 1950 ) ; and in the leaves of two species of Tropaeolum and their hybrid 
by Whaley and Whaley ( 1942 ) . In this last piece of work it was found 
that the pattern was essentially determined by early differential cell 
division in certain regions, the final form being attained by uniform 
cell expansion. Much more work of this sort, even at the purely descrip- 
tive level, needs to be done, for it will doubtless pose more sharply the 
problems which have to be solved and may suggest new methods of 
attacking them. 

Fig. 19-6. Acetabularia. A, A. mediterranea. B, A. wettsteinii. C, a stalk piece (shaded) 
of the former species grafted to a rhizoid of the latter, which contains the nucleus. The 
regenerating "hat" resembles that of A. wettsteinii and thus seems to be determined 
by the nucleus. ( After Hiimmerling. ) 

Acetabularia. In a few cases more direct proof of gene control over 
form characters has been obtained. One of the most notable of these is 
presented by the coenocytic marine green alga Acetabularia (p. 136). This 
plant has a branching, rhizoidal base from which rises a stalk several 
centimeters high, surmounted by a "hat" something like the "umbrella" 
of a mushroom. At most stages of its life history, Acetabularia has but a 
single, large nucleus, located in one of the basal rhizoids. Two species 
of the genus differ in size and especially in the form of the hat. If a 
long piece of the stalk of the taller species ( A. mediterranea ) is cut out 
and grafted to a decapitated basal portion of the other (A. wettsteinii), 
which contains the nucleus, a new hat will be regenerated at the apex 
of the grafted stalk. At first this hat will resemble that of the species 
which contributed the stalk but at length it comes to be like that of the 

428 Morpho genetic Factors 

species at the base and thus like the one to which the nucleus belongs. 
Evidently the nucleus, with its genes, determines the form of the hat 
(Fig. 19-6). The delay in expressing this determination is thought to be 
due to the persistence for a time of specific substances in the cytoplasm 
of the stalk, produced previously by the nucleus of the species from which 
it came. Acetabularia has provided material for many experiments im- 
portant for genetics and physiology as well as for morphogenesis, most 
of them by Hammerling ( see, for example, 1946 ) . 

Genes and Sex Structures. A particularly complex problem in genetics 
and one of much importance in morphogenesis is the inheritance of sex 
and the determination of the structures in which sexual differences are 

In animals most individuals definitely belong to one sex or the other, 
and the production of both male and female gametes by the same or- 
ganism is rare. In plants, however, it is much commoner than the uni- 
sexual condition. Among higher forms, staminate and pistillate flowers 
may be separate but on the same individual (the monoecious condition) 
or the flowers may be perfect and hermaphroditic. There are a consider- 
able number of cases, however, where the two types of sexual structures 
are borne on different individuals (the dioecious condition). Such forms 
are strictly comparable to unisexual animals. 

In determining just what the sexual character of a plant will be, how- 
ever, the environment has a much greater effect than it does in animals. 
Sex reversal or the production of one sort of sexual structures rather 
than another due to physiological or environmental changes is rare in ani- 
mals but in plants this is relatively easy to accomplish. In monoecious 
angiosperms the ratio of staminate and pistillate flowers to each other 
or to the perfect flowers which sometimes occur on such plants may be 
determined by light or by growth substances, or it may be an aspect 
of the general phasic development of the plant (as in cucurbit flowers; 
Nitsch, Kurtz, Liverman, and Went, 1952 ) . The problem of sex expression 
in plants is therefore in large measure a developmental rather than a 
genetic one. Although there is a definite genetic basis for most of the 
sexual differences, this wide variability in its expression makes genetic 
analysis particularly difficult. There is a large literature on this subject 
only a very small part of which can be mentioned here. 

With the rediscovery of Mendel's law there were many attempts to 
analyze sexual differences in dioecious plants in mendelian terms. Among 
the early workers were Correns (1907), Bateson (1909), and Shull 
(1910). The plants on which they first worked were Melandrium (Lych- 
nis) and Bryonia. Correns concluded from his experiments that in Melan- 
drium the egg cells all carry a tendency toward femaleness and that the 
male is heterozygous for sex, half the male gametes bearing female 

Genetic Factors 


tendency and half male. This plant could therefore be regarded as hav- 
ing the XY type of sex inheritance and to be comparable to Drosophila. 
The later discovery by Blackburn (1923) of an unequal pair of chromo- 
somes in Melandrium strongly supported this conclusion, and it is now 
generally accepted. It has been strengthened by the fact that definitely 
sex-linked traits have been found here, notably a difference in leaf shape 
(Fig. 19-7). 

Sex chromosomes are not confined to vascular plants. In the liverwort 
Sphaerocarpos Donnellii Allen (1919) reported that the four spores of 
each tetrad produce two male and two female gametophytes. The females 
have a very large chromosome, apparently the X, and the males its much 
smaller homolog, apparently the Y ( Fig. 19-8 ) . This is the XY type of 

Fig. 19-7. A sex-linked trait in Melandrium. At left, normal plant. At right, a narrow- 
leaved mutant, the gene for which is located in the X chromosome. ( From Shull. ) 

sex determination but expressed in the gametophyte generation. A good 
many other dioecious liverworts and mosses have been found to possess 
a similar pair of sex chromosomes. It is significant that when such 
gametophytes are made diploid they become monoecious, evidently be- 
cause they now possess both types of chromosomes, though in such 
cases the gametes usually fail to function. The genetics of bryophytes has 
been reviewed by Allen ( 1935, 1945 ) . 

There are a number of complications in the chromosome theory of sex 
determination in the higher plants, however. Bryonia, which, like 
Melandrium, is clearly male-heterozygous on breeding evidence, has no 
unequal chromosome pair, and it turns out that visibly unequal chromo- 
somes, presumably sex chromosomes, are present in only about half 
the genera of dioecious plants. In some cases, also, like Dioscorea, the 
female is XX and the male XO, with one chromosome less than the female. 

430 Morphogenetic Factors 

In others, like the dioecious species of the strawberry, Fragaria, it is the 
female that is the heterozygous sex and all male gametes are alike. Other 
difficulties appear. In Rumex, for example, the female has two X chromo- 
somes and the male one, but the male has two different Y chromosomes. 
In Humulus lupulus, hops, the female apparently has two pairs of dif- 
ferent X chromosomes and the male has one of each of these plus two 
different Y chromosomes. 

The situation is so complex and the results reported often so conflict- 
ing that some botanists, among them Schaffner and Yampolsky, have 
entirely repudiated the chromosome theory of sex, particularly since in 
some cases, notably in Cannabis, sex can readily be reversed by various 
environmental factors, as Schaffner was able to do by altering the 

Fig. 19-8. Chromosomes of Sphaerocarpos. From female gametophyte, above, showing 
X chromosome, and from male, below, showing Y. ( From C. E. Allen. ) 

photoperiod (p. 317). McPhee (1924) obtained similar results. He showed 
that this does not invalidate the genetic basis for sex but simply demon- 
strates that the range of expression for the genotype in hemp in response 
to the environment is very wide. 

The early ideas that two X chromosomes produce a female and one X 
a male are clearly too simple. The modern view of sex determination 
conceives of a balance between several, probably many, genes of which 
some are in the so-called sex chromosomes and others may be in the 
autosomes. This theory of balance is well shown by the results of several 
workers (Warmke and Blakeslee, 1940; Westergaard, 1940; and Warmke, 
1946) with a dioecious race of Melandrium which had been made tetra- 
ploid by colchicine treatment. Here the female had four X chromosomes 
and four sets of autosomes, 4A + XXXX, and the male 4A -f- XXYY. By 
crosses among these and with diploids, the investigators were able to pro- 

Genetic Factors 431 

duce types with one or with two Y chromosomes combined with two, 
three, or four X's, the number of autosomes being kept the same in each 
combination. The Y was thus the only variable. They also combined one, 
two, three, and four X chromosomes with one Y, again with the same 
number of autosomes, so that the number of X chromosomes was the 
variable. The results as to male, female, and hermaphrodite flowers led 
to the conclusion that the autosomes have little sexual tendency ( as they 
do have in Drosophila), that the Y contains strong male-determining 
genes, and the X, weaker female-determining ones. The ratio of X to Y 
chromosomes determines what the sex will be. It can also be shown 
that there are at least three genes for maleness in the Y chromosome. 

Experiments like these support the hypothesis that genes with tenden- 
cies to produce male structures or female ones are carried by the X, the 
Y, or the autosomes and that the balance between them determines the 
particular expression of sex. These genes may be present in chromosomes 
where there is no morphological difference between members of the pair, 
and such difference is obviously not significant for sex determination. The 
sharp distinctions between maleness and femaleness in animals and the 
relative scarcity of intersexual forms among them make genetic analysis 
of sex much less difficult there than in plants. 

Aside from the determination of sex in individuals as a whole, the de- 
gree of sex expression within the individual has also been shown to be 
under gene control. Typical illustrations of this are the following: In 
several cucurbits the monoecious condition is dominant over the andro- 
monoecious (where some flowers are hermaphroditic and some stami- 
nate), and a single gene is involved (Rosa, 1928). Poole and Grimball 
(1939) extended this for Cucumis by demonstrating a two-factor dif- 
ference between hermaphroditic and monoecious, since a cross between 
these types gave in F 2 nine monoecious: three gynomonoecious: three 
andromonoecious: one hermaphroditic. In Carica, Hofmeyr (1938) re- 
ports that three alleles, M l5 M 2 , and m, are responsible for maleness, 
hermaphroditism, and femaleness, respectively (M^m is staminate; M<jn, 
hermaphroditic; and mm, pistillate). Homozygous Mi or M> are lethal. 
Janick and Stevenson (1955) find that the monoecious character in 
spinach is allelic to the XY pair of genes concerned in sex determination. 

A notable example of the effect of genes on sex determination which 
could lead directly to the production of a dioecious condition from a 
monoecious one was demonstrated by Jones (1934) with maize. In this 
plant there are several recessive genes for tassel seed (ts) on chromo- 
some 1, in which the staminate florets are replaced by pistillate ones 
and the plant is thus essentially female. The recessive silkless gene (sk) 
in chromosome 2 sterilizes the female flowers and thus produces a plant 
essentially male. The silkless gene has no effect in the presence of tassel 

432 Morpho genetic Factors 

seed and the double recessive sksk tsts is thus female. If this is crossed 
with a plant sksk Tsts, which is male, half the offspring are Tsts (male) 
and half tsts (female). Interbreeding such plants, which are incapable 
of self-fertilization, will continue to produce offspring of which half are 
staminate and half pistillate, so that these plants, if prevented from 
crossing with other types of maize, will constitute a dioecious race. In this 
case chromosome 1, on which tassel seed is located, functions as a sex 
chromosome although no morphological difference is visible. 

A detailed review of the genetic basis for sex expression in flowering 
plants has been written by Allen (1940). 

Genes and Growth Substances. Since growth substances so powerfully 
affect growth and development it is natural to expect that in many cases 
gene action will involve the production and distribution of these sub- 
stances, and in a considerable number of cases this has proved to be the 
case. Thus in the profound changes in growth habit of "lazy" maize 
(p. 390) the character is due to the fact that auxin, instead of accumu- 
lating on the lower side of a horizontal stem, remains evenly distributed 
so that the stem does not turn upward. Its failure to do so is not the re- 
sult of mechanical weakness but of abnormal auxin relations. 

Mention has already been made (p. 264) of the single-gene difference 
between the annual and the biennial varieties of Hyoscyamus niger and 
the fact that this is apparently due to a growth substance which can 
be transmitted by grafting, to make the biennial form flower in its first 
season. It is probable that a growth substance may also be operative in 
other similar cases, like those of beets and white sweet clover, where the 
difference between annual and biennial forms has been shown to be due 
to a single gene. In the single-gene mutants reported by Stein (p. 265) 
which were grafted to normal stock, some of the effects of the gene seemed 
able to cross the graft union but others did not. 

A particularly interesting case is reported by Scheibe ( 1956 ) for peas, 
where a recessive fasciated mutant, differing from normal by a single 
gene, has a higher concentration of natural auxin than the normal. 
Furthermore, fasciation can be produced in the normal type by appli- 
cation of indoleacetic or naphthaleneacetic acid. Here the difference be- 
tween the two genes seems to be in their ability to produce a growth 

Genes are also concerned with photoperiodic response. The Mammoth 
mutant of tobacco differs from most strains by a single gene which, 
among other effects, has changed the normal day-neutral type to one 
that flowers only under short day-lengths. Chandraratna ( 1955 ) has 
shown that, in rice, races sensitive to photoperiod differ from day-neutral 
types by a single gene. 

Goodwin ( 1944 ) crossed several races of a short-day species of Solidago 

Genetic Factors 433 

differing in flowering time and found that a considerable number of 
genes were involved in the determination of this character, probably 
distributed among all nine chromosome pairs. 

Since developmental traits involve reaction of the genetic constitu- 
tion of the organism to various other environmental factors— water, tem- 
perature, light, mechanical factors, and others— it is obvious that genes 
or their combinations must take part in these reactions, and although 
there are not yet many cases in which the action of individual genes 
has yet been analyzed as successfully here as it has in the biochemical 
genetics of lower organisms, this will doubtless be accomplished and 
will give information of value for the solution of the problems of de- 
velopment. This knowledge will be second in significance only to that 
which may be gained as to the mechanisms by which these innumerable 
gene reactions are so organized in the growth and activity of the indi- 
vidual that it becomes an organism. 


The fact that an entire and normal plant may be produced by regenera- 
tion from a single cell (p. 253) or a group of similar cells in different 
regions and from different tissues is evidence that all the cells of the plant 
are genetically alike. This implies that all the nuclei are similar, a conclu- 
sion supported by the fact that in all cells (save in cases of polysomaty) 
the number and structure of the chromosomes are constant. 

If this conclusion is valid, the basis of differentiation would appear to 
lie in the extranuclear portion of the cell, the cytoplasm. Much less is 
known about the cytoplasm than about the nucleus, and events in it are 
not as dramatic and easily observed. It is clear, however, that beneath 
its relatively homogeneous superficial appearance there must be a high 
degree of chemical and physical diversity, an understanding of which 
is necessary before the problems of differentiation can be mastered. The 
basis for cytoplasmic differentiation is doubtless at the submicroscopic 
level. Weiss (1956), Schmitt (1956), and some others have shown that 
the elements of the macromolecular pattern are markedly different from 
one another in size and distribution (see also Tartar, p. 455). A wide field 
of research at this level is now developing. 

There are a few cases where the cytoplasm can be shown to be im- 
mediately concerned in determination of structural characters. This is 
particularly true of the cell wall, which is directly produced by the 
cytoplasm and is a primary element in many differences between cells. 
Examples of this are seen in the regenerative conversion of thin-walled 
parenchyma cells into xylem cells with reticulate lignified walls. The pat- 
tern of these lignified thickenings is preceded in development by an 

434 Morphogenetic Factors 

identical pattern of granular cytoplasmic strands on which the thickenings 
are laid down (p. 193). Other markings in the cell wall have also been 
traced to cytoplasmic differences. 

The cytoplasm may be concerned in the development of characters 
above the cellular level. In the cells of air roots of orchids, the bands 
of wall thickening ( p. 201 ) that keep these cells from collapsing are laid 
down by the cytoplasm in each cell. They are not isolated structures in 
single cells, however, but a thickening in one cell is directly adjacent to 
one in the next, so that a continuous system is produced extending from 
cell to cell and forming a histological pattern over a considerable mass of 
tissue. The differential fiber patterns in Luff a (p. 197) and similar cases 
doubtless originate in the cytoplasm. 

The distribution and configuration of the cytoplasm probably have a 
more deeply seated relation to development, however, than in these ex- 
amples of cellular patterns. The plane in which a cell divides, at least 
in vacuolate cells, is foreshadowed by the orientation of a cytoplasmic 
plate some time before the axis of the spindle is established (p. 25), a 
fact which suggests that cell polarity, and thus the direction of growth 
and ultimately organic form itself, may have their immediate basis in the 
distribution and patterning of the cytoplasmic body. 

The relation of cytoplasm to such differences in form may be well 
seen in the coenocytic bodies of many red algae, notably the genus 
Caulerpa. Here there are no cellular barriers to the passage of cytoplasm 
from one part of the plant to another, and much streaming takes place. 
Differences may be observed in various parts of the plant body as to the 
character of the cytoplasm, part of which is fixed to the inner wall and 
does not stream. In such plants the differential distribution of the cyto- 
plasm seems to be related to the differentiation of the plant body, though 
the mechanisms involved are unknown. The disadvantage in such or- 
ganisms and the probable reason why they have never been able to de- 
velop very highly differentiated bodies is their difficulty in keeping the 
various components of their living material sufficiently isolated so that 
physiological differences can be maintained effectively and a high degree 
of organization thus made possible. 

Even in multicellular plants visible differences in distribution of cyto- 
plasm are related to differentiation. This is especially evident in cases of 
unequal cell division, as in the formation of trichoblasts in many roots 
(p. 190). In the mother cell, which is to divide unequally to form a 
small trichoblast and a larger hairless cell, much of the cytoplasm (and 
the nucleus) moves toward the end at which the trichoblast will be cut 
off, so that even before division there is a difference in cytoplasmic dis- 
tribution. The formation of the new wall finally separates two regions 
which had already become cytoplasmically different. In the formation 

Genetic Factors 435 

of stomatal mother cells, initials of trichosclereids, and many other ex- 
amples of unequal cell division the same differential distribution of cyto- 
plasm, before division, is evident. 

In such cases, unlike the coenocytes, the differences which arise in the 
cytoplasm cannot become distributed beyond the limits of the original 
mother cell. If in each cell division, however, there were a quantitatively 
or qualitatively unequal distribution of cytoplasm between the two 
daughter cells, cellular differentiation would result. It seems reasonable 
to suppose that many cell divisions are thus cytoplasmically unequal 
even though the differences are not visible and may be at the submicro- 
scopic or chemical level. Such inequality would provide the necessary 
cytoplasmic basis for differentiation. 

If differentiation proves to be primarily a matter of cytoplasmic distribu- 
tion, the mechanism by which this distribution is controlled must evi- 
dently be one of the major problems of morphogenesis. In some cases 
the cause may be ascribed to polarity. Certainly polarity is involved, as 
we have seen, in many axes other than the major one of the plant body 
and is evident in many developmental patterns. It may be that the mecha- 
nisms which are effective in the extreme and conspicuous cases of un- 
equal and polar cell division may also be involved in all differential di- 
visions (Bunning, 1958). 

That the cytoplasm contributes to the determination of developmental 
processes through inheritance is clear from a number of facts, particularly 
in cases where the offspring of reciprocal crosses are unlike. Where the 
offspring tends to resemble the maternal parent this difference is evi- 
dently due to that which only this parent contributes to it, the cytoplasm. 
In traits where plastid differences are involved the influence of the 
cytoplasm is clear, since the plastid primordia are carried in it. In other 
cases it is more difficult to see what the mechanism of transmission is. 

Crosses in Epilobium, where reciprocal hybrids are often markedly dif- 
ferent in size, have been studied intensively. Lehmann (1936) showed 
that in such cases the smaller hybrids have a lower concentration of auxin 
than the larger ones. When Schlenker and Mittmann (1936) applied 
auxin to the smaller plants their size was considerably increased. These 
facts suggest that something carried in the cytoplasm stimulates the 
synthesis of auxin. It may be, as has sometimes been suggested, that 
sensitivity to auxin is determined by the genes but that auxin synthesis 
is carried on in the cytoplasm. Michaelis (1938) disagrees with Lehmann 
and believes that the facts can be explained by specific interactions be- 
tween genes and cytoplasm. This Epilobium work, however, and the 
great body of evidence obtained by von Wettstein (1924) and his col- 
leagues in experiments with mosses show that the genes are not inde- 
pendent in their effects but that what they do is determined to a con- 

436 Morphogenetic Factors 

siderable degree by the cytoplasm with which they are associated. 
Whether the specificity of the cytoplasm results from self-perpetuating 
bodies such as the often postulated plasmagenes or from persisting 
effects of genes on the cytoplasm (Mather, 1948) has not been de- 
termined. This question is primarily of genetic rather than of morpho- 
genetic importance. 

When a knowledge of the cytoplasm is more complete, that part of the 
cell, somewhat neglected by genetic investigations in the past, will doubt- 
less contribute much more significantly to our understanding of develop- 
ment and differentiation. What has been called protoplasmatic plant 
anatomy is concerned with some of these problems. Its contributions have 
been summarized by Reuter ( 1955 ) . 


The control of development and form lies chiefly with the genes and 
their reaction to the environment, but it must also be recognized that 
differences in the number and character of the chromosomes, apart from 
the genes they contain, may be of considerable morphogenetic signifi- 

Polyploidy. Most plants in the sporophyte generation are diploid, the 
cells containing two sets of chromosomes, one coming from the male 
parent and one from the female, each chromosome belonging to a pair 
of homologous ones. In some plants, however, the number of sets has 
been multiplied so that every chromosome is represented by more than 
two homologs. Such plants are polyploids. There are many cases where 
the number of sets is doubled, to form tetraploids. Hexaploids, octoploids, 
and many other polyploid types are known, though polyploidy cannot 
be increased indefinitely throughout the plant because of loss of vigor in 
higher members of the series. Individual cells, however, or groups of 
cells may become very highly polyploid. 

Polyploids are often found in nature, many species belonging to so- 
called polyploid series where each species has a particular multiple of a 
basic number of chromosomes. 

Various ways of producing tetraploids artificially are known, and many 
polyploids used in experimental work have arisen in this way. One 
effective means is treatment with colchicine or certain other chemicals 
which check mitosis after chromosome division but before the new 
nuclear membrane is formed, so that the two daughter cells have the 
double chromosome number. Colchicine may be applied to seeds or to 
the whole plant. Growth after the latter treatment, as compared with 
normal development, has been described for cranberry by Dermen 
(1944). Many large cells in normal plants are polyploid, and a bud 

Genetic Factors 437 

developing in such a tissue will be a tetraploid or higher (Jorgensen, 
1928). In mosses and ferns gametophytes may be regenerated from 
diploid tissue under favorable conditions (p. 234) and are thus diploid 
in character. From them tetraploid sporophytes may arise. Haploid 
sporophytes in higher plants have been produced by various chemical 
and physical treatments and sometimes occur in twin seedlings ( Christen- 
sen and Bamford, 1943). These cases prove that the differentiation into 
sporophyte and gametophyte does not result simply from difference in 
chromosome number. 

The primary effect of a multiplication of chromosome number is an 
increase in the volume of the nucleus and the cell. Most other distinctive 
traits of polyploids follow from this one. The relationship between num- 
ber of chromosome sets and cell size is not always a simple proportional- 
ity, however. A study of such a series as that in Datura stramonium (Sin- 
nott, Houghtaling, and Blakeslee, 1934), for example, where In, 2n, 3n, 
and 4n plants can be compared directly, shows that the increase in cell 
size is different in different tissues. In epidermal cells it is not far from 
1:2:4:8. In xylem cells the increase is a little greater, but in the paren- 
chymatous cells of the fundamental tissue in the petiole, the tetraploid 
is usually much more than eight times the diploid. 

In such cases, each added chromosome complement does not simply 
add an amount proportional to the increase in chromosome number but 
multiplies cell size by a certain amount. In other words, the addition is 
geometric rather than arithmetic. This is evident to some degree in the 
pedicel cells of Datura but particularly in the large ones of the petiole. 
It is well seen in mosses, where von Wettstein ( 1924 ) produced diploid 
gametophytes by regeneration of protonemata from sporophyte (2n) 
tissue so that haploid and diploid gametophytes could be directly com- 
pared. When this was done, Tobler (1931) found that the increase in 
cell size of diploid over haploid was different for different races and 
that in crosses between them it was a character which seemed to segregate. 
The effect of polyploidy may be different in related species ( H. H. Smith, 

In some polyploid series, like that reported by Harriet E. Smith (1946) 
for races of Sedum pulchellum with two, four, and six chromosome sets, 
cell size increased with number of sets. In many cases, however, mem- 
bers of such a polyploid series in nature do not differ appreciably in cell 
size. An observation of von Wettstein's (1938) may indicate the reason 
for this. He grew a diploid race of Bryum, which he named Bryum 
corrensii, from a regenerated diploid protonema. It had large leaves and 
cells about twice the haploid size and was quite sterile. Under vege- 
tative propagation its size gradually became reduced until after 11 
years it had returned to a practically normal condition as to leaf and 

438 Morpho genetic Factors 

cell size and fertility, although its chromosome number was still diploid. 
This suggests that regulation to a physiologically optimum cell size had 
taken place. In polyploid series in the higher plants it may be that regu- 
lation, through natural selection or otherwise, has produced an optimum 
cell size even with widely different numbers of chromosomes. In poly- 
ploid series of recent origin, however, such as those produced by colchi- 








Fig. 19-9. Flower of diploid (left) and tetraploid (right) in Antirrhinum majus. 
( Courtesy W. Atlee Burpee Co. ) 

cine, there is almost always a close relationship between chromosome 
number and cell size. 

The increased cell size of polyploids may be reflected in larger plant 
size (Fig. 19-9). Oenothera gigas, which proved to be a tetraploid, was 
named for its size, and "gigas" tetraploids of many species are now known. 
Frequently, however, the tetraploid plant is little larger than the diploid, 
and it may even be smaller. It usually has stouter stems and thicker 

Genetic Factors 


leaves, and its flowers are larger. Haploid plants are universally smaller 
than diploid ones. Triploids may be intermediate between tetraploids and 
diploids but are often indistinguishable from the latter. 

Genetic differences may sometimes determine the effect of polyploidy 
on plant size. Flax ( Linum usitatissimum ) has been selected commercially 
in two directions, toward the production of linseed oil and of flax fiber. 
Pandey (1956) compared the tetraploid with the diploid forms of both 
types of plants in this species and found that in the linseed type the An 
is a gigas form whereas in the flax type it is actually smaller than the 2n. 
The linseed tetraploid grows faster than its diploid but the flax tetraploid 
grows more slowly. The two tetraploids also show certain morphological 

APEX 10 

LEAF 10 

APEX 10 

LEAF 10 

Q 10 mm. 

Fig. 19-10. Shoot apex and leaf primordium in Zea mays. Left, diploid; right, tetra- 
ploid. Difference in size is due entirely to larger cells of the tetraploid. (From Ran- 
dolph, Abbe, and Einset. ) 

The origin of size differences in members of a polyploid series has 
been studied developmentally in a few cases. The apical meristem of the 
tetraploid is always broader than that of the diploid. Sometimes it is 
relatively flat, as in Vinca rosea (Cross and Johnson, 1941), or it may 
be the same shape as the diploid but doubled in size, as in maize 
(Randolph, Abbe, and Einset, 1944; Fig. 19-10). In the development 
of the cucurbit fruit, Sinnott and Franklin (1943) found that the ovary, 
from primordium to the time of flowering, was almost twice the volume 
in An as in 2n and thus was proportional to the volume of its cells (Fig. 
19-11). In these plants, growth of the ovary into the fruit is chiefly by 
cell expansion. This second phase of growth is much less extensive in An 
than in In, so that the mature fruit is almost the same size in both, as 
are the cells of which it is composed. In these cases the flower is "gigas" 
but the fruit is not. 

440 Morpho genetic Factors 

Members of a polyploid series often differ in other respects than size. 
There is a general tendency in tetraploids for organs to be relatively 
shorter and wider than in diploids. This difference is well shown in the 
series of capsules in Datura from In to An (Blakeslee, 1934). The 
haploid has a slender capsule and becomes progressively flatter in the 
upper members of the series. Fruits of tetraploid cucurbits produced by 
colchicine were in every case changed toward a flatter, or at least a less 
elongate, form. The leaves of tetraploid varieties of most plants show 
the same shape changes in comparison with the diploid. Straub (1940) 
observed that flower size changes in the same way, flowers from higher 
members of a polyploid series being relatively wider. In Torenia the 
position of the anthers with reference to the corolla is changed in the 


60 ao ioo 

8 10 20 

Fig. 19-11. Graph showing general relations between cell size and ovary size in de- 
veloping fruits of diploid and tetraploid cucurbits. In early stages the 4N is larger in 
both respects but after flowering the growth of the 2N is greater and at fruit ma- 
turity (vertical bar) the two are essentially alike. (From Sinnott and Franklin.) 

This shape difference probably originates at the meristem itself. 
Cross and Johnson found that in Vinca rosea the tetraploid apex was con- 
siderably wider but no deeper and that the increase in size of its com- 
ponent cells was also chiefly in width. Riidiger (1952) has shown that 
in various plants the cells of the tetraploid are somewhat shorter and 
wider than those of the diploid. Organ shape may thus be a reflection 
of cell shape, though in other cases cell shape seems not to be markedly 
different in An and 2n. Why there should be such a shape difference, 
either in cell or organ, is not clear. Cell size alone is not enough to ac- 
count for it, for there are large-celled races which do not differ in 
shape from small-celled ones of the same species. 

Certain more general biological facts are related to polyploidy. Steb- 
bins (1938), for example, has studied polyploidy in a large number of 
woody and herbaceous genera and finds that polyploid series are more 

Genetic Factors 


abundant in perennial herbs than in annuals or woody plants (Fig. 
19-12). The basic chromosome numbers, however, are significantly 
higher in woody genera. 

Much work on polyploidy has been done by students of ecology and 
plant distribution ( Miintzing, 1936 ) , though most of this has little direct 
concern with morphogenesis. In general, polyploids can endure extremes 
of climate better than diploids. The distribution of polyploids as to 
latitude has been discussed and the literature reviewed by Love and Love 

Polyploids are of importance for evolution not only in matters of selec- 
tion and distribution but from the fact that by their means sterile 
hybrids can become fertile and genetic lines, separated by incom- 








nln hi 





5 & 7 8 9 


Fig. 19-12. Distribution of basic chromosome numbers in herbaceous and woody 
genera of dicotyledons. Dotted lines indicate genera known only tentatively. Herba- 
ceous types tend to have smaller numbers. ( From Stebbins. ) 

patibilities, thus be brought together. Many species will hybridize but 
the offspring are usually sterile. If their chromosome number is doubled 
(to form an allopolyploid), as sometimes happens in nature, fertility 
is restored since there are now pairs of homologous chromosomes and 
normal meiosis can take place. 

Somatic Polyploidy ( Polysomaty ) . Polyploidy is concerned with an 
important aspect of differentiation, for investigation has shown that many 
somatic cells, particularly the larger or physiologically more active 
ones, are polyploid and often to a rather high degree. This condition is 
termed polysomaty. In some cases, such as the meristematic region of 
the root of spinach (Gentcheff and Gustafsson, 1939; Berger, 1941), 
chromosomes of certain of the dividing cells, particularly in the periblem, 
are twice, or sometimes four times, the normal diploid number. This seems 
to be the result of an additional doubling in the prophase before the 

442 Morphogenetic Factors 

mitotic figure is formed. Polysomaty of this type has been reported in 
other cases, as in Cucumis (Ervin, 1941). 

In many instances it has been shown that a process of endomitosis 
takes place in certain cells by which the chromosome number is doubled 
(or further multiplied) even though the cell is mature, the nuclear 
membrane intact, and no mitotic figure has been formed (Geitler, 1949). 
That this doubling has taken place is indicated by the fact that the num- 
ber of visible chromocenters, presumably corresponding to the chromo- 
somes, is doubled. The chromosome number may be definitely determined 
by inducing these cells to go into typical mitosis, either through wound- 
ing or by application of growth substances. Under such conditions the 
chromosomes can readily be counted. Grafl (1939) was thus able to 
prove that among mature and normally differentiated cells of Sauromatum 
guttatum some were tetraploid, some octaploid, and some 16-ploid. This 
situation has now been found in many other cases. D'Amato (1950), by 
the use of 2,4-D, observed it in roots of a number of monocotyledons, 
and Holzer (1952), by treatment with indoleacetic acid, in the roots of 
27 species of angiosperms. Holzer found that the distribution of these 
polyploid cells was not at random but formed a pattern which was similar 
in groups of related plants. Often it is not the number of chromosomes 
(or chromosome centers) that increases but the number of strands per 

The connection between polyploidy and the volume of cell and 
nucleus has important morphogenetic implications. In certain animal 
tissues, among them the cells of the developing salivary glands of 
Drosophila, it has been observed that nuclear volume falls into definite 
classes, each approximately twice the volume of the one next below it. 
Such "rhythmic" distribution was observed in plants by Monschau ( 1930). 
The relation between polysomaty and nuclear volume has now become 
well established (Bradley, 1954, and others). This makes it possible to 
determine with some accuracy the degree of polyploidy in a mature and 
differentiated cell by measuring the volume of its nucleus or even of the 
cell itself, in comparison with related cells and ultimately with those in 
which the chromosome number can be determined directly. Thus in 
various angiosperm species, Tschermak-Woess and Hasitschka (1953«) 
have estimated the degree of polyploidy of certain cells in a tissue as 
various multiples of the basic number, up to 256-ploid (Fig. 19-13). 
Somatic polyploidy, as estimated either by direct count or by nuclear 
size, commonly increases with distance from the apical meristem (Fig. 
19-14) and with age (Fig. 19-15). 

All this obviously has a Very important bearing on the problem of 
differentiation. Wipf and Cooper (1940) for example, found a close rela- 
tion between the presence of naturally occurring tetraploid cells and the 

Fig. 19-13. Cereus spachianus. Cells showing increasing chromosome numbers during 
development, a and b, 2n ; c, 32n; d, 64n; e, highly polyploid resting nucleus from an 
older portion of the cortex; /, young polyploid resting nucleus with chromocenters, 
2 to 3 mm. behind the shoot apex; g, a part of the differentiated cortex showing marked 
differences in cell size, presumably because of somatic polyploidy. ( From Fenzl and 
Tschermak-Woess. ) 





100 lM MO 230 300 350 400 450 500 550 600 630 700 750 870 830 600 050 1000 1050 1100 1130 l!00 1250 ~I300 


Fig. 19-14. Frequency distribution of mitoses with one, two, or four times the basic 
number of chromosomes, in the root of Cucumis. Chromosome number increases with 
distance from the apex ( left. ) ( From Ervin. ) 



























14 20 



Fig. 19-15. Changes in nuclear volume (cubic 
microns ) during development of maize endo- 
sperm. C, central region. AB and AD, surface 
layers. Nuclear growth here is due to endomi- 
tosis, with increase in number and size of 
strands per chromosome. Number of chromo- 
somes presumably does not increase. ( From 
Duncan and Ross. ) 


Fig. 19-16. Endopolyploidy in tri- 
chomes of Bryonia, a, glandular 
hair with basal cell 256-ploid; 
b, hair of filament, the basal cell 
128-ploid, and others 16-ploid. The 
epidermis is diploid. ( From Tscher- 
mak-Woess and Hasitschka. ) 

Genetic Factors 445 

origin of root nodules in legumes. Von Witsch and Fliigel (1952) found 
that meristematic tissue of Kalanchoe Blossfeldiana was diploid but that 
as the leaves differentiated the mesophyll became polyploid, as could 
be shown by wounding and observing mitoses in the wound callus. Inci- 
dentally, the mesophyll was only 8-ploid if the leaf had developed under 
long days but 32-ploid under short-day conditions. Steffen (1956) esti- 
mated from nuclear volume that chalazal haustoria in Pedicularis were 
96-ploid and micropylar ones 384-ploid. An illuminating study of endo- 
mitotic polyploidy in the differentiation of the trichomes of angiosperms 
has been made by Tschermak-Woess and Hasitschka (1954; Fig. 19-16). 

Somatic polyploidy has now been shown to be so frequent as to make it 
probable that much of the differentiation of plant cells, so far as cell size 
is concerned, is related to it, cell size being roughly proportional to the 
degree of polyploidy. Exceptions have been found to this relationship, 
and no firm generalization about it can yet be made. Obviously, too, 
there is much more to differentiation than change in size. However, the 
possibility certainly exists of learning much about the mechanism of 
cellular differentiation and of developing what has been called "karyologi- 
cal plant anatomy," a subject outlined in some detail by Tschermak-Woess 
(1956). D'Amato (1952) has also reviewed the field of polyploidy in 
differentiation, and Geitler's book (1953) discusses endomitosis more 

Other Effects of Chromosome Differences. There is sometimes a rela- 
tion between cell size and total chromosome bulk. In 13 species of Crepis 
differing in number and length of chromosomes Navashin ( 1931 ) meas- 
ured total chromosome length at a comparable stage of mitosis, using 
this as an indication of chromatin mass. In each species he plotted this 
against the volume of comparable cells in the root meristem and found 
a close correlation between the two, suggesting that the total bulk of 
chromosome material affects the size of the cell (Fig. 19-17). In some 
cases, notably in mutants of Primula and Phragmites (p. 35), both 
chromosomes and cells are markedly larger than normal, and this is also 
reflected in the size of the plant itself. Neither of these types is polyploid. 
In some plants there are "accessory" chromosomes which seem to have 
little or no genetic effect, but Miintzing and Akdik (1948) find that 
their presence causes an increase in the size of stomatal cells. 

The influence of extra chromosomes was studied in Crepis tectorum by 
Schkwarnikow ( 1934 ) . This species has four pairs of chromosomes called 
A, B, C, and D, and four races were available in each of which one of 
these chromosomes was represented by three instead of two members. 
Plants in which B or C was present as a trisome had cells larger than the 
normal diploid. Those in which A or D were the extra ones, however, 
had smaller cells than normal. Here evidently something more than bulk 


Morpho genetic Factors 



1 i 

'Z 4000- 











* 3000- 





■ i i i i i 

i i 

i i i 

i i i i 

Chromosome length (relative units) 

Fig. 19-17. Correlation between average cell volume (of dermatogen cells in root tip) 
and average total chromosome length (chromatin mass) in 13 species of Crepis. The 
length of bar for each species indicates the probable error for that determination. 
( Redrawn from Navashin. ) 

of chromosomes was concerned with cell size. This presumably was de- 
termined by the genes the chromosomes carried (Table 19-1). 

Table 19-1. Average Cross-sectional Area (in Square Microns) of Primary 

Dermatogen Cells in Root of Diploid Crepis tectorum and in 

Four of Its Trisomies * 

Diploid 260.0 ± 5.3 

Trisomic A 235.5 ± 3.6 

Trisomic B 282.6 ± 4.0 

Trisomic C 271.0 ± 2.4 

Trisomic D 228.5 ± 5.5 

° From Schkwarnikow (1934). 

A somewhat similar result was found by Sinnott, Houghtaling, and 
Blakeslee (1934) in the primary chromosome mutants of Datura stra- 
monium. This species has 12 pairs of chromosomes, and in each mutant 
race there are three representatives of one of these chromosomes. The 12 
trisomic races differed very considerably from one another, both ex- 

Genetic Factors 


ternally and internally. In some the cells in certain tissues were larger 
and in some smaller than in the diploid. Such traits as the size of the 
bundles, the amount of internal phloem, and the development of peri- 
cycle fibers differed among the mutants. It is significant that for each 
trait the diploid was approximately the average of the 12 mutants, a good 



I 2- 


3 3 

— 3-4 

5 6 — 



1112 — 



14 14-- 

15 16- 

17 17 — 

23 24- 


6 6 

■7 8 


9 3 





2 N 



210 2.58 3.06 3.54 4 02 451 5.00 

Fig. 19-18. Genie balance in Datura stramonium. Cross-sectional area of flower stalk 
in various chromosomal types (in square millimeters). Vertical line marks normal 
value for the diploid. Areas larger or smaller than this are shown by the lengths of the 
horizontal lines for each type. IN, 3N, and 4N are below. Above, the primary mutants 
have solid lines, the secondaries dotted ones. Each chromosome is numbered by its 
two ends, the primaries being 1.2, 3.4, etc., and the secondaries 2.2, 3.3, etc. The 
geometrical mean of the primaries is almost exactly the value for the diploid. (From 
Sinnott, Houghtaling, and Blakeslee. ) 

example of genie balance (Fig. 19-18). Evidently each chromosome 
contributed something to the character of the plant, and when this 
chromosome was represented by three instead of two, this contribution 
was increased. Such a result seems clearly to be due to the genes which 
the chromosome carries, however, rather than to something specific in 

448 Morphogenetic Factors 

the chromosomes as such. It may be that all the effects now attributed to 
chromosomes may ultimately be found to be due to their genes. 

It must regretfully be admitted that not very much of substantial impor- 
tance has yet been contributed by genetic analysis to a solution of the 
problems of morphogenesis. Its most significant addition to our knowl- 
edge of development thus far is perhaps the discovery of the relation be- 
tween the size of a cell and the degree of its internal polyploidy, with the 
bearing this fact has on the control of histological differentiation. Behind 
the other problems of morphogenesis still lurks the unanswered question 
of how genes control the development of form and structure and thus 
the orderly and integrated growth of an individual as an organism. En- 
vironmental factors have an important influence on the character of this 
organism, but the organizing process itself seems to be centered in proto- 
plasm and thus to be under the control of the directive and self-multi- 
plying elements in protoplasm, the genes. It is to this general problem of 
organization that attention in our final chapter is directed. 



Underlying the various phenomena of morphogenesis that have here been 
discussed stands a single basic problem: how a mass of living stuff is 
organized into a system, so well termed an organism. Organization is evi- 
dent in various ways but most vividly in the development of form in 
living things. Form is not simply a trait to be described and classified. 
It is also the visible expression of a self-regulatory equilibrium which 
tends to be attained in development, maintained during life, and restored 
when disturbed. Every individual has a specific equilibrium of this sort, 
a morphogenetic norm, so to speak, to which it tends to conform. This is 
the unifying factor that gives continuity to an organism. It is a pattern 
for development in which every part, in its growth and activity, is related 
to all the others and by which the fate of each is determined by its posi- 
tion in the organized whole. 

The manifestations of a given norm are various. It is not constant in 
expression but may change, in a precise and regular fashion, from embryo 
to maturity. It may produce very different results under different en- 
vironmental conditions. Its basis is established in the genetic constitu- 
tion of the individual. The nature of this norm, how it is modified in ex- 
pression by factors inside and outside the organism, and how develop- 
ment is regulated in conformity to it are the basic questions with which 
the science of morphogenesis must deal. 

To attack this problem hopefully one should break it down, if possible, 
into simpler components. These have already been discussed in earlier 
pages, but it will be useful here to bring them briefly together into focus 
on the main problem. 

One of these components is polarity. All plants, at least at certain 
stages, are organized around an axis which provides a basis along which 
development takes place and in relation to which the lateral organs are 
formed. The two poles of this axis (save in rare cases) are unlike. The 
axis may be vertical or horizontal, and if lateral axes arise from it, these 
often have specific orientations, so that the plant body is really a pattern 
of polarities. In many cases transverse as well as longitudinal ones may 


450 Morphogenetic Factors 

be seen. Polarity is manifest in physiological activities as well as in struc- 

The universal presence of polar phenomena in plants suggests that 
there is in living stuff an innate tendency toward polarization, although 
most eggs and undifferentiated cells at first are unpolarized. This provides 
a concrete point of attack on developmental problems. The suggestion is 
obvious that organic polarities may be related to those evident in the 
inorganic world, and especially to chemical and electrical ones, but this 
has proved somewhat difficult to establish experimentally. Polarity has 
been shown in some cases to be modified by specific factors in the en- 

Not only are the two ends of the polar axis unlike but other differences 
appear in graded series backward from each apex. These gradients are 
evident in the rate of various metabolic processes and in the form and 
structure of lateral organs arising successively at the growing point. Here 
is a simple place to study the origin of differentiation. One school of 
biologists regards these axial gradients as the most important factors in 

The polar axis is not only a gradient axis but an axis of symmetry. In- 
deed, the symmetrical arrangement of parts around it is a conspicuous 
aspect of axiation. Organic symmetry, an expression of the basic regu- 
larity in the arrangement of plant structures, is manifest at every level 
from the internal structure of a cell to the configuration of a tree. It is 
obviously an important element in the orderly formativeness that living 
organisms display. 

Radial symmetry— the regular spiral distribution of lateral structures 
around an axis— is best displayed in vertical plant axes where the stimuli 
of gravity and light affect all sides of the axis almost equally. Under these 
conditions, the symmetrical arrangement of the lateral organs seems to be 
traceable to an inherent spirality in living stuff itself. This has as yet 
received little experimental study, but its widespread occurrence in 
phyllotaxy, the spiral grain of wood, the spiral character of the cell wall 
and of protoplasmic streaming, and the spiral movements of the plant 
body suggest that spirality is another basic fact in organization. 

Spirality is masked to a considerable extent in those axes which are 
horizontal and thus exposed differentially to gravity and light, but the 
dorsiventral symmetry of these structures provides good material for a 
study of the interaction of polarity, spirality, and environmental factors 
in the development of plant form. 

We may therefore recognize several components in the general phe- 
nomenon of plant organization: polarity, differential gradients, symmetry, 
and spirality. These seem to be distinct characteristics and may have dif- 
ferent bases in protoplasm. They certainly can be investigated separately. 

Organization 451 

All of them, or their rudiments, seem to be present in all plants. They 
provide the basic ingredients, so to speak, out of which the developmental 
norm is produced. Just what a specific norm will be depends on the 
interaction between these inherent protoplasmic traits and two other 
factors: the genetic constitution of the individual and the environment 
in which it develops. 

The genetic constitution is the complement of genes in the organism. 
These act on the protoplasmic traits just mentioned to produce the form 
characteristic of that organism. This form, however, is not a specific pat- 
tern of polarities, gradients, spiralities, and symmetries but results from 
a specific reaction to a specific environment. Neither genes nor environ- 
ment alone determines what an organism is, for their action is comple- 
mentary and one cannot be separated from the other. In practical experi- 
mental work, however, much can be learned by studying the effects of 
different genotypes under the same environment or of the same genotype 
in different environments. The latter method has so far been much more 
fruitful, as is shown by the vast literature in the fields of the morpho- 
genetic effects of light, water, temperature, mechanical factors, and 
various chemical substances. There is still opportunity for much fruitful 
work in all these fields. A study of gene action, on the other hand, al- 
though actively pursued, has thus far been concerned chiefly with the 
effects of genes on metabolic processes or on the synthesis of specific 
substances. How genes control developmental relationships, and thus 
the production of organic form, is almost unexplored territory. 

In a given individual, therefore, through the interaction of its genotype 
and the particular environment in which it lives, both acting on the basic 
tendencies toward polarization, gradients, symmetry, and spirality, there 
is at any given stage of its development a norm to which it conforms. 
This involves more than a mere interaction between organism and en- 
vironment. What emerges from the developmental process is an organized 
system in which the various parts are related and mutually interdependent 
and which controls its own development by a process of self -regulation. 
This is to be seen most clearly in the familiar phenomena of growth. A 
plant or animal exists in an environment of which the chemical constitu- 
ents (atoms, molecules, or larger particles) are a heterogeneous mass and 
dispersed at random. When these particles are drawn into the organism 
they lose this randomness and each now comes to occupy a particular 
place in the living system. By some means this orderly disposition of new 
material into an organized whole is controlled. When death ensues, the 
control disappears and the dispersive tendencies of lifeless matter break 
down the system. This system is specific and is different in every indi- 
vidual. The mass of data now accumulated from studies of regeneration 
suggests that all the cells, at least at their beginnings, are totipotent and 

452 Morphogenetic Factors 

that the basis for the norm of the organism is present in the living stuff 
of every cell. 

The omnipresence of the developmental norm is suggested by that 
quality in development termed by Driesch equifinality, the attainment of 
the same developmental goal in very different ways. There is no single 
or linear progression of steps by which a structure is formed, but the or- 
ganism may shift its course of development according to circumstances. 
This is much more difficult to explain than is a linear step-by-step series 
of changes, each a precursor of the next. A regulatory mechanism of some 
sort must be involved. What gives unity to the individual is not so much 
its unchanging genetic constitution, important as this is, but rather this 
developmental norm, immanent in the organism from the first and often 
reached over different routes. 

It should be emphasized that what is involved here is not inherent 
adaptability by which an organism naturally reacts in a favorable way to 
environmental changes, for a given norm may hinder survival. The "lazy" 
mutation in maize, for example, which causes the plant to grow flat on 
the ground, would not persist in nature. Most normative reactions of 
plants evident today tend to be favorable since they are the ones that 
have survived in the winnowing process of natural selection, but the fact 
of normativeness has no relation to adaptability. 

The self-regulatory character of living things has often been observed 
and discussed. In bodily activities it is the basis of the homeostatic re- 
actions so evident in physiology. Its most conspicuous manifestations, 
however, are structural and are seen in those cases where a portion of an 
organism is experimentally isolated from the rest and then proceeds to 
restore its missing parts so that a single whole is produced. Each of the 
first two cells from a Fucus egg, or the first two blastomeres in a frog 
embryo, if isolated, will develop not into half an organism but into a whole 
one. Single cells from the epidermis of a plant, under appropriate condi- 
tions, will develop into whole individuals. Cuttings restore lost roots. 
Missing shoots are replaced by others. Severed vascular strands are united 
by the growth of connecting bundles. The literature of plant regeneration 
is full of such examples. Many of these were brought together by Ungerer 
( 1926 ) , but botanists have in general been less concerned with this prob- 
lem than have zoologists. Regeneration is such a dramatic fact that many 
attempts have been made to account for it. Actually, however, regenerative 
development is no more and no less difficult to explain than normal 
development. The real problem is not regeneration but self-regulatory, 
normative development. 

Many suggestions have been made to explain this central enigma of 
morphogenesis, but it must be admitted that none has yet been proposed 
which is generally acceptable. The problem is enormously difficult and 

Organization 453 

doubtless cannot be solved by any simple or single hypothesis. It is essen- 
tially one of synthesis, in which evidence from many sources must be 
coordinated. For discussions of it the reader is referred to the publications 
of Agar (1951), von Bertalanffy (1952), Child (1941), Driesch (1937), 
Holmes (1948), Lillie (1945), Meyer (1935), Needham (1936), Reinke 
(1922), E. S. Russell (1933), Smuts (1926), Troll (1928), Ungerer 
(1926), Wardlaw (1955c), Weiss (1950), Whyte (1954), and Woodger 
(1929). Woodger's discussion of the concept of organism (1930, 1931) 
is particularly useful. 

The position that a biologist assumes toward this problem will usually 
be determined more by his attitudes and predilections than by the con- 
flicting and inconclusive evidence that is now available. To those who 
assume that all organic traits must have been produced by natural selec- 
tion, both normal and regenerative development will be regarded as the 
result of a long-continued selective process. Holmes (1948) and others 
have supported this view, and it is probably held by a majority of biol- 
ogists who have considered the matter. Aside from the general presump- 
tion in its favor, there is some positive evidence for this position in the 
fact that organized, regulatory development is not invariable but some- 
times breaks down. There are many examples of this in the various types 
of abnormal growth. Sometimes, as in teratological structures, only the 
last developmental stages become confused and irregular, but when the 
breakdown is more complete, tumors, galls, and other amorphous struc- 
tures are produced. Finally, in tissue culture, all traces of multicellular 
organization seem to have vanished. There evidently are various levels 
of organization, and it is reasonable to suggest that the more complex 
ones have gradually evolved from the simpler because of the presumptive 
advantages that a highly organized system has. 

There are some difficulties with this hypothesis, however. In the break- 
down of visible regulation the organizing capacity itself has not been 
lost, for such abnormal structures as fasciations may revert to normal 
growth again, and in amorphous galls and tissue cultures growing points 
may appear which develop into typical plants. Single cells from a tissue 
culture may produce normal organisms. There is no necessary connection 
between the genetic constitution of the individual (which is what is 
presumably modified by natural selection) and the appearance, or lack 
of it, of a visible state of organization. Furthermore, it must be remem- 
bered that even where gross visible organization has broken down, the 
living cells themselves are small organized systems with a complex 
though often submicroscopic structure and with a very considerable de- 
gree of physiological self -regulation. Indeed, if all organization disappears, 
death ensues. Organization seems to be a fundamental quality of living 
things, explain it in whatever way we can, rather than a simple trait 

454 Morpho genetic Factors 

comparable to those upon which natural selection is effective. One cannot 
discount the possibility, of course, that in the very beginning this regula- 
tory normativeness may have arisen by a selective process and later be- 
came established as a general characteristic of all life. 

An essential aspect of organization and of the organic forms that result 
from it is that these involve much more than a series of successive chemi- 
cal steps, for form is concerned with relationships. In an earlier chapter 
evidence has been presented that relationships, and thus form, are 
genetically determined, but how genes act to produce this trait is far 
from clear. An organized system is a complex pattern of such relation- 
ships and one that is not static but changes during development and 
restores itself if altered. Whatever its origin, the problem still remains as 
to how, in terms of protoplasmic activity, such a self-regulating, pat- 
terned system is produced. 

One of the simplest explanations is that proposed by Child ( 1941 ) and 
others who point out that an organized system does not develop unless 
a polar axis is first established in originally homogeneous material by the 
influence of an environment that differs on the two sides and that a 
gradient arises along this axis. The essential importance, at least at the 
beginning of development, of an asymmetric environment must not be 
forgotten, an environment which, so to speak, "lines up" the undif- 
ferentiated egg or mass of tissue in one direction and so orients it that 
it can then organize its developmental pattern in an orderly fashion. Grav- 
ity and unilateral light are the commonest asymmetric environmental 
factors for plants. They make one side of the structure different from 
the other and thus begin the continuing process of differentiation. Plants 
grown where these factors are uniform on all sides are usually amorphous 
(p. 137). The interaction between the polar gradient and the genes of the 
organism is an important factor in producing a specific form. For some 
biologists this explanation is sufficient to account for organic form, but 
it does not give a clear picture of how such interaction works nor does it 
explain the complex correlations that occur during development. 

Among other suggested explanations one of the most promising is the 
concept of a biological "field" in conformity to which development takes 
place. Gurwitsch ( 1923 ) assumes such a field to be present around a 
developing organic structure, but this is difficult to describe in physical 
terms. In Gurwitsch's hands it acquires an almost mystical character and 
seems to resemble nothing so much as an astral body immanent in and 
around the growing organism. Where such a beautifully precise structure 
as a fungus fruiting body molds itself out of a complex and intertwining 
tangle of sliding hyphae, or where from a throng of individual myx- 
amoebae a specifically fashioned sorocarp develops through their inter- 
action, we must assume that in the mass of living stuff there is a formative 

Organization 455 

factor of some sort, but it should be possible to describe this in more pre- 
cise terms than does Gurwitsch. 

Fields of various sorts have often been postulated in animal develop- 
ment, but in a somewhat more descriptive sense, as the developmental 
influence of a given region over structures in it. If the rudiment of a 
young and growing amphibian tail, for example, is transplanted at an 
early age into the region of a leg it will grow into a leg since it is now in a 
leg field. If it is somewhat older before transplanting it will become a 
tail, since its own tail field is now operative. This conception of a morpho- 
genetic field recognizes the formative influence acting within a given 
region or throughout the embryo but offers little explanation of this 
action. Fields have been discussed by Weiss (1950), Raven (1943), 
and many other experimental embryologists. In botanical morphogenesis 
the field concept has been employed by the Snows and Wardlaw to ac- 
count for the localized development of lateral structures at the apical 

More specific is the suggestion of Burr and others (1932) that the 
morphogenetic field is a bioelectric one. Burr has found that around a de- 
veloping structure, such as a fertilized egg in animals or a developing 
ovary primordium in plants, a micropotentiometer will reveal an orderly 
pattern of potential differences that is a correlate of the form which will 
develop from them. Burr and Northrop (1935) support the view that 
the primary entities in nature are fields and not particles and that the 
former determine the activities of the latter instead of the other way 
around. Both physical and biological phenomena certainly are electrical 
in their ultimate nature, but Burr's theory goes much further than that 
in assuming the organized biological pattern, manifest to our eyes, to 
be the visible expression of an underlying bioelectrical pattern. The ori- 
gin of such a pattern and what determines the changes in it are yet un- 

One of the difficulties in accounting for an organic pattern is to see how 
it can arise in a semifluid and formless protoplasmic system. How, one 
asks, can such a flowing and unstable material as protoplasm produce 
the very specific forms which come out of it? It is obvious that proto- 
plasm, homogeneous though it seems to be, must ha